Intensive Care Med

DOI 10.1007/s00134-017-4683-6

CONFERENCE REPORTS AND EXPERT PANEL

Surviving Sepsis Campaign:

International Guidelines forManagement
ofSepsis andSeptic Shock: 2016
AndrewRhodes1*, LauraE.Evans2, WaleedAlhazzani3, MitchellM.Levy4, MassimoAntonelli5, RicardFerrer6,
AnandKumar7, JonathanE.Sevransky8, CharlesL.Sprung9, MarkE.Nunnally2, BramRochwerg3,
GordonD.Rubenfeld10, DerekC.Angus11, DjillaliAnnane12, RichardJ.Beale13, GeoffreyJ.Bellinghan14,
GordonR.Bernard15, JeanDanielChiche16, CraigCoopersmith8, DanielP.DeBacker17, CraigJ.French18,
SeitaroFujishima19, HerwigGerlach20, JorgeLuisHidalgo21, StevenM.Hollenberg22, AlanE.Jones23,
DilipR.Karnad24, RuthM.Kleinpell25, YounsukKoh26, ThiagoCostaLisboa27, FlaviaR.Machado28,
JohnJ.Marini29, JohnC.Marshall30, JohnE.Mazuski31, LauralynA.McIntyre32, AnthonyS.McLean33,
SangeetaMehta34, RuiP.Moreno35, JohnMyburgh36, PaoloNavalesi37, OsamuNishida38, TiffanyM.Osborn31,
AndersPerner39, ColleenM.Plunkett25, MarcoRanieri40, ChristaA.Schorr22, MaureenA.Seckel41,
ChristopherW.Seymour42, LisaShieh43, KhalidA.Shukri44, StevenQ.Simpson45, MervynSinger46,
B.TaylorThompson47, SeanR.Townsend48, ThomasVan derPoll49, JeanLouisVincent50, W.JoostWiersinga49,
JaniceL.Zimmerman51 andR.PhillipDellinger22

2017 SCCM and ESICM

Abstract
Objective: To provide an update to Surviving Sepsis Campaign Guidelines for Management of Sepsis and Septic
Shock: 2012.
Design: A consensus committee of 55 international experts representing 25 international organizations was con
vened. Nominal groups were assembled at key international meetings (for those committee members attending
the conference). A formal conflict-of-interest (COI) policy was developed at the onset of the process and enforced
throughout. A stand-alone meeting was held for all panel members in December 2015. Teleconferences and
electronic-based discussion among subgroups and among the entire committee served as an integral part of the
development.
Methods: The panel consisted of five sections: hemodynamics, infection, adjunctive therapies, metabolic, and
ventilation. Population, intervention, comparison, and outcomes (PICO) questions were reviewed and updated as
needed, and evidence profiles were generated. Each subgroup generated a list of questions, searched for best avail
able evidence, and then followed the principles of the Grading of Recommendations Assessment, Development, and
Evaluation (GRADE) system to assess the quality of evidence from high to very low, and to formulate recommenda
tions as strong or weak, or best practice statement when applicable.

*Correspondence: andrewrhodes@nhs.net
1
St. Georges Hospital, London, England, UK
Full author information is available at the end of the article

This article is being simultaneously published in Critical Care Medicine

(DOI: 10.1097/CCM.0000000000002255) and Intensive Care Medicine.
 Results: The Surviving Sepsis Guideline panel provided 93 statements on early management and resuscitation of
patients with sepsis or septic shock. Overall, 32 were strong recommendations, 39 were weak recommendations, and
18 were best-practice statements. No recommendation was provided for four questions.
Conclusions: Substantial agreement exists among a large cohort of international experts regarding many strong
recommendations for the best care of patients with sepsis. Although a significant number of aspects of care have rela
tively weak support, evidence-based recommendations regarding the acute management of sepsis and septic shock
are the foundation of improved outcomes for these critically ill patients with high mortality.
Keywords: Evidence-based medicine, Grading of Recommendations Assessment, Development, and Evaluation
criteria, Guidelines, Infection, Sepsis, Sepsis bundles, Sepsis syndrome, Septic shock, Surviving Sepsis Campaign

INTRODUCTION History ofthe guidelines

Sepsis is life-threatening organ dysfunction caused by These clinical practice guidelines are a revision of the
a dysregulated host response to infection [13]. Sepsis 2012 Surviving Sepsis Campaign (SSC) guidelines for the
and septic shock are major healthcare problems, affect- management of severe sepsis and septic shock [8, 9]. The
ing millions of people around the world each year, and initial SSC guidelines were first published in 2004 [10],
killing as many as one in four (and often more) [46]. and revised in 2008 [11, 12] and 2012 [8, 9]. The current
Similar to polytrauma, acute myocardial infarction, or iteration is based on updated literature searches incor-
stroke, early identification and appropriate manage- porated into the evolving manuscript through July 2016.
ment in the initial hours after sepsis develops improves A summary of the 2016 guidelines appears in Appendix
outcomes. 1. A comparison of recommendations from 2012 to 2016
The recommendations in this document are appears in Appendix 2. Unlike previous editions, the
intended to provide guidance for the clinician caring SSC pediatric guidelines will appear in a separate docu-
for adult patients with sepsis or septic shock. Recom- ment, also to be published by the Society of Critical Care
mendations from these guidelines cannot replace the Medicine (SCCM) and the European Society of Intensive
clinicians decision-making capability when presented Care Medicine (ESICM).
with a patients unique set of clinical variables. These
guidelines are appropriate for the sepsis patient in a Sponsorship
hospital setting. These guidelines are intended to be Funding for the development of these guidelines was
best practice (the committee considers this a goal for provided by SCCM and ESICM. In addition, sponsor-
clinical practice) and not created to represent standard ing organizations provided support for their members
of care. involvement.

METHODOLOGY Selection andorganization ofcommittee members

Below is a summary of the important methodologic con- The selection of committee members was based on
siderations for developing these guidelines. expertise in specific aspects of sepsis. Co-chairs were
appointed by the SCCM and ESICM governing bodies.
Definitions Each sponsoring organization appointed a representative
As these guidelines were being developed, new defini- who had sepsis expertise. Additional committee mem-
tions for sepsis and septic shock (Sepsis-3) were pub- bers were appointed by the co-chairs and the SSC Guide-
lished. Sepsis is now defined as life-threatening organ lines Committee Oversight Group to balance continuity
dysfunction caused by a dysregulated host response to and provide new perspectives with the previous commit-
infection. Septic shock is a subset of sepsis with circula- tees membership as well as to address content needs. A
tory and cellular/metabolic dysfunction associated with patient representative was appointed by the co-chairs.
a higher risk of mortality [3]. The Sepsis-3 definition Methodologic expertise was provided by the GRADE
also proposed clinical criteria to operationalize the new Methodology Group.
definitions; however, in the studies used to establish the
evidence for these guidelines, patient populations were Question development
primarily characterized by the previous definition of sep- The scope of this guideline focused on early manage-
sis, severe sepsis, and septic shock stated in the 1991 and ment of patients with sepsis or septic shock. The guide-
2001 consensus documents [7]. line panel was divided into five sections (hemodynamics,
infection, adjunctive therapies, metabolic, and ventila- followed by assessment of the balance between ben-
tion). The group designations were the internal work efit and harm, patients values and preferences, cost and
structure of the guidelines committee. Topic selection resources, and feasibility and acceptability of the inter-
was the responsibility of the co-chairs and group heads, vention. The final recommendations formulated by the
with input from the guideline panel in each group. Pri- guideline panel are based on the assessment of these fac-
oritization of the topics was completed by discussion tors. The GRADE assessment of the quality of evidence is
through e-mails, teleconferences, and face-to-face meet- presented in Table1.
ings. All guideline questions were structured in PICO RCTs begin as high-quality evidence that could be
format, which described the population, intervention, downgraded due to limitations in any of the aforemen-
control, and outcomes. tioned categories. While observational (nonrandomized)
Questions from the last version of the SSC guidelines studies begin as low-quality evidence, the quality level
were reviewed; those that were considered important could be upgraded on the basis of a large magnitude of
and clinically relevant were retained. Questions that were effect or other factors. The GRADE methodology clas-
considered less important or of low priority to clinicians sifies recommendations as strong or weak. The factors
were omitted, and new questions that were considered influencing this determination are presented in Table 2.
high priority were added. The decision regarding ques- The guideline committee assessed whether the desir-
tion inclusion was reached by discussion and consensus able effects of adherence would outweigh the undesirable
among the guideline panel leaders with input from panel effects, and the strength of a recommendation reflects the
members and the methodology team in each group. groups degree of confidence in that balance assessment.
GRADE methodology was applied in selecting only Thus, a strong recommendation in favor of an interven-
outcomes that were considered critical from a patients tion reflects the panels opinion that the desirable effects
perspective [13]. All PICO questions with supporting evi- of adherence to a recommendation will clearly outweigh
dence are presented in Supplemental Digital Content 1 the undesirable effects. A weak recommendation in favor
(ESM 1). of an intervention indicates the judgment that the desir-
able effects of adherence to a recommendation prob-
Search strategy ably will outweigh the undesirable effects, but the panel
With the assistance of professional librarians, an inde- is not confident about these trade-offseither because
pendent literature search was performed for each defined some of the evidence is low quality (and thus uncertainty
question. The panel members worked with group heads, remains regarding the benefits and risks) or the benefits
methodologists, and librarians to identify pertinent and downsides are closely balanced. A strong recommen-
search terms that included, at a minimum, sepsis, severe dation is worded as we recommend and a weak recom-
sepsis, septic shock, sepsis syndrome, and critical illness, mendation as we suggest. An alphanumeric scheme was
combined with appropriate key words specific to the used in previous editions of the SSC guidelines. Table 3
question posed. provides a comparison to the current grading system.
For questions addressed in the 2012 SSC guidelines, The implications of calling a recommendation strong
the search strategy was updated from the date of the are that most patients would accept that intervention
last literature search. For each of the new questions, an and that most clinicians should use it in most situations.
electronic search was conducted of a minimum of two Circumstances may exist in which a strong recommenda-
major databases (e.g., Cochrane Registry, MEDLINE, or tion cannot or should not be followed for an individual
EMBASE) to identify relevant systematic reviews and because of that patients preferences or clinical charac-
randomized clinical trials (RCTs). teristics that make the recommendation less applicable.
These are described in Table 4. A strong recommenda-
Grading ofrecommendations tion does not imply standard of care.
Grading of Recommendations Assessment, Develop- A number of best practice statements (BPSs) appear
ment, and Evaluation (GRADE) system principles guided throughout the document; these statements represent
assessment of quality of evidence from high to very low ungraded strong recommendations and are used under
and were used to determine the strength of recommen- strict criteria. A BPS would be appropriate, for example,
dations (Tables 1, 2) [14]. The GRADE methodology is when the benefit or harm is unequivocal, but the evidence
based on assessment of evidence according to six catego- is hard to summarize or assess using GRADE method-
ries: (1) risk of bias, (2) inconsistency, (3) indirectness, (4) ology. The criteria suggested by the GRADE Working
imprecision, (5) publication bias, and (6) other criteria, Group in Table5 were applied in issuing BPSs [15].
Table1 Determination ofthe quality ofevidence

Underlying methodology

1. High: RCTs
2. Moderate: Downgraded RCTs or upgraded observaonal studies
3. Low: Well-done observaonal studies with RCTs
4. Very Low: Downgraded controlled studies or expert opinion or other evidence
Factors that may decrease the strength of evidence

1. Methodologic features of available RCTs suggesng high likelihood of bias

2. Inconsistency of results, including problems with subgroup analyses

3. Indirectness of evidence (differing populaon, intervenon, control, outcomes,

comparison)

4. Imprecision of results

5. High likelihood of reporng bias

Main factors that may increase the strength of evidence

1. Large magnitude of effect (direct evidence, relave risk > 2 with no plausible
confounders)
2. Very large magnitude of effect with relave risk > 5 and no threats to validity (by
two levels)
3. Dose-response gradient
RCT = randomized clinical trial

Voting process Conflictofinterest policy

Following formulation of statements through discussion No industry input into guidelines development occurred,
in each group and deliberation among all panel members and no industry representatives were present at any of
during face-to-face meetings at which the groups pre- the meetings. No member of the guidelines committee
sented their draft statements, all panel members received received honoraria for any role in the guidelines process.
links to polls created using SurveyMonkey, Inc. (Palo The process relied solely on personal disclosure, and no
Alto, CA) to indicate agreement or disagreement with attempt was made by the group to seek additional confir-
the statement, or abstention. Acceptance of a statement mation. The co-chairs, COI chair, and group heads adju-
required votes from 75% of the panel members with an dicated this to the best of their abilities.
80% agreement threshold. Voters could provide feed- On initial review, 31 financial COI disclosures and five
back for consideration in revising statements that did not nonfinancial disclosures were submitted by committee
receive consensus in up to three rounds of voting. members; others reported no COI. Panelists could have
Table2 Factors determining strong vs. weak recommendation

What Should Be Considered Recommended Process

High or moderate evidence The higher the quality of evidence, the more likely a
strong recommendaon
(Is there high-or moderate-
quality evidence?)

Certainty about the balance of The larger the difference between the desirable and
benefits vs. harms and burdens undesirable consequences and the certainty around that
difference, the more likely a strong recommendaon. The
(Is there certainty?)
smaller the net benefit and the lower the certainty for
that benefit, the more likely a weak recommendaon.

Certainty in, or similar, values The more certainty or similarity in values and preferences,
the more likely a strong recommendaon.
(Is there certainty or similarity?)

Resource implicaons The lower the cost of an intervenon compared to the

alternave and other costs related to the decision (i.e.,
(Are resources worth expected
fewer resources consumed), the more likely a strong
benefits?)
recommendaon.

Table3 Comparison of2016 grading terminology withprevious alphanumeric descriptors

2016 Descriptor 2012 Descriptor

Strength Strong 1

Weak 2

Quality High A

Moderate B

Low C

Very Low D

Ungraded strong recommendaon Best Pracce Statement Ungraded

Table4 Implications ofthe strength ofrecommendation

Strong Recommendaon Weak Recommendaon

Most individuals in this The majority of individuals in

For paents situaon would want the this situaon would want the
recommended course of suggested course of acon,
acon, and only a small but many would not.
proporon would not.

Most individuals should Different choices are likely to

receive the recommended be appropriate for different
course of acon. Adherence to paents, and therapy should
this recommendaon be tailored to the individual
according to the guideline paents circumstances. These
For clinicians could be used as a quality circumstances may include the
criterion or performance paents or familys values and
indicator. Formal decision aids preferences.
are not likely to be needed to
help individuals make
decisions consistent with their
values and preferences.

The recommendaon can be Policy-making will require

adapted as policy in most substanal debates and
situaons, including for use as involvement of many
performance indicators. stakeholders. Policies are also
For policy makers more likely to vary between
regions. Performance
indicators would have to focus
on the fact that adequate
deliberaon about the
management opons has
taken place.

both financial and nonfinancial COI. Declared COI dis- as the described restrictions on voting on recommen-
closures from 11 members were determined by the COI dations in areas of potential COI. One individual was
subcommittee to be not relevant to the guidelines con- asked to step down from the committee. All panelists
tent process. Fifteen who were determined to have COI with COI were required to work within their group with
(financial and nonfinancial) were adjudicated by a man- full disclosure when a topic for which they had relevant
agement plan that required adherence to SSC COI policy COI was discussed, and they were not allowed to serve
limiting discussion or voting at any committee meetings as group head. At the time of final approval of the docu-
during which content germane to their COI was dis- ment, an update of the COI statement was required. No
cussed. Five were judged as having conflicts that were additional COI issues were reported that required further
managed through reassignment to another group as well adjudication.
Table5 Criteria forBest practice statements

Criteria for Best Pracce Statements

1 Is the statement clear and aconable?

2 Is the message necessary?

3 Is the net benefit (or harm) unequivocal?

4 Is the evidence difficult to collect and summarize?

5 Is the raonale explicit?

6 Is this be
er to be formally GRADEd?

GRADE = Grading of Recommendaons Assessment, Development, and Evaluaon

Modified from Guya
et al (15).

A summary of all statements determined by the septic shock requiring vasopressors (strong rec-
guidelines panel appears in Appendix 1. All evidence ommendation, moderate quality of evidence).
summaries and evidence profiles that informed the rec- 7. We suggest guiding resuscitation to normalize
ommendations and statements appear in ESM 2. Links to lactate in patients with elevated lactate levels as
specific tables and figures appear within the relevant text. a marker of tissue hypoperfusion (weak recom-
mendation, low quality of evidence).
A. INITIAL RESUSCITATION
Rationale Early effective fluid resuscitation is cru-
1. Sepsis and septic shock are medical emergencies, cial for stabilization of sepsis-induced tissue hypoper-
and we recommend that treatment and resuscita- fusion or septic shock. Sepsis-induced hypoperfusion
tion begin immediately (BPS). may be manifested by acute organ dysfunction and/
2. We recommend that, in the resuscitation from or decreased blood pressure and increased serum
sepsis-induced hypoperfusion, at least 30 mL/kg lactate. Previous iterations of these guidelines have
of IV crystalloid fluid be given within the first 3h recommended a protocolized quantitative resuscita-
(strong recommendation, low quality of evidence). tion, otherwise known as early goal-directed therapy
3. We recommend that, following initial fluid resus- (EGDT), which was based on the protocol published by
citation, additional fluids be guided by frequent Rivers [16]. This recommendation described the use of
reassessment of hemodynamic status (BPS). a series of goals that included central venous pressure
Remarks Reassessment should include a thorough (CVP) and central venous oxygen saturation (Scvo2). This
clinical examination and evaluation of available phys- approach has now been challenged following the failure
iologic variables (heart rate, blood pressure, arterial to show a mortality reduction in three subsequent large
oxygen saturation, respiratory rate, temperature, multicenter RCTs [1719]. No harm was associated with
urine output, and others, as available) as well as other the interventional strategies; thus, the use of the previous
noninvasive or invasive monitoring, as available. targets is still safe and may be considered. Of note, the
4. We recommend further hemodynamic assess- more recent trials included less severely ill patients (lower
ment (such as assessing cardiac function) to baseline lactate levels, Scvo2 at or above the target value
determine the type of shock if the clinical exami- on admission, and lower mortality in the control group).
nation does not lead to a clear diagnosis (BPS). Although this protocol cannot now be recommended
5. We suggest that dynamic over static variables be from its evidence base, bedside clinicians still need guid-
used to predict fluid responsiveness, where avail- ance as to how to approach this group of patients who
able (weak recommendation, low quality of evi- have significant mortality and morbidity. We recom-
dence). mend, therefore, that these patients be viewed as having
6. We recommend an initial target mean arterial a medical emergency that necessitates urgent assessment
pressure (MAP) of 65 mm Hg in patients with and treatment. As part of this, we recommend that initial
fluid resuscitation begin with 30 mL/kg of crystalloid applicability of the parameters to all situations may influ-
within the first 3 h. This fixed volume of fluid enables ence the routine use of dynamic indices [22, 25].
clinicians to initiate resuscitation while obtaining more MAP is the driving pressure of tissue perfusion. While
specific information about the patient and while await- perfusion of critical organs such as the brain or kidney may
ing more precise measurements of hemodynamic status. be protected from systemic hypotension by autoregulation
Although little literature includes controlled data to sup- of regional perfusion, below a threshold MAP, tissue per-
port this volume of fluid, recent interventional studies fusion becomes linearly dependent on arterial pressure. In
have described this as usual practice in the early stages a single-center trial [26], dose titration of norepinephrine
of resuscitation, and observational evidence supports the from 65 to 75 and 85 mm Hg raised cardiac index (from
practice [20, 21]. The average volume of fluid pre-rand- 4.70.5 to 5.50.6 L/min/m2) but did not change urinary
omization given in the PROCESS and ARISE trials was flow, arterial lactate levels, oxygen delivery and consump-
approximately 30 mL/kg, and approximately 2 L in the tion, gastric mucosal Pco2, RBC velocity, or skin capillary
PROMISE trial [1719]. Many patients will require more flow. Another single-center [27] trial compared, in norep-
fluid than this, and for this group we advocate that fur- inephrine-treated septic shock, dose titration to maintain
ther fluid be given in accordance with functional hemo- MAP at 65 mm Hg versus achieving 85 mm Hg. In this
dynamic measurements. trial, targeting high MAP increased cardiac index from
One of the most important principles to understand in 4.8 (3.86.0) to 5.8 (4.36.9)L/min/m2 but did not change
the management of these complex patients is the need renal function, arterial lactate levels, or oxygen consump-
for a detailed initial assessment and ongoing reevalua- tion. A third single-center trial [28] found improved micro-
tion of the response to treatment. This evaluation should circulation, as assessed by sublingual vessel density and the
start with a thorough clinical examination and evalua- ascending slope of thenar oxygen saturation after an occlu-
tion of available physiologic variables that can describe sion test, by titrating norepinephrine to a MAP of 85mm
the patients clinical state (heart rate, blood pressure, Hg compared to 65mm Hg. Only one multicenter trial that
arterial oxygen saturation, respiratory rate, temperature, compared norepinephrine dose titration to achieve a MAP
urine output, and others as available). Echocardiography of 65mm Hg versus 85mm Hg had mortality as a primary
in recent years has become available to many bedside outcome [29]. There was no significant difference in mor-
clinicians and enables a more detailed assessment of the tality at 28days (36.6% in the high-target group and 34.0%
causes of the hemodynamic issues [22]. in the low-target group) or 90days (43.8% in the high-tar-
The use of CVP alone to guide fluid resuscitation can get group and 42.3% in the low-target group). Targeting a
no longer be justified [22] because the ability to predict MAP of 85 mm Hg resulted in a significantly higher risk
a response to a fluid challenge when the CVP is within a of arrhythmias, but the subgroup of patients with previ-
relatively normal range (812mm Hg) is limited [23]. The ously diagnosed chronic hypertension had a reduced need
same holds true for other static measurements of right for renal replacement therapy (RRT) at this higher MAP. A
or left heart pressures or volumes. Dynamic measures of recent pilot trial of 118 septic shock patients [30] suggested
assessing whether a patient requires additional fluid have that, in the subgroup of patients older than 75years, mor-
been proposed in an effort to improve fluid management tality was reduced when targeting a MAP of 6065 versus
and have demonstrated better diagnostic accuracy at pre- 7580mm Hg. The quality of evidence was moderate (ESM
dicting those patients who are likely to respond to a fluid 4) due to imprecise estimates (wide confidence intervals).
challenge by increasing stroke volume. These techniques As a result, the desirable consequences of targeting MAP
encompass passive leg raises, fluid challenges against of 65 mm Hg (lower risk of atrial fibrillation, lower doses
stroke volume measurements, or the variations in systolic of vasopressors, and similar mortality) led to a strong rec-
pressure, pulse pressure, or stroke volume to changes in ommendation favoring an initial MAP target of 65mm Hg
intrathoracic pressure induced by mechanical ventilation over higher MAP targets. When a better understanding of
[24]. Our review of five studies of the use of pulse pres- any patients condition is obtained, this target should be
sure variation to predict fluid responsiveness in patients individualized to the pertaining circumstances.
with sepsis or septic shock demonstrated a sensitivity of Serum lactate is not a direct measure of tissue perfu-
0.72 (95% CI 0.610.81) and a specificity of 0.91 (95% CI sion [31]. Increases in the serum lactate level may repre-
0.830.95); the quality of evidence was low due to impre- sent tissue hypoxia, accelerated aerobic glycolysis driven
cision and risk of bias (ESM 3) [24]. A recent multicenter by excess beta-adrenergic stimulation, or other causes
study demonstrated limited use of cardiac function moni- (e.g., liver failure). Regardless of the source, increased
tors during fluid administration in the ICUs. Even though lactate levels are associated with worse outcomes [32].
data on the use of these monitors in the emergency Because lactate is a standard laboratory test with pre-
department are lacking, the availability of the devices and scribed techniques for its measurement, it may serve as
a more objective surrogate for tissue perfusion as com- Improvement [46]. The SSC bundles that are based on
pared with physical examination or urine output. Five previous guidelines have been adopted by the U.S.-based
randomized controlled trials (647 patients) have evalu- National Quality Forum and have also been adapted by
ated lactate-guided resuscitation of patients with septic the U.S. healthcare systems regulatory agencies for pub-
shock [3337]. A significant reduction in mortality was lic reporting. To align with emerging evidence and U.S.
seen in lactate-guided resuscitation compared to resusci- national efforts, the SSC bundles were revised in 2015.
tation without lactate monitoring (RR 0.67; 95% CI 0.53 While specifics vary widely among different programs, a
0.84; low quality). There was no evidence for difference common theme is the drive toward improvement in com-
in ICU length of stay (LOS) (mean difference 1.51days; pliance with sepsis bundles and practice guidelines such
95% CI 3.65 to 0.62; low quality). Two other meta-anal- as SSC [8]. A meta-analysis of 50 observational studies
yses of the 647 patients who were enrolled in these trials demonstrated that performance improvement programs
demonstrate moderate evidence for reduction in mor- were associated with a significant increase in compliance
tality when an early lactate clearance strategy was used, with the SSC bundles and a reduction in mortality (OR
compared with either usual care (nonspecified) or with a 0.66; 95% CI 0.610.72) [47]. The largest study to date
Scvo2 normalization strategy [38, 39]. examined the relationship between compliance with the
SSC bundles (based on the 2004 guidelines) and mortal-
B. SCREENING FOR SEPSIS AND PERFORMANCE ity. A total of 29,470 patients in 218 hospitals in the United
IMPROVEMENT States, Europe, and South America were examined over
a 7.5-year period [21]. Lower mortality was observed in
1. We recommend that hospitals and hospital sys- hospitals with higher compliance. Overall hospital mor-
tems have a performance improvement program tality decreased 0.7% for every 3 months a hospital par-
for sepsis, including sepsis screening for acutely ticipated in the SSC, associated with a 4% decreased LOS
ill, high-risk patients (BPS). for every 10% improvement in compliance with bundles.
This benefit has also been shown across a wide geographic
Rationale Performance improvement efforts for sep- spectrum. A study of 1794 patients from 62 countries with
sis are associated with improved patient outcomes [40]. severe sepsis (now termed sepsis after the Sepsis-3 defi-
Sepsis performance improvement programs should opti- nition [1] or septic shock demonstrated a 3640% reduc-
mally have multiprofessional representation (physicians, tion of the odds of dying in the hospital with compliance
nurses, affiliate providers, pharmacists, respiratory thera- with either the 3- or 6-h SSC bundles [48]. This recom-
pists, dietitians, administrators) with stakeholders from mendation met the prespecified criteria for a BPS. The
all key disciplines represented in their development and specifics of performance improvement methods varied
implementation. Successful programs should include pro- markedly between studies; thus, no single approach to per-
tocol development and implementation, targeted metrics formance improvement could be recommended (ESM 5).
to be evaluated, data collection, and ongoing feedback
to facilitate continuous performance improvement [41]. C. DIAGNOSIS
In addition to traditional continuing education efforts
to introduce guidelines into clinical practice, knowledge 1. We recommend that appropriate routine micro-
translation efforts can be valuable in promoting the use of biologic cultures (including blood) be obtained
high-quality evidence in changing behavior [42]. before starting antimicrobial therapy in patients
Sepsis performance improvement programs can be with suspected sepsis or septic shock if doing so
aimed at earlier recognition of sepsis via a formal screen- results in no substantial delay in the start of anti-
ing effort and improved management of patients once microbials (BPS).
they are identified as being septic. Because lack of recog- Remarks Appropriate routine microbiologic cultures
nition prevents timely therapy, sepsis screening is associ- always include at least two sets of blood cultures
ated with earlier treatment [43, 44]. (aerobic and anaerobic).
Notably, sepsis screening has been associated with
decreased mortality in several studies [20, 45]. The imple- Rationale Sterilization of cultures can occur within
mentation of a core set of recommendations (bundle) has minutes to hours after the first dose of an appropriate
been a cornerstone of sepsis performance improvement antimicrobial [49, 50]. Obtaining cultures prior to the
programs aimed at improving management [46]. Note administration of antimicrobials significantly increases
that the SSC bundles have been developed separately from the yield of cultures, making identification of a pathogen
the guidelines in conjunction with an educational and more likely. Isolation of an infecting organism(s) allows
improvement partnership with the Institute for Healthcare for de-escalation of antimicrobial therapy first at the
point of identification and then again when susceptibili- simultaneous peripheral blood cultures). This is done
ties are obtained. De-escalation of antimicrobial therapy to assist in the diagnosis of a potential catheter-related
is a mainstay of antibiotic stewardship programs and is bloodstream infection. Data are inconsistent regarding
associated with less resistant microorganisms, fewer side the utility of differential time to blood culture positivity
effects, and lower costs [51]. Several retrospective stud- (i.e., equivalent volume blood culture from the vascular
ies have suggested that obtaining cultures prior to anti- access device positive more than 2 h before the periph-
microbial therapy is associated with improved outcome eral blood culture) in suggesting that the vascular access
[52, 53]. Similarly, de-escalation has also been associated device is the source of the infection [6365].
with improved survival in several observational studies It is important to note that drawing blood cultures from
[54, 55]. The desire to obtain cultures prior to initiating an intravascular catheter in case of possible infection of
antimicrobial therapy must be balanced against the mor- the device does not eliminate the option of removing the
tality risk of delaying a key therapy in critically ill patients catheter (particular nontunneled catheters) immediately
with suspected sepsis or septic shock who are at signifi- afterward.
cant risk of death [56, 57]. In patients without a suspicion of catheter-associated
We recommend that blood cultures be obtained prior to infection and in whom another clinical infection site is
initiating antimicrobial therapy if cultures can be obtained suspected, at least one blood culture (of the two or more
in a timely manner. However, the risk/benefit ratio favors that are required) should be obtained peripherally. How-
rapid administration of antimicrobials if it is not logisti- ever, no recommendation can be made as to where addi-
cally possible to obtain cultures promptly. Therefore, in tional blood cultures should be drawn. Options include:
patients with suspected sepsis or septic shock, appropri- (a) all cultures drawn peripherally via venipuncture,
ate routine microbiologic cultures should be obtained (b) cultures drawn through each separate intravascu-
before initiation of antimicrobial therapy from all sites lar device but not through multiple lumens of the same
considered to be potential sources of infection if it results intravascular catheter, or (c) cultures drawn through
in no substantial delay in the start of antimicrobials. This multiple lumens in an intravascular device [6670].
may include blood, cerebrospinal fluid, urine, wounds, In the near future, molecular diagnostic methods may offer
respiratory secretions, and other body fluids, but does not the potential to diagnose infections more quickly and more
normally include samples that require an invasive proce- accurately than current techniques. However, varying tech-
dure such as bronchoscopy or open surgery. The decision nologies have been described, clinical experience remains
regarding which sites to culture requires careful consid- limited, and additional validation is needed before recom-
eration from the treatment team. Pan culture of all sites mending these methods as an adjunct to or replacement for
that could potentially be cultured should be discouraged standard blood culture techniques [7173]. In addition, sus-
(unless the source of sepsis is not clinically apparent), ceptibility testing is likely to require isolation and direct test-
because this practice can lead to inappropriate antimi- ing of viable pathogens for the foreseeable future.
crobial use [58]. If history or clinical examination clearly
indicates a specific anatomic site of infection, cultures of D. ANTIMICROBIAL THERAPY
other sites (apart from blood) are generally unnecessary.
We suggest 45min as an example of what may be consid- 1. We recommend that administration of IV antimi-
ered to be no substantial delay in the initiation of antimi- crobials be initiated as soon as possible after rec-
crobial therapy while cultures are being obtained. ognition and within 1h for both sepsis and septic
Two or more sets (aerobic and anaerobic) of blood cul- shock (strong recommendation, moderate quality
tures are recommended before initiation of any new anti- of evidence; grade applies to both conditions).
microbial in all patients with suspected sepsis [59]. All
necessary blood cultures may be drawn together on the Rationale The rapidity of administration is central to
same occasion. Blood culture yield has not been shown the beneficial effect of appropriate antimicrobials. In the
to be improved with sequential draws or timing to tem- presence of sepsis or septic shock, each hour delay in
perature spikes [60, 61]. Details on appropriate methods administration of appropriate antimicrobials is associated
to draw and transport blood culture samples are enumer- with a measurable increase in mortality [57, 74]. Further,
ated in other guidelines [61, 62]. several studies show an adverse effect on secondary end
In potentially septic patients with an intravascular points (e.g., LOS [75], acute kidney injury [76], acute lung
catheter (in place >48h) in whom a site of infection is not injury [77], and organ injury assessed by Sepsis-Related
clinically apparent or a suspicion of intravascular cathe- Organ Assessment score [78] with increasing delays.
ter-associated infection exists, at least one blood culture Despite a meta-analysis of mostly poor-quality studies
set should be obtained from the catheter (along with that failed to demonstrate a benefit of rapid antimicrobial
therapy, the largest and highest-quality studies support prompt administration. Many antimicrobials will not
giving appropriate antimicrobials as soon as possible in remain stable if premixed in a solution. This issue must
patients with sepsis with or without septic shock [57, 74, be taken into consideration in institutions that rely on
7981]. The majority of studies within the meta-analy- premixed solutions for rapid antimicrobial availability. In
sis were of low quality due to a number of deficiencies, choosing the antimicrobial regimen, clinicians should be
including small study size, using an initial index time aware that some antimicrobial agents (notably -lactams)
of an arbitrary time point such as emergency depart- have the advantage of being able to be safely adminis-
ment arrival, and indexing of outcome to delay in time tered as a bolus or rapid infusion, while others require a
to the first antimicrobial (regardless of activity against lengthy infusion. If vascular access is limited and many
the putative pathogen) [82, 83]. Other negative studies different agents must be infused, drugs that can be
not included in this meta-analysis are compromised by administered as a bolus or rapid infusion may offer an
equating bacteremia with sepsis (as currently defined to advantage for rapid achievement of therapeutic levels for
include organ failure) and septic shock [8487]. Many of the initial dose.
these studies are also compromised by indexing delays While establishing vascular access and initiating
to easily accessible but nonphysiologic variables such as aggressive fluid resuscitation are very important when
time of initial blood culture draw (an event likely to be managing patients with sepsis or septic shock, prompt
highly variable in timing occurrence). IV infusion of antimicrobial agents is also a priority. This
While available data suggest that the earliest possible may require additional vascular access ports. Intraosse-
administration of appropriate IV antimicrobials follow- ous access, which can be quickly and reliably established
ing recognition of sepsis or septic shock yields optimal (even in adults), can be used to rapidly administer the
outcomes, 1h is recommended as a reasonable minimal initial doses of any antimicrobial [90, 91]. In addition,
target. The feasibility of achieving this target consistently, intramuscular preparations are approved and available
however, has not been adequately assessed. Practical for several first-line -lactams, including imipenem/
considerations, for example, challenges with clinicians cilastatin, cefepime, ceftriaxone, and ertapenem. Sev-
early identification of patients or operational complexi- eral additional first-line -lactams can also be effectively
ties in the drug delivery chain, represent poorly studied administered intramuscularly in emergency situations if
variables that may affect achieving this goal. A number vascular and intraosseous access is unavailable, although
of patient and organizational factors appear to influence regulatory approval for intramuscular administration for
antimicrobial delays [88]. these drugs is lacking [9294]. Intramuscular absorption
Accelerating appropriate antimicrobial delivery insti- and distribution of some of these agents in severe ill-
tutionally starts with an assessment of causes of delays ness has not been studied; intramuscular administration
[89]. These can include an unacceptably high frequency should be considered only if timely establishment of vas-
of failure to recognize the potential existence of sepsis or cular access is not possible.
septic shock and of inappropriate empiric antimicrobial
initiation (e.g., as a consequence of lack of appreciation of 2. We recommend empiric broad-spectrum therapy
the potential for microbial resistance or recent previous with one or more antimicrobials for patients pre-
antimicrobial use in a given patient). In addition, unrec- senting with sepsis or septic shock to cover all
ognized or underappreciated administrative or logistic likely pathogens (including bacterial and poten-
factors (often easily remedied) may be found. Possible tially fungal or viral coverage) (strong recom-
solutions to delays in antimicrobial initiation include mendation, moderate quality of evidence).
use of stat orders or including a minimal time element 3. We recommend that empiric antimicrobial ther-
in antimicrobial orders, addressing delays in obtaining apy be narrowed once pathogen identification
blood and site cultures pending antimicrobial adminis- and sensitivities are established and/or adequate
tration, and sequencing antimicrobial delivery optimally clinical improvement is noted (BPS).
or using simultaneous delivery of key antimicrobials, as
well as improving supply chain deficiencies. Improving Rationale The initiation of appropriate antimicrobial
communication among medical, pharmacy, and nursing therapy (i.e., with activity against the causative patho-
staff can also be highly beneficial. gen or pathogens) is one of the most important facets of
Most issues can be addressed by quality improvement effective management of life-threatening infections caus-
initiatives, including defined order sets. If antimicrobial ing sepsis and septic shock. Failure to initiate appropriate
agents cannot be mixed and delivered promptly from the empiric therapy in patients with sepsis and septic shock
pharmacy, establishing a supply of premixed drugs for is associated with a substantial increase in morbidity
urgent situations is an appropriate strategy for ensuring and mortality [79, 9597]. In addition, the probability
of progression from gram-negative bacteremic infection antimicrobial regimens in patients with sepsis and sep-
to septic shock is increased [98]. Accordingly, the initial tic shock is complex and cannot be reduced to a simple
selection of antimicrobial therapy must be broad enough table. Several factors must be assessed and used in deter-
to cover all likely pathogens. The choice of empiric anti- mining the appropriate antimicrobial regimen at each
microbial therapy depends on complex issues related to medical center and for each patient. These include:
the patients history, clinical status, and local epidemio-
logic factors. Key patient factors include the nature of the (a) The anatomic site of infection with respect to the
clinical syndrome/site of infection, concomitant underly- typical pathogen profile and to the properties of indi-
ing diseases, chronic organ failures, medications, indwell- vidual antimicrobials to penetrate that site.
ing devices, the presence of immunosuppression or other (b) Prevalent pathogens within the community, hospital,
form of immunocompromise, recent known infection or and even hospital ward.
colonization with specific pathogens, and the receipt of (c) The resistance patterns of those prevalent pathogens.
antimicrobials within the previous three months. In addi- (d) The presence of specific immune defects such as
tion, the patients location at the time of infection acqui- neutropenia, splenectomy, poorly controlled HIV
sition (i.e., community, chronic care institution, acute infection and acquired or congenital defects of
care hospital), local pathogen prevalence, and the suscep- immunoglobulin, complement or leukocyte function
tibility patterns of those common local pathogens in both or production.
the community and hospital must be factored into the (e) Age and patient comorbidities including chronic ill-
choice of therapy. Potential drug intolerances and toxicity ness (e.g., diabetes) and chronic organ dysfunction
must also be considered. (e.g., liver or renal failure), the presence of invasive
The most common pathogens that cause septic shock devices (e.g., central venous lines or urinary catheter)
are gram-negative bacteria, gram-positive, and mixed that compromise the defense to infection.
bacterial microorganisms. Invasive candidiasis, toxic
shock syndromes, and an array of uncommon pathogens In addition, the clinician must assess risk factors for
should be considered in selected patients. Certain spe- infection with multidrug-resistant pathogens including
cific conditions put patients at risk for atypical or resist- prolonged hospital/chronic facility stay, recent antimi-
ant pathogens. For example, neutropenic patients are at crobial use, prior hospitalization, and prior colonization
risk for an especially wide range of potential pathogens, or infection with multidrug-resistant organisms. The
including resistant gram-negative bacilli and Candida occurrence of more severe illness (e.g., septic shock) may
species. Patients with nosocomial acquisition of infec- be intrinsically associated with a higher probability of
tion are prone to sepsis with methicillin-resistant Staph- resistant isolates due to selection in failure to respond to
ylococcus aureus (MRSA) and vancomycin-resistant earlier antimicrobials.
Enterococci. Given the range of variables that must be assessed, the
Historically, critically ill patients with overwhelming recommendation of any specific regimen for sepsis and
infection have not been considered a unique subgroup septic shock is not possible. The reader is directed to
comparable to neutropenic patients for purposes of guidelines that provide potential regimens based on ana-
selection of antimicrobial therapy. Nonetheless, critically tomic site of infection or specific immune defects [67,
ill patients with severe and septic shock are, like neutro- 99109].
penic patients, characterized by distinct differences from However, general suggestions can be provided. Since
the typical infected patient that impact on the optimal the vast majority of patients with severe sepsis and septic
antimicrobial management strategy. Primary among shock have one or more forms of immunocompromise,
these differences are a predisposition to infection with the initial empiric regimen should be broad enough to
resistant organisms and a marked increase in frequency cover most pathogens isolated in healthcare-associated
of death and other adverse outcomes if there is a failure infections. Most often, a broad-spectrum carbapenem
of rapid initiation of effective antimicrobial therapy. (e.g., meropenem, imipenem/cilastatin or doripenem)
Selection of an optimal empiric antimicrobial regimen or extended-range penicillin/-lactamase inhibitor com-
in sepsis and septic shock is one of the central deter- bination (e.g., piperacillin/tazobactam or ticarcillin/
minants of outcome. Survival may decrease as much as clavulanate) is used. However, several third- or higher-
fivefold for septic shock treated with an empiric regimen generation cephalosporins can also be used, especially
that fails to cover the offending pathogen [95]. Because as part of a multidrug regimen. Of course, the specific
of the high mortality associated with inappropriate ini- regimen can and should be modified by the anatomic site
tial therapy, empiric regimens should err on the side of infection if it is apparent and by knowledge of local
of over-inclusiveness. However, the choice of empiric microbiologic flora.
 Multidrug therapy is often required to ensure a suf- of such tests is not high enough to justify dependence on
ficiently broad spectrum of empiric coverage initially. these tests for primary decision-making.
Clinicians should be cognizant of the risk of resistance Superior empiric coverage can be obtained using local
to broad-spectrum -lactams and carbapenems among and unit-specific antibiograms [113, 114] or an infectious
gram-negative bacilli in some communities and health- diseases consultation [115117]. Where uncertainty
care settings. The addition of a supplemental gram-neg- regarding appropriate patient-specific antimicrobial ther-
ative agent to the empiric regimen is recommended for apy exists, infectious diseases consultation is warranted.
critically ill septic patients at high risk of infection with Early involvement of infectious diseases specialists can
such multidrug-resistant pathogens (e.g., Pseudomonas, improve outcome in some circumstances (e.g., S. aureus
Acinetobacter, etc.) to increase the probability of at least bacteremia) [113115].
one active agent being administered [110]. Similarly, Although restriction of antimicrobials is an impor-
insituations of a more-than-trivial risk for other resistant tant strategy to reduce both the development of patho-
or atypical pathogens, the addition of a pathogen-specific gen resistance and cost, it is not an appropriate strategy
agent to broaden coverage is warranted. Vancomycin, in the initial therapy for this patient population. Patients
teicoplanin, or another anti-MRSA agent can be used with sepsis or septic shock generally warrant empiric
when risk factors for MRSA exist. A significant risk of broad-spectrum therapy until the causative organism
infection with Legionella species mandates the addition and its antimicrobial susceptibilities are defined. At that
of a macrolide or fluoroquinolone. point, the spectrum of coverage should be narrowed
Clinicians should also consider whether Candida spe- by eliminating unneeded antimicrobials and replacing
cies are likely pathogens when choosing initial therapy. broad-spectrum agents with more specific agents [118].
Risk factors for invasive Candida infections include However, if relevant cultures are negative, empiric nar-
immunocompromised status (neutropenia, chemo- rowing of coverage based on a good clinical response
therapy, transplant, diabetes mellitus, chronic liver is appropriate. Collaboration with antimicrobial stew-
failure, chronic renal failure), prolonged invasive vas- ardship programs is encouraged to ensure appropriate
cular devices (hemodialysis catheters, central venous choices and rapid availability of effective antimicrobials
catheters), total parenteral nutrition, necrotizing pan- for treating septic patients.
creatitis, recent major surgery (particularly abdominal), In situations in which a pathogen is identified, de-
prolonged administration of broad-spectrum antibiotics, escalation to the narrowest effective agent should be
prolonged hospital/ICU admission, recent fungal infec- implemented for most serious infections. However, approx-
tion, and multisite colonization [111, 112]. If the risk of imately one-third of patients with sepsis do not have a
Candida sepsis is sufficient to justify empiric antifun- causative pathogen identified [95, 119]. In some cases, this
gal therapy, the selection of the specific agent should may be because guidelines do not recommend obtaining
be tailored to the severity of illness, the local pattern cultures (e.g., community-acquired abdominal sepsis with
of the most prevalent Candida species, and any recent bowel perforation) [108]. In others, cultures may have fol-
exposure to antifungal drugs. Empiric use of an echino- lowed antimicrobial therapy. Further, almost half of patients
candin (anidulafungin, micafungin, or caspofungin) is with suspected sepsis in one study have been adjudicated in
preferred in most patients with severe illness, especially post hoc analysis to lack infection or represent only pos-
in those patients with septic shock, who have recently sible sepsis [120]. Given the adverse societal and individ-
been treated with other antifungal agents, or if Candida ual risks to continued unnecessary antimicrobial therapy,
glabrata or Candida krusei infection is suspected from we recommend thoughtful de-escalation of antimicrobials
earlier culture data [100, 105]. Triazoles are acceptable based on adequate clinical improvement even if cultures
in hemodynamically stable, less ill patients who have not are negative. When infection is found not to be present,
had previous triazole exposure and are not known to be antimicrobial therapy should be stopped promptly to mini-
colonized with azole-resistant species. Liposomal formu- mize the likelihood that the patient will become infected
lations of amphotericin B are a reasonable alternative to with an antimicrobial-resistant pathogen or develop a
echinocandins in patients with echinocandin intolerance drug-related adverse effect. Thus, the decisions to continue,
or toxicity [100, 105]. Knowledge of local resistance pat- narrow, or stop antimicrobial therapy must be made on the
terns to antifungal agents should guide drug selection basis of clinician judgment and clinical information.
until fungal susceptibility test results, if available, are
received. Rapid diagnostic testing using -d-glucan or 4. We recommend against sustained systemic antimi-
rapid polymerase chain reaction assays to minimize inap- crobial prophylaxis in patients with severe inflam-
propriate anti-Candida therapy may have an evolving matory states of noninfectious origin (e.g., severe
supportive role. However, the negative predictive value pancreatitis, burn injury) (BPS).
 Rationale A systemic inflammatory response without with resistant organisms. Perhaps most importantly
infection does not mandate antimicrobial therapy. Exam- with respect to initial empiric antimicrobial dosing is an
ples of conditions that may exhibit acute inflammatory increased volume of distribution for most antimicrobials,
signs without infection include severe pancreatitis and in part due to the rapid expansion of extracellular volume
extensive burn injury. Sustained systemic antimicrobial as a consequence of aggressive fluid resuscitation. This
therapy in the absence of suspected infection should be results in an unexpectedly high frequency of suboptimal
avoided in these situations to minimize the likelihood drug levels with a variety of antimicrobials in patients
that the patient will become infected with an antimicro- with sepsis and septic shock [125128]. Early attention
bial-resistant pathogen or will develop a drug-related to appropriate antimicrobial dosing is central to improv-
adverse effect. ing outcome given the marked increase in mortality and
Although the prophylactic use of systemic antimicro- other adverse outcomes if there is a failure of rapid initia-
bials for severe necrotizing pancreatitis has been rec- tion of effective therapy. Antimicrobial therapy in these
ommended in the past, recent guidelines have favored patients should always be initiated with a full, high end-
avoidance of this approach [121]. loading dose of each agent used.
The current position is supported by meta-analyses that Different antimicrobials have different required plasma
demonstrate no clinical advantage of prophylactic antibi- targets for optimal outcomes. Failure to achieve peak
otics that would outweigh their long-term adverse effects plasma targets on initial dosing has been associated with
[122]. Similarly, prolonged systemic antimicrobial proph- clinical failure with aminoglycosides [129]. Similarly,
ylaxis has been used in the past for patients with severe inadequate early vancomycin trough plasma concen-
burns. However, recent meta-analyses suggest questiona- trations (in relation to pathogen minimum inhibitory
ble clinical benefit with this approach [123, 124]. Current concentration [MIC]) have been associated with clini-
guidelines for burn management do not support sus- cal failure for serious MRSA infections [130] (including
tained antimicrobial prophylaxis [101]. Summarizing the nosocomial pneumonia [131] and septic shock [132].
evidence is challenging due to the diversity of the popu- The clinical success rate for treatment of serious infec-
lation. The quality of evidence was low for mortality in tions correlates with higher peak blood levels (in rela-
pancreatitis [122] and low for burns; therefore, we believe tion to pathogen MIC) of fluoroquinolones (nosocomial
this recommendation is better addressed as a BPS, in pneumonia and other serious infections) [133135] and
which the alternative of administering antibiotics without aminoglycosides (gram-negative bacteremia, nosoco-
indicators of infection is implausible [122124]. Despite mial pneumonia, and other serious infections) [129, 136].
our recommendation against sustained systemic antimi- For -lactams, superior clinical and microbiologic cures
crobial prophylaxis generally, brief antibiotic prophylaxis appear to be associated with a longer duration of plasma
for specific invasive procedures may be appropriate. In concentration above the pathogen MIC, particularly in
addition, if there is a strong suspicion of concurrent sep- critically ill patients [137140].
sis or septic shock in patients with a severe inflammatory The optimal dosing strategy for aminoglycosides and
state of noninfectious origin (despite overlapping clinical fluoroquinolones involves optimizing peak drug plasma
presentations), antimicrobial therapy is indicated. concentrations. For aminoglycosides, this can most eas-
ily be attained with once daily dosing (57 mg/kg daily
5. We recommend that dosing strategies of antimi- gentamicin equivalent). Once-daily dosing yields at least
crobials be optimized based on accepted phar- comparable clinical efficacy with possibly decreased renal
macokinetic/pharmacodynamic principles and toxicity compared to multiple daily dosing regimens
specific drug properties in patients with sepsis or [141, 142]. Once-daily dosing of aminoglycosides is used
septic shock (BPS). for patients with preserved renal function. Patients with
chronically mildly impaired renal function should still
Rationale Early optimization of antimicrobial phar- receive a once-daily-equivalent dose but would normally
macokinetics can improve the outcome of patients with have an extended period (up to 3 days) before the next
severe infection. Several considerations should be made dose. This dosing regimen should not be used in patients
when determining optimal dosing for critically ill patients with severe renal function in whom the aminoglycoside
with sepsis and septic shock. These patients have distinct is not expected to clear within several days. Therapeutic
differences from the typical infected patient that affect drug monitoring of aminoglycosides in this context is
the optimal antimicrobial management strategy. These primarily meant to ensure that trough concentrations are
differences include an increased frequency of hepatic sufficiently low to minimize the potential for renal toxic-
and renal dysfunction, a high prevalence of unrecognized ity. For fluoroquinolones, an approach that optimizes the
immune dysfunction, and a predisposition to infection dose within a nontoxic range (e.g., ciprofloxacin, 600mg
every 12 h, or levofloxacin, 750 mg every 24 h, assum- patients with sepsis [140, 156158]. A recent individual
ing preserved renal function) should provide the high- patient data meta-analysis of randomized controlled tri-
est probability of a favorable microbiologic and clinical als comparing continuous versus intermittent infusion of
response [127, 143, 144]. -lactam antibiotics in critically ill patients with severe
Vancomycin is another antibiotic whose efficacy is sepsis demonstrated an independent protective effect of
at least partially concentration-dependent. Dosing to continuous therapy after adjustment for other correlates
a trough target of 1520 mg/L is recommended by sev- of outcome [140].
eral authorities to maximize the probability of achieving While the weight of evidence supports pharmacoki-
appropriate pharmacodynamic targets, improve tissue netically optimized antimicrobial dosing strategies in
penetration, and optimize clinical outcomes [145147]. critically ill patients with sepsis and septic shock, this
Pre-dose monitoring of trough concentrations is recom- is difficult to achieve on an individual level without a
mended. For sepsis and septic shock, an IV loading dose broader range of rapid therapeutic drug monitoring
of 2530 mg/kg (based on actual body weight) is sug- options than currently available (i.e., vancomycin, teico-
gested to rapidly achieve the target trough drug concen- planin and aminoglycosides). The target group of criti-
tration. A loading dose of 1 g of vancomycin will fail to cally ill, septic patients exhibit a variety of physiologic
achieve early therapeutic levels for a significant subset of perturbations that dramatically alter antimicrobial phar-
patients. In fact, loading doses of antimicrobials with low macokinetics. These include unstable hemodynamics,
volumes of distribution (teicoplanin, vancomycin, colis- increased cardiac output, increased extracellular volume
tin) are warranted in critically ill patients to more rapidly (markedly increasing volume of distribution), variable
achieve therapeutic drug levels due to their expanded kidney and hepatic perfusion (affecting drug clearance)
extracellular volume related to volume expansion follow- and altered drug binding due to reduced serum albumin
ing fluid resuscitation [148152]. [159]. In addition, augmented renal clearance is a recently
Loading doses are also recommended for -lactams described phenomenon that may lead to decreased serum
administered as continuous or extended infusions to antimicrobial levels in the early phase of sepsis [160
accelerate accumulation of drug to therapeutic levels 162]. These factors make individual assessment of opti-
[153]. Notably, the required loading dose of any antimi- mal drug dosing difficult in critically ill patients. Based
crobial is not affected by alterations of renal function, on studies with therapeutic drug monitoring, under-dos-
although this may affect frequency of administration ing (particularly in the early phase of treatment) is com-
and/or total daily dose. mon in critically ill, septic patients, but drug toxicity such
For -lactams, the key pharmacodynamics correlate to as central nervous system irritation with -lactams and
microbiologic and clinical response is the time that the renal injury with colistin is also seen [163166]. These
plasma concentration of the drug is above the pathogen problems mandate efforts to expand access to therapeutic
MIC relative to the dosing interval (T > MIC). A mini- drug monitoring for multiple antimicrobials for critically
mum T > MIC of 60% is generally sufficient to allow a ill patients with sepsis.
good clinical response in mild to moderate illness. How-
ever, optimal response in severe infections, including 6. We suggest empiric combination therapy (using
sepsis, may be achieved with a T > MIC of 100% [139]. at least two antibiotics of different antimicro-
The simplest way to increase T>MIC is to use increased bial classes) aimed at the most likely bacterial
frequency of dosing (given an identical total daily dose). pathogen(s) for the initial management of septic
For example, piperacillin/tazobactam can be dosed at shock (weak recommendation, low quality of evi-
either 4.5 g every 8 h or 3.375 g every 6 h for serious dence).
infections; all things being equal, the latter would achieve Remarks Readers should review Table 6 for defini-
a higher T>MIC. We suggested earlier that initial doses tions of empiric, targeted/definitive, broad-spectrum,
of -lactams can be given as a bolus or rapid infusion to combination, and multidrug therapy before reading
rapidly achieve therapeutic blood levels. However, fol- this section.
lowing the initial dose, an extended infusion of drug over
several hours (which increases T>MIC) rather than the
standard 30min has been recommended by some author- 7. We suggest that combination therapy not be rou-
ities [154, 155]. In addition, some meta-analyses suggest tinely used for ongoing treatment of most other
that extended/continuous infusion of -lactams may be serious infections, including bacteremia and sep-
more effective than intermittent rapid infusion, particu- sis without shock (weak recommendation, low
larly for relatively resistant organisms and in critically ill quality of evidence).
 Remarks This does not preclude the use of multidrug Remarks This does not preclude the use of multidrug
therapy to broaden antimicrobial activity. therapy to broaden antimicrobial activity.

8. We recommend against combination therapy for 9. If combination therapy is initially used for septic
the routine treatment of neutropenic sepsis/bac- shock, we recommend de-escalation with discon-
teremia (strong recommendation, moderate qual- tinuation of combination therapy within the first
ity of evidence). few days in response to clinical improvement and/

Table6 Important terminology forantimicrobial recommendations

Empiric therapy Inial therapy started in the absence of definive

microbiologic pathogen idenficaon. Empiric therapy may
be mono-, combinaon, or broad-spectrum, and/or
muldrug in nature.

Targeted/definive therapy Therapy targeted to a specific pathogen (usually aer

microbiologic idenficaon). Targeted/definive therapy
may be mono- or combinaon, but is not intended to be
broad-spectrum.
Broad-spectrum therapy The use of one or more anmicrobial agents with the
specific intent of broadening the range of potenal
pathogens covered, usually during empiric therapy (e.g.,
piperacillin/tazobactam, vancomycin, and anidulafungin;
each is used to cover a different group of pathogens).
Broad-spectrum therapy is typically empiric since the usual
purpose is to ensure anmicrobial coverage with at least
one drug when there is uncertainty about the possible
pathogen. On occasion, broad-spectrum therapy may be
connued into the targeted/definive therapy phase if
mulple pathogens are isolated.

Muldrug therapy Therapy with mulple anmicrobials to deliver broad-

spectrum therapy (i.e., to broaden coverage) for empiric
therapy (i.e., where pathogen is unknown) or to potenally
accelerate pathogen clearance (combinaon therapy) with
respect to a specific pathogen(s) where the pathogen(s) is
known or suspected (i.e., for both targeted or empiric
therapy). This term therefore includes combinaon therapy.
Combinaon therapy The use of mulple anbiocs (usually of different
mechanisc classes) with the specific intent of covering the
known or suspected pathogen(s) with more than one
anbioc (e.g., piperacillin/tazobactam and an
aminoglycoside or fluoroquinolone for gram-negave
pathogens) to accelerate pathogen clearance rather than to
broaden anmicrobial coverage. Other proposed
applicaons of combinaon therapy include inhibion of
bacterial toxin producon (e.g., clindamycin with -lactams
for streptococcal toxic shock) or potenal immune
modulatory effects (macrolides with a -lactam for
pneumococcal pneumonia).
 or evidence of infection resolution. This applies to A number of other recent observational studies and
both targeted (for culture-positive infections) and some small, prospective trials also support initial com-
empiric (for culture-negative infections) combina- bination therapy for selected patients with specific
tion therapy (BPS). pathogens (e.g., severe pneumococcal infection, multid-
rug-resistant gram-negative pathogens) [172, 176182].
Rationale In light of the increasing frequency of path- Unfortunately, in most cases and pending the develop-
ogen resistance to antimicrobial agents in many parts of ment of rapid bedside pathogen detection techniques,
the world, the initial use of multidrug therapy is often the offending pathogen is not known at the time of pres-
required to ensure an appropriately broad-spectrum entation. Therefore, specifying combination therapy to
range of coverage for initial empiric treatment. The use of specific identified pathogens is useful only if more pro-
multidrug therapy for this purpose in severe infections is longed targeted combination therapy is contemplated. In
well understood. addition, with respect to multidrug-resistant pathogens,
The phrase combination therapy in the context of this both individual studies and meta-analyses yield variable
guideline connotes the use of two different classes of anti- results depending on the pathogen and the clinical sce-
biotics (usually a -lactam with a fluoroquinolone, ami- nario [179184]. Infectious diseases consultation may be
noglycoside, or macrolide) for a single putative pathogen advisable if multidrug-resistant pathogens are suspected.
expected to be sensitive to both, particularly for purposes One area of broad consensus on the use of a specific form
of accelerating pathogen clearance. The term is not used of combination therapy is for streptococcal toxic shock
where the purpose of a multidrug strategy is to strictly syndrome, for which animal models and uncontrolled,
broaden the range of antimicrobial activity (e.g., vanco- clinical experience demonstrate a survival advantage with
mycin added to ceftazidime, metronidazole added to an penicillin and clindamycin, the latter as a transcriptional
aminoglycoside or an echinocandin added to a -lactam). inhibitor to pyrogenic exotoxin superantigens [109, 185,
A propensity-matched analysis and a meta-analysis/ 186].
meta-regression analysis have demonstrated that com- Despite evidence suggesting benefit of combination
bination therapy produces higher survival in severely ill therapy in septic shock, this approach has not been
septic patients with a high risk of death, particularly in shown to be effective for ongoing treatment of most
those with septic shock [167, 168]. A meta-regression other serious infections, including bacteremia and sepsis
study [167] suggested benefit with combination therapy without shock [168, 174, 175]. The term ongoing treat-
in patients with a mortality risk greater than 25%. Sev- ment includes extended empiric therapy for culture-
eral observational studies have similarly shown a survival negative infections and extended definitive/targeted
benefit in very ill patients [169172]. However, the afore- therapy where a pathogen is identified. In the case of
mentioned meta-regression analysis also suggested the neutropenia in the absence of septic shock, studies using
possibility of increased mortality risk with combination modern broad-spectrum antibiotics consistently suggest
therapy in low-risk (<15% mortality risk) patients with- that, while multidrug therapy to broaden pathogen cov-
out septic shock [167]. One controlled trial suggested erage (e.g., to include Candida species) may be useful,
that, when using a carbapenem as empiric therapy in a combination therapy using a -lactam and an aminogly-
population at low risk for infection with resistant micro- coside for purposes of accelerating pathogen clearance is
organisms, the addition of a fluoroquinolone does not not beneficial for less severely ill low-risk patients [187].
improve patients outcomes [173]. A close examination Combination therapy of this sort for even high-risk
of the results, however, demonstrates findings consistent neutropenic patients (inclusive of hemodynamic insta-
with the previously mentioned meta-regression (trend bility and organ failure) with sepsis is inconsistently sup-
to benefit in septic shock with an absence of benefit in ported by several international expert groups [106, 188].
sepsis without shock). Despite the overall favorable evi- This position against combination therapy for a single
dence for combination therapy in septic shock, direct pathogen in any form of neutropenic infection emphati-
evidence from adequately powered RCTs is not avail- cally does not preclude the use of multidrug therapy for
able to validate this approach definitively. Nonetheless, the purpose of broadening the spectrum of antimicrobial
in clinical scenarios of severe clinical illness (particularly treatment.
septic shock), several days of combination therapy is High-quality data on clinically driven de-escalation of
biologically plausible and is likely to be clinically useful antimicrobial therapy for severe infections are limited
[152, 167, 168] even if evidence has not definitively dem- [189]. Early de-escalation of antimicrobial therapy in the
onstrated improved clinical outcome in bacteremia and context of combination therapy as described here has
sepsis without shock [174, 175]. Thus, we issue a weak not been studied. However, observational studies have
recommendation based on low quality of evidence. shown that early de-escalation of multidrug therapy is
associated with equivalent or superior clinical outcomes demonstrated, although cumulative antimicrobial toxic-
in sepsis and septic shock [54, 190192]; despite this, at ity; the occurrence of antimicrobial-associated secondary
least one study has indicated an increased frequency of infections (e.g., C. difficile colitis); and selection of, and
superinfection and longer ICU stay [192]. superinfection with, multidrug-resistant pathogens are
In addition to institutional benefit with respect to lim- all potential contributors.
iting a driver of antimicrobial resistance, early de-esca- Although patient factors will influence the length of
lation can also benefit the individual patient [193195]. antibiotic therapy, a treatment duration of 710 days (in
Although the data are not entirely consistent, on bal- the absence of source control issues) is generally ade-
ance, an approach that emphasizes early de-escalation is quate for most serious infections [103, 197199]. Cur-
favored when using combination therapy. rent guidelines recommend a 7-day course of therapy for
While substantial consensus on the need for early de- nosocomial pneumonia [both hospital-acquired and ven-
escalation of combination therapy exists, agreement is tilator-associated pneumonia (VAP)] [103]. Recent data
lacking on precise criteria for triggering de-escalation. suggest that some serious infections may be treated with
Among approaches used by panel members are de-esca- shorter courses especially if there is a need for and suc-
lation based on: (a) clinical progress (shock resolution, cessful provision of source control [200, 201].
decrease in vasopressor requirement, etc.), (b) infection Subgroup analysis of the most critically ill subjects
resolution as indicated by biomarkers (especially pro- [Acute Physiologic and Chronic Health Evaluation
calcitonin), and (c) a relatively fixed duration of combi- (APACHE) II score greater than either 15 or 20] in the
nation therapy. This lack of consensus on de-escalation short course of antimicrobials in the intra-abdominal
criteria for combination therapy reflects the lack of solid sepsis study of Sawyer etal. demonstrated no difference
data addressing this issue (notwithstanding procalcitonin in outcome based on the duration of therapy (as with
data relating to general de-escalation). the overall group) [200, 202]. A treatment duration of
35days or fewer was as effective as a duration of up to
10. We suggest that an antimicrobial treatment dura- 10 days. Similarly, studies have shown that a treatment
tion of 710 days is adequate for most serious duration of <7days is as effective as longer durations in
infections associated with sepsis and septic shock the management of acute pyelonephritis with or with-
(weak recommendation, low quality of evidence). out bacteremia [201], uncomplicated cellulitis [203], and
11. We suggest that longer courses are appropri- spontaneous bacterial peritonitis [204]. Some conditions
ate in patients who have a slow clinical response, are generally thought to require more prolonged antimi-
undrainable foci of infection, bacteremia with crobial therapy. These include situations in which there
S. aureus, some fungal and viral infections, or is a slow clinical response, undrainable foci of infection,
immunologic deficiencies, including neutropenia bacteremia with S. aureus (particularly MRSA) [67, 104],
(weak recommendation, low quality of evidence). candidemia/invasive candidiasis [105] and other fungal
12. We suggest that shorter courses are appropriate in infections, some viral infections (e.g., herpes, cytomeg-
some patients, particularly those with rapid clini- alovirus), and immunologic deficiencies, including neu-
cal resolution following effective source control of tropenia [188].
intra-abdominal or urinary sepsis and those with Assessment of the required duration of therapy in criti-
anatomically uncomplicated pyelonephritis (weak cally ill patients should include host factors, particularly
recommendation, low quality of evidence). immune status. For example, patients with neutropenic
13. We recommend daily assessment for de-escala- infection and sepsis usually require therapy for at least
tion of antimicrobial therapy in patients with sep- the duration of their neutropenia. The nature of the
sis and septic shock (BPS). infecting pathogen also plays a role. In particular, uncom-
plicated S. aureus bacteremia requires at least 14 days
Rationale Unnecessarily prolonged administration of of therapy, while complicated bacteremia requires treat-
antimicrobials is detrimental to society and to the indi- ment as an endovascular infection with 6weeks of ther-
vidual patient. For society, excessive antimicrobial use apy. Uncomplicated bacteremia has been defined as: (1)
drives antimicrobial resistance development and dis- exclusion of endocarditis, (2) no implanted prostheses,
semination [196]. For individual patients, prolonged 3) negative results of follow-up blood cultures drawn
antibiotic therapy is associated with specific illnesses 24 days after the initial set, (4) defervescence within
such as Clostridium difficile colitis [195] and, more 72h after the initiation of effective antibiotic therapy, and
broadly, an increased mortality risk [54]. The basis of the (5) no evidence of metastatic infection [104].
increased mortality with unnecessarily prolonged and Patients with candidemia (whether or not catheter-
broad antimicrobial therapy has not been convincingly associated) and deep Candida infections, whether or not
associated with sepsis, require more prolonged therapy 15. We suggest that procalcitonin levels can be used
[105, 205]. Highly resistant gram-negative pathogens to support the discontinuation of empiric anti-
with marginal sensitivity to utilized antimicrobials may biotics in patients who initially appeared to have
be slow to clear and represent another example. The sepsis, but subsequently have limited clinical evi-
nature and site of infection may also affect duration of dence of infection (weak recommendation, low
therapy. Larger abscesses and osteomyelitis have limited quality of evidence).
drug penetration and require longer therapy. Although it
is well known that endocarditis requires prolonged anti- Rationale During the past decade, the role of bio-
microbial therapy, severe disease more typically presents markers to assist in the diagnosis and management of
as cardiac failure/cardiogenic shock and emboli rather infections has been extensively explored. The use of
than as sepsis or septic shock [206, 207]. A variety of galactomannan and -d-glucan to assist in the assess-
other factors may play a role in determining the optimal ment of invasive aspergillus (and a broad range of fungal
duration of therapy, particularly in critically ill infected pathogens) has become well accepted [209, 210].
patients. If the clinician is uncertain, infectious diseases Similarly, measurement of serum procalcitonin is com-
consultation should be sought. monly used in many parts of the world to assist in the
Few of the studies noted focused on patients with sep- diagnosis of acute infection and to help define the dura-
tic shock, sepsis with organ failure, or even critical illness. tion of antimicrobial therapy. Various procalcitonin-based
To an extent, standard recommendations on duration of algorithms have been used to direct de-escalation of anti-
therapy in this document depend on inferences from less microbial therapy in severe infections and sepsis [211
ill cohorts. Therefore, decisions to narrow or stop antimi- 216]. However, it is not clear that any particular algorithm
crobial therapy must ultimately be made on the basis of provides a clinical advantage over another. A large body of
sound clinical judgment. literature suggests that use of such algorithms can speed
There are many reasons for unnecessarily prolonged safe antimicrobial de-escalation compared to standard
antimicrobial therapy. For complicated, critically ill clinical approaches with reduced antimicrobial consump-
patients admitted with serious infections, noninfectious tion without an adverse effect on mortality. Recently, a
concurrent illness and medical interventions may pro- large randomized trial on procalcitonin use in critically ill
duce signs and symptoms consistent with active infec- patients with presumed bacterial infection demonstrated
tion (even following control of infection). For example, evidence of a reduction in duration of treatment and daily
pulmonary infiltrates and shortness of breath may be defined doses of antimicrobials [217]. However, given
caused by pulmonary edema in addition to pneumonia; the design of the study, the reduction could have been
an elevated white cell count may occur as a consequence related to a prompting effect as seen in other studies [55,
of corticosteroid administration or physiologic stress; 218]. In addition, the procalcitonin group showed a sig-
fever may be associated with certain drugs, including nificant reduction in mortality. This finding is congruent
-lactams and phenytoin. In addition, there is a natural with studies demonstrating an association between early
tendency to want to continue a therapy that is often seen antimicrobial de-escalation and survival in observational
as benign long enough to be confident of cure. However, studies of sepsis and septic shock [54, 55].
as discussed, antimicrobials are not an entirely benign This benefit is uncertain, though, because another
therapy. In low-risk patients, the adverse effects can out- meta-analysis of randomized controlled studies of de-
weigh any benefit. escalation failed to demonstrate a similar survival advan-
Given the potential harm associated with unnecessar- tage [219]. Meta-analyses also suggest that procalcitonin
ily prolonged antimicrobial therapy, daily assessment for can also be used to assist in differentiating infectious
de-escalation of antimicrobial therapy is recommended and noninfectious conditions at presentation [211, 214,
in patients with sepsis and septic shock. Studies have 216]. The strongest evidence appears to relate to bacterial
shown that daily prompting on the question of antimi- pneumonia versus noninfectious pulmonary pathology
crobial de-escalation is effective and may be associated [216, 220], where meta-analysis suggests that procalci-
with improved mortality rates [55, 208]. tonin may assist in predicting the presence of bacteremia,
particularly in ICU patients [221].
14. We suggest that measurement of procalcitonin No evidence to date demonstrates that the use of pro-
levels can be used to support shortening the dura- calcitonin reduces the risk of antibiotic-related diarrhea
tion of antimicrobial therapy in sepsis patients from C. difficile. However, the occurrence of C. diffi-
(weak recommendation, low quality of evidence). cile colitis is known to be associated with cumulative
antibiotic exposure in individual patients [195], so such a Clinical experience suggests that, without adequate
benefit is likely. In addition, although prevalence of anti- source control, some more severe presentations will not
microbial resistance has not been shown to be reduced stabilize or improve despite rapid resuscitation and pro-
by the use of procalcitonin, the emergence of antimicro- vision of appropriate antimicrobials. In view of this fact,
bial resistance is known to be associated with total anti- prolonged efforts at medical stabilization prior to source
microbial consumption in large regions [196]. control for severely ill patients, particularly those with
It is important to note that procalcitonin and all other septic shock, are generally not warranted [108].
biomarkers can provide only supportive and supplemen- The selection of optimal source control methods must
tal data to clinical assessment. Decisions on initiating, weigh the benefits and risks of the specific intervention,
altering, or discontinuing antimicrobial therapy should risks of transfer for the procedure, potential delays associ-
never be made solely on the basis of changes in any bio- ated with a specific procedure, and the probability of the
marker, including procalcitonin. procedures success. Source control interventions may
cause further complications, such as bleeding, fistulas,
E. SOURCE CONTROL or inadvertent organ injury. In general, the least invasive
effective option for source control should be pursued.
1. We recommend that a specific anatomic diagno- Open surgical intervention should be considered when
sis of infection requiring emergent source con- other interventional approaches are inadequate or cannot
trol be identified or excluded as rapidly as pos- be provided in a timely fashion. Surgical exploration may
sible in patients with sepsis or septic shock, and also be indicated when diagnostic uncertainty persists
that any required source control intervention be despite radiologic evaluation or when the probability of
implemented as soon as medically and logisti- success with a percutaneous procedure is uncertain and
cally practical after the diagnosis is made (BPS). the mortality risk as a consequence of a failed procedure
2. We recommend prompt removal of intravascular causing delays is high. Specific clinical situations require
access devices that are a possible source of sep- consideration of available choices, the patients prefer-
sis or septic shock after other vascular access has ences, and the clinicians expertise. Logistic factors unique
been established (BPS). to each institution, such as surgical or interventional staff
availability, may also play a role in the decision.
Rationale The principles of source control in the man- Intravascular devices such as central venous catheters
agement of sepsis and septic shock include rapid diagno- can be the source of sepsis or septic shock. An intravas-
sis of the specific site of infection and determination of cular device suspected to be a source of sepsis should gen-
whether that infection site is amenable to source control erally be removed promptly after establishing another site
measures (specifically the drainage of an abscess, debride- for vascular access. In the absence of both septic shock and
ment of infected necrotic tissue, removal of a potentially fungemia, some implanted, tunneled catheter infections
infected device, and definitive control of a source of ongo- may be able to be treated effectively with prolonged anti-
ing microbial contamination) [222]. Foci of infection microbial therapy if removal of the catheter is not practical
readily amenable to source control include intra-abdomi- [67]. However, catheter removal (with antimicrobial ther-
nal abscesses, gastrointestinal perforation, ischemic bowel apy) is definitive and is preferred where possible.
or volvulus, cholangitis, cholecystitis, pyelonephritis asso-
ciated with obstruction or abscess, necrotizing soft tissue F. FLUID THERAPY
infection, other deep space infection (e.g., empyema or
septic arthritis), and implanted device infections. 1. We recommend that a fluid challenge technique
Infectious foci suspected to cause septic shock should be applied where fluid administration is contin-
be controlled as soon as possible following successful ued as long as hemodynamic factors continue to
initial resuscitation [223, 224]. A target of no more than improve (BPS).
612 h after diagnosis appears to be sufficient for most 2. We recommend crystalloids as the fluid of choice
cases [223229]. Observational studies generally show for initial resuscitation and subsequent intravas-
reduced survival beyond that point. The failure to show cular volume replacement in patients with sepsis
benefit with even earlier source control implementation and septic shock (strong recommendation, mod-
may be a consequence of the limited number of patients erate quality of evidence).
in these studies. Therefore, any required source control 3. We suggest using either balanced crystalloids or
intervention in sepsis and septic shock should ideally be saline for fluid resuscitation of patients with sep-
implemented as soon as medically and logistically practi- sis or septic shock (weak recommendation, low
cal after the diagnosis is made. quality of evidence).
 4. We suggest using albumin in addition to crys- No cost-effectiveness studies compare balanced and
talloids for initial resuscitation and subsequent unbalanced crystalloid solutions. Therefore, we consid-
intravascular volume replacement in patients ered the desirable and undesirable consequences to be
with sepsis and septic shock when patients comparable for both solutions, and issued a weak recom-
require substantial amounts of crystalloids (weak mendation to use either solution. Hyperchloremia should
recommendation, low quality of evidence). be avoided, however, and thus close scrutiny of serum
5. We recommend against using hydroxyethyl chloride levels is advised, whichever fluid solutions are
starches (HESs) for intravascular volume replace- used.
ment in patients with sepsis or septic shock The SAFE study indicated that albumin administra-
(strong recommendation, high quality of evi- tion was safe and equally effective as 0.9% saline in
dence). ICU patients requiring fluid administration [239]. A
6. We suggest using crystalloids over gelatins when meta-analysis aggregated data from 17 randomized tri-
resuscitating patients with sepsis or septic shock als (n = 1977) of albumin versus other fluid solutions
(weak recommendation, low quality of evidence). in patients with sepsis or septic shock [240]; 279 deaths
occurred among 961 albumin-treated patients (29%) ver-
Rationale The use of IV fluids in the resuscitation of sus 343 deaths among 1016 patients (34%) treated with
patients is a cornerstone of modern therapy. Despite this, other fluids, favoring albumin (OR 0.82; 95% CI 0.67
there is little available evidence from RCTs to support 1.00). When albumin-treated patients were compared
its practice; this is an area in which research is urgently with those receiving crystalloids (seven trials, n = 144),
needed. One trial of children (mostly with malaria) in the odds ratio of dying was significantly reduced for albu-
Africa, in a setting where escalation to mechanical ven- min-treated patients (OR 0.78; 95% CI 0.620.99).
tilation and other organ support was limited, questioned Since the 2012 SSC guideline publication, six system-
this practice [230]. We believe that the extrapolation of atic reviews/meta-analyses [237, 241245] were pub-
these data to patients in better-resourced settings is not lished assessing the use of albumin solutions in the
valid and thus recommend that clinicians restore euv- management of patients with sepsis or septic shock. Each
olemia with IV fluids, more urgently initially, and then meta-analysis included different populations (adult/child,
more cautiously as the patient stabilizes. There is some septic/nonseptic, and acute resuscitation/maintenance),
evidence that a sustained positive fluid balance during different comparators and different duration of exposure
ICU stay is harmful [231235]. We do not recommend, to the intervention (hours, days), which made combining
therefore, that fluid be given beyond initial resuscitation data challenging (ESM 7).
without some estimate of the likelihood that the patient Xu et al. [242] evaluated albumin compared to crys-
will respond positively. talloid as a resuscitation fluid. Five studies, encompass-
The absence of any clear benefit following the adminis- ing 3658 sepsis and 2180 septic shock patients, were
tration of colloid compared to crystalloid solutions in the included. Albumin use resulted in reduced septic shock
combined subgroups of sepsis, in conjunction with the 90-day mortality (OR 0.81; 95% CI 0.670.97) and
expense of albumin, supports a strong recommendation trended toward reduced 90-day mortality in sepsis (OR
for the use of crystalloid solutions in the initial resuscita- 0.88; 95% CI 0.761.01; p=0.08). Jiang etal. [245] evalu-
tion of patients with sepsis and septic shock. ated albumin in a mixed population of sepsis severity
We were unable to recommend one crystalloid solution including adults and children. Three septic shock stud-
over another because no direct comparisons have been ies, encompassing 1931 patients, were included. Albu-
made between isotonic saline and balanced salt solu- min use resulted in decreased mortality (OR 0.89; 95% CI
tions in patients with sepsis. One before-after study in all 0.800.99) with low heterogeneity (I2=0%). A mortality
ICU patients suggested increased rates of acute kidney reduction trend was reported for albumin administration
injury and RRT in patients managed with a chloride-lib- compared to crystalloids when given less than 6 h from
eral strategy compared to a chloride-restrictive strategy identification (11 studies; n = 5515; OR 0.94; 95% CI
[236]. There is indirect low-quality evidence from a net- 0.861.03).
work meta-analysis suggesting improved outcome with Patel etal. [244] evaluated mixed populations, includ-
balanced salt solutions as compared to saline in patients ing resuscitation and maintenance. Additionally, a series
with sepsis [237] (ESM 6). In addition, the neutral result of studies excluded from other meta-analyses due to
of the SPLIT cluster RCT in ICU patients (mainly surgi- accuracy concerns was included in this evaluation [246
cal patients) in four New Zealand ICUs lowered our con- 248]. When comparing crystalloid and albumin, the
fidence in recommending one solution over the other authors report a combined mortality benefit of albumin
[238]. as compared to crystalloid (seven studies, n=3878; OR
0.93; 95% CI 0.861.00), but it was not consistent across available evidence resulted in a strong recommendation
individual severity subgroups. Use of albumin in septic against the use of HES in resuscitation of patients with
shock trended toward mortality benefit (four studies; sepsis or septic shock.
n = 1949; OR 0.91; 95% CI 0.821.01; p = 0.06), and Gelatin is another synthetic colloid that can be used for
the use of albumin in sepsis was not significant (four fluid resuscitation; however, high-quality studies compar-
studies; n = 1929; OR 0.96; 95% CI 0.831.10). Evalua- ing gelatins to other fluids in patients with sepsis or septic
tion of treatment within 24h also trended toward mor- shock are lacking. Trials conducted in critically ill patients
tality benefit (four studies; n = 3832; RR 0.93; 95% CI were summarized in a recent meta-analysis [251]. Gelatin
0.861.01). Rochwerg 2014 et al. [237] evaluated resus- use in critically ill adult patients did not increase mor-
citative fluid use in a network meta-analysis of 14 trials, tality (RR 1.10; 95% CI 0.851.43; low-quality evidence)
encompassing 18,916 patients. When comparing albu- or acute kidney injury (RR 1.35; 95% CI 0.583.14; very
min to crystalloid, there was no significant reduction low-quality evidence) compared to albumin or crystalloid.
in mortality with moderate quality of evidence in both These results are limited by indirectness, since the studies
the four- and six-node analyses (four-node: OR 0.83; did not focus on critically ill patients. The aforementioned
credible interval [CrI] 0.651.04; six-node OR 0.82; CrI network meta-analysis by Rochwerg etal. did not identify
0.651.04). any RCTs comparing gelatins to crystalloids or albumin;
The ALBIOS trial [249] showed no mortality benefit of therefore, the generated estimates were imprecise and
albumin in combination with crystalloids compared to were based on indirect comparisons [237]. Given the low
crystalloids alone in patients with sepsis or septic shock quality of the available data and the cost associated with
(RR 0.94; 95% CI 0.851.05); a subgroup analysis sug- gelatin use, we issued a weak recommendation favoring
gested that the albumin group was associated with lower the use of crystalloids over gelatins.
90-day mortality in patients with septic shock (RR 0.87;
95% CI 0.770.99). Fluid administration continued for G. VASOACTIVE MEDICATIONS
28days or until discharge and was not targeted for acute
resuscitation. In addition, the amount of 20% albumin was 1. We recommend norepinephrine as the first-
guided by serum albumin level with the ultimate goal of choice vasopressor (strong recommendation,
achieving levels >30g/L. These results are limited by sig- moderate quality of evidence).
nificant indirectness and imprecision, resulting in low 2. We suggest adding either vasopressin (up to
quality of evidence. 0.03 U/min) (weak recommendation, moderate
HESs are colloids for which there are safety concerns quality of evidence) or epinephrine (weak rec-
in patients with sepsis. A meta-analysis of nine trials ommendation, low quality of evidence) to nor-
(3456 patients) comparing 6% HES 130/0.380.45 solu- epinephrine with the intent of raising MAP to
tions to crystalloids or albumin in patients with sepsis target, or adding vasopressin (up to 0.03U/min)
showed no difference in all-cause mortality (RR 1.04; (weak recommendation, moderate quality of evi-
95% CI 0.891.22) [250]. However, when low risk of dence) to decrease norepinephrine dosage.
bias trials were analyzed separately, HES use resulted in 3. We suggest using dopamine as an alternative
higher risk of death compared to other fluids (RR 1.11; vasopressor agent to norepinephrine only in
95% CI 1.011.22; high-quality evidence), which trans- highly selected patients (e.g., patients with low
lates to 34 more deaths per 1000 patients. Furthermore, risk of tachyarrhythmias and absolute or relative
HES use led to a higher risk of RRT (RR 1.36; 95% CI bradycardia) (weak recommendation, low quality
1.081.72; high-quality evidence) [250]. A subsequent of evidence).
network meta-analysis focused on acute resuscitation of 4. We recommend against using low-dose dopa-
patients with sepsis or septic shock and found that HES mine for renal protection (strong recommenda-
resulted in higher risk of death (10 RCTs; OR 1.13; CrI, tion, high quality of evidence).
0.991.30; high-quality evidence) and need for RRT (7 5. We suggest using dobutamine in patients who
RCTs; OR 1.39; CrI, 1.171.66; high-quality evidence) show evidence of persistent hypoperfusion
compared to crystalloids. When comparing albumin to despite adequate fluid loading and the use of
HES, albumin resulted in lower risk of death (OR 0.73; vasopressor agents (weak recommendation, low
CrI, 0.560.93; moderate-quality evidence) and a trend quality of evidence).
toward less need for RRT (OR 0.74; CrI, 0.531.04; low- Remarks If initiated, vasopressor dosing should be
quality evidence) [237]. Overall, the undesirable con- titrated to an end point reflecting perfusion, and the
sequences of using HES (increased risk of death and agent reduced or discontinued in the face of worsen-
need for RRT) along with moderate to high quality of ing hypotension or arrhythmias.
 Rationale The physiologic effects of vasopressors and plus vasopressin at 0.03 U/min, showed no difference in
combined inotrope/vasopressor selection in septic shock outcome in the intent-to-treat population [274]. An a
are outlined in an extensive number of literature reviews priori defined subgroup analysis demonstrated improved
[252261]. Norepinephrine increases MAP due to its survival among patients receiving <15 g/min norepi-
vasoconstrictive effects, with little change in heart rate nephrine at randomization with the addition of vasopres-
and less increase in stroke volume compared with dopa- sin; however, the pretrial rationale for this stratification
mine. Dopamine increases MAP and cardiac output, pri- was based on exploring potential benefit in the popula-
marily due to an increase in stroke volume and heart rate. tion requiring 15g/min norepinephrine. Higher doses
Norepinephrine is more potent than dopamine and may of vasopressin have been associated with cardiac, digi-
be more effective at reversing hypotension in patients tal, and splanchnic ischemia and should be reserved for
with septic shock. Dopamine may be particularly useful situations in which alternative vasopressors have failed
in patients with compromised systolic function but causes [275]. In the VANISH trial, 409 patients with septic shock
more tachycardia and may be more arrhythmogenic than were randomized in a factorial (22) design to receive
norepinephrine [262]. It may also influence the endocrine vasopressin with placebo or hydrocortisone, or norepi-
response via the hypothalamic pituitary axis and may nephrine with placebo or hydrocortisone. There was no
have immunosuppressive effects [263]. However, a recent significant difference in kidney failure-free days or death;
systematic review and meta-analysis that included 11 however, the vasopressin group had less use of RRT
randomized trials (n=1710) comparing norepinephrine [276]. We conducted an updated meta-analysis to include
to dopamine does not support the routine use of dopa- the results of the VANISH trial. Data from nine trials
mine in the management of septic shock [264]. Indeed, (n=1324 patients with septic shock), comparing norepi-
norepinephrine use resulted in lower mortality (RR 0.89; nephrine with vasopressin (or terlipressin) demonstrated
95% CI 0.810.98, high-quality evidence) and lower risk no significant difference in mortality (RR 0.89; 95% CI
of arrhythmias (RR 0.48; 95% CI 0.400.58; high-quality 0.791.00; moderate-quality evidence) (ESM 10) [268,
evidence) compared with dopamine (ESM 8). 271, 272, 277279]. Results were similar after exclud-
Human and animal studies suggest that the infusion of ing trials that used a combination of norepinephrine and
epinephrine may have deleterious effects on the splanch- vasopressin in the intervention arm (RR 0.89; 95% CI
nic circulation and produces hyperlactatemia. However, 0.771.02). Large studies comparing vasopressin to other
clinical trials do not demonstrate worsening of clinical vasopressors in septic shock are lacking; most of the data
outcomes. One RCT comparing norepinephrine to epi- regarding vasopressin support a sparing effect on norepi-
nephrine demonstrated no difference in mortality but an nephrine dose, and there is uncertainty about the effect
increase in adverse drug-related events with epinephrine of vasopressin on mortality. Norepinephrine, therefore,
[265]. Similarly, a meta-analysis of four randomized tri- remains the first-choice vasopressor to treat patients with
als (n = 540) comparing norepinephrine to epinephrine septic shock. We do not recommend the use of vasopres-
found no significant difference in mortality (RR 0.96; CI sin as a first-line vasopressor for the support of MAP and
0.771.21; low-quality evidence) (ESM 9) [264]. Epineph- would advocate caution when using it in patients who are
rine may increase aerobic lactate production via stimu- not euvolemic or at doses higher than 0.03U/min.
lation of skeletal muscle 2-adrenergic receptors and Phenylephrine is a pure -adrenergic agonist. Clini-
thus may preclude the use of lactate clearance to guide cal trial data in sepsis are limited. Phenylephrine has the
resuscitation. potential to produce splanchnic vasoconstriction [280].
Vasopressin levels in septic shock have been reported A network meta-analysis resulted in imprecise esti-
to be lower than anticipated for a shock state [266]. Low mates (wide confidence intervals) when phenylephrine
doses of vasopressin may be effective in raising blood was compared to other vasopressors [281]. Therefore,
pressure in patients refractory to other vasopressors and the impact on clinical outcomes is uncertain, and phe-
may have other potential physiologic benefits [266271]. nylephrine use should be limited until more research is
Terlipressin has similar effects, but is long-acting [272]. available.
Studies show that vasopressin concentrations are elevated A large randomized trial and meta-analysis compar-
in early septic shock, but decrease to normal range in the ing low-dose dopamine to placebo found no difference in
majority of patients between 24 and 48h as shock contin- need for RRT, urine output, time to renal recovery, sur-
ues [273]. This finding has been called relative vasopres- vival, ICU stay, hospital stay, or arrhythmias [282, 283].
sin deficiency because, in the presence of hypotension, Thus, the available data do not support administration of
vasopressin would be expected to be elevated. The sig- low doses of dopamine solely to maintain renal function.
nificance of this finding is unknown. The VASST trial, an Myocardial dysfunction consequent to infection occurs
RCT comparing norepinephrine alone to norepinephrine in a subset of patients with septic shock, but cardiac
output is usually preserved by ventricular dilation, tach- patients with septic shock and acute respiratory distress
ycardia, and reduced vascular resistance [284]. Some syndrome (ARDS) randomized to levosimendan or pla-
portion of these patients may have diminished cardiac cebo, levosimendan improved right ventricular perfor-
reserve, and may not be able to achieve a cardiac out- mance and mixed venous oxygen saturation compared
put adequate to support oxygen delivery. Recognition of to placebo [291]. Trials comparing levosimendan with
such reduced cardiac reserve can be challenging; imag- dobutamine are limited but show no clear advantage for
ing studies that show decreased ejection fraction may not levosimendan [292]. Levosimendan is more expensive
necessarily indicate inadequate cardiac output. Concomi- than dobutamine and is not available in many parts of
tant measurement of cardiac output along with a meas- the world. Six small RCTs (116 patients in total) com-
ure of the adequacy of perfusion is preferable. pared levosimendan to dobutamine; pooled estimates
Routinely increasing cardiac output to predeter- showed no significant effect on mortality (RR 0.83; 95%
mined supranormal levels in all patients clearly does CI 0.661.05; low quality) (ESM 11). Given the low-qual-
not improve outcomes, as shown by two large prospec- ity evidence available and the higher cost associated with
tive clinical trials of critically ill ICU patients with sepsis levosimendan, dobutamine remains the preferred choice
treated with dobutamine [285287]. in this population. An RCT enrolled 516 patients with
Some patients, however, may have improved tissue per- septic shock who were randomized to receive either levo-
fusion with inotropic therapy aimed at increasing oxygen simendan or placebo; there was no difference in mortal-
delivery. In this situation, dobutamine is the first-choice ity. However, levosimendan led to significantly higher risk
inotrope for patients with measured or suspected low of supraventricular tachyarrhythmia than placebo (abso-
cardiac output in the presence of adequate left ventric- lute difference, 2.7%; 95% CI 0.15.3%) [293]. The results
ular filling pressure (or clinical assessment of adequate of this trial question the systematic use of this agent in
fluid resuscitation) and adequate MAP. Monitoring the patients with septic shock. Of note, cardiac function was
response of indices of perfusion to measured increases not evaluated in that trial, and inotropic stimulation may
in cardiac output is the best way to target such a therapy be of benefit in patients with a low cardiac output due to
[287]. impaired cardiac function.
The data supporting dobutamine are primarily physio-
logic, with improved hemodynamics and some improve- 6. We suggest that all patients requiring vasopres-
ment in indices of perfusion, which may include clinical sors have an arterial catheter placed as soon as
improvement, decreasing lactate levels, and improve- practical if resources are available (weak recom-
ment in Scvo2. No randomized controlled trials have mendation, very low quality of evidence).
compared the effects of dobutamine versus placebo on
clinical outcomes. Mortality in patients randomized to Rationale In shock states, estimation of blood pres-
dobutamine added to norepinephrine was no different sure using a cuff, especially an automated measurement
compared to epinephrine [287], although the trial may system, may be inaccurate. Use of an arterial cannula
have been underpowered. Dobutamine was used as the provides a more accurate and reproducible measure-
first-line inotrope as part of standard care in clinical trials ment of arterial pressure [287, 294] and also allows beat-
of EGDT [16, 19, 288, 289], and adverse effects on mor- to-beat analysis so that decisions regarding therapy can
tality were not detected with its use. be based on immediate and reproducible blood pressure
Although there are only a few studies, alternative ino- information [295]. Insertion of radial arterial catheters is
tropic agents might be used to increase cardiac out- generally safe; a systematic review of observational stud-
put in specific situations. Phosphodiesterase inhibitors ies showed an incidence of limb ischemia and bleeding
increase intracellular cyclic AMP and thus have ino- to be less than 1%, with the most common complication
tropic effects independent of -adrenergic receptors. being localized hematoma (14%) [296]. Complication
The phosphodiesterase inhibitor milrinone was shown to rates may be lower if an ultrasound-guided technique is
increase cardiac output in one small randomized trial of used [297]. A recent systematic review showed higher
12 pediatric patients, but the trial was underpowered for risk of infections when femoral arterial catheters were
assessment of outcomes [290]. Levosimendan increases used compared to radial artery catheters (RR 1.93; 95%
cardiac myocyte calcium responsiveness and also opens CI 1.322.84), and the overall pooled incidence of blood-
ATP-dependent potassium channels, giving the drug stream infection was 3.4 per 1000 catheters [298]. Large
both inotropic and vasodilatory properties. Given the randomized trials that compare arterial blood pressure
potential role for abnormal calcium handling in sepsis- monitoring versus noninvasive methods are lacking.
induced myocardial depression, the use of levosimendan In view of the low complication rate and likely bet-
has been proposed in septic shock as well. In a trial of 35 ter estimation of blood pressure but potentially limited
resources in some countries, and the lack of high qual- performance and detection bias; 27 were at low risk of attri-
ity studies, the benefits of arterial catheters probably tion bias; and 14 were at low risk of selective reporting. Cor-
outweigh the risks. Therefore, we issued a weak recom- ticosteroids reduced 28-day mortality (27 trials; n = 3176;
mendation in favor of arterial catheter placement. Arte- RR 0.87; 95% CI 0.761.00). Treatment with a long course
rial catheters should be removed as soon as continuous of low-dose corticosteroids significantly reduced 28-day
hemodynamic monitoring is not required to minimize mortality (22 trials; RR 0.87; 95% CI 0.780.97). Corticos-
the risk of complications. teroids also reduced ICU mortality (13 trials; RR 0.82; 95%
CI 0.681.00) and in hospital mortality (17 trials; RR 0.85;
H. CORTICOSTEROIDS 95% CI 0.730.98). Corticosteroids increased the propor-
tion of shock reversal by day 7 (12 trials; RR 1.31; 95% CI
1. We suggest against using IV hydrocortisone 1.141.51) and by day 28 (seven trials; n=1013; RR 1.11;
to treat septic shock patients if adequate fluid 95% CI 1.021.21). Finally, an additional systematic review
resuscitation and vasopressor therapy are able by Volbeda etal. including a total of 35 trials randomizing
to restore hemodynamic stability. If this is not 4682 patients has been published (all but two trials had high
achievable, we suggest IV hydrocortisone at a risk of bias) [305]. Conversely, in this review, no statisti-
dose of 200 mg per day (weak recommendation, cally significant effect on mortality was found for any dose
low quality of evidence). of steroids versus placebo or for no intervention at maximal
follow-up. The two trials with low risk of bias also showed
Rationale The response of septic shock patients to no statistically significant difference (random-effects model
fluid and vasopressor therapy seems to be an important RR 0.38; 95% CI 0.062.42). Similar results were obtained in
factor in selection of patients for optional hydrocorti- subgroups of trials stratified according to hydrocortisone (or
sone therapy. One French multicenter RCT of patients equivalent) at high (>500mg) or low (500mg) doses [RR
in vasopressor-unresponsive septic shock (systolic blood 0.87; trial sequential analysis (TSA)-adjusted CI; 0.381.99;
pressure <90mm Hg despite fluid resuscitation and vaso- and RR 0.90; TSA-adjusted CI 0.491.67, respectively]. No
pressors for more than 1 h) showed significant shock statistically significant effects on serious adverse events
reversal and reduction of mortality rate in patients with other than mortality were reported (RR 1.02; TSA-adjusted
relative adrenal insufficiency [defined as a maximal post- CI 0.71.48). In the absence of convincing evidence of ben-
adrenocorticotropic hormone (ACTH) cortisol increase efit, we issue a weak recommendation against the use of cor-
9g/dL] [299]. Two smaller RCTs also showed signifi- ticosteroids to treat septic shock patients if adequate fluid
cant effects on shock reversal with steroid therapy [300, resuscitation and vasopressor therapy are able to restore
301]. In contrast, a large, European multicenter trial hemodynamic stability.
(CORTICUS) that enrolled patients with systolic blood In one study, the observation of a potential interac-
pressure of <90 mm Hg despite adequate fluid replace- tion between steroid use and ACTH test was not sta-
ment or need for vasopressors had a lower risk of death tistically significant [306]. Furthermore, no evidence of
than the French trial and failed to show a mortality ben- this distinction was observed between responders and
efit with steroid therapy [302]. There was no difference in nonresponders in a recent multicenter trial [302]. Ran-
mortality in groups stratified by ACTH response. dom cortisol levels may still be useful for absolute adre-
Several systematic reviews have examined the use of low- nal insufficiency; however, for septic shock patients who
dose hydrocortisone in septic shock with contradictory have relative adrenal insufficiency (no adequate stress
results. Annane etal. [299] analyzed the results of 12 stud- response), random cortisol levels have not been demon-
ies and calculated a significant reduction in 28-day mortality strated to be useful. Cortisol immunoassays may over-
with prolonged low-dose steroid treatment in adult septic or underestimate the actual cortisol level, affecting the
shock patients (RR 0.84; 95% CI 0.720.97; p=0.02). In par- assignment of patients to responders or nonresponders
allel, Sligl etal. [303] used a similar technique, but identified [307]. Although the clinical significance is not clear, it is
only eight studies for their meta-analysis, six of which had a now recognized that etomidate, when used for induction
high-level RCT design with low risk of bias. In contrast to for intubation, will suppress the hypothalamicpitui-
the aforementioned review, this analysis revealed no statis- taryadrenal axis [308, 309]. Moreover, a subanalysis of
tically significant difference in mortality (RR 1.00; 95% CI the CORTICUS trial revealed that the use of etomidate
0.841.18). Both reviews, however, confirmed the improved before application of low-dose steroids was associated
shock reversal by using low-dose hydrocortisone. More with an increased 28-day mortality rate [302].
recently, Annane etal. included 33 eligible trials (n=4268) There has been no comparative study between a fixed-
in a new systematic review [304]. Of these 33 trials, 23 duration and clinically guided regimen or between taper-
were at low risk of selection bias; 22 were at low risk of ing and abrupt cessation of steroids. Three RCTs used a
fixed-duration protocol for treatment [300, 302, 306], in two of the three treatment arms in the Protocol-Based
and therapy was decreased after shock resolution in two Care for Early Septic Shock (ProCESS) trial were a sub-
RCTs [301, 310]. In four studies, steroids were tapered part of a more comprehensive sepsis management strategy
over several days [300302, 310] and steroids were with- [18]. The EGDT group received transfusion at a hemato-
drawn abruptly in two RCTs [306, 311]. One crossover crit <30% (hemoglobin 10g/dL) when the Scvo2 was <70%
study showed hemodynamic and immunologic rebound after initial resuscitation interventions compared to the
effects after abrupt cessation of corticosteroids [312]. protocol-based standard care group that received blood
Further, one study revealed no difference in outcome of transfusion only when the hemoglobin was <7.5 g/dL.
septic shock patients if low-dose hydrocortisone is used No significant differences were found between the two
for 3 or 7days; hence, we suggest tapering steroids when groups for 60-day in-hospital mortality or 90-day mortal-
vasopressors are no longer needed [313]. ity. Although the ProCESS trial is a less direct assessment
Steroids may be indicated when there is a history of of blood transfusion therapy, it does provide important
steroid therapy or adrenal dysfunction, but whether information in regard to transfusion in the acute resuscita-
low-dose steroids have a preventive potency in reducing tive phase of sepsis. We judge the evidence to be high cer-
the incidence of sepsis and septic shock in critically ill tainty that there is little difference in mortality, and, if there
patients cannot be answered. A recent large multicenter is, that it would favor lower hemoglobin thresholds.
RCT demonstrated no reduction in the development of
septic shock in septic patients treated with hydrocorti- 2. We recommend against the use of erythropoietin
sone versus placebo [314]; steroids should not be used in for treatment of anemia associated with sepsis
septic patients to prevent septic shock. Additional studies (strong recommendation, moderate quality of evi-
are underway that may provide additional information to dence).
inform clinical practice.
Several randomized trials on the use of low-dose Rationale No specific information regarding eryth-
hydrocortisone in septic shock patients revealed a signifi- ropoietin use in septic patients is available, and clini-
cant increase of hyperglycemia and hypernatremia [306] cal trials of erythropoietin administration in critically
as side effects. A small prospective study demonstrated ill patients show a small decrease in red cell transfusion
that repetitive bolus application of hydrocortisone leads requirement with no effect on mortality [317, 318]. The
to a significant increase in blood glucose; this peak effect effect of erythropoietin in sepsis and septic shock would
was not detectable during continuous infusion. Further, not be expected to be more beneficial than in other criti-
considerable inter-individual variability was seen in this cal conditions. Erythropoietin administration may be
blood glucose peak after the hydrocortisone bolus [315]. associated with an increased incidence of thrombotic
Although an association of hyperglycemia and hyper- events in the critically ill. Patients with sepsis and septic
natremia with patient outcome measures could not be shock may have coexisting conditions that meet indica-
shown, good practice includes strategies for avoidance tions for the use of erythropoietin or similar agents.
and/or detection of these side effects.
3. We suggest against the use of fresh frozen plasma
I. BLOOD PRODUCTS to correct clotting abnormalities in the absence
of bleeding or planned invasive procedures (weak
1. We recommend that RBC transfusion occur only recommendation, very low quality of evidence).
when hemoglobin concentration decreases to
<7.0 g/dL in adults in the absence of extenuat- Rationale No RCTs exist related to prophylactic fresh
ing circumstances, such as myocardial ischemia, frozen plasma transfusion in septic or critically ill patients
severe hypoxemia, or acute hemorrhage (strong with coagulation abnormalities. Current recommendations
recommendation, high quality of evidence). are based primarily on expert opinion that fresh frozen
plasma be transfused when there is a documented defi-
Rationale Two clinical trials in septic patients evalu- ciency of coagulation factors (increased prothrombin time,
ated specific blood transfusion thresholds. The Transfu- international normalized ratio, or partial thromboplastin
sion Requirements In Septic Shock (TRISS) trial addressed time) and the presence of active bleeding or before surgi-
a transfusion threshold of 7 versus 9g/dL in septic shock cal or invasive procedures [319]. In addition, transfusion of
patients after admission to the ICU [316]. Results showed fresh frozen plasma usually fails to correct the prothrombin
similar 90-day mortality, ischemic events, and use of life time in nonbleeding patients with mild abnormalities. No
support in the two treatment groups with fewer transfu- studies suggest that correction of more severe coagulation
sions in the lower-threshold group. The hemoglobin targets abnormalities benefits patients who are not bleeding.
4. We suggest prophylactic platelet transfusion 95% CI 0.841.24). Similarly, Laupland etal. [331] found
when counts are <10,000/mm3 (10109/L) in the a significant reduction in mortality with the use of IVIg
absence of apparent bleeding and when counts are treatment (OR 0.66; 95% CI 0.530.83; p<0.005). When
<20,000/mm3 (20109/L) if the patient has a sig- only high-quality studies were pooled, the results were no
nificant risk of bleeding. Higher platelet counts longer statistically significant (OR 0.96); OR for mortality
[50,000/mm3 (50109/L)] are advised for active was 0.96 (95% CI 0.711.3; p=0.78). Two meta-analyses
bleeding, surgery, or invasive procedures (weak that used less strict criteria to identify sources of bias or
recommendation, very low quality of evidence). did not state their criteria for the assessment of study
quality found significant improvement in patient mortal-
Rationale No RCTs of prophylactic platelet transfu- ity with IVIg treatment [333335]. Finally, there are no
sion in septic or critically ill patients exist. Current rec- cutoffs for plasma IgG levels in septic patients, for which
ommendations and guidelines for platelet transfusion are substitution with IVIgG improves outcome data [334].
based on clinical trials of prophylactic platelet transfu- Most IVIg studies are small, and some have a high risk of
sion in patients with therapy-induced thrombocytopenia bias; the only large study (n=624) showed no effect [328].
(usually leukemia and stem cell transplant) [320327]. Subgroup effects between IgM-enriched and non-enriched
Thrombocytopenia in sepsis is likely due to a different formulations reveal significant heterogeneity. Indirectness
pathophysiology of impaired platelet production and and publication bias were considered, but not invoked in
increased platelet consumption. Factors that may increase grading this recommendation. The low certainty of evi-
the bleeding risk and indicate the need for a higher plate- dence led to the grading as a weak recommendation. The
let count are frequently present in patients with sepsis. statistical information that comes from the high-quality
trials does not support a beneficial effect of polyclonal
J. IMMUNOGLOBULINS IVIg. We encourage conduct of large multicenter studies
to further evaluate the effectiveness of other IV polyclonal
1. We suggest against the use of IV immunoglobu- immunoglobulin preparations in patients with sepsis.
lins in patients with sepsis or septic shock (weak
recommendation, low quality of evidence). K. BLOOD PURIFICATION

Rationale There were no new studies informing this 1. We make no recommendation regarding the use
guideline recommendation. One larger multicenter RCT of blood purification techniques.
(n = 624) [328] in adult patients found no benefit for
IV immunoglobulin (IVIg). The most recent Cochrane Rationale Blood purification includes various tech-
meta-analysis [329] differentiates between standard poly- niques, such as high-volume hemofiltration and hemoad-
clonal IV immunoglobulins (IVIgG) and immunoglobulin sorption (or hemoperfusion), where sorbents, removing
M-enriched polyclonal Ig (IVIgGM). In ten studies with either endotoxin or cytokines, are placed in contact with
IVIgG (1430 patients), mortality between 28 and 180days blood; plasma exchange or plasma filtration, through
was 29.6% in the IVIgG group and 36.5% in the placebo- which plasma is separated from whole blood, removed,
group (RR 0.81; 95% CI 0.700.93), and for the seven and replaced with normal saline, albumin, or fresh frozen
studies with IVIgGM (528 patients), mortality between 28 plasma; and the hybrid system: coupled plasma filtration
and 60days was 24.7% in the IVIgGM group and 37.5% in adsorption (CPFA), which combines plasma filtration and
the placebo-group (RR 0.66; 95% CI 0.510.85). The cer- adsorption by a resin cartridge that removes cytokines.
tainty of the studies was rated as low for the IVIgG trials, When these modalities of blood purification are con-
based on risk of bias and heterogeneity, and as moderate sidered versus conventional treatment, the available trials
for the IVIgGM trials, based on risk of bias. Comparable are, overall, small, unblinded, and with high risk of bias.
results were found in other meta-analyses [330]. However, Patient selection was unclear and differed with the various
after excluding low-quality trials, the recent Cochrane techniques. Hemoadsorption is the technique most largely
analysis [329] revealed no survival benefit. investigated, in particular with polymyxin B-immobilized
These findings are in accordance with those of two older polystyrene-derived fibers to remove endotoxin from the
meta-analyses [331, 332] from other Cochrane authors. blood. A recent meta-analysis demonstrated a favorable
One systematic review [332] included a total of 21 trials effect on overall mortality with this technique [336]. The
and showed a reduction in death with immunoglobu- composite effect, however, depends on a series of studies
lin treatment (RR 0.77; 95% CI 0.680.88); however, the performed in a single country (Japan), predominantly by
results of only high-quality trials (total of 763 patients) one group of investigators. A recent large RCT performed
did not show a statistically significant difference (RR 1.02; on patients with peritonitis related to organ perforation
within 12 h after emergency surgery found no benefit systematic reviews showed a potential survival benefit
of polymyxin B hemoperfusion on mortality and organ of heparin in patients with sepsis without an increase in
failure, as compared to standard treatment [337]. Illness major bleeding [344]. However, overall impact remains
severity of the study patients, however, was low overall, uncertain, and heparin cannot be recommended until
which makes these findings questionable. A multicenter further RCTs are performed.
blinded RCT is ongoing, which should provide stronger Recombinant activated protein C, which was originally
evidence regarding this technique [338]. recommended in the 2004 and 2008 SSC guidelines, was
Few RCTs evaluated plasma filtration, alone or com- not shown to be effective for adult patients with septic
bined with adsorption for cytokine removal (CPFA). A shock by the PROWESS-SHOCK trial, and was with-
recent RCT comparing CPFA with standard treatment drawn from the market [345].
was stopped for futility [339]. About half of the patients
randomized to CPFA were undertreated, primarily M. MECHANICAL VENTILATION
because of clotting of the circuit, which raises doubts
about CPFA feasibility. 1. We recommend using a target tidal volume of
In consideration of all these limitations, our con- 6mL/kg predicted body weight (PBW) compared
fidence in the evidence is very low either in favor of or with 12 mL/kg in adult patients with sepsis-
against blood purification techniques; therefore, we induced ARDS (strong recommendation, high
do not provide a recommendation. Further research is quality of evidence).
needed to clarify the clinical benefit of blood purification 2. We recommend using an upper limit goal for pla-
techniques. teau pressures of 30 cmH2O over higher plateau
pressures in adult patients with sepsis-induced
L. ANTICOAGULANTS severe ARDS (strong recommendation, moderate
quality of evidence).
1. We recommend against the use of antithrom-
bin for the treatment of sepsis and septic shock Rationale This recommendation is unchanged from
(strong recommendation, moderate quality of the previous guidelines. Of note, the studies that guide
evidence). the recommendations in this section enrolled patients
using criteria from the AmericanEuropean Consen-
Rationale Antithrombin is the most abundant antico- sus Criteria Definition for Acute Lung Injury and ARDS
agulant circulating in plasma. The decrease of its plasma [346]. For the current document, we used the 2012 Berlin
activity at onset of sepsis correlates with disseminated intra- definition and the terms mild, moderate, and severe ARDS
vascular coagulation (DIC) and lethal outcome. However, a (Pao2/Fio2 300, 200, and 100mm Hg, respectively)
phase III clinical trial of high-dose antithrombin for adults [347]. Several multicenter randomized trials have been
with sepsis and septic shock as well as systematic reviews of performed in patients with established ARDS to evaluate
antithrombin for critically ill patients did not demonstrate the effects of limiting inspiratory pressure through mod-
any beneficial effect on overall mortality. Antithrombin was eration of tidal volume [348351]. These studies showed
associated with an increased risk of bleeding [340, 341]. differing results, which may have been caused by differ-
Although post hoc subgroup analyses of patients with sep- ences in airway pressures in the treatment and control
sis associated with DIC showed better survival in patients groups [347, 351, 353]. Several meta-analyses suggest
receiving antithrombin, this agent cannot be recommended decreased mortality in patients with a pressure- and vol-
until further clinical trials are performed. ume-limited strategy for established ARDS [353, 354].
The largest trial of a volume- and pressure-limited
2. We make no recommendation regarding the use strategy showed 9% absolute decrease in mortality in
of thrombomodulin or heparin for the treatment ARDS patients ventilated with tidal volumes of 6mL/kg
of sepsis or septic shock. compared with 12 mL/kg PBW, and aiming for plateau
pressure 30 cmH2O [350]. The use of lung-protective
Rationale Most RCTs of recombinant soluble throm- strategies for patients with ARDS is supported by clini-
bomodulin have been targeted for sepsis associated cal trials and has been widely accepted; however, the pre-
with DIC, and a systematic review suggested a beneficial cise tidal volume for an individual ARDS patient requires
effect on survival without an increase of bleeding risk adjustment for factors such as the plateau pressure, the
[342, 343]. A phase III RCT is ongoing for sepsis associ- selected positive end-expiratory pressure (PEEP), thora-
ated with DIC. The guideline panel has elected to make coabdominal compliance, and the patients breathing
no recommendation pending these new results. Two effort. Patients with profound metabolic acidosis, high
minute ventilation, or short stature may require addi- plateau pressures are in use. Three large multicenter tri-
tional manipulation of tidal volumes. Some clinicians als and a pilot trial using higher versus lower levels of
believe it may be safe to ventilate with tidal volumes PEEP in conjunction with low tidal volumes did not
>6mL/kg PBW as long as plateau pressure can be main- show benefit or harm [363366]. A patient-level meta-
tained 30cmH2O [355, 356]. The validity of this ceiling analysis showed no benefit in all patients with ARDS;
value will depend on the patients effort, because those however, patients with moderate or severe ARDS (Pao2/
who are actively breathing generate higher transpulmo- Fio2 200 mm Hg) had decreased mortality with the
nary pressures for a given plateau pressure than patients use of higher PEEP, whereas those with mild ARDS did
who are passively inflated. Conversely, patients with not [367]. A patient-level analysis of two of the rand-
very stiff chest/abdominal walls and high pleural pres- omized PEEP trials suggested a survival benefit if Pao2/
sures may tolerate plateau pressures >30cmH2O because Fio2 increased with higher PEEP and harm if Pao2/Fio2
transpulmonary pressures will be lower. A retrospective fell [368]. A small randomized trial suggested that adjust-
study suggested that tidal volumes should be lowered ing PEEP to obtain a positive transpulmonary pressure as
even with plateau pressures 30 cmH2O [357] because estimated by esophageal manometry improved outcomes;
lower plateau pressures were associated with reduced a confirmatory trial is underway [369]. An analysis of
hospital mortality [358]. A recent patient-level media- nearly all the randomized trials of lung-protective ventila-
tion analysis suggested that a tidal volume that results tion suggested a benefit of higher PEEP if driving pressure
in a driving pressure (plateau pressure minus set PEEP) fell with increased PEEP, presumably indicating increased
below 1215 cmH2O may be advantageous in patients lung compliance from opening of lung units [359].
without spontaneous breathing efforts [359]. Prospective While moderate-quality evidence suggests that higher
validation of tidal volume titration by driving pressure is PEEP improves outcomes in moderate to severe ARDS,
needed before this approach can be recommended. the optimal method for selecting a higher PEEP level is
High tidal volumes coupled with high plateau pressures unclear. One option is to titrate PEEP according to bed-
should be avoided in ARDS. Clinicians should use as a start- side measurements of thoracopulmonary compliance
ing point the objective of reducing tidal volume over 12h with the objective of obtaining the best compliance or
from its initial value toward the goal of a low tidal volume lowest driving pressure, reflecting a favorable balance
(6 mL/kg PBW) achieved in conjunction with an end- of lung recruitment and overdistension [370]. The sec-
inspiratory plateau pressure 30 cmH2O. If plateau pres- ond option is to titrate PEEP upward on a tidal volume
sure remains >30cmH2O after reduction of tidal volume to of 6 mL/kg PBW until the plateau airway pressure is
6 mL/kg PBW, tidal volume may be further reduced to as 28 cmH2O [365]. A third option is to use a PEEP/Fio2
low as 4mL/kg PBW. Respiratory rate should be increased titration table that titrates PEEP based on the combina-
to a maximum of 35 breaths/min during tidal volume reduc- tion of Fio2 and PEEP required to maintain adequate
tion to maintain minute ventilation. Volume- and pressure- oxygenation [350, 363365, 368]. A PEEP >5 cmH2O is
limited ventilation may lead to hypercapnia even with these usually required to avoid lung collapse [371].
maximum tolerated set respiratory rates; this appears to be
tolerated and safe in the absence of contraindications (e.g., 4. We suggest using recruitment maneuvers in adult
high intracranial pressure, sickle cell crisis). patients with sepsis-induced, severe ARDS (weak
No single mode of ventilation (pressure control, vol- recommendation, moderate quality of evidence).
ume control) has consistently been shown to be advan-
tageous when compared with any other that respects the Rationale Many strategies exist for treating refractory
same principles of lung protection. hypoxemia in patients with severe ARDS [372]. Tem-
porarily raising transpulmonary pressure may facilitate
3. We suggest using higher PEEP over lower PEEP opening atelectatic alveoli to permit gas exchange [371],
in adult patients with sepsis-induced moderate to but could also overdistend aerated lung units, leading to
severe ARDS (weak recommendation, moderate ventilator-induced lung injury and transient hypoten-
quality of evidence). sion. The application of sustained continuous positive
airway pressure (CPAP) appears to improve survival (RR
Rationale Raising PEEP in ARDS may open lung units 0.84; 95% CI 0.740.95) and reduce the occurrence of
to participate in gas exchange. This may increase Pao2 severe hypoxia requiring rescue therapy (RR 0.76; 95% CI
when PEEP is applied through either an endotracheal 0.411.40) in patients with ARDS. Although the effects
tube or a face mask [360362]. In animal experiments, of recruitment maneuvers improve oxygenation ini-
avoidance of end-expiratory alveolar collapse helps mini- tially, the effects can be transient [373]. Selected patients
mize ventilator-induced lung injury when relatively high with severe hypoxemia may benefit from recruitment
maneuvers in conjunction with higher levels of PEEP, in patients randomized to HFOV. An increase in baro-
but little evidence supports the routine use in all ARDS trauma was seen in patients receiving HFOV (RR 1.19;
patients [373]. Any patient receiving this therapy should 95% CI 0.831.72); however, this was based on very low-
be monitored closely and recruitment maneuvers discon- quality evidence.
tinued if deterioration in clinical variables is observed. The role of HFOV as a rescue technique for refractory
ARDS remains unclear; however, we recommend against
5. We recommend using prone over supine position its early use in moderate-severe ARDS given the lack of
in adult patients with sepsis-induced ARDS and demonstrated benefit and a potential signal for harm.
a Pao2/Fio2 ratio <150 (strong recommendation,
moderate quality of evidence). 7. We make no recommendation regarding the use
of noninvasive ventilation (NIV) for patients with
Rationale In patients with ARDS and a Pao2/Fio2 ratio sepsis-induced ARDS.
<150, the use of prone compared with supine position
within the first 36 h of intubation, when performed for Rationale NIV may have theoretical benefits in
>16h a day, showed improved survival [374]. Meta-anal- patients with sepsis-induced respiratory failure, such
ysis including this study demonstrated reduced mortal- as better communication abilities, reduced need for
ity in patients treated with prone compared with supine sedation, and avoidance of intubation. However, NIV
position (RR 0.85; 95% CI 0.711.01) as well as improved may preclude the use of low tidal volume ventilation or
oxygenation as measured by change in Pao2/Fio2 ratio achieving adequate levels of PEEP, two ventilation strat-
(median 24.03 higher, 95% CI 13.334.7 higher) [375]. egies that have shown benefit even in mild-moderate
Most patients respond to the prone position with ARDS [365, 386]. Also, in contrast to indications such
improved oxygenation and may also have improved lung as cardiogenic pulmonary edema or chronic obstructive
compliance [374, 376379]. While prone position may be pulmonary disease exacerbation where NIV use is brief,
associated with potentially life-threatening complications ARDS often takes days or weeks to improve, and pro-
including accidental removal of the endotracheal tube, longed NIV use may lead to complications such as facial
this was not evident in pooled analysis (RR 1.09; 95% CI skin breakdown, inadequate nutritional intake, and fail-
0.851.39). However, prone position was associated with ure to rest respiratory muscles.
an increase in pressure sores (RR 1.37; 95% CI 1.051.79) A few small RCTs have shown benefit with NIV for
[375], and some patients have contraindications to the early or mild ARDS or de novo hypoxic respiratory fail-
prone position [374]. ure; however, these were in highly selected patient popu-
In patients with refractory hypoxia, alternative strate- lations [387, 388]. More recently, a larger RCT in patients
gies, including airway pressure release ventilation and with hypoxemic respiratory failure compared NIV to
extracorporeal membrane oxygenation, may be con- traditional oxygen therapy or high-flow nasal cannula
sidered as rescue therapies in experienced centers [372, [389]. This study demonstrated improved 90-day survival
380383]. with high-flow oxygen compared with standard therapy
or NIV; however, the NIV technique was not standard-
6. We recommend against using high-frequency ized and the experience of the centers varied. Although
oscillatory ventilation (HFOV) in adult patients high-flow oxygen has not been addressed here, it is pos-
with sepsis-induced ARDS (strong recommenda- sible that this technique may play a more prominent role
tion, moderate quality of evidence). in the treatment of hypoxic respiratory failure and ARDS
moving forward.
Rationale HFOV has theoretical advantages that make Given the uncertainty regarding whether clinicians can
it an attractive ventilator mode for patients with ARDS. identify ARDS patients in whom NIV might be benefi-
Two large RCTs evaluating routine HFOV in moderate- cial, we have not made a recommendation for or against
severe ARDS have been recently published [384, 385]. this intervention. If NIV is used for patients with ARDS,
One trial was stopped early because the mortality was we suggest close monitoring of tidal volumes.
higher in patients randomized to HFOV [384]. Including
these recent studies, a total of five RCTs (1580 patients) 8. We suggest using neuromuscular blocking agents
have examined the role of HFOV in ARDS. Pooled analy- (NMBAs) for 48h in adult patients with sepsis-
sis demonstrates no effect on mortality (RR 1.04; 95% induced ARDS and a Pao2/Fio2 ratio <150mm Hg
CI 0.831.31) and an increased duration of mechani- (weak recommendation, moderate quality of evi-
cal ventilation (MD, 1.1 days higher; 95% CI 0.032.16) dence).
 Rationale The most common indication for NMBA (strong recommendation, moderate quality of evi-
use in the ICU is to facilitate mechanical ventilation dence).
[390]. When appropriately used, these agents may
improve chest wall compliance, prevent respiratory dys- Rationale Mechanisms for the development of pul-
synchrony, and reduce peak airway pressures [391]. Mus- monary edema in patients with ARDS include increased
cle paralysis may also reduce oxygen consumption by capillary permeability, increased hydrostatic pressure,
decreasing the work of breathing and respiratory muscle and decreased oncotic pressure [403]. Small prospec-
blood flow [392]. However, a placebo-controlled RCT tive studies in patients with critical illness and ARDS
in patients with severe sepsis demonstrated that oxy- have suggested that low weight gain is associated with
gen delivery, oxygen consumption, and gastric intramu- improved oxygenation [404] and fewer days of mechani-
cosal pH were not improved during deep neuromuscular cal ventilation [405, 406]. A fluid-conservative strategy
blockade [393]. to minimize fluid infusion and weight gain in patients
An RCT of continuous infusions of cisatracurium in with ARDS, based on either a CVP or a pulmonary
patients with early ARDS and a Pao2/Fio2 <150 mm artery (PA) catheter (PA wedge pressure) measure-
Hg showed improved adjusted survival rates and more ment, along with clinical variables to guide treatment,
organ failure-free days without an increased risk in led to fewer days of mechanical ventilation and reduced
ICU-acquired weakness compared with placebo-treated ICU LOS without altering the incidence of renal failure
patients [394]. The investigators used a high fixed dose of or mortality rates [407]. This strategy was only used in
cisatracurium without train-of-four monitoring; half of patients with established ARDS, some of whom had
the patients in the placebo group received at least a sin- shock during their ICU stay, and active attempts to
gle NMBA dose. Of note, groups in both the intervention reduce fluid volume were conducted only outside peri-
and control groups were ventilated with volume-cycled ods of shock.
and pressure-limited mechanical ventilation. Although
many of the patients in this trial appeared to meet sep- 10. We recommend against the use of -2 agonists
sis criteria, it is not clear whether similar results would for the treatment of patients with sepsis-induced
occur in sepsis patients or in patients ventilated with ARDS without bronchospasm (strong recom-
alternate modes. Pooled analysis including three trials mendation, moderate quality of evidence).
that examined the role of NMBAs in ARDS, including the
one above, showed improved survival (RR 0.72; 95% CI Rationale Patients with sepsis-induced ARDS often
0.580.91) and a decreased frequency of barotrauma (RR develop increased vascular permeability; preclinical data
0.43; 95% CI 0.200.90) in those receiving NMBAs [395]. suggest that -adrenergic agonists may hasten resorption
An association between NMBA use and myopathies of alveolar edema [408]. Three RCTs (646 patients) evalu-
and neuropathies has been suggested by case studies ated -agonists in patients with ARDS [408410]. In two
and prospective observational studies in the critical care of these trials, salbutamol (15g/kg of ideal body weight)
population [391, 396399], but the mechanisms by which delivered intravenously [408, 409] was compared with
NMBAs produce or contribute to myopathies and neu- placebo, while the third trial compared inhaled albuterol
ropathies in these patients are unknown. Pooled analysis versus placebo [410]. Group allocation was blinded in all
of the RCT data did not show an increase in neuromus- three trials, and two trials were stopped early for futil-
cular weakness in those who received NMBAs (RR 1.08; ity or harm [409411]. More than half of the patients
95% CI 0.831.41); however, this was based on very low enrolled in all three trials had pulmonary or non-pulmo-
quality of evidence [395]. Given the uncertainty that nary sepsis as the cause of ARDS.
still exists pertaining to these important outcomes and Pooled analysis suggests -agonists may reduce survival
the balance between benefits and potential harms, the to hospital discharge in ARDS patients (RR 1.22; 95% CI
panel decided that a weak recommendation was most 0.951.56) while significantly decreasing the number of
suitable. If NMBAs are used, clinicians must ensure ade- ventilator-free days (MD, 2.19; 95% CI 3.68 to 0.71)
quate patient sedation and analgesia [400, 401]; recently [412]. -Agonist use also led to more arrhythmias (RR
updated clinical practice guidelines are available for spe- 1.97; 95% CI 0.705.54) and more tachycardia (RR 3.95;
cific guidance [402]. 95% CI 1.4111.06).
-2 agonists may have specific indications in the criti-
9. We recommend a conservative fluid strategy for cally ill, such as the treatment of bronchospasm and
patients with established sepsis-induced ARDS hyperkalemia. In the absence of these conditions, we
who do not have evidence of tissue hypoperfusion recommend against the use of -agonists, either in IV or
aerosolized form, for the treatment of patients with sep- respiratory failure, shorten LOS, and result in fewer post-
sis-induced ARDS. operative episodes of sepsis [423]. Subgroup analysis of
only the studies that enrolled critically ill patients [424]
11. We recommend against the routine use of the PA suggests similar benefits of low tidal volume ventilation
catheter for patients with sepsis-induced ARDS on duration of mechanical ventilation and development
(strong recommendation, high quality of evi- of ARDS, but is further limited by imprecision given the
dence). small number of studies included. Despite these method-
ologic concerns, the benefits of low tidal volume ventila-
Rationale This recommendation is unchanged from tion in patients without ARDS are thought to outweigh
the previous guidelines. Although insertion of a PA cath- any potential harm. Planned RCTs may inform future
eter may provide useful information regarding volume practice.
status and cardiac function, these benefits may be con-
founded by differences in interpretation of the results 13. We recommend that mechanically ventilated sep-
[413, 414], poor correlation of PA occlusion pressures sis patients be maintained with the head of the
with clinical response [415], and lack of a PA cathe- bed elevated between 30 and 45 to limit aspira-
ter-based strategy demonstrated to improve patient tion risk and to prevent the development of VAP
outcomes [416]. Pooled analysis of two multicenter rand- (strong recommendation, low quality of evidence).
omized trials, one with 676 patients with shock or ARDS
[417] and another with 1000 patients with ARDS [418], Rationale The semi-recumbent position has been
failed to show any benefit associated with PA catheter demonstrated to decrease the incidence of VAP [425].
use on mortality (RR 1.02; 95% CI 0.961.09) or ICU Enteral feeding increased the risk of developing VAP;
LOS (mean difference 0.15days longer; 95% CI 0.74days 50% of the patients who were fed enterally in the supine
fewer1.03 days longer) [407, 419421] This lack of position developed VAP, compared with 9% of those fed
demonstrated benefit must be considered in the context in the semi-recumbent position [425]. However, the bed
of the increased resources required. Notwithstanding, position was monitored only once a day, and patients
selected sepsis patients may be candidates for PA catheter who did not achieve the desired bed elevation were not
insertion if management decisions depend on informa- included in the analysis [425]. One study did not show a
tion solely obtainable from PA catheter measurements. difference in incidence of VAP between patients main-
tained in supine and semi-recumbent positions [426];
12. We suggest using lower tidal volumes over higher patients assigned to the semi-recumbent group did not
tidal volumes in adult patients with sepsis- consistently achieve the desired head-of-bed eleva-
induced respiratory failure without ARDS (weak tion, and the head-of-bed elevation in the supine group
recommendation, low quality of evidence). approached that of the semi-recumbent group by day 7
[426]. When necessary, patients may be laid flat when
Rationale Low tidal volume ventilation (46 mL/kg) indicated for procedures, hemodynamic measurements,
has been shown to be beneficial in patients with estab- and during episodes of hypotension. Patients should not
lished ARDS [422] by limiting ventilator-induced lung be fed enterally while supine. There were no new pub-
injury. However, the effect of volume- and pressure-lim- lished studies since the last guidelines that would inform
ited ventilation is less clear in patients with sepsis who a change in the strength of the recommendation for the
do not have ARDS. Meta-analysis demonstrates the ben- current iteration. The evidence profile for this recom-
efits of low tidal volume ventilation in patients without mendation demonstrated low quality of evidence. The
ARDS, including a decrease in the duration of mechani- lack of new evidence, along with the low harms of head-
cal ventilation (MD, 0.64 days fewer; 95% CI 0.490.79) of-bed and high feasibility of implementation given the
and the decreased development of ARDS (RR 0.30; 95% frequency of the practice resulted in the strong recom-
CI 0.160.57) with no impact on mortality (RR 0.95; 95% mendation. There is a small subgroup of patients, such as
CI 0.641.41). Importantly, the certainty in this data is trauma patients with a spine injury, for whom this rec-
limited by indirectness because the included studies var- ommendation would not apply.
ied significantly in terms of populations enrolled, mostly
examining perioperative patients and very few focusing 14. We recommend using spontaneous breathing tri-
on ICU patients. The use of low tidal volumes in patients als in mechanically ventilated patients with sepsis
who undergo abdominal surgery, which may include sep- who are ready for weaning (strong recommenda-
sis patients, has been shown to decrease the incidence of tion, high quality of evidence).
 Rationale Spontaneous breathing trial options include depends on the existing local culture and practice [435,
a low level of pressure support, CPAP (5cmH2O), or use 436]. Another option for systematically limiting the use of
of a T-piece. A recently published clinical practice guide- sedation is the administration of intermittent rather than
line suggests the use of inspiratory pressure augmentation continuous sedation [437, 438]. Daily sedation interrup-
rather than T-piece or CPAP for an initial spontaneous tion (DSI) was associated with improved outcomes in a
breathing trial for acutely hospitalized adults on mechani- single-center randomized trial compared with usual care
cal ventilation for more than 24h [427]. Daily spontane- [430]; however, in a multicenter RCT there was no advan-
ous breathing trials in appropriately selected patients tage to DSI when patients were managed with a sedation
reduce the duration of mechanical ventilation and wean- protocol, and nurses perceived a higher workload [439].
ing duration both in individual trials as well as with A recent Cochrane meta-analysis did not find strong evi-
pooled analysis of the individual trials [428430]. These dence that DSI alters the duration of mechanical ventila-
breathing trials should be conducted in conjunction with tion, mortality, ICU or hospital LOS, adverse event rates,
a spontaneous awakening trial [431]. Successful comple- or drug consumption for critically ill adults receiving
tion of spontaneous breathing trials leads to a high like- mechanical ventilation compared to sedation strategies
lihood of successful early discontinuation of mechanical that do not include DSI; however, interpretation of the
ventilation with minimal demonstrated harm. results is limited by imprecision and clinical heterogeneity
[440]. Another strategy is the primary use of opioids alone
15. We recommend using a weaning protocol in and avoidance of sedatives, which was shown to be feasi-
mechanically ventilated patients with sepsis- ble in the majority of ventilated patients in a single-center
induced respiratory failure who can tolerate trial, and was associated with more rapid liberation from
weaning (strong recommendation, moderate mechanical ventilation [441]. Finally, the use of short-
quality of evidence). acting drugs such as propofol and dexmedetomidine may
result in better outcomes than the use of benzodiazepines
Rationale Protocols allow for standardization of clini- [442444]. Recent pain, agitation, and delirium guidelines
cal pathways to facilitate desired treatment [432]. These provide additional detail on implementation of sedation
protocols may include both spontaneous breathing trials, management, including nonpharmacologic approaches
gradual reduction of support, and computer-generated for the management of pain, agitation, and delirium [445].
weaning. Pooled analysis demonstrates that patients Regardless of approach, a large body of indirect evidence
treated with protocolized weaning compared with usual is available demonstrating the benefit of limiting sedation
care experienced shorter weaning duration (39 h; 95% in those requiring mechanical ventilation and without
CI 67h to 11h), and shorter ICU LOS (9h; 95% CI contraindication. As such, this should be best practice for
15 to 2). There was no difference between groups in any critically ill patient, including those with sepsis.
ICU mortality (OR 0.93; 95% CI 0.581.48) or need for
reintubation (OR 0.74; 95% CI 0.441.23) [428]. O. GLUCOSE CONTROL

N. SEDATION AND ANALGESIA 1. We recommend a protocolized approach to blood

glucose management in ICU patients with sepsis,
1. We recommend that continuous or intermittent commencing insulin dosing when two consecu-
sedation be minimized in mechanically venti- tive blood glucose levels are >180 mg/dL. This
lated sepsis patients, targeting specific titration approach should target an upper blood glucose
end points (BPS). level 180 mg/dL rather than an upper target
blood glucose level 110 mg/dL (strong recom-
Rationale Limiting the use of sedation in critically ill mendation, high quality of evidence).
ventilated patients reduces the duration of mechanical 2. We recommend that blood glucose values be
ventilation and ICU and hospital LOS, and allows ear- monitored every 12 h until glucose values and
lier mobilization [433, 434]. While these data arise from insulin infusion rates are stable, then every 4 h
studies performed in a wide range of critically ill patients, thereafter in patients receiving insulin infusions
there is little reason to believe that septic patients will not (BPS).
derive the same benefits from sedation minimization. 3. We recommend that glucose levels obtained with
Several strategies have been shown to reduce sedative point-of-care testing of capillary blood be inter-
use and the duration of mechanical ventilation. Nurse- preted with caution because such measurements
directed protocols that incorporate a sedation scale may not accurately estimate arterial blood or
likely result in improved outcomes; however, the benefit plasma glucose values (BPS).
 4. We suggest the use of arterial blood rather than associated with higher mortality [466471]. The continu-
capillary blood for point-of-care testing using ation of insulin infusions, especially with the cessation of
glucose meters if patients have arterial catheters nutrition, has been identified as a risk factor for hypo-
(weak recommendation, low quality of evidence). glycemia [454]. Balanced nutrition may be associated
with a reduced risk of hypoglycemia [472]. Hyperglyce-
Rationale A large single-center RCT in 2001 dem- mia and glucose variability seem to be unassociated with
onstrated a reduction in ICU mortality with intensive increased mortality rates in diabetic patients compared
IV insulin (Leuven protocol) targeting blood glucose to to nondiabetic patients [473475]. Patients with diabe-
80110mg/dL [446]. A second randomized trial of inten- tes and chronic hyperglycemia, end-stage renal failure, or
sive insulin therapy using the Leuven protocol enrolled medical versus surgical ICU patients may require higher
medical ICU patients with an anticipated ICU LOS of blood glucose ranges [476, 477].
more than three days in three medical ICUs; overall mor- Several factors may affect the accuracy and reproduc-
tality was not reduced [447]. ibility of point-of-care testing of blood capillary blood
Since these studies [446, 447] appeared, several RCTs glucose, including the type and model of the device used,
[448455] and meta-analyses [456462] of intensive user expertise, and patient factors, including hemato-
insulin therapy have been performed. The RCTs studied crit (false elevation with anemia), Pao2, and drugs [478].
mixed populations of surgical and medical ICU patients Plasma glucose values by capillary point-of-care test-
and found that intensive insulin therapy did not signifi- ing have been found to be potentially inaccurate, with
cantly decrease mortality, whereas the NICE-SUGAR frequent false elevations [479481] over the range of
trial demonstrated an increased mortality [451]. All stud- glucose levels, but especially in the hypoglycemic and
ies reported a much higher incidence of severe hypo- hyperglycemic ranges [482] and in shock patients (receiv-
glycemia (glucose 40 mg/dL) (629%) with intensive ing vasopressors) [478, 480]. A review of studies found
insulin therapy. Several meta-analyses confirmed that the accuracy of glucose measurements by arterial blood
intensive insulin therapy was not associated with a mor- gas analyzers and glucose meters by using arterial blood
tality benefit in surgical, medical, or mixed ICU patients. significantly higher than measurements with glucose
The meta-analysis by Song et al. [462] evaluated only meters using capillary blood [480].
septic patients and found that intensive insulin therapy The U.S. Food and Drug Administration has stated that
did not change 28- or 90-day mortality, but was associ- critically ill patients should not be tested with a glucose
ated with a higher incidence of hypoglycemia. The trig- meter because results may be inaccurate, and Centers for
ger to start an insulin protocol for blood glucose levels Medicare and Medicaid Services have plans to enforce
>180 mg/dL with an upper target blood glucose level the prohibition of off-label use of point-of-care capillary
<180mg/dL derives from the NICE-SUGAR trial, which blood glucose monitor testing in critically ill patients
used these values for initiating and stopping therapy. The [483]. Several medical experts have stated the need for
NICE-SUGAR trial is the largest, most compelling study a moratorium on this plan [484]. Despite the attempt to
to date on glucose control in ICU patients given its inclu- protect patients from harm because of inaccurate capil-
sion of multiple ICUs and hospitals and a general patient lary blood testing, a prohibition might cause more harm
population. Several medical organizations, including because a central laboratory test may take significantly
the American Association of Clinical Endocrinologists, longer to provide results than point-of-care glucometer
American Diabetes Association, American Heart Asso- testing.
ciation, American College of Physicians, and Society A review of 12 published insulin infusion protocols
of Critical Care Medicine, have published consensus for critically ill patients showed wide variability in dose
statements for glycemic control of hospitalized patients recommendations and variable glucose control [485].
[463, 465]. These statements usually targeted glucose This lack of consensus about optimal dosing of IV insu-
levels between 140 and 180 mg/dL. Because there is no lin may reflect variability in patient factors (severity of
evidence that targets between 140 and 180 mg/dL are illness, surgical versus medical settings), or practice pat-
different from targets of 110140 mg/dL, the present terns (e.g., approaches to feeding, IV dextrose) in the
recommendations use an upper target blood glucose environments in which these protocols were developed
180mg/dL without a lower target other than hypogly- and tested. Alternatively, some protocols may be more
cemia. Stricter ranges, such as 110140 mg/dL, may be effective than others, a conclusion supported by the wide
appropriate for selected patients if this can be achieved variability in hypoglycemia rates reported with protocols.
without significant hypoglycemia [463, 465]. Treatment Thus, the use of established insulin protocols is impor-
should avoid hyperglycemia (>180 mg/dL), hypoglyce- tant not only for clinical care, but also for the conduct of
mia, and wide swings in glucose levels that have been clinical trials to avoid hypoglycemia, adverse events, and
premature termination of trials before the efficacy sig- support of continuous therapies in sepsis independent of
nal, if any, can be determined. Several studies have sug- renal replacement needs.
gested that computer-based algorithms result in tighter For this revision of the guidelines, no additional RCTs
glycemic control with a reduced risk of hypoglycemia evaluating the hemodynamic tolerance of continuous
[486, 487]. Computerized decision support systems and versus intermittent RRT were identified. Accordingly, the
fully automated closed-loop systems for glucose control limited and inconsistent evidence persists. Two prospec-
are feasible, but not yet standard care. Further study of tive trials [497, 502] have reported a better hemodynamic
validated, safe, and effective protocols and closed-loop tolerance with continuous treatment, with no improve-
systems for controlling blood glucose concentrations and ment in regional perfusion [502] and no survival benefit
variability in the sepsis population is needed. [497]. Four other studies did not find any significant dif-
ference in MAP or drop in systolic pressure between the
P. RENAL REPLACEMENT THERAPY two methods [498, 500, 501, 503]. Two studies reported
a significant improvement in goal achievement with
1. We suggest that either continuous RRT (CRRT) continuous methods [497, 499] regarding fluid balance
or intermittent RRT be used in patients with sep- management.
sis and acute kidney injury (weak recommenda- Two additional RCTs reporting the effect of dose of
tion, moderate quality of evidence). CRRT on outcomes in patients with acute renal fail-
2. We suggest using CRRT to facilitate management ure were identified in the current literature review [504,
of fluid balance in hemodynamically unstable 505]. Both studies enrolled patients with sepsis and acute
septic patients (weak recommendation, very low kidney injury and did not demonstrate any difference
quality of evidence). in mortality associated with a higher dose of RRT. Two
3. We suggest against the use of RRT in patients large, multicenter, randomized trials comparing the dose
with sepsis and acute kidney injury for increase of renal replacement (Acute Renal Failure Trial Network
in creatinine or oliguria without other definitive in the United States and RENAL Study in Australia and
indications for dialysis (weak recommendation, New Zealand) also failed to show benefit of more aggres-
low quality of evidence). sive renal replacement dosing [506, 507]. A meta-anal-
ysis of the sepsis patients included in all relevant RCTs
Rationale Although numerous nonrandomized stud- (n=1505) did not demonstrate any significant relation-
ies have reported a nonsignificant trend toward improved ship between dose and mortality; the point estimate,
survival using continuous methods [488494], two however, favors CRRT doses >30 mL/kg/h. Because of
meta-analyses [495, 496] reported the absence of signifi- risk of bias, inconsistency, and imprecision, confidence in
cant differences in hospital mortality between patients the estimate is very low; further research is indicated. A
who receive CRRT and intermittent RRT. This absence typical dose for CRRT would be 2025mL/kg/h of efflu-
of apparent benefit of one modality over the other per- ent generation.
sists even when the analysis is restricted to RCTs [496]. One small trial from 2002 [504] evaluated early ver-
To date, five prospective RCTs have been published sus late or delayed initiation of RRT; it included only
[497501]; four found no significant difference in mor- four patients with sepsis and did not show any benefit
tality [497, 498, 500, 501], whereas one found signifi- of early CRRT. Since then, two relevant RCTs [508, 509]
cantly higher mortality in the continuous treatment were published in 2016. Results suggest the possibil-
group [499]; but imbalanced randomization had led to ity of either benefit [509] or harm [508] for mortality,
a higher baseline severity of illness in this group. When increased use of dialysis, and increased central line
a multivariable model was used to adjust for severity of infections with early RRT. Enrollment criteria and tim-
illness, no difference in mortality was apparent between ing of initiation of RRT differed in the two trials. Results
the groups. Most studies comparing modes of RRT in the were judged to be of low certainty based on indirectness
critically ill have included a small number of outcomes (many nonseptic patients) and imprecision for mortal-
and had a high risk of bias (e.g., randomization failure, ity. The possibility of harm (e.g., central line infections)
modifications of therapeutic protocol during the study pushes the balance of risk and benefit against early ini-
period, combination of different types of CRRT, small tiation of RRT. Meanwhile, the undesirable effects and
number of heterogeneous groups of enrollees). The most costs appear to outweigh the desirable consequences;
recent and largest RCT [501] enrolled 360 patients and therefore, we suggest not using RRT in patients with
found no significant difference in survival between the sepsis and acute kidney injury for increase in creati-
continuous and intermittent groups. We judged the over- nine or oliguria without other definitive indications for
all certainty of the evidence to be moderate and not in dialysis.
Q. BICARBONATE THERAPY Rationale ICU patients are at risk for deep vein throm-
bosis (DVT) as well as pulmonary embolism (PE). The
1. We suggest against the use of sodium bicarbo- incidence of DVT acquired in the ICU may be as high
nate therapy to improve hemodynamics or to as 10% [512]; the incidence of acquired PE may be 24%
reduce vasopressor requirements in patients [513, 514]. Patients with sepsis and septic shock are likely
with hypoperfusion-induced lactic acidemia with at increased risk for this complication. Vasopressor use,
pH 7.15 (weak recommendation, moderate which is frequent in these patients, has been found to be
quality of evidence). an independent risk factor for ICU-acquired DVT.
A meta-analysis of pharmacologic prophylaxis with UFH
Rationale Although sodium bicarbonate therapy may or LMWH in critically ill patients showed significant reduc-
be useful in limiting tidal volume in ARDS in some situ- tions in both DVT and PE, with no significant increase in
ations of permissive hypercapnia, no evidence supports bleeding complications. Mortality was lower in the patients
the use of sodium bicarbonate therapy in the treatment receiving prophylaxis, although this did not reach statistical
of hypoperfusion-induced lactic acidemia associated significance [514]. All studies included in the meta-analy-
with sepsis. Two blinded, crossover RCTs that compared sis were cited in the 2012 guideline, which recommended
equimolar saline and sodium bicarbonate in patients with pharmacologic prophylaxis. No additional prospective ran-
lactic acidosis failed to reveal any difference in hemody- domized controlled trials related to this topic have been
namic variables or vasopressor requirements [510, 511]. identified since the meta-analysis and the previous guide-
The number of patients with <7.15 pH in these stud- line were published (ESM 12). Data in support of phar-
ies was small, and we downgraded the certainty of evi- macologic prophylaxis are considered somewhat indirect.
dence for serious imprecision; further, patients did not Except for a large prospective randomized controlled trial
have exclusively septic shock, but also had other diseases, comparing VTE in septic patients treated with drotrecogin
such as mesenteric ischemia. Bicarbonate administration alfa who were randomized to receive placebo versus UFH
has been associated with sodium and fluid overload, an versus LWMH [515], all studies have been in an undiffer-
increase in lactate and Paco2, and a decrease in serum entiated population of critically ill patients. Overall, we
ionized calcium, but the directness of these variables made a strong recommendation in favor of pharmacologic
to outcome is uncertain. The effect of sodium bicarbo- prophylaxis against VTE in critically ill patients based on
nate administration on hemodynamics and vasopressor the overall efficacy of this intervention, although the evi-
requirements at lower pH, as well as the effect on clinical dence was downgraded to moderate because of indirect-
outcomes at any pH level, is unknown. No studies have ness of the populations studied.
examined the effect of bicarbonate administration on A number of studies have also compared use of LMWH
outcomes. This recommendation is unchanged from the to UFH for prevention of VTE prophylaxis in critically ill
2012 guidelines. patients. Four trials were included in the meta-analysis of
Alhazzani etal. [514]. We did not identify any new trials
R. VENOUS THROMBOEMBOLISMPROPHYLAXIS since then. In this meta-analysis, the overall rate of DVT
was lower in patients receiving LWMH compared to
1. We recommend pharmacologic prophylaxis UFH, and overall mortality was reduced by 7%; however,
[unfractionated heparin (UFH) or low-molec- these differences did not reach statistical significance.
ular-weight heparin (LMWH)] against venous In those trials evaluating PE, the rates were significantly
thromboembolism (VTE) in the absence of con- lower in patients receiving LWMH. As with all studies
traindications to the use of these agents (strong of pharmacologic VTE prophylaxis, only one trial [515]
recommendation, moderate quality of evidence). was restricted to septic patients, and that trial utilized
2. We recommend LMWH rather than UFH for drotrecogin alfa in all patients. An additional meta-anal-
VTE prophylaxis in the absence of contraindica- ysis found that LWMH was more effective than UFH in
tions to the use of LMWH (strong recommenda- reducing the incidence of DVT and PE in critically ill
tion, moderate quality of evidence). patients [516]. However, the authors of this meta-analysis
3. We suggest combination pharmacologic VTE included studies of critically ill trauma patients.
prophylaxis and mechanical prophylaxis, when- All studies of LMWH have compared these agents
ever possible (weak recommendation, low quality against UFH administered twice daily. No high-quality
of evidence). studies in critically ill patients have directly compared
4. We suggest mechanical VTE prophylaxis when LWMH against UFH administered thrice daily. An indi-
pharmacologic VTE is contraindicated (weak rect comparison meta-analysis published in 2011 failed
recommendation, low quality of evidence). to identify a significant difference in efficacy between
twice-daily and thrice-daily heparin in medical patients guidelines made no recommendation regarding the use of
[517]. However, another review and meta-analysis (also combined modality in critically ill patients, but do suggest
using indirect comparison) suggested greater efficacy but use of combined mechanical and pharmacologic prophy-
higher rates of bleeding with thrice-daily UFH [518]. laxis in high-risk surgical patients [525, 526].
A Cochrane review demonstrated a substantial decrease A significant number of septic patients may have rela-
in the incidence of HIT in postoperative patients receiving tive contraindications to the use of pharmacologic proph-
LMWH compared to UFH [519], although the studies were ylaxis. These patients may be candidates for mechanical
not specific to either septic or critically ill patients. Finally, prophylaxis using IPC and/or GCS. However, relatively
a cost-effectiveness analysis based on one trial of LMWH little data exist regarding this approach in critically ill
versus UFH [520] suggested that use of LMWH resulted patients. Two meta-analyses have been published com-
in an overall decrease in costs of care, despite the higher paring use of mechanical prophylaxis with no prophylaxis
acquisition cost of the pharmaceutical agent [521]. Overall, in combined patient groups, primarily those undergoing
the desirable consequences (i.e., reduction in PE, HIT, cost orthopedic surgery [527, 528]. The former meta-analysis
savings, and ease of administration) of using LMWH clearly focused on use of GCS and the latter on use of IPC. In
outweigh the undesirable consequences; therefore, we made these analyses, both modalities appeared more effective
a strong recommendation in favor of LMWH instead of than no mechanical prophylaxis, but variable numbers
UFH, whenever feasible. However, the evidence for this was of patients received pharmacologic prophylaxis in both
considered only of moderate quality because of indirect- arms, making this evidence indirect. A cohort study
ness, both with respect to the populations studied and also of 798 patients using propensity scores for risk adjust-
because LMWH has only been systematically compared to ment concluded that IPC was the only effective means
UFH administered twice daily, and not thrice daily. for mechanical VTE prophylaxis in critically ill patients;
Precautions are generally suggested regarding use of however, there was heavy use of pharmacologic prophy-
LMWH in patients with renal dysfunction. In a prelimi- laxis in all groups [529]. Overall, based on these data,
nary trial, no accumulation of anti-Xa levels was dem- we made a weak recommendation for using mechanical
onstrated with dalteparin in patients with a calculated prophylaxis in critically ill septic patients with contrain-
creatinine clearance <30 mL/min [522]. Thus, these dications to use of pharmacologic prophylaxis. Very lim-
patients were included in the PROTECT study [520]. In ited evidence indicates that IPC may be more effective
the actual trial, 118 patients with renal failure were ana- than GCS alone in critically ill patients, making it the
lyzed, 60 of whom were randomized to dalteparin and 58 preferred modality for mechanical prophylaxis.
to UFH. There was no evidence of untoward reactions in
patients receiving dalteparin compared to UFH. How- S. STRESS ULCER PROPHYLAXIS
ever, dalteparin was not more efficacious than UFH in this
small number of patients. These investigators speculated 1. We recommend that stress ulcer prophylaxis be
that other types of LMWH might be safe to use in patients given to patients with sepsis or septic shock who
with renal failure, but acknowledged no other high-qual- have risk factors for gastrointestinal (GI) bleed-
ity data to support this theory. Thus, use of LMWH in ing (strong recommendation, low quality of evi-
septic patients with renal dysfunction might be an option, dence).
but data in support of that remain quite limited. 2. We suggest using either proton pump inhibi-
Combined pharmacologic prophylaxis and mechani- tors (PPIs) or histamine-2 receptor antagonists
cal prophylaxis with intermittent pneumatic compression (H2RAs) when stress ulcer prophylaxis is indi-
(IPC) and/or graduated compression stockings (GCS) is cated (weak recommendation, low quality of evi-
a potential option in critically ill patients with sepsis and dence).
septic shock. No high-quality studies of this approach 3. We recommend against stress ulcer prophylaxis
in septic patients, or even critically ill patients in gen- in patients without risk factors for GI bleeding
eral, exist; however, further research is ongoing [523]. A (BPS).
Cochrane review [524] of 11 studies in surgical patients
suggested that combined prophylaxis was more effective Rationale Stress ulcers develop in the GI tract of criti-
than either modality used alone. However, the quality of cally ill patients and can be associated with significant
evidence was low due to indirectness of population and morbidity and mortality [530]. The exact mechanism is
imprecision of estimates. Therefore, we can make only a not completely understood, but is believed to be related
weak recommendation for combined modality therapy for to disruption of protective mechanisms against gastric
VTE prophylaxis in critically ill patients with sepsis or sep- acid, gastric mucosal hypoperfusion, increased acid pro-
tic shock. Recent American College of Chest Physicians duction, and oxidative injury to the digestive track [531].
The strongest clinical predictors of GI bleeding risk in reporting patients values and preferences concerning the
critically ill patients are mechanical ventilation for >48h efficacy and safety of these agents are essentially lacking.
and coagulopathy [532]. A recent international cohort Furthermore, cost-effectiveness analyses reached differ-
study showed that preexisting liver disease, need for RRT, ent conclusions [547, 548].
and higher organ failure scores were independent predic- Consequently, the benefit of preventing GI bleeding
tors of GI bleeding risk [533]. A multicenter prospective (moderate-quality evidence) must be weighed against the
cohort study found the incidence of clinically important potential increase in infectious complications (very low-
GI bleeding to be 2.6% (95% CI 1.63.6%) in critically to low-quality evidence). The choice of prophylactic agent
ill patients [533]; however, other observational studies will largely depend on individual patients characteristics;
showed lower rates of GI bleeding [534537]. patients values; and the local prevalence of GI bleed-
A recent systematic review and meta-analysis of 20 RCTs ing, pneumonia, and C. difficile infection. Because of the
examined the efficacy and safety of stress ulcer prophylaxis uncertainties, we did not recommend one agent over the
[538]. Moderate quality of evidence showed that prophylaxis other. Ongoing trials aim to investigate the benefit and
with either H2RAs or PPIs reduced the risk of GI bleeding harm of withholding stress ulcer prophylaxis (clinicaltri-
compared to no prophylaxis (RR 0.44; 95% CI 0.280.68; als.gov registration NCT02290327, NCT02467621). The
low quality of evidence showed a nonsignificant increase in results of these trials will inform future recommendations.
pneumonia risk (RR 1.23; 95% CI 0.861.78) [538]. Recently,
a large, retrospective cohort study examined the effect of T. NUTRITION
stress ulcer prophylaxis in patients with sepsis and found no
significant difference in the risk of C difficile infection com- 1. We recommend against the administration of
pared to no prophylaxis [539] (ESM 13). The choice of pro- early parenteral nutrition alone or parenteral
phylactic agent should depend on patients characteristics, nutrition in combination with enteral feedings
patients values and preferences, and the local incidence of (but rather initiate early enteral nutrition) in
C. difficile infections and pneumonia. critically ill patients with sepsis or septic shock
Although published RCTs did not exclusively include who can be fed enterally (strong recommenda-
septic patients, risk factors for GI bleeding are frequently tion, moderate quality of evidence).
present in patients with sepsis and septic shock [532];
therefore, using the results to inform our recommenda- Rationale Parenteral nutrition delivery can secure
tions is acceptable. Based on the available evidence, the the intended amount of calories. This may represent an
desirable consequences of stress ulcer prophylaxis out- advantage over enteral nutrition, especially when patients
weigh the undesirable consequences; therefore, we made may be underfed due to GI intolerance, which may be
a strong recommendation in favor of using stress ulcer pertinent over the first days of care in the ICU. However,
prophylaxis in patients with risk factors. Patients without parenteral delivery is more invasive and has been asso-
risk factors are unlikely to develop clinically important ciated with complications, including an increased risk of
GI bleeding during their ICU stay [532]; therefore, stress infections. Further, purported physiologic benefits are
ulcer prophylaxis should only be used when risk factors associated with enteral feeding, which make this strategy
are present, and patients should be periodically evaluated the mainstay of care [549]. To address the question of the
for the continued need for prophylaxis. superiority of parenteral nutrition for patients with sepsis
While there is variation in practice worldwide, several and septic shock, we evaluated the evidence for patients
surveys showed that PPIs are the most frequently used who could be fed enterally early versus those for whom
agents in North America, Australia, and Europe, followed early enteral feeding was not feasible.
by H2RAs [540544]. A recent meta-analysis including Our first systematic review examined the impact of an
19 RCTs (n = 2177) showed that PPIs were more effec- early parenteral feeding strategy alone or in combination
tive than H2RAs in preventing clinically important GI with enteral feeding versus enteral feeding alone on mor-
bleeding (RR 0.39; 95% CI 0.210.71; p=0.002; moder- tality in patients who could be fed enterally. We identified
ate quality), but led to a nonsignificant increase in pneu- a total of 10 trials with 2888 patients that were conducted
monia risk (RR 1.17; 95% CI 0.881.56; p = 0.28; low in heterogeneous critically ill and surgical patients,
quality) [544] prior meta-analyses reached a similar con- trauma and traumatic brain injury, and those with severe
clusion [545, 546]. None of the RCTs reported the risk of acute pancreatitis [550559]. No evidence showed that
C. difficile infection; nonetheless, a large retrospective early parenteral nutrition reduced mortality (RR 0.97;
cohort study demonstrated a small increase in the risk of 95% CI 0.871.08; n = 2745) or infection risk (RR 1.52;
C. difficile infection with PPIs compared to H2RAs (2.2 95% CI 0.882.62; n=2526), but ICU LOS was increased
vs. 3.8%; p < 0.001; very low-quality evidence). Studies (MD, 0.90; 95% CI 0.381.42; n=46). The quality of the
evidence was graded as moderate to very low. In the larg- reported less muscle wasting and fat loss in the early par-
est randomized trial that addressed this study question enteral nutrition group according to a Subjective Global
(CALORIES, n = 2400), there were fewer episodes of Assessment Score [564]. In summary, due to the lack of
hypoglycemia and vomiting in the early parenteral group, mortality benefit, the increased risk of infection, and the
but no differences in death between the study groups extra cost for parenteral nutrition in the absence of clini-
[553, 560]. Due to the lack of mortality benefit, the added cal benefit [568], current evidence does not support the
cost of parenteral nutrition in absence of clinical benefit initiation of early parenteral nutrition over the first 7days
[550, 551, 555, 560], and the potential physiologic ben- of care for patients with contraindications or intolerance
efits of enteral feeding [549, 561, 562], we recommend to enteral nutrition. Specific patient groups may benefit
early enteral nutrition as the preferred route of adminis- more or incur more harm with early initiation of par-
tration in patients with sepsis or septic shock who can be enteral nutrition in this context. We encourage future
fed enterally. research according to individual patient level meta-anal-
yses to characterize these subgroups and plan for future
2. We recommend against the administration of par- randomized trials. It is important to note that patients
enteral nutrition alone or in combination with who were malnourished were either excluded or rarely
enteral feeds (but rather to initiate IV glucose and represented in the included trials from our systematic
advance enteral feeds as tolerated) over the first review. Since so few malnourished patients were enrolled,
7days in critically ill patients with sepsis or septic evidence to guide practice is lacking. Malnourished
shock for whom early enteral feeding is not feasi- patients may represent a subgroup of critically ill patients
ble (strong recommendation, moderate quality of for whom the clinician may consider initiating parenteral
evidence). nutrition early when enteral feeding is not feasible.

Rationale In some patients with sepsis or septic shock, 3. We suggest the early initiation of enteral feeding
feeding enterally early may not be feasible because of con- rather than a complete fast or only IV glucose in
traindications related to surgery or feeding intolerance. critically ill patients with sepsis or septic shock
These patients represent another subgroup of critically who can be fed enterally (weak recommendation,
ill patients for whom the clinician may question whether low quality of evidence).
to start parenteral nutrition early with or without some 4. We suggest either early trophic/hypocaloric or
enteral feeding to meet nutritional goals, versus trophic/ early full enteral feeding in critically ill patients
hypocaloric enteral feeding alone, or nothing except the with sepsis or septic shock; if trophic/hypocaloric
addition of IV glucose/dextrose for the provision of some feeding is the initial strategy, then feeds should
calories. To address this question, we conducted a system- be advanced according to patient tolerance (weak
atic review, which included a total of four trials and 6087 recommendation, moderate quality of evidence).
patients [563566]. Two of the included trials accounted
for 98.5% of the patients included in the review and, of Rationale The early administration of enteral nutri-
these trials, more than 65% of the patients were surgical tion in patients with sepsis and septic shock has potential
critically ill patients [564, 567]. Seven (20%) of the patients physiologic advantages related to the maintenance of gut
from these two trials were considered septic and patients integrity and prevention of intestinal permeability, damp-
with malnourishment were either excluded or repre- ening of the inflammatory response, and modulation of
sented a very small fraction (n=46, 3.3%) of the included metabolic responses that may reduce insulin resistance
patients. In three of the included trials, parenteral nutri- [561, 562]. To examine evidence for this nutrition strat-
tion was initiated if enteral feeding was not tolerated after egy, we asked if early full feeding (started within the
the first 7days of care [564, 566, 567]. Our review found first 48h and feeding goals to be met within 72h of ICU
that early parenteral nutrition with or without supple- admission or injury) as compared to a delayed strategy
mentation of enteral nutrition was not associated with (feeds delayed for at least 48 h) improved the outcome
reduced mortality (RR 0.96; 95% CI 0.791.16; n=6087; of our critically ill patients. In our systematic review, we
moderate-quality evidence), but was associated with identified a total of 11 trials in heterogeneous critically ill
increased risk of infection (RR 1.12; 95% CI 1.021.24; 3 patient populations (n = 412 patients) [569579]. Only
trials; n = 6054; moderate-quality evidence) (ESM 14). one trial was specifically conducted in patients with sep-
Length of ventilation outcomes were reported divergently sis (n=43 patients) [577]. The risk of death was not sig-
in the two large trials, with one suggesting an increase nificantly different between the groups (RR 0.75; 95% CI
[567] and the other a decrease [564] in ventilation time 0.431.31; n=188 patients), and infections were not sig-
associated with early parenteral nutrition. One trial also nificantly reduced (RR 0.60; 95% CI 0.3412.07; n=122
patients). Other recent systematic reviews in the criti- target goal. Study intervention periods ranged from 6
cally ill focused specifically on trauma (three trials, 126 to 14days (or until ICU discharge). In three of the tri-
patients) or more heterogeneous critically ill populations als, protein (0.81.5 g/kg/days) was administered to
(6 trials, n = 234 patients) and found that early enteral the trophic/hypocaloric group to meet protein require-
feeding reduced death and pneumonia [580, 581]. How- ments [584, 586, 587]. Overall, there were no differences
ever, in contrast to our systematic review, these latter in mortality (RR 0.95; 95% CI 0.821.10; n=2665; high-
reviews did not include studies in which enteral feeding quality evidence), infections (RR 0.96; 95% CI 0.83
in the intervention arm was both early and full and where 1.12; n = 2667; moderate-quality evidence), or ICU
the control arm feeding strategy was delayed for at least LOS (MD, 0.27days; 95% CI 1.40 to 0.86, n=2567;
the first 48 h. We also examined whether the provision moderate-quality evidence between the study groups)
of an early trophic/hypocaloric feeding strategy (defined (ESM 15). One trial that instituted hypocaloric feeding
by enteral feeding started within the first 48h and up to (goal 4060% target feeds for up to 14 days) reported
70% of target caloric goals for at least 48h) was superior a subgroup of 292 patients with sepsis; there were also
to a delayed enteral feeding strategy. In the two trials that no detectable differences in death at 90 days between
fit these criteria, there were no statistical differences in the study groups (RR 0.95; 95% CI 0.711.27; p = 0.82
death (RR 0.67; 95% CI 0.351.29; n = 229; low-qual- for interaction) [584]. A recently published systematic
ity evidence) or infection (RR 0.92; 95% CI 0.611.37; review of normocaloric versus hypocaloric feeding also
n=229; very low-quality evidence) between the groups found no differences in hospital mortality, infections,
[582, 583]. Since the present evidence does not suggest ICU LOS, or ventilator-free days between the study
harm with early versus delayed institution of enteral groups [585]. Some evidence also suggests a lack of
feeding, and there is possible benefit from physiologic adverse consequences even with longer-term outcomes.
evidence suggesting reduced gut permeability, inflam- A trophic/hypocaloric feeding trial of 525 patients,
mation, and infection risk, the committee issued a weak which instituted the most significant restrictions in
recommendation to start feeding early in patients with enteral feeding (20% of caloric goal) for up to 6 days,
sepsis and septic shock. found no differences in muscle strength, muscle mass,
Some evidence suggests that intentional early under- and 6-min walk test at 6 months or 1 year, although
feeding as compared to early full feeding of critically ill patients in the trophic/hypocaloric feeding group were
patients may lead to immune hyporesponsiveness and more likely to be admitted to a rehabilitation facility
an increase in infectious complications [549]. Further, during the first 12months of follow-up [592]. The cur-
because critical illness is associated with loss of skeletal rent evidence base would suggest that a trophic/hypoca-
mass, it is possible that not administering adequate pro- loric or early full enteral feeding strategy is appropriate.
tein may lead to challenges weaning from the ventilator However, for patients with sepsis or septic shock who
and more general weakness. However, a biological ration- are not tolerating enteral feeds, trophic/hypocaloric
ale for a trophic/hypocaloric or hypocaloric feeding strat- feeding may be preferred, with feeds titrated over time
egy exists, at least as the initial approach to feeding the according to patient tolerance. There is insufficient
critically ill as compared to a fully fed strategy. Limiting evidence to confirm that a trophic/hypocaloric feed-
caloric intake stimulates autophagy, which is considered ing strategy is effective and safe in patients who are
a defense mechanism against intracellular organisms and malnourished (body mass index <18.5) because these
therefore raises the possibility that this approach could patients were either excluded or rarely represented in
reduce infection risk [584, 585]. the clinical trials from our systematic review. Until fur-
We defined feeds as trophic/hypocaloric if goal feeds ther clinical evidence is generated for this subpopula-
were 70% or less of standard caloric targets for at least tion, the clinician may consider titrating enteral feeds
a 48-hour period before they were titrated toward goal. more aggressively in accordance with patient tolerance
Our systematic review identified seven randomized tri- while monitoring for re-feeding syndrome. Current evi-
als and 2665 patients studied [584, 586591]. Patient dence did not specifically address patients with high
populations included heterogeneous critically ill vasopressor requirements, and the decision about with-
patients and those with acute lung injury and/or ARDS. holding the feeds should be individualized.
Patients who were malnourished were excluded from
four of the trials [588591] and the average body mass 5. We recommend against the use of omega-3 fatty
index in the remaining three trials ranged from 28 to 30 acids as an immune supplement in critically ill
[584, 586, 587]. Targets for trophic/hypocaloric feed- patients with sepsis or septic shock (strong rec-
ing groups ranged from 10 to 20kcal/h to up to 70% of ommendation, low quality of evidence).
 Rationale Use of omega-3 fatty acids in the con- or pneumonia [597603]. In our systematic review, we
text of clinical trials in the critically ill has been a sub- identified one multicenter non-inferiority trial of 452
ject of interest during the past several years because of critically ill patients who were randomized to not moni-
the immunomodulatory potential [593]. However, sys- toring GRVs versus monitoring GRVs at 6-h intervals
tematic reviews of parenteral or enteral omega-3 sup- [602]. Intolerance to feeds was defined as vomiting in
plementation in critically ill and ARDS patients have not the intervention group versus a GRV of >250mL, vomit-
confirmed their therapeutic benefit [594, 595]. Further, a ing, or both in the control group. Although vomiting was
recent randomized trial of 272 patients with acute lung more frequent (39.6 versus 27%; median difference, 12.6;
injury found excess harm related to mortality as well as 95% CI 5.419.9) in the group in which GRVs were not
fewer ventilator- and ICU-free days in the omega-3 arm monitored, a strategy of not monitoring GRVs was found
as compared to the control arm [596]. A limitation of to be non-inferior compared to monitoring at 6-h inter-
this trial as well as several other omega-3 trials is that the vals with regard to the primary outcome of VAP (16.7
intervention arm also contained vitamins and trace min- versus 15.8% respectively; difference, 0.9%; 95% CI 4.8
eral supplementation, making omega-3 fatty acids alone to 6.7%). No detectable differences in death were shown
difficult to isolate as the cause for harm or benefit. For between the study groups at 28 and 90days. Patients who
these reasons, we conducted a systematic review of clini- had surgery up to one month prior to study eligibility
cal trials in the critically ill that administered omega-3 were not included in this study, so these results should
alone in the intervention arm. In a total of 16 trials not be applied to surgical critically ill patients. How-
(n = 1216 patients), there were no significant reduc- ever, the results of this trial question the need to meas-
tions in death (RR 0.86; 95% CI 0.711.03; low quality ure GRVs as a method to reduce aspiration pneumonia
evidence); however, ICU LOS was significantly reduced in all critically ill patients. Due to the absence of harm
in the omega-3 group (MD, 3.84days; 95% CI 5.57 to and the potential reduction in nursing resources needed
2.12, very low-quality evidence). The overall quality of to monitor patients, we suggest against routine monitor-
the evidence was graded as low. Due to the uncertainty ing of GRVs in all patients with sepsis unless the patient
of benefit, the potential for harm, and the excess cost has demonstrated feeding intolerance (e.g., vomiting,
and varied availability of omega-3 fatty acids, we make a reflux of feeds into the oral cavity) or for patients who
strong recommendation against the use of omega-3 fatty are considered to be at high risk for aspiration (e.g., sur-
acids for patients with sepsis and septic shock outside the gery, hemodynamic instability). We recommend the gen-
conduct of RCTs. eration of further evidence through the conduct of future
randomized controlled trials targeted to higher-risk
6. We suggest against routinely monitoring gastric patient groups such as the surgical population or those
residual volumes (GRVs) in critically ill patients in shock to determine the threshold and frequency with
with sepsis or septic shock (weak recommenda- which GRVs should be monitored.
tion, low quality of evidence). However, we sug-
gest measurement of gastric residuals in patients 7. We suggest the use of prokinetic agents in criti-
with feeding intolerance or who are considered to cally ill patients with sepsis or septic shock and
be at high risk of aspiration (weak recommenda- feeding intolerance (weak recommendation, low
tion, very low quality of evidence). quality of evidence).
Remarks This recommendation refers to nonsurgical
critically ill patients with sepsis or septic shock. Rationale Feeding intolerance is defined as vomiting,
aspiration of gastric contents, or high GRVs. For multi-
Rationale Critically ill patients are at significant risk ple reasons, feeding intolerance commonly develops in
for GI dysmotility, which may then predispose them to critically ill patients. Patients with preexisting gastro-
regurgitation or vomiting, aspiration, and the develop- paresis or diabetes or those who are receiving sedatives
ment of aspiration pneumonia. The rationale for meas- and vasopressors are at risk. Prokinetic agents, includ-
urement of GRVs is to reduce the risk for aspiration ing metoclopramide, domperidone, and erythromycin,
pneumonia by either ceasing or modifying the enteral are frequently used in the ICU. Each of these agents has
feeding strategy based on the detection of excess gas- different pharmacodynamics and pharmacokinetic prop-
tric residuals. The inherent controversy is that observa- erties; however, these agents may be associated with pro-
tional and interventional studies have not consistently longation of QT interval and ventricular arrhythmias. A
confirmed a relationship between the measurement of large casecontrol study in non-ICU patients showed a
GRVs (with thresholds ranging from 200mL to no moni- threefold increase in risk of sudden cardiac death with
toring of GRVs) and outcomes of vomiting, aspiration, domperidone use at doses >30 mg/day [604]. Another
retrospective cohort study showed that outpatient use of opioids, or NMBAs), gastric hypoperfusion in the context
erythromycin is associated with a twofold increase in the of shock, hyperglycemia, or vasopressor use [610612].
risk of sudden cardiac death, especially if concomitantly Post-pyloric tubes have the theoretical advantage of
used with other CYP3A inhibitors [605]. The impact on improving feeding intolerance in patients with gastropa-
ventricular arrhythmias in ICU patients is less clear. resis, consequently improving the delivery of nutrition
A recent systematic review and meta-analysis included into the gut. Post-pyloric feeding tubes, although safe,
13 RCTs enrolling 1341 critically ill patients showed that are not always available, and require technical skill for
prokinetic agent use was associated with lower risk of successful insertion. Gastric air insufflation and proki-
feeding intolerance (RR 0.73; 95% CI 0.550.97; mod- netic agents are both effective strategies to facilitate the
erate-quality evidence). This was equivalent to an abso- insertion of post-pyloric tubes in critically ill patients
lute risk reduction of 17%. The use of prokinetic agents [613]. Endoscopy and an external magnet device can also
did not significantly increase mortality (RR 0.97; 95% be used to guide post-pyloric tube insertion, but are not
CI 0.811.1; low-quality evidence); however, the inci- always available, are expensive, and require a higher level
dence of fatal or nonfatal cardiac arrhythmias was not of expertise.
consistently reported across studies. There was no sig- We conducted a systematic review and meta-anal-
nificant effect on the risk of pneumonia or vomiting. The ysis of randomized trials to examine the effect of
majority of trials examined the effect of metoclopramide post-pyloric (compared to gastric) feeding on patient-
or erythromycin; subgroup analysis by drug class was important outcomes. We identified 21 eligible RCTs
underpowered to detect important subgroup differences enrolling 1579 patients. Feeding via post-pyloric tube
[606]. We considered the desirable consequences (lower reduced the risk of pneumonia compared to gastric
risk of feeding intolerance) and the low quality of evi- tube feeding (RR 0.75; 95% CI 0.590.94; low-quality
dence showing no difference in mortality or pneumonia, evidence). This translates into a 2.5% (95% CI 0.64.1%)
and issued a weak recommendation for using proki- absolute reduction in pneumonia risk. However, there
netic agents (metoclopramide or erythromycin) to treat was no significant effect on the risk of death, aspiration,
feeding intolerance in patients with sepsis. Future large or vomiting (ESM 16). This is consistent with the results
comparative trials are needed to determine the relative of older meta-analyses [614, 615]. Although the use of
efficacy and safety of different agents. post-pyloric tubes reduced risk of pneumonia, the qual-
Monitoring the QT interval with serial electrocardio- ity of evidence was low, the magnitude of benefit was
grams is required when these agents are used in the ICU, small, and there was uncertainty about the effect on
especially if concomitantly used with other agents that other patient-important outcomes. Cost-effectiveness
could prolong the QT interval [607]. The need for proki- studies that describe the economic consequences of
netic agents should be assessed daily, and they should be using post-pyloric feeding tubes are lacking. There-
stopped when clinically not indicated. fore, we decided that the balance between desirable
and undesirable consequences was unclear in low-risk
8. We suggest placement of post-pyloric feeding patients; however, the use of post-pyloric feeding tubes
tubes in critically ill patients with sepsis or septic may be justified in patients at high risk of aspiration
shock with feeding intolerance or who are consid- (i.e., patients with history of recurrent aspiration, severe
ered to be at high risk of aspiration (weak recom- gastroparesis, feeding intolerance, or refractory medical
mendation, low quality of evidence). treatment).

Rationale Feeding intolerance is defined as vomiting, 9. We recommend against the use of IV selenium to
abdominal distention, or high GRVs that result in inter- treat sepsis and septic shock (strong recommen-
ruption of enteral nutrition. Critically ill patients are at dation, moderate quality of evidence).
risk of gastroparesis and feeding intolerance; evidence of
delayed gastric emptying can be found in approximately Rationale Selenium was administered in the hope
50% of critically ill patients [608]. The proportion of that it could correct the known reduction of selenium
patients who will progress to develop clinical symptoms concentration in sepsis patients and provide a pharma-
is less clear. Feeding intolerance can result in interrup- cologic effect through an antioxidant defense. Although
tion of nutritional support, vomiting, aspiration of gas- some RCTs are available, the evidence for the use of IV
tric contents, or pneumonia [609]. The pathophysiology selenium is not convincing. Two recent meta-analyses
is not completely understood and is likely to be multi- suggest, with weak findings, a potential benefit of sele-
factorial. Gastroparesis can be caused by pharmacologic nium supplementation in sepsis [616, 617]. However, a
agents that are frequently used in the ICU (e.g., sedatives, recent large RCT also examined the effect on mortality
rates [618]. Overall pooled odds ratio (0.94; CI 0.771.15) 12. We make no recommendation about the use of
suggests no significant impact on mortality with sepsis. carnitine for sepsis and septic shock.
Also, no differences in secondary outcomes of develop-
ment of nosocomial pneumonia or ICU LOS were found. Rationale Massive disruption in energy metabolism
When updating our meta-analysis to include the results contributes to sepsis severity and end organ failure.
of this recent study, there was no difference in mortality The magnitude of the energy shift, and, possibly more
between both groups (ESM 17). importantly, the hosts metabolic adaptiveness to the
shift in energy demand, likely influence patient survival.
10. We suggest against the use of arginine to treat Carnitine, endogenously manufactured from lysine and
sepsis and septic shock (weak recommendation, methionine, is required for the transport of long-chain
low quality of evidence). fatty acids into the mitochondria and the generation of
energy. As such, carnitine utilization is essential for ena-
Rationale Arginine availability is reduced in sepsis, bling the switch from glucose to long-chain fatty acid
which can lead to reduced nitric oxide synthesis, loss of metabolism during the sepsis energy crisis. This is the
microcirculatory regulation, and enhanced production basis for the rationale of employing l-carnitine as a ther-
of superoxide and peroxynitrite. However, arginine sup- apeutic in sepsis. One small randomized trial in patients
plementation could lead to unwanted vasodilation and with sepsis reported a 28-day mortality decrease in septic
hypotension [619, 620]. Human trials of l-arginine sup- shock patients treated with IV l-carnitine therapy within
plementation have generally been small and reported 24h of shock onset; however, the trial was underpowered
variable effects on mortality [621624]. The only study to detect such a difference [637]. Larger, ongoing trials
in septic patients showed improved survival, but had should provide more evidence of the usefulness of carni-
limitations in study design [623]. Other studies suggested tine supplementation.
no benefit or possible harm in the subgroup of septic
patients [621, 624, 625]. Some authors found improve- U. SETTING GOALS OF CARE
ment in secondary outcomes in septic patients, such as
reduced infectious complications) and hospital LOS, but 1. We recommend that goals of care and prognosis
the relevance of these findings in the face of potential be discussed with patients and families (BPS).
harm is unclear. 2. We recommend that goals of care be incorpo-
rated into treatment and end-of-life care plan-
11. We recommend against the use of glutamine to ning, utilizing palliative care principles where
treat sepsis and septic shock (strong recommen- appropriate (strong recommendation, moderate
dation, moderate quality of evidence) quality of evidence).
3. We suggest that goals of care be addressed as
Rationale Glutamine levels are also reduced during early as feasible, but no later than within 72h of
critical illness. Exogenous supplementation can improve ICU admission (weak recommendation, low qual-
gut mucosal atrophy and permeability, possibly lead- ity of evidence).
ing to reduced bacterial translocation. Other potential
benefits are enhanced immune cell function, decreased Rationale Patients with sepsis and multiple organ
proinflammatory cytokine production, and higher lev- system failure have a high mortality rate; some will not
els of glutathione and antioxidative capacity [619, 620]. survive or will have a poor quality of life. Although the
However, the clinical significance of these findings is not outcome of intensive care treatment in critically ill
clearly established. patients may be difficult to prognosticate accurately,
Although a previous meta-analysis showed mortal- establishing realistic ICU treatment goals is paramount
ity reduction [626], several other meta-analyses did not [638], especially because inaccurate expectations about
[627630]. Four recent well-designed studies also failed prognosis are common among surrogates [639]. Non-
to show a mortality benefit in the primary analyses, beneficial ICU advanced life-prolonging treatment is
although none focused specifically on septic patients not consistent with setting goals of care [640, 641]. Mod-
[631634]. Two small studies on septic patients showed els for structuring initiatives to enhance care in the ICU
no benefit in mortality rates [635, 636], but showed a sig- highlight the importance of incorporating goals of care,
nificant reduction in infectious complications [636] and a along with prognosis, into treatment plans [642]. The use
faster recovery of organ dysfunction. of proactive family care conferences to identify advance
directives and treatment goals within 72h of ICU admis- 9
Hadassah Hebrew University Medical Center, Jerusalem, Israel. 10Sunnybrook
Health Sciences Centre, Toronto, ON, Canada. 11University ofPittsburgh
sion has been demonstrated to promote communica- Critical Care Medicine CRISMA Laboratory, Pittsburgh, PA, USA. 12Hospital Ray
tion and understanding between the patients family and mond Poincare, Garches, France. 13Saint Thomas Hospital, London, England,
the treating team; improve family satisfaction; decrease UK. 14University College London Hospitals, London, England, UK. 15Vanderbilt
University Medical Center, Nashville, TN, USA. 16Service de Reanimation Medi
stress, anxiety, and depression in surviving relatives; cale, Paris, France. 17CHIREC Hospitals, Braine LAlleud, Belgium. 18Western
facilitate end-of-life decision-making; and shorten ICU Hospital, Victoria, Australia. 19Keio University School ofMedicine, Tokyo, Japan.
LOS for patients who die in the ICU [643, 644]. Promot- 20
Vivantes-Klinikum Neuklln, Berlin, Germany. 21Karl Heusner Memorial Hos
pital, Belize Healthcare Partners, Belize City, Belize. 22Cooper Health System,
ing shared-decision-making with patients and families Camden, NJ, USA. 23University ofMississippi Medical Center, Jackson, MS, USA.
is beneficial in ensuring appropriate care in the ICU and 24
Jupiter Hospital, Thane, India. 25Rush University Medical Center, Chicago,
that futile care is avoided [641, 645, 646]. IL, USA. 26ASAN Medical Center, University ofUlsan College ofMedicine,
Seoul, South Korea. 27Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil.
Palliative care is increasingly accepted as an essen- 28
Federal University ofSao Paulo, Sao Paulo, Brazil. 29Regions Hospital, St.
tial component of comprehensive care for critically ill Paul, MN, USA. 30Saint Michaels Hospital, Toronto, ON, Canada. 31Washington
patients regardless of diagnosis or prognosis [642, 647]. University School ofMedicine, St. Louis, MO, USA. 32Ottawa Hospital, Ottawa,
ON, Canada. 33Nepean Hospital, University ofSydney, Penrith, NSW, Australia.
Use of palliative care in the ICU enhances the ability to 34
Mount Sinai Hospital, Toronto, ON, Canada. 35UCINC, Centro Hospitalar de
recognize pain and distress; establish the patients wishes, Lisboa Central, Lisbon, Portugal. 36University ofNew South Wales, Sydney,
beliefs, and values, and their impact on decision-making; NSW, Australia. 37Universit dellla Magna Graecia, Catanzaro, Italy. 38Fujita
Health University School ofMedicine, Toyoake, Aich, Japan. 39Rigshospita
develop flexible communication strategies; conduct fam- let, Copenhagen, Denmark. 40Universit Sapienza, Rome, Italy. 41Christiana
ily meetings and establish goals of care; provide family Care Health Services, Newark, DE, USA. 42University ofPittsburgh School
support during the dying process; help resolve team con- ofMedicine, Pittsburgh, PA, USA. 43Stanford University School ofMedicine,
Stanford, CA, USA. 44Kaust Medical Services, Thuwal, Saudi Arabia. 45University
flicts; and establish reasonable goals for life support and ofKansas Medical Center, Kansas City, KS, USA. 46Wolfson Institute ofBiomedi
resuscitation [648]. cal Research, London, England, UK. 47Massachusetts General Hospital, Boston,
A recent systematic review of the effect of palliative MA, USA. 48California Pacific Medical Center, San Francisco, CA, USA. 49Uni
versity ofAmsterdam, Amsterdam, Netherlands. 50Erasm University Hospital,
care interventions and advanced care planning on ICU Brussels, Belgium. 51Houston Methodist Hospital, Houston, TX, USA.
utilization identified that, despite wide variation in study
type and quality among nine randomized control trials Acknowledgements
We would like to acknowledge the members of the systematic review team:
and 13 nonrandomized controlled trials, patients who Drs. Emile Belley-Cote, Fayez Alshamsi, Sunjay Sharma, Eric Duan, Kim Lewis,
received advance care planning or palliative care inter- and Clara Lu for their invaluable help in the systematic review process. We also
ventions consistently showed a pattern toward decreased would like to acknowledge professors Gordon Guyatt and Roman Jaeschke for
sharing their methodology expertise. Finally, we thank Deborah McBride for
ICU admissions and reduced ICU LOS [649]. the incredible editorial support.
However, significant inter-hospital variation in ratings
and delivery of palliative care is consistent with prior Endorsing Organizations The following sponsoring organizations (with
formal liaison appointees) endorse this guideline: American College of Chest
studies showing variation in intensity of care at the end Physicians, American College of Emergency Physicians, American Thoracic So
of life [650]. Despite differences in geographic location, ciety, Asia Pacific Association of Critical Care Medicine, Associao de Medicina
legal system, religion, and culture, there is worldwide Intensiva Brasileira, Australian and New Zealand Intensive Care Society, Cons
orcio Centroamericano y del Caribe de Terapia Intensiva, European Society of
professional consensus for key end-of-life practices in the Clinical Microbiology and Infectious Diseases, German Sepsis Society, Indian
ICU [651]. Society of Critical Care Medicine, International Pan Arab Critical Care Medicine
Promoting patient- and family-centered care in the Society, Japanese Association for Acute Medicine, Japanese Society of Inten
sive Care Medicine, Latin American Sepsis Institute, Scandinavian Critical Care
ICU has emerged as a priority and includes implemen- Trials Group, Society for Academic Emergency Medicine, Society of Hospital
tation of early and repeated care conferencing to reduce Medicine, Surgical Infection Society, World Federation of Critical Care Nurses,
family stress and improve consistency in communication; World Federation of Societies of Intensive and Critical Care Medicine.
The following non-sponsoring organizations (without formal liaison ap
open flexible visitation; family presence during clini- pointees) endorse this guideline: Academy of Medical Royal Colleges, Chinese
cal rounds, resuscitation, and invasive procedures; and Society of Critical Care Medicine, Asociacin Colombiana de Medicina Crtica y
attention to cultural and spiritual support [652655]. Cuidado Intensivo, Emirates Intensive Care Society, European Society of Paedi
atric and Neonatal Intensive Care, European Society for Emergency Medicine,
Electronic supplementary material Federacin Panamericana e Ibrica de Medicina Crtica y Terapia Intensiva,
The online version of this article (doi:10.1007/s00134-017-4683-6) contains Sociedad Peruana de Medicina Intensiva, Shock Society, Sociedad Argentina
supplementary material, which is available to authorized users. de Terapia Intensiva, World Federation of Pediatric Intensive and Critical Care
Societies.
Author details
1
St. Georges Hospital, London, England, UK. 2New York University School Governance of Surviving Sepsis Campaign Guidelines Committee
ofMedicine, New York, NY, USA. 3McMaster University, Hamilton, ON, Canada. SSC Executive and Steering Committees http://www.survivingsepsis.org/
4
Brown University School ofMedicine, Providence, RI, USA. 5Instituto di About-SSC/Pages/Leadership.aspx.
Anestesiologia e Rianimazione, Universit Cattolica del Sacro Cuore, Rome,
Italy. 6Vall dHebron University Hospital, Barcelona, Spain. 7University ofMani SSC Guidelines Committee Oversight Group
toba, Winnipeg, MB, Canada. 8Emory University Hospital, Atlanta, GA, USA. Andrew Rhodes, Laura Evans, Mitchell M. Levy.
SSC Guidelines Committee Group Heads Mazuski received funding from Actavis (Allergan) (consultant), Astra-Zeneca
Massimo Antonelli (Hemodynamics), Ricard Ferrer (Adjunctive therapies), (consultant), Bayer (consultant), and from Cubist (now part of Merck)
Anand Kumar (Infection), Jonathan E. Sevransky (Ventilation), Charles L. (consultant); he received research grant funding from Astra-Zeneca, Bayer, and
Sprung (Metabolic). from Merck; and participates in Surgical Infection Society (President-elect and
Chair of Task Force on Guidelines for the Management of Intra-abdominal
GRADE Methodology Group Infection) and in the American College of Surgeons (speaker at Annual
Waleed Alhazzani (chair), Mark E. Nunnally, Bram Rochwerg. Congress, member of Trusted Medical Information Commission). Dr. Mehta
participates in ATS activities. Dr. Moreno participates in the Portuguese and
Brasilian Societies of Intensive Care Medicine. Dr. Myburghs institution
Compliance with ethical standards received unrestricted grant funding, logistical support and reimbursement
from Fresenius Kabi for travel expenses to conduct a randomized controlled
Conflicts of interest trial of fluid resuscitation (CHEST study): 2008-2012: A$7,600,000 (US$
Dr. Rhodes is a past-president of the European Society of Intensive Care 5,000,000); an unrestricted grant for partial funding from Baxter Healthcare of
Medicine. Dr. Levy received consulting fees from ImmuneExpress. Dr. Antonelli an international observational study of patterns of fluid resuscitation (FLUID
received funding from Pfizer, MSD, Cubist, Maquet, Drager, Toray, and Baxter; TRIPS study) in 2014: A$70,000 (US$ 50,000); honoraria and travel reimburse
he participates in ESA and SIAARTI. Dr. Kumar received scientific consulting ments from Baxter Healthcare for participation in Advisory Board meetings in
fees from Baxter, Isomark, and Opsonix on diagnostic technologies; he Sydney (2013), Paris (2014) and China (2014); and an unrestricted grant for
received grant funding from GSK in the area of influenza. Dr. Ferrer Roca partial funding from CSL Bioplasma for an international observational study of
received funding from Estor, MSD, Astra-Zeneca, and Grifols and participates in patterns of fluid resuscitation (FLUID TRIPS study) in 2014: A$10,000 (US$
ESICM and SEMICYUC. Dr. Sevransky is an Associate Editor for Critical Care 7500); he also participates as a council member in the World Federation of
Medicine. Dr. Sprung received funding from Asahi Kasei Pharma America Societies of Intensive and Critical Care Medicine. Dr. Navalesi participates in
Corporation (consultant, Data Safety and Monitoring Committee) and the European Respiratory Society (Head of Assembly Respiratory Intensive
LeukoDx Ltd. (consultant; PI, research study on biomarkers of sepsis). He Care), is a member of ESICM (European Society of Intensive Care Medicine)
participates in International Sepsis Forum (board member). Dr. Angus received and ESA (European Society of Anaesthesiology), and is in the Scientific
funding Ferring Inc (consulting fees for serving on the Trial Steering Committee of SIAARTI (the Italian Association of Anesthesia and Intensive
Committee of a Phase 2/3 trial of selepressin for septic shock), and from Ibis Care). Dr. Nishida participates in The Japanese Society of Intensive Care
and Genmark (both for consulting fees regarding diagnostic strategies in Medicine (vice chairman of the executive boards), the Japanese Guidelines for
sepsis). He is a contributing editor for JAMA, has conducted committee the Management of Sepsis and Septic Shock 2016 (chairman), The Japanese
membership work for the American Thoracic Society, and has contributed to Guidelines for Nutrition Support Therapy in the Adult and Pediatric Critically Ill
an IOM workshop on regulatory science. Dr. Angus provided expert testimony Patients (board), The Japanese Guidelines for the Management of Acute
in medical malpractice cases. Dr. Beales institution received funding from Kidney Injury 2016 (board), The Expert Consensus of the Early Rehabilitation in
Roche (consulting regarding sepsis diagnostics); he received funding from Critical Care (board), The sepsis registry organization in Japan (member). Dr.
Quintiles (consulting on routes to license for a potential ARDS therapy); he Osborn received funding from Cheetah (speaker related to fluid resuscitation
participates in the UK National Institute for Clinical and Healthcare Excellence and use of NICOM); she participates in American College of Emergency
Sepsis Guideline Development Group; he has served as an expert witness, Physicians (Representative to SCC), consultant for national database
disclosing that he is approached from time to time regarding expert witness development, CDC sepsis task force, IHI consultant. Dr. Perner is the editor of
testimony for ICU cases, which may involve patients who have sepsis and the ICM; his department received research funding from CSL Behring and
testimony relates to generally accepted current standards of care, and formal Fresenius Kabi. Dr. Ranieri participates in ESICM. Dr. Seckel received funding
guidance, as it currently pertains within the UK. Dr. Bellingan received funding from American Association of Critical-Care Nurses (AACN) (honorarium for
from Faron (research into interferon in lung injury) and Athersys (stem cells in speaker at 2016 annual conference; AACN Online Web based Essentials of
lung injury). Dr. Chiche received funding for consulting activities and Critical Care Orientation); she participates as a volunteer for AACN, and served
honoraria for lectures from GE Healthcare, monitoring and IT solutions; he as AACN liaison to the ATS/ESICM/SCCM CPG: Mechanical Ventilation in Adult
received funding from Nestl Healthsciences (consulting activities and Patients with ARDS. Dr. Shieh participates in Society of Hospital Medicine
honorarium), and from Abbott diagnostics (consulting activities). Dr. Faculty for Sepsis Workshop, SHM-SCCM Moore Foundation collaborative
Coopersmith is on the fellowship committee of Surgical Infection Society. Dr. faculty. Dr. Shukri participates in the International Pan Arab Critical Care
De Backer received funding from Edwards Healthcare, Fresenius Kabi, and Society educational activities. Dr. Simpson participates in CHEST Regent at
Grifols. Dr. Dellinger provided expert testimony for alleged malpractice in Large (board of directors), and is an ATS member. Dr. Singer received funding
critical care. Dr. French participates in Australian and New Zealand Intensive from Deltex Medical, Bayer, Biotest, and MSD; he participates in the UK
Care Society Clinical Trials Group (chair). Dr. Fujishima participates in the Intensive Care Society research and Meeting committees; he has provided
Japanese Association for Acute Medicine (board member, Japanese Guidelines expert testimony, disclosing: I do medicolegal work (6 cases/year) as an
for the management of sepsis) and Japanese Respiratory Society (board independent expert, 80% on behalf of the defendant. Dr. Thompson received
member, Japanese Guidelines for the management of ARDS); he received funding from serving on DSMBs trials sponsored by Ferring Pharmaceuticals,
funding from Asahi Kasei Co (lecture). Dr. Hollenberg participates in the ACC/ Farron Labs, and Roche Genentec; also received funding from Asahi Kasei
AHA PCI and Heart Failure guidelines, CHEST editorial board, ACCP-SEEK, and Pharma America (consulting), UpToDate (wrote two chapters on pulmonary
CHEST CV Network chair. Dr. Jones participates in ACEP and SAEM, and has embolism diagnosis), and was a pro bono consultant for BioAegis; participates
served as an expert witness on various cases. Dr. Karnad received funding as a member of the American Thoracic Society committee to develop the ATS/
from Quintiles Cardiac Safety Services (consultant) and from Bharat Serum and ESICM/SCCM Clinical Practice Guideline: Mechanical Ventilation in Adult
Vaccines Ltd (consultant). He participates in the Indian Society of Critical Care Patients with Acute Respiratory Distress Syndrome. Dr. Vincent participates in
Medicine and the Association of Physicians of India. Dr. Kleinpell participates World Federation of Societies of Intensive and Critical Care Societies
in Critical Care Medicine American Board of Internal Medicine (board (president) and Critical Care Foundation (president). Dr. Wiersinga is treasurer
member), Institute of Medicine of Chicago (board member), and the of both the ESCMID Study Group for Bloodstream Infections and Sepsis
Commission on Collegiate Nursing Education (board member). Dr. Koh (ESGBIS) and the Dutch Working Party on Antibiotic Policy (SWAB), Academic
participates in The Korean Society of Critical Care Medicine, The European Medical Center, University of Amsterdam (all non-profit). Dr. Zimmerman
Society of Intensive Care Medicine, and The Korean Society of Medical Ethics. participates in ACCP, ACP, WFSICCM, and PAIF; she has provided expert
Dr. Lisboa participates in ILAS, AMIB, and ESICM. Dr. Machado participates in testimony on loss of digits due to DIC, mesenteric ischemia. Dr. Nunnally
the Latin America Sepsis Institution (CEO). Dr. Marshall received funding from participates in SOCCA (bpoard), ASA (committee), NYSSA, IARS, and AUA. Dr.
Member Data Safety Monitoring Committee AKPA Pharma; he participates in Rochwerg participates as a methodologist for ATS, ESCIM, and Canadian Blood
International Forum for Acute Care Trialists (Chair) and World Federation of services. The remaining authors have disclosed that they do not have any
Societies of Intensive and Critical Care Medicine (Secretary-General). Dr. potential conflicts of interest.
Appendix 1
Recommendations and best practice statements

A. INITIAL RESUSCITATION

1. Sepsis and sepc shock are medical emergencies, and we recommend that treatment and resuscitaon begin
immediately (BPS).
2. We recommend that, in the resuscitaon from sepsis-induced hypoperfusion, at least 30 mL/kg of IV crystalloid fluid
be given within the first 3 hours (strong recommendaon, low quality of evidence).
3. We recommend that, following inial fluid resuscitaon, addional fluids be guided by frequent reassessment of
hemodynamic status (BPS).
Remarks: Reassessment should include a thorough clinical examinaon and evaluaon of available physiologic
variables (heart rate, blood pressure, arterial oxygen saturaon, respiratory rate, temperature, urine output, and
others, as available) as well as other noninvasive or invasive monitoring, as available.
4. We recommend further hemodynamic assessment (such as assessing cardiac function) to determine the type of shock
if the clinical examinaon does not lead to a clear diagnosis (BPS).
5. We suggest that dynamic over stac variables be used to predict fluid responsiveness, where available (weak
recommendaon, low quality of evidence).
6. We recommend an inial target mean arterial pressure of 65 mm Hg in paents with sepc shock requiring
vasopressors (strong recommendaon, moderate quality of evidence).
7. We suggest guiding resuscitaon to normalize lactate in paents with elevated lactate levels as a marker of ssue
hypoperfusion (weak recommendaon, low quality of evidence).

B. SCREENING FOR SEPSIS AND PERFORMANCE IMPROVEMENT

1. We recommend that hospitals and hospital systems have a performance improvement program for sepsis, including
sepsis screening for acutely ill, high risk paents (BPS).

C. DIAGNOSIS

1. We recommend that appropriate roune microbiologic cultures (including blood) be obtained before starng
anmicrobial therapy in paents with suspected sepsis or sepc shock if doing so results in no substanal delay in the
start of anmicrobials (BPS).
Remarks: Appropriate roune microbiologic cultures always include at least two sets of blood cultures (aerobic and
anaerobic).

D. ANTIMICROBIAL THERAPY

1. We recommend that administraon of IV anmicrobials should be iniated as soon as possible aer recognion and
within one hour for both sepsis and sepc shock (strong recommendaon, moderate quality of evidence).
2. We recommend empiric broad-spectrum therapy with one or more anmicrobials for paents presenng with sepsis
or sepc shock to cover all likely pathogens (including bacterial and potenally fungal or viral coverage) (strong
recommendaon, moderate quality of evidence).
3. We recommend that empiric anmicrobial therapy be narrowed once pathogen idenficaon and sensivies are
established and/or adequate clinical improvement is noted (BPS).
4. We recommend against sustained systemic anmicrobial prophylaxis in paents with severe inflammatory states of
noninfecous origin (e.g., severe pancreas, burn injury) (BPS).
5. We recommend that dosing strategies of anmicrobials be opmized based on accepted
pharmacokinec/pharmacodynamic principles and specific drug properes in paents with sepsis or sepc shock
 (BPS).
6. We suggest empiric combinaon therapy (using at least two anbiocs of different anmicrobial classes) aimed at the
most likely bacterial pathogen(s) for the inial management of sepc shock (weak recommendaon, low quality of
evidence).
Remarks: Readers should review Table 6 for definions of empiric, targeted/definive, broad-spectrum, combinaon,
and muldrug therapy before reading this secon.
7. We suggest that combinaon therapy not be rounely used for ongoing treatment of most other serious infecons,
including bacteremia and sepsis without shock (weak recommendaon, low quality of evidence).
Remarks: This does not preclude the use of muldrug therapy to broaden anmicrobial acvity.
8. We recommend against combinaon therapy for the roune treatment of neutropenic sepsis/bacteremia (strong
recommendaon, moderate quality of evidence).
Remarks: This does not preclude the use of muldrug therapy to broaden anmicrobial acvity.
9. If combinaon therapy is used for sepc shock, we recommend de-escalaon with disconnuaon of combinaon
therapy within the first few days in response to clinical improvement and/or evidence of infecon resoluon. This
applies to both targeted (for culture-positive infecons) and empiric (for culture-negave infecons) combinaon
therapy (BPS).
10. We suggest that an anmicrobial treatment duraon of 7 to 10 days is adequate for most serious infecons
associated with sepsis and sepc shock (weak recommendaon, low quality of evidence).
11. We suggest that longer courses are appropriate in paents who have a slow clinical response, undrainable foci of
infecon, bacteremia with Staphylococcus aureus, some fungal and viral infecons, or immunologic deficiencies,
including neutropenia (weak recommendaon, low quality of evidence).
12. We suggest that shorter courses are appropriate in some paents, parcularly those with rapid clinical resoluon
following effecve source control of intra-abdominal or urinary sepsis and those with anatomically uncomplicated
pyelonephris (weak recommendaon, low quality of evidence).
13. We recommend daily assessment for de-escalaon of anmicrobial therapy in paents with sepsis and sepc shock
(BPS).
14. We suggest that measurement of procalcitonin levels can be used to support shortening the duraon of anmicrobial
therapy in sepsis paents (weak recommendaon, low quality of evidence).
15. We suggest that procalcitonin levels can be used to support the disconnuaon of empiric anbiocs in paents who
inially appeared to have sepsis, but subsequently have limited clinical evidence of infecon (weak recommendaon,
low quality of evidence).

E. SOURCE CONTROL

1. We recommend that a specific anatomic diagnosis of infecon requiring emergent source control should be idenfied
or excluded as rapidly as possible in paents with sepsis or sepc shock, and that any required source control
intervenon should be implemented as soon as medically and logiscally praccal aer the diagnosis is made (BPS).
2. We recommend prompt removal of intravascular access devices that are a possible source of sepsis or sepc shock
aer other vascular access has been established (BPS).

F. FLUID THERAPY

1. We recommend that a fluid challenge technique be applied where fluid administraon is connued as long as
hemodynamic factors connue to improve (BPS).
2. We recommend crystalloids as the fluid of choice for inial resuscitaon and subsequent intravascular volume
replacement in paents with sepsis and sepc shock (strong recommendaon, moderate quality of evidence).
3. We suggest using either balanced crystalloids or saline for fluid resuscita on of pa ents with sepsis or sep c shock
(weak recommenda on, low quality of evidence).
4. We suggest using albumin in addi on to crystalloids for ini al resuscita on and subsequent intravascular volume
replacement in pa ents with sepsis and sep c shock, when pa ents require substan al amounts of crystalloids (weak
recommenda on, low quality of evidence).
5. We recommend against using hydroxyethyl starches for intravascular volume replacement in pa ents with sepsis or
sep c shock (strong recommenda on, high quality of evidence).
6. We suggest using crystalloids over gela ns when resuscita ng pa ents with sepsis or sep c shock (weak
recommenda on, low quality of evidence).

G. VASOACTIVE MEDICATIONS

1. We recommend norepinephrine as the first-choice vasopressor (strong recommenda on, moderate quality of
evidence).
2. We suggest adding either vasopressin (up to 0.03 U/min) (weak recommenda on, moderate quality of evidence) or
epinephrine (weak recommenda on, low quality of evidence) to norepinephrine with the intent of raising mean
arterial pressure to target, or adding vasopressin (up to 0.03 U/min) (weak recommenda on, moderate quality of
evidence) to decrease norepinephrine dosage.
3. We suggest using dopamine as an alterna ve vasopressor agent to norepinephrine only in highly selected pa ents
(e.g., pa ents with low risk of tachyarrhythmias and absolute or rela ve bradycardia) (weak recommenda on, low
quality of evidence).
4. We recommend against using low-dose dopamine for renal protec on (strong recommenda on, high quality of
evidence).
5. We suggest using dobutamine in pa ents who show evidence of persistent hypoperfusion despite adequate fluid
loading and the use of vasopressor agents (weak recommenda on, low quality of evidence).
Remarks: If ini ated, dosing should be trated to an end point reflec ng perfusion, and the agent reduced or
discon nued in the face of worsening hypotension or arrhythmias.
6. We suggest that all pa ents requiring vasopressors have an arterial catheter placed as soon as prac cal if resources
are available (weak recommenda on, very low quality of evidence).

H. CORTICOSTEROIDS

1. We suggest against using IV hydrocor sone to treat sep c shock pa ents if adequate fluid resuscita on and
vasopressor therapy are able to restore hemodynamic stability. If this is not achievable, we suggest IV hydrocor sone
at a dose of 200 mg per day (weak recommenda on, low quality of evidence).

I. BLOOD PRODUCTS

1. We recommend that RBC transfusion occur only when hemoglobin concentra on decreases to < 7.0 g/dL in adults in
the absence of extenua ng circumstances, such as myocardial ischemia, severe hypoxemia, or acute hemorrhage
(strong recommenda on, high quality of evidence).
2. We recommend against the use of erythropoie n for treatment of anemia associated with sepsis (strong
recommenda on, moderate quality of evidence).
3. We suggest against the use of fresh frozen plasma to correct clong abnormali es in the absence of bleeding or
planned invasive procedures (weak recommenda on, very low quality of evidence).
3 9
4. We suggest prophylac c platelet transfusion when counts are < 10,000/mm (10 10 /L) in the absence of apparent
 3 9
bleeding and when counts are < 20,000/mm (20 10 /L) if the paent has a significant risk of bleeding. Higher
3 9
platelet counts ( 50,000/mm [50 x 10 /L]) are advised for acve bleeding, surgery, or invasive procedures (weak
recommendaon, very low quality of evidence).

J. IMMUNOGLOBULINS

1. We suggest against the use of IV immunoglobulins in paents with sepsis or sepc shock (weak recommendaon, low
quality of evidence).

K. BLOOD PURIFICATION

1. We make no recommendaon regarding the use of blood purificaon techniques.

L. ANTICOAGULANTS

1. We recommend against the use of anthrombin for the treatment of sepsis and sepc shock (strong recommendaon,
moderate quality of evidence).
2. We make no recommendaon regarding the use of thrombomodulin or heparin for the treatment of sepsis or sepc
shock.

M. MECHANICAL VENTILATION

1. We recommend using a target dal volume of 6 mL/kg predicted body weight compared with 12 mL/kg in adult
paents with sepsis-induced acute respiratory distress syndrome (ARDS) (strong recommendaon, high quality of
evidence).
2. We recommend using an upper limit goal for plateau pressures of 30 cm H2O over higher plateau pressures in adult
paents with sepsis-induced severe ARDS (strong recommendaon, moderate quality of evidence).
3. We suggest using higher posive end-expiratory pressure (PEEP) over lower PEEP in adult paents with sepsis-induced
moderate to severe ARDS (weak recommendaon, moderate quality of evidence).
4. We suggest using recruitment maneuvers in adult paents with sepsis-induced, severe ARDS (weak recommendaon,
moderate quality of evidence).
5. We recommend using prone over supine posion in adult paents with sepsis-induced ARDS and a PaO2/FIO2 rao <
150 (strong recommendaon, moderate quality of evidence).
6. We recommend against using high-frequency oscillatory venlaon in adult paents with sepsis-induced ARDS (strong
recommendaon, moderate quality of evidence).
7. We make no recommendaon regarding the use of noninvasive venlaon for paents with sepsis-induced ARDS.
8. We suggest using neuromuscular blocking agents for 48 hours in adult paents with sepsis-induced ARDS and a
PaO2/FIO2 rao < 150 mm Hg (weak recommendaon, moderate quality of evidence).
9. We recommend a conservave fluid strategy for paents with established sepsis-induced ARDS who do not have
evidence of ssue hypoperfusion (strong recommendaon, moderate quality of evidence).
10. We recommend against the use of -2 agonists for the treatment of paents with sepsis-induced ARDS without
bronchospasm (strong recommendaon, moderate quality of evidence).
11. We recommend against the roune use of the pulmonary artery catheter for paents with sepsis-induced ARDS
(strong recommendaon, high quality of evidence).
12. We suggest using lower dal volumes over higher dal volumes in adult paents with sepsis-induced respiratory
failure without ARDS (weak recommendaon, low quality of evidence).
13. We recommend that mechanically venlated sepsis paents be maintained with the head of the bed elevated
between 30 and 45 degrees to limit aspiraon risk and to prevent the development of venlator-associated
pneumonia (strong recommendaon, low quality of evidence).
14. We recommend using spontaneous breathing trials in mechanically venlated paents with sepsis who are ready for
weaning (strong recommendaon, high quality of evidence).
15. We recommend using a weaning protocol in mechanically venlated paents with sepsis-induced respiratory failure
who can tolerate weaning (strong recommendaon, moderate quality of evidence).
N. SEDATION AND ANALGESIA

1. We recommend that connuous or intermient sedaon be minimized in mechanically venlated sepsis paents,
targeng specific traon end points (BPS).

O. GLUCOSE CONTROL

1. We recommend a protocolized approach to blood glucose management in ICU paents with sepsis, commencing
insulin dosing when two consecuve blood glucose levels are > 180 mg/dL. This approach should target an upper
blood glucose level 180 mg/dL rather than an upper target blood glucose level 110 mg/dL (strong
recommendaon, high quality of evidence).
2. We recommend that blood glucose values be monitored every 1 to 2 hours unl glucose values and insulin infusion
rates are stable, then every 4 hours thereaer in paents receiving insulin infusions (BPS).
3. We recommend that glucose levels obtained with point-of-care tesng of capillary blood be interpreted with cauon
because such measurements may not accurately esmate arterial blood or plasma glucose values (BPS).
4. We suggest the use of arterial blood rather than capillary blood for point-of-care tesng using glucose meters if
paents have arterial catheters (weak recommendaon, low quality of evidence).

P. RENAL REPLACEMENT THERAPY

1. We suggest that either connuous or intermient renal replacement therapy (RRT) be used in paents with sepsis
and acute kidney injury (weak recommendaon, moderate quality of evidence).

2. We suggest using connuous therapies to facilitate management of fluid balance in hemodynamically unstable sepc
paents (weak recommendaon, very low quality of evidence).

3. We suggest against the use of RRT in paents with sepsis and acute kidney injury for increase in creanine or oliguria
without other definive indicaons for dialysis (weak recommendaon, low quality of evidence).

Q. BICARBONATE THERAPY

1. We suggest against the use of sodium bicarbonate therapy to improve hemodynamics or to reduce vasopressor
requirements in paents with hypoperfusion-induced lacc acidemia with pH 7.15 (weak recommendaon,
moderate quality of evidence).

R. VENOUS THROMBOEMBOLISM PROPHYLAXIS

1. We recommend pharmacologic prophylaxis (unfraconated heparin [UFH] or low-molecular-weight heparin [LMWH])

 against venous thromboembolism (VTE) in the absence of contraindicaons to the use of these agents (strong
recommendation, moderate quality of evidence).
2. We recommend LMWH rather than UFH for VTE prophylaxis in the absence of contraindicaons to the use of LMWH
(strong recommendaon, moderate quality of evidence).
3. We suggest combinaon pharmacologic VTE prophylaxis and mechanical prophylaxis, whenever possible (weak
recommendaon, low quality of evidence).
4. We suggest mechanical VTE prophylaxis when pharmacologic VTE is contraindicated (weak recommendaon, low
quality of evidence).

S. STRESS ULCER PROPHYLAXIS

1. We recommend that stress ulcer prophylaxis be given to paents with sepsis or sepc shock who have risk factors for
gastrointesnal (GI) bleeding (strong recommendaon, low quality of evidence).
2. We suggest using either proton pump inhibitors or histamine-2 receptor antagonists when stress ulcer prophylaxis is
indicated (weak recommendaon, low quality of evidence).
3. We recommend against stress ulcer prophylaxis in paents without risk factors for GI bleeding (BPS).

T. NUTRITION

1. We recommend against the administraon of early parenteral nutrion alone or parenteral nutrion in combinaon
with enteral feedings (but rather iniate early enteral nutrion) in crically ill paents with sepsis or sepc shock wh o
can be fed enterally (strong recommendaon, moderate quality of evidence).
2. We recommend against the administraon of parenteral nutrion alone or in combinaon with enteral feeds (but
rather to iniate IV glucose and advance enteral feeds as tolerated) over the first 7 days in crically ill paents with
sepsis or sepc shock for whom early enteral feeding is not feasible (strong recommendaon, moderate quality of
evidence).
3. We suggest the early iniaon of enteral feeding rather than a complete fast or only IV glucose in crically ill paents
with sepsis or sepc shock who can be fed enterally (weak recommendaon, low quality of evidence).
4. We suggest either early trophic/hypocaloric or early full enteral feeding in crically ill paents with sepsis or sepc
shock; if trophic/hypocaloric feeding is the inial strategy, then feeds should be advanced according to paent
tolerance (weak recommendaon, moderate quality of evidence).
5. We recommend against the use of omega-3 fay acids as an immune supplement in crically ill paents with sepsis or
sepc shock (strong recommendaon, low quality of evidence).
6. We suggest against rounely monitoring gastric residual volumes in crically ill paents with sepsis or sepc shock
(weak recommendaon, low quality of evidence). However, we suggest measurement of gastric residuals in paents
with feeding intolerance or who are considered to be at high risk of aspiraon (weak recommendaon, very low
quality of evidence).
Remarks: This recommendaon refers to nonsurgical crically ill paents with sepsis or sepc shock.
7. We suggest the use of prokinec agents in crically ill paents with sepsis or sepc shock and feeding intolerance
(weak recommendaon, low quality of evidence).
8. We suggest placement of post-pyloric feeding tubes in crically ill paents with sepsis or sepc shock with feeding
intolerance or who are considered to be at high risk of aspiraon (weak recommendaon, low quality of evidence).
9. We recommend against the use of IV selenium to treat sepsis and sepc shock (strong recommendaon, moderate
quality of evidence).
10. We suggest against the use of arginine to treat sepsis and sepc shock (weak recommendaon, low quality of
evidence).
 11. We recommend against the use of glutamine to treat sepsis and sep
c shock (strong recommenda
on, moderate
quality of evidence).
12. We make no recommenda
on about the use of carni
ne for sepsis and sep
c shock.

U. SETTING GOALS OF CARE

1. We recommend that goals of care and prognosis be discussed with pa

ents and families (BPS).
2. We recommend that goals of care be incorporated into treatment and end-of-life care planning, u
lizing pallia
ve
care principles where appropriate (strong recommenda
on, moderate quality of evidence).
3. We suggest that goals of care be addressed as early as feasible, but no later than within 72 hours of ICU admission
(weak recommenda
on, low quality of evidence).

Appendix 2
Comparison of recommendations from 2012 to 2016

2012 RECOMMENDATIONS 2016 RECOMMENDATIONS

A. INITIAL RESUSCITATION A. INITIAL RESUSCITATION

1. Protocolized, quantave resuscitaon of paents with sepsis-induced ssue 1. Sepsis and sepc shock are medical emergencies, and we recommend that
hypoperfusion (defined in this document as hypotension persisng aer inial treatment and resuscitaon begin immediately (BPS).
fluid challenge or blood lactate concentraon 4 mmol/L). Goals during the first 2. We recommend that, in the resuscitaon from sepsis-induced hypoperfusion, at
6 hours of resuscitaon: least 30 mL/kg of IV crystalloid fluid be given within the first 3 hours (strong
a. Central venous pressure 812 mm Hg recommendaon, low quality of evidence).
b. Mean arterial pressure 65 mm Hg 3. We recommend that, following inial fluid resuscitaon, addional fluids be
c. Urine output 0.5 mL/kg/hr guided by frequent reassessment of hemodynamic status (BPS).
d. Central venous (superior vena cava) or mixed venous oxygen saturaon Remarks: Reassessment should include a thorough clinical examinaon and
70% or 65%, respecvely (grade 1C). evaluaon of available physiologic variables (heart rate, blood pressure, arterial
2. In paents with elevated lactate levels,targeng resuscitaon to normalize oxygen saturaon, respiratory rate, temperature, urine output, and others, as
lactate (grade 2C). available) as well as other noninvasive or invasive monitoring,as available.
4. We recommend further hemodynamic assessment (such as assessing cardiac
funcon) to determine the type of shock if the clinical examinaon does not lead
to a clear diagnosis (BPS).
5. We suggest that dynamic over stac variables be used to predict fluid
responsiveness, where available (weak recommendaon, low quality of
evidence).
6. We recommend an inial target mean arterial pressure of 65 mmHg in paents
with sepc shock requiring vasopressors (strong recommendaon, moderate
quality of evidence).
7. We suggest guiding resuscitaon to normalize lactate in paents with elevated
lactate levels as a marker of ssue hypoperfusion (weak recommendaon, low
quality of evidence).

B. SCREENING FOR SEPSIS AND PERFORMANCE IMPROVEMENT B. SCREENING FOR SEPSIS AND PERFORMANCE IMPROVEMENT

1. Roune screening of potenally infected seriously ill paents for severe sepsis to 1. We recommend that hospitals and hospital systems have a performance
allow earlier implementaon of therapy (grade 1C). improvement program for sepsis,including sepsis screening for acutely ill, high-
2. Hospital-based performance improvement efforts in severe sepsis (UG). risk paents (BPS).

C. DIAGNOSIS C. DIAGNOSIS

1. Cultures as clinically appropriate before anmicrobial therapy if no significant 1. We recommend that appropriate roune microbiologic cultures (including
delay (> 45 min) in the start of anmicrobials (grade 1C). At least 2 sets of blood blood) be obtained before starng anmicrobial therapy in paents with
cultures (both aerobic and anaerobic boles) be obtained before anmicrobial suspected sepsis or sepc shock if doing so results in no substanal delay in the
therapy with at least 1 drawn percutaneously and 1 drawn through each vascular start of anmicrobials (BPS).
access device, unless the device was recently (<48 hrs) inserted (grade 1C). Remarks: Appropriate roune microbiologic cultures always include at least two
2. Use of the 1,3--D-glucan assay (grade 2B), mannan and an-mannan anbody sets of blood cultures (aerobic and anaerobic).
assays (2C), if available,and invasive candidiasis in differenal diagnosis of cause
of infecon.
3. Imaging studies performed promptly to confirm a potenal source of infecon
(UG).
D. ANTIMICROBIAL THERAPY D. ANTIMICROBIAL THERAPY

1. Administraon of effecve IV anmicrobials within the first hour of recognion 1. We recommend that administraon of IV anmicrobials be iniated as soon as
of sepc shock (grade 1B) and severe sepsis without sepc shock (grade 1C) as possible aer recognion and within one hour for both sepsis and sepc shock
the goal of therapy. (strong recommendaon, moderate quality of evidence).
2. Inial empiric aninfecve therapy of one or more drugs that have activity 2. We recommend empiric broad-spectrum therapy with one or more
against all likely pathogens (bacterial and/or fungal or viral) and that penetrate in anmicrobials for paents presenng with sepsis or sepc shock to cover all
adequate concentraons into ssues presumed to be the source of sepsis (grade likely pathogens (including bacterial and potenally fungal or viral coverage)
1B). (strong recommendaon, moderate quality of evidence).
3. Anmicrobial regimen should be reassessed daily for potenal de-escalaon 3. We recommend that anmicrobial therapy is narrowed once pathogen
(grade 1B). idenficaon and sensivies are established and/or adequate clinical
4. Use of low procalcitonin levels or similar biomarkers to assist the clinician in the improvement is noted (BPS).
disconnuaon of empiric anbiocs in paents who inially appeared sepc, 4. We recommend against sustained systemic anmicrobial prophylaxis in paents
but have no subsequent evidence of infecon (grade 2C). with severe inflammatory states of noninfecous origin (e.g., severe pancreas,
5. Combinaon empirical therapy for neutropenic paents with severe sepsis burn injury) (BPS).
(grade 2B) and for paents with difficult-to-treat, muldrug-resistant bacterial 5. We recommend that dosing strategies of anmicrobials be opmized based on
pathogens such as Acinetobacter and Pseudomonas species (grade 2B). For accepted pharmacokinec/pharmacodynamic principles and specific drug
paents with severe infecons associated with respiratory failure and sepc properes in paents with sepsis or sepc shock (BPS).
shock, combinaon therapy with an extended-spectrum -lactam and either an 6. We suggest empiric combinaon therapy (using at least two anbiocs of
aminoglycoside or a fluoroquinolone for Pseudomonas aeruginosa bacteremia different anmicrobial classes) aimed at the most likely bacterial pathogen(s) for
(grade 2B). A combinaon of -lactam and macrolide for paents with sepc the inial management of sepc shock (weak recommendaon, low quality of
shock from bacteremic Streptococcus pneumoniae infecons (grade 2B). evidence).
6. Empiric combinaon therapy should not be administered for more than 3 to 5 Remarks: Readers should review Table 6 for definions of empiric,
days. De-escalaon to the most appropriate single therapy should be performed targeted/definive, broad-spectrum, combinaon, and muldrug therapy before
as soon as the suscepbility profile is known (grade 2B). reading this secon.
7. Duraon of therapy typically 7 to 10 days; longer courses may be appropriate in 7. We suggest that combinaon therapy not be rounely used for ongoing
paents who have a slow clinical response, undrainable foci of infecon, treatment of most other serious infecons, including bacteremia and sepsis
bacteremia with Staphylococcus aureus, some fungal and viral infecons, or without shock (weak recommendaon, low quality of evidence).
immunologic deficiencies, including neutropenia (grade 2C). Remarks: This does not preclude the use of muldrug therapy to broaden
8. Anviral therapy iniated as early as possible in paents with severe sepsis or anmicrobial acvity.
sepc shock of viral origin (grade 2C). 8. We recommend against combinaon therapy for the roune treatment of
9. Anmicrobial agents should not be used in paents with severe inflammatory neutropenic sepsis/bacteremia (strong recommendaon, moderate quality of
states determined to be of noninfecous cause (UG). evidence).
Remarks: This does not preclude the use of muldrug therapy to broaden
anmicrobial acvity.
9. If combinaon therapy is used for sepc shock, we recommend de-escalaon
with disconnuaon of combinaon therapy within the first few days in
response to clinical improvement and/or evidence of infecon resoluon. This
applies to both targeted (for culture-posive infecons) and empiric (for culture-
negave infecons) combinaon therapy (BPS).
10. We suggest that an anmicrobial treatment duraon of 7 to 10 days is adequate
for most serious infecons associated with sepsis and sepc shock (weak
recommendaon, low quality of evidence).
11. We suggest that longer courses are appropriate in paents who have a slow
clinical response, undrainable foci of infecon, bacteremia with Staphylococcus
aureus, some fungal and viral infecons, or immunologic deficiencies, including
neutropenia (Weak recommendaon, low quality of evidence).
12. We suggest that shorter courses are appropriate in some paents, parcularly
those with rapid clinical resoluon following effecve source control of intra-
abdominal or urinary sepsis and those with anatomically uncomplicated
pyelonephris (weak recommendaon, low quality of evidence).
13. We recommend daily assessment for de-escalaon of anmicrobial therapy in
paents with sepsis and sepc shock (BPS).
14. We suggest that measurement of procalcitonin levels can be used to support
shortening the duraon of anmicrobial therapy in sepsis paents (weak
recommendaon, low quality of evidence).
15. We suggest that procalcitonin levels can be used to support the disconnuaon
of empiric anbiocs in paents who inially appeared to have sepsis, but
subsequently have limited clinical evidence of infecon (weak recommendaon,
low quality of evidence).
E. SOURCE CONTROL E. SOURCE CONTROL

1. A specific anatomic diagnosis of infecon requiring consideraon for emergent 1. We recommend that a specific anatomic diagnosis of infecon requiring
source control be sought and diagnosed or excluded as rapidly as possible, and emergent source control should be idenfied or excluded as rapidly as possible
intervenon be undertaken for source control within the first 12 hours aer the in paents with sepsis or sepc shock, and that any required source control
diagnosis is made, if feasible (grade 1C). intervenon should be implemented as soon as medically and logiscally
2. When infected peripancreac necrosis is idenfied as a potenal source of praccal aer the diagnosis is made (BPS).
infecon, definive intervenon is best delayed unl adequate demarcaon of 2. We recommend prompt removal of intravascular access devices that are a
viable and nonviable ssues has occurred (grade 2B). possible source of sepsis or sepc shock aer other vascular access has been
3. When source control in a severely sepc paent is required, the effecve established (BPS).
intervenon associated with the least physiologic insult should be used (e.g.,
percutaneous rather than surgical drainage of an abscess) (UG).
4. If intravascular access devices are a possible source of severe sepsis or sepc
shock, they should be removed promptly aer other vascular access has been
established (UG).

F. FLUID THERAPY F. FLUID THERAPY

1. Crystalloids as the inial fluid of choice in the resuscitaon of severe sepsis and 1. We recommend that a fluid challenge technique be applied where fluid
sepc shock (grade 1B). administraon is connued as long as hemodynamic factors connue to improve
2. Against the use of hydroxyethyl starches for fluid resuscitaon of severe sepsis (BPS).
and sepc shock (grade 1B). 2. We recommend crystalloids as the fluid of choice for inial resuscitaon and
3. Albumin in the fluid resuscitaon of severe sepsis and sepc shock when subsequent intravascular volume replacement in paents with sepsis and sepc
paents require substanal amounts of crystalloids (grade 2C). shock (strong recommendaon, moderate quality of evidence).
4. Inial fluid challenge in paents with sepsis-induced ssue hypoperfusion with 3. We suggest using either balanced crystalloids or saline for fluid resuscitaon of
suspicion of hypovolemia to achieve a minimum of 30 mL/kg of crystalloids (a paents with sepsis or sepc shock (weak recommendaon, low quality of
poron of this may be albumin equivalent). More rapid administraon and evidence).
greater amounts of fluid may be needed in some paents (grade 1C). 4. We suggest using albumin in addion to crystalloids for inial resuscitaon and
5. Fluid challenge technique be applied wherein fluid administraon is connued as subsequent intravascular volume replacement in paents with sepsis and sepc
long as there is hemodynamic improvement either based on dynamic (e.g., shock, when paents require substanal amounts of crystalloids (weak
change in pulse pressure, stroke volume variaon) or stac (e.g., arterial recommendaon, low quality of evidence).
pressure, heart rate) variables (UG). 5. We recommend against using hydroxyethyl starches for intravascular volume
replacement in paents with sepsis or sepc shock (strong recommendaon,
high quality of evidence).
6. We suggest using crystalloids over gelans when resuscitang paents with
sepsis or sepc shock (weak recommendaon, low quality of evidence).

G. VASOACTIVE MEDICATIONS G. VASOACTIVE MEDICATIONS

1. Vasopressor therapy inially to target a mean arterial pressure (MAP) of 65 mm 1. We recommend norepinephrine as the first-choice vasopressor (strong
Hg (grade 1C). recommendaon, moderate quality of evidence).
2. Norepinephrine as the first-choice vasopressor (grade 1B). 2. We suggest adding either vasopressin (up to 0.03 U/min) (weak
3. Epinephrine (added to and potenally substuted for norepinephrine) when an recommendaon, moderate quality of evidence) or epinephrine (weak
addional agent is needed to maintain adequate blood pressure (grade 2B). recommendaon, low quality of evidence) to norepinephrine with the intent of
4. Vasopressin, 0.03 units/minute, can be added to norepinephrine with intent of raising mean arterial pressure to target, or adding vasopressin (up to 0.03
either raising MAP or decreasing norepinephrine dosage (UG). U/min) (weak recommendaon, moderate quality of evidence) to decrease
5. Low-dose vasopressin is not recommended as the single inial vasopressor for norepinephrine dosage.
treatment of sepsis-induced hypotension, and vasopressin doses higher than 3. We suggest using dopamine as an alternave vasopressor agent to
0.030.04 units/minute should be reserved for salvage therapy (failure to norepinephrine only in highly selected paents (e.g., paents with low risk of
achieve adequate MAP with other vasopressor agents) (UG). tachyarrhythmias and absolute or relave bradycardia) (weak recommendaon,
6. Dopamine as an alternave vasopressor agent to norepinephrine only in highly low quality of evidence).
selected paents (e.g., paents with low risk of tachyarrhythmias and absolute 4. We recommend against using low-dose dopamine for renal protecon (strong
or relave bradycardia) (grade 2C). recommendaon, high quality of evidence).
7. Phenylephrine is not recommended in the treatment of sepc shock except in 5. We suggest using dobutamine in paents who show evidence of persistent
circumstances where (a) norepinephrine is associated with serious arrhythmias, hypoperfusion despite adequate fluid loading and the use of vasopressor agents
(b) cardiac output is known to be high and blood pressure persistently low, or (c) (weak recommendaon, low quality of evidence).
as salvage therapy when combined inotrope/vasopressor drugs and low-dose Remarks: If iniated, dosing should be trated to an end point reflecng
vasopressin have failed to achieve MAP target (grade 1C). perfusion, and the agent reduced or disconnued in the face of worsening
8. Low-dose dopamine should not be used for renal protecon (grade 1A). hypotension or arrhythmias.
9. All paents requiring vasopressors have an arterial catheter placed as soon as 6. We suggest that all paents requiring vasopressors have an arterial catheter
praccal if resources are available (UG). placed as soon as praccal if resources are available (weak recommendaon,
10. A trial of dobutamine infusion up to 20 g/kg/min be administered or added to very low quality of evidence).
 vasopressor (if in use) in the presence of (a) myocardial dysfuncon as suggested
by elevated cardiac filling pressures and low cardiac output or (b) ongoing signs
of hypoperfusion, despite achieving adequate intravascular volume and
adequate MAP (grade 1C).
11. Not using a strategy to increase cardiac index to predetermined supranormal
levels (grade 1B).

H. CORTICOSTEROIDS H. CORTICOSTEROIDS

1. Not using IV hydrocorsone to treat adult sepc shock paents if adequate fluid 1. We suggest against using IV hydrocorsone to treat sepc shock paents if
resuscitaon and vasopressor therapy are able to restore hemodynamic stability adequate fluid resuscitaon and vasopressor therapy are able to restore
(see goals for Inial Resuscitaon). In case this is not achievable, we suggest IV hemodynamic stability. If this is not achievable, we suggest IV hydrocorsone at
hydrocorsone alone at a dose of 200 mg/day (grade 2C). a dose of 200 mg per day (weak recommendaon, low quality of evidence).
2. Not using the adrenocorcotropic hormone smulaon test to idenfy adults
with sepc shock who should receive hydrocorsone (grade 2B).
3. In treated paents, hydrocorsone tapered when vasopressors are no longer
required (grade 2D).
4. Corcosteroids not be administered for the treatment of sepsis in the absence of
shock (grade 1D).
5. When hydrocorsone is given, use connuous flow (grade 2D).

I. BLOOD PRODUCTS I. BLOOD PRODUCTS

1. Once ssue hypoperfusion has resolved and in the absence of extenuang 1. We recommend that RBC transfusion occur only when hemoglobin
circumstances, such as myocardial ischemia, severe hypoxemia, acute concentraon decreases to < 7.0 g/dL in adults in the absence of extenuang
hemorrhage, or ischemic heart disease, we recommend that RBC transfusion circumstances, such as myocardial ischemia, severe hypoxemia, or acute
occur only when hemoglobin concentraon decreases to < 7.0 g/dL to target a hemorrhage (trong recommendaon, high quality of evidence).
hemoglobin concentraon of 7.09.0 g/dL in adults (grade 1B). 2. We recommend against the use of erythropoien for treatment of anemia
2. Not using erythropoien as a specific treatment of anemia associated with associated with sepsis (strong recommendaon, moderate quality of evidence).
severe sepsis (grade 1B). 3. We suggest against the use of fresh frozen plasma to correct clong
3. Fresh frozen plasma not be used to correct laboratory clong abnormalies in abnormalies in the absence of bleeding or planned invasive procedures (weak
the absence of bleeding or planned invasive procedures (grade 2D). recommendaon, very low quality of evidence).
3
4. Not using anthrombin for the treatment of severe sepsis and sepc shock 4. We suggest prophylacc platelet transfusion when counts are < 10,000/mm (10
9 3
(grade 1B). 10 /L) in the absence of apparent bleeding and when counts are < 20,000/mm
9
5. In paents with severe sepsis, administer platelets prophylaccally when counts (20 10 /L) if the paent has a significant risk of bleeding. Higher platelet counts
3 9 3 9
are < 10,000/mm (10 10 /L) in the absence of apparent bleeding. We suggest ( 50,000/mm [50 10 /L]) are advised for acve bleeding, surgery, or invasive
3 9
prophylacc platelet transfusion when counts are < 20,000/mm (20 10 /L) if procedures (weak recommendaon, very low quality of evidence).
the paent has a significant risk of bleeding. Higher platelet counts (
3 9
50,000/mm [50 10 /L]) are advised for acve bleeding, surgery, or invasive
procedures (grade 2D).

J. IMMUNOGLOBULINS J. IMMUNOGLOBULINS

1. Not using IV immunoglobulins in adult paents with severe sepsis or sepc shock 1. We suggest against the use of IV immunoglobulins in paents with sepsis or
(grade 2B). sepc shock (weak recommendaon, low quality of evidence).

K. BLOOD PURIFICATION

1. We make no recommendaon regarding the use of blood purificaon

techniques.

L. ANTICOAGULANTS

1. We recommend against the use of anthrombin for the treatment of sepsis and
sepc shock (strong recommendaon, moderate quality of evidence).
2. We make no recommendaon regarding the use of thrombomodulin or heparin
for the treatment of sepsis or sepc shock.
M. MECHANICAL VENTILATION M. MECHANICAL VENTILATION

1. Target a dal volume of 6 mL/kg predicted body weight in paents with sepsis- 1. We recommend using a target dal volume of 6 mL/kg predicted body weight
induced acute respiratory distress syndrome (ARDS) (grade 1A vs. 12 mL/kg). compared with 12 mL/kg in adult paents with sepsis-induced acute respiratory
2. Plateau pressures be measured in paents with ARDS and inial upper-limit goal distress syndrome (ARDS) (strong recommendaon, high quality of evidence).
for plateau pressures in a passively inflated lung be 30 cm H2O (grade 1B). 2. We recommend using an upper limit goal for plateau pressures of 30 cm H2O
3. Posive end-expiratory pressure (PEEP) be applied to avoid alveolar collapse at over higher plateau pressures in adult paents with sepsis-induced severe ARDS
end-expiraon (atelectotrauma) (grade 1B). (strong recommendaon, moderate quality of evidence).
4. Strategies based on higher rather than lower levels of PEEP be used for paents 3. We suggest using higher posive end-expiratory pressure (PEEP) over lower
with sepsis-induced moderate or severe ARDS (grade 2C). PEEP in adult paents with sepsis-induced moderate to severe ARDS (weak
5. Recruitment maneuvers be used in sepsis paents with severe refractory recommendaon, moderate quality of evidence).
hypoxemia (grade 2C). 4. We suggest using recruitment maneuvers in adult paents with sepsis-induced
6. Prone posioning be used in sepsis-induced ARDS paents with a PaO2/FIO2 rao severe ARDS (weak recommendaon, moderate quality of evidence).
100 mm Hg in facilies that have experience with such pracces (grade 2B). 5. We recommend using prone over supine posion in adult paents with sepsis-
7. Mechanically venlated sepsis paents be maintained with the head of the bed induced ARDS and a PaO2/FIO2 rao < 150 (strong recommendaon, moderate
elevated to 3045 degrees to limit aspiraon risk and to prevent the quality of evidence).
development of venlator-associated pneumonia (grade 1B). 6. We recommend against using high-frequency oscillatory venlaon in adult
8. Noninvasive mask venlaon (NIV) be used in that minority of sepsis-induced paents with sepsis-induced ARDS (strong recommendaon, moderate quality of
ARDS paents in whom the benefits of NIV have been carefully considered and evidence).
are thought to outweigh the risks (grade 2B). 7. We make no recommendaon regarding the use of noninvasive venlaon for
9. A weaning protocol be in place, and that mechanically venlated paents with paents with sepsis-induced ARDS.
severe sepsis undergo spontaneous breathing trials regularly to evaluate their 8. We suggest using neuromuscular blocking agents for 48 hours in adult paents
ability to disconnue mechanical venlaon when they sasfy the following with sepsis-induced ARDS and a PaO2/FIO2 rao < 150 mm Hg (weak
criteria: a) arousable, b) hemodynamically stable (without vasopressor agents), recommendaon, moderate quality of evidence).
c) no new potenally serious condions, d) low venlatory and end-expiratory 9. We recommend a conservave fluid strategy for paents with established
pressure requirements, and e) low FIO2 requirements that can be met safely sepsis-induced ARDS who do not have evidence of ssue hypoperfusion (strong
delivered with a face mask or nasal cannula. If the spontaneous breathing trial is recommendaon, moderate quality of evidence).
successful, consideraon should be given for extubaon (grade 1A). 10. We recommend against the use of -2 agonists for the treatment of paents
10. Against the roune use of the pulmonary artery catheter for paents with with sepsis-induced ARDS without bronchospasm (strong recommendaon,
sepsis-induced ARDS (grade 1A). moderate quality of evidence).
11. A conservave rather than liberal fluid strategy for paents with established 11. We recommend against the roune use of the pulmonary artery catheter for
sepsis-induced ARDS who do not have evidence of ssue hypoperfusion (grade paents with sepsis-induced ARDS (strong recommendaon, high quality of
1C). evidence).
12. In the absence of specific indicaons such as bronchospasm, not using -2 12. We suggest using lower dal volumes over higher tidal volumes in adult paents
agonists for treatment of sepsis-induced ARDS (grade 1B). with sepsis-induced respiratory failure without ARDS (weak recommendaon,
low quality of evidence).
13. We recommend that mechanically venlated sepsis paents be maintained with
the head of the bed elevated between 30 and 45 degrees to limit aspiraon risk
and to prevent the development of venlator-associated pneumonia (strong
recommendaon, low quality of evidence).
14. We recommend using spontaneous breathing trials in mechanically venlated
paents with sepsis who are ready for weaning (strong recommendaon, high
quality of evidence).
15. We recommend using a weaning protocol in mechanically venlated paents
with sepsis-induced respiratory failure who can tolerate weaning (strong
recommendaon, moderate quality of evidence).

N. SEDATION and ANALGESIA N. SEDATION AND ANALGESIA

1. Connuous or intermient sedaon be minimized in mechanically venlated 1. We recommend that connuous or intermient sedaon be minimized in
sepsis paents, targeng specific traon end points (grade 1B). mechanically venlated sepsis paents, targeng specific traon endpoints
2. Neuromuscular blocking agents (NMBAs) be avoided if possible in sepc paents (BPS).
without ARDS due to the risk of prolonged neuromuscular blockade following
disconnuaon. If NMBAs must be maintained, either intermient bolus as
required or connuous infusion with train-of-four monitoring of the depth of
blockade should be used (grade 1C).
3. A short course of NMBA of not greater than 48 hours for paents with early
sepsis-induced ARDS and a PaO2/FIO2 rao < 150 mm Hg (grade 2C).
O. GLUCOSE CONTROL O. GLUCOSE CONTROL

1. A protocolized approach to blood glucose management in ICU paents with 1. We recommend a protocolized approach to blood glucose management in ICU
severe sepsis commencing insulin dosing whenconsecuve blood glucose levels paents with sepsis, commencing insulin dosing whentwo consecuve blood
are > 180 mg/dL. This protocolized approach should target an upper blood glucose levels are > 180 mg/dL. This approach should target an upper blood
glucose level 180 mg/dL rather than an upper target blood glucoselevel 110 glucose level 180 mg/dL rather than an upper target blood glucose level 110
mg/dL (grade 1A). mg/dL (strong recommendaon, high quality of evidence).
2. Blood glucose values be monitored every 1to 2 hours unl glucose values and 2. We recommend that blood glucose values be monitored every 1 to 2hours unl
insulin infusion rates are stable and then every 4 hours thereaer (grade 1C). glucose values and insulin infusion rates are stable, then every 4 hours
3. Glucose levels obtained with point-of-care tesng of capillary blood be thereaer in paents receiving insulin infusions (BPS).
interpreted with cauon because such measurements may not accurately 3. We recommend that glucose levels obtained with point-of-care tesng of
esmate arterial blood or plasma glucose values (UG). capillary blood be interpreted with caution because such measurements may not
accurately esmate arterial blood or plasma glucose values (BPS).
4. We suggest the use of arterial blood rather than capillary blood for point-of-care
tesng using glucose meters if paents have arterial catheters (weak
recommendaon, low quality of evidence).

P. RENAL REPLACEMENT THERAPY P. RENAL REPLACEMENT THERAPY

1. Connuous renal replacement therapies and intermient hemodialysis are 1. We suggest that either connuous or intermient renal replacement therapy
equivalent in paents with severe sepsis and acute renal failure (grade 2B). (RRT) be used in paents with sepsis and acute kidneyinjury (weak
2. Use connuous therapies to facilitate management of fluid balance in recommendaon, moderate quality of evidence).
hemodynamically unstable sepc paents (grade 2D). 2. We suggest using connuous therapies to facilitate management of fluid balance
in hemodynamically unstable sepc paents (weak recommendaon, very low
quality of evidence).
3. We suggest against the use of RRT in paents with sepsis and acute kidney injury
for increase in creanine or oliguria without other definive indicaons for
dialysis (weak recommendaon, low quality of evidence).

Q. BICARBONATE THERAPY Q. BICARBONATE THERAPY

1. Not using sodium bicarbonate therapy for the purpose of improving 1. We suggest against the use of sodium bicarbonate therapy toimprove
hemodynamics or reducing vasopressor requirements in paents with hemodynamics or to reduce vasopressor requirements in paents with
hypoperfusion-induced lacc acidemia with pH 7.15 (grade 2B). hypoperfusion-induced lacc acidemia with pH 7.15 (weak recommendaon,
moderate quality of evidence).

R. VENOUS THROMBOEMBOLISM (VTE) PROPHYLAXIS R. VENOUS THROMBOEMBOLISM PROPHYLAXIS

1. Paents with severe sepsisreceive daily pharmacoprophylaxis against venous 1. We recommend pharmacologic prophylaxis (unfraconated heparin [UFH] or
thromboembolism (VTE) (grade 1B). This should be accomplished with daily low-molecular-weight heparin [LMWH]) against venous thromboembolism (VTE)
subcutaneous low-molecular-weight heparin (LMWH) (grade 1B versus twice in the absence of contraindicaons to the use of these agents (strong
daily unfraconated heparin [UFH], grade 2C versus three mes daily UFH). If recommendaon, moderate quality of evidence).
creanine clearance is < 30 mL/min, use dalteparin (grade 1A) or another form 2. We recommend LMWH rather than UFH for VTE prophylaxis in the absence of
of LMWH that has a low degree of renal metabolism (grade 2C) or UFH (grade contraindicaons to the use of LMWH (strong recommendaon, moderate
1A). quality of evidence).
2. Paents with severe sepsis be treated with a combinaon of pharmacologic 3. We suggest combinaon pharmacologic VTE prophylaxis and mechanical
therapy and intermient pneumac compression devices whenever possible prophylaxis, whenever possible (weak recommendaon, low quality of
(grade 2C). evidence).
3. Sepc paents who have a contraindicaon for heparin use (e.g., 4. We suggest mechanical VTE prophylaxis when pharmacologic VTE is
thrombocytopenia, severe coagulopathy, acve bleeding, recent intracerebral contraindicated (weak recommendaon, low quality of evidence).
hemorrhage) not receive pharmacoprophylaxis (grade 1B), but receive
mechanical prophylacc treatment, such as graduated compression stockings or
intermient compression devices (grade 2C), unless contraindicated. When the
risk decreases, start pharmacoprophylaxis (grade 2C).
S. STRESS ULCER PROPHYLAXIS S. STRESS ULCER PROPHYLAXIS

1. Stress ulcer prophylaxis using histamine-2 blocker or proton pump inhibitor be 1. We recommend that stress ulcer prophylaxis be given to paents with sepsis or
given to paents with severe sepsisor sepc shock who have bleeding risk sepc shock who have risk factors for gastrointesnal (GI) bleeding (strong
factors (grade 1B). recommendaon, low quality of evidence).
2. When stress ulcer prophylaxis is used, proton pump inhibitors rather than 2. We suggest using either proton pump inhibitors orhistamine-2 receptor
histamine-2 receptor antagonists (grade 2D). antagonists when stress ulcer prophylaxis is indicated w
( eak recommendaon,
3. Paents without risk factors do not receive prophylaxis (grade 2B). low quality of evidence).
3. We recommend against stress ulcer prophylaxis in paents without risk factors
for GI bleeding (BPS).

T. NUTRITION T. NUTRITION

1. Administer oral or enteral (if necessary) feedings, as tolerated, rather than either 1. We recommend against the administraon of early parenteral nutrion alone or
complete fasng or provision of onlyIV glucose within the first 48 hours aer a parenteral nutrion in combinaon with enteral feedings (but ratheriniate
diagnosis of severe sepsis or sepc shock (grade 2C). early enteral nutrion) in crically ill paents with sepsis or sepc shock who can
2. Avoid mandatory full caloric feeding in thefirst week but rather suggest low- be fed enterally (strong recommendaon, moderate quality of evidence).
dose feeding (e.g., up to 500 calories per day), advancing only as tolerated 2. We recommend against the administraon of parenteral nutrion alone or in
(grade 2B). combinaon with enteral feeds (but rather to iniate IV glucose and advance
3. Use IV glucose and enteral nutrion rather than total parenteral nutrion alone enteral feeds as tolerated) over the first 7 days in crically ill paents with sepsis
or parenteral nutrion in conjuncon with enteralfeeding in the first 7 days or sepc shock for whom early enteral feeding is not feasible (strong
aer a diagnosis of severe sepsisor sepc shock (grade 2B). recommendaon, moderate quality of evidence).
4. Use nutrion with no specific immunomodulang supplementaon rather than 3. We suggest the early iniaon of enteral feeding rather than a complete fast or
nutrion providing specific immunomodulang supplementaon in paents with only IV glucose in crically ill paents with sepsis or sepc shock who can be fed
severe sepsis (grade 2C). enterally (weak recommendaon, low quality of evidence).
5. Not using IV selenium for the treatment of severe sepsis (grade 2C). 4. We suggest either early trophic/hypocaloric or early full enteral feeding in
crically ill paents with sepsis or sepc shock; iftrophic/hypocaloric feeding is
the inial strategy, then feeds should be advanced according to paent
tolerance (weak recommendaon, moderate quality of evidence).
5. We recommend against the use of omega-3 fay acids as an immune
supplement in crically ill paents with sepsis or sepc shock s( trong
recommendaon, low quality of evidence).
6. We suggest against rounely monitoring gastric residual volumes in crically ill
paents with sepsis or sepc shock (weak recommendaon, low quality of
evidence). However, we suggest measurement of gastric residuals in paents
with feeding intolerance or who are considered to beat high risk of aspiraon
(weak recommendaon, very low quality of evidence).
Remarks: This recommendaon refers to nonsurgical crically ill paents with
sepsis or sepc shock.
7. We suggest the use of prokinec agents in crically ill paents with sepsis or
sepc shock and feeding intolerance (weak recommendaon, low quality of
evidence).
8. We suggest placement of post-pyloric feeding tubes in crically ill paents with
sepsis or sepc shock with feeding intolerance or who are considered to be at
high risk of aspiraon (weak recommendaon, low quality of evidence).
9. We recommend against the use of IV selenium to treat sepsis and sepc shock
(strong recommendaon, moderate quality of evidence).
10. We suggest against the use of arginine to treat sepsis and sepc shock w ( eak
recommendaon, low quality of evidence).
11. We recommend against the use of glutamine to treat sepsis and sepc shock
(strong recommendaon, moderate quality of evidence).
12. We make no recommendaon about the use of carnine for sepsis and sepc
shock.

U. SETTING GOALS OF CARE U. SETTING GOALS OF CARE

1. Discuss goals of care and prognosis with paents and families (grade 1B). 1. We recommend that goals of care and prognosis be discussed with paents and
2. Incorporate goals of care into treatment andend-of-life care planning, ulizing families (BPS).
palliave care principles where appropriate (grade 1B). 2. We recommend that goals of care be incorporated into treatment and end-of-
3. Address goals of care as early as feasible, but no later than within 72 hours of life care planning, ulizing palliave care principles where appropriate s(trong
ICU admission (grade 2C). recommendaon, moderate quality of evidence).
3. We suggest that goals of care be addressed as early as feasible, but no later than
within 72 hours of ICU admission (weak recommendaon, low quality of
evidence).
Received: 2 January 2017 Accepted: 6 January 2017 European Society of Intensive Care Medicine. Intensive Care Med
40(12):17951815
23. Eskesen TG, Wetterslev M, Perner A (2016) Systematic review including
re-analyses of 1148 individual data sets of central venous pressure as a
predictor of fluid responsiveness. Intensive Care Med 42(3):324332
24. Monnet X, Marik P, Teboul JL (2016) Passive leg raising for predicting
References fluid responsiveness: a systematic review and meta-analysis. Intensive
1. Singer M, Deutschman CS, Seymour CW etal (2016) The Third Inter Care Med 42(12):19351947
national Consensus Definitions for Sepsis and Septic Shock (Sepsis-3). 25. Cecconi M, Hofer C, Teboul JL etal (2015) Fluid challenges in intensive
JAMA 315(8):801810 care: the FENICE study: a global inception cohort study. Intensive Care
2. Shankar-Hari M, Phillips GS, Levy ML etal (2016) Developing a new Med 41(9):15291537
definition and assessing new clinical criteria for septic shock: for the 26. LeDoux D, Astiz ME, Carpati CM, Rackow EC (2000) Effects of perfu
Third International Consensus Definitions for Sepsis and Septic Shock sion pressure on tissue perfusion in septic shock. Crit Care Med
(Sepsis-3). JAMA 315(8):775787 28(8):27292732
3. Seymour CW, Liu VX, Iwashyna TJ etal (2016) Assessment of clinical 27. Bourgoin A, Leone M, Delmas A, Garnier F, Albanese J, Martin C (2005)
criteria for sepsis: for the Third International Consensus Definitions for Increasing mean arterial pressure in patients with septic shock: effects
Sepsis and Septic Shock (Sepsis-3). JAMA 315(8):762774 on oxygen variables and renal function. Crit Care Med 33(4):780786
4. Angus DC, Linde-Zwirble WT, Lidicker J, Clermont G, Carcillo J, Pinsky 28. Thooft A, Favory R, Salgado DR etal (2011) Effects of changes in
MR (2001) Epidemiology of severe sepsis in the United States: analysis arterial pressure on organ perfusion during septic shock. Crit Care
of incidence, outcome, and associated costs of care. Crit Care Med 15(5):R222
29(7):13031310 29. Asfar P, Meziani F, Hamel JF etal (2014) High versus low blood-pressure
5. Dellinger RP (2003) Cardiovascular management of septic shock. Crit target in patients with septic shock. N Engl J Med 370(17):15831593
Care Med 31(3):946955 30. Lamontagne F, Meade MO, Hebert PC etal (2016) Higher versus lower
6. Martin GS, Mannino DM, Eaton S, Moss M (2003) The epidemiology blood pressure targets for vasopressor therapy in shock: a multicentre
of sepsis in the United States from 1979 through 2000. N Engl J Med pilot randomized controlled trial. Intensive Care Med 42(4):542550
348(16):15461554 31. Levy B (2006) Lactate and shock state: the metabolic view. Curr Opin
7. Levy MM, Fink MP, Marshall JC etal (2003) 2001 SCCM/ESICM/ACCP/ Crit Care. 12(4):315321
ATS/SIS International Sepsis Definitions Conference. Crit Care Med 32. Casserly B, Phillips GS, Schorr C etal (2015) Lactate measurements in
31(4):12501256 sepsis-induced tissue hypoperfusion: results from the Surviving Sepsis
8. Dellinger RP, Levy MM, Rhodes A etal (2013) Surviving Sepsis Cam Campaign database. Crit Care Med 43(3):567573
paign: international guidelines for management of severe sepsis and 33. Jansen TC, van Bommel J, Schoonderbeek FJ etal (2010) Early
septic shock, 2012. Intensive Care Med 39(2):165228 lactate-guided therapy in intensive care unit patients: a multicenter,
9. Dellinger RP, Levy MM, Rhodes A etal (2013) Surviving Sepsis Cam open-label, randomized controlled trial. Am J Respir Crit Care Med
paign: international guidelines for management of severe sepsis and 182(6):752761
septic shock: 2012. Crit Care Med 41(2):580637 34. Jones AE, Shapiro NI, Trzeciak S etal (2010) Lactate clearance vs central
10. Dellinger RP, Carlet JM, Masur H etal (2004) Surviving Sepsis Campaign venous oxygen saturation as goals of early sepsis therapy: a rand
guidelines for management of severe sepsis and septic shock. Crit Care omized clinical trial. JAMA 303(8):739746
Med 32(3):858873 35. Lyu X, Xu Q, Cai G, Yan J, Yan M (2015) Efficacies of fluid resuscitation as
11. Dellinger RP, Levy MM, Carlet JM (2008) Surviving Sepsis Campaign: guided by lactate clearance rate and central venous oxygen saturation
international guidelines for management of severe sepsis and septic in patients with septic shock. Zhonghua Yi Xue Za Zhi. 95(7):496500
shock: 2008. Crit Care Med 36(1):296327 36. Tian HH, Han SS, Lv CJ etal (2012) The effect of early goal lactate
12. Dellinger RP, Levy MM, Carlet JM etal (2008) Surviving Sepsis Cam clearance rate on the outcome of septic shock patients with severe
paign: international guidelines for management of severe sepsis and pneumonia. Zhongguo Wei Zhong Bing Ji Jiu Yi Xue 24(1):4245
septic shock: 2008. Crit Care Med 36(1):296327 37. Yu B, Tian HY, Hu ZJ etal (2013) Comparison of the effect of fluid resus
13. Guyatt GH, Oxman AD, Kunz R etal (2011) GRADE guidelines: 2. citation as guided either by lactate clearance rate or by central venous
Framing the question and deciding on important outcomes. J Clin oxygen saturation in patients with sepsis. Zhonghua Wei Zhong Bing Ji
Epidemiol 64(4):395400 Jiu Yi Xue. 25(10):578583
14. Guyatt GH, Oxman AD, Vist GE etal (2008) GRADE: an emerging con 38. Gu WJ, Zhang Z, Bakker J (2015) Early lactate clearance-guided therapy
sensus on rating quality of evidence and strength of recommendations. in patients with sepsis: a meta-analysis with trial sequential analysis of
BMJ 336(7650):924926 randomized controlled trials. Intensive Care Med 41(10):18621863
15. Guyatt GH, Schunemann HJ, Djulbegovic B, Akl EA (2015) Guideline 39. Simpson SQ, Gaines M, Hussein Y, Badgett RG (2016) Early goal-directed
panels should not GRADE good practice statements. J Clin Epidemiol therapy for severe sepsis and septic shock: a living systematic review. J
68(5):597600 Crit Care 36:4348
16. Rivers E, Nguyen B, Havstad S etal (2001) Early goal-directed therapy 40. Dellinger RP (2015) Foreword. The future of sepsis performance
in the treatment of severe sepsis and septic shock. N Engl J Med improvement. Crit Care Med 43(9):17871789
345(19):13681377 41. Murphy DJ, Ogbu OC, Coopersmith CM (2015) ICU director data: using
17. Peake SL, Delaney A, Bailey M etal (2014) Goal-directed resuscitation data to assess value, inform local change, and relate to the external
for patients with early septic shock. N Engl J Med 371(16):14961506 world. Chest 147(4):11681178
18. Yealy DM, Kellum JA, Huang DT etal (2014) A randomized trial of proto 42. Black MD, Schorr C, Levy MM (2012) Knowledge translation and the
col-based care for early septic shock. N Engl J Med 370(18):16831693 multifaceted intervention in the intensive care unit. Crit Care Med
19. Mouncey PR, Osborn TM, Power GS etal (2015) Trial of early, goal- 40(4):13241328
directed resuscitation for septic shock. N Engl J Med 372(14):13011311 43. Gatewood MO, Wemple M, Greco S, Kritek PA, Durvasula R (2015) A
20. Levy MM, Dellinger RP, Townsend SR etal (2010) The Surviving Sepsis quality improvement project to improve early sepsis care in the emer
Campaign: results of an international guideline-based performance gency department. BMJ Qual Saf 24(12):787795
improvement program targeting severe sepsis. Crit Care Med 44. Hayden GE, Tuuri RE, Scott R etal (2016) Triage sepsis alert and sepsis
38(2):367374 protocol lower times to fluids and antibiotics in the ED. Am J Emerg Med
21. Levy MM, Rhodes A, Phillips GS etal (2015) Surviving Sepsis Campaign: 34(1):19
association between performance metrics and outcomes in a 7.5-year 45. Jones SL, Ashton CM, Kiehne L etal (2015) Reductions in sepsis
study. Crit Care Med 43(1):312 mortality and costs after design and implementation of a nurse-based
22. Cecconi M, De Backer D, Antonelli M etal (2014) Consensus on early recognition and response program. Jt Comm J Qual Patient Saf.
circulatory shock and hemodynamic monitoring. Task force of the 41(11):483491
 46. Levy MM, Pronovost PJ, Dellinger RP etal (2004) Sepsis change bundles: 66. OGrady NP, Barie PS, Bartlett JG etal (2008) Guidelines for evaluation of
converting guidelines into meaningful change in behavior and clinical new fever in critically ill adult patients: 2008 update from the American
outcome. Crit Care Med 32(11 Suppl):S595S597 College of Critical Care Medicine and the Infectious Diseases Society of
47. Damiani E, Donati A, Serafini G etal (2015) Effect of performance America. Crit Care Med 36(4):13301349
improvement programs on compliance with sepsis bundles and mor 67. Mermel LA, Allon M, Bouza E etal (2009) Clinical practice guidelines for
tality: a systematic review and meta-analysis of observational studies. the diagnosis and management of intravascular catheter-related infec
PLoS One 10(5):e0125827 tion: 2009 update by the Infectious Diseases Society of America. Clin
48. Rhodes A, Phillips G, Beale R etal (2015) The Surviving Sepsis Campaign Infect Dis 49(1):145
bundles and outcome: results from the International Multicentre 68. Boyce JM, Nadeau J, Dumigan D etal (2013) Obtaining blood cultures
Prevalence Study on Sepsis (the IMPreSS study). Intensive Care Med by venipuncture versus from central lines: impact on blood culture
41(9):16201628 contamination rates and potential effect on central line-associated
49. Zadroga R, Williams DN, Gottschall R etal (2013) Comparison of 2 blood bloodstream infection reporting. Infect Control Hosp Epidemiol
culture media shows significant differences in bacterial recovery for 34(10):10421047
patients on antimicrobial therapy. Clin Infect Dis 56(6):790797 69. Beekmann SE, Diekema DJ, Huskins WC etal (2012) Diagnosing and
50. Kanegaye JT, Soliemanzadeh P, Bradley JS (2001) Lumbar puncture in reporting of central line-associated bloodstream infections. Infect
pediatric bacterial meningitis: defining the time interval for recovery of Control Hosp Epidemiol 33(9):875882
cerebrospinal fluid pathogens after parenteral antibiotic pretreatment. 70. Garcia RA, Spitzer ED, Beaudry J etal (2015) Multidisciplinary team
Pediatrics 108(5):11691174 review of best practices for collection and handling of blood cultures to
51. Pollack LA, van Santen KL, Weiner LM, Dudeck MA, Edwards JR, determine effective interventions for increasing the yield of true-posi
Srinivasan A (2016) Antibiotic stewardship programs in U.S. acute care tive bacteremias, reducing contamination, and eliminating false-posi
hospitals: findings from the 2014 National Healthcare Safety Network tive central line-associated bloodstream infections. Am J Infect Control
Annual Hospital Survey. Clin Infect Dis. 63(4):443449 43(11):12221237
52. Cardoso T, Carneiro A, Ribeiro O, Teixeira-Pinto A, Costa-Pereira A (2010) 71. Vincent JL, Brealey D, Libert N etal (2015) Rapid diagnosis of infection
Reducing mortality in severe sepsis with the implementation of a core in the critically ill, a multicenter study of molecular detection in blood
6-hour bundle: results from the Portuguese community-acquired sepsis stream infections, pneumonia, and sterile site infections. Crit Care Med
study (SACiUCI study). Crit Care 14(3):R83 43(11):22832291
53. De Sousa AG, Fernandes Junior CJ, Santos GPD etal (2008) The 72. Makristathis A, Riss S, Hirschl AM (2014) A novel fluorescence insitu
impact of each action in the Surviving Sepsis Campaign measures on hybridization test for rapid pathogen identification in positive blood
hospital mortality of patients with severe sepsis/septic shock. Einstein. cultures. Clin Microbiol Infect 20(10):O760O763
6(3):323327 73. Tissari P, Zumla A, Tarkka E etal (2010) Accurate and rapid identification
54. Garnacho-Montero J, Gutirrez-Pizarraya A, Escoresca-Ortega A etal of bacterial species from positive blood cultures with a DNA-based
(2013) De-escalation of empirical therapy is associated with lower microarray platform: an observational study. Lancet 375(9710):224230
mortality in patients with severe sepsis and septic shock. Intensive Care 74. Ferrer R, Martin-Loeches I, Phillips G etal (2014) Empiric antibiotic
Med 40(1):32-40. doi:10.1007/s00134-013-3077-7 treatment reduces mortality in severe sepsis and septic shock from the
55. Weiss CH, Persell SD, Wunderink RG, Baker DW (2012) Empiric antibiotic, first hour: results from a guideline-based performance improvement
mechanical ventilation, and central venous catheter duration as poten program. Crit Care Med 42(8):17491755
tial factors mediating the effect of a checklist prompting intervention 75. Zhang D, Micek ST, Kollef MH (2015) Time to appropriate antibiotic ther
on mortality: an exploratory analysis. BMC Health Serv Res. 12(1):1 apy is an independent determinant of postinfection ICU and hospital
56. Ferrer R, Artigas A, Suarez D etal (2009) Effectiveness of treatments lengths of stay in patients with sepsis. Crit Care Med 43(10):21332140
for severe sepsis: a prospective, multicenter, observational study. Am J 76. Bagshaw SM, Lapinsky S, Dial S etal (2009) Acute kidney injury in septic
Respir Crit Care Med 180(9):861866 shock: clinical outcomes and impact of duration of hypotension prior
57. Kumar A, Roberts D, Wood KE etal (2006) Duration of hypotension to initiation of antimicrobial therapy. Intensive Care Med 35(5):871881
before initiation of effective antimicrobial therapy is the critical determi 77. Iscimen R, Cartin-Ceba R, Yilmaz M etal (2008) Risk factors for the
nant of survival in human septic shock. Crit Care Med 34(6):15891596 development of acute lung injury in patients with septic shock: an
58. Vaughn VM, Chopra V (2016) Revisiting the panculture. BMJ Qual Saf. observational cohort study. Crit Care Med 36(5):15181522
doi:10.1136/bmjqs-2015-004821 (Epub ahead of print) 78. Garnacho-Montero J, Aldabo-Pallas T, Garnacho-Montero C etal (2006)
59. Weinstein MP, Reller LB, Murphy JR, Lichtenstein KA (1983) The clinical Timing of adequate antibiotic therapy is a greater determinant of
significance of positive blood cultures: a comprehensive analysis of outcome than are TNF and IL-10 polymorphisms in patients with sepsis.
500 episodes of bacteremia and fungemia in adults. I. Laboratory and Crit Care 10(4):R111
epidemiologic observations. Rev Infect Dis 5(1):3553 79. Barie PS, Hydo LJ, Shou J, Larone DH, Eachempati SR (2005) Influence
60. Li J, Plorde JJ, Carlson LG (1994) Effects of volume and periodicity on of antibiotic therapy on mortality of critical surgical illness caused or
blood cultures. J Clin Microbiol 32(11):28292831 complicated by infection. Surg Infect. 6(1):4154
61. Baron EJ, Miller JM, Weinstein MP etal (2013) A guide to utilization of 80. Barochia AV, Cui X, Vitberg D etal (2010) Bundled care for septic shock:
the microbiology laboratory for diagnosis of infectious diseases: 2013 an analysis of clinical trials. Crit Care Med 38(2):668678
recommendations by the Infectious Diseases Society of America (IDSA) 81. Gaieski DF, Mikkelsen ME, Band RA etal (2010) Impact of time to antibi
and the American Society for Microbiology (ASM)(a). Clin Infect Dis otics on survival in patients with severe sepsis or septic shock in whom
57(4):e22e121 early goal-directed therapy was initiated in the emergency depart
62. OGrady NP, Alexander M, Burns LA etal (2011) Guidelines for the ment. Crit Care Med 38(4):10451053
prevention of intravascular catheter-related infections. Clin Infect Dis 82. Kumar A (2016) Systematic bias in meta-analyses of time to antimicro
52(9):e162e193 bial in sepsis studies. Crit Care Med 44(4):e234e235
63. Blot F, Schmidt E, Nitenberg G etal (1998) Earlier positivity of central 83. Shirakura Y, Kuriyama A (2016) Timing of antibiotic administration in
venous versus peripheral blood cultures is highly predictive of catheter- sepsis and septic shock: the impact that a meta-analysis does not
related sepsis. J Clin Microbiol 36(1):105109 depict. Crit Care Med 44(10):e1004
64. Kaasch AJ, Rieg S, Hellmich M, Kern WV, Seifert H (2014) Differential 84. Kaasch AJ, Rieg S, Kuetscher J etal (2013) Delay in the administra
time to positivity is not predictive for central line-related Staphylo- tion of appropriate antimicrobial therapy in Staphylococcus aureus
coccus aureus bloodstream infection in routine clinical care. J Infect bloodstream infection: a prospective multicenter hospital-based cohort
68(1):5861 study. Infection 41(5):979985
65. Malgrange VB, Escande MC, Theobald S (2001) Validity of earlier positiv 85. Corona A, Bertolini G, Lipman J, Wilson AP, Singer M (2010) Antibiotic
ity of central venous blood cultures in comparison with peripheral use and impact on outcome from bacteraemic critical illness: the
blood cultures for diagnosing catheter-related bacteremia in cancer BActeraemia Study in Intensive Care (BASIC). J Antimicrob Chemother
patients. J Clin Microbiol 39(1):274278 65(6):12761285
 86. Giner AM, Kuster SP, Zbinden R, Ruef C, Ledergerber B, Weber R (2011) working party of the German Society of Hematology and Oncology.
Initial management of and outcome in patients with pneumococcal Ann Oncol 22(5):10191029
bacteremia: a retrospective study at a Swiss university hospital, 2003- 108. Solomkin JS, Mazuski JE, Bradley JS etal (2010) Diagnosis and manage
2009. Infection 39(6):519526 ment of complicated intra-abdominal infection in adults and children:
87. Lin MY, Weinstein RA, Hota B (2008) Delay of active antimicrobial guidelines by the Surgical Infection Society and the Infectious Diseases
therapy and mortality among patients with bacteremia: impact of Society of America. Surg Infect. 11(1):79109
severe neutropenia. Antimicrob Agents Chemother 52(9):31883194 109. Stevens DL, Bisno AL, Chambers HF etal (2014) Practice guidelines for
88. Amaral AC, Fowler RA, Pinto R etal (2016) Patient and organizational the diagnosis and management of skin and soft tissue infections: 2014
factors associated with delays in antimicrobial therapy for septic shock. update by the Infectious Diseases Society of America. Clin Infect Dis
Crit Care Med 44(12):21452153 59(2):147159
89. Funk DJ, Kumar A (2011) Antimicrobial therapy for life-threatening 110. Micek ST, Welch EC, Khan J etal (2010) Empiric combination antibiotic
infections: speed is life. Crit Care Clin 27(1):5376 therapy is associated with improved outcome in gram-negative sepsis:
90. Petitpas F, Guenezan J, Vendeuvre T, Scepi M, Oriot D, Mimoz O (2016) a retrospective analysis. Antimicrob Agents Chemother (Bethesda).
Use of intra-osseous access in adults: a systematic review. Crit Care 54(5):17421748
20(1):102 111. Pittet D, Monod M, Suter PM etal (1994) Candida colonization and
91. Buck ML, Wiggins BS, Sesler JM (2007) Intraosseous drug administra subsequent infections in critically ill surgical patients. Ann Surg
tion in children and adults during cardiopulmonary resuscitation. Ann 220(6):751758
Pharmacother 41(10):16791686 112. Blumberg HM, Jarvis WR, Soucie JM etal (2001) Risk factors for candidal
92. Romanelli G, Cravarezza P, Group IIMS (1995) Intramuscular meropenem bloodstream infections in surgical intensive care unit patients: the
in the treatment of bacterial infections of the urinary and lower respira NEMIS prospective multicenter study. The National Epidemiology of
tory tracts. J Antimicrob Chemother 36(suppl A):109119 Mycosis Survey. Clin Infect Dis 33(2):177186
93. Cormio L, Berardi B, Callea A etal (2002) Antimicrobial prophylaxis for 113. Green DL (2005) Selection of an empiric antibiotic regimen for hospital-
transrectal prostatic biopsy: a prospective study of ciprofloxacin vs acquired pneumonia using a unit and culture-type specific antibio
piperacillin/tazobactam. BJU Int 90(7):700702 gram. J Intensive Care Med. 20(5):296301
94. Barbhaiya RH, Knupp CA, Tenney J, Martin RR, Weidler DJ, Pittman 114. Kaufman D, Haas CE, Edinger R, Hollick G (1998) Antibiotic susceptibil
KA (1990) Safety, tolerance, and pharmacokinetics of cefepime ity in the surgical intensive care unit compared with the hospital-wide
administered intramuscularly to healthy subjects. J Clin Pharmacol antibiogram. Arch Surg 133(10):10411045
30(10):900910 115. Kerremans JJ, Verbrugh HA, Vos MC (2012) Frequency of microbiologi
95. Kumar A, Ellis P, Arabi Y etal (2009) Initiation of inappropriate antimicro cally correct antibiotic therapy increased by infectious disease consulta
bial therapy results in a five-fold reduction of survival in human septic tions and microbiological results. J Clin Microbiol 50(6):20662068
shock. Chest 136(5):12371248 116. Raineri E, Pan A, Mondello P, Acquarolo A, Candiani A, Crema L (2008)
96. Ibrahim EH, Sherman G, Ward S, Fraser VJ, Kollef MH (2000) The influ Role of the infectious diseases specialist consultant on the appropriate
ence of inadequate antimicrobial treatment of bloodstream infections ness of antimicrobial therapy prescription in an intensive care unit. Am
on patient outcomes in the ICU setting. Chest 118(1):146155 J Infect Control 36(4):283290
97. Paul M, Shani V, Muchtar E, Kariv G, Robenshtok E, Leibovici L (2010) 117. Bai AD, Showler A, Burry L etal (2015) Impact of infectious disease
Systematic review and meta-analysis of the efficacy of appropriate consultation on quality of care, mortality, and length of stay in Staphy
empiric antibiotic therapy for sepsis. Antimicrob Agents Chemother lococcus aureus bacteremia: results from a large multicenter cohort
54(11):48514863 study. Clin Infect Dis 60(10):14511461
98. Kreger BE, Craven DE, McCabe WR (1980) Gram-negative bacteremia. IV. 118. Guo Y, Gao W, Yang H, Ma C, Sui S (2016) De-escalation of empiric anti
Re-evaluation of clinical features and treatment in 612 patients. Am J biotics in patients with severe sepsis or septic shock: a meta-analysis.
Med 68(3):344355 Heart Lung 45(5):454459
99. Mermel LA, Maki DG (1993) Detection of bacteremia in adults: conse 119. Bernard GR, Vincent JL, Laterre P etal (2001) Efficacy and safety of
quences of culturing an inadequate volume of blood. Ann Intern Med recombinant human activated protein C for severe sepsis. N Engl J Med
119(4):270272 344(10):699709
100. Bow EJ, Evans G, Fuller J etal (2010) Canadian clinical practice guide 120. Klein Klouwenberg PM, Cremer OL, van Vught LA etal (2015) Likelihood
lines for invasive candidiasis in adults. Can J Infect Dis Med Microbiol. of infection in patients with presumed sepsis at the time of intensive
21(4):e122e150 care unit admission: a cohort study. Crit Care 19:319
101. Connolly S (2011) Clinical Practice Guidelines: Burn Patient Manage 121. Working Gropu IAP/APA Acute Pancreatitis Guidelines (2013) IAP/APA
ment. ACI Statewide Burn Injury Service. Chatswood, NSW, Australia: evidence-based guidelines for the management of acute pancreatitis.
NSW Agency for Clinical Innovation Pancreatology. 13(4):e1e15
102. Cornely OA, Bassetti M, Calandra T etal (2012) ESCMID guideline for 122. Wittau M, Mayer B, Scheele J, Henne-Bruns D, Dellinger EP, Isenmann R
the diagnosis and management of Candida diseases 2012: non-neutro (2011) Systematic review and meta-analysis of antibiotic prophylaxis in
penic adult patients. Clin Microbiol Infect 18(Suppl 7):1937 severe acute pancreatitis. Scand J Gastroenterol 46(3):261270
103. Kalil AC, Metersky ML, Klompas M etal (2016) Management of adults 123. Avni T, Levcovich A, Ad-El DD, Leibovici L, Paul M (2010) Prophylactic
with hospital-acquired and ventilator-associated pneumonia: 2016 Clin antibiotics for burns patients: systematic review and meta-analysis. BMJ
ical Practice Guidelines by the Infectious Diseases Society of America 340:c241
and the American Thoracic Society. Clin Infect Dis 63(5):e61e111 124. BarajasNava LA, LpezAlcalde J, Roqu i Figuls M, Sol I, Bonfill Cosp
104. Liu C, Bayer A, Cosgrove SE etal (2011) Clinical Practice Guidelines X (2013) Antibiotic prophylaxis for preventing burn wound infection.
by the Infectious Diseases Society of America for the treatment of Cochrane Database Syst Rev 6:CD008738
methicillin-resistant Staphylococcus aureus infections in adults and 125. Chelluri L, Jastremski MS (1987) Inadequacy of standard ami
children. Clin Infect Dis 52(3):e18e55 noglycoside loading doses in acutely ill patients. Crit Care Med
105. Pappas PG, Kauffman CA, Andes DR etal (2016) Clinical practice guide 15(12):11431145
line for the management of candidiasis: 2016 update by the Infectious 126. Pletz M, Bloos F, Burkhardt O etal (2010) Pharmacokinetics of moxi
Diseases Society of America. Clin Infect Dis 62(4):e1e50 floxacin in patients with severe sepsis or septic shock. Intensive Care
106. Penack O, Becker C, Buchheidt D etal (2014) Management of sepsis in Med 36(6):979983
neutropenic patients: 2014 updated guidelines from the Infectious Dis 127. Van Zanten AR, Polderman KH, van Geijlswijk IM, van der Meer GY,
eases Working Party of the German Society of Hematology and Medical Schouten MA, Girbes AR (2008) Ciprofloxacin pharmacokinetics in criti
Oncology (AGIHO). Ann Hematol 93(7):10831095 cally ill patients: a prospective cohort study. J Crit Care 23(3):422430
107. Penack O, Buchheidt D, Christopeit M etal (2011) Management of 128. Blot S, Koulenti D, Akova M etal (2014) Does contemporary vanco
sepsis in neutropenic patients: guidelines from the infectious diseases mycin dosing achieve therapeutic targets in a heterogeneous clinical
 cohort of critically ill patients? Data from the multinational DALI study. population pharmacokinetics, protein binding, and prediction of bacte
Crit Care 18(3):112 rial kill. Antimicrob Agents Chemother 56(8):42414249
129. Moore RD, Smith CR, Lietman PS (1984) Association of aminoglycoside 149. Pea F, Brollo L, Viale P, Pavan F, Furlanut M (2003) Teicoplanin therapeutic
plasma levels with therapeutic outcome in gram-negative pneumonia. drug monitoring in critically ill patients: a retrospective study empha
Am J Med 77(4):657662 sizing the importance of a loading dose. J Antimicrob Chemother
130. Men P, Li HB, Zhai SD, Zhao RS (2016) Association between the AUC0- 51(4):971975
24/MIC ratio of vancomycin and its clinical effectiveness: a systematic 150. Pea F, Viale P (2009) Bench-to-bedside review: appropriate antibiotic
review and meta-analysis. PLoS One 11(1):e0146224 therapy in severe sepsis and septic shockdoes the dose matter? Crit
131. Moise-Broder PA, Forrest A, Birmingham MC, Schentag JJ (2004) Care 13(3):214
Pharmacodynamics of vancomycin and other antimicrobials in patients 151. Wang JT, Fang CT, Chen YC, Chang SC (2001) Necessity of a loading
with Staphylococcus aureus lower respiratory tract infections. Clin dose when using vancomycin in critically ill patients. J Antimicrob
Pharmacokinet 43(13):925942 Chemother 47(2):246
132. Zelenitsky S, Rubinstein E, Ariano R etal (2013) Vancomycin pharmaco 152. Kumar A (2014) An alternate pathophysiologic paradigm of sepsis
dynamics and survival in patients with methicillin-resistant Staphy and septic shock: implications for optimizing antimicrobial therapy.
lococcus aureus-associated septic shock. Int J Antimicrob Agents Virulence 5(1):8097
41(3):255260 153. Rhodes NJ, MacVane SH, Kuti JL, Scheetz MH (2014) Impact of loading
133. Forrest A, Nix DE, Ballow CH, Goss TF, Birmingham MC, Schentag JJ doses on the time to adequate predicted beta-lactam concentrations
(1993) Pharmacodynamics of intravenous ciprofloxacin in seriously ill in prolonged and continuous infusion dosing schemes. Clin Infect Dis
patients. Antimicrob Agents Chemother (Bethesda). 37(5):10731081 59(6):905907
134. Preston SL, Drusano GL, Berman AL etal (1998) Pharmacodynam 154. Lodise TP, Lomaestro BM, Drusano GL (2007) Piperacillin-tazobactam
ics of levofloxacin: a new paradigm for early clinical trials. JAMA for Pseudomonas aeruginosa infection: clinical implications of an
279(2):125129 extended-infusion dosing strategy. Clin Infect Dis 44:357363
135. Drusano GL, Preston SL, Fowler C, Corrado M, Weisinger B, Kahn J (2004) 155. Yost RJ, Cappelletty DM, Bennett JH etal (2011) The retrospective
Relationship between fluoroquinolone area under the curve: minimum cohort of extended-infusion piperacillin-tazobactam (RECEIPT) study: a
inhibitory concentration ratio and the probability of eradication of the multicenter study. Pharmacotherapy. 31(8):767775
infecting pathogen, in patients with nosocomial pneumonia. J Infect 156. Falagas ME, Tansarli GS, Ikawa K, Vardakas KZ (2013) Clinical outcomes
Dis 189(9):15901597 with extended or continuous versus short-term intravenous infusion
136. Kashuba AD, Nafziger AN, Drusano GL, Bertino JS Jr (1999) Optimiz of carbapenems and piperacillin/tazobactam: a systematic review and
ing aminoglycoside therapy for nosocomial pneumonia caused by meta-analysis. Clin Infect Dis 56(2):272282
gram-negative bacteria. Antimicrob Agents Chemother (Bethesda). 157. Yusuf E, Spapen H, Pirard D (2014) Prolonged vs intermittent infusion
43(3):623629 of piperacillin/tazobactam in critically ill patients: a narrative and
137. Schentag JJ, Smith IL, Swanson DJ etal (1984) Role for dual individuali systematic review. J Crit Care 29(6):10891095
zation with cefmenoxime. Am J Med 77(6A):4350 158. Mah GT, Mabasa VH, Chow I, Ensom MH (2012) Evaluating outcomes
138. Crandon JL, Bulik CC, Kuti JL, Nicolau DP (2010) Clinical pharmacody associated with alternative dosing strategies for piperacillin/tazobac
namics of cefepime in patients infected with pseudomonas aeruginosa. tam: a qualitative systematic review. Ann Pharmacother 46(2):265275
Antimicrob Agents Chemother (Bethesda). 54(3):11111116 159. Roberts JA, Abdul-Aziz MH, Lipman J etal (2014) Individualised antibi
139. McKinnon PS, Paladino JA, Schentag JJ (2008) Evaluation of area under otic dosing for patients who are critically ill: challenges and potential
the inhibitory curve (AUIC) and time above the minimum inhibitory solutions. Lancet Infect Dis. 14(6):498509
concentration (T>MIC) as predictors of outcome for cefepime and 160. Baptista JP, Sousa E, Martins PJ, Pimentel JM (2012) Augmented renal
ceftazidime in serious bacterial infections. Int J Antimicrob Agents clearance in septic patients and implications for vancomycin optimisa
31(4):345351 tion. Int J Antimicrob Agents 39(5):420423
140. Roberts JA, Abdul-Aziz MH, Davis JS etal (2016) Continuous versus 161. Hobbs AL, Shea KM, Roberts KM, Daley MJ (2015) Implications of aug
intermittent beta-lactam infusion in severe sepsis. a meta-analysis of mented renal clearance on drug dosing in critically ill patients: a focus
individual patient data from randomized trials. Am J Respir Crit Care on antibiotics. Pharmacotherapy. 35(11):10631075
Med 194(6):681691 162. Udy AA, Varghese JM, Altukroni M etal (2012) Subtherapeutic initial
141. Barza M, Ioannidis JP, Cappelleri JC, Lau J (1996) Single or multiple daily beta-lactam concentrations in select critically ill patients: association
doses of aminoglycosides: a meta-analysis. BMJ 312(7027):338344 between augmented renal clearance and low trough drug concentra
142. Hatala R, Dinh T, Cook DJ (1996) Once-daily aminoglycoside dos tions. Chest 142(1):3039
ing in immunocompetent adults: a meta-analysis. Ann Intern Med 163. Blot S, Koulenti D, Akova M etal (2014) Does contemporary vanco
124(8):717725 mycin dosing achieve therapeutic targets in a heterogeneous clinical
143. Zelenitsky SA, Ariano RE (2010) Support for higher ciprofloxacin AUC cohort of critically ill patients? Data from the multinational DALI study.
24/MIC targets in treating Enterobacteriaceae bloodstream infection. J Crit Care 18(3):R99
Antimicrob Chemother 65(8):17251732 164. Roberts JA, Paul SK, Akova M etal (2014) DALI: defining antibiotic levels
144. Dunbar LM, Wunderink RG, Habib MP etal (2003) High-dose, short- in intensive care unit patients: are current beta-lactam antibiotic doses
course levofloxacin for community-acquired pneumonia: a new treat sufficient for critically ill patients? Clin Infect Dis 58(8):10721083
ment paradigm. Clin Infect Dis 37(6):752760 165. Taccone FS, Laterre PF, Spapen H etal (2010) Revisiting the loading
145. Rybak M, Lomaestro B, Rotschafer J etal (2009) Vancomycin therapeu dose of amikacin for patients with severe sepsis and septic shock. Crit
tic guidelines: a summary of consensus recommendations from the Care 14(2):R53
infectious diseases Society of America, the American Society of Health- 166. Rea RS, Capitano B, Bies R, Bigos KL, Smith R, Lee H (2008) Subopti
System Pharmacists, and the Society of Infectious Diseases Pharmacists. mal aminoglycoside dosing in critically ill patients. Ther Drug Monit
Clin Infect Dis 49(3):325327 30(6):674681
146. Matsumoto K, Takesue Y, Ohmagari N etal (2013) Practice guidelines 167. Kumar A, Safdar N, Kethireddy S, Chateau D (2010) A survival benefit of
for therapeutic drug monitoring of vancomycin: a consensus review combination antibiotic therapy for serious infections associated with
of the Japanese Society of Chemotherapy and the Japanese Society of sepsis and septic shock is contingent only on the risk of death: a meta-
Therapeutic Drug Monitoring. J Infect Chemother. 19(3):365380 analytic/meta-regression study. Crit Care Med 38(8):16511665
147. Steinmetz T, Eliakim-Raz N, Goldberg E, Leibovici L, Yahav D (2015) 168. Kumar A, Zarychanski R, Light B etal (2010) Early combination
Association of vancomycin serum concentrations with efficacy in antibiotic therapy yields improved survival compared with mono
patients with MRSA infections: a systematic review and meta-analysis. therapy in septic shock: a propensity-matched analysis. Crit Care Med
Clin Microbiol Infect 21(7):665673 38(9):17731785
148. Mohamed AF, Karaiskos I, Plachouras D etal (2012) Application of 169. Al-Hasan MN, Wilson JW, Lahr BD etal (2009) Beta-lactam and fluo
a loading dose of colistin methanesulfonate in critically ill patients: roquinolone combination antibiotic therapy for bacteremia caused
 by gram-negative bacilli. Antimicrob Agents Chemother (Bethesda). 189. Gomes Silva BN, Andriolo RB, Atallah AN, Salomo R (2010) De-escala
53(4):13861394 tion of antimicrobial treatment for adults with sepsis, severe sepsis or
170. Delannoy PY, Boussekey N, Alfandari S etal (2012) Impact of combina septic shock. Cochrane Database Syst Rev 12:CD007934
tion therapy with aminoglycosides on the outcome of ICU-acquired 190. Morel J, Casoetto J, Josp R etal (2010) De-escalation as part of a global
bacteraemias. Eur J Clin Microbiol Infect Dis 31(9):22932299 strategy of empiric antibiotherapy management. A retrospective study
171. Diaz-Martin A, Martinez-Gonzalez ML, Ferrer R etal (2012) Antibiotic in a medico-surgical intensive care unit. Crit Care 14(6):R225
prescription patterns in the empiric therapy of severe sepsis: combina 191. Joung MK, Lee JA, Moon SY etal (2011) Impact of de-escalation therapy
tion of antimicrobials with different mechanisms of action reduces on clinical outcomes for intensive care unit-acquired pneumonia. Crit
mortality. Crit Care 16(6):R223 Care 15(2):R79
172. Martin-Loeches I, Lisboa T, Rodriguez A etal (2010) Combination 192. Leone M, Bechis C, Baumstarck K etal (2014) De-escalation versus
antibiotic therapy with macrolides improves survival in intubated continuation of empirical antimicrobial treatment in severe sepsis: a
patients with community-acquired pneumonia. Intensive Care Med multicenter non-blinded randomized noninferiority trial. Intensive Care
36(4):612620 Med 40(10):13991408
173. Brunkhorst FM, Oppert M, Marx G etal (2012) Effect of empirical treat 193. Riccio LM, Popovsky KA, Hranjec T etal (2014) Association of exces
ment with moxifloxacin and meropenem vs meropenem on sepsis- sive duration of antibiotic therapy for intra-abdominal infection with
related organ dysfunction in patients with severe sepsis: a randomized subsequent extra-abdominal infection and death: a study of 2,552
trial. JAMA 307(22):23902399 consecutive infections. Surg Infect. 15(4):417424
174. Safdar N, Handelsman J, Maki DG (2004) Does combination antimicro 194. Aarts MA, Brun-Buisson C, Cook DJ etal (2007) Antibiotic man
bial therapy reduce mortality in Gram-negative bacteraemia? A meta- agement of suspected nosocomial ICU-acquired infection: does
analysis. Lancet Infect Dis. 4(8):519527 prolonged empiric therapy improve outcome? Intensive Care Med
175. Paul M, Silbiger I, Grozinsky S, Soares-Weiser K, Leibovici L (2006) Beta 33(8):13691378
lactam antibiotic monotherapy versus beta lactam-aminoglycoside 195. Stevens V, Dumyati G, Fine LS, Fisher SG, Van Wijngaarden E (2011)
antibiotic combination therapy for sepsis. Cochrane Database Syst Rev Cumulative antibiotic exposures over time and the risk of Clostridium
1:CD003344 difficile infection. Clin Infect Dis 53(1):4248
176. Rodriguez A, Mendia A, Sirvent JM etal (2007) Combination antibiotic 196. Goossens H (2009) Antibiotic consumption and link to resistance. Clin
therapy improves survival in patients with community-acquired pneu Microbiol Infect 15(Suppl 3):1215
monia and shock. Crit Care Med 35(6):14931498 197. Chastre J, Wolff M, Fagon JY etal (2003) Comparison of 8 vs 15days
177. Baddour LM, Yu VL, Klugman KP etal (2004) Combination antibiotic of antibiotic therapy for ventilator-associated pneumonia in adults: a
therapy lowers mortality among severely ill patients with pneumococ randomized trial. JAMA 290(19):25882598
cal bacteremia. Am J Respir Crit Care Med 170:440444 198. Choudhury G, Mandal P, Singanayagam A, Akram AR, Chalmers JD,
178. Hilf M, Yu VL, Sharp J, Zuravleff JJ, Korvick JA, Muder RR (1989) Antibiotic Hill AT (2011) Seven-day antibiotic courses have similar efficacy to
therapy for Pseudomonas aeruginosa bacteremia: outcome correlations prolonged courses in severe community-acquired pneumonia-a
in a prospective study of 200 patients. Am J Med 87(5):540546 propensity-adjusted analysis. Clin Microbiol Infect 17(12):18521858
179. Tumbarello M, Viale P, Viscoli C etal (2012) Predictors of mortality in 199. Pugh R, Grant C, Cooke RP, Dempsey G (2015) Shortcourse versus
bloodstream infections caused by KPC-producing Klebsiella pneumo- prolongedcourse antibiotic therapy for hospitalacquired pneumonia
niae: importance of combination therapy. Clin Infect Dis 55(7):943 in critically ill adults. Cochrane Database Syst Rev 8:CD007577
180. Bass SN, Bauer SR, Neuner EA, Lam SW (2015) Impact of combination 200. Sawyer RG, Claridge JA, Nathens AB etal (2015) Trial of short-course
antimicrobial therapy on mortality risk for critically ill patients with antimicrobial therapy for intraabdominal infection. N Engl J Med
carbapenem-resistant bacteremia. Antimicrob Agents Chemother 372(21):19962005
(Bethesda). 59(7):37483753 201. Eliakim-Raz N, Yahav D, Paul M, Leibovici L (2013) Duration of antibiotic
181. Poulikakos P, Tansarli G, Falagas M (2014) Combination antibiotic treat treatment for acute pyelonephritis and septic urinary tract infec
ment versus monotherapy for multidrug-resistant, extensively drug- tion7days or less versus longer treatment: systematic review and
resistant, and pandrug-resistant Acinetobacter infections: a systematic meta-analysis of randomized controlled trials. J Antimicrob Chemother
review. Eur J Clin Microbiol Infect Dis 33(10):16751685 68(10):21832191
182. Falagas ME, Lourida P, Poulikakos P, Rafailidis PI, Tansarli GS (2014) Anti 202. Rattan R, Allen CJ, Sawyer RG etal (2016) Patients with complicated
biotic treatment of infections due to carbapenem-resistant Enterobac intra-abdominal infection presenting with sepsis do not require longer
teriaceae: systematic evaluation of the available evidence. Antimicrob duration of antimicrobial therapy. J Am Coll Surg 222(4):440446
Agents Chemother (Bethesda). 58(2):654663 203. Hepburn MJ, Dooley DP, Skidmore PJ, Ellis MW, Starnes WF, Hasewin
183. Hu Y, Li L, Li W etal (2013) Combination antibiotic therapy versus kle WC (2004) Comparison of short-course (5days) and standard
monotherapy for Pseudomonas aeruginosa bacteraemia: a meta-anal (10days) treatment for uncomplicated cellulitis. Arch Intern Med
ysis of retrospective and prospective studies. Int J Antimicrob Agents 164(15):16691674
42(6):492496 204. Chaudhry ZI, Nisar S, Ahmed U, Ali M (2000) Short course of antibiotic
184. Vardakas KZ, Tansarli GS, Bliziotis IA, Falagas ME (2013) -lactam plus treatment in spontaneous bacterial peritonitis: a randomized controlled
aminoglycoside or fluoroquinolone combination versus -lactam study. J Coll Physicians Surg Pak 10(8):284288
monotherapy for Pseudomonas aeruginosa infections: a meta-analysis. 205. Jack L, Bal AM, Harte S, Collier A (2016) International guidelines: the
Int J Antimicrob Agents 41(4):301310 need to standardize the management of candidaemia. Infect Dis
185. Stevens DL, Tanner MH, Winship J etal (1989) Severe group A strep (Lond). 48(1112):779781
tococcal infections associated with a toxic shock-like syndrome and 206. Baddour LM, Wilson WR, Bayer AS etal (2015) Infective endocarditis
scarlet fever toxin A. N Engl J Med 321(1):17 in adults: diagnosis, antimicrobial therapy, and management of com
186. Zimbelman J, Palmer A, Todd J (1999) Improved outcome of clindamy plications: a scientific statement for healthcare professionals from the
cin compared with beta-lactam antibiotic treatment for invasive Strep American Heart Association. Circulation 132(15):14351486
tococcus pyogenes infection. Pediatr Infect Dis J. 18(12):10961100 207. Habib G, Lancellotti P, Antunes MJ etal (2015) 2015 ESC Guide
187. Paul M, Soares-Weiser K, Leibovici L (2003) -lactam monotherapy lines for the management of infective endocarditis. Eur Heart J
versus -lactam-aminoglycoside combination therapy for fever 36(44):30753128
with neutropenia: systematic review and meta-analysis. Br Med J 208. Weiss CH, Moazed F, McEvoy CA etal (2011) Prompting physicians to
326(7399):11111118 address a daily checklist and process of care and clinical outcomes: a
188. Freifeld AG, Bow EJ, Sepkowitz KA etal (2011) Clinical practice guideline single-site study. Am J Respir Crit Care Med 184(6):680
for the use of antimicrobial agents in neutropenic patients with cancer: 209. Aguado JM, Vzquez L, Fernndez-Ruiz M etal (2015) Serum galac
2010 update by the Infectious Diseases Society of America. Clin Infect tomannan versus a combination of galactomannan and polymerase
Dis 52(4):e56e93 chain reaction-based Aspergillus DNA detection for early therapy of
 invasive aspergillosis in high-risk hematological patients: a randomized patients with cholangitis-associated septic shock. Aliment Pharmacol
controlled trial. Clin Infect Dis 60(3):405414 Ther 44(7):755766
210. Hou TY, Wang SH, Liang SX, Jiang WX, Luo DD, Huang DH (2015) The 230. Maitland K, Kiguli S, Opoka RO etal (2011) Mortality after fluid bolus in
screening performance of serum 1,3-beta-D-glucan in patients with African children with severe infection. N Engl J Med 364(26):24832495
invasive fungal diseases: a meta-analysis of prospective cohort studies. 231. Acheampong A, Vincent JL (2015) A positive fluid balance is an inde
PLoS One 10(7):e0131602 pendent prognostic factor in patients with sepsis. Crit Care 19:251
211. Schuetz P, Briel M, Christ-Crain M etal (2012) Procalcitonin to guide 232. Brotfain E, Koyfman L, Toledano R etal (2016) Positive fluid balance as a
initiation and duration of antibiotic treatment in acute respiratory major predictor of clinical outcome of patients with sepsis/septic shock
infections: an individual patient data meta-analysis. Clin Infect Dis after ICU discharge. Am J Emerg Med 34(11):21222126
55(5):651662 233. Mitchell KH, Carlbom D, Caldwell E, Leary PJ, Himmelfarb J, Hough
212. Matthaiou DK, Ntani G, Kontogiorgi M, Poulakou G, Armaganidis A, CL (2015) Volume overload: prevalence, risk factors, and func
Dimopoulos G (2012) An ESICM systematic review and meta-analysis of tional outcome in survivors of septic shock. Ann Am Thorac Soc.
procalcitonin-guided antibiotic therapy algorithms in adult critically ill 12(12):18371844
patients. Intensive Care Med 38(6):940949 234. De Oliveira FS, Freitas FG, Ferreira EM etal (2015) Positive fluid balance
213. Prkno A, Wacker C, Brunkhorst FM, Schlattmann P (2013) Procalcitonin- as a prognostic factor for mortality and acute kidney injury in severe
guided therapy in intensive care unit patients with severe sepsis sepsis and septic shock. J Crit Care 30(1):97101
and septic shocka systematic review and meta-analysis. Crit Care 235. Malbrain ML, Marik PE, Witters I etal (2014) Fluid overload, de-resus
17(6):R291 citation, and outcomes in critically ill or injured patients: a systematic
214. Westwood M, Ramaekers B, Whiting P etal (2015) Procalcitonin testing review with suggestions for clinical practice. Anaesthesiol Intensive
to guide antibiotic therapy for the treatment of sepsis in intensive care Ther. 46(5):361380
settings and for suspected bacterial infection in emergency depart 236. Yunos NM, Bellomo R, Hegarty C, Story D, Ho L, Bailey M (2012) Associa
ment settings: a systematic review and cost-effectiveness analysis. tion between a chloride-liberal vs chloride-restrictive intravenous fluid
Health Technol Assess 19(96):vxxv, 1236 administration strategy and kidney injury in critically ill adults. JAMA
215. Wacker C, Prkno A, Brunkhorst FM, Schlattmann P (2013) Procalcitonin 308(15):15661572
as a diagnostic marker for sepsis: a systematic review and meta-analysis. 237. Rochwerg B, Alhazzani W, Sindi A etal (2014) Fluid resuscitation in
Lancet Infect Dis. 13(5):426435 sepsis: a systematic review and network meta-analysis. Ann Intern Med
216. Soni NJ, Samson DJ, Galaydick JL etal (2013) Procalcitonin-guided 161(5):347355
antibiotic therapy: a systematic review and meta-analysis. J Hosp Med. 238. Young P, Bailey M, Beasley R etal (2015) Effect of a buffered crystal
8(9):530540 loid solution vs saline on acute kidney injury among patients in
217. De Jong E, van Oers JA, Beishuizen A etal (2016) Efficacy and safety of the intensive care unit: the SPLIT randomized clinical trial. JAMA
procalcitonin guidance in reducing the duration of antibiotic treatment 314(16):17011710
in critically ill patients: a randomised, controlled, open-label trial. Lancet 239. Finfer S, Norton R, Bellomo R, Boyce N, French J, Myburgh J (2004) The
Infect Dis. 16(7):819827 SAFE study: saline vs. albumin for fluid resuscitation in the critically ill.
218. Lesprit P, Landelle C, Girou E, Brun-Buisson C (2010) Reassessment of Vox Sang 87(Suppl 2):123131
intravenous antibiotic therapy using a reminder or direct counselling. J 240. Delaney AP, Dan A, McCaffrey J, Finfer S (2011) The role of albumin as
Antimicrob Chemother 65(4):789795 a resuscitation fluid for patients with sepsis: a systematic review and
219. Paul M, Dickstein Y, Raz-Pasteur A (2016) Antibiotic de-escalation for meta-analysis. Crit Care Med 39(2):386391
bloodstream infections and pneumonia: systematic review and meta- 241. Rochwerg B, Alhazzani W, Gibson A etal (2015) Fluid type and the
analysis. Clin Microbiol Infect 22(12):960967 use of renal replacement therapy in sepsis: a systematic review and
220. Schuetz P, Kutz A, Grolimund E etal (2014) Excluding infection through network meta-analysis. Intensive Care Med 41(9):15611571
procalcitonin testing improves outcomes of congestive heart failure 242. Xu JY, Chen QH, Xie JF etal (2014) Comparison of the effects of albumin
patients presenting with acute respiratory symptoms: results from the and crystalloid on mortality in adult patients with severe sepsis and sep
randomized ProHOSP trial. Int J Cardiol 175(3):464472 tic shock: a meta-analysis of randomized clinical trials. Crit Care 18(6):702
221. Hoeboer SH, van der Geest PJ, Nieboer D, Groeneveld AB (2015) The 243. Uhlig C, Silva PL, Deckert S, Schmitt J, de Abreu MG (2014) Albumin ver
diagnostic accuracy of procalcitonin for bacteraemia: a systematic sus crystalloid solutions in patients with the acute respiratory distress
review and meta-analysis. Clin Microbiol Infect 21(5):474481 syndrome: a systematic review and meta-analysis. Crit Care 18(1):R10
222. Jimenez MF, Marshall JC (2001) Source control in the management of 244. Patel A, Laffan MA, Waheed U, Brett SJ (2014) Randomised trials
sepsis. Intensive Care Med 27:S49S62 of human albumin for adults with sepsis: systematic review and
223. Azuhata T, Kinoshita K, Kawano D etal (2014) Time from admission to meta-analysis with trial sequential analysis of all-cause mortality. BMJ
initiation of surgery for source control is a critical determinant of sur 349:g4561
vival in patients with gastrointestinal perforation with associated septic 245. Jiang L, Jiang S, Zhang M, Zheng Z, Ma Y (2014) Albumin versus other
shock. Crit Care 18(3):R87 fluids for fluid resuscitation in patients with sepsis: a meta-analysis.
224. Bloos F, Thomas-Rddel D, Rddel H etal (2014) Impact of compliance PLoS One 9(12):e114666
with infection management guidelines on outcome in patients with 246. Boldt J, Heesen M, Muller M, Pabsdorf M, Hempelmann G (1996) The
severe sepsis: a prospective observational multi-center study. Crit Care effects of albumin versus hydroxyethyl starch solution on cardiores
18(2):1 piratory and circulatory variables in critically ill patients. Anesth Analg
225. Moss RL, Musemeche CA, Kosloske AM (1996) Necrotizing fasciitis in 83(2):254261
children: prompt recognition and aggressive therapy improve survival. 247. Boldt J, Heesen M, Welters I, Padberg W, Martin K, Hempelmann G
J Pediatr Surg 31(8):11421146 (1995) Does the type of volume therapy influence endothelial-related
226. Wong CH, Chang HC, Pasupathy S, Khin LW, Tan JL, Low CO (2003) coagulation in the critically ill? Br J Anaesth 75(6):740746
Necrotizing fasciitis: clinical presentation, microbiology, and determi 248. Boldt J, Muller M, Heesen M, Heyn O, Hempelmann G (1996) Influence
nants of mortality. J Bone Joint Surg Am 85(8):14541460 of different volume therapies on platelet function in the critically ill.
227. Chao WN, Tsai CF, Chang HR etal (2013) Impact of timing of surgery Intensive Care Med 22(10):10751081
on outcome of Vibrio vulnificus-related necrotizing fasciitis. Am J Surg 249. Caironi P, Tognoni G, Masson S etal (2014) Albumin replace
206(1):3239 ment in patients with severe sepsis or septic shock. N Engl J Med
228. Buck DL, Vester-Andersen M, Mller MH, Danish Clinical Register of 370(15):14121421
Emergency Surgery (2013) Surgical delay is a critical determinant of 250. Haase N, Perner A, Hennings LI etal (2013) Hydroxyethyl starch
survival in perforated peptic ulcer. Br J Surg 100(8):10451049 130/0.38-0.45 versus crystalloid or albumin in patients with sepsis:
229. Karvellas CJ, Abraldes JG, Zepeda-Gomez S etal (2016) The impact systematic review with meta-analysis and trial sequential analysis. BMJ
of delayed biliary decompression and antimicrobial therapy in 260 346:f839
 251. Moeller C, Fleischmann C, Thomas-Rueddel D etal (2016) How 274. Russell JA, Walley KR, Singer J etal (2008) Vasopressin versus nor
safe is gelatin? A systematic review and meta-analysis of gelatin- epinephrine infusion in patients with septic shock. N Engl J Med
containing plasma expanders vs crystalloids and albumin. J Crit Care 358(9):877887
35:7583 275. Dunser MW, Mayr AJ, Tur A etal (2003) Ischemic skin lesions as a
252. Day NP, Phu NH, Bethell DP etal (1996) The effects of dopamine and complication of continuous vasopressin infusion in catecholamine-
adrenaline infusions on acid-base balance and systemic haemodynam resistant vasodilatory shock: incidence and risk factors. Crit Care Med
ics in severe infection. Lancet 348(9022):219223 31(5):13941398
253. De Backer D, Creteur J, Silva E, Vincent JL (2003) Effects of dopamine, 276. Gordon AC, Mason AJ, Thirunavukkarasu N etal (2016) Effect of early
norepinephrine, and epinephrine on the splanchnic circulation in vasopressin vs norepinephrine on kidney failure in patients with septic
septic shock: which is best? Crit Care Med 31(6):16591667 shock: the VANISH randomized clinical trial. JAMA 316(5):509518
254. Martin C, Papazian L, Perrin G, Saux P, Gouin F (1993) Norepinephrine 277. Albanese J, Leone M, Delmas A, Martin C (2005) Terlipressin or norepi
or dopamine for the treatment of hyperdynamic septic shock. Chest nephrine in hyperdynamic septic shock: a prospective, randomized
103(6):18261831 study. Crit Care Med 33(9):18971902
255. Martin C, Viviand X, Leone M, Thirion X (2000) Effect of norepinephrine 278. Morelli A, Ertmer C, Lange M etal (2008) Effects of short-term simulta
on the outcome of septic shock. Crit Care Med 28(8):27582765 neous infusion of dobutamine and terlipressin in patients with septic
256. Bollaert PE, Bauer P, Audibert G, Lambert H, Larcan A (1990) Effects of shock: the DOBUPRESS study. Br J Anaesth 100(4):494503
epinephrine on hemodynamics and oxygen-metabolism in dopamine- 279. Morelli A, Ertmer C, Rehberg S etal (2009) Continuous terlipressin
resistant septic shock. Chest 98(4):949953 versus vasopressin infusion in septic shock (TERLIVAP): a randomized,
257. Levy B, Bollaert PE, Charpentier C etal (1997) Comparison of nor controlled pilot study. Crit Care 13(4):R130
epinephrine and dobutamine to epinephrine for hemodynamics, 280. Zhou F, Mao Z, Zeng X etal (2015) Vasopressors in septic shock: a
lactate metabolism, and gastric tonometric variables in septic shock: a systematic review and network meta-analysis. Ther Clin Risk Manag
prospective, randomized study. Intensive Care Med 23(3):282287 11:10471059
258. Zhou SX, Qiu HB, Huang YZ, Yang Y, Zheng RQ (2002) Effects of norepi 281. De Backer D, Aldecoa C, Njimi H, Vincent JL (2012) Dopamine versus
nephrine, epinephrine, and norepinephrine-dobutamine on systemic norepinephrine in the treatment of septic shock: a meta-analysis. Crit
and gastric mucosal oxygenation in septic shock. Acta Pharmacol Sin Care Med 40(3):725730
23(7):654658 282. De Backer D, Biston P, Devriendt J etal (2010) Comparison of dopa
259. Mackenzie SJ, Kapadia F, Nimmo GR, Armstrong IR, Grant IS (1991) mine and norepinephrine in the treatment of shock. N Engl J Med
Adrenaline in treatment of septic shock: effects on hemodynamics and 362(9):779789
oxygen-transport. Intensive Care Med 17(1):3639 283. Parker MM, Shelhamer JH, Bacharach SL etal (1984) Profound but
260. Moran JL, OFathartaigh MS, Peisach AR, Chapman MJ, Leppard P (1993) reversible myocardial depression in patients with septic shock. Ann
Epinephrine as an inotropic agent in septic shock: a dose-profile analy Intern Med 100:483490
sis. Crit Care Med 21(1):7077 284. Gattinoni L, Brazzi L, Pelosi P etal (1995) A trial of goal-oriented
261. Yamazaki T, Shimada Y, Taenaka N, Oshumi H, Takezawa J, Yoshiya I hemodynamic therapy in critically ill patients. N Engl J Med
(1982) Circulatory responses to afterloading with phenylephrine in 333(16):10251032
hyperdynamic sepsis. Crit Care Med 10(7):432435 285. Hayes MA, Timmins AC, Yau EH, Palazzo M, Hinds CJ, Watson D (1994)
262. Regnier B, Rapin M, Gory G, Lemaire F, Teisseire B, Harari A (1977) Elevation of systemic oxygen delivery in the treatment of critically ill
Haemodynamic effects of dopamine in septic shock. Intensive Care patients. N Engl J Med 330(24):17171722
Med 3(2):4753 286. Hollenberg SM, Ahrens TS, Annane D etal (2004) Practice parameters
263. Beck GCh, Brinkkoetter P, Hanusch C etal (2004) Clinical review: immu for hemodynamic support of sepsis in adult patients: 2004 update. Crit
nomodulatory effects of dopamine in general inflammation. Crit Care Care Med 32(9):19281948
8(6):485491 287. Annane D, Vignon P, Renault A etal (2007) Norepinephrine plus dobu
264. Avni T, Lador A, Lev S, Leibovici L, Paul M, Grossman A (2015) Vasopres tamine versus epinephrine alone for management of septic shock: a
sors for the treatment of septic shock: systematic review and meta- randomised trial. Lancet 370(9588):676684
analysis. PLoS One 10(8):e0129305 288. ProCess Investigators, Yealy DM, Kellum JA etal (2014) A randomized
265. Myburgh JA, Higgins A, Jovanovska A etal (2008) A comparison of trial of protocol-based care for early septic shock. N Engl J Med
epinephrine and norepinephrine in critically ill patients. Intensive Care 370(18):16831693
Med 34(12):22262234 289. ARISE Investigators, Anzics Clinical Trials Group, Peake SL etal (2014)
266. Landry DW, Levin HR, Gallant EM etal (1997) Vasopressin defi Goal-directed resuscitation for patients with early septic shock. N Engl J
ciency contributes to the vasodilation of septic shock. Circulation Med 371(16):14961506
95(5):11221125 290. Barton P, Garcia J, Kouatli A etal (1996) Hemodynamic effects of IV
267. Patel BM, Chittock DR, Russell JA, Walley KR (2002) Beneficial effects of milrinone lactate in pediatric patients with septic shock. A prospective,
short-term vasopressin infusion during severe septic shock. Anesthesi double-blinded, randomized, placebo-controlled, interventional study.
ology 96(3):576582 Chest 109(5):13021312
268. Dunser MW, Mayr AJ, Ulmer H etal (2003) Arginine vasopressin in 291. Morelli A, Teboul JL, Maggiore SM etal (2006) Effects of levosimendan
advanced vasodilatory shock: a prospective, randomized, controlled on right ventricular afterload in patients with acute respiratory distress
study. Circulation 107(18):23132319 syndrome: a pilot study. Crit Care Med 34(9):22872293
269. Lauzier F, Levy B, Lamarre P, Lesur O (2006) Vasopressin or norepineph 292. Morelli A, De Castro S, Teboul JL etal (2005) Effects of levosimendan on
rine in early hyperdynamic septic shock: a randomized clinical trial. systemic and regional hemodynamics in septic myocardial depression.
Intensive Care Med 32(11):17821789 Intensive Care Med 31(5):638644
270. Holmes CL, Walley KR, Chittock DR, Lehman T, Russell JA (2001) The 293. Gordon AC, Perkins GD, Singer M etal Levosimendan for the preven
effects of vasopressin on hemodynamics and renal function in severe tion of acute organ dysfunction in sepsis. N Engl J Med. doi:10.1056/
septic shock: a case series. Intensive Care Med 27(8):14161421 NEJMoa1609409
271. Malay MB, Ashton RC, Landry DW, Townsend RN (1999) Low-dose 294. Cohn JN (1967) Blood pressure measurement in shock. Mecha
vasopressin in the treatment of vasodilatory septic shock. J Trauma nism of inaccuracy in ausculatory and palpatory methods. JAMA
47(4):699703 199(13):118122
272. OBrien A, Clapp L, Singer M (2002) Terlipressin for norepinephrine- 295. Hollenberg SM, Parrillo JE (1997) Shock. In: Braunwald E, Isselbacher KJ,
resistant septic shock. Lancet 359(9313):12091210 Wilson JD etal (eds) Harrisons principles of internal medicine, 14th edn.
273. Sharshar T, Blanchard A, Paillard M, Raphael JC, Gajdos P, Annane D McGraw-Hill, New York, pp 214222
(2003) Circulating vasopressin levels in septic shock. Crit Care Med 296. Scheer B, Perel A, Pfeiffer UJ (2002) Clinical review: complications and
31(6):17521758 risk factors of peripheral arterial catheters used for haemodynamic
 monitoring in anaesthesia and intensive care medicine. Crit Care 320. British Committee for Standards in Haematology (2003) Blood Transfu
6(3):199204 sion Task Force. Guidelines for the use of platelet transfusions. Br J
297. Gu WJ, Wu XD, Wang F, Ma ZL, Gu XP (2016) Ultrasound guidance facili Haematol 122(1):1023
tates radial artery catheterization: a meta-analysis with trial sequential 321. Diedrich B, Remberger M, Shanwell A, Svahn BM, Ringdn O (2005)
analysis of randomized controlled trials. Chest 149(1):166179 A prospective randomized trial of a prophylactic platelet transfu
298. OHoro JC, Maki DG, Krupp AE, Safdar N (2014) Arterial catheters as a sion trigger of 1010(9) per L versus 3010(9) per L in allogeneic
source of bloodstream infection: a systematic review and meta-analysis. hematopoietic progenitor cell transplant recipients. Transfusion.
Crit Care Med 42(6):13341339 45(7):10641072
299. Annane D, Bellissant E, Bollaert PE etal (2009) Corticosteroids in the 322. Kaufman RM, Djulbegovic B, Gernsheimer T etal (2015) Platelet
treatment of severe sepsis and septic shock in adults: a systematic transfusion: a clinical practice guideline from the AABB. Ann Intern Med
review. JAMA 301(22):23622375 162(3):205213
300. Bollaert PE, Charpentier C, Levy B, Debouverie M, Audibert G, Larcan 323. Schiffer CA, Anderson KC, Bennett CL etal (2001) Platelet transfusion
A (1998) Reversal of late septic shock with supraphysiologic doses of for patients with cancer: clinical practice guidelines of the American
hydrocortisone. Crit Care Med 26(4):645650 Society of Clinical Oncology. J Clin Oncol 19(5):15191538
301. Briegel J, Forst H, Haller M etal (1999) Stress doses of hydrocortisone 324. Stanworth SJ, Estcourt LJ, Llewelyn CA, Murphy MF, Wood EM,
reverse hyperdynamic septic shock: a prospective, randomized, double- TOPPS Study Investigators (2014) Impact of prophylactic platelet
blind, single-center study. Crit Care Med 27(4):723732 transfusions on bleeding events in patients with hematologic
302. Sprung CL, Annane D, Keh D etal (2008) Hydrocortisone therapy for malignancies: a subgroup analysis of a randomized trial. Transfusion.
patients with septic shock. N Engl J Med 358(2):111124 54(10):23852393
303. Sligl WI, Milner DA Jr, Sundar S, Mphatswe W, Majumdar SR (2009) 325. Stanworth SJ, Estcourt LJ, Powter G etal (2013) A no-prophylaxis
Safety and efficacy of corticosteroids for the treatment of septic shock: platelet-transfusion strategy for hematologic cancers. N Engl J Med
a systematic review and meta-analysis. Clin Infect Dis 49(1):93101 368(19):17711780
304. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y (2015) 326. Wandt H, Schaefer-Eckart K, Wendelin K etal (2012) Therapeutic platelet
Corticosteroids for treating sepsis. Cochrane Database Syst Rev transfusion versus routine prophylactic transfusion in patients with
12:CD002243 haematological malignancies: an open-label, multicentre, randomised
305. Volbeda M, Wetterslev J, Gluud C, Zijlstra JG, van der Horst IC, Keus study. Lancet 380(9850):13091316
F (2015) Glucocorticosteroids for sepsis: systematic review with 327. Zumberg MS, del Rosario ML, Nejame CF etal (2002) A prospective
meta-analysis and trial sequential analysis. Intensive Care Med randomized trial of prophylactic platelet transfusion and bleed
41(7):12201234 ing incidence in hematopoietic stem cell transplant recipients:
306. Annane D, Sebille V, Charpentier C etal (2002) Effect of treatment 10,000/L versus 20,000/microL trigger. Biol Blood Marrow Transplant
with low doses of hydrocortisone and fludrocortisone on mortality in 8(10):569576
patients with septic shock. JAMA 288(7):862871 328. Werdan K, Pilz G, Bujdoso O, Fraunberger P etal (2007) Score-based
307. Briegel J, Sprung CL, Annane D etal (2009) Multicenter comparison of immunoglobulin G therapy of patients with sepsis: the SBITS study. Crit
cortisol as measured by different methods in samples of patients with Care Med 35(12):26932701
septic shock. Intensive Care Med 35(12):21512156 329. Alejandria MM, Lansang MA, Dans LF, Mantaring JB 3rd (2013) Intra
308. Allolio B, Dorr H, Stuttmann R, Knorr D, Engelhardt D, Winkelmann W venous immunoglobulin for treating sepsis, severe sepsis and septic
(1985) Effect of a single bolus of etomidate upon 8 major corticosteroid shock. Cochrane Database Syst Rev 9:CD001090
hormones and plasma ACTH. Clin Endocrinol 22(3):281286 330. Soares MO, Welton NJ, Harrison DA etal (2012) An evaluation of the
309. Jabre P, Combes X, Lapostolle F etal (2009) Etomidate versus ketamine feasibility, cost and value of information of a multicentre randomised
for rapid sequence intubation in acutely ill patients: a multicentre controlled trial of intravenous immunoglobulin for sepsis (severe sepsis
randomised controlled trial. Lancet 374(9686):293300 and septic shock): incorporating a systematic review, meta-analysis and
310. Oppert M, Schindler R, Husung C etal (2005) Low-dose hydrocortisone value of information analysis. Health Technol Assess 16(7):1186
improves shock reversal and reduces cytokine levels in early hyperdy 331. Laupland KB, Kirkpatrick AW, Delaney A (2007) Polyclonal intravenous
namic septic shock. Crit Care Med 33(11):24572464 immunoglobulin for the treatment of severe sepsis and septic shock in
311. Yildiz O, Doganay M, Aygen B, Gven M, Kelestimur F, Tutu A (2002) critically ill adults: a systematic review and meta-analysis. Crit Care Med
Physiological-dose steroid therapy in sepsis [ISRCTN36253388]. Crit 35(12):26862692
Care 6(3):251259 332. Pildal J, Gotzsche PC (2004) Polyclonal immunoglobulin for treatment
312. Keh D, Boehnke T, Weber-Cartens S etal (2003) Immunologic and of bacterial sepsis: a systematic review. Clin Infect Dis 39(1):3846
hemodynamic effects of low-dose hydrocortisone in septic shock: a 333. Kreyrnann KG, de Heer G, Nierhaus A, Kluge S (2007) Use of polyclonal
double-blind, randomized, placebo-controlled, crossover study. Am J immunoglobulins as adjunctive therapy for sepsis or septic shock. Crit
Respir Crit Care Med 167(4):512520 Care Med 35(12):26772685
313. Huh JW, Choi H-S, Lim C-M etal (2011) Low-dose hydrocortisone treat 334. Shankar-Hari M, Culshaw N, Post B etal (2015) Endogenous IgG
ment for patients with septic shock: a pilot study comparing 3days hypogammaglobulinaemia in critically ill adults with sepsis: systematic
with 7 days. Respirology 16(7):10881095 review and meta-analysis. Intensive Care Med 41(8):13931401
314. Keh D, Trips E, Marx G etal (2016) Effect of hydrocortisone on devel 335. Turgeon AF, Hutton B, Fergusson DA etal (2007) Meta-analysis: intra
opment of shock among patients with severe sepsis: the HYPRESS venous immunoglobulin in critically ill adult patients with sepsis. Ann
randomized clinical trial. JAMA 316(17):17751785 Intern Med 146(3):193203
315. Weber-Carstens S, Deja M, Bercker S etal (2007) Impact of bolus appli 336. Zhou F, Peng Z, Murugan R, Kellum JA (2013) Blood purification and
cation of low-dose hydrocortisone on glycemic control in septic shock mortality in sepsis: a meta-analysis of randomized trials. Crit Care Med
patients. Intensive Care Med 33(4):730733 41(9):22092220
316. Holst LB, Haase N, Wetterslev J etal (2014) Lower versus higher 337. Payen DM, Guilhot J, Launey Y etal (2015) Early use of polymyxin B
hemoglobin threshold for transfusion in septic shock. N Engl J Med hemoperfusion in patients with septic shock due to peritonitis: a multi
371(15):13811391 center randomized control trial. Intensive Care Med 41(6):975984
317. Corwin HL, Gettinger A, Pearl RG etal (2002) Efficacy of recombinant 338. Klein DJ, Foster D, Schorr CA, Kazempour K, Walker PM, Dellinger RP
human erythropoietin in critically ill patients: a randomized controlled (2014) The EUPHRATES trial (evaluating the use of polymyxin B hemop
trial. JAMA 288(22):28272835 erfusion in a randomized controlled trial of adults treated for endotox
318. Corwin HL, Gettinger A, Rodriguez RM etal (1999) Efficacy of recom emia and septic shock): study protocol for a randomized controlled
binant human erythropoietin in the critically ill patient: a randomized, trial. Trials. 15:218
double-blind, placebo-controlled trial. Crit Care Med 27(11):23462350 339. Livigni S, Bertolini G, Rossi C etal (2014) Efficacy of coupled plasma
319. Liumbruno G, Bennardello F, Lattanzio A etal (2009) Recommendations filtration adsorption (CPFA) in patients with septic shock: a multicenter
for the transfusion of plasma and platelets. Blood Transfus. 7(2):132137 randomised controlled clinical trial. BMJ Open. 4(1):e003536
 340. Allingstrup M, Wetterslev J, Ravn FB, Moller AM, Afshari A (2016) positive pressure breathing during the adult respiratory distress syn
Antithrombin III for critically ill patients. Cochrane Database Syst Rev drome. Bull Eur Physiopathol Respir. 21(3):275279
2:CD005370 361. Marini JJ, Ravenscraft SA (1992) Mean airway pressure: physiological
341. Warren BL, Eid A, Singer P etal (2001) High-dose antithrombin III in determinants and clinical importancePart 1: physiological determi
severe sepsis: a randomized controlled trial. JAMA 286(15):18691878 nants and measurements. Crit Care Med 20(10):14611472
342. Vincent JL, Ramesh MK, Ernest D etal (2013) A randomized, double- 362. Pesenti A, Marcolin R, Prato P, Borelli M, Riboni A, Gattinoni L (1985)
blind, placebo-controlled, phase 2b study to evaluate the safety and Mean airway pressure vs positive end-expiratory pressure during
efficacy of recombinant human soluble thrombomodulin, ART-123, in mechanical ventilation. Crit Care Med 13(1):3437
patients with sepsis and suspected disseminated intravascular coagula 363. Brower RG, Lanken PN, MacIntyre N etal (2004) Higher versus lower
tion. Crit Care Med 41(9):20692079 positive end-expiratory pressures in patients with the acute respiratory
343. Yamakawa K, Ogura H, Fujimi S etal (2013) Recombinant human distress syndrome. N Engl J Med 351(4):327336
soluble thrombomodulin in sepsis-induced disseminated intravascular 364. Meade MO, Cook DJ, Guyatt GH etal (2008) Ventilation strategy using
coagulation: a multicenter propensity score analysis. Intensive Care low tidal volumes, recruitment maneuvers, and high positive end-
Med 39(4):644652 expiratory pressure for acute lung injury and acute respiratory distress
344. Zarychanski R, Abou-Setta AM, Kanji S etal (2015) The efficacy and syndrome: a randomized controlled trial. JAMA 299(6):637645
safety of heparin in patients with sepsis: a systematic review and 365. Mercat A, Richard JC, Vielle B etal (2008) Positive end-expiratory
metaanalysis. Crit Care Med 43(3):511518 pressure setting in adults with acute lung injury and acute res
345. Ranieri VM, Thompson BT, Barie PS etal (2012) Drotrecogin alfa (acti piratory distress syndrome: a randomized controlled trial. JAMA
vated) in adults with septic shock. N Engl J Med 366(22):20552064 299(6):646655
346. Bernard GR, Artigas A, Brigham KL etal (1994) The American-European 366. Villar J, Kacmarek RM, Perez-Mendez L, Aguirre-Jaime A (2006) A high
consensus conference on ARDS. Definitions, mechanisms, relevant positive end-expiratory pressure, low tidal volume ventilatory strategy
outcomes, and clinical-trial coordination. Am J Respir Crit Care Med improves outcome in persistent acute respiratory distress syndrome: a
149(3):818824 randomized, controlled trial. Crit Care Med 34(5):13111318
347. ARDS Definition Task Force, Ranieri VM, Rubenfeld GD etal (2012) 367. Briel M, Meade M, Mercat A etal (2010) Higher vs lower positive
Acute respiratory distress syndrome: the Berlin Definition. JAMA end-expiratory pressure in patients with acute lung injury and acute
307(23):25262533 respiratory distress syndrome: systematic review and meta-analysis.
348. Amato MB, Barbas CS, Medeiros DM etal (1998) Effect of a protective- JAMA 303(9):865873
ventilation strategy on mortality in the acute respiratory distress 368. Goligher EC, Kavanagh BP, Rubenfeld GD etal (2014) Oxygenation
syndrome. N Engl J Med 338(6):347354 response to positive end-expiratory pressure predicts mortality in acute
349. Brochard L, Roudot-Thoraval F, Roupie E etal (1998) Tidal volume reduc respiratory distress syndrome. A secondary analysis of the LOVS and
tion for prevention of ventilator-induced lung injury in acute respiratory ExPress trials. Am J Respir Crit Care Med 190(1):7076
distress syndrome. Am J Respir Crit Care Med 158(6):18311838 369. Talmor D, Sarge T, Malhotra A etal (2008) Mechanical ventilation
350. Brower RG, Matthay MA, Morris A etal (2000) Ventilation with lower guided by esophageal pressure in acute lung injury. N Engl J Med
tidal volumes as compared with traditional tidal volumes for acute 359(20):20952104
lung injury and the acute respiratory distress syndrome. N Engl J Med 370. Amato MB, Barbas CS, Medeiros DM etal (1995) Beneficial effects of the
342(18):13011308 open lung approach with low distending pressures in acute respira
351. Brower RG, Shanholtz CB, Fessler HE etal (1999) Prospective, rand tory distress syndrome. A prospective randomized study on mechanical
omized, controlled clinical trial comparing traditional versus reduced ventilation. Am J Respir Crit Care Med 152(6):18351846
tidal volume ventilation in acute respiratory distress syndrome patients. 371. Gattinoni L, Caironi P, Cressoni M etal (2006) Lung recruitment in
Crit Care Med 27(8):14921498 patients with the acute respiratory distress syndrome. N Engl J Med
352. Eichacker PQ, Gerstenberger EP, Banks SM, Cui XZ, Natanson C (2002) 354(17):17751786
Meta-analysis of acute lung injury and acute respiratory distress 372. Pipeling MR, Fan E (2010) Therapies for refractory hypoxemia in acute
syndrome trials testing low tidal volumes. Am J Respir Crit Care Med respiratory distress syndrome. JAMA 304(22):25212527
166(11):15101514 373. Fan E, Wilcox ME, Brower RG etal (2008) Recruitment maneuvers for
353. Burns KEA, Adhikari NK, Slutsky AS etal (2011) Pressure and volume acute lung injury: a systematic review. Am J Respir Crit Care Med
limited ventilation for the ventilatory management of patients with 178(11):11561163
acute lung injury: a systematic review and meta-analysis. PLoS One 374. Guerin C, Reignier J, Richard JC etal (2013) Prone positioning in severe
6(1):e14623 acute respiratory distress syndrome. N Engl J Med 368(23):21592168
354. Putensen C, Theuerkauf N, Zinserling J, Wrigge H, Pelosi P (2009) 375. Beitler JR, Shaefi S, Montesi SB etal (2014) Prone positioning reduces
Meta-analysis: ventilation strategies and outcomes of the acute mortality from acute respiratory distress syndrome in the low tidal
respiratory distress syndrome and acute lung injury. Ann Intern Med volume era: a meta-analysis. Intensive Care Med 40(3):332341
151(8):566576 376. Jolliet P, Bulpa P, Chevrolet JC (1998) Effects of the prone position on
355. Marini JJ, Gattinoni L (2004) Ventilatory management of acute gas exchange and hemodynamics in severe acute respiratory distress
respiratory distress syndrome: a consensus of two. Crit Care Med syndrome. Crit Care Med 26(12):19771985
32(1):250255 377. Lamm WJE, Graham MM, Albert RK (1994) Mechanism by which the
356. Tobin MJ (2000) Culmination of an era in research on the acute respira prone position improves oxygenation in acute lung injury. Am J Respir
tory distress syndrome. N Engl J Med 342(18):13601361 Crit Care Med 150(1):184193
357. Hager DN, Krishnan JA, Hayden DL, Brower RG, ARDS Clinical Trials 378. Stocker R (1997) Prone positioning and low-volume pressure-limited
Network (2005) Tidal volume reduction in patients with acute lung ventilation improve survival in patients with severe ARDS. Chest
injury when plateau pressures are not high. Am J Respir Crit Care Med 111(4):1008
172(10):12411245 379. Guerin C, Gaillard S, Lemasson S etal (2004) Effect of systematic prone
358. Checkley W, Brower R, Korpak A, Thompson BT, Acute Respiratory positioning in hypoxemic acute respiratory failure: a randomized con
Distress Syndrome Network Investigators (2008) Effects of a clinical trial trolled trial. JAMA 292(19):23792387
on mechanical ventilation practices in patients with acute lung injury. 380. Peek GJ, Mugford M, Tiruvoipati R (2009) Efficacy and economic
Am J Respir Crit Care Med 177(11):12151222 assessment of conventional ventilatory support versus extracorporeal
359. Amato MB, Meade MO, Slutsky AS etal (2015) Driving pressure and membrane oxygenation for severe adult respiratory failure (CESAR): a
survival in the acute respiratory distress syndrome. N Engl J Med multicentre randomised controlled trial. Lancet 374(9698):13511363
372(8):747755 381. Sud S, Sud M, Friedrich JO etal (2010) High frequency oscillation in
360. Gattinoni L, Marcolin R, Caspani ML, Fumagalli R, Mascheroni D, Pesenti patients with acute lung injury and acute respiratory distress syndrome
A (1985) Constant mean airway pressure with different patterns of (ARDS): systematic review and meta-analysis. BMJ 340:c2327
 382. Noah MA, Peek GJ, Finney SJ etal (2011) Referral to an extracorporeal 404. Martin GS, Mangialardi RJ, Wheeler AP, Dupont WD, Morris JA, Bernard
membrane oxygenation center and mortality among patients with GR (2002) Albumin and furosemide therapy in hypoproteinemic
severe 2009 influenza A (H1N1). JAMA 306(15):16591668 patients with acute lung injury. Crit Care Med 30(10):21752182
383. Checkley W (2011) Extracorporeal membrane oxygenation as a first-line 405. Mitchell JP, Schuller D, Calandrino FS, Schuster DP (1992) Improved
treatment strategy for ARDS: is the evidence sufficiently strong? JAMA outcome based on fluid management in critically ill patients requiring
306(15):17031704 pulmonary-artery catheterization. Am Rev Respir Dis. 145(5):990998
384. Ferguson ND, Cook DJ, Guyatt GH etal (2013) High-frequency oscil 406. Schuller D (1991) Fluid balance during pulmonary edema. Is fluid gain a
lation in early acute respiratory distress syndrome. N Engl J Med marker or a cause of poor outcome? Chest 100(4):1068
368(9):795805 407. Wiedemann HP, Wheeler AP, Bernard GR etal (2006) Comparison of
385. Young D, Lamb SE, Shah S etal (2013) High-frequency oscillation for two fluid-management strategies in acute lung injury. N Engl J Med
acute respiratory distress syndrome. N Engl J Med 368(9):806813 354(24):25642575
386. Meade MO, Cook DJ, Guyatt GH etal (2008) Ventilation strategy using 408. Perkins GD, McAuley DF, Thickett DR, Gao F (2006) The beta-agonist
low tidal volumes, recruitment maneuvers, and high positive end- lung injury trial (BALTI): a randomized placebo-controlled clinical trial.
expiratory pressure for acute lung injury and acute respiratory distress Am J Respir Crit Care Med 173(3):281287
syndrome: a randomized controlled trial. JAMA 299(6):637645 409. Gao Smith F, Perkins GD, Gates S etal (2012) Effect of intravenous
387. Antonelli M, Conti G, Rocco M etal (1998) A comparison of noninvasive beta-2 agonist treatment on clinical outcomes in acute respiratory
positive-pressure ventilation and conventional mechanical ventilation distress syndrome (BALTI-2): a multicentre, randomised controlled trial.
in patients with acute respiratory failure. N Engl J Med 339(7):429435 Lancet 379(9812):229235
388. Ferrer M, Esquinas A, Leon M, Gonzalez G, Alarcon A, Torres A (2003) 410. Matthay MA, Brower RG, Carson S etal (2011) Randomized, placebo-
Noninvasive respiratory ventilation in severe hypoxemic failure: a rand controlled clinical trial of an aerosolized beta(2)-agonist for treatment
omized clinical trial. Am J Respir Crit Care Med 168(12):14381444 of acute lung injury. Am J Respir Crit Care Med 184(5):561568
389. Frat JP, Thille AW, Mercat A etal (2015) High-flow oxygen through 411. Matthay MA, Brower RG, Carson S etal (2011) Randomized, placebo-
nasal cannula in acute hypoxemic respiratory failure. N Engl J Med controlled clinical trial of an aerosolized beta(2)-agonist for treatment
372(23):21852196 of acute lung injury. Am J Respir Crit Care Med 184(5):561568
390. Klessig HT, Geiger HJ, Murray MJ, Coursin DB (1992) A national survey 412. Singh B, Tiwari AK, Singh K etal (2014) Beta2 agonist for the treatment
on the practice patterns of anesthesiologist intensivists in the use of of acute lung injury: a systematic review and meta-analysis. Respir Care.
muscle-relaxants. Crit Care Med 20(9):13411345 59(2):288296
391. Murray MJ, Cowen J, DeBlock H etal (2002) Clinical practice guidelines 413. Connors AF Jr, Speroff T, Dawson NV etal (1996) The effectiveness of
for sustained neuromuscular blockade in the adult critically ill patient. right heart catheterization in the initial care of critically ill patients. SUP
Crit Care Med 30(1):142156 PORT Investigators. JAMA. 276(11):889897
392. Hansen-Flaschen JH, Brazinsky S, Basile C, Lanken PN (1991) Use of 414. Iberti TJ, Fischer EP, Leibowitz AB, Panacek EA, Silverstein JH, Albertson
sedating drugs and neuromuscular blocking agents in patients requir TE (1990) A multicenter study of physicians knowledge of the pulmo
ing mechanical ventilation for respiratory failure. A national survey. nary artery catheter. JAMA 264(22):29282932
JAMA. 266(20):28702875 415. Osman D, Ridel C, Ray P etal (2007) Cardiac filling pressures are not
393. Freebairn RC, Derrick J, Gomersall CD, Young RJ, Joynt GM (1997) Oxy appropriate to predict hemodynamic response to volume challenge.
gen delivery, oxygen consumption, and gastric intramucosal pH are not Crit Care Med 35(1):6468
improved by a computer-controlled, closed-loop, vecuronium infusion 416. Richard C, Warszawski J, Anguel N etal (2003) Early use of the pulmo
in severe sepsis and septic shock. Crit Care Med 25(1):7277 nary artery catheter and outcomes in patients with shock and acute
394. Papazian L, Forel JM, Gacouin A etal (2010) Neuromuscular block respiratory distress syndrome: a randomized controlled trial. JAMA
ers in early acute respiratory distress syndrome. N Engl J Med 290(20):27132720
363(12):11071116 417. Wheeler AP, Bernard GR, Thompson BT etal (2006) Pulmonary artery
395. Alhazzani W, Alshahrani M, Jaeschke R etal (2013) Neuromuscular versus central venous catheter to guide treatment of acute lung injury.
blocking agents in acute respiratory distress syndrome: a systematic N Engl J Med 354(21):22132224
review and meta-analysis of randomized controlled trials. Crit Care 418. Harvey S, Harrison DA, Singer M etal (2005) Assessment of the clinical
17(2):R43 effectiveness of pulmonary artery catheters in management of patients
396. Forel JM, Roch A, Marin V etal (2006) Neuromuscular blocking agents in intensive care (PAC-Man): a randomised controlled trial. Lancet
decrease inflammatory response in patients presenting with acute 366(9484):472477
respiratory distress syndrome. Crit Care Med 34(11):27492757 419. Rhodes A, Cusack RJ, Newman PJ, Grounds RM, Bennett ED (2002) A
397. Shapiro BA, Warren J, Egol AB etal (1995) Practice parameters for randomised, controlled trial of the pulmonary artery catheter in criti
sustained neuromuscular blockade in the adult critically ill patient: an cally ill patients. Intensive Care Med 28(3):256264
executive summary. Society of Critical Care Medicine. Crit Care Med 420. Sandham JD, Hull RD, Brant RF etal (2003) A randomized, controlled
23(9):16011605 trial of the use of pulmonary-artery catheters in high-risk surgical
398. Meyer KC, Prielipp RC, Grossman JE, Coursin DB (1994) Prolonged weak patients. N Engl J Med 348(1):514
ness after infusion of atracurium in two intensive care unit patients. 421. Shah MR, Hasselblad V, Stevenson LW etal (2005) Impact of the pulmo
Anesth Analg 78(4):772774 nary artery catheter in critically ill patients: meta-analysis of randomized
399. Lacomis D, Petrella JT, Giuliani MJ (1998) Causes of neuromuscular clinical trials. JAMA 294(13):16641670
weakness in the intensive care unit: a study of ninety-two patients. 422. Petrucci N, De Feo C (2013) Lung protective ventilation strategy for
Muscle Nerve 21(5):610617 the acute respiratory distress syndrome. Cochrane Database Syst Rev
400. Johnson KL, Cheung RB, Johnson SB, Roberts M, Niblett J, Manson D 2:CD003844
(1999) Therapeutic paralysis of critically ill trauma patients: perceptions 423. Futier E, Constantin JM, Paugam-Burtz C etal (2013) A trial of intraop
of patients and their family members. Am J Crit Care 8(1):490498 erative low-tidal-volume ventilation in abdominal surgery. N Engl J Med
401. Ballard N, Robley L, Barrett D, Fraser D, Mendoza I (2006) Patients recol 369(5):428437
lections of therapeutic paralysis in the intensive care unit. Am J Crit 424. Pinheiro de Oliveira R, Hetzel MP, dos Anjos Silva M, Dallegrave D, Fried
Care 15(1):8694 (quiz 5) man G (2010) Mechanical ventilation with high tidal volume induces
402. Murray MJ, DeBlock H, Erstad B etal (2016) Clinical practice guidelines inflammation in patients without lung disease. Crit Care 14(2):R39
for sustained neuromuscular blockade in the adult critically ill patient. 425. Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogue S, Ferrer N
Crit Care Med 44(11):20792103 (1999) Supine body position as a risk factor for nosocomial pneumo
403. Sibbald WJ, Short AK, Warshawski FJ, Cunningham DG, Cheung H (1985) nia in mechanically ventilated patients: a randomised trial. Lancet
Thermal dye measurements of extravascular lung water in critically ill 354(9193):18511858
patients. Intravascular Starling forces and extravascular lung water in 426. Van Nieuwenhoven CA, Vandenbroucke-Grauls C, van Tiel FH etal
the adult respiratory distress syndrome. Chest 87(5):585592 (2006) Feasibility and effects of the semirecumbent position to prevent
 ventilator-associated pneumonia: a randomized study. Crit Care Med 448. Arabi YM, Dabbagh OC, Tamim HM etal (2008) Intensive versus con
34(2):396402 ventional insulin therapy: a randomized controlled trial in medical and
427. Ouellette DR, Patel S, Girard TD etal (2016) Liberation from mechanical surgical critically ill patients. Crit Care Med 36(12):31903197
ventilation: an official American College of Chest Physicians/American 449. Brunkhorst FM, Engel C, Bloos F etal (2008) Intensive insulin
Thoracic Society clinical practice guideline: inspiratory pressure aug therapy and pentastarch resuscitation in severe sepsis. N Engl J Med
mentation during spontaneous breathing trials, protocols minimizing 358(2):125139
sedation, and non-invasive ventilation immediately after extubation. 450. De La Rosa Gdel C, Donado JH, Restrepo AH etal (2008) Strict glycae
Chest 151(1):166180. doi:10.1016/j.chest.2016.10.036 mic control in patients hospitalised in a mixed medical and surgical
428. Blackwood B, Burns KE, Cardwell CR, OHalloran P (2014) Protocol intensive care unit: a randomised clinical trial. Crit Care 12(5):R120
ized versus non-protocolized weaning for reducing the duration of 451. Finfer S, Blair D, Bellomo R etal (2009) Intensive versus conventional
mechanical ventilation in critically ill adult patients. Cochrane Database glucose control in critically ill patients. N Engl J Med 360(13):12831297
Syst Rev 11:CD006904 452. Annane D, Cariou A, Maxime V etal (2010) Corticosteroid treatment
429. Ely EW, Baker AM, Dunagan DP etal (1996) Effect on the duration of and intensive insulin therapy for septic shock in adults: a randomized
mechanical ventilation of identifying patients capable of breathing controlled trial. JAMA 303(4):341348
spontaneously. N Engl J Med 335(25):18641869 453. Kalfon P, Giraudeau B, Ichai C etal (2014) Tight computerized versus
430. Kress JP, Pohlman AS, OConnor MF, Hall JB (2000) Daily interruption conventional glucose control in the ICU: a randomized controlled trial.
of sedative infusions in critically ill patients undergoing mechanical Intensive Care Med 40(2):171181
ventilation. N Engl J Med 342(20):14711477 454. Preiser JC, Devos P, Ruiz-Santana S etal (2009) A prospective ran
431. Girard TD, Kress JP, Fuchs BD etal (2008) Efficacy and safety of a paired domised multi-centre controlled trial on tight glucose control by
sedation and ventilator weaning protocol for mechanically ventilated intensive insulin therapy in adult intensive care units: the Glucontrol
patients in intensive care (Awakening and Breathing Controlled trial): a study. Intensive Care Med 35(10):17381748
randomised controlled trial. Lancet 371(9607):126134 455. Zhang RH, W; Li, T etal (2008) Evaluation of optimal goal of glucose
432. Sevransky JE, Checkley W, Herrera P etal (2015) Protocols and hospital control in critically ill patients. Chin J Clin Nutr 16:204208
mortality in critically ill patients: the United States Critical Illness and 456. Friedrich JO, Chant C, Adhikari NK (2010) Does intensive insulin therapy
Injury Trials Group Critical Illness Outcomes Study. Crit Care Med really reduce mortality in critically ill surgical patients? A reanalysis of
43(10):20762084 meta-analytic data. Crit Care 14(5):324
433. Schweickert WD, Pohlman MC, Pohlman AS etal (2009) Early physi 457. Griesdale DE, de Souza RJ, van Dam RM etal (2009) Intensive insulin
cal and occupational therapy in mechanically ventilated, critically ill therapy and mortality among critically ill patients: a meta-analysis
patients: a randomised controlled trial. Lancet 373(9678):18741882 including NICE-SUGAR study data. CMAJ 180(8):821827
434. Shehabi Y, Bellomo R, Reade MC etal (2012) Early intensive care seda 458. Kansagara D, Fu R, Freeman M, Wolf F, Helfand M (2011) Intensive insulin
tion predicts long-term mortality in ventilated critically ill patients. Am J therapy in hospitalized patients: a systematic review. Ann Intern Med
Respir Crit Care Med 186(8):724731 154(4):268282
435. Brook AD, Ahrens TS, Schaiff R etal (1999) Effect of a nursing-imple 459. Marik PE, Preiser JC (2010) Toward understanding tight glycemic control
mented sedation protocol on the duration of mechanical ventilation. in the ICU: a systematic review and metaanalysis. Chest 137(3):544551
Crit Care Med 27(12):26092615 460. Wiener RS, Wiener DC, Larson RJ (2008) Benefits and risks of tight glu
436. Bucknall TK, Manias E, Presneill JJ (2008) A randomized trial of protocol- cose control in critically ill adults: a meta-analysis. JAMA 300(8):933944
directed sedation management for mechanical ventilation in an 461. Ling Y, Li X, Gao X (2012) Intensive versus conventional glucose control
Australian intensive care unit. Crit Care Med 36(5):14441450 in critically ill patients: a meta-analysis of randomized controlled trials.
437. Kollef MH, Levy NT, Ahrens TS, Schaiff R, Prentice D, Sherman G (1998) Eur J Intern Med. 23(6):564574
The use of continuous IV sedation is associated with prolongation of 462. Song F, Zhong LJ, Han L etal (2014) Intensive insulin therapy for septic
mechanical ventilation. Chest 114(2):541548 patients: a meta-analysis of randomized controlled trials. Biomed Res
438. Carson SS, Kress JP, Rodgers JE etal (2006) A randomized trial of inter Int. 2014:698265
mittent lorazepam versus propofol with daily interruption in mechani 463. American Diabetes Association (2014) Standards of medical care in
cally ventilated patients. Crit Care Med 34(5):13261332 diabetes2014. Diabetes Care 37(Suppl 1):S14S80
439. Mehta S, Burry L, Cook D etal (2012) Daily sedation interruption in 464. Marvin MR, Inzucchi SE, Besterman BJ (2013) Computerization of
mechanically ventilated critically ill patients cared for with a sedation the Yale insulin infusion protocol and potential insights into causes
protocol: a randomized controlled trial. JAMA 308(19):19851992 of hypoglycemia with intravenous insulin. Diabetes Technol Ther.
440. Jansen JP, Naci H (2013) Is network meta-analysis as valid as standard 15(3):246252
pairwise meta-analysis? It all depends on the distribution of effect 465. Qaseem A, Chou R, Humphrey LL (2014) Shekelle P; Clinical Guidelines
modifiers. BMC Med 11:159 Committee of the American College of Physicians. Inpatient glycemic
441. Strom T, Martinussen T, Toft P (2010) A protocol of no sedation for criti control: best practice advice from the Clinical Guidelines Committee of
cally ill patients receiving mechanical ventilation: a randomised trial. the American College of Physicians. Am J Med Qual 29(2):9598
Lancet 375(9713):475480 466. Siegelaar SE, Hermanides J, Oudemans-van Straaten HM etal (2010)
442. Lonardo NW, Mone MC, Nirula R etal (2014) Propofol is associ Mean glucose during ICU admission is related to mortality by a
ated with favorable outcomes compared with benzodiazepines in U-shaped curve in surgical and medical patients: a retrospective cohort
ventilated intensive care unit patients. Am J Respir Crit Care Med study. Crit Care 14(6):R224
189(11):13831394 467. Badawi O, Waite MD, Fuhrman SA, Zuckerman IH (2012) Association
443. Fraser GL, Devlin JW, Worby CP etal (2013) Benzodiazepine versus between intensive care unit-acquired dysglycemia and in-hospital
nonbenzodiazepine-based sedation for mechanically ventilated, criti mortality. Crit Care Med 40(12):31803188
cally ill adults: a systematic review and meta-analysis of randomized 468. Finfer S, Liu B, Chittock DR etal (2012) Hypoglycemia and risk of death
trials. Crit Care Med 41(9 Suppl 1):S30S38 in critically ill patients. N Engl J Med 367(12):11081118
444. Reade MC, Eastwood GM, Bellomo R etal (2016) Effect of dexmedeto 469. Kalfon P, Le Manach Y, Ichai C etal (2015) Severe and multiple hypo
midine added to standard care on ventilator-free time in patients with glycemic episodes are associated with increased risk of death in ICU
agitated delirium: a randomized clinical trial. JAMA 315(14):14601468 patients. Crit Care 19:153
445. Barr J, Fraser GL, Puntillo K etal (2013) Clinical practice guidelines for 470. Krinsley JS (2008) Glycemic variability: a strong independent predictor
the management of pain, agitation, and delirium in adult patients in of mortality in critically ill patients. Crit Care Med 36(11):30083013
the intensive care unit. Crit Care Med 41(1):263306 471. Todi S, Bhattacharya M (2014) Glycemic variability and outcome in criti
446. Van den Berghe G, Wouters P, Weekers F etal (2001) Intensive insulin cally ill. Indian J Crit Care Med. 18(5):285290
therapy in critically ill patients. N Engl J Med 345(19):13591367 472. Kauffmann RM, Hayes RM, Jenkins JM etal (2011) Provision of balanced
447. Van den Berghe G, Wilmer A, Hermans G etal (2006) Intensive insulin nutrition protects against hypoglycemia in the critically ill surgical
therapy in the medical ICU. N Engl J Med 354(5):449461 patient. JPEN J Parenter Enteral Nutr 35(6):686694
 473. Egi M, Bellomo R, Stachowski E etal (2008) Blood glucose concentra 495. Kellum JA, Angus DC, Johnson JP etal (2002) Continuous versus inter
tion and outcome of critical illness: the impact of diabetes. Crit Care mittent renal replacement therapy: a meta-analysis. Intensive Care Med
Med 36(8):22492255 28(1):2937
474. Krinsley JS (2009) Glycemic variability and mortality in critically ill 496. Tonelli M, Manns B, Feller-Kopman D (2002) Acute renal failure in the
patients: the impact of diabetes. J Diabetes Sci Technol. 3(6):12921301 intensive care unit: a systematic review of the impact of dialytic modal
475. Krinsley JS, Preiser JC (2015) Time in blood glucose range 70 to 140mg/ ity on mortality and renal recovery. Am J Kidney Dis 40(5):875885
dl>80% is strongly associated with increased survival in non-diabetic 497. Augustine JJ, Sandy D, Seifert TH, Paganini EP (2004) A randomized con
critically ill adults. Crit Care 19:179 trolled trial comparing intermittent with continuous dialysis in patients
476. Egi M, Bellomo R, Stachowski E etal (2011) The interaction of chronic with ARF. Am J Kidney Dis 44(6):10001007
and acute glycemia with mortality in critically ill patients with diabetes. 498. Gasparovic V, Filipovic-Grcic I, Merkler M, Pisl Z (2003) Continuous renal
Crit Care Med 39(1):105111 replacement therapy (CRRT) or intermittent hemodialysis (IHD)What
477. Sandler V, Misiasz MR, Jones J, Baldwin D (2014) Reducing the risk of is the procedure of choice in critically ill patients? Ren Fail 25(5):855862
hypoglycemia associated with intravenous insulin: experience with a 499. Mehta RL, McDonald B, Gabbai FB etal (2001) A randomized clinical
computerized insulin infusion program in 4 adult intensive care units. J trial of continuous versus intermittent dialysis for acute renal failure.
Diabetes Sci Technol. 8(5):923929 Kidney Int 60(3):11541163
478. Pereira AJ, Correa TD, de Almeida FP etal (2015) Inaccuracy of venous 500. Uehlinger DE, Jakob SM, Ferrari P etal (2005) Comparison of continu
point-of-care glucose measurements in critically ill patients: a cross- ous and intermittent renal replacement therapy for acute renal failure.
sectional study. PLoS ONE 10(6):e0129568 Nephrol Dial Transplant 20(8):16301637
479. Hoedemaekers CWE, Gunnewiek JMTK, Prinsen MA, Willems JL, 501. Vinsonneau C, Camus C, Combes A etal (2006) Continuous veno
Van der Hoeven JG (2008) Accuracy of bedside glucose measure venous haemodiafiltration versus intermittent haemodialysis for acute
ment from three glucometers in critically ill patients. Crit Care Med renal failure in patients with multiple-organ dysfunction syndrome: a
36(11):30623066 multicentre randomised trial. Lancet 368(9533):379385
480. Inoue S, Egi M, Kotani J, Morita K (2013) Accuracy of blood-glucose 502. John S, Griesbach D, Baumgartel M, Weihprecht H, Schmieder RE,
measurements using glucose meters and arterial blood gas analyzers in Geiger H (2001) Effects of continuous haemofiltration vs intermittent
critically ill adult patients: systematic review. Crit Care 17(2):R48 haemodialysis on systemic haemodynamics and splanchnic regional
481. Kanji S, Buffie J, Hutton B etal (2005) Reliability of point-of-care perfusion in septic shock patients: a prospective, randomized clinical
testing for glucose measurement in critically ill adults. Crit Care Med trial. Nephrol Dial Transplant 16(2):320327
33(12):27782785 503. Misset B, Timsit JF, Chevret S, Renaud B, Tamion F, Carlet J (1996) A
482. Khan AI, Vasquez Y, Gray J, Wians FH Jr, Kroll MH (2006) The variability of randomized cross-over comparison of the hemodynamic response
results between point-of-care testing glucose meters and the central to intermittent hemodialysis and continuous hemofiltration in ICU
laboratory analyzer. Arch Pathol Lab Med 130(10):15271532 patients with acute renal failure. Intensive Care Med 22(8):742746
483. Rice MJ, Coursin DB (2016) Glucose meters: here today, gone tomor 504. Bouman CSC, Oudemans-van Straaten HM, Tijssen JGP, Zandstra
row? Crit Care Med 44(2):e97e100 DF, Kesecioglu J (2002) Effects of early high-volume continuous
484. Klonoff DC, Draznin B, Drincic A etal (2015) PRIDE Statement on the venovenous hemofiltration on survival and recovery of renal function
need for a moratorium on the CMS plan to cite hospitals for perform in intensive care patients with acute renal failure: a prospective, rand
ing point-of-care capillary blood glucose monitoring on critically ill omized trial. Crit Care Med 30(10):22052211
patients. J Clin Endocrinol Metab 100(10):36073612 505. Ronco C, Bellomo R, Homel P etal (2000) Effects of different doses in
485. Wilson M, Weinreb J, Hoo GW (2007) Intensive insulin therapy in critical continuous veno-venous haemofiltration on outcomes of acute renal
care: a review of 12 protocols. Diabetes Care 30(4):10051011 failure: a prospective randomised trial. Lancet 356(9223):2630
486. Dortch MJ, Mowery NT, Ozdas A etal (2008) A computerized insulin 506. Bellomo R, Cass A, Cole L etal (2009) Intensity of continuous
infusion titration protocol improves glucose control with less hypogly renal-replacement therapy in critically ill patients. N Engl J Med
cemia compared to a manual titration protocol in a trauma intensive 361(17):16271638
care unit. JPEN J Parenter Enteral Nutr 32(1):1827 507. Palevsky PM, Zhang JH, OConnor TZ etal (2008) Intensity of renal
487. Newton CA, Smiley D, Bode BW etal (2010) A comparison study of support in critically ill patients with acute kidney injury. N Engl J Med
continuous insulin infusion protocols in the medical intensive care unit: 359(1):720
computer-guided vs. standard column-based algorithms. J Hosp Med. 508. Gaudry S, Hajage D, Schortgen F etal (2016) Initiation strategies for
5(8):432437 renal replacement therapy in the intensive care unit. N Engl J Med
488. Bartlett RH, Mault JR, Dechert RE, Palmer J, Swartz RD, Port FK (1986) 375(2):122133
Continuous arteriovenous hemofiltration: improved survival in surgical 509. Zarbock A, Kellum JA, Schmidt C etal (2016) Effect of early vs delayed
acute renal failure. Surgery. 100(2):400408 initiation of renal replacement therapy on mortality in critically ill
489. Bellomo R, Farmer M, Parkin G, Wright C, Boyce N (1995) Severe acute patients with acute kidney injury: the ELAIN randomized clinical trial.
renal failure: a comparison of acute continuous hemodiafiltration and JAMA 315(20):21902199
conventional dialytic therapy. Nephron. 71(1):5964 510. Cooper DJ, Walley KR, Wiggs BR, Russell JA (1990) Bicarbonate does not
490. Bellomo R, Mansfield D, Rumble S, Shapiro J, Parkin G, Boyce N (1992) improve hemodynamics in critically ill patients who have lactic acidosis:
Acute renal failure in critical illness. Conventional dialysis versus acute a prospective, controlled clinical study. Ann Intern Med 112(7):492498
continuous hemodiafiltration. ASAIO J 38(3):M654M657 511. Mathieu D, Neviere R, Billard V, Fleyfel M, Wattel F (1991) Effects of bicar
491. Kierdorf H (1991) Continuous versus intermittent treatment: clinical bonate therapy on hemodynamics and tissue oxygenation in patients
results in acute renal failure. In: Sieberth HG, Mann H, Stummvoll HK with lactic acidosis: a prospective, controlled clinical study. Crit Care
(eds) Continuous hemofiltration. Karger, Basel, pp 112 Med 19(11):13521356
492. Mauritz W, Sporn P, Schindler I, Zadrobilek E, Roth E, Appel W (1986) 512. Cook D, Crowther M, Meade M etal (2005) Deep venous thrombosis
Acute renal failure in abdominal infection. Comparison of hemodialysis in medical-surgical critically ill patients: prevalence, incidence, and risk
and continuous arteriovenous hemofiltration. Anasth Intensivther factors. Crit Care Med 33(7):15651571
Notfallmed. 21(4):212217 513. Kahn SR, Lim W, Dunn AS etal (2012) Prevention of VTE in nonsurgical
493. Guerin C, Girard R, Selli JM, Ayzac L (2002) Intermittent versus continu patients: antithrombotic Therapy and Prevention of Thrombosis, 9th ed:
ous renal replacement therapy for acute renal failure in intensive care American College of Chest Physicians evidence-based clinical practice
units: results from a multicenter prospective epidemiological survey. guidelines. Chest 141(2 Suppl):e195Se226S
Intensive Care Med 28(10):14111418 514. Alhazzani W, Lim W, Jaeschke RZ, Murad MH, Cade J, Cook DJ (2013)
494. Van Bommel E, Bouvy ND, So KL etal (1995) Acute dialytic support for Heparin thromboprophylaxis in medical-surgical critically ill patients: a
the critically iII: intermittent hemodialysis versus continuous arterio systematic review and meta-analysis of randomized trials. Crit Care Med
venous hemodiafiltration. Am J Nephrol 15(3):192200 41(9):20882098
 515. Levi M, Levy M, Williams MD etal (2007) Prophylactic heparin in 536. DAncona G, Baillot R, Poirier B etal (2003) Determinants of
patients with severe sepsis treated with drotrecogin alfa (activated). Am gastrointestinal complications in cardiac surgery. Tex Heart Inst J
J Respir Crit Care Med 176(5):483490 30(4):280285
516. Beitland S, Sandven I, Kjaervik LK, Sandset PM, Sunde K, Eken T (2015) 537. Faisy C, Guerot E, Diehl JL, Iftimovici E, Fagon JY (2003) Clinically signifi
Thromboprophylaxis with low molecular weight heparin versus unfrac cant gastrointestinal bleeding in critically ill patients with and without
tionated heparin in intensive care patients: a systematic review with meta- stress ulcer prophylaxis. Intensive Care Med 29(8):13061313
analysis and trial sequential analysis. Intensive Care Med 41(7):12091219 538. Krag M, Perner A, Wetterslev J, Wise MP, Hylander Moller M (2014)
517. Phung OJ, Kahn SR, Cook DJ, Murad MH (2011) Dosing frequency of Stress ulcer prophylaxis versus placebo or no prophylaxis in critically ill
unfractionated heparin thromboprophylaxis: a meta-analysis. Chest patients. A systematic review of randomised clinical trials with meta-
140(2):374381 analysis and trial sequential analysis. Intensive Care Med 40(1):1122
518. Mahan CE, Pini M, Spyropoulos AC (2010) Venous thromboembolism 539. Sasabuchi Y, Matsui H, Lefor AK, Fushimi K, Yasunaga H (2016) Risks and
prophylaxis with unfractionated heparin in the hospitalized medical benefits of stress ulcer prophylaxis for patients with severe sepsis. Crit
patient: the case for thrice daily over twice daily dosing. Intern Emerg Care Med 44(7):e464e469
Med 5(4):299306 540. Eastwood GM, Litton E, Bellomo R etal (2014) Opinions and practice of
519. Junqueira DR, Perini E, Penholati RR Carvalho MG (2012) Unfractionated stress ulcer prophylaxis in Australian and New Zealand intensive care
heparin versus low molecular weight heparin for avoiding heparin- units. Crit Care Resusc 16(3):170174
induced thrombocytopenia in postoperative patients. Cochrane 541. Krag M, Perner A, Wetterslev J etal (2015) Stress ulcer prophylaxis in the
Database Syst Rev 9:CD007557 intensive care unit: an international survey of 97 units in 11 countries.
520. Cook D, Meade M, Guyatt G etal (2011) Dalteparin versus unfraction Acta Anaesthesiol Scand 59(5):576585
ated heparin in critically ill patients. N Engl J Med 364(14):13051314 542. Preslaski CR, Mueller SW, Kiser TH, Fish DN, MacLaren R (2014) A
521. Fowler RA, Mittmann N, Geerts W etal (2014) Cost-effectiveness of survey of prescriber perceptions about the prevention of stress-
dalteparin vs unfractionated heparin for the prevention of venous related mucosal bleeding in the intensive care unit. J Clin Pharm Ther
thromboembolism in critically ill patients. JAMA 312(20):21352145 39(6):658662
522. Douketis J, Cook D, Meade M etal (2008) Prophylaxis against deep 543. Shears M, Alhazzani W, Marshall JC etal (2016) Stress ulcer prophylaxis
vein thrombosis in critically ill patients with severe renal insufficiency in critical illness: a Canadian survey. Can J Anaesth 63(6):718724
with the low-molecular-weight heparin dalteparin: an assessment of 544. Alshamsi F, Belley-Cote E, Cook D etal (2016) Efficacy and safety of pro
safety and pharmacodynamics: the DIRECT study. Arch Intern Med ton pump inhibitors for stress ulcer prophylaxis in critically ill patients:
168(16):18051812 a systematic review and meta-analysis of randomized trials. Crit Care
523. Arabi YM, Alsolamy S, Al-Dawood A etal (2016) Thromboprophylaxis 20(1):120
using combined intermittent pneumatic compression and pharmaco 545. Alhazzani W, Alenezi F, Jaeschke RZ, Moayyedi P, Cook DJ (2013) Proton
logic prophylaxis versus pharmacologic prophylaxis alone in critically ill pump inhibitors versus histamine 2 receptor antagonists for stress ulcer
patients: study protocol for a randomized controlled trial. Trials. 17:390 prophylaxis in critically ill patients: a systematic review and meta-analy
524. Kakkos SK, Caprini JA, Geroulakos G, Nicolaides AN, Stansby GP, Reddy sis. Crit Care Med 41(3):693705
DJ (2008) Combined intermittent pneumatic leg compression and 546. Barkun AN, Bardou M, Pham CQ, Martel M (2012) Proton pump inhibi
pharmacological prophylaxis for prevention of venous thromboembo tors vs. histamine 2 receptor antagonists for stress-related mucosal
lism in high-risk patients. Cochrane Database Syst Rev 4:CD005258 bleeding prophylaxis in critically ill patients: a meta-analysis. Am J
525. Falck-Ytter Y, Francis CW, Johanson NA etal (2012) Prevention of VTE in Gastroenterol 107(4):507520 (quiz 21)
orthopedic surgery patients: antithrombotic Therapy and Prevention of 547. Barkun AN, Adam V, Martel M, Bardou M (2013) Cost-effectiveness
Thrombosis, 9th ed: American College of Chest Physicians evidence- analysis: stress ulcer bleeding prophylaxis with proton pump inhibitors,
based clinical practice guidelines. Chest 141(2 Suppl):e278Se325S H2 receptor antagonists. Value Health. 16(1):1422
526. Gould MK, Garcia DA, Wren SM etal (2012) Prevention of VTE in non 548. MacLaren R, Campbell J (2014) Cost-effectiveness of histamine recep
orthopedic surgical patients: antithrombotic Therapy and Prevention of tor-2 antagonist versus proton pump inhibitor for stress ulcer prophy
Thrombosis, 9th ed: American College of Chest Physicians Evidence- laxis in critically ill patients. Crit Care Med 42(4):809815
Based Clinical Practice Guidelines. Chest 141(2 Suppl):e227Se277S 549. Villet S, Chiolero RL, Bollmann MD etal (2005) Negative impact of
527. Pavon JM, Adam SS, Razouki ZA etal (2016) Effectiveness of intermit hypocaloric feeding and energy balance on clinical outcome in ICU
tent pneumatic compression devices for venous thromboembolism patients. Clin Nutr. 24(4):502509
prophylaxis in high-risk surgical patients: a systematic review. J Arthro 550. Adams S, Dellinger EP, Wertz MJ, Oreskovich MR, Simonowitz D,
plasty 31(2):524532 Johansen K (1986) Enteral versus parenteral nutritional support follow
528. Sachdeva A, Dalton M, Amaragiri SV, Lees T (2014) Graduated compres ing laparotomy for trauma: a randomized prospective trial. J Trauma
sion stockings for prevention of deep vein thrombosis. Cochrane 26(10):882891
Database Syst Rev 12:CD001484 551. Borzotta AP, Pennings J, Papasadero B etal (1994) Enteral versus paren
529. Arabi YM, Khedr M, Dara SI etal (2013) Use of intermittent pneumatic teral nutrition after severe closed head injury. J Trauma 37(3):459468
compression and not graduated compression stockings is associated 552. Dunham CM, Frankenfield D, Belzberg H, Wiles C, Cushing B, Grant Z
with lower incident VTE in critically ill patients: a multiple propensity (1994) Gut failurepredictor of or contributor to mortality in mechani
scores adjusted analysis. Chest 144(1):152159 cally ventilated blunt trauma patients? J Trauma 37(1):3034
530. Cook DJ, Griffith LE, Walter SD etal (2001) The attributable mortality 553. Harvey SE, Parrott F, Harrison DA etal (2014) Trial of the route
and length of intensive care unit stay of clinically important gastrointes of early nutritional support in critically ill adults. N Engl J Med
tinal bleeding in critically ill patients. Crit Care 5(6):368375 371(18):16731684
531. Bardou M, Quenot JP, Barkun A (2015) Stress-related mucosal disease in 554. Justo Meirelles CM, de Aguilar-Nascimento JE (2011) Enteral or paren
the critically ill patient. Nat Rev Gastroenterol Hepatol. 12(2):98107 teral nutrition in traumatic brain injury: a prospective randomised trial.
532. Cook DJ, Fuller HD, Guyatt GH etal (1994) Risk factors for gastrointesti Nutr Hosp 26(5):11201124
nal bleeding in critically ill patients. N Engl J Med 330(6):377381 555. Kalfarentzos F, Kehagias J, Mead N, Kokkinis K, Gogos CA (1997) Enteral
533. Krag M, Perner A, Wetterslev J etal (2015) Prevalence and outcome of nutrition is superior to parenteral nutrition in severe acute pancreatitis:
gastrointestinal bleeding and use of acid suppressants in acutely ill results of a randomized prospective trial. Br J Surg 84(12):16651669
adult intensive care patients. Intensive Care Med 41(5):833845 556. Moore FA, Moore EE, Jones TN, McCroskey BL, Peterson VM (1989)
534. Andersson B, Nilsson J, Brandt J, Hoglund P, Andersson R (2005) Gastro TEN versus TPN following major abdominal traumareduced septic
intestinal complications after cardiac surgery. Br J Surg 92(3):326333 morbidity. J Trauma 29(7):916922 (discussion 2223)
535. Bruno JJ, Canada TW, Wakefield CD, Nates JL (2009) Stress-related 557. Peterson VM, Moore EE, Jones TN etal (1988) Total enteral nutrition
mucosal bleeding in critically ill oncology patients. J Oncol Pharm Pract. versus total parenteral nutrition after major torso injury: attenuation of
15(1):916 hepatic protein reprioritization. Surgery. 104(2):199207
 558. Sun JK, Mu XW, Li WQ, Tong ZH, Li J, Zheng SY (2013) Effects of early 581. Doig GS, Heighes PT, Simpson F, Sweetman EA, Davies AR (2009) Early
enteral nutrition on immune function of severe acute pancreatitis enteral nutrition, provided within 24h of injury or intensive care unit
patients. World J Gastroenterol 19(6):917922 admission, significantly reduces mortality in critically ill patients: a
559. Wang G, Wen J, Xu L etal (2013) Effect of enteral nutrition and ecoim meta-analysis of randomised controlled trials. Intensive Care Med
munonutrition on bacterial translocation and cytokine production in 35(12):20182027
patients with severe acute pancreatitis. J Surg Res 183(2):592597 582. Malhotra A, Mathur AK, Gupta S (2004) Early enteral nutrition after
560. Harvey SE, Parrott F, Harrison DA etal (2016) A multicentre, randomised surgical treatment of gut perforations: a prospective randomised study.
controlled trial comparing the clinical effectiveness and cost-effectiveness J Postgrad Med 50(2):102106
of early nutritional support via the parenteral versus the enteral route in 583. Pupelis G, Austrums E, Jansone A, Sprucs R, Wehbi H (2000) Ran
critically ill patients (CALORIES). Health Technol Assess 20(28):1144 domised trial of safety and efficacy of postoperative enteral feeding
561. Kudsk KA (2002) Current aspects of mucosal immunology and its influ in patients with severe pancreatitis: preliminary report. Eur J Surg
ence by nutrition. Am J Surg 183(4):390398 166(5):383387
562. McClave SA, Heyland DK (2009) The physiologic response and associ 584. Arabi YM, Aldawood AS, Haddad SH etal (2015) Permissive underfeed
ated clinical benefits from provision of early enteral nutrition. Nutr Clin ing or standard enteral feeding in critically ill adults. N Engl J Med
Pract. 24(3):305315 372(25):23982408
563. Casaer MP, Mesotten D, Hermans G etal (2011) Early versus late paren 585. Marik PE, Hooper MH (2016) Normocaloric versus hypocaloric feeding
teral nutrition in critically ill adults. N Engl J Med 365(6):506517 on the outcomes of ICU patients: a systematic review and meta-analy
564. Doig GS, Simpson F, Sweetman EA etal (2013) Early parenteral sis. Intensive Care Med 42(3):316323
nutrition in critically ill patients with short-term relative contraindica 586. Arabi YM, Tamim HM, Dhar GS etal (2011) Permissive underfeeding
tions to early enteral nutrition: a randomized controlled trial. JAMA and intensive insulin therapy in critically ill patients: a randomized
309(20):21302138 controlled trial. Am J Clin Nutr 93(3):569577
565. Rapp RP, Donaldson ES, Bivins BA (1983) Parenteral nutrition in a patient 587. Charles EJ, Petroze RT, Metzger R etal (2014) Hypocaloric compared
with familial type IV hypertriglyceridemia: a dilemma. Drug Intell Clin with eucaloric nutritional support and its effect on infection rates in
Pharm. 17(6):458460 a surgical intensive care unit: a randomized controlled trial. Am J Clin
566. Young B, Ott L, Haack D etal (1987) Effect of total parenteral nutri Nutr 100(5):13371343
tion upon intracranial pressure in severe head injury. J Neurosurg 588. Ibrahim EH, Mehringer L, Prentice D etal (2002) Early versus late enteral
67(1):7680 feeding of mechanically ventilated patients: results of a clinical trial.
567. Sadique Z, Grieve R, Harrison D, Rowan K (2015) Cost-effectiveness of JPEN J Parenter Enteral Nutr 26(3):174181
early parenteral versus enteral nutrition in critically ill patients. Value 589. National Heart, Lung, and Blood Institute Acute Respiratory Distress
Health. 18(7):A532 Syndrome (ARDS) Clinical Trials Network, Rice TW, Wheeler AP etal
568. Chiarelli A, Enzi G, Casadei A, Baggio B, Valerio A, Mazzoleni F (1990) (2012) Initial trophic vs full enteral feeding in patients with acute lung
Very early nutrition supplementation in burned patients. Am J Clin Nutr injury: the EDEN randomized trial. JAMA 307(8):795803
51(6):10351039 590. Petros S, Horbach M, Seidel F, Weidhase L (2016) Hypocaloric vs nor
569. Dvorak MF, Noonan VK, Blanger etal (2004) Early versus late enteral mocaloric nutrition in critically ill patients: a prospective randomized
feeding in patients with acute cervical spinal cord injury: a pilot study. pilot trial. JPEN J Parenter Enteral Nutr 40(2):242249
Spine (Phila Pa 1976) 29(9):E175E180 591. Rice TW, Mogan S, Hays MA, Bernard GR, Jensen GL, Wheeler AP (2011)
570. Eyer SD, Micon LT, Konstantinides FN etal (1993) Early enteral feeding Randomized trial of initial trophic versus full-energy enteral nutrition in
does not attenuate metabolic response after blunt trauma. J Trauma mechanically ventilated patients with acute respiratory failure. Crit Care
34(5):639644 Med 39(5):967974
571. Grahm TW, Zadrozny DB, Harrington T (1989) The benefits of early 592. Needham DM, Dinglas VD, Bienvenu OJ etal (2013) One year outcomes
jejunal hyperalimentation in the head-injured patient. Neurosurgery. in patients with acute lung injury randomised to initial trophic or full
25(5):729735 enteral feeding: prospective follow-up of EDEN randomised trial. BMJ
572. Hasse JM, Blue LS, Liepa GU etal (1995) Early enteral nutrition support 346:f1532
in patients undergoing liver transplantation. JPEN J Parenter Enteral 593. Garcia de Acilu M, Leal S, Caralt B, Roca O, Sabater J, Masclans JR (2015)
Nutr 19(6):437443 The role of omega-3 polyunsaturated fatty acids in the treatment of
573. Minard G, Kudsk KA, Melton S, Patton JH, Tolley EA (2000) Early versus patients with acute respiratory distress syndrome: a clinical review.
delayed feeding with an immune-enhancing diet in patients with Biomed Res Int 2015:653750
severe head injuries. JPEN J Parenter Enteral Nutr 24(3):145149 594. Manzanares W, Dhaliwal R, Jurewitsch B, Stapleton RD, Jeejeebhoy KN,
574. Moore EE, Jones TN (1986) Benefits of immediate jejunostomy feeding Heyland DK (2014) Parenteral fish oil lipid emulsions in the critically
after major abdominal-traumaa prospective, randomized study. J ill: a systematic review and meta-analysis. JPEN J Parenter Enteral Nutr
Trauma 26(10):874881 38(1):2028
575. Nguyen NQ, Fraser RJ, Bryant LK etal (2008) The impact of delay 595. Zhu D, Zhang Y, Li S, Gan L, Feng H, Nie W (2014) Enteral omega-3
ing enteral feeding on gastric emptying, plasma cholecystokinin, fatty acid supplementation in adult patients with acute respiratory
and peptide YY concentrations in critically ill patients. Crit Care Med distress syndrome: a systematic review of randomized controlled trials
36(5):14691474 with meta-analysis and trial sequential analysis. Intensive Care Med
576. Peng YZ, Yuan ZQ, Xiao GX (2001) Effects of early enteral feeding on the 40(4):504512
prevention of enterogenic infection in severely burned patients. Burns. 596. Rice TW, Wheeler AP, Thompson BT, deBoisblanc BP, Steingrub J, Rock P
27(2):145149 (2011) Enteral omega-3 fatty acid, gamma-linolenic acid, and antioxi
577. Singh G, Ram RP, Khanna SK (1998) Early postoperative enteral feeding dant supplementation in acute lung injury. JAMA 306(14):15741581
in patients with nontraumatic intestinal perforation and peritonitis. J 597. McClave SA, DeMeo MT, DeLegge MH etal (2002) North American
Am Coll Surg 187(2):142146 Summit on Aspiration in the Critically Ill Patient: consensus statement.
578. Chuntrasakul C, Chinswangwatanakul V, Chockvivatanavanit S, Siltharm JPEN J Parenter Enteral Nutr 26(6 Suppl):S80S85
S, Pongprasobchai T, Bunnak A (1996) Early nutritional support in severe 598. McClave SA, Lukan JK, Stefater JA etal (2005) Poor validity of residual
traumatic patients. J Med Assoc Thai 79(1):2126 volumes as a marker for risk of aspiration in critically ill patients. Crit
579. Chourdakis M, Kraus MM, Tzellos T etal (2012) Effect of early compared Care Med 33(2):324330
with delayed enteral nutrition on endocrine function in patients with 599. Metheny NA, Schallom L, Oliver DA, Clouse RE (2008) Gastric residual
traumatic brain injury: an open-labeled randomized trial. JPEN J Par volume and aspiration in critically ill patients receiving gastric feedings.
enter Enteral Nutr 36(1):108116 Am J Crit Care 17(6):512519 (quiz 20)
580. Doig GS, Heighes PT, Simpson F, Sweetman EA (2011) Early enteral 600. Montejo JC, Minambres E, Bordeje L etal (2010) Gastric residual volume
nutrition reduces mortality in trauma patients requiring intensive care: during enteral nutrition in ICU patients: the REGANE study. Intensive
a meta-analysis of randomised controlled trials. Injury. 42(1):5056 Care Med 36(8):13861393
 601. Poulard F, Dimet J, Martin-Lefevre L etal (2010) Impact of not meas 622. Caparros T, Lopez J, Grau T (2001) Early enteral nutrition in critically ill
uring residual gastric volume in mechanically ventilated patients patients with a high-protein diet enriched with arginine, fiber, and antioxi
receiving early enteral feeding: a prospective before-after study. JPEN J dants compared with a standard high-protein diet. The effect on nosoco
Parenter Enteral Nutr 34(2):125130 mial infections and outcome. JPEN J Parenter Enteral Nutr 25(6):299309
602. Reignier J, Mercier E, Le Gouge A etal (2013) Effect of not monitoring 623. Galban C, Montejo JC, Mesejo A etal (2000) An immune-enhancing
residual gastric volume on risk of ventilator-associated pneumonia in enteral diet reduces mortality rate and episodes of bacteremia in septic
adults receiving mechanical ventilation and early enteral feeding: a intensive care unit patients. Crit Care Med 28(3):643648
randomized controlled trial. JAMA 309(3):249256 624. Santora R, Kozar RA (2010) Molecular mechanisms of pharmaconutri
603. Elke G, Felbinger TW, Heyland DK (2015) Gastric residual volume in criti ents. J Surg Res 161(2):288294
cally ill patients: a dead marker or still alive? Nutr Clin Pract. 30(1):5971 625. Kieft H, Roos AN, van Drunen JDE, Bindels A, Bindels JG, Hofman Z
604. Van Noord C, Dieleman JP, van Herpen G, Verhamme K, Sturkenboom (2005) Clinical outcome of immunonutrition in a heterogeneous inten
MC (2010) Domperidone and ventricular arrhythmia or sudden cardiac sive care population. Intensive Care Med 31(4):524532
death: a population-based case-control study in the Netherlands. Drug 626. Marik PE, Zaloga GP (2001) Early enteral nutrition in acutely ill patients:
Saf 33(11):10031014 a systematic review. Crit Care Med 29(12):22642270
605. Ray WA, Murray KT, Meredith S, Narasimhulu SS, Hall K, Stein CM (2004) 627. Avenell A (2006) Glutamine in critical care: current evidence from
Oral erythromycin and the risk of sudden death from cardiac causes. N systematic reviews. Proc Nutr Soc. 65(3):236241
Engl J Med 351(11):10891096 628. Avenell A (2009) Hot topics in parenteral nutrition. Current evidence
606. Lewis K, Alqahtani Z, McIntyre L etal (2016) The efficacy and safety of and ongoing trials on the use of glutamine in critically-ill patients and
prokinetic agents in critically ill patients receiving enteral nutrition: patients undergoing surgery. Proc Nutr Soc. 68(3):261268
a systematic review and meta-analysis of randomized trials. Crit Care 629. Jiang H, Chen W, Hu W, Cai B, Liao RJ (2009) The impact of glutamine-
20(1):259 enhanced enteral nutrition on clinical outcome of patients with critical
607. Freeman BD, Dixon DJ, Coopersmith CM, Zehnbauer BA, Buchman illness: a systematic review of randomized controlled trials. Zhonghua
TG (2008) Pharmacoepidemiology of QT-interval prolonging drug Shao Shang Za Zhi. 25(5):325330
administration in critically ill patients. Pharmacoepidemiol Drug Saf 630. Novak F, Heyland DK, Avenell A, Drover JW, Su XY (2002) Glutamine
17(10):971981 supplementation in serious illness: a systematic review of the evidence.
608. Fruhwald S, Kainz J (2010) Effect of ICU interventions on gastrointestinal Crit Care Med 30(9):20222029
motility. Curr Opin Crit Care. 16(2):159164 631. Grau T, Bonet A, Minambres E etal (2011) The effect of L-alanyl-l-glu
609. Mentec H, Dupont H, Bocchetti M, Cani P, Ponche F, Bleichner G (2001) tamine dipeptide supplemented total parenteral nutrition on infectious
Upper digestive intolerance during enteral nutrition in critically ill morbidity and insulin sensitivity in critically ill patients. Crit Care Med
patients: frequency, risk factors, and complications. Crit Care Med 39(6):12631268
29(10):19551961 632. Wang Z, Forceville X, Van Antwerpen P etal (2009) A large-bolus injec
610. Dive A, Foret F, Jamart J, Bulpa P, Installe E (2000) Effect of dopamine tion, but not continuous infusion of sodium selenite improves outcome
on gastrointestinal motility during critical illness. Intensive Care Med in peritonitis. Shock. 32(2):140146
26(7):901907 633. Wernerman J, Kirketeig T, Andersson B etal (2011) Scandinavian
611. Dive A, Miesse C, Jamart J, Evrard P, Gonzalez M, Installe E (1994) glutamine trial: a pragmatic multi-centre randomised clinical trial of
Duodenal motor response to continuous enteral feeding is impaired in intensive care unit patients. Acta Anaesthesiol Scand 55(7):812818
mechanically ventilated critically ill patients. Clin Nutr. 13(5):302306 634. Heyland D, Muscedere J, Wischmeyer PE etal (2013) A randomized
612. Zaloga GP, Marik P (2000) Promotility agents in the intensive care unit. trial of glutamine and antioxidants in critically ill patients. N Engl J Med
Crit Care Med 28(7):26572659 368(16):14891497
613. Tiancha H, Jiyong J, Min Y (2015) How to promote bedside placement 635. Beale RJ, Sherry T, Lei K etal (2008) Early enteral supplementation with
of the postpyloric feeding tube: a network meta-analysis of randomized key pharmaconutrients improves sequential organ failure assessment
controlled trials. JPEN J Parenter Enteral Nutr 39(5):521530 score in critically ill patients with sepsis: outcome of a randomized,
614. Alhazzani W, Almasoud A, Jaeschke R etal (2013) Small bowel feeding controlled, double-blind trial. Crit Care Med 36(1):131144
and risk of pneumonia in adult critically ill patients: a systematic review 636. Fuentes-Orozco C, Anaya-Prado R, Gonzalez-Ojeda A etal (2004)
and meta-analysis of randomized trials. Crit Care 17(4):R127 L-alanyl-l-glutamine-supplemented parenteral nutrition improves
615. Deane AM, Dhaliwal R, Day AG, Ridley EJ, Davies AR, Heyland DK (2013) infectious morbidity in secondary peritonitis. Clin Nutr. 23(1):1321
Comparisons between intragastric and small intestinal delivery of 637. Puskarich MA, Kline JA, Krabill V, Claremont H, Jones AE (2014) Prelimi
enteral nutrition in the critically ill: a systematic review and meta-analy nary safety and efficacy of L-carnitine infusion for the treatment of
sis. Crit Care 17(3):R125 vasopressor-dependent septic shock: a randomized control trial. JPEN J
616. Alhazzani W, Jacobi J, Sindi A etal (2013) The effect of selenium therapy Parenter Enteral Nutr 38(6):736743
on mortality in patients with sepsis syndrome: a systematic review 638. White DB, Engelberg RA, Wenrich MD, Lo B, Curtis JR (2010) The
and meta-analysis of randomized controlled trials. Crit Care Med language of prognostication in intensive care units. Med Decis Making
41(6):15551564 30(1):7683
617. Valenta J, Brodska H, Drabek T, Hendl J, Kazda A (2011) High-dose 639. Chiarchiaro J, Buddadhumaruk P, Arnold RM, White DB (2015) Quality of
selenium substitution in sepsis: a prospective randomized clinical trial. communication in the ICU and surrogates understanding of prognosis.
Intensive Care Med 37(5):808815 Crit Care Med 43(3):542548
618. Bloos F, Trips E, Nierhaus A etal (2016) Effect of sodium selenite admin 640. Downar J, You JJ, Bagshaw SM etal (2015) Nonbeneficial treatment
istration and procalcitonin-guided therapy on mortality in patients with Canada: definitions, causes, and potential solutions from the perspec
severe sepsis or septic shock: a randomized clinical trial. JAMA Int Med. tive of healthcare practitioners. Crit Care Med 43(2):270281
176(9):12661276 641. Kon AA, Shepard EK, Sederstrom NO etal (2016) Defining futile and
619. Bertolini G, Iapichino G, Radrizzani D etal (2003) Early enteral potentially inappropriate interventions: a policy statement from the
immunonutrition in patients with severe sepsis: results of an interim Society of Critical Care Medicine Ethics Committee. Crit Care Med
analysis of a randomized multicentre clinical trial. Intensive Care Med 44(9):17691774
29(5):834840 642. Nelson JE, Curtis JR, Mulkerin C etal (2013) Choosing and using screen
620. Suchner U, Kuhn KS, Furst P (2000) The scientific basis of immunonutri ing criteria for palliative care consultation in the ICU: a report from the
tion. Proc Nutr Soc. 59(4):553563 Improving Palliative Care in the ICU (IPAL-ICU) Advisory Board. Crit Care
621. Bower RH, Cerra FB, Bershadsky B etal (1995) Early enteral administra Med 41(10):23182327
tion of a formula (Impact) supplemented with arginine, nucleotides, 643. Detering KM, Hancock AD, Reade MC, Silvester W (2010) The impact
and fish oil in intensive care unit patients: results of a multicenter, of advance care planning on end of life care in elderly patients: ran
prospective, randomized, clinical trial. Crit Care Med 23(3):436449 domised controlled trial. BMJ 340:c1345. doi:10.1136/bmj.c1345
 644. Scheunemann LP, McDevitt M, Carson SS, Hanson LC (2011) Rand 651. Sprung CL, Truog RD, Curtis JR etal (2014) Seeking worldwide profes
omized, controlled trials of interventions to improve communication in sional consensus on the principles of end-of-life care for the critically
intensive care: a systematic review. Chest 139(3):543554 ill. The Consensus for Worldwide End-of-Life Practice for Patients in
645. Bosslet GT, Pope TM, Rubenfeld GD etal (2015) An official ATS/AACN/ Intensive Care Units (WELPICUS) study. Am J Respir Crit Care Med
ACCP/ESICM/SCCM policy statement: responding to requests for 190(8):855866
potentially inappropriate treatments in intensive care units. Am J Respir 652. Davidson JE (2013) Family presence on rounds in neonatal, pediatric,
Crit Care Med 191(11):13181330 and adult intensive care units. Ann Am Thorac Soc. 10(2):152156
646. Kon AA, Davidson JE, Morrison W, Danis M, White DB (2016) Shared 653. Flanders SA, Strasen JH (2014) Review of evidence about family pres
decision making in ICUs: an American College of Critical Care Medicine ence during resuscitation. Crit Care Nurs Clin North Am. 26(4):533550
and American Thoracic Society policy statement. Crit Care Med 654. Oczkowski SJ, Mazzetti I, Cupido C, Fox-Robichaud AE (2015) The offer
44(1):188201 ing of family presence during resuscitation: a systematic review and
647. Aslakson R, Cheng J, Vollenweider D, Galusca D, Smith TJ, Pronovost meta-analysis. J Intensive Care. 3:41
PJ (2014) Evidence-based palliative care in the intensive care unit: a 655. Oczkowski SJ, Mazzetti I, Cupido C, Fox-Robichaud AE (2015) Canadian
systematic review of interventions. J Palliat Med. 17(2):219235 Critical Care Society: family presence during resuscitation: A Canadian
648. Schulz V, Novick RJ (2013) The distinct role of palliative care in the surgi Critical Care Society position paper. Can Respir J 22(4):201205
cal intensive care unit. Semin Cardiothorac Vasc Anesth 17(4):240248
649. Khandelwal N, Kross EK, Engelberg RA, Coe NB, Long AC, Curtis JR
(2015) Estimating the effect of palliative care interventions and advance
care planning on ICU utilization: a systematic review. Crit Care Med
43(5):11021111
650. DeCato TW, Engelberg RA, Downey L etal (2013) Hospital variation and
temporal trends in palliative and end-of-life care in the ICU. Crit Care
Med 41(6):14051411