REVIEW

Rational Fluid Resuscitation in Sepsis for the

Hospitalist: A Narrative Review
Adam Timothy Ladzinski, DO; Guramrinder Singh Thind, MD;
and Matthew T. Siuba, DO

Abstract

Administration of fluid is a cornerstone of supportive care for sepsis. Current guidelines suggest a
protocolized approach to fluid resuscitation in sepsis despite a lack of strong physiological or clinical
evidence to support it. Both initial and ongoing fluid resuscitation requires careful consideration, as
fluid overload has been shown to be associated with increased risk for mortality. Initial fluid resus-
citation should favor balanced crystalloids over isotonic saline, as the former is associated with
decreased risk of renal dysfunction. Traditionally selected resuscitation targets, such as lactate
elevation, are fraught with limitations. For developing or established septic shock, a focused hemo-
dynamic assessment is needed to determine if fluid is likely to be beneficial. When initial fluid therapy
is unable to achieve the blood pressure goal, initiation of early vasopressors and admission to intensive
care should be favored over repetitive administration of fluid.
ª 2021 Mayo Foundation for Medical Education and Research n Mayo Clin Proc. 2021;nn(n):1-10

nuances of IV fluid therapy in sepsis for From the Department of

S
epsis is characterized by a maladaptive
Internal Medicine, Adoles-
systemic inflammatory response to the acute care internist. cent and Internal Medi-
infection that results in organ cine, Western Michigan
University, Homer Stryker
dysfunction. In 2017, an estimated 48.9 BASIC RATIONALE OF FLUID THERAPY IN M.D. School of Medicine,
million cases occurred worldwide.1 From SEPSIS Kalamazoo, MI (A.T.L.);
the original early goal-directed therapy trial The primary objective of fluid therapy is to and the Department of
Critical Care Medicine,
to the Surviving Sepsis Campaign, fluid ther- increase ventricular preload. If the heart is Respiratory Institute,
apy has played a central role in the treatment on the steep part of the Frank-Starling rela- Cleveland Clinic, Cleve-
land, OH (G.S.T., M.T.S.).
of sepsis. Although current guidelines tionship, an increase in preload is expected
recommend a standardized prescription for to cause a rise in cardiac output (Figure 1).
fluid therapy, this recommendation is not In sepsis, ventricular preload can be reduced
supported by strong clinical or physiological by 2 distinct mechanisms: intravascular vol-
data. As such, it is critical to understand the ume depletion and venodilation. Both factors
rationale behind the use of fluids in sepsis. result in a fall in venous return and preload.
The general awareness of sepsis has Intravascular depletion may be caused by
grown considerably in both professional actual fluid losses such as diarrhea in gastro-
and lay circles. Subsequently, both public intestinal infection, insensible loss from
and private insurances have developed tachypnea, and anorexia with poor oral
sepsis-focused metrics that emphasize a intake before presentation. Furthermore,
standardized approach to fluid management. some intravascular volume may be trans-
Hospitalists find themselves at the forefront ferred to the interstitial compartment
of managing sepsis, and, as such, it is imper- because of capillary leak. Cytokine-
ative to have a thorough grasp on the ratio- mediated damage to the vascular endothe-
nale for intravenous (IV) administration of lium leads to fluid extravasation from the
fluids. Choosing which IV fluid to admin- intravascular to extravascular compartment.2
ister, how much, and when are critical deci- The second cause of reduced ventricular pre-
sion points. This review discusses the load in sepsis is venodilation that increases

Mayo Clin Proc. n XXX 2021;nn(n):1-10 n https://doi.org/10.1016/j.mayocp.2021.05.020 1

www.mayoclinicproceedings.org n ª 2021 Mayo Foundation for Medical Education and Research
 MAYO CLINIC PROCEEDINGS

in part, to the large mortality benefit shown

ARTICLE HIGHLIGHTS in this single-center trial, EGDT was
endorsed in the initial surviving sepsis
d Fluid resuscitation in sepsis requires careful consideration, as
guidelines.4
fluid overload is associated with increased risk for mortality.
Subsequent multicenter studies were un-
d Balanced crystalloids are associated with decreased risk of renal able to reproduce the benefits seen in the
dysfunction and should be favored over isotonic saline. Rivers trial. For instance, the 3-arm
d Traditional resuscitation targets, such as serum lactate, are Protocol-Based Care for Early Septic Shock
fraught with limitations. (ProCESS) trial demonstrated no mortality
d For developing or established septic shock, a focused hemody- benefit to protocol-based therapy (EGDT or
not) in comparison with a usual-care arm.5
namic assessment is needed to determine if fluid is likely to be
The Prospective Multicenter Imaging Study
beneficial. When targeted fluid therapy is unable to reverse for Evaluation of Chest Pain (ProMISe) trial
hypotension or shock, initiation of early vasopressors and similarly showed no mortality benefit to
admission to intensive care should be favored over repetitive EGDT compared with usual care.6 The Aus-
administration of fluid. tralasian Resuscitation in Sepsis Evaluation
(ARISE) trial also compared EGDT with
usual care, with patients in the intervention
venous capacitance, thereby leading to arm ultimately receiving more fluids, blood
sequestration of blood in the venous transfusions, and dobutamine. Overall,
compartment. The relative contribution of EGDT did not decrease 90-day mortality
these mechanisms may vary among patients. compared with usual care.7 Subsequently,
Objectively, patients with true intravascular the 2016 surviving sepsis guidelines
depletion would benefit most with fluid ther- removed the recommendation for EGDT,
apy, whereas vasopressor therapy can help instead making a strong recommendation
reverse venodilation. These distinctions are for the use of 30-mL/kg fluid bolus, based
very difficult to make clinically, and there on low quality of evidence.8 It is important
is often significant overlap in individual to note, however, that participants in the 3
patients. trials described here all received at least 2 li-
ters of fluid before enrollment. Most had also
INITIAL RESUSCITATION already received antibiotics (Table).
Initial administration of fluids in sepsis has A recent randomized control trial, the
been studied in a number of clinical trials. simplified severe sepsis protocol 2 (SSSP-
The hallmark study using a protocolized 2), studied the effects of a resuscitation pro-
treatment of sepsis and septic shock was tocol in patients with sepsis and hypotension
the Rivers trial of early goal-directed therapy in Zambia.9 Notably, this population was
(EGDT). The protocol’s targets included he- predominantly HIV positive, which may
moglobin of 10 mg/dL, central venous pres- limit external validity of the findings. In
sure (CVP) of at least 8 to 12 mm Hg, and the intervention group, an initial 2-liter
central venous oxygen saturation of at least bolus of isotonic crystalloid was followed
70%.3 On average, patients in the interven- with an additional 2 liters over the next 4
tion arm received 13.4 liters of IV fluids.3 hours. Patients in the intervention arm
The practice of fluid loading until a certain received a median of 1.5 liters additional cu-
CVP goal is achieved lacks sound physiolog- mulative fluid compared with the usual-care
ical rationale. A CVP of 8 to 12 is likely path- arm. The intervention-arm mortality rate
ologically high. In the absence of cardiac was 15% higher than the usual-care arm, in
dysfunction, this is most likely reflective of spite of the fact that lactate clearance was
intravascular volume overload. As this was better in this group.
the first major effort to protocolize sepsis There are several pathophysiological rea-
treatment, the Rivers trial introduced a new sons why fluid overload may lead to worse
era of time-sensitive bundled care. Owing, outcomes. Broadly, these can be grouped
n n
2 Mayo Clin Proc. XXX 2021;nn(n):1-10 https://doi.org/10.1016/j.mayocp.2021.05.020
www.mayoclinicproceedings.org
RATIONAL FLUID RESUSCITATION IN SEPSIS

into hemodynamic consequences of intravas-

cular congestion and adverse effects of tissue
edema. The effects of pulmonary congestion
are more obvious, especially in patients with
concomitant acute respiratory distress syn-
drome. Manifestations of systemic conges-
tion may be more insidious but are crucial
to appreciate. Renal congestion reduces renal

Stroke volume
perfusion pressure and may cause acute kid-
ney injury; high CVP values have been asso-
ciated with increased risk of kidney injury as
well as mortality.16,17 Intestinal edema may
lead to increased intraabdominal pressure
and further reduction in renal perfusion.18
The various adverse consequences of fluid
overload underscore the potential benefit of
an individualized approach.
A B C D
Preload
FLUID SELECTION
The use of colloids is not supported by FIGURE 1. Example Frank-Starling relationships demonstrating the impor-
strong evidence and is associated with a tance of fluid responsiveness assessment. On the upper cardiac function
higher cost, so the discussion will focus on curve (green solid line), additional preload will translate into higher stroke
crystalloid selection. The Isotonic Solutions volume if the patient moves from point A to point B, and to a lesser extent
and Major Adverse Renal Events (SMART) when moving from C to D. The patient with the lower cardiac function
trial was a cluster-randomized trial in 5 curve (dashed brown line) will have a modest increase in stroke volume
intensive care units that compared normal moving from point A to point B but not when moving from C to D.
Additional administration of fluid past that point is likely to lead to organ
saline with balanced crystalloids (lactated
congestion and adverse outcomes.
Ringers or Plasma-Lyte) in a heterogeneous
patient population. This trial demonstrated
a reduction in 30-day major adverse kidney
crystalloids should be favored except in
events (a composite of death, new renal-
rare circumstances, such as hyponatremia
replacement therapy, or persistent renal
requiring correction, neurologically injured
dysfunction) with balanced crystalloids.10 A
patients, or concurrent metabolic alkalosis.
secondary analysis of patients enrolled in
SMART with sepsis in medical intensive
care units affirmed this finding. In addition, RESUSCITATION TARGETS IN SEPSIS AND
patients receiving balanced crystalloids had THE ROLE OF FLUID THERAPY
lower risk of death (adjusted odds ratio While discussing goals of resuscitation in
0.74; 95% confidence interval, 0.59 to sepsis, it is helpful to cognitively decouple
0.93), as well as an increase in days free of shock and hypotension. Shock can be
renal replacement therapy and described as a state of global impairment in
vasopressors.11 oxygen delivery or use. Hypotension, on
The term normal saline is certainly a the other hand, is a hemodynamic problem
misnomer. Aggressive resuscitation with in which systemic blood pressure drops
0.9% sodium chloride may result in hyper- below critical limit of autoregulatory poten-
chloremic acidosis. Hyperchloremia has tial of vital organs (eg, heart, brain, kidneys).
been shown to cause renal vasoconstriction It is possible to have shock with a normal (or
in preclinical models19 and may be the high) blood pressure such as normotensive
mechanism behind increased incidence of cardiogenic shock. Also, anesthesia-induced
acute kidney injury with normal saline. hypotension is commonly treated with vaso-
Based on the available evidence, balanced pressors, but these patients typically do not
Mayo Clin Proc. n XXX 2021;nn(n):1-10 n https://doi.org/10.1016/j.mayocp.2021.05.020 3
www.mayoclinicproceedings.org
 4

TABLE. Summary of Key Trials of Fluid Therapy and Resuscitation Targets in Patients With Sepsis and Septic Shock
Median fluid Antibiotic adminis-
Outcome administered tration before
Study Population Intervention Comparison (primary) before enrollment enrollment Notes
5
ProCESS 1341 adults with Protocolized Usual care Protocolized care Approximately Approximately Higher amount of
septic shock (3 EGDT or did not improve 2100 to 2200 76% of patients fluids in protocol
arms) in the protocolized, 60-day mortality mL in each arm in each arm arms; higher rate
United States non-EGDT of vasoactive
medications in
intervention
arms
ProMISe6 1260 adults with EGDT Usual care EGDT did not Approximately 100% of patients in Higher rate of
septic shock in improve 90-day 2000 mL in each each arm administration of
the United mortality arm vasoactive
Kingdom medications and
transfusions in
the intervention
arm
ARISE7 1600 adults with EGDT Usual care EGDT did not Approximately Median time to first Higher rate of
Mayo Clin Proc.

septic shock in improve 90-day 2500 mL in each antibiotic administration of

Australia and mortality arm administered did vasoactive
New Zealand not differ medications in
between arms the intervention
(approximately arm
n
XXX 2021;nn(n):1-10

70 minutes)
SSSP-29 212 adults with Early fluid Usual care Absolute risk Not clearly Median time to first Higher rate of
sepsis and resuscitation increase 15.1% reported antibiotic administration of
hypotension in protocol (2000 (95% CI, 2.0% to Median time to administered did vasoactive
Zambia mL fluid initially, 28.3%) of enrollment not differ medications in
repeated hospital approximately between arms the intervention
n

over the next 4 mortality for the 70 minutes in (approximately arm

https://doi.org/10.1016/j.mayocp.2021.05.020

hours) intervention arm each arm 1.5 to 2 hours) Approximately

MAYO CLINIC PROCEEDINGS

one-half of
control patients
www.mayoclinicproceedings.org

received no fluid
89.5% of patients
enrolled were
HIV positive
SMART10 15,802 unselected Balanced crystalloid 0.9% saline Reduced odds of Not applicable Secondary
critically ill adults major adverse analysis11 of
Continued on next page
 TABLE. Continued
www.mayoclinicproceedings.org
Mayo Clin Proc. n XXX 2021;nn(n):1-10

RATIONAL FLUID RESUSCITATION IN SEPSIS

Median fluid Antibiotic adminis-
Outcome administered tration before
Study Population Intervention Comparison (primary) before enrollment enrollment Notes
kidney events in Approximately 800 1641 patients in
intervention arm mL in each arm medical ICU
(OR, 0.91; 95% (mean) with sepsis
CI 0.84 to 0.99), demonstrated
lower mortality,
vasopressor
days, renal
replacement
n
https://doi.org/10.1016/j.mayocp.2021.05.020

therapy days in
the balanced
arm
CENSER12 310 adults with Early Usual care 27.7% absolute Approximately 800 Not clearly Fewer episodes of
sepsis and norepinephrine increase in rate mL in each arm reported arrhythmia and
hypotension in of achieving pulmonary
Thailand target blood edema in the
pressure at 6 intervention arm
hours in the
intervention arm
ANDROMEDA- 424 adults with Capillary refill time Lactate-guided 8.5% absolute Approximately 25 Median time to first Bayesian re-
SHOCK13 septic shock in guided resuscitation reduction in 28- to 30 mL/kg in antibiotic analysis14
South America resuscitation day mortality each arm administered did demonstrated
(95% CI, not differ >90% chance of
1.2% to 18.2%) between arms 28-day mortality
in intervention (approximately benefit of
arm, did not 1.5 to 2 hours) intervention arm
reach statistical for all priors
significance
FRESH15 124 adults with Dynamic fluid Usual care 1.37 liters lower Approximately Not clearly Fewer patients in
septic shock in responsiveness fluid balance in 2100 to 2400 reported the intervention
the United assessment intervention arm mL in each arm arm required
States and at 72 hours (mean, renal
United Kingdom calculated from replacement
supplement) therapy and
mechanical
ventilation
CI, confidence interval; EGDT, early goal directed therapy; ICU, intensive care unit; kg, kilogram; OR, odds ratio.
5
 MAYO CLINIC PROCEEDINGS

meet the criteria for shock. It is widely receptors are stimulated by endogenous
appreciated that patients with sepsis have epinephrine, leading to cyclic-AMPemedi-
reduced arterial tone, which may cause hy- ated generation of pyruvate, leading to
potension. Sepsis can also cause shock by excessive production of lactate. Therefore,
diverse mechanisms. Overall, the goal in sepsis-associated lactic acidosis has been
sepsis resuscitation should be to correct proposed to reflect severity of illness. It is
both hypotension and tissue hypoxia. likely that both type A and type B lactic
Clinically, the most common trigger for acidosis play a role in sepsis.
fluid therapy is hypotension. As highlighted The ANDROMEDA-SHOCK trial was a
earlier, fluid therapy in a preload-responsive multicenter study comparing a resuscitation
heart would result in an increase in cardiac strategy guided by capillary refill time to a
output. Although this increase in cardiac lactate-based strategy, with the primary
output would cause some increase in blood outcome of all-cause 28-day mortality.13
pressure, this effect would be blunted in The study was unable to detect a statistically
the presence of reduced arterial tone.20 significant difference in outcomes between
This, of course, is a common scenario in the 2 strategies; however, a post hoc
sepsis. Anecdotally, it is not uncommon to Bayesian analysis of the trial suggested lower
see multiple fluid boluses given in a futile mortality and faster resolution of organ
attempt to treat hypotension in patients dysfunction when targeting capillary refill
with reduced arterial tone. To that effect, if time.14 Importantly, yet another post hoc
the initial augmentation of preload does analysis of the same study demonstrated an
not resolve hypotension, a prompt consider- association between lactate-based resuscita-
ation of vasopressor therapy should be made. tion and higher mortality if capillary refill
Timely initiation of vasopressor therapy has time was normal at enrollment.23 This
been associated with improved outcomes. finding calls into question the current prac-
In the Early Use of Norepinephrine in Septic tice of attempting to normalize lactate in pa-
Shock Resuscitation (CENSER) trial, early tients with sepsis and septic shock, although
administration of norepinephrine showed a confirmatory studies are needed. If capillary
higher rate of resolution of shock at 6 hours refill time is chosen as a resuscitation target,
and lower incidence of pulmonary edema it should be performed in a systematic
compared with standard treatment, although fashion as in the original trial. First, apply
no statistically detectable difference in mor- firm pressure to the ventral surface of the in-
tality was seen.12 This was also supported dex finger with a glass microscope slide until
by a prospective observational study in the skin blanches, then maintain pressure for
which norepinephrine was shown to in- 10 seconds. Finally, release and measure the
crease cardiac preload and cardiac output time for color to return to normal. Abnormal
in septic shock when administered early.21 capillary refill time is defined as longer than
Although blood pressure is easily 3 seconds.
measured, the available markers of global Apart from hypotension, the second
perfusion are imprecise. The most trigger for fluid therapy is evidence of shock,
commonly used perfusion marker is serum as detected with the aforementioned perfu-
lactate level. The inherent assumption in sion markers. Intuitively, the mechanism
this practice is that a raised lactate level re- by which fluids may improve perfusion is
sults from increased anaerobic glycolysis by augmenting cardiac output given a
resulting from tissue hypoxia: type A lactic preload-responsive heart. A retrospective
acidosis. All causes of increase in serum analysis of patients with sepsis and septic
lactate level independent of the aforemen- shock demonstrated an association of
tioned mechanism are considered type B lac- improved mortality if fluid was given in the
tic acidosis. There is significant evidence that first 3 hours, rather than later.24 Further
several pathways in sepsis may result in type analysis of the ANDROMEDA-SHOCK trial
B lactic acidosis;22 for example, beta-2 provides important insights in this regard.
n n
6 Mayo Clin Proc. XXX 2021;nn(n):1-10 https://doi.org/10.1016/j.mayocp.2021.05.020
www.mayoclinicproceedings.org
RATIONAL FLUID RESUSCITATION IN SEPSIS

Inpatient with sepsis AND

hypotension and/or signs of poor perfusion
(CRT, urine output, encephalopathy,
lactate elevation)

Initial resuscitation (e.g.

30 ml/kg fluid) complete?

Yes No

Administer 30 ml/kg
Strong history of balanced crystalloid
hypovolemia? unless overt symptomatic
Yes No hypervolemia

Resolution of poor
Consider 1-2 additional 500 No perfusion or Yes Exit algorithm
ml boluses targeted to CRT hypotension?
and/or urine output*

Assess fluid responsiveness

Resolution of poor (e.g. increase in CO or SV with
perfusion or No
rapid bolus or PLR)
hypotension? IVC exam not recommended

Yes

Consider 1-2 additional 500

Exit algorithm Fluid Responsive? Yes ml boluses targeted to CRT

Yes
No or unable to determine

Consider prompt ICU Resolution of poor

transfer and initiation No perfusion or
of vasopressors hypotension?

Yes

Exit algorithm

FIGURE 2. Proposed fluid management approach for the inpatient with sepsis and hypotension or signs of poor perfusion. *Urine
output may be an unreliable resuscitation target in the setting of acute kidney injury; use with caution. CRT, capillary refill time, CO,
cardiac output (measured by echocardiography or noninvasive monitor); IVC, inferior vena cava; PLR, passive leg raise test; SV, stroke
volume (monitoring in the same fashion as cardiac output).

Mayo Clin Proc. n XXX 2021;nn(n):1-10 n https://doi.org/10.1016/j.mayocp.2021.05.020 7

www.mayoclinicproceedings.org
 MAYO CLINIC PROCEEDINGS

Interestingly, only 242 of 424 patients change in cardiac output. A positive

(57.1%) in the study were fluid responsive response is usually considered as an increase
at baseline. After 4 hours, only 48 of 424 pa- in cardiac output by 10%. In essence, this is
tients (11.3%) were still fluid responsive.13 a test of the heart’s position on the Frank-
By extension, it would be reasonable to as- Starling curve (Figure 1). A change in blood
sume that most inpatients would not be fluid pressure or heart rate should not be used as a
responsive after initial resuscitation in the surrogate measure, as it does not provide
emergency department. In addition, if fluid sufficient accuracy.30 This highlights the ma-
therapy fails to increase cardiac output (ie, jor challenge in assessing fluid responsive-
in nonresponders), the resulting hemodilu- ness: performing reliable dynamic cardiac
tion may, in fact, reduce oxygen delivery.25 output monitoring. On the medical floor,
In the face of ongoing hypotension or shock, point-of-care ultrasonography may be the
early transfer to an intensive care unit for most effective tool available to meet this
timely vasopressor therapy should be seri- end. Noninvasive cardiac output monitors
ously considered. are also an option, although not routinely
available outside of intensive care units or
RATIONALE FOR DETERMINING FLUID operating rooms.
RESPONSIVENESS Bedside echocardiography can be per-
Fluid responsiveness is generally defined as formed for real-time estimation of cardiac
an increase in cardiac output after receiving output. The technique involves measuring
intravascular volume expansion.26 Deter- dimensions of the left-ventricular outflow
mining this in a patient can be challenging, tract and using pulsed-wave Doppler to sam-
even in the resource-abundant intensive ple a velocity-time integral (VTI) at the left-
care unit. Testing for fluid responsiveness ventricular outflow tract during systole.31 A
involves 2 sequential steps: augmenting combination of these 2 data points would
ventricular preload and measuring the provide the clinician with an estimated
change in stroke volume or cardiac output. stroke volume. This can then be multiplied
Ventricular preload can be augmented by by the heart rate to obtain cardiac output.
either a 100-mL to 500-mL rapid fluid The major disadvantage of this method is
bolus27 or by performing passive leg raise the advanced training and experience
(PLR); PLR is performed in a semirecum- needed to perform and interpret it appropri-
bent position,28 with the patient’s legs ately. Another potentially more feasible
passively raised to 45 degrees above the pa- method involves using carotid Doppler to
tient for 30 to 90 seconds. This maneuver trend VTI and assessing the response to pre-
has been shown to mobilize approximately load augmentation.32 Reportedly, approxi-
300 cc of venous blood that augments car- mately 15% to 26 % for the proportion of
diac preload. The obvious advantage is cardiac output (CO) is dedicated to cerebral
that no external fluid infusion is necessary perfusion. It thus follows that trends in ca-
for PLR. The stroke volume or cardiac rotid flows are expected to correlate with
output should be measured before the CO trends and can therefore be used as a
PLR, during the maneuver, and after surrogate.32,33 This method is appealing, as
returning the patient to the upright posi- the necessary images are much easier to ac-
tion to ensure that it has returned to its quire, although, overall, it has a smaller evi-
original value. Notably, the volume of dence base.
blood mobilized may be lower in cases of Perhaps the most commonly used mea-
severe hypovolemia and intra-abdominal sure of fluid responsiveness is also the
hypertension,28,29 resulting in false most inaccurate: inferior vena cava (IVC)
negatives. collapsibility. In spontaneously breathing pa-
Once ventricular preload has been tients, baseline IVC size and the degree of
augmented, the next step is to measure the collapsibility with inspiration have been

n n
8 Mayo Clin Proc. XXX 2021;nn(n):1-10 https://doi.org/10.1016/j.mayocp.2021.05.020
www.mayoclinicproceedings.org
RATIONAL FLUID RESUSCITATION IN SEPSIS

shown to be useful predominantly for esti- Abbreviations and Acronyms: CVP = central venous
mation of central venous pressure.34 Howev- pressure; EGDT = early goal-directed therapy; IV = intra-
venous; IVC = inferior vena cava; PLR = passive leg raise
er, IVC collapsibility in spontaneously VTI = velocity-time integral
breathing patients has not found to be useful
in predicting fluid responsiveness.35 Lack of
standardization of inspiratory effort is likely Potential Competing Interests: The authors report no
competing interests.
one of the major reasons for this. On the
other hand, CVP itself has not shown to pro- Correspondence: Address to Matthew T. Siuba, DO,
vide any useful information regarding fluid Department of Critical Care Medicine, Respiratory Institute,
Cleveland Clinic, 9500 Euclid Ave, L2-330, Cleveland, OH
responsiveness and should not be used for 44195 (siubam@ccf.org; Twitter: @msiuba).
this purpose.26
Determining fluid responsiveness is a ORCID
Matthew T. Siuba: https://orcid.org/0000-0002-4321-
challenge, and yet it is important to define
4944
if additional administration of fluid is
considered. Patients in the Fluid Response
Evaluation in Sepsis Hypotension and Shock REFERENCES
(FRESH) trial randomized to fluid respon- 1. Rudd KE, Johnson SC, Agesa KM, et al. Global, regional, and na-
siveness assessment (by noninvasive cardiac tional sepsis incidence and mortality, 1990e2017: analysis for
the Global Burden of Disease Study. Lancet. 2020;
output monitoring) demonstrated lower net 395(10219):200-211.
fluid balance and less need for renal replace- 2. Pober JS, Sessa WC. Inflammation and the blood microvascular
system. Cold Spring Harb Perspect Biol. 2015;7(1):a016345.
ment therapy or invasive mechanical ventila- 3. Rivers E, Nguyen B, Havstad S, et al. Early goal-directed therapy
tion. This difference was demonstrated in the treatment of severe sepsis and septic shock. N Engl J Med.
despite the fact that the administration of 2001;345(19):1368-1377.
4. Dellinger R, Carlet J, Masur H, et al. Surviving Sepsis Campaign
fluid between groups only differed by less guidelines for management of severe sepsis and septic shock.
than 2 liters at 72 hours.15 Crit Care Med. 2004;32(3):858-873.
Overall, fluid should be given in a simpli- 5. Yearly D, Kellum J, Huang D, et al. A randomized trial of
protocol-based care for early septic shock. N Engl J Med.
fied, physiologically goal-directed fashion as 2014;370(18):1683-1693.
described here. Accordingly, there is no clear 6. Mouncey PR, Osborn TM, Power GS, et al. Trial of early, goal-
directed resuscitation for septic shock. N Engl J Med. 2015;
rationale for maintenance IV fluids in hospi- 372(14):1301-1311.
talized patients if the intent is purely to treat 7. ARISE Investigators; ANZICS Clinical Trials Group, Peake SL,
sepsis or septic shock; this will likely lead to et al. Goal-directed resuscitation for patients with early septic
shock. N Engl J Med. 2014;371(16):1496-1506.
higher cumulative fluid balance without 8. Rhodes A, Evans LE, Alhazzani W, et al. Surviving Sepsis
clear hemodynamic benefits. A comprehen- Campaign: International Guidelines for Management of Sepsis
sive approach to fluid management as out- and Septic Shock: 2016. Crit Care Med. 2017;45(3):486-552.
9. Andrews B, Semler MW, Muchemwa L, et al. Effect of an early
lined in this review is detailed in Figure 2. resuscitation protocol on in-hospital mortality among adults
with sepsis and hypotension: a randomized clinical trial. JAMA.
2017;318(13):1233.
CONCLUSION 10. Semler MW, Self WH, Wanderer JP, et al. Balanced crystalloids
Fluid management in the patient with sepsis versus saline in critically ill adults. N Engl J Med. 2018;378(9):
829-839.
and signs of malperfusion poses a unique 11. Brown RM, Wang L, Coston TD, et al. Balanced crystalloids
challenge for the hospitalist. Recent trials versus saline in sepsis: a secondary analysis of the SMART clin-
have demonstrated that the natural history ical trial. Am J Respir Crit Care Med. 2019;200(12):1487-1495.
12. Permpikul C, Tongyoo S, Viarasilpa T, Trainarongsakul T,
of fluid responsiveness is short lived, and a Chakorn T, Udompanturak S. Early Use of Norepinephrine in
cautious approach is necessary to avoid the Septic Shock Resuscitation (CENSER): a randomized trial. Am
harms of fluid overload. Furthermore, the J Respir Crit Care Med. 2019;199(9):1097-1105.
13. Hernández G, Ospina-Tascón GA, Damiani LP, et al. Effect of a
multifactorial nature of lactate elevation in resuscitation strategy targeting peripheral perfusion status vs
sepsis makes it a poor perfusion target. If hy- serum lactate levels on 28-day mortality among patients with
septic shock: the ANDROMEDA-SHOCK randomized clinical
potension persists despite hemodynamically trial. JAMA. 2019;321(7):654-664.
targeted administration of fluid, ventricular 14. Zampieri FG, Damiani LP, Bakker J, et al. Effects of a resuscita-
preload is less likely the issue, and early tion strategy targeting peripheral perfusion status vs serum
lactate levels among patients with septic shock: a Bayesian rean-
transfer to intensive care units and initiation alysis of the ANDROMEDA-SHOCK trial. Am J Respir Crit Care
of vasopressors is advised. Med. 2020;201(4):423-429.

Mayo Clin Proc. n XXX 2021;nn(n):1-10 n https://doi.org/10.1016/j.mayocp.2021.05.020 9

www.mayoclinicproceedings.org
 MAYO CLINIC PROCEEDINGS

15. Douglas IS, Alapat PM, Corl KA, et al. Fluid response evaluation resuscitation is associated with reduced mortality: a retrospec-
in sepsis hypotension and shock: a randomized clinical trial. tive cohort study. Chest. 2014;146(4):908-915.
Chest. 2020;158(4):1431-1445. 25. Perel A. Iatrogenic hemodilution: a possible cause for avoidable
16. Boyd JH, Forbes J, Nakada T, Walley KR, Russell JA. Fluid resus- blood transfusions? Crit Care. 2017;21(1):291.
citation in septic shock: a positive fluid balance and elevated 26. Marik PE, Cavallazzi R. Does the central venous pressure pre-
central venous pressure are associated with increased mortality. dict fluid responsiveness? An updated meta-analysis and a
Crit Care Med. 2011;39(2):259-265. plea for some common sense. Crit Care Med. 2013;41(7):
17. Chen X, Wang X, Honore PM, Spapen HD, Liu D. Renal failure 1774-1781.
in critically ill patients, beware of applying (central venous) pres- 27. Biais M, de Courson H, Lanchon R, et al. Mini-fluid challenge of
sure on the kidney. Ann Intensive Care. 2018;8(1):91. 100 ml of crystalloid predicts fluid responsiveness in the oper-
18. Malbrain MLNG, Van Regenmortel N, Saugel B, et al. Principles ating room. Anesthesiology. 2017;127(3):450-456.
of fluid management and stewardship in septic shock: it is time 28. Jabot J, Teboul J-L, Richard C, Monnet X. Passive leg raising for
to consider the four D’s and the four phases of fluid therapy. predicting fluid responsiveness: importance of the postural
Ann Intensive Care. 2018;8(1):66. change. Intensive Care Med. 2008;35(1):85.
19. Chowdhury AH, Cox EF, Francis ST, Lobo DN. A randomized, 29. Monnet X, Teboul J-L. Passive leg raising: five rules, not a drop
controlled, double-blind crossover study on the effects of 2-L of fluid!. Crit Care. 2015;19(1):18.
infusions of 0.9% saline and plasma-lyte(R) 148 on renal blood 30. Ait-Hamou Z, Teboul J-L, Anguel N, Monnet X. How to detect
flow velocity and renal cortical tissue perfusion in healthy vol- a positive response to a fluid bolus when cardiac output is not
unteers. Ann Surg. 2013;258(6):1118. measured? Ann Intensive Care. 2019;9(1):138.
20. Monge García MI, Gil Cano A, Gracia Romero M. Dynamic arte- 31. Miller A, Mandeville J. Predicting and measuring fluid responsive-
rial elastance to predict arterial pressure response to volume ness with echocardiography. Echo Res Pract. 2016;3(2):G1-G12.
loading in preload-dependent patients. Crit Care. 2011;15(1):R15. 32. Gassner M, Killu K, Bauman Z, Coba V, Rosso K, Blyden D.
21. Hamzaoui O, Georger J-F, Monnet X, et al. Early administration Feasibility of common carotid artery point of care ultrasound
of norepinephrine increases cardiac preload and cardiac output in cardiac output measurements compared to invasive
in septic patients with life-threatening hypotension. Crit Care. methods. J Ultrasound. 2015;18(2):127-133.
2010;14(4):R142. 33. Adedipe AA, Fly DL, Schwitz SD, et al. Carotid Doppler blood
22. Garcia-Alvarez M, Marik P, Bellomo R. Sepsis-associated hyper- flow measurement during cardiopulmonary resuscitation is
lactatemia. Crit Care. 2014;18(5):503. feasible: a first in man study. Resuscitation. 2015;96:121-125.
23. Kattan E, Hernández G, Ospina-Tascón G, et al. A lactate-tar- 34. Vecchis RD, Baldi C, Giandomenico G, Maio MD, Giasi A,
geted resuscitation strategy may be associated with higher mor- Cioppa C. Estimating right atrial pressure using ultrasounds:
tality in patients with septic shock and normal capillary refill an old issue revisited with new methods. J Clin Med Res.
time: a post hoc analysis of the ANDROMEDA-SHOCK study. 2016;8(8):569-574.
Ann Intensive Care. 2020;10(1):114. 35. Kory P. COUNTERPOINT: Should acute fluid resuscitation be
24. Lee SJ, Ramar K, Park JG, Gajic O, Li G, Kashyap R. Increased guided primarily by inferior vena cava ultrasound for patients in
fluid administration in the first three hours of sepsis shock? No. Chest. 2017;151(3):533-536.

n n
10 Mayo Clin Proc. XXX 2021;nn(n):1-10 https://doi.org/10.1016/j.mayocp.2021.05.020
www.mayoclinicproceedings.org