Original Paper

Cerebrovasc Dis 2015;40:10–17 Received: January 14, 2015

Accepted: April 21, 2015
DOI: 10.1159/000430871
Published online: May 27, 2015

Outcome Predictors of Acute

Stroke Patients in Need of Intensive
Care Treatment
Angelika Alonso a Anne D. Ebert a Rolf Kern a Simone Rapp b
       

Michael G. Hennerici a Marc Fatar a    

a
Department of Neurology, and b First Department of Medicine, Universitätsmedizin Mannheim,
   

University of Heidelberg, Mannheim, Germany

Key Words terventional procedures requiring mechanical ventilation

Stroke · Intracerebral hemorrhage · Intensive care unit · (11%). In all, 231/347 patients (66.6%) were mechanically
Mechanical ventilation · Prognosis · Mortality ventilated (mean 84 h). In-hospital mortality (143/347; 41.2%)
was associated with old age, poor NIHSS score at admission,
intracerebral hemorrhage and mechanical ventilation (p <
Abstract 0.001 in all). Further, admission to ICU because of stroke-re-
Background: The prognosis of stroke patients admitted to lated impairment of consciousness increased in-hospital
intensive care units (ICU) is commonly regarded to be poor. mortality (p < 0.001). Similarly, poor outcome after rehabilita-
However, only limited data regarding outcome predictors tion was associated with old age (p = 0.029) and mechanical
are available. Patients and Methods: Out of 4,958 consecu- ventilation (p < 0.001). In patients ≥80 years with either intra-
tive patients admitted to our stroke unit with the diagnosis cerebral hemorrhage or need of mechanical ventilation, out-
of acute stroke, after analysis we identified 347 patients (164 come was unfavorable in nearly any case. However, the over-
male) in need of ICU management. In-hospital and post-re- all post-rehabilitation outcome did not differ between pa-
habilitation mortality as well as functional outcome at dis- tients with intracerebral hemorrhage and ischemic stroke
charge and after rehabilitation were analyzed. Results: Isch- (p = 0.275). Conclusion: The stroke population in our study
emic stroke was diagnosed in 252 patients (72.6%) and intra- was associated with an increased early mortality; however,
cerebral hemorrhage occurred in 95 patients (27.4%). The given the same conditions, it was old with a high percentage
mean age in our cohort was considerably high (70.8 years). of patients requiring mechanical ventilation. This did not re-
One hundred patients were comatose at admission. The me- sult in increased in-hospital mortality rates compared to
dian NIHSS score at admission in the remaining patients was younger and less severely affected cohorts. Thus, ICU man-
12. Apart from stroke-related disturbances of consciousness agement is a life-saving initiative even among the elderly.
(47.1%), the most common reasons for ICU treatment were However, the functional outcome was poor in older patients,
cardiac (23.4%) and respiratory (12.1%) complications or in- thus limiting the benefits of ICU care in these patients.
© 2015 S. Karger AG, Basel

© 2015 S. Karger AG, Basel Angelika Alonso, MD

1015–9770/15/0402–0010$39.50/0 Department of Neurology, Universitätsklinikum Mannheim
Theodor Kutzer Ufer 1–3
E-Mail karger@karger.com
DE–68167 Mannheim (Germany)
www.karger.com/ced
E-Mail alonso @ neuro.ma.uni-heidelberg.de
 Introduction have been recommended by Kirkman and colleagues,
namely, large MCA infarct volume that predicts a malig-
Stroke remains the major cause of severe disability and nant course, postoperatively following decompressive
the second most common cause of death despite several craniectomy or management of organ support [3]. Nota-
advances in prevention and treatment over the last two bly, a recent study by Cereda and colleagues suggested
decades [1]. Management in a stroke care unit has been that stroke patients admitted to a semi-intensive stroke
proven to be beneficial, with a significant reduction of unit (SI-SU) showed a lower proportion of unfavourable
both death and poor outcome [2]. However, the percent- outcome at three months compared to those admitted to
age of stroke patients being admitted to an intensive care an ICU with a mobile stroke team [13]. The authors sug-
unit (ICU) for monitoring and/or management of post- gest that the continuity of the primary neurology team, a
stroke complications is increasing [3]. The growing num- harmonized neuro-rehabilitation program as well as
ber of ICU-admitted stroke patient parallels the advances comprehensive medical, nursing protocols and treatment
in intensive care management as well as a trend toward algorithms on SI-SU might account for this finding. Pa-
more aggressive and invasive treatments. An increasingly tients with ‘relative’ indications for ICU treatment as de-
aging population with growing incidence of elderly pa- fined by Kirkman and colleagues [3] might therefore
tients suffering a stroke may also account for incremental rather profit from treatment in a stroke unit.
rates of stroke patients treated in ICUs. However, even in patients with indisputable indica-
Despite the high socioeconomic impact, the actual ben- tions for ICU treatment, criteria when ICU treatment
efit of ICU treatment of stroke patients remains a matter does not provide reasonable chances for good functional
of discussion, not in the least due to the scarcity of reliable outcome are still lacking.
data. The reported mortality rates for stroke patients re- Our study now aimed at defining factors in patients
quiring ICU management vary widely, with a short-term with acute hemorrhagic or ischemic stroke in need of ICU
mortality of more than 50% in some studies [4, 5]. Newer management that may allow a prognosis of the short-
data of a multicenter retrospective cohort study, however, term outcome.
report on lower mortality rates both in the short- and long-
term in older ICU patients with acute ischemic stroke [6].
The wide range of mortality rates is explained by several Methods
factors, including ICU admission criteria and percentage This study was approved by the local Institutional Review
of patients with intracerebral hemorrhage in the respective Board (Medizinische Ethikkommission II der Medizinischen
cohorts. Regardless of this, some factors like comatose state Fakultät Mannheim, University of Heidelberg). Patient consent
at admission to ICU or mechanical ventilation have been was not required due to the retrospective nature of the study and
shown to predict a poor prognosis in several studies [4, 7, the lack of patient interaction.
From January 2007 to July 2011, 4,958 patients were admitted to
8]. How to determine suitability for ICU admission in the Department of Neurology with the diagnosis of ischemic stroke
acute stroke patients is a key question that still remains. or intracerebral hemorrhage (ICH). The diagnosis was confirmed in
Most surprisingly, the actual ‘Guidelines for the Early all patients by cranial computed tomography (CT) or magnetic res-
Management of Patients With Acute Ischemic Stroke’, onance tomography (MRT). Of these, 347 required treatment in an
released by the AHA/ASA in 2013, do not answer this intensive care unit (ICU) and were included in the study.
All patients underwent stroke workup according to a standard-
question in detail [9], while the ‘Guidelines for manage- ized protocol including assessment of risk factors (history of hy-
ment of ischaemic stroke and transient ischaemic attack pertension, diabetes, dyslipidemia, smoking, coronary artery dis-
2008’ by the European Stroke Organisation (ESO) do not ease, atrial fibrillation and previous stroke), neurological status
address this issue at all [10, 11]. The ‘Guidelines for ICU (National Institute of Health Stroke Scale Score, NIHSS; Richmond
Admission, Discharge, and Triage’, published by the Agitation Sedation Scale, RASS, if applicable), extra-/transcranial
ultrasound and continuous monitoring (ECG, blood pressure) on
American College of Critical Care Medicine of the Society our Stroke Unit (SU) or ICU.
of Critical Care Medicine as far as 15 years ago [12], have The reasons for ICU admission were classified into neurologi-
been recalled, and a revision is still in preparation. Indis- cal (stroke-related disturbances of consciousness with reduced
putable criteria for ICU management in acute stroke pa- brain stem reflexes), cardiac (therapy refractory tachy- or brady-
tients include decreased level of consciousness, need for cardia, acute cardiac insufficiency, acute myocardial infarction, re-
suscitation), respiratory (aspiration, pneumonia, pulmonary ede-
mechanical ventilation and invasive monitoring and were ma, decompensated chronic obstructive lung disease, pulmonary
taken as a basis for ICU treatment indication in our study. embolism) complications, infections (pulmonary, abdominal,
Recently, possible further indications for ICU admission CNS), interventional procedures requiring mechanical ventilation

Outcome of Stroke Patients on ICU Cerebrovasc Dis 2015;40:10–17 11

DOI: 10.1159/000430871
and other reasons. Multiple assignments were allowed. Indication Table 1. Patient characteristics (n = 347)
for mechanical ventilation was classified as coma, respiratory fail-
ure or procedure-related. The duration time of ventilation was as- Age, years, mean (range) 70.8 (28–95)
sessed in terms of hours. Sex, male/female 164/183
In order to assess the outcome after rehabilitation, the medical Diagnosis, %
records of patients transferred to rehabilitation hospitals were Ischemic stroke 252 (72.6)
evaluated. In-hospital and post-rehabilitation mortality as well as Thrombolysis performed in 97/252 (38.5)
modified Rankin scale (mRS) at discharge and after rehabilitation Intracerebral hemorrhage 95 (27.4)
were analysed as outcome parameters. A mRS score of ≤3 was de- Risk factors, %
fined as favorable functional outcome. Atrial fibrillation 140 (40.4)
Arterial hypertension 290 (83.6)
Statistical Analysis Diabetes 125 (36.0)
In order to analyze possible predictors of mortality, we per- Dyslipidemia 114 (32.9)
formed χ2-test for categorical and t-test for metric variables. The Current smoker 53 (15.3)
Spearman-Rho rank correlation coefficient was used to calculate Coronary artery disease 79 (22.8)
nonparametric correlations between age or duration of rehabilita- Previous stroke 62 (17.9)
tion on the one hand and the outcome after rehabilitation (mRS) Neurological status at admission, %
on the other hand. The impact of mechanical ventilation, stroke Comatose (RASS –5) 100 (28.8)
type and thrombolysis on the outcome after rehabilitation was cal- Non-comatose 247 (71.2)
culated by the Mann-Whitney U test. A p value <0.05 was consid- NIHSS score at admission, mean, IQR 12 (5–18)
ered to indicate statistical significance. Statistical analysis was per- Indication for ICU management, %
formed using the Statistical Package for the Social Sciences (SPSS), Neurological complications 162 (46.7)
version 22.0.0.0 (IBM, USA). Cardiac complications 81 (23.3)
Respiratory complications 42 (12.1)
Intervention requiring mechanical
ventilation 39 (11.2)
Other 23 (6.6)
Results Mechanical ventilation, % 231 (66.6)
Duration, mean (range) 84 h (1 h to 56 days)
Out of 4,958 consecutive patients admitted to our
stroke unit with the diagnosis of acute stroke, 347 patients RASS = Richmond Agitation Sedation Scale; NIHSS = Nation-
al Institute of Health Stroke Scale; ICU = intensive care unit; IQR =
(164 male, 183 female, mean age 70.8, range 28–95 years)
interquartile range.
required ICU admission at any time point during their
index hospitalization. Of these, 174 patients (50.5%) were
initially admitted to ICU. Ischemic stroke was diagnosed
in 252/347 (72.6%); ICH occurred in 95/347 (27.4%). tients (15.3%) were current smokers. Coronary artery dis-
Out of 347 patients, 100 had been intubated and sedated ease was known in 79 patients (22.8%), and 62 patients
in the prehospital setting and were comatose (Richmond (17.9%) had a history of previous stroke.
Agitation Sedation Scale –5 in all) at admission. In the re- The most common reasons for ICU treatment were
maining 247 patients, the median NIHSS score at admis- neurological complications (46.7%), followed by cardiac
sion was 12 (interquartile range 5–18). Of the 252 patients (23.3%) and respiratory (12.1%) complications. An 11.2%
with ischemic stroke, 97 patients (38.5%) underwent of patients were admitted to ICU because of intervention-
thrombolysis: intravenous thrombolysis with rtPA (85 pa- al procedures requiring mechanical ventilation. Patients
tients), mechanical thrombectomy (5 patients) or a combi- numbering 231 (66.6%) were mechanically ventilated
nation of both according to the ‘bridging’ concept (7 pa- with a mean duration of ventilation of 84 h (range 1 h to
tients). The number of patients undergoing thrombolysis 56 days). An overview of the patient characteristics is giv-
was significantly higher in this collective than in ischemic en in table 1.
stroke patients not requiring ICU management over the One hundred and forty three patients (41.2%) died
same time period (23%, p = 0.041). Cardiac embolism due during hospitalization. Of these, 64 patients suffered
to newly detected or known atrial fibrillation was the most from ICH (37 with additional intraventricular hemor-
common stroke etiology and was documented in 140/347 rhage), and 79 patients had ischemic stroke, resulting in
patients (40.4%). Patients numbering 290 (83.6%) had a a significantly higher mortality in patients with ICH (p <
history of arterial hypertension requiring antihypertensive 0.001). In-hospital mortality was further associated with
medication; diabetes was diagnosed in 125 patients (36.0%) old age (p < 0.001) and severity of neurological deficit at
and dyslipidemia in 114 patients (32.9%). Fifty-three pa- admission (p < 0.001). Moreover, admission to ICU be-

12 Cerebrovasc Dis 2015;40:10–17 Alonso/Ebert/Kern/Rapp/Hennerici/Fatar

DOI: 10.1159/000430871
cause of stroke-related impairment of consciousness as treatment of severely affected stroke patients is yet uncer-
well as the need for mechanical ventilation was related to tain. In our cohort of 347 acute stroke patients in need of
in-hospital mortality (p < 0.001 for both). In-hospital ICU management, we could demonstrate that the overall
mortality in mechanically ventilated patients was 57.1%. in-hospital mortality was high, with 41.2% not surviving
Patients with atrial fibrillation (AF) were more likely the acute phase. In-hospital mortality rates reported in
to develop cardiac complications (p < 0.001) and to die the literature vary widely, highly depending on the pa-
from non-neurological reasons (p = 0.044); overall in- tient characteristics.
hospital mortality, however, was not increased in patients A small Australian study (n = 58) including both isch-
with AF (p = 0.225). In-hospital mortality due to non- emic stroke and ICH patients provided a slightly higher
neurological reasons was further associated with the oc- mortality rate of 47%; however, the analysed data dates
currence of cardiac and respiratory complications, infec- back to the 1990s [5]. The Evascan project, a prospective
tion and resuscitation (p < 0.001). In contrast, patients observational study in 132 stroke patients admitted to the
who experienced neurological complications like recur- ICU, also as far back as in 1999, reported a lower mortal-
rent stroke, secondary hemorrhage, brain edema or epi- ity rate of 33% [4]. Newer studies with recruitment of pa-
leptic seizures were more likely to die as a consequence of tients after 12/1999 are scarce; an overview is given in
the neurological disease (p < 0.001). table 2. Interestingly, the reported mortality rates in these
Median mRS after rehabilitation was mRS 4. The out- studies vary substantially depending on the originating
come after rehabilitation was associated with patient age, country, with rather low short-term mortality in Asian
with younger patients achieving lower mRS scores than countries [16, 17] and markedly higher mortality rates up
older patients (p = 0.008). However, patients attaining a to 70% in studies originating from European countries
good functional outcome (mRS ≤3) were not younger [7] and the United States [18]. Two possible factors may
than patients with a less favorable outcome (p = 0.16). account for this finding: first, although the incidence of
Improvement during rehabilitation was associated with ICH is known to be higher in Asians than in Caucasians,
duration of rehabilitation, with longer duration of stay a recent study on racial/ethnic differences in the outcome
predicting better recovery (p = 0.022). Nevertheless, ab- of ICH patients found lower risk-adjusted, in-hospital
solute outcome defined as post-rehabilitation mRS was mortality for Asian than white patients with ICH [19].
independent of the duration of rehabilitation. Thus, me- Second, ICU admission criteria in Asian countries may
chanical ventilation remained a major risk factor for poor vary considerably from those in Europe/United States,
outcome, even after rehabilitation (p < 0.001). Although not at least because of the limited access to tertiary care
in-hospital mortality was significantly associated with centers in many Asian countries [20]. Mortality rates
ICH, the outcome after rehabilitation did not differ be- should therefore be interpreted with caution and consid-
tween patients with ICH and ischemic stroke (p = 0.275). er the patients’ characteristics of the respective cohort.
In patients with ischemic stroke, thrombolysis (per- We analysed several factors as possible outcome pre-
formed vs. not performed) did not have an influence on dictors and could identify the following negative outcome
the post-rehabilitation outcome mRS (p = 0.921) or on markers: age; coma at admission; severe neurological def-
the risk of death (p = 0.421). icit at admission given by the NIHSS score, need for me-
In patients ≥80 years, one additional risk factor for chanical ventilation, and hemorrhagic stroke.
poor outcome resulted in an unfavourable outcome in Age is a widely accepted predictor of poor outcome in
nearly any case: of 65 patients ≥80 years requiring me- stroke patients requiring ICU treatment [21]. Schielke
chanical ventilation, 64 had a post-rehabilitation mRS of and colleagues found a higher probability of death after
≥4, with 47 fatalities. Likewise, 23/27 of the ≥80 year-old 2 months and 2 years post-ischemic stroke in ischemic
patients with ICH stroke did not survive the acute phase, stroke patients >60 years [22]. In mechanically ventilated
the remaining 4 patients were severely disabled. patients suffering from either ischemic or hemorrhagic
stroke, an age of >65 years indicated a poor outcome [23].
Notably, both cut-off values are far below the mean age of
Discussion and Review of the Literature patients in our cohort with 70.8 years. In their large mul-
ticenter retrospective cohort study in 31,301 ischemic
There is plenty of evidence that management of both stroke patients (26% of those in need of ICU admission),
ischemic stroke and ICH patients on specialized stroke Golestanian and coworkers could demonstrate that age
units is beneficial [14, 15]. However, the benefit of ICU was associated with a gradually increasing risk of death,

Outcome of Stroke Patients on ICU Cerebrovasc Dis 2015;40:10–17 13

DOI: 10.1159/000430871
 14
Table 2. Overview on studies addressing the prognosis and outcome of stroke patients admitted to intensive care units

Author Country Age, y, Stroke Thrombolysis, NIHSS MV, % In-hospital Predictors for Predictors
mean subtype, % % admission mortality, % in-hospital for poor
mortality outcome

Alonso Germany 347 70.8 IS 252 (72.6) 97 (38.5) median 231 (66.6) IS 79 (31.3) age, MV, ICH, age, MV
et al., 2015 ICH 95 (27.4) 12 ICH 64 (67.4) NIHSS
Lahiri et al., USA 99,782 67.2 IS 50,871 (51) n.a. n.a. 99,782 (100) IS 46.8 n.a. n.a.
2014 [18] ICH 32,967 (33) ICH 61.0
Moon et al., Korea 498 56 IS 198 (39.8) n.a. mean n.a. 131 (26.3) NIHSS, GCS, n.a.
2015 [25] ICH 300 (60.2) 21.63 APACHE II,
SAPS II
Riachy et al., Lebanon 62 65.8 IS 45 (62.6) 10 (16.1) n.a. 23 (37) 16 (25.8) age, APACHE II, n.a.

DOI: 10.1159/000430871
2008 [21] ICH 17 (27.4) GCS, MV (a.o.)

Cerebrovasc Dis 2015;40:10–17

Jeng et al., Taiwan 850 65.3 IS 508 (59.8) 70 (13.8) median 278 (32.7) 139 (16.3) n.a. age, MV,
2008 [16] ICH 342 (40.2) 17 (3-m-mortality: NIHSS
NIHSS, MV) (a.o.)
Handschu Germany 90 64.3 IS 41 (45.6) n.a. n.a. 90 (100) n.a. n.a. n.a.
et al., 2005 [26] ICH 49 (54.4) (10-d- (10-d-mortality:
mortality: 29 GCS, SAPS I/II)
(32.2))
Golestanian USA 8,185 80.2 IS 8,185 (100) n.a. n.a. 876 (10.7) n.a. n.a. n.a.
et al., 2009 [6] (30-d- (30-d-mortality:
mortality: age, MV)
1,748 (21.4))
Lan et al., Taiwan 231 68 (survivors), IS 231 (100) n.a. n.a. n.a. 34 (14.7) impaired n.a.
2006 [17] 71 (deceased), consciousness
n.s.
Milhaud France 50 58.7 IS 50 (100) n.a. mean 19.9 50 (100) 35 (70) n.a. n.a.
et al., 2004 [7] (survivors) (1-y-survival:
to 21.8 absence of
(deceased), complete MCA
n.s. infarction)

Only studies with patient recruitment after 12/1999 were considered.

Y = Years; NIHSS = National Institute of Health Stroke Scale; MV = mechanical ventilation; IS = ischemic stroke; ICH = intracerebral hemorrhage; n.a. = not avail-
able; APACHE = Acute Physiology and Chronic Health Evaluation (score); GCS = Glasgow Coma Scale; m = months; a.o. = and others; MCA = middle cerebral artery; d =
days; SAPS = Simplified Acute Physiology Score.

Alonso/Ebert/Kern/Rapp/Hennerici/Fatar
resulting in a three-fold increase in mortality hazard for with our own results. In our cohort, the in-hospital mor-
patients >85 years [6]. This is quite in line with our re- tality of mechanically ventilated patients was 57.1%,
sults, showing that patients >80 years with additional which is quite in line with previously published data [18,
negative outcome markers either do not survive the acute 23]. Strikingly, the percentage of mechanically ventilated
phase or are left in a completely dependent state. The patients in our study was exceptionally high (66.6%) as
markedly higher mean age of our patients compared to compared to previous studies [6, 8]. This finding might
the most previous studies reflects the actual epidemio- be attributable to the multidisciplinary competencies on
logical changes with an aging population. With an overall our stroke unit in the management of severe stroke and
life-expectancy of about 80 years for US citizens and even its complications, thus reducing the need for ICU trans-
slightly more for inhabitants of European states (WHO, ferrals in non-ventilated patients.
2012), we will have to face an increasing number of aged We further found that ICH was a predictor for in-hos-
patients with multiple comorbidities suffering a stroke. pital mortality, while the functional outcome after reha-
Of importance, mortality rates for ischemic stroke pa- bilitation did not differ between ischemic and hemor-
tients in our study did not differ substantially from those rhagic stroke patients. A higher short-term mortality for
reported in younger stroke populations [4, 24]. However, ICH patients treated on specialized stroke units has al-
the functional outcome after rehabilitation in the acute ready been shown by Chambers and coworkers in 1987
stroke survivors was mainly influenced by the patients’ [29]; in stroke patients requiring ICU management, the
age. data are more conflicting. Most studies agree with our
Impaired consciousness at admission is probably the results regarding an increased risk of mortality in ICH
most accepted negative outcome predictor in stroke pa- patients [4, 5]. After correction for age and APACHE III
tients needing ICU management. In our study, coma at score, patients with ICH had a mortality that was 4.1
admission was a strong predictor of in-hospital mortality. times higher than that of ischemic stroke patients in a
In line with this finding, several studies demonstrated a Spanish prospective observational study [4]. However,
correlation of initial GCS and short-term mortality [21, several studies also linked the presence of intracerebral
25, 26]. However, only two studies including >100 pa- hemorrhage to a worse functional outcome [5]. The lack
tients – both originating from Asian countries – evalu- of association in our study might be attributable to two
ated the NIHSS score in their patient collectives as a factors: first, severely affected ischemic stroke patients
stroke-specific severity marker [16, 25]. Consistent with with high NIHSS at admission and only reduced potential
our results, NIHSS was found to predict both in-hospital for improvement might be overrepresented in our study.
mortality as well as functional outcome. In contrast, the Second, hemorrhagic stroke patients have been shown to
two large US studies with inclusion of 8,185 and 99,782 experience even more significant recovery than ischemic
stroke patients treated on ICU, respectively, could only stroke patients [30]. This observation might be attribut-
draw on administrative data and therefore lacked specific able to an improvement of the initial neurological deficit
information on stroke severity or stroke treatment [6, 18]. due to brain compression after the resolution of the he-
Whether to initiate mechanical ventilation in severely matoma [31].
affected stroke patients is still a matter of debate, and, to Notably, the proportion of ischemic stroke patients re-
date, trials that verify the utility of mechanical ventilation ceiving thrombolytic therapy was very high (38.5%), pos-
in severe stroke are still lacking [3]. Undisputedly, the sibly raising the concern that thrombolytic therapy itself
need for mechanical ventilation is a potent predictor of might be an additional risk for ICU admission. However,
mortality and poor outcome [16] and leads to a five-fold whether thrombolysis was performed or not did not af-
increase in mortality hazard in the short-term [6]. The fect the mortality rate or functional outcome. We believe
restraint of many stroke physicians is based on large stud- that, apart from early admission to our stroke unit, the
ies conducted in the 1990s showing catastrophic fatality overrepresentation of severely affected stroke patients in
rates of about 90% in acute stroke patients requiring me- our cohort may explain the high thrombolysis rates. Sup-
chanical ventilation [27, 28]. More recent studies report porting this theory, Faigle and colleagues demonstrated
somewhat lower rates, however, still reaching about 70% that high NIHSS score at admission predicts ICU needs
mortality after 1 year [7, 8, 18]. In a small prospective following tPA for acute ischemic stroke [32]. This finding
study including 58 patients with ischemic stroke in need should not discourage stroke physicians to perform
of mechanical ventilation, the presence of stupor or coma thrombolysis even in severely affected and older stroke
was an independent predictor of mortality [8], consistent patients [33]. Although octogenarians undergoing throm-

Outcome of Stroke Patients on ICU Cerebrovasc Dis 2015;40:10–17 15

DOI: 10.1159/000430871
bolytic treatment may exhibit a higher mortality rate horts. This finding may reflect the recent advances in
compared with patients <80 years [34], recent data from intensive care management. However, functional out-
the third international stroke trial (IST-3) demonstrated come was poor especially in older patients with either
that patients >80 years achieved similar benefit when hemorrhagic stroke or need of MV, leaving barely all
compared to younger patients [35]. Likewise, the study survivors in a dependent state. Several studies have
identified significant trends towards larger effects of shown that retrospective consent to neurointensive care
treatment in patients with more severe strokes [35] de- depends on the functional outcome [36, 37]. In a cohort
spite an increasing risk of hemorrhagic complications. of 704 patients admitted to a neurocritical care unit, only
In summary, we analysed the short-term outcome 19% of patients with an outcome mRS of 4 or 5 retro-
and outcome predictors in acute stroke patients in need spectively consented to ICU management [36]. Further
of ICU treatment. In comparison to previously analysed decisions about when to admit stroke patients to ICU
cohorts, the stroke population in our study was remark- will thus have to focus on the potential functional out-
ably older with a very high percentage of patients requir- come, taking account of accepted predictors of poor out-
ing mechanical ventilation. Although patient age and come. Consideration of the putative patient’s will or pa-
MV are main predictors of negative outcome, in-hospi- tient’s advance directives will become increasingly im-
tal mortality rates in our study were not significantly portant [38], and efforts to improve end-of-life care
higher than in younger and less severely affected co- should be made [39].

References
1 Donnan GA, Fisher M, Macleod M, Davis 9 Jauch EC, Saver JL, Adams HP Jr, Bruno A, laboration: Stroke unit care benefits patients
SM: Stroke. Lancet 2008;371:1612–1623. Connors JJ, Demaerschalk BM, Khatri P, with intracerebral hemorrhage: systematic re-
2 Langhorne P, Dey P, Woodman M, Kalra L, McMullan PW Jr, Qureshi AI, Rosenfield K, view and meta-analysis. Stroke 2013; 44:
Wood-Dauphinee S, Patel N, Hamrin E: Is Scott PA, Summers DR, Wang DZ, Winter- 3044–3049.
stroke unit care portable? A systematic review mark M, Yonas H; American Heart Associa- 15 Stroke Unit Trialists’ Collaboration: Organ-
of the clinical trials. Age Ageing 2005;34:324– tion Stroke Council; Council on Cardiovascu- ised inpatient (stroke unit) care for stroke.
330. lar Nursing; Council on Peripheral Vascular Cochrane Database Syst Rev 2013;
3 Kirkman MA, Citerio G, Smith M: The inten- Disease; Council on Clinical Cardiology: 9:CD000197.
sive care management of acute ischemic Guidelines for the early management of pa- 16 Jeng JS, Huang SJ, Tang SC, Yip PK: Predic-
stroke: an overview. Intensive Care Med 2014; tients with acute ischemic stroke: a guideline tors of survival and functional outcome in
40:640–653. for healthcare professionals from the acute stroke patients admitted to the stroke
4 Navarrete-Navarro P, Rivera-Fernández R, American heart association/American stroke intensive care unit. J Neurol Sci 2008;270:60–
López-Mutuberría MT, Galindo I, Murillo F, association. Stroke 2013;44:870–947. 66.
Dominguez JM, Muñoz A, Jimenez-Moragas 10 European Stroke Organisation (ESO) Execu- 17 Lan MY, Wu SJ, Chang YY, Chen WH, Lai SL,
JM, Nacle B, Vázquez-Mata G: Outcome pre- tive Committee; ESO Writing Committee: Liu JS: Neurologic and non-neurologic pre-
diction in terms of functional disability and Guidelines for management of ischaemic dictors of mortality in ischemic stroke pa-
mortality at 1 year among ICU-admitted se- stroke and transient ischaemic attack 2008. tients admitted to the intensive care unit. J
vere stroke patients: a prospective epidemio- Cerebrovasc Dis 2008;25:457–507. Formos Med Assoc 2006;105:653–658.
logical study in the south of the European 11 Kern R, Nagayama M, Toyoda K, Steiner T, 18 Lahiri S, Mayer SA, Fink ME, Lord AS, Rosen-
Union (Evascan project, Andalusia, Spain). Hennerici MG, Shinohara Y: Comparison of gart A, Mangat HS, Segal AZ, Claassen J, Ka-
Intensive Care Med 2003;29:1237–1244. the European and Japanese guidelines for the mel H: Mechanical ventilation for acute
5 Fanshawe M, Venkatesh B, Boots RJ: Out- management of ischemic stroke. Cerebrovasc stroke: a multi-state population-based study.
come of stroke patients admitted to intensive Dis 2013;35:402–418. Neurocrit Care 2014, Epub ahead of print.
care: experience from an Australian teaching 12 Guidelines for intensive care unit admission, 19 Xian Y, Holloway RG, Smith EE, Schwamm
hospital. Anaesth Intensive Care 2002; 30: discharge, and triage. Task force of the LH, Reeves MJ, Bhatt DL, Schulte PJ, Cox M,
628–632. American college of critical care medicine, so- Olson DM, Hernandez AF, Lytle BL, Anstrom
6 Golestanian E, Liou JI, Smith MA: Long-term ciety of critical care medicine. Crit Care Med KJ, Fonarow GC, Peterson ED: Racial/ethnic
survival in older critically ill patients with 1999;27:633–638. differences in process of care and outcomes
acute ischemic stroke. Crit Care Med 2009;37: 13 Cereda CW, George PM, Pelloni LS, Gandol- among patients hospitalized with intracere-
3107–3113. fi-Decristophoris P, Mlynash M, Biancon bral hemorrhage. Stroke 2014;45:3243–3250.
7 Milhaud D, Popp J, Thouvenot E, Heroum C, Montaperto L, Limoni C, Stojanova V, Malac- 20 Wasay M, Khatri IA, Kaul S: Stroke in South
Bonafé A: Mechanical ventilation in ischemic rida R, Städler C, Bassetti CL: Beneficial ef- Asian countries. Nat Rev Neurol 2014; 10:
stroke. J Stroke Cerebrovasc Dis 2004;13:183– fects of a semi-intensive stroke unit are 135–143.
188. beyond the monitor. Cerebrovasc Dis 2015; 21 Riachy M, Sfeir F, Sleilaty G, Hage-Chahine S,
8 Santoli F, De Jonghe B, Hayon J, Tran B, Pipe- 39:102–109. Dabar G, Bazerbachi T, Aoun-Bacha Z, Khay-
raud M, Merrer J, Outin H: Mechanical ven- 14 Langhorne P, Fearon P, Ronning OM, Kaste at G, Koussa S: Prediction of the survival and
tilation in patients with acute ischemic stroke: M, Palomaki H, Vemmos K, Kalra L, Indreda- functional ability of severe stroke patients af-
survival and outcome at one year. Intensive vik B, Blomstrand C, Rodgers H, Dennis MS, ter ICU therapeutic intervention. BMC Neu-
Care Med 2001;27:1141–1146. Al-Shahi Salman R; Stroke Unit Trialists’ Col- rol 2008;8:24.

16 Cerebrovasc Dis 2015;40:10–17 Alonso/Ebert/Kern/Rapp/Hennerici/Fatar

DOI: 10.1159/000430871
22 Schielke E, Busch MA, Hildenhagen T, Holt- stroke patients undergoing mechanical venti- Murray V, Berge E, Slot KB, Hankey GJ, Cor-
kamp M, Küchler I, Harms L, Masuhr F: lation. Intensive Care Med 1994;20:32–36. reia M, Peeters A, Matz K, Lyrer P, Gubitz G,
Functional, cognitive and emotional long- 29 Chambers BR, Norris JW, Shurvell BL, Phillips SJ, Arauz A: The benefits and harms
term outcome of patients with ischemic Hachinski VC: Prognosis of acute stroke. of intravenous thrombolysis with recombi-
stroke requiring mechanical ventilation. J Neurology 1987;37:221–225. nant tissue plasminogen activator within 6 h
Neurol 2005;252:648–654. 30 Katrak PH, Black D, Peeva V: Do stroke pa- of acute ischaemic stroke (the third interna-
23 Steiner T, Mendoza G, De Georgia M, tients with intracerebral hemorrhage have a tional stroke trial [IST-3]): a randomised con-
Schellinger P, Holle R, Hacke W: Prognosis of better functional outcome than patients with trolled trial. Lancet 2012;379:2352–2363.
stroke patients requiring mechanical ventila- cerebral infarction? PM R 2009;1:427–433. 36 Kiphuth IC, Köhrmann M, Kuramatsu JB,
tion in a neurological critical care unit. Stroke 31 Mendelow AD: Mechanisms of ischemic Mauer C, Breuer L, Schellinger PD, Schwab S,
1997;28:711–715. brain damage with intracerebral hemorrhage. Huttner HB: Retrospective agreement and
24 Rordorf G, Koroshetz W, Efird JT, Cramer Stroke 1993; 24(12 suppl):I115–I117; discus- consent to neurocritical care is influenced by
SC: Predictors of mortality in stroke patients sion I118–I119. functional outcome. Crit Care 2010;14:R144.
admitted to an intensive care unit. Crit Care 32 Faigle R, Sharrief A, Marsh EB, Llinas RH, Ur- 37 Foerch C, Lang JM, Krause J, Raabe A, Sitzer
Med 2000;28:1301–1305. rutia VC: Predictors of critical care needs after M, Seifert V, Steinmetz H, Kessler KR: Func-
25 Moon BH, Park SK, Jang DK, Jang KS, Kim IV thrombolysis for acute ischemic stroke. tional impairment, disability, and quality of
JT, Han YM: Use of APACHE II and SAPS II PLoS One 2014;9:e88652. life outcome after decompressive hemicrani-
to predict mortality for hemorrhagic and 33 Kurre W, Aguilar-Pérez M, Niehaus L, Fisch- ectomy in malignant middle cerebral artery
ischemic stroke patients. J Clin Neurosci er S, Schmid E, Bäzner H, Henkes H: Predic- infarction. J Neurosurg 2004;101:248–254.
2015;22:111–115. tors of outcome after mechanical thrombec- 38 Hartog CS, Peschel I, Schwarzkopf D, Curtis
26 Handschu R, Haslbeck M, Hartmann A, Fell- tomy for anterior circulation large vessel oc- JR, Westermann I, Kabisch B, Pfeifer R,
giebel A, Kolominsky-Rabas P, Schneider D, clusion in patients aged ≥80 years. Cerebrovasc Guenther A, Michalsen A, Reinhart K: Are
Berrouschot J, Erbguth F, Reulbach U: Mor- Dis 2013;36:430–436. written advance directives helpful to guide
tality prediction in critical care for acute 34 Bhatnagar P, Sinha D, Parker RA, Guyler P, end-of-life therapy in the intensive care unit?
stroke: severity of illness-score or coma-scale? O’Brien A: Intravenous thrombolysis in acute A retrospective matched-cohort study. J Crit
J Neurol 2005;252:1249–1254. ischaemic stroke: a systematic review and me- Care 2014;29:128–133.
27 el-Ad B, Bornstein NM, Fuchs P, Korczyn ta-analysis to aid decision making in patients 39 Truog RD, Campbell ML, Curtis JR, Haas CE,
AD: Mechanical ventilation in stroke pa- over 80 years of age. J Neurol Neurosurg Psy- Luce JM, Rubenfeld GD, Rushton CH, Kaufman
tients – is it worthwhile? Neurology 1996; 47: chiatry 2011;82:712–717. DC: Recommendations for end-of-life care in
657–659. 35 Sandercock P, Wardlaw JM, Lindley RI, Den- the intensive care unit: a consensus statement by
28 Burtin P, Bollaert PE, Feldmann L, Nace L, nis M, Cohen G, Murray G, Innes K, Venables the American college [corrected] of critical care
Lelarge P, Bauer P, Larcan A: Prognosis of G, Czlonkowska A, Kobayashi A, Ricci S, medicine. Crit Care Med 2008;36:953–963.

Outcome of Stroke Patients on ICU Cerebrovasc Dis 2015;40:10–17 17

DOI: 10.1159/000430871