Intensive Care Med (2017) 43:1075–1084

DOI 10.1007/s00134-017-4807-z

ORIGINAL

Potentially modifiable factors

contributing to sepsis‑associated
encephalopathy
Romain Sonneville1,2*, Etienne de Montmollin3,4, Julien Poujade1, Maïté Garrouste‑Orgeas3,5,
Bertrand Souweine6, Michael Darmon8,9, Eric Mariotte7, Laurent Argaud10, François Barbier11,
Dany Goldgran‑Toledano12, Guillaume Marcotte13, Anne‑Sylvie Dumenil14, Samir Jamali15, Guillaume Lacave16,
Stéphane Ruckly3, Bruno Mourvillier1,3 and Jean‑François Timsit1,3

© 2017 Springer-Verlag Berlin Heidelberg and ESICM

Abstract 
Purpose:  Identifying modifiable factors for sepsis-associated encephalopathy may help improve patient care and
outcomes.
Methods:  We conducted a retrospective analysis of a prospective multicenter database. Sepsis-associated encepha‑
lopathy (SAE) was defined by a score on the Glasgow coma scale (GCS) <15 or when features of delirium were noted.
Potentially modifiable risk factors for SAE at ICU admission and its impact on mortality were investigated using multi‑
variate logistic regression analysis and Cox proportional hazard modeling, respectively.
Results:  We included 2513 patients with sepsis at ICU admission, of whom 1341 (53%) had sepsis-associated
encephalopathy. After adjusting for baseline characteristics, site of infection, and type of admission, the following
factors remained independently associated with sepsis-associated encephalopathy: acute renal failure [adjusted odds
ratio (aOR) = 1.41, 95% confidence interval (CI) 1.19–1.67], hypoglycemia <3 mmol/l (aOR = 2.66, 95% CI 1.27–5.59),
hyperglycemia >10 mmol/l (aOR = 1.37, 95% CI 1.09–1.72), hypercapnia >45 mmHg (aOR = 1.91, 95% CI 1.53–2.38),
hypernatremia >145 mmol/l (aOR = 2.30, 95% CI 1.48–3.57), and S. aureus (aOR = 1.54, 95% CI 1.05–2.25). Sepsis-asso‑
ciated encephalopathy was associated with higher mortality, higher use of ICU resources, and longer hospital stay.
After adjusting for age, comorbidities, year of admission, and non-neurological SOFA score, even mild alteration of
mental status (i.e., a score on the GCS of 13–14) remained independently associated with mortality (adjusted hazard
ratio = 1.38, 95% CI 1.09–1.76).
Conclusions:  Acute renal failure and common metabolic disturbances represent potentially modifiable factors con‑
tributing to sepsis-associated encephalopathy. However, a true causal relationship has yet to be demonstrated. Our
study confirms the prognostic significance of mild alteration of mental status in patients with sepsis.
Keywords:  Sepsis, Delirium, Encephalopathy, Acute renal failure, Brain injury, Outcome

*Correspondence: romain.sonneville@aphp.fr
1
Department of Intensive Care Medicine and Infectious Diseases, Hôpital
Bichat-Claude Bernard, 46 rue Henri Huchard, 75877 Paris Cedex, France
Full author information is available at the end of the article

Take‑home message:  Acute renal failure and common metabolic

disturbances represent potentially modifiable factors contributing to
sepsis-associated encephalopathy.

A list of members of the OUTCOMEREA study group is given in the

Acknowledgements.
1076
Introduction
Sepsis is frequently complicated by an encephalopa-
thy ranging from confusion to coma, which is associ-
ated with brain injury and adverse outcome [1–4]. Many
mechanisms have been incriminated, including cerebral
endothelial alterations with increased blood–brain bar-
rier permeability, and microglial activation in brain
parenchyma [5–9]. Neuroimaging and autopsy studies
suggest that sepsis is associated with brain injury, in the
form of ischemic lesions, white matter damage, micro
abscesses, and brain atrophy involving the frontal cor-
tex and hippocampus [5, 8, 10–13]. Data on potentially
modifiable factors contributing to sepsis-associated
encephalopathy (SAE) are scarce. SAE might notably
be aggravated by severe hypoxia, systemic hypotension,
common metabolic, and use of several drugs, i.e., seda-
tive agents, antibiotics with significant neurotoxicity and
steroids [14, 15].
In a landmark study conducted in non-sedated septic
patients, SAE, defined by a score on the Glasgow coma
scale (GCS) <15 at ICU admission, was observed in 54%
of patients [1]. The severity of encephalopathy correlated
with bacteremia and renal and hepatic dysfunctions. A
recent single-center study found several parameters to be
associated with SAE, including severity of illness, meta-
bolic disturbances, site of infection, and type of micro-
organism [16]. These studies had limitations, including
a relatively low sample size, a single-center design, and
a lack of adjustment for important risk factors for ICU
delirium, i.e., age, hypertension, pre-existing cognitive
impairment, type of admission, and long-term use of psy-
choactive drugs [17].
Previously published data suggest an important impact
of SAE on outcomes, with mortality rates over 60% in
comatose patients at ICU admission [1]. The recent defi-
nitions for sepsis and septic shock (Sepsis-3) suggest that
altered mentation, defined by a score on the GCS  ≤13,
in patients with suspected infection outside of the ICU
is independently associated with hospital mortality [18].
In this retrospective analysis of a prospective multi-
center database, we aimed to identify the determinants
and prognostic value of SAE in ICU patients. Specifically,
we focused on identifying early and potentially modifi-
able factors for SAE at ICU admission.
Methods
Patients
This study was conducted using data from the French
prospective multicenter (n  =  12 ICUs) OUTCOM-
EREA database. The study used data from patients
included between 1997 and 2014. OUTCOMEREA has
already been extensively described elsewhere [19]. This
database has been approved by the French Advisory
Committee for Data Processing in Health Research
(CCTIRS) and the French Informatics and Liberty
Commission (CNIL) [19, 20]. The study was approved
by the ethics committee of Clermont-Ferrand, France.
We included adult patients with a diagnosis of severe
sepsis or septic shock at ICU admission. We used
parameters collected prospectively by our data-capture
software to select patients with severe sepsis, defined as
systemic inflammatory response syndrome combined
with an infectious episode and dysfunction of at least
one organ, as previously described [21]. Data on sepsis
definitions are provided in Online resource 1. Exposure
to antibiotic agents with known neurotoxicity [namely
beta-lactams (penicillins, cephalosporins, carbapenems)
and/or fluoroquinolones], exposure to benzodiazepines,
propofol, neuroleptics, barbiturates and/or opioids, and
presence of metabolic disturbances at ICU admission
(i.e., hypernatremia (>145  mmol/l) or hyponatremia
(<135  mmol/l), hyperglycemia (>10  mmol/l) or hypo-
glycemia (<3  mmol/l), and hypercapnia (>45  mmHg))
were also prospectively collected. We excluded patients
with acute brain injury as the cause of ICU admission
(e.g., meningitis, encephalitis, status epilepticus, trau-
matic brain injury, or stroke). For each patient, we ret-
rospectively collected the following data in the medical
charts: (1) medical history, including common risk fac-
tors for delirium: chronic alcohol consumption, long-
term use of psychoactive drugs (i.e., benzodiazepines,
neuroleptics, serotonin uptake receptor inhibitors) and
statins, chronic psychiatric disorder, and chronic neu-
rological disease; (2) infection-related parameters: site
of infection, microorganism, and presence or absence
of associated-bacteremia; (3) neurological presentation
at ICU admission: hypoactive or hyperactive (agitated)
presentation, focal sign(s), and/or convulsive seizure(s).
Severity of illness was graded at ICU admission with
use of a modified simplified acute physiology score
(SAPS II), excluding the GCS score [22]. Organ failure
was assessed with each component of the sepsis-related
organ failure assessment (SOFA) score, and a modified
SOFA score was calculated, excluding the neurological
component [23, 24]. For each component of the SOFA
score, organ failure was defined by a SOFA >2. Lengths
of ICU and hospital stays were determined starting at
ICU admission.
Sepsis‑associated encephalopathy
SAE was defined at ICU admission as a score on the
GCS  <15, or when abnormal neurological findings con-
sistent with delirium (i.e., inattention, disorientation,
altered thinking, psychomotor slowing, and/or agita-
tion) were noted in the medical charts. For each patient,
we used the lowest score on the GCS available at ICU

admission entered in the OUTCOMEREA database.
For patients sedated before ICU admission, a score on
the GCS before any administration of sedation is always
recorded in the database. For those patients, we used the
assumed score on the GCS for analyses, i.e., the score
measured before any administration of sedative/relaxant
drug [25]. For postoperative patients, we used the score
on the GCS measured before surgery.
Statistical analysis
Data are presented as numbers and percentages or
median and interquartile ranges. Accordingly, Chi
square tests for equal proportions or the Wilcoxon rank-
sum test was used to test differences. To identify vari-
ables associated with SAE at ICU admission, we used
univariate logistic regression after stratification by
center. Clinically relevant variables and variables sig-
nificantly associated with SAE in univariate analysis
(p  <  0.1) were entered in the multivariate model. All
variables entered in multivariate models were collected
at ICU admission. Missing data are described in table
footnotes. For each continuous variable, the linearity of
the logit was tested using additive models. Collinearity
between variables and 2 by 2 interactions were tested.
Multivariate logistic regression was performed with a
stepwise selection procedure. Sensitivity analyses were
conducted in medical patients, in the whole cohort after
exclusion of SAE patients with a score on the GCS of
9–15, and in patients without history of neurological
disease or cognitive impairment.
Cox proportional hazard modeling was used to inves-
tigate the relationship between SAE and mortality. Sur-
vival analyses were censored at day 30. Overall, 290/2513
(11.5%) patients were lost to follow-up before day 30 and
were censored at hospital discharge. Proportional hazards
assumption was tested graphically with the log of negative
log of estimated survivor functions versus the log of time.
To identify the impact of the severity of SAE on outcome,
the GCS variable was categorized into clinically relevant
categories, adapted from a previous study [1]: a score on
the GCS of 15 and features of delirium; a score on the
GCS of 13–14; a score on the GCS of 9–12; and a score on
the GCS of 3–8, indicating coma. Patients with a score on
the GCS of 15 and without features of delirium were con-
sidered to have no encephalopathy (non-SAE patients).
The analysis was stratified by center and adjusted for
clinically relevant parameters associated with outcome,
i.e., age, chronic liver disease, year of admission, and non-
neurological SOFA score. Kaplan–Meier survival curves
were computed and compared with use of the log-rank
test. All statistical analyses were conducted with SAS 9.4
(SAS Institute Inc., Cary, NC, USA). A p value of 0.05 and
lower was considered statistically significant.
1077
Results
Patients
Among 18,713 ICU admissions over the study period,
we identified 2647 patients with sepsis at ICU admission.
We excluded 134 patients with acute brain injury and
2513 patients were included in the final cohort (Online
resource 2). Patients’ characteristics are presented in
Tables  1, 2. Patients were predominantly male, mainly
admitted for medical reasons. Sepsis had mainly a pul-
monary or an intra-abdominal origin, and bacteremia
was diagnosed a quarter of the patients. On the day of
ICU admission, the estimated score on the GCS was 15
(10–15), 2208 (88%) patients had septic shock, and 1071
(43%) patients required invasive mechanical ventilation.
Sepsis‑associated encephalopathy
Overall, SAE was observed in 1341 (53%) patients, with
an estimated score on the GCS of 11 (4–14). Among
SAE patients, 251/1341 (19%) patients had a score on the
GCS of 15 and features of delirium, 308/1341 (23%) had
a score on the GCS of 13–14, 241/1341 (18%) a score on
the GCS of 9–12, and 541/1341 (40%) had a score on the
GCS of 3–8, indicating coma. SAE patients most often
presented with clinical features of hypoactive delirium,
whereas agitation (n  =  137/1341, 10%), convulsive sei-
zures (n  =  21/1341, 2%), and focal signs (n  =  17/1341,
1%) were rarely observed.
Univariate analysis of baseline characteristics associ-
ated with SAE is presented in Table 1. Compared to non-
SAE patients, SAE patients were older and had a more
frequent history of chronic psychiatric disease, neuro-
logical disease (i.e., stroke, epilepsy), and pre-existing
chronic cognitive impairment. Long-term use of psy-
choactive drugs, alcohol abuse, chronic liver disease, and
immunodepression were also more frequently observed
in SAE patients.
Univariate analysis of ICU admission characteristics
is presented in Table  2. SAE was associated with medi-
cal admission, reason for ICU admission, higher SAPS 2
and SOFA scores, respiratory failure, liver failure, renal
failure, infection source, bacteremia, S. aureus infection,
and septic shock. Although cardiovascular SOFA scores
did not differ between SAE and non-SAE patients, SAE
patients had more severe circulatory failure, as assessed
by lower systolic blood pressure, higher heart rates, and
higher lactate levels on the day of ICU admission. By
contrast, ­PaO2 was comparable between the two groups.
Common metabolic disturbances (i.e., hypoglycemia,
hyperglycemia, hypercapnia, and hypernatremia) and use
of propofol, benzodiazepines, opioids, and steroids were
also significantly associated with SAE at ICU admission.
After adjusting for baseline characteristics, site of
infection, and type of admission, the following factors
1078

Table 1  Patients’ baseline characteristics

Variable All patients No SAE SAE p value
n = 2513 n = 1172 n = 1341

Age, years 66.2 [54.3; 76.5] 64.1 [50.8; 74.7] 68.2 [57.4; 77.2] <0.01
Male sex 1528 (60.8) 699 (59.6) 829 (61.8) 0.26
Comorbidities, n (%)
 Hypertension 927 (36.9) 415 (35.4) 512 (38.2) 0.15
 Chronic liver disease 183 (7.3) 55 (4.7) 128 (9.5) <0.01
 Grade IV NYHA cardiomyopathy 360 (14.3) 153 (13.1) 207 (15.4) 0.09
 COPD 249 (9.9) 103 (8.8) 146 (10.9) 0.08
 Hemodialysis 178 (7.1) 83 (7.1) 95 (7.1) 1.00
 Immunodepression 750 (29.8) 412 (35.2) 338 (25.2) <0.01
 Diabetes 439 (17.5) 175 (14.9) 264 (19.7) <0.01
 Chronic alcohol abuse 189 (7.5) 42 (3.6) 147 (11) <0.01
 Psychiatric disease 39 (1.6) 10 (0.9) 29 (2.2) <0.01
 History of neurological disease 278 (11.1) 96 (8.2) 182 (13.6) <0.01
  Stroke 145 (5.8) 46 (3.9) 99 (7.4) <0.01
  Traumatic brain injury 23 (0.9) 8 (0.7) 15 (1.1) 0.25
  Epilepsy 47 (1.9) 12 (1) 35 (2.6) <0.01
  Other 92 (3.7) 39 (3.3) 53 (4) 0.41
 Pre-existing cognitive impairment 49 (1.9) 10 (0.9) 39 (2.9) <0.01
Chronic treatments, n (%)
 Statins 355 (14.1) 153 (13.1) 202 (15.1) 0.15
 Steroids 168 (6.7) 83 (7.1) 85 (6.3) 0.46
 Long-term use of psychoactive ­drugsa 483 (19.2) 187 (16) 296 (22.1) <0.01
  Benzodiazepines 232 (9.2) 82 (7) 150 (11.2) <0.01
  Serotonin uptake receptors inhibitors 111 (4.4) 43 (3.7) 68 (5.1) 0.09
  Antipsychotics 78 (3.1) 21 (1.8) 57 (4.3) <0.01
a
  Each patient could receive more than one psychoactive drug
SAE sepsis-associated encephalopathy, NYHA New York Heart Association, COPD chronic obstructive pulmonary disease

were associated with SAE on the day of ICU admis-
sion (Table  3): acute renal failure [adjusted odds ratio
(aOR)  =  1.41, 95% confidence interval (CI) 1.19–1.67],
hypoglycemia <3  mmol/l (aOR  =  2.66, 95% CI 1.27–
5.59), hyperglycemia >10  mmol/l (aOR  =  1.37, 95% CI
1.09–1.72), hypercapnia >45 mmHg (aOR = 1.91, 95% CI
1.53–2.38), hypernatremia >145 mmol/l (aOR = 2.3, 95%
CI 1.48–3.57), and S. aureus (aOR = 1.54, 95% CI 1.05–
2.25). Sensitivity analyses conducted in medical patients,
in the whole cohort after exclusion of SAE patients with
a score on the GCS of 9–15, and in patients without his-
tory of neurological disease or cognitive impairment are
presented in Online resource 3.
Outcomes. Interventions and outcomes are described
in Online resource 4. Overall, 290 patients were lost to
follow-up and censored before day 30. Presence of SAE
at ICU admission was associated with higher mortality,
higher use of ICU resources, and longer hospital stay.
The severity of SAE remained associated with ICU and
hospital mortality (Online resource 5). Kaplan–Meier’s
survival estimates of patients according to the presence
or absence of SAE at ICU admission, and according to
the severity of encephalopathy at ICU admission are
presented in Online resource 6 and Fig.  1, respectively.
After adjustment for age, chronic liver disease, year of
admission and non-neurological SOFA score, even mild
alteration of mental status (GCS 13–14) remained inde-
pendently associated with mortality (Table 4). A stronger
independent association was observed in a sensitivity
analysis conducted in patients without history of neuro-
logical disease or cognitive impairment (Online resource
7). Histograms of GCS scores are provided in Online
resource 8.
Discussion
In this retrospective analysis of a large prospective mul-
ticenter database, we found that one out of two patients
admitted with sepsis had encephalopathy at ICU admis-
sion, as defined by a score on the GCS <15 or presence of
abnormal neurological findings consistent with delirium.
The identified subset of patients more susceptible to SAE
was older patients with a history of chronic alcohol abuse,
 1079

Table 2  Patients’ characteristics at ICU admission

Variable All patients No SAE SAE p value
n = 2513 n = 1172 n = 1341

Type of admission, n (%)

 Medical 1845 (73.6) 812 (69.5) 1033 (77.1) <0.01
 Scheduled surgery 119 (4.7) 67 (5.7) 52 (3.9)
 Unscheduled surgery 544 (21.7) 290 (24.8) 254 (19)
Time from hospital to ICU admission, days 1 [0; 5] 1 [0; 4] 1 [0; 6] 0.07
Reason for ICU admission, n (%) <0.01
 Acute respiratory failure 468 (18.6) 204 (17.4) 264 (19.7)
 Circulatory failure 1650 (65.7) 750 (64) 900 (67.1)
 Acute renal failure 83 (3.3) 42 (3.6) 41 (3.1)
 Coma 59 (2.3) 1 (0.1) 58 (4.3)
 Other 253 (10.1) 175 (14.9) 78 (5.8)
Highest SBP (mmHg) 135 [120; 154] 135 [120; 152] 136 [120; 156] 0.20
Lowest SBP (mm Hg) 80 [64; 91] 83 [70; 94] 76 [58; 90] <0.01
Heart rate ­(min−1) 120 [103; 139] 118 [102; 136] 121 [104; 140] <0.01
Temperature (°C) 38.2 [37.5; 39.1] 38.4 [37.6; 39.1] 38.1 [37.4; 39.0] <0.01
Modified SAPS 2 score 51 [38; 67] 43 [33; 53] 61 [46; 79] <0.01
Modified SOFA score 7 [4, 9] 6 [3, 8] 7 [5, 10] <0.01
Organ failures (SOFA score >2), n (%)
 Respiration, >2 1618 (64.4) 683 (58.3) 935 (69.7) <0.01
 Coagulation, >2 1035 (41.2) 485 (41.4) 550 (41) 0.85
 Liver, >2 656 (26.1) 274 (23.4) 382 (28.5) <0.01
 Cardiovascular, >2 1930 (76.8) 885 (75.5) 1045 (77.9) 0.15
 Renal, >2 1431 (56.9) 591 (50.4) 840 (62.6) <0.01
Septic shock, n (%) 2208 (87.9) 978 (83.4) 1230 (91.7) <0.01
Infection source, n (%)
 Lung 717 (28.5) 288 (24.6) 429 (32) <0.01
 Intra-abdominal 640 (25.5) 326 (27.8) 314 (23.4) 0.01
 Catheter-related 79 (3.1) 51 (4.4) 28 (2.1) <0.01
 Skin and soft tissue 144 (5.7) 85 (7.3) 59 (4.4) <0.01
 Urinary tract 322 (12.8) 164 (14) 158 (11.8) 0.10
 Endocarditis 41 (1.6) 12 (1) 29 (2.2) 0.02
Bacteremia, n (%) 656 (26.1) 328 (28) 328 (24.5) 0.04
Main microorganisms, n (%)
 Staphylococcus aureus 143 (5.7) 51 (4.4) 92 (6.9) <0.01
 Streptococci 318 (12.7) 154 (13.1) 164 (12.2) 0.49
 Pneumococci 117 (4.7) 57 (4.9) 60 (4.5) 0.64
 Coagulase-negative staphylococci 137 (5.5) 70 (6) 67 (5) 0.28
 Enterobacteriaceae 736 (29.3) 356 (30.4) 380 (28.3) 0.26
 Pseudomonas aeruginosa 168 (6.7) 71 (6.1) 97 (7.2) 0.24
 Anaerobes 68 (2.7) 33 (2.8) 35 (2.6) 0.75
 Candida sp. 99 (3.9) 46 (3.9) 53 (4) 0.97
Biological parameters
 PaO2 (mmHg) 96 [75; 140] 99 [76; 140] 95 [75; 140] 0.61
 Leukocytes (G/l) 12.6 [5.3; 19.4] 11.9 [4.5; 18.8] 13.2 [6.3; 20.5] <0.01
 Hematocrit, (%) 31 [26, 36] 30 [26, 35] 31 [27, 36] 0.03
 Prothrombin time (%) 59 [44; 71] 61 [47; 71] 57 [41; 71] <0.01
 Lactates (mmol/l) 2.5 [1.5; 4.5] 2.2 [1.5; 3.8] 2.8 [1.6; 5.1] <0.01
 Hypoglycemia (<3 mmol/l), n (%) 46 (1.8) 10 (0.9) 36 (2.7) <0.01
 Hyperglycemia (>10 mmol/l), n (%) 430 (17.1) 174 (14.8) 256 (19.1) <0.01
1080

Table 2  continued
Variable All patients No SAE SAE p value
n = 2513 n = 1172 n = 1341
 Hypercapnia (> 45 mmHg), n (%) 486 (19.3) 157 (13.4) 329 (24.5) <0.01
 Hyponatremia (<135 mmol/l), n (%) 781 (31.1) 373 (31.8) 408 (30.4) 0.45
 Hypernatremia (> 145 mmol/l), n (%) 124 (4.9) 31 (2.6) 93 (6.9) <0.01
Medications, n (%)
 Beta-lactams 1853 (73.7) 884 (75.4) 969 (72.3) 0.07
 Fluoroquinolones 24 (1) 7 (0.6) 17 (1.3) 0.08
 Propofol 245 (9.7) 97 (8.3) 148 (11) 0.02
 Benzodiazepines 723 (28.8) 213 (18.2) 510 (38) <0.01
 Opioids 890 (35.4) 304 (25.9) 586 (43.7) <0.01
 Neuroleptics 7 (0.3) 2 (0.2) 5 (0.4) 0.34
 Steroids (acute use) 555 (22.1) 230 (19.6) 325 (24.2) <0.01
SAE sepsis-associated encephalopathy, ICU Intensive Care Unit, SBP systolic blood pressure, SAPS simplified acute physiology score, SOFA sequential organ failure
assessment

Table 3 Multivariate analysis of  factors associated

1.0
with sepsis-associated encephalopathy
Variable OR 95% CI p value

0.8
Age, per 1-year increment 1.02 1.01 1.02 <0.01
Survival probability
Chronic alcohol abuse 3.38 2.34 4.89 <0.01
0.6
History of neurological disease 1.56 1.18 2.06 <0.01
Pre-existing cognitive impairment 2.25 1.09 4.67 0.03
0.4

Long-term use of psychoactive drugs 1.37 1.11 1.70 <0.01

Medical ­admissiona 1.75 1.43 2.14 <0.01 Glasgow 3−8
0.2

Glasgow 9−12
Renal SOFA > 2 1.41 1.19 1.67 <0.01 Glasgow 13−14
Glasgow 15 SAE + Logrank p<.0001
Hypoglycemia, <3 mmol/l 2.66 1.27 5.59 <0.01 Glasgow 15 SAE −
0.0

Hyperglycemia, >10 mmol/l 1.37 1.09 1.72 <0.01

Hypercapnia, >45 mmHg 1.91 1.53 2.38 <0.01 0 5 10 15 20 25 30
Hypernatremia, >145 mmol/l 2.30 1.48 3.57 <0.01 Days after Inclusion
No. of patients at risk
Staphylococcus aureus 1.54 1.05 2.25 0.03 Glasgow 3−8 541 329 277 237 213 191 174
Glasgow 9−12 241 204 178 156 149 137 121
Catheter-related infection 0.53 0.32 0.88 0.01
Glasgow 13−14 308 260 239 216 199 189 181
Skin and soft tissue infection 0.57 0.39 0.82 <0.01 Glasgow 15 SAE + 251 222 208 195 186 174 167
Glasgow 15 SAE − 1172 1077 998 932 868 821 782
OR odds ratio, CI confidence interval, SOFA sequential organ failure assessment
Fig. 1  Kaplan–Meier’s survival estimates of patients according to the
AUC = 0.698; Hosmer and Lemeshow: p value = 0.92
severity of encephalopathy at ICU admission
a
  5 missing values for type of admission were imputed to the mode

neurological disease, pre-existing cognitive impairment,
and long-term use of psychoactive drugs. We identified
clinically relevant and potentially modifiable factors for
SAE, in the form of acute renal failure and hypoglyce-
mia, hyperglycemia, hypercapnia, and hypernatremia.
Our study confirms the prognostic value of SAE at ICU
admission. Interestingly, even mild alteration in mental
status at ICU admission, defined by a score on the GCS
of 13–14, was independently associated with mortality.
The independent association between acute renal fail-
ure and SAE at ICU admission may have several expla-
nations. Acute renal failure is associated with biological
alterations that may impact on brain function, includ-
ing severe acidosis and uremia. Renal failure may also
induce accumulation of neurotoxic substances, such as
antibiotics and hypnotics. It might also be that unknown
factors caused both SAE and renal dysfunction without
any reciprocal influence. Unfortunately, whether correc-
tion of renal failure is associated with a reduction in the
occurrence or in the duration of SAE could not be deter-
mined from our study.
Our results suggest that “systemic insults” (i.e., hyper-
capnia, hypernatremia, hypoglycemia, and persistent
hyperglycemia) might play a role in the pathophysiology
 1081

Table 4 Cox multivariate analysis of  factors associated and survival. Interestingly, a recent multicenter placebo-
with mortality controlled trial conducted in patients with severe sepsis
Variable HR 95% CI p value found that use of hydrocortisone at sepsis onset resulted
in a lower incidence of delirium at 28  days, suggesting
Age, per 1-year increment 1.03 1.02 1.03 <0.01 neuroprotection [32]. We also observed a significant
Chronic immunodepression 1.58 1.34 1.87 <0.01 association between SAE and use of common medica-
Chronic cardiac disease 1.33 1.10 1.61 0.003 tions with potential neurotoxic effects, such as propofol,
Chronic respiratory disease 1.19 0.96 1.48 0.10 midazolam, and opioids. However, temporality between
Chronic liver disease 1.92 1.54 2.39 <0.01 exposure to medications and SAE on the day of ICU
Year of admission admission could not be accurately determined, preclud-
 <2006 Ref. <0.01 ing the inclusion of these variables in our multivariate
 2006–2008 0.94 0.76 1.16 model.
 2009–2010 0.67 0.52 0.85 Contrary to previously published data [16], we found
 >2010 0.84 0.67 1.06 that S.  aureus was the only pathogen associated with
Non-neurological SOFA score, per 1 1.10 1.08 1.12 <0.01 SAE, irrespective of infection source or bacteremia.
point increment
Pathophysiology of S.  aureus associated with SAE is
Sepsis-associated encephalopathy, score on GCS
probably multifactorial. For instance, staphylococcal
 GCS 3–8 3.37 2.82 4.03 <0.01
toxin α-hemolysin has been identified as neurotoxic [33].
 GCS 9–12 1.80 1.41 2.29 Staphylococcus aureus bacteremia is associated with an
 GCS 13–14 1.38 1.09 1.76 increased risk of arterial thromboembolic events, which
 GCS 15, features of delirium 1.06 0.80 1.41 could have an impact on neurological status [34]. Staphy-
 GCS 15, no feature of delirium Ref. lococcus aureus is also an independent risk factor for
Variables included in the multivariate model had no missing data
neurologic complication in patients with severe infective
HR hazard ratio, CI confidence interval, GCS Glasgow coma scale, SOFA sequential endocarditis [35]. However, the respective roles played by
organ failure assessment these factors cannot be disentangled by our data, and this
question deserves further investigation.
Eidelman et  al. showed in a landmark study that SAE
of SAE [15]. Blood carbon dioxide tension is a well- was observed in 54% of non-sedated septic patients at
known determinant of cerebral blood flow, and hyper- ICU admission, with a mortality rate of 63% when the
capnia is likely to induce significant brain vasodilation score on the GCS dropped below 8 [1]. Since then, lim-
and increased intracranial pressure as a consequence in ited data has been published on the epidemiology of
septic patients. Hypernatremia has also been previously SAE [16]. Moreover, progress in critical care and sep-
described as an independent risk factor for agitated delir- sis management over the past 20 years has led to a dra-
ium and mortality in ICU patients [26]. Neuropathologi- matic decrease of sepsis mortality [36]. Our multicenter
cal studies conducted in non-survivors of critical illness study, which provides a valuable update on outcomes of
and in experimental models highlighted the detrimental SAE, shows that mortality of septic patients comatose at
role of uncontrolled hyperglycemia on neurons and glial admission has not decreased significantly.
cells [27–30]. Common classes of medications may also The strengths of our study include a large multicenter
play a role in the pathophysiology of SAE, including anti- population from a high-quality database with severe
biotics and steroids. In our study, use of fluoroquinolo- sepsis or septic shock patients of various origins. In this
nes tended to be associated with SAE, in line with recent regard, our study may represent one of the largest to date
reviews highlighting their neurotoxicity [14]. However, a on risk factors and outcomes of SAE. The large sample
true causal relationship between antibiotic exposure and size of the study allowed for adjustment with risk factors
SAE also remains to be demonstrated. As neurological for delirium in the ICU [17].
side effects of antibiotics are frequently associated with Our study has important limitations. First, a causal
overdosage, systematic therapeutic drug monitoring may relationship between the identified risk factors and SAE
be proposed in patients with SAE [31]. In our study, acute cannot be determined from this observational study. Sec-
use of steroids was associated with SAE. On the one ond, the definition of SAE may not meet the most recent
hand, this association may be explained by well-known validated criteria for delirium (DSM-5). Of note, the
side effects of steroids on the brain, including delirium proportion of patients with mild neurological dysfunc-
and/or agitation. On the other hand, SAE is character- tion represented a minority of patients in our study, pre-
ized by excessive and prolonged neuroinflammation cluding analysis of the attributable mortality of delirium
that may have deleterious effects on neuronal function in the context of sepsis [37]. A prospective study using
1082
appropriate tools for delirium detection is warranted.
Third, SAE was defined according to the score on the
GCS, which may be confounded by several parameters in
ICU patients, including invasive mechanical ventilation,
sedative, and neuromuscular blocking agents. However,
the strong association between severity of encephalopa-
thy and outcomes of our study is similar to that found
in a previous study conducted in non-sedated patients
with sepsis [1]. Moreover, this association remained
significant after adjustment for other predictors of out-
come in the ICU population. Fourth, we did not analyze
the impact of transient (1  day) or persistent (more than
1 day) SAE. Fifth, we focused our analysis on the occur-
rence of SAE at ICU admission, and thus did not evaluate
the impact of SAE developing during ICU stay. Sixth, the
absence of neuroimaging data in our study does not allow
for assessment of the impact of brain injury on outcome.
Brain MRI is obviously relevant for prognosis, as MRI
studies conducted in survivors of critical illness dem-
onstrated important structural changes, including brain
volume reduction and white matter disruption that cor-
related with cognitive outcomes [11–13]. Finally, long-
term follow-up was not available in the OUTCOMEREA
database. Therefore, the exact impact of SAE on long-
term neurologic outcomes remains to be investigated.
In conclusion, in this retrospective analysis of a pro-
spective multicenter database, we found that half of
patients admitted to the ICU with sepsis had encephalop-
athy. The identified subset of patients more susceptible
to SAE was older patients with a history of chronic alco-
hol abuse, neurological disease, pre-existing cognitive
impairment, and long-term use of psychoactive drugs.
Potentially modifiable factors associated with SAE at
ICU admission included acute renal failure and common
metabolic disturbances, namely hypoglycemia, hypergly-
cemia, hypercapnia, and hypernatremia. Although these
factors are likely to play a key role in SAE pathophysiol-
ogy, the existence of a true causal relationship remains
to be investigated. Finally, our study confirms the prog-
nostic significance of mild alteration of mental status in
patients with sepsis.
Electronic supplementary material
The online version of this article (doi:10.1007/s00134-017-4807-z) contains
supplementary material, which is available to authorized users.
Author details
1 Melun); Jean-Baptiste Joffredo (ICU, G Montpied Hospital, Clermont-Ferrand,
 Department of Intensive Care Medicine and Infectious Diseases, Hôpital
Bichat-Claude Bernard, 46 rue Henri Huchard, 75877 Paris Cedex, France.
2 laume Lacave (ICU, Versailles Hospital, Versailles, France), Alexandre Lautrette
 UMR1148, LVTS, Sorbonne Paris Cité, Inserm/Paris Diderot University,
Paris, France. 3 UMR 1137, IAME Team 5, DeSCID: Decision Sciences in Infec‑
tious Diseases, Control and Care, Sorbonne Paris Cité, Inserm/Paris Diderot
University, Paris, France. 4 Medical‑Surgical Intensive Care Unit, Delafontaine
Hospital, Saint‑Denis, France. 5 Medical‑Surgical Intensive Care Unit, Saint-
Joseph Hospital Network, Paris, France. 6 Medical Intensive Care Unit, Gabriel
Montpied University Hospital, Clermont‑Ferrand, France. 7 Medical Intensive
Care Unit, AP-HP, Saint Louis Hospital, Paris, France. 8 Medical Intensive Care
Unit, Saint Etienne University Hospital, Saint‑Etienne, France. 9 Jacques Lisfranc
Medicine University, Jean Monnet University, Saint‑Etienne, France. 10 Medical
Intensive Care Unit, Lyon University Hospital, Lyon, France. 11 Medical Intensive
Care Unit, La Source Hospital, Orléans, France. 12 Medical‑Surgical Intensive
Care Unit, Gonesse Hospital, Gonesse, France. 13 Surgical ICU, Edouard Herriot
University Hospital, Lyon, France. 14 Medical‑Surgical Intensive Care Unit, AP-
HP, Antoine Béclère University Hospital, Clamart, France. 15 Medical‑Surgical
Intensive Care Medicine Unit, Dourdan Hospital, Dourdan, France. 16 Medical
Intensive Care Unit, André Mignot Hospital, Versailles, France.
Acknowledgements
Members of the OUTCOMEREA Study Group.
Scientific committee: Jean-François Timsit (Medical and Infectious Diseases
ICU, Bichat-Claude Bernard Hospital, Paris, France; UMR 1137 Inserm–Paris
Diderot university IAME, 75018, Paris); Elie Azoulay (Medical ICU, Saint Louis
Hospital, Paris, France); Maïté Garrouste-Orgeas (ICU, Saint-Joseph Hospital,
Paris, France); Jean-Ralph Zahar (Infection Control Unit, Angers Hospital,
Angers, France); Christophe Adrie (Physiology, Cochin Hospital, Paris, France);
Michael Darmon (Medical ICU, Saint Etienne University Hospital, St Etienne,
France); and Christophe Clec’h (ICU, Avicenne Hospital, Bobigny, and UMR
1137 Inserm –Paris Diderot university IAME, 75018, Paris, France).
Biostatistical and information system expertise: Jean-Francois Timsit (Medical
and Infectious Diseases ICU, Bichat-Claude Bernard Hospital, Paris, France; UMR
1137 Inserm–Paris Diderot university IAME, 75018, Paris); Corinne Alberti (Med‑
ical Computer Sciences and Biostatistics Department, Robert Debré Hospital,
Paris, France); Adrien Français (Integrated Research Center U823, Grenoble,
France); Aurélien Vesin (OUTCOMEREA organization and Integrated Research
Center U823, Grenoble, France); Stephane Ruckly (OUTCOMEREA organization
and Inserm UMR 1137 IAME, 75018, Paris); Sébastien Bailly (Grenoble university
hospital Inserm UMR 1137 IAME, 75018, Paris) and Christophe Clec’h (ICU,
Avicenne Hospital, Bobigny, and Inserm UMR 1137 IAME, 75018, Paris, France);
Frederik Lecorre (Supelec, France); Didier Nakache (Conservatoire National
des Arts et Métiers, Paris, France); and Aurélien Vannieuwenhuyze (Tourcoing,
France).
Investigators of the OUTCOMEREA database: Christophe Adrie (ICU, CH
Melun, and Physiology, Cochin Hospital, Paris, France); Carole Agasse (medical
ICU, university hospital Nantes, France); Bernard Allaouchiche (ICU, Pierre
benite Hospital, Lyon, France); Olivier Andremont (ICU, Bichat Hospital, Paris,
France); Pascal Andreu (CHU Dijon, Dijon, France); Laurent Argaud (Medical
ICU, Hospices Civils de Lyon, Lyon, France); Claire Ara-Somohano (Medical ICU,
University Hospital, Grenoble, France); Elie Azoulay (Medical ICU, Saint Louis
Hospital, Paris, France); Francois Barbier (medical-surgical ICU, Orleans, France),
Jean-Pierre Bedos (ICU, Versailles Hospital, Versailles, France); Jérome Bedel
(ICU, Versailles Hospital, Versailles, France), Julien Bohé (ICU, Hôpital Pierre Ben‑
ite, Lyon France), Lila Bouadma (ICU, Bichat Hospital, Paris, France); Noel Brule
(medical ICU, university hospital Nantes, France); Cédirc Brétonnière (medical
ICU, university hospital Nantes, France); Christine Cheval (ICU, Hyeres Hospital,
Hyeres, France); Christophe Clec’h (ICU, Avicenne Hospital, Bobigny, France);
Elisabeth Coupez (ICU, G Montpied Hospital, Clermont-Ferrand, France);
Michael Darmon (ICU, Saint Etienne Hospital, Saint Etienne, France); Etienne
de Montmollin (ICU, Delafontaine Hospital, Saint Denis), Loa Dopeux (ICU, G
Montpied Hospital, Clermont-Ferrand, France); Anne-Sylvie Dumenil (Antoine
Béclère Hospital, Clamart, France); Claire Dupuis (Bichat hospital and UMR
1137 Inserm–Paris Diderot university IAME, 75018, Paris, France), Jean-Marc
Forel (AP HM, Medical ICU, Hôpital Nord Marseille), Marc Gainnier (la Timone
hospital, Marseille, France), Charlotte Garret (medical ICU, university hospital
Nantes, France); Antoine Gros (ICU, Versailles Hospital, Versailles, France),
Akim Haouache (Surgical ICU, H Mondor Hospital, Creteil, France); Romain
Hernu (Medical ICU, Hospices Civils de Lyon, Lyon, France); Samir Jamali (ICU,
Dourdan, Dourdan Hospital, Dourdan, France); Sébastien Jochmans (ICU, CH
France); Hatem Khallel (ICU, Cayenne General Hospital, Cayenne, France); Guil‑
(ICU, G Montpied Hospital, Clermont-Ferrand, France); Virgine Lemiale (Medical
ICU, Saint Louis Hospital, Paris, France); Mathilde Lermuzeaux (ICU, Bichat
Hospital, Paris, France), Guillaume Marcotte (Surgical ICU, Hospices Civils de
Lyon, Lyon, France); Eric Le Miere (ICU, Louis Mourier Hospital, Colombes,
France); Jordane Lebut (ICU, Bichat Hospital, Paris, France); Maxime Lugosi
(Medical ICU, University Hospital Grenoble, Grenoble, France); Eric Magalhaes
(ICU, Bichat Hospital, Paris, France), Sibylle Merceron (ICU, Versailles Hospital,
Versailles, France), Bruno Mourvillier (ICU, Bichat Hospital, Paris, France); Benoît

Misset (ICU, Saint-Joseph Hospital, Paris, France); Delphine Moreau (ICU,
Saint-Louis Hospital, Paris, France); Bruno Mourvillier (ICU, Bichat Hospital, Paris,
France); Mathild Neuville (ICU, Bichat Hospital, Paris, France), Laurent Nicolet
(medical ICU, university hospital Nantes, France); Laurent Papazian (Hopital
Nord, Marseille, France), Benjamin Planquette (pulmonology ICU, George
Pompidou hospital Hospital, Paris, France); Jean-Pierre Quenot Pascal Andreu
(CHU Dijon, Dijon, France); Aguila Radjou (ICU, Bichat Hospital, Paris, France),
Romain Sonneville (ICU, Bichat Hospital, Paris, France), Jean Reignier (medical
ICU, university hospital Nantes, France); Bertrand Souweine (ICU, G Mont‑
pied Hospital, Clermont-Ferrand, France); Carole Schwebel (ICU, A Michallon
Hospital, Grenoble, France); Roland Smonig (ICU, Bichat Hospital, Paris, France);
Gilles Troché (ICU, Antoine Béclère Hospital, Clamart, France); Marie Thuong
(ICU, Delafontaine Hospital, Saint Denis, France); Guillaume Thierry (ICU, Saint-
Louis Hospital, Paris, France); Dany Toledano (ICU, Gonesse Hospital, Gonesse,
France); Guillaume Van Der Meersch –Medical Surgical ICU, university hospital
Avicenne), Marion Venot (Medical ICU, Saint Louis Hospital, Paris, France);
Olivier Zambon (medical ICU, university hospital Nantes, France);.
Study monitors: Julien Fournier, Caroline Tournegros, Stéphanie Bagur,
Mireille Adda, Vanessa Vindrieux, Sylvie de la Salle, Loic Ferrand, Stéphane
Guessens, Helene Merle, Nadira Kaddour, Boris Berthe, Samir Bekkhouche,
Kaouttar Mellouk, Mélaine Lebrazic, Carole Ouisse, Diane Maugars, Sylvie
Conrozier, Igor Theodose, Manal Nouacer, Veronique Deiler, Myriam Moussa,
Atika Mouaci Nassima Viguier and Sophie Letrou.
Authors’ contributions
Design and conduct of the study: R. Sonneville, S. Ruckly, and J.F. Timsit. Col‑
lection of the data: J. Poujade, M. Garrouste-Orgeas, B. Souweine, M. Darmon,
E. Mariotte, L. Argaud, F. Barbier, D. Goldgran-Toledano, G. Marcotte, AS Dume‑
nil, S. Jamali, V. Laurent, B. Mourvillier, and JF. Timsit. Management, analysis and
interpretation of the data: R. Sonneville, E. de Montmollin, S. Ruckly, and J.F.
Timsit. Manuscript preparation: R. Sonneville, E. de Montmollin, S. Ruckly, and
J.F. Timsit. Critical revision and approval of the manuscript: All authors.
Received: 2 December 2016 Accepted: 17 April 2017
Published online: 2 May 2017
References
1. Eidelman LA, Putterman D, Putterman C, Sprung CL (1996) The spectrum
of septic encephalopathy. Definitions, etiologies, and mortalities. JAMA
275:470–473
2. Ebersoldt M, Sharshar T, Annane D (2007) Sepsis-associated delirium.
Intensive Care Med 33:941–950
3. Annane D, Sharshar T (2015) Cognitive decline after sepsis. Lancet Respir
Med 3:61–69
4. Widmann CN, Schewe JC, Heneka MT (2014) Sepsis-associated encepha‑
lopathy versus sepsis-induced encephalopathy–authors’ reply. Lancet
Neurol 13:968–969
5. Stubbs DJ, Yamamoto AK, Menon DK (2013) Imaging in sepsis-associated
encephalopathy–insights and opportunities. Nat Rev Neurol 9:551–561
6. van Gool WA, van de Beek D, Eikelenboom P (2010) Systemic infec‑
tion and delirium: when cytokines and acetylcholine collide. Lancet
375:773–775
7. Sharshar T, Carlier R, Bernard F, Guidoux C, Brouland JP, Nardi O, de la
Grandmaison GL, Aboab J, Gray F, Menon D, Annane D (2007) Brain
lesions in septic shock: a magnetic resonance imaging study. Intensive
Care Med 33:798–806
8. Polito A, Eischwald F, Maho AL, Polito A, Azabou E, Annane D, Chretien F,
Stevens RD, Carlier R, Sharshar T (2013) Pattern of brain injury in the acute
setting of human septic shock. Crit Care 17:R204
9. Smith M, Meyfroidt G (2017) Critical illness: the brain is always in the line
of fire. Intensive Care Med. doi:10.1007/s00134-017-4791-3
10. Sharshar T, Annane D, de la Grandmaison GL, Brouland JP, Hopkinson
NS, Francoise G (2004) The neuropathology of septic shock. Brain Pathol
14:21–33
11. Morandi A, Rogers BP, Gunther ML, Merkle K, Pandharipande P, Girard
TD, Jackson JC, Thompson J, Shintani AK, Geevarghese S, Miller RR 3rd,
Canonico A, Cannistraci CJ, Gore JC, Ely EW, Hopkins RO (2012) The
1083
relationship between delirium duration, white matter integrity, and
cognitive impairment in intensive care unit survivors as determined
by diffusion tensor imaging: the VISIONS prospective cohort magnetic
resonance imaging study. Crit Care Med 40:2182–2189
12. Gunther ML, Morandi A, Krauskopf E, Pandharipande P, Girard TD, Jackson
JC, Thompson J, Shintani AK, Geevarghese S, Miller RR 3rd, Canonico
A, Merkle K, Cannistraci CJ, Rogers BP, Gatenby JC, Heckers S, Gore JC,
Hopkins RO, Ely EW, VISIONS Investigation, VISualizing Icu SurvivOrs
Neuroradiological Sequelae (2012) The association between brain vol‑
umes, delirium duration, and cognitive outcomes in intensive care unit
survivors: the VISIONS cohort magnetic resonance imaging study. Crit
Care Med 40:2022–2032
13. Semmler A, Widmann CN, Okulla T, Urbach H, Kaiser M, Widman G, Mor‑
mann F, Weide J, Fliessbach K, Hoeft A, Jessen F, Putensen C, Heneka MT
(2013) Persistent cognitive impairment, hippocampal atrophy and EEG
changes in sepsis survivors. J Neurol Neurosurg Psychiatry 84:62–69
14. Bhattacharyya S, Darby RR, Raibagkar P, Gonzalez Castro LN, Berkowitz AL
(2016) Antibiotic-associated encephalopathy. Neurology 86:963–971
15. Sonneville R, Verdonk F, Rauturier C, Klein IF, Wolff M, Annane D, Chretien
F, Sharshar T (2013) Understanding brain dysfunction in sepsis. Ann
Intensive Care 3:15
16. Zhang LN, Wang XT, Ai YH, Guo QL, Huang L, Liu ZY, Yao B (2012) Epide‑
miological features and risk factors of sepsis-associated encephalopathy
in intensive care unit patients: 2008-2011. Chin Med J (Engl) 125:828–831
17. Zaal IJ, Devlin JW, Peelen LM, Slooter AJ (2015) A systematic review of risk
factors for delirium in the ICU. Crit Care Med 43:40–47
18. Seymour CW, Liu VX, Iwashyna TJ, Brunkhorst FM, Rea TD, Scherag A,
Rubenfeld G, Kahn JM, Shankar-Hari M, Singer M, Deutschman CS,
Escobar GJ, Angus DC (2016) Assessment of clinical criteria for sepsis: for
the third international consensus definitions for sepsis and septic shock
(Sepsis-3). JAMA 315:762–774
19. Lautrette A, Garrouste-Orgeas M, Bertrand PM, Goldgran-Toledano D,
Jamali S, Laurent V, Argaud L, Schwebel C, Mourvillier B, Darmon M,
Ruckly S, Dumenil AS, Lemiale V, Souweine B, Timsit JF, Outcomerea
Study Group (2015) Respective impact of no escalation of treatment,
withholding and withdrawal of life-sustaining treatment on ICU patients’
prognosis: a multicenter study of the Outcomerea Research Group.
Intensive Care Med 41:1763–1772
20. Truche AS, Darmon M, Bailly S, Clec’h C, Dupuis C, Misset B, Azoulay E,
Schwebel C, Bouadma L, Kallel H, Adrie C, Dumenil AS, Argaud L, Marcotte
G, Jamali S, Zaoui P, Laurent V, Goldgran-Toledano D, Sonneville R, Souweine
B, Timsit JF, OUTCOMEREA Study Group (2016) Continuous renal replace‑
ment therapy versus intermittent hemodialysis in intensive care patients:
impact on mortality and renal recovery. Intensive Care Med 42:1408–1417
21. Zahar JR, Timsit JF, Garrouste-Orgeas M, Francais A, Vesin A, Descorps-
Declere A, Dubois Y, Souweine B, Haouache H, Goldgran-Toledano D,
Allaouchiche B, Azoulay E, Adrie C (2011) Outcomes in severe sepsis and
patients with septic shock: pathogen species and infection sites are not
associated with mortality. Crit Care Med 39:1886–1895
22. Le Gall JR, Lemeshow S, Saulnier F (1993) A new simplified acute physiol‑
ogy score (SAPS II) based on a European/North American multicenter
study. JAMA 270:2957–2963
23. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H,
Reinhart CK, Suter PM, Thijs LG (1996) The SOFA (sepsis-related organ fail‑
ure assessment) score to describe organ dysfunction/failure. On behalf of
the Working Group on Sepsis-Related Problems of the European Society
of Intensive Care Medicine. Intensive Care Med 22:707–710
24. Timsit JF, Fosse JP, Troche G, De Lassence A, Alberti C, Garrouste-Orgeas
M, Bornstain C, Adrie C, Cheval C, Chevret S, OUTCOMEREA Study Group,
France (2002) Calibration and discrimination by daily logistic organ dys‑
function scoring comparatively with daily sequential organ failure assess‑
ment scoring for predicting hospital mortality in critically ill patients. Crit
Care Med 30:2003–2013
25. Teasdale G, Maas A, Lecky F, Manley G, Stocchetti N, Murray G (2014) The
Glasgow coma scale at 40 years: standing the test of time. Lancet Neurol
13:844–854
26. Jaber S, Chanques G, Altairac C, Sebbane M, Vergne C, Perrigault PF,
Eledjam JJ (2005) A prospective study of agitation in a medical-surgical
ICU: incidence, risk factors, and outcomes. Chest 128:2749–2757
27. Sonneville R, den Hertog HM, Guiza F, Gunst J, Derese I, Wouters PJ,
Brouland JP, Polito A, Gray F, Chretien F, Charlier P, Annane D, Sharshar
1084
T, Van den Berghe G, Vanhorebeek I (2012) Impact of hyperglycemia on
neuropathological alterations during critical illness. J Clinic Endocrinol
Metab 97:2113–2123
28. Sonneville R, Derese I, Marques MB, Langouche L, Derde S, Chatre L,
Chretien F, Annane D, Sharshar T, Van den Berghe G, Vanhorebeek I
(2015) Neuropathological correlates of hyperglycemia during prolonged
polymicrobial sepsis in mice. Shock 44:245–251
29. Sonneville R, Vanhorebeek I, den Hertog HM, Chretien F, Annane D, Shar‑
shar T, Van den Berghe G (2015) Critical illness-induced dysglycemia and
the brain. Intensive Care Med 41:192–202
30. Polito A, Brouland JP, Porcher R, Sonneville R, Siami S, Stevens RD, Gui‑
doux C, Maxime V, de la Grandmaison GL, Chretien FC, Gray F, Annane
D, Sharshar T (2011) Hyperglycaemia and apoptosis of microglial cells in
human septic shock. Crit Care 15:R131
31. De Waele JJ, Lipman J, Akova M, Bassetti M, Dimopoulos G, Kaukonen
M, Koulenti D, Martin C, Montravers P, Rello J, Rhodes A, Udy AA, Starr T,
Wallis SC, Roberts JA (2014) Risk factors for target non-attainment during
empirical treatment with beta-lactam antibiotics in critically ill patients.
Intensive Care Med 40:1340–1351
32. Keh D, Trips E, Marx G, Wirtz SP, Abduljawwad E, Bercker S, Bogatsch H,
Briegel J, Engel C, Gerlach H, Goldmann A, Kuhn SO, Huter L, Meier-
Hellmann A, Nierhaus A, Kluge S, Lehmke J, Loeffler M, Oppert M, Resener
K, Schadler D, Schuerholz T, Simon P, Weiler N, Weyland A, Reinhart
K, Brunkhorst FM, SepNet–Critical Care Trials Group (2016) Effect of
hydrocortisone on development of shock among patients with severe
sepsis: the HYPRESS randomized clinical trial. JAMA 316:1775–1785
33. Dahlberg D, Mariussen E, Goverud IL, Tonjum T, Maehlen J, Antal EA,
Hassel B (2015) Staphylococcal alpha-hemolysin is neurotoxic and causes
lysis of brain cells in vivo and in vitro. Neurotoxicology 48:61–67
34. Mejer N, Gotland N, Uhre ML, Westh H, Schonheyder HC, Petersen A,
Jensen AG, Larsen AR, Skov R, Benfield T, Danish Staphylococcal Bactere‑
mia Study Group (2015) Increased risk of arterial thromboembolic events
after Staphylococcus aureus bacteremia: a matched cohort study. J Infect
71:167–178
35. Sonneville R, Mirabel M, Hajage D, Tubach F, Vignon P, Perez P, Lavoue
S, Kouatchet A, Pajot O, Mekontso Dessap A, Tonnelier JM, Bollaert PE,
Frat JP, Navellou JC, Hyvernat H, Hssain AA, Tabah A, Trouillet JL, Wolff M,
ENDOcardite en REAnimation Study Group (2011) Neurologic complica‑
tions and outcomes of infective endocarditis in critically ill patients: the
ENDOcardite en REAnimation prospective multicenter study. Crit Care
Med 39:1474–1481
36. Kaukonen KM, Bailey M, Suzuki S, Pilcher D, Bellomo R (2014) Mortality
related to severe sepsis and septic shock among critically ill patients in
Australia and New Zealand, 2000–2012. JAMA 311:1308–1316
37. Klein Klouwenberg PM, Zaal IJ, Spitoni C, Ong DS, van der Kooi AW,
Bonten MJ, Slooter AJ, Cremer OL (2014) The attributable mortality of
delirium in critically ill patients: prospective cohort study. BMJ 349:g6652