ALT indicates alanine aminotransferase; AST, aspartate aminotransferase; CK, creatine kinase;
ICU, intensive care unit; VAP, ventilator-associated pneumonia.
eTable 1. Factors Influencing the Risk of Ventilator-Associated Pneumonia
eTable 2. Diagnosis of Ventilator-Associated Pneumonia and Antibiotic Therapy
eTable 3. Pathogens Identified in the Specimens Collected at Enrollment From
Patients With Probable Ventilator-Associated Pneumonia
eTable 4. Course of Organ Dysfunctions
eTable 5. Infections Within the First 28 days After Enrollment
Papazian L, Roch A, Charles P, Penot-Ragon C, Perrin G, Roulier P, Goutorbe P, Lefrant J, Wiramus S, Jung B, Perbet S, Hernu R, Nau A, Baldesi O, Allardet-Servent J, Baumstarck K, Jouve E, Moussa M, Hraiech S, Guervilly C, Forel J, . Effect of Statin Therapy on Mortality in Patients With Ventilator-Associated
PneumoniaA Randomized Clinical Trial. JAMA. 2013;310(16):1692-1700. doi:10.1001/jama.2013.280031
Copyright 2013 American Medical Association. All Rights Reserved. Applicable
FARS/DFARS Restrictions Apply to Government Use.
Observational studies have reported that statin use may be associated with improved outcomes of
various infections. Ventilator-associated pneumonia (VAP) is the most common infection in the
intensive care unit (ICU) and is associated with substantial mortality.
To determine whether statin therapy can decrease day-28 mortality in patients with VAP.
Design, Setting, and Participants
Randomized, placebo-controlled, double-blind, parallel-group, multicenter trial performed in 26
intensive care units in France from January 2010 to March 2013. For power to detect an 8% absolute
reduction in the day-28 mortality rate, we planned to enroll 1002 patients requiring invasive
mechanical ventilation for more than 2 days and having suspected VAP, defined as a modified Clinical
Pulmonary Infection Score of 5 or greater. The futility stopping rules were an absolute increase in
day-28 mortality of at least 2.7% with simvastatin compared with placebo after enrollment of the
first 251 patients.
Participants were randomized to receive simvastatin (60 mg) or placebo, started on the same day
as antibiotic therapy and given until ICU discharge, death, or day 28, whichever occurred first.
Main Outcomes and Measures
Primary outcome was day-28 mortality. Day-14, ICU, and hospital mortality rates were determined,
as well as duration of mechanical ventilation and Sequential Organ Failure Assessment (SOFA) scores
on days 3, 7, and 14.
The study was stopped for futility at the first scheduled interim analysis after enrollment of
300 patients, of whom all but 7% in the simvastatin group and 11% in the placebo group were naive to
statin therapy at ICU admission. Day-28 mortality was not lower in the simvastatin group (21.2% [95%
CI, 15.4% to 28.6%) than in the placebo group (15.2% [95% CI, 10.2% to 22.1%];
P = .10; hazard ratio, 1.45 [95% CI, 0.83 to 2.51]); the between-group
difference was 6.0% (95% CI, −3.0% to 14.9%). In statin-naive patients, day-28 mortality was
21.5% (95% CI, 15.4% to 29.1%) with simvastatin and 13.8% (95% CI, 8.8% to 21.0%) with placebo
(P = .054) (between-group difference, 7.7% [95%CI, −1.8% to
16.8%). There were no significant differences regarding day-14, ICU, or hospital mortality rates;
duration of mechanical ventilation; or changes in SOFA score.
Conclusions and Relevance
In adults with suspected VAP, adjunctive simvastatin therapy compared with placebo did not
improve day-28 survival. These findings do not support the use of statins with the goal of improving
clinicaltrials.gov Identifier: NCT01057758
Statins are widely used to lower cholesterol levels for prevention of cardiovascular disease.
Statins also exert anti-inflammatory and immunomodulating effects, have been reported to counteract
the deleterious effects of sepsis on coagulation,1,2 and may directly inhibit pathogenic microorganisms.3,4 Many clinical studies have been performed to evaluate whether these
pleiotropic effects are beneficial in infections, although most used an observational design. A
meta-analysis identified 20 studies, including only 1 randomized controlled trial, and found lower
mortality (including pneumonia-related mortality) with statin use, defined as taking any statin for
any reason.5 Another meta-analysis suggested efficacy of
statins for treating and preventing infections,6 but a
meta-analysis including only randomized controlled trials found no evidence of preventive
These data indicate a need for randomized controlled trials in uniform patient populations,
distinguishing de novo from continued statin therapy. In the few such trials available, de novo
atorvastatin therapy decreased the rate of progression to severe sepsis among ward patients with
sepsis8; simvastatin decreased levels of tumor necrosis
factor α and IL-6 in patients with acute bacterial infection9; continued atorvastatin therapy improved survival in patients with severe
sepsis, although de novo atorvastatin therapy did not10; and
de novo simvastatin therapy improved nonpulmonary organ dysfunction without affecting survival in
patients with acute lung injury.11 Although these findings
seem promising, the toxicities and renal excretion of statins in critically ill patients may differ
from those in relatively healthy individuals, and very high peak plasma statin levels have been
reported in critically ill patients.12 Thus, the
risk-benefit ratio of statin therapy in the intensive care unit (ICU) remains unclear.
Ventilator-associated pneumonia (VAP) is diagnosed in approximately 8% to 28% of ICU patients
receiving mechanical ventilation13,14 and remains
associated with increased mortality rates and high health care costs.15 Thus, new adjunctive treatments are needed to improve the outcomes of VAP.
The primary objective of this trial was to determine whether adjunctive statin therapy decreased
day-28 mortality among ICU patients with VAP.
This 1:1 randomized, parallel-group, placebo-controlled, double-blind trial was monitored by an
independent data and safety monitoring board (DSMB). A steering committee designed and executed the
study, analyzed the data, interpreted the findings, wrote the manuscript, and holds the data. The
study was conducted in accordance with the protocol and statistical analysis plan, which were
approved for all centers by the ethics committee of the Nice University Hospital (Comité de
Protection des Personnes Sud Méditerranée V). According to French law, written informed
consent was obtained from the patients or their proxies either before study inclusion or, for
patients not competent to provide consent and with no available proxies, at recovery of
The study was conducted in 26 French ICUs from January 2010 to March 2013. Consecutive adults who
had received mechanical ventilation in the ICU for at least 2 days were eligible if they had
suspected VAP defined as a modified Clinical Pulmonary Infection Score (CPIS)16 of at least 5 and if they underwent quantitative bacteriological cultures of
bronchoalveolar lavage (BAL) fluid, a protected telescopic catheter (PTC), or an endotracheal
aspirate. The modified CPIS16 is based on body temperature,
blood leukocyte count, amount and appearance of tracheal secretions, ratio of
Pao2 to fraction of inspired oxygen, acute respiratory distress syndrome (ARDS),
and infiltrates on chest radiography. The total can range from 1 to 10 points. Patients were
included only for the first episode of suspected VAP. Noninclusion criteria were statin therapy at
intubation, previous VAP episode during the same hospitalization, known pregnancy, immunodepression
with bone marrow aplasia, imminent death (Simplified Acute Physiology Score II of 75 or greater,
calculated over the last 6 hours), treatment-limitation decisions, nothing-by-mouth order and no
nasogastric tube, continuous gastric aspiration, known chronic intestinal malabsorption, known
simvastatin hypersensitivity, acute hepatic failure,17 use
of CYP3A4 inhibitors or cyclosporine, creatine kinase level greater than 5 times the upper limit of
normal, alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels greater than 3
times the upper limit of normal (as recommended by the Agence Nationale de la Sécurité du
Médicament from the French Ministry of Health), and enrollment in another trial within the
previous 30 days.
Patients were randomly allocated to simvastatin (60 mg) or placebo given via a nasogastric tube
or orally from study inclusion to ICU discharge, death, or day 28, whichever occurred first.
Simvastatin or placebo was started on the same day as antibiotic therapy for suspected VAP. The
simvastatin dosage was halved in patients with renal failure (creatinine clearance <30 mL/min).
Randomization was stratified by center. A computer-generated random-number table was prepared by
statisticians to assign patients in blocks of 4 to receive either simvastatin or placebo. Block size
was unknown to the investigators, who enrolled the patients and then called the statistics
department to obtain the randomization and treatment numbers after checking the inclusion and
noninclusion criteria. To ensure blinding, placebo tablets identical to the simvastatin tablets were
manufactured (same appearance, color, and time to dilution in 30 mL of water). Simvastatin or
placebo was administered via the nasogastric tube or orally. Patients, clinicians, evaluators,
monitors, and data analysts were blinded to the study treatment.
Definite bacterial VAP was defined as a positive pleural fluid culture or rapid cavitation of the
lung infiltrate as determined by computed tomography, biopsy, or both, or an autopsy specimen
showing histological evidence of pneumonia (consolidation with large numbers of neutrophils in the
bronchioles and adjacent alveoli in several adjacent low-power fields, with or without tissue
necrosis).18 Probable bacterial VAP was defined as a
modified CPIS of 5 or greater, combined with BAL cultures with greater than 104
colony-forming units (CFU)/mL for at least 1 organism, PTC cultures with greater than 103
CFU/mL for at least 1 organism, or endotracheal aspirate culture with greater than 105
CFU/mL for at least 1 organism. Possible VAP was defined as absence of the above-listed criteria
with a modified CPIS of 5 or greater. Adequate empirical therapy, ARDS, and septic shock are defined
in the eMethods in the Supplement; the eMethods also describe the use of antibiotics.
We recorded demographic data, physiological variables, Simplified Acute Physiology Score II at
admission and radiologic score,19 antibiotics used, and
relevant diagnostic and therapeutic interventions in the ICU. The Sequential Organ Failure
Assessment (SOFA) score20 and CPIS16 were calculated on the day of enrollment (day 1) and then on days 3, 7, and 14.
Patients were monitored daily for evidence of infection. The duration of mechanical ventilation,
length of ICU stay, and length of hospital stay were recorded. The occurrence of myocardial ischemia
or infarction was assessed until day 28. Serum levels of creatine kinase, ALT, and AST were measured
on days 1, 3, 7, 14, and 21.
The primary outcome was the day-28 mortality rate. Secondary outcomes were day-14, day-90
(exploratory), and ICU mortality rates; number of days outside the ICU between day 1 and day 28;
mortality rates in the subgroups with definite and probable VAP; and number of ventilator-free days
(after successful weaning) between day 1 and both day 28 and day 90 (exploratory). Successful
weaning was defined as spontaneous breathing for at least 48 hours after disconnection of the
ventilator. Other secondary outcomes were the SOFA score on days 3, 7, and 14, the occurrence of
subsequent VAP episodes, bacteremia/fungemia, urinary tract infections (exploratory), and
catheter-related infections (exploratory) from inclusion to day 28. All these outcomes were
prespecified in the protocol except day-90 mortality, day-90 ventilator-free days, and incidences of
urinary tract and catheter-related infections from inclusion to day 28, which were exploratory but
nevertheless diagnosed and recorded prospectively. The main safety end points were the number of
adverse events and the proportions of patients with alterations in creatine kinase levels, levels of
ALT and AST, or both.
The DSMB requested a change in the statistical methodology initially chosen for analyzing the
primary outcome. The DSMB further suggested that obtaining valid support for our hypothesis that
simvastatin was superior over the placebo for the main objective required the use of an appropriate
statistical methodology with a unilateral hypothesis and asymmetric boundaries. This change was
made, and the final statistical plan was then approved by the ethics committee in May 2012. Two
interim analyses were planned. Data management and analysis were each conducted in blinded fashion
by 2 separate biostatistician teams.
Assuming a 28-day mortality rate of 30%, we needed 1002 patients (501 in each group) to obtain
80% power for detecting an 8% absolute reduction in the day-28 mortality rate (26.7% relative
reduction) with a 1-sided α risk of .025 (SAS version 9.2 [SAS Institute Inc]).18,21 The study was monitored using group sequential
testing; the stopping rules were efficacy (significant survival improvement with simvastatin) and
futility (low probability of demonstrating such an improvement) according to
O’Brien-Fleming–type asymmetric boundaries. Interim analyses of the primary outcome
(day-28 mortality) were to be performed after enrollment of 251 and 502 patients. Asymmetric
stopping boundaries were designed using α and β spending boundaries as described by Lan
and DeMets.22 The efficacy stopping rules were an absolute
decrease in day-28 mortality of at least 21.3% with simvastatin compared with placebo after
enrollment of 251 patients and an absolute decrease at least 10.6% after enrollment of 502 patients;
for the final analysis, an absolute decrease of at least 5.3% established efficacy. The futility
stopping rules were an absolute increase in day-28 mortality of at least 2.7% with simvastatin
compared with placebo after enrollment of 251 patients and an absolute decrease in day-28 mortality
no greater than 2.6% with simvastatin compared with placebo after enrollment of 502 patients. The
independent DSMB conducted the first interim analysis in December 2012 based on data from the first
Continuous variables are reported as mean (SD) or median (interquartile range) and categorical
variables as number (%). Between-group differences were assessed using t test,
Wilcoxon test, χ2 test, or Fisher exact test, as appropriate. A Cox multivariate
proportional hazards model was built using 5 predefined covariates: simvastatin use, baseline SOFA
score, age, sex, and presence of a fatal underlying disease at hospital admission. We also planned
to include all variables showing an imbalance between the 2 groups at baseline; the only such
variable was antibiotic use within 3 days before inclusion. A Kaplan-Meier curve was plotted for
time to death until day 28.
All analyses were performed using SAS version 9.2 (SAS Institute Inc) on an intention-to-treat
basis. No patients were lost to follow-up for the main outcome or any of the outcomes related to
mortality, duration of mechanical ventilation, stay lengths, or new infections. Variables with
missing values (for determination of the SOFA score and for enzyme levels, both after day 3) were
handled by confining the analyses to those patients with available data.
P ≤ .05 (2-sided) was considered significant, except for the
primary outcome, for which P ≤ .025 (1-sided) was considered
The trial was stopped for futility at the first scheduled interim analysis, based on results
reported by the DSMB after breaking the randomization code.
Between January 2010 and December 2012, we screened 1303 patients for eligibility; of these, 300
were enrolled (Figure 1). All patients received at
least 1 dose of statin or placebo. Consent was denied by 6 patients and withdrawn by 10; these
patients were not included in the analysis. The only clinically significant baseline difference
between groups was a higher proportion of patients receiving antibiotics in the simvastatin group
(Table 1). Only 11 patients (7%) in the simvastatin
group and 15 patients (11%) in the placebo group had received statins in the past month; statins
were stopped at ICU admission (P = .33). VAP risk factors and
preventive measures are listed in eTable 1 in the Supplement. In all patients, simvastatin or placebo was started within
24 hours after the first dose of antibiotics prescribed for VAP.
No patients had definite bacterial VAP. Probable VAP was diagnosed in 106 patients (73%) in the
simvastatin group and 105 (76%) in the placebo group (P = .50) (eTable
2 in the Supplement). The
diagnosis was based mainly on cultures of BAL fluid and PTCs, which recovered the organisms listed
in eTable 3 in the Supplement.
There were no between-group differences in the rates of multidrug-resistant or high-risk organisms.
Adequacy of empirical treatment did not differ significantly between the 2 groups (86% for the
simvastatin group and 77% for the placebo group).
Day-28 mortality was not significantly decreased by simvastatin therapy (21.2% [95% CI, 15.4% to
28.6%] vs 15.2% [95% CI, 10.2% to 22.1%]; P = .10; hazard ratio, 1.45
[95% CI, 0.83-2.51]) (Figure 2). The between-group
difference in day-28 mortality was 6.0% (95% CI, −3.0% to 14.9%). Restricting the analysis to
patients with probable VAP did not change the results: day-28 mortality was 22.6% (95% CI, 15.7% to
31.5%) with simvastatin vs 14.3% (95% CI, 8.9% to 22.2%) with placebo
(P = .06; between-group difference, 8.3% [95% CI, −2.2% to
18.7%]). In the subgroup naive to statins at admission, day-28 mortality was 21.5% (95% CI, 15.4% to
29.1%) with simvastatin and 13.8% (95% CI, 8.8% to 21.0%) with placebo
(P = .054; between-group difference, 7.7% [95% CI, −1.8% to
16.8%]). After adjustment, simvastatin was not significantly associated with day-28 mortality. Age,
fatal underlying disease, and SOFA score were associated with day-28 mortality, but sex and baseline
antibiotic therapy were not.
There were no significant between-group differences for day-14, ICU, or hospital mortality rates;
mechanical ventilation duration; number of ventilator-free days by day 28; coronary events; or ARDS
within 28 days after enrollment (Table 2). Neither
did the groups differ regarding the course of organ dysfunctions (total SOFA score, SOFA subscores,
and CPIS) or the development of kidney dysfunction (eTable 4 in the Supplement). At least 1 new
nosocomial infection occurred after enrollment in 46 patients (31% [95% CI, 25% to 39%]) in the
simvastatin group and 52 (38% [95% CI, 30% to 46%]) in the placebo group (between-group difference,
7% [95% CI, 5% to 17%]; P = .27) (eTable 5 in the Supplement). The mean number of
antibiotic-free days was 12.5 (SD, 8.4) days in the simvastatin group and 12.5 (SD, 8.0) days in the
placebo group (P = .99).
Simvastatin (60 mg/d) was well tolerated, with no increases in rates of elevated creatine kinase,
ALT, or AST levels (Table 3). There were no
between-group differences in creatinine levels at days 3, 7, 14, or 21 (Table 3). No unexpected serious adverse reactions occurred during the
study. The study treatment was interrupted for at least 24 hours because of an adverse event in 55
patients (19%) (31 [21%] in the simvastatin group and 24 [17%] in the placebo group).
The use of simvastatin (60 mg) for the adjunctive treatment of VAP was not associated with a
reduction in day-28 mortality in this trial. None of the secondary outcomes were significantly
improved by simvastatin. Nonetheless, it should be emphasized that, although our study was not
designed to test whether a placebo was superior to simvastatin, there was a nearly 6% absolute
increase in day-28 mortality in the simvastatin group overall and a nearly 8% absolute increase in
the statin-naive subgroup. Because de novo statin therapy does not constitute standard practice in
ICU patients with infection, designing a study to test hypothetical superiority of a placebo over a
statin would not have been relevant. We therefore tested the hypothesis that simvastatin was
superior to a placebo. Our results do not support the use of adjunctive statin therapy in ICU
patients with VAP, and this conclusion probably deserves to be extended to ICU patients with any
type of nosocomial infection.
A large body of experimental data supports the use of statins in sepsis,4,23- 25 and many clinical cohort studies suggest
a role for adjunctive statin therapy in severe infections.26- 30 Studies conducted among
patients with community-acquired pneumonia have produced conflicting results.31- 38 All of these studies used an observational design and evaluated the effects of statins prescribed
for lowering lipid levels.
Three randomized controlled trials8,10,39 evaluated statins in infections, but only 1 was
conducted in ICU patients and its primary outcome was the plasma IL-6 level, which was not
significantly affected by statin therapy.39 To our
knowledge, ours is the first randomized, placebo-controlled, double-blind trial evaluating
adjunctive statin therapy in a specific infection. In a 2-center, randomized, open-label trial,
pravastatin (40 mg/d) was compared with a placebo in patients receiving mechanical ventilation and
having ICU stays longer than 48 hours.40 Six patients
(8.45%) in the pravastatin group and 16 (19.85%) in the control group died during the 30-day
treatment period (P = .06).40
The conflicting results reported to date may be related in part to differences in the proportions of
patients taking statins at baseline, because statin discontinuation in the placebo group might
explain the improved outcomes seen with statin therapy in some studies. Our choice of simvastatin,
the statin on the market the longest, as the study drug may explain the results. We chose
simvastatin because of its well-established immunomodulatory properties.41 Also, in a mouse model of acute Chlamydia pneumoniae
infection, simvastatin decreased viable C pneumoniae counts and increased
inflammatory cell infiltrates in lung tissue, suggesting not only immunomodulatory properties but
also potential antimicrobial effects.42 The dose used in the
present study may have been lower than required. However, in a study of healthy volunteers, plasma
levels of inflammatory mediators were not significantly different between 40-mg and 80-mg
simvastatin doses; consequently, using a higher simvastatin dose would probably not have changed our
Strengths of our study include the multicenter design, which supports the external validity of
our findings. Furthermore, VAP was diagnosed based on criteria used in many previous studies18 requiring bacteriological confirmation by quantitative
However, our study has several limitations. Because of the early trial termination recommended by
the DSMB based on the interim analysis, we cannot completely rule out marginal benefits from
simvastatin therapy in ICU patients with infection, given the CI for the between-group difference.
However, it would have been ethically unacceptable to continue the trial after the interim analysis,
which showed higher day-28 mortality in the simvastatin group, even though the increase was not
statistically significant. The focus on VAP resulted in a uniform patient population, although it
also limited the relevance of our findings to other infections. Statins are available only as oral
preparations, and the pharmacokinetic profile of statins in ICU patients with sepsis,
gastrointestinal intolerance, or both is unclear. However, all statins are absorbed rapidly after
administration, with a time to peak plasma concentration of 4 hours.43,44 Very high plasma atorvastatin concentrations were documented in
ICU patients with sepsis after a single oral dose of 20 mg.12 A study of the pharmacokinetics of simvastatin (60 mg) in ICU patients is in progress. In
addition, our results may have limited relevance to non–ICU-acquired infections. Our patients
received simvastatin (or placebo) after several days in the ICU and had at least 1 organ
dysfunction. In contrast, many of the earlier studies focused on statin therapy to prevent severe
sepsis or to reduce mortality after sepsis related to community-acquired infections.
We assessed levels of creatine kinase, ALT, and AST to evaluate the tolerance of simvastatin (60
mg/d) in ICU patients. No serious adverse events occurred, and the rates of elevated creatine
kinase, ALT, and AST levels were comparable in the 2 groups. Similarly, in a randomized
placebo-controlled trial of simvastatin (80 mg/d) in patients with acute lung injury, the rates of
elevated creatine kinase, ALT, and AST levels were not significantly different between the 2
groups.11 Although simvastatin has been reported to improve
sepsis-induced acute kidney injury via direct effects on the renal vasculature, reversal of tubular
hypoxia, and a systemic anti-inflammatory effect,45 in our
trial renal function was not better in the simvastatin group compared with placebo. The rates of
adverse events requiring treatment discontinuation showed no significant differences between the 2
groups, in keeping with earlier data.11 High simvastatin
doses did not appear to increase the risk of adverse effects in this population.
In adults with suspected VAP, adjunctive simvastatin therapy compared with placebo did not
improve day-28 survival. These findings do not support the use of statins with the goal of improving
Corresponding Author: Laurent Papazian, MD, PhD,
Réanimation des détresses respiratoires et infections sévères, Hôpital
Nord, Chemin des Bourrely, 13015 Marseille, France (email@example.com).
Published Online: October 9, 2013.
Author Contributions: Drs Baumstarck and Jouve
had full access to all of the data in the study and take responsibility for the integrity of the
data and the accuracy of the data analysis.
Study concept and design: Papazian, Roch, Baumstarck, Forel.
Acquisition of data: Roch, Charles, Perrin, Roulier, Goutorbe, Lefrant, Wiramus,
Jung, Perbet, Hernu, Nau, Baldesi, Allardet-Servent, Moussa, Hraiech, Guervilly, Forel.
Analysis and interpretation of data: Papazian, Penot-Ragon, Baumstarck,
Drafting of the manuscript: Papazian, Roulier, Wiramus, Jouve.
Critical revision of the manuscript for important intellectual content: Roch,
Charles, Penot-Ragon, Perrin, Goutorbe, Lefrant, Jung, Perbet, Hernu, Nau, Baldesi,
Allardet-Servent, Baumstarck, Moussa, Hraiech, Guervilly, Forel.
Statistical analysis: Baumstarck, Jouve.
Obtained funding: Papazian.
Administrative, technical, or material support: Papazian, Roch, Penot-Ragon,
Roulier, Moussa, Forel.
Study supervision: Papazian, Roch, Penot-Ragon, Moussa, Forel.
Conflict of Interest Disclosures: All authors
have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr
Wiramus reported receiving travel expenses from Fresenius Kabi and MSD. Dr Jung reported receiving
payment for lectures from Merck France and Pfizer. Dr Guervilly reported receiving travel expenses
from Gilead and MSD. Dr Papazian reported receiving payment for providing expert testimony from
Faron and receiving travel expenses from Air Liquide Santé. No other authors reported
Funding/Support: This study was supported by a
grant from the French Ministry of Health (Programme
Hospitalier de Recherche Clinique 2007).
Role of the Sponsor: The study sponsor (Assistance Publique–Hôpitaux de
Marseille) took full administrative responsibility but had no role in the recruitment of patients;
the management, analysis, or interpretation of the data; or the preparation, review, or approval of
the manuscript. The French Ministry of Health had no role in the design and conduct of the study;
the collection, management, analysis, and interpretation of the data; the preparation, review, or
approval of the manuscript; or the decision to submit the manuscript for publication.
Data and Safety Monitoring Board: J. Chastre (Service de Réanimation
Médicale, Institut de Cardiologie, Hôpital Pitié-Salpêtrière, Paris,
France); J. Pugin (Service de Réanimation, Hôpitaux Universitaires de Genève,
Genève, Suisse) ; P. Auquier (Laboratoire de Santé Publique, Faculté de
Médecine, Aix-Marseille Université, Marseille, France).
Group Information: Members of the STATIN-VAP
Study Group: Corsica: Réanimation CH Ajaccio, Ajaccio (B. Lecomte, M. Mattys);
Réanimation CH Bastia, Bastia (P. Mercury). France: Réanimation
médicale, CHU Hôpital Edouard Herriot (Lyon L. Argaud; S. Conrozier); Réanimation
Centre Ambroise Paré, Marseille (J.-M. Seghboyan); Réanimation médicale, CHU
Hôpital Bocage, Dijon (P.-E. Charles); J.-P. Quenot; T. Devaux Réanimation Centre
Clairval, Marseille (M. Codde); Réanimation chirurgicale, CHU Hôtel Dieu, Clermont-Ferrand
(J.-M. Constantin); Réanimation polyvalente, CH Aix En Provence (O. Baldesi); Réanimation
HIA, Toulon (P. Goutorbet, H. Boret); Réanimation CH Aubagne, Aubagne (C. Agostini, D.
Almosnino, O. Darchen, G. Grossmith); Réanimation chirurgicale Hôpital Saint-Roch, Nice
(C. Ichai); Réanimation CH Cannes, Cannes (A. Freche); Réanimation chirurgicale, CHU
Hôpital Caremeau, Nïmes (J.-Y. Lefrant, G. Louart, L. Muller, S. Lloret); Réanimation
polyvalente CHU Conception, Marseille (F. Heraud); Réanimation CHU Nord, Marseille (M. Leone,
C. Martin, J. Textoris); Réanimation des Urgences et Médicale, CHU Timone, Marseille (G.
Perrin); Réanimation CH Salon de Provence, Salon de Provence (A. Mofredj); Réanimation
Institut Paoli-Calmettes, Marseille (D. Mokart); Réanimation HIA Laveran, Marseille (A. Nau);
Réanimation chirurgicale, CHU Saint-Eloi, Montpellier (F. Belafia, S. Jaber, A. Prades);
Réanimation CH Perpignan, Perpignan (P. Roulier); Réanimation CHU Hôpital Roger
Salengro, Lille (L. Robriquet); Réanimation Hôpital Desbief, Marseille (J.-M. Seghboyan);
Réanimation chirurgicale CHU Pontchaillou, Rennes (P. Seguin, D. Deborah); Réanimation
Médicale, Hôpital Nord, CHU Saint Etienne (P. Dominique); Réanimation des
Détresses Respiratoires et Infections Sévères, Hôpital Nord, Marseille (F.
Xeridat, M. Castanier, M. Adda, S. Dizier); CPCET, Marseille (E. Charles, J. Micaleff); Laboratoire
de Santé Publique, Faculté de Médecine de Marseille (A. Loundou); Direction de la
Recherche et de l’Innovation APHM, Marseille (J.C. Reynier, A. Giuliani, M. Seux, P.
Additional Contributions: We are indebted to Antoinette Wolfe, MD, for assistance in
preparing and reviewing the manuscript. Dr Wolfe is a self-employed independent medical writer with
no affiliations to any public or private institutions or corporations. She was paid for her work on
our manuscript by the ADEREM, a French nonprofit donation-funded organization that supports medical
research. We are also indebted to Didier Raoult, MD, PhD (Unité de Recherche sur les Maladies
Infectieuses et Tropicales Emergentes, Faculté de Médecine, Aix-Marseille
Université), for his helpful comments during the preparation of the study. Dr Raoult received
no compensation for his contributions. We are grateful to all medical staff, staff nurses, and
research nurses at the 26 sites, who contributed greatly to the successful conduct of the study.