ABSTRACT
BACKGROUND: Our objective was to systematically review the effect of pharmacological therapies on mortality in patients with acute respiratory distress syndrome (ARDS), focusing on randomized controlled trials (RCTs) published since a previous review in 2004. METHODS: We updated previous searches and searched OVID versions of MEDLINE, EMBASE and CENTRAL (to January 2013) and proceedings from conferences and bibliographies of included studies. We included RCTs of pharmacologic therapies compared with placebo or no therapy for adult patients with ARDS, using authors' definitions, which reported on mortality (≤ 3 months after randomization). We excluded subgroups of patients with ARDS reported in RCTs enrolling other populations and RCTs of therapies to prevent ARDS, nutritional or fluid interventions, inhaled nitric oxide, therapies coupled to a mechanical ventilation strategy, or oxygen. Two reviewers independently screened citations, selected articles for inclusion, and abstracted clinical and methodological data from included studies with disagreements resolved by a third reviewer. Mortality data were pooled using random-effects models. RESULTS: From 13461 citations, 58 trials (6635 patients) of 21 classes of medications met selection criteria; 26 trials (3880 patients) were published after 2003. Meta-analyses showed reduced 28-day mortality with a 48-hour infusion of cis-atracurium in early ARDS (relative risk 0.66, 95% confidence interval 0.50 to 0.87; 431 patients, 138 deaths). There was no effect on mortality with granulocyte-macrophage colony stimulating factor, late low-dose methylprednisolone, neutrophil elastase inhibitors, intravenous salbutamol, surfactant, or N-acetylcysteine; each meta-analysis included ≥ 1 trial published after 2003. Seven single trials of other treatments published after 2003 showed no effect. Meta-analysis of older trials of prostaglandin E1 also showed no effect. CONCLUSION: Effective pharmacotherapy for ARDS remains extremely limited. Cis-atracurium is a promising treatment for early moderate-severe ARDS (using Berlin definition nomenclature) and merits further investigation in a large RCT.
