• Geert Koster, Jørn Wetterslev, Christian Gluud, Jan G Zijlstra, Thomas W L Scheeren, Iwan C C van der Horst, and Frederik Keus.
• Department of Critical Care, University of Groningen, University Medical Centre Groningen, P.O. Box 30.001, 9700 RB, Groningen, The Netherlands, g.koster@umcg.nl.
• Intensive Care Med. 2015 Feb 1; 41 (2): 203-21.

PurposeTo assess the benefits and harms of levosimendan for low cardiac output syndrome in critically ill patients.MethodsWe conducted a systematic review with meta-analyses and trial sequential analyses (TSA) of randomised clinical trials comparing levosimendan with any type of control. Two reviewers independently assessed studies for inclusion. The Cochrane Collaboration methodology was used. Random-effects risk ratios (RR) and 95 % confidence intervals (CI) were derived for the principal primary outcome mortality at maximal follow-up.ResultsA total of 88 trials were included in the systematic review and 49 trials (6,688 patients) in the meta-analysis. One trial had low risk of bias and nine trials (2,490 patients) were considered lower risk of bias. Trials compared levosimendan with placebo, control interventions, and other inotropes. Pooling all trials including heterogenous populations was considered inappropriate. Pooled analysis of 30 trials including critically ill patients not having cardiac surgery showed an association between levosimendan and mortality (RR 0.83, TSA-adjusted 95 % CI 0.59-0.97), while trials with lower risk of bias showed no significant difference (RR 0.83, TSA-adjusted 95 % CI 0.48-1.55). Conventional meta-analysis of all 14 trials including cardiac surgery patients showed an association, while lower risk of bias trials showed no association between levosimendan and mortality (RR 0.52, 95 % CI 0.37-0.73 versus RR 1.02, 95 % CI 0.48-2.16).ConclusionsThe available evidence is inconclusive whether or not levosimendan may have a beneficial effect on mortality due to risks of systematic errors and random errors. Further well-designed randomised trials are needed.

17 keywords...

hide…