• Hervé Dupont, Yazine Mahjoub, Taieb Chouaki, Emmanuel Lorne, and Elie Zogheib.
• 1Department of Anesthesiology and Critical Care Medicine, CHU Amiens Picardie, Amiens, France. 2INSERM UMR 1088, University of Picardy Jules Verne, Amiens, France. 3Mycology Laboratory, CHU Amiens Picardie, Amiens, France.
• Crit. Care Med. 2017 Nov 1; 45 (11): 1937-1945.

ObjectivesThe aim of this study was to identify the impact of antifungal prevention in critically ill immunocompetent adult patients on mortality and subsequent infection.Data SourcesA systematic review and meta-analysis of randomized controlled trials comparing any antifungal use versus placebo to prevent candidiasis in ICU patients were performed.Study SelectionSearches were performed on PubMed, Embase, Scopus, main conference proceedings, and ClinicalTrials.gov, as well as reference lists.Data ExtractionThe primary outcomes were mortality and invasive candidiasis. The secondary outcome was the rate of Candida albicans and nonalbicans strains after treatment. A random effect model was used, and sensitivity analysis was performed for both outcomes. Results are expressed as risk ratios and their 95% CIs.Data SynthesisNineteen trials (10 with fluconazole, four with ketoconazole, one with itraconazole, three with micafungin, and one with caspofungin) including 2,792 patients were identified. No individual trial showed a decreased mortality rate. Combined analysis showed that preventive antifungal did not decrease mortality (risk ratio, 0.88; 95% CI, 0.74-1.04; p = 0.14) but significantly decreased secondary fungal infections by 50% (risk ratio, 0.49; 95% CI, 0.35-0.68; p = 0.0001). No shift across nonalbicans strains was observed during treatment (risk ratio, 0.62; 95% CI, 0.19-1.97; p = 0.42). However, publication biases preclude any definite conclusions for prevention of infection.ConclusionsAntifungal prevention of systemic candidiasis in immunocompetent critically ill adults did not reduce mortality and may have decreased secondary fungal infection rates. However, significant publication bias was present.

10 keywords...

hide…