BackgroundThe most optimal antifungal agent for empiric treatment of invasive fungal diseases (IFDs) in febrile neutropenia is controversial. Our objective was evaluate the relative efficacy of antifungals for all-cause mortality, fungal infection-related mortality and treatment response in this population.MethodsPubmed, Embase and Cochrane Library were searched to identify randomized controlled trials (RCTs). Two reviewers performed the quality assessment and extracted data independently. Pairwise meta-analysis and network meta-analysis were conducted to compare the antifungals.ResultsSeventeen RCTs involving 4583 patients were included. Risk of bias of included studies was moderate. Pairwise meta-analysis indicated the treatment response rate of itraconazole was significantly better than conventional amphotericin B (RR = 1.33, 95%CI 1.10–1.61). Network meta-analysis showed that amphotericin B lipid complex, conventional amphotericin B, liposomal amphotericin B, itraconazole and voriconazole had a significantly lower rate of fungal infection-related mortality than no antifungal treatment. Other differences in outcomes among antifungals were not statistically significant. From the rank probability plot, caspofungin appeared to be the most effective agent for all-cause mortality and fungal infection-related mortality, whereas micafungin tended to be superior for treatment response. The results were stable after excluding RCTs with high risk of bias, whereas micafungin had the lowest fungal infection-related mortality.ConclusionsOur results highlighted the necessity of empiric antifungal treatment and indicates that echinocandins appeared to be the most effective agents for empiric treatment of febrile neutropenic patients based on mortality and treatment response. However, more studies are needed to determine the best antifungal agent for empiric treatment. Our systematic review has been prospectively registered in PROSPERO and the registration number was CRD42015026629.