Nine trials (2,029 infants) were included in the review. Seven of the nine trials reported an adequate sequence generation, five reported adequate allocation concealment, four reported blinding on all outcomes, eight addressed incomplete outcome data and were free of selective reporting.
Fluconazole significantly reduced the incidence of invasive fungal infection in very low birthweight babies compared to placebo (RR 0.36, 95% CI 0.15 to 0.89; six trials, 840 infants; significant heterogeneity Ι²=66%). The number needed to treat was nine babies (95% CI 6.7 to 16.3) to prevent one invasive fungal infection. Mortality was not significantly different between treatment groups (RR 0.76, 95% CI 0.54 to 1.08). No significant toxicity was reported from fluconazole and no babies were withdrawn from the studies due to adverse events.
Nystatin significantly reduced the incidence of invasive fungal infection in very low birthweight babies compared to placebo or no drug (RR 0.16, 95% CI 0.11 to 0.23; three trials, 1,200 infants; no significant heterogeneity). The number needed to treat was four babies (95% CI 3.1 to 4.3) to prevent one invasive fungal infection. Mortality was not significantly different between treatment groups (RR 0.86, 95% CI 0.59 to 1.26). No significant toxicity was reported from oral nystatin and no babies were withdrawn from the studies due to adverse events.
There was no significant difference in the incidence of invasive fungal infections (RR 0.54, 95% CI 0.19 to 1.56) or mortality (RR 0.43, 95% CI 0 to 4.31; two trials, 257 infants) when fluconazole was compared with nystatin. No significant toxicity was reported from either agent in the studies reviewed.
The results of the publication bias test were not reported.