Design: Cochrane systematic review of RCTs and quasi-RCTs
Population: 229 newborns with suspected or confirmed invasive fungal infection (6 trials)
Primary outcome: All-cause death before discharge
Effect: All comparisons very low certainty; e.g. caspofungin vs amphotericin B RR 0.43 (0.13–1.48)

The updated Cochrane review of antifungal treatment for newborns with suspected or confirmed invasive fungal infection found six randomised trials with 229 infants, published this week. Comparisons were amphotericin B against fluconazole, micafungin or caspofungin, and micafungin against fluconazole.

Every result was very low certainty. For death before discharge, caspofungin versus amphotericin B gave RR 0.43 (95% CI 0.13 to 1.48), micafungin versus amphotericin B RR 1.20 (0.29 to 4.90) and fluconazole versus amphotericin B RR 0.73 (0.26 to 2.05). No trial reported neurodevelopmental outcome.

The choice of agent in neonatal candidiasis therefore rests on pharmacology, local resistance and toxicity rather than comparative trial evidence. Local NICU susceptibility data matter more than any ranking of agents.

Base empirical antifungal choice on local NICU susceptibility data.
Check for meningitis and end-organ involvement, which shape drug choice.
Monitor renal function and electrolytes with amphotericin B.
Remove central lines where feasible in confirmed candidaemia.

Why it matters

Common NICU antifungal choices have almost no comparative trial evidence behind them.

The statistics, in plain English

Every interval here is very wide and crosses 1, so each comparison is compatible with large benefit or large harm. With 23 to 86 infants per comparison, the trials were far too small to answer the question.