Systemic fungal infections are associated with substantial case-morbidity and fatality rates in premature infants. Considerable evidence indicates that prophylaxis with fluconazole given to premature infants reduces the risk of invasive fungal infection. There is scant information from developing countries. A comparative study of 2 years, one with fluconazole prophylaxis and the other without was conducted in all premature babies weighing less than 1250 g at birth. Fluconazole was administered in 3 mg/kg doses, given every 48 hours, starting on day 3 of life, for a period of 6 weeks. Documented systemic Candida infection was the primary outcome. A total of 271 and 252 patients, respectively, were evaluated during the year before (control group) and after (treatment group) routine fluconazole prophylaxis. The control group developed 21 Candida infections (7.7%) while the treatment group had only 3 Candida infections (1.1%). This difference was statistically significant (P = 0.007; odds ratio, 0.13; 95% confidence interval, 0.03-0.47). The number needed to treat to prevent one case was 7. Although case-fatality rates for documented Candida infection were similar in both periods (76% vs. 67%), fewer deaths attributed to the fungal infection were noted in the prophylaxis year (6% vs. 1%, P = 0.003). Routine fluconazole prophylaxis given to premature infants of less than 1250 g at birth is associated with a significant impact on frequency of documented systemic Candida infections.