BACKGROUND: Intravenous (i.v.) hydrocortisone (HC) has been used recently in selected preterm infants for hypotension soon after birth. During the same time period that HC was used, there was a marked increase in the incidence of disseminated candidal infections (DCIs). OBJECTIVE: To determine whether there is an association between DCI in the first 35 days of life and i.v. HC in preterm infants. RESEARCH DESIGN: A hospital case-control study comparing the exposure of HC between preterm infants with DCI and matched infants without DCI. SETTING: A tertiary level intensive care nursery in a major teaching hospital in San Francisco, CA. PATIENTS: Seventeen preterm infants with DCI and 25 infants without DCI, with gestational age younger than 28 weeks and birth weight less than 1000 g, inborn and outborn admitted to the intensive care nursery between January 1992 and September 1993. METHODS: All preterm infants diagnosed with DCI at younger than 35 days of age were identified using a perinatal and neonatal database. DCI was defined as a blood, cerebrospinal fluid, or two urine cultures positive for Candida requiring antifungal therapy. A control group of uninfected infants matched for the major risk factors for DCI (gestational age, birth weight, duration of intubation, broad-spectrum antibiotics, and i.v. alimentation, including lipids and central venous catheters) admitted during the same period was identified using the same database. Postmatching comparison was performed for several other factors to detect any other differences between the groups. RESULTS: The infants with DCI (n = 17) and control infants (n = 25) had no statistical difference in exposure to the major risk factors for DCI or in postmatching comparison. Ten (59%) of the infants with DCI were receiving HC at the time of infection, whereas four (16%) of the control infants received HC during the first 35 days of life. Infants with DCI were 7.5 times as likely as control infants (95% confidence interval, 5 to 11) to have received IV HC before the onset of fungal infection. CONCLUSION: We conclude that the administration of i.v. HC significantly increases the risk of DCI in susceptible preterm infants younger than 35 days of age. The potentially serious risks of DCI should be considered particularly in the patient selection process for administration of i.v. HC.