BACKGROUND The prevalence of nephrotoxicity in neonates receiving amphotericin B deoxycholate (amphoB) is not well-defined. While some studies report a lack of toxicity, others claim a frequency as high as 85%. METHODS We reviewed medical records of all infants < or = 90 days of age in the neonatal intensive care unit who received at least 3 doses of amphoB between January 1990 and December 2004. A standardized form was used to collect demographic, therapeutic, microbiologic, and laboratory data for each patient. Nephrotoxicity was defined as a rise in serum creatinine (SCr) of at least 0.4 mg/dL any time during amphoB therapy. RESULTS A total of 92 infants met entry criteria. Median gestational age was 26 (range: 23-41) weeks and median birth weight was 863 (range: 546-4000) grams. Overall, 15 (16%) infants experienced nephrotoxicity, and 16 (17%) developed hypokalemia (<3.0 mmol/L). There were no differences between infants who did or did not develop nephrotoxicity in terms of gestational age, birth weight, gender, underlying medical conditions, or use of other potentially nephrotoxic medications. AmphoB exposure and duration of therapy were similar between infants who developed nephrotoxicity and those who did not, with a mean cumulative dose of 13.5 +/- 9.6 mg/kg and duration of 16.3 +/- 10.4 days. With the exception of 1 infant, the elevated SCr values resolved in all infants by the end of amphoB therapy. CONCLUSIONS AmphoB administration does not appear to be associated with lasting measurable nephrotoxicity in neonates. Because of changes in serum creatinine and potassium, renal function and potassium levels should be monitored closely in infants receiving amphoB.