This study was designed to assess: (1) whether furosemide modifies the incidence of failure to close a symptomatic patent ductus arteriosus (PDA) in response to indomethacin in premature infants, (2) whether furosemide decreases renal and hydromineral side effects of indomethacin, and (3) whether the effects of furosemide on renal function depend on initial extracellular volume [assessed by blood urea nitrogen (BUN)/creatinine ratio]. We did a systematic review and meta-analysis of all published controlled trials assessing either ductal closure or renal function after randomized allocation to treatment with indomethacin and furosemide versus indomethacin alone. All of the three studies meeting entry criteria were small and had methodological limitations. The number of patients was too small to rule out a 10% risk increase in failure of ductal closure. After the first dose of indomethacin, patients receiving furosemide had higher urine output, fractional excretion of sodium, and osmolar clearance than controls. Among patients with initial BUN/creatinine ratio <20, those on furosemide had a higher glomerular filtration rate (GFR) than controls. Among patients with initial BUN/creatinine of 20-30, those on furosemide had a lower GFR than controls. Thus, dehydration appears to be a contraindication for furosemide administration in premature infants treated with indomethacin for symptomatic PDA. The risk-benefit ratio of administering furosemide in well-hydrated patients treated with indomethacin for symptomatic PDA could only be assessed by a large randomized clinical trial.