To determine the therapeutic range of plasma indomethacin levels for ductus closure, we evaluated the ductus response and renal side effects on two therapeutic regimens using different dosage; regimen I received 0.3 mg/kg q 24 h for a maximum of 3 doses, and regimen II received 0.1, 0.2 and 0.3 mg/kg at 24-hour intervals, for a maximum of 3 doses if needed. Infants in regimen I had significantly higher plasma indomethacin and higher ductus response rate than infants in regimen II. Urine output (U/O) was comparable between the regimens, but serum sodium was lower in regimen I than in regimen II. In both regimens, U/O and serum sodium return to normal by 72 h. The plasma indomethacin levels at 12 h after 1 dose correlated significantly with ductus response and hyponatremia. There appeared to be a minimal level of plasma indomethacin above which U/O decreased; with a plasma level greater than 170 ng/ml the majority (greater than 97%) of infants showed a decrease in U/O. While there was a 50% or greater chance that ductus would close when the plasma levels reached 600 ng/ml or more, a great proportion of infants would also develop renal side effects. Thus, a safe therapeutic range of plasma indomethacin appeared to be very narrow. However, when the dose of indomethacin is increased to optimize constrictive response, there is no significant increase in incidence and severity of renal adverse effects. In view of the transient nature of renal side effects, they should not hinder indomethacin therapy if ductus closure is indicated.