Doxorubicin (Adriamycin) was administered by continuous infusion to reduce peak plasma levels and thus lessen cardiac toxicity. Cardiotoxicity was monitored by noninvasive methods, and endomyocardial biopsy specimens were studied by electronmicroscopy. Cardiotoxicity was compared in 21 patients receiving doxorubicin intravenously over 48 or 96 hours and in 30 control patients treated by standard intravenous injection. Both groups were studied prospectively and were well matched by risk factors for doxorubicin cardiotoxicity. The median cumulative dose for those receiving continuous infusion was 600 mg/m2 body surface area (range, 360 to 1500 mg/m2) compared with 465 mg/m2 (range 290 to 680 mg/m2) in the control group (p = 0.002). Fourteen of the 30 patients in the control group showed severe morphologic changes in the biopsy specimens, precluding further doxorubicin administration, as compared with two of 21 patients receiving the drug by continuous infusion (p less than 0.02). The mean pathologic score for the infusion group, 0.9, was lower than the mean for the control group, 1.6 (p = 0.004). Antitumor activity was not compromised. Decreasing peak plasma levels of doxorubicin by continuous infusion reduces cardiotoxicity.