We conducted a randomized prospective trial in 19 disease-free soft tissue sarcoma patients with doxorubicin-induced cardiomyopathy identified by ECG radionuclide angiography at rest and during exercise to determine the efficacy of the free radical scavenger, N-Acetyl Cysteine (NAC), in reversing the drug's cardiotoxic effect. Of the 19 patients, 11 received oral NAC (5.5 gm/m2 daily for 30 days) and eight patients served as controls. Patients were stratified for age less than greater than 45 years, time from final dose of doxorubicin to randomization less than greater than 8 months, and history of treatment with mediastinal irradiation. The two groups were well-matched for all parameters. Cumulative mean doxorubicin dose (523 mg/m2 and 532 mg/m2) and range 500-600 mg/m2 was comparable. Left ventricular (LV) ejection fraction before randomization was not significantly different between the two groups either at rest (39 +/- 10% control, 38 +/- 13% NAC) or during exercise (38 +/- 12% control, 35 +/- 11% NAC). Neither rest nor exercise ejection fraction values changed significantly in either group between prerandomization and 1-month postrandomization studies. Late studies performed in seven NAC patients 3-5 months after randomization revealed no difference in LV ejection fraction compared to 1-month postrandomization values. Clinical course in patients with overt congestive heart failure was similar in both groups. LV function did not return to normal in any patient in either group. We conclude that N-Acetyl Cysteine has no effect in reversing long standing doxorubicin-induced cardiomyopathy.