A randomized prospective trial for rescue therapy from acute myocardial rejection was undertaken utilizing Minnesota antilymphoblastic globulin (n = 15) versus murine monoclonal anti-CD3 antibody therapy (OKT3) (n = 14). Patients included in the study had moderate rejection unresponsive to bolus high-dose steroid therapy, or moderate-to-severe rejection with hemodynamic instability. Analysis was performed using the t test and chi-square, significance was P less than 0.05. Patient age, sex, interval from transplant to treatment, and number of unresponsive patients vs. hemodynamically unstable patients were similar in both groups (P greater than 0.05). Initial resolution occurred in 9/15 MALG- vs. 14/14 OKT3-treated patients (P = 0.017). Secondary resolution following repeat treatment occurred in 5/6 remaining MALG patients, for a final resolution of 14/15 MALG vs. 14/14 OKT3 patients (P = NS). Rebound rejection was not significantly different (1/14 MALG vs. 4/13 OKT3). However, 7/14 OKT3-treated patients developed life-threatening infections (1 CMV pancreatitis, 2 CMV pneumonias, 1 systemic candidiasis, 3 CMV viremia) vs. 1/15 MALG-treated patients (CMV viremia) (P = 0.014). Death occurred in 4/14 OKT3- (infection) vs. 1/14 MALG- (rejection) treated patients (P = NS). There were no significant differences in the rate of resolution, rebound, infection, or outcome between unresponsive or hemodynamically unstable patients within either group. Although initial rescue is significantly better with OKT3, final resolution is the same in both groups. Since there was a significant incidence of life-threatening infections (7/14) leading to 4 deaths with OKT3 treatment, we recommend MALG for rescue therapy of refractory acute myocardial rejection if this immunosuppressive regimen is to be used.