The natural course of mild acute cardiac allograft rejection (MAR) under cyclosporine-based therapy is generally considered benign, and usually antirejection therapy is not instituted. The present study was undertaken to determine the frequency of and the risk factors for progression of MAR into a clinically significant (moderate or severe) rejection on subsequent endomyocardial biopsy (EMB). Among 167 cardiac recipients, transplanted from 3/1984 to 4/1990, MAR under cyclosporine-based therapy was diagnosed on 220 EMBs. Depending upon the outcome on the subsequent EMB, MAR was categorized as progressive or nonprogressive. This served as the dependent variable for a stepwise logistic regression analysis evaluating 11 covariates as potential risk factors: perioperative antibody prophylaxis (ATG vs. OKT3), maintenance therapy, underlying disease, HLA-mismatches for A- and B + DR-loci, serum creatinine (mg/dl) and cyclosporine HPLC blood level (ng/ml) at diagnosis of MAR and at subsequent biopsy, recipient age, donor age. 40 (18.2%) of 220 MARs became progressive as opposed to 37 (7.3%) of a control cohort of 507 negative EMBs (P less than 0.0001). Stepwise logistic regression yielded the type of maintenance therapy (P = 0.0019) and serum creatinine level at diagnosis of MAR (P = 0.0615) as independent predictors of progression of MAR. After adjustment for influence of maintenance therapy and serum creatinine none of the cyclosporine variables provided any additional information. MARs without maintenance steroids and low serum creatinine levels had the highest risk (37.2% observed incidence) to develop moderate or severe rejection on subsequent EMB. This analysis supports evidence that diagnosis of MAR on EMB is associated with a considerable high progression rate into clinically significant rejection when compared to negative EMBs. Progression particularly occurs in MAR under steroid-free maintenance therapy and suggests early augmentation of immunosuppression. In terms of progression of MAR serum creatinine values, obviously indicating cyclosporine nephrotoxicity, appear to reflect the extent of cyclosporine-mediated immunosuppressive activity more properly than parameters of its bioavailability by measuring cyclosporine HPLC blood levels.