Endocardial mononuclear cell infiltrates were studied in 2,350 consecutive biopsies from 172 patients over a period ranging from 4 to 16 months post cardiac transplantation. The patients, otherwise unselected, were equally subdivided into four groups based upon the specific type of maintenance immunosuppression used. This was to allow for comparison of the effects of four separate commonly used recipient immunosuppression protocols, which could potentially influence the characteristics of the infiltrates. With azathiaprine-corticosteroid immunosuppression, endocardial infiltrates in otherwise normal biopsies were exceedingly rare, very minor, and invariably unifocal. Mild and moderate rejection were associated with a highly significant stepwise increase in incidence, prominence, and multifocality of endocardial infiltrates. In contrast, with each of the three cyclosporine-based recipient immunosuppression protocols which were evaluated, approximately 15% of biopsies with no evidence of rejection were associated with endocardial infiltrates. There was a wide range of variation in the prominence of the endocardial infiltrates present. Multifocal infiltrates were frequently encountered, the incidence of which was exclusively dependent upon the specific cyclosporine-based immunosuppression protocol used. With mild and with moderate rejection there was a significant stepwise increase in overall biopsy incidence of all endocardial infiltrates in each of the three groups, although there was no variation in relative prominence of the infiltrates, or in incidence of multifocality when biopsies without rejection were compared. The presence of conspicuous vascular spaces within endocardial infiltrates and significant extension of the infiltrates into the adjacent myocardium, with or without associated myofiber necrosis, were characteristic features of the most prominent endocardial infiltrates in all three cyclosporine-based immunosuppression groups. This constellation of changes has sometimes been referred to as "Quilty" effect. The relative incidence with which these particular features were encountered in association with endocardial infiltrates did not vary with rejection category of the biopsies. This study has shown that the presence of all forms of endocardial infiltrates, in the absence of concomitant rejection, is a characteristic manifestation of cyclosporine-based recipient immunosuppression, regardless of the specific protocol and cyclosporine dosage schedule. Under azathiaprine-based immunosuppression, endocardial infiltrates are almost invariably associated with rejection. It is postulated that cyclosporine-related endocardial mononuclear cell infiltration, in the absence of overt rejection, may result from a low level alloimmune response secondary to fluctuations in cyclosporine drug levels or related factors, and that the incidence with these infiltrates occur can be augmented during acute rejection episodes when the strength of the recipient immune response is magnified.