Histological and immunohistochemical analyses were made of biopsy specimens from 50 consecutive patients who experienced putative graft rejection. The mean age of the patients was 44.5 years (range, 17-69 years) and 26 were men. There were 67 evaluable allograft specimens, which were grouped according to the histological diagnosis: group 1, acute tubulointerstitial rejection (n = 42); group 2, acute vascular rejection (n = 18); and group 3, diffuse thrombosis (n = 7). Over a follow-up period of 21-57 months, the mean number of rejection episodes was 1.7, 2.8, and 3.3 in groups 1, 2, and 3, respectively. Allograft loss occurred in 7 out of 30, 10 out of 16, and 4 out of 4 patients in groups 1, 2, and 3, respectively. The following histological parameters differed significantly (P < 0.05) among the groups: interstitial edema, congestion of peritubular capillaries, glomerular thrombosis, and glomerular ischemia (group 3 > group 2 > group 1). Interstitial bleeding was seen more often in group 2 and 3 tissues than in group 1 specimens (P < 0.01). Immunohistochemical analyses showed that vascular rejection was associated with WT14 staining for monocytes and macrophages around the tubuli and with interstitial deposition of complement factor 3. With regard to serology, positive anti-endothelial cell antibody-dependent cellular cytotoxicity was associated with vascular rejection and thrombosis of the graft in all patients tested, and with graft loss in 75%. Pre-existent positive anti-IgG immunofluorescence on peritubular capillaries in pretransplant biopsy specimens incubated with patient serum was found in only 3 of the 50 patients, but was associated with graft loss in 2 of the 3. Cytomegalovirus infection was associated with a higher percentage of graft loss. There were significant intergroup differences in panel reactive antibodies before transplantation (P < 0.001), with higher titers in groups 2 and 3. The findings in relation to interstitial rejection are compatible with cellular rejection, while the data on vascular rejection support a humorally mediated pathogenesis.