The incidence of acute humoral rejection (AHR) in renal allograft biopsies has been difficult to determine because widely accepted diagnostic criteria have not been established. C4d deposition in peritubular capillaries (PTC) of renal allografts has been proposed as a useful marker for AHR. This study was designed to test the relative value of C4d staining, histology, and serology in the diagnosis of AHR. Of 232 consecutive kidney transplants performed at a single institution from July 1995 to July 1999, all patients (n = 67) who developed acute rejection within the first 3 mo and had a renal biopsy with available frozen tissue at acute rejection onset, as well as posttransplant sera within 30 d of the biopsy, were included in this study. Hematoxylin and eosin and periodic acid-Schiff stained sections were scored for glomerular, vascular, and tubulointerstitial pathology. C4d staining of cryostat sections was done by a sensitive three-layer immunofluorescence method. Donor-specific antibodies (DSA) were detected in posttransplant recipient sera using antihuman-globulin-enhanced T cell and B cell cytotoxicity assays and/or flow cytometry. Widespread C4d staining in PTC was present in 30% (20 of 67) of all acute rejection biopsies. The initial histologic diagnoses of the C4d(+) acute rejection cases were as follows: AHR only, 30%; acute cellular rejection (ACR) and AHR, 45%; ACR (CCTT types 1 or 2) alone, 15%; and acute tubular injury (ATI), 10%. The distinguishing morphologic features in C4d(+) versus C4d(-) acute rejection cases included the following: neutrophils in PTC, 65% versus 9%; neutrophilic glomerulitis, 55% versus 4%; neutrophilic tubulitis, 55% versus 9%; severe ATI, 75% versus 9%; and fibrinoid necrosis in glomeruli, 20% versus 0%, or arteries, 25% versus 0%; all P < 0.01. Mononuclear cell tubulitis was more common in the C4d(-) group (70% versus 100%; P < 0.01). No significant difference between C4d(+) and C4d(-) acute rejection was noted for endarteritis, 25% versus 32%; interstitial inflammation (mean % cortex), 27.2 +/- 27% versus 38 +/- 21%; interstitial hemorrhage, 25% versus 15%; or infarcts, 5% versus 2%. DSA were present in 90% (18 of 20) of the C4d(+) cases compared with 2% (1 of 47) in the C4d(-) acute rejection cases (P < 0.001). The pathology of the C4d(+) but DSA(-) cases was not distinguishable from the C4d(+), DSA(+) cases. The C4d(+) DSA(-) cases may be due to non-HLA antibodies or subthreshold levels of DSA. The sensitivity of C4d staining is 95% in the diagnosis of AHR compared with the donor-specific antibody test (90%). Overall, eight grafts were lost to acute rejection in the first year, of which 75% (6 of 8) had AHR. The 1-yr graft failure rate was 27% (4 of 15) for those AHR cases with only capillary neutrophils versus 40% (2 of 5) for those who also had fibrinoid necrosis of arteries. In comparison, the 1-yr graft failure rates were 3% and 7%, respectively, in ACR 1 (Banff/CCTT type 1) and ACR 2 (Banff/CCTT type 2) C4d(-) groups. A substantial fraction (30%) of biopsy-confirmed acute rejection episodes have a component of AHR as judged by C4d staining; most (90%), but not all, have detectable DSA. AHR may be overlooked in the presence of ACR or ATI by histology or negative serology, arguing for routine C4d staining of renal allograft biopsies. Because AHR has a distinct therapy and prognosis, we propose that it should be classified separately from ACR, with further sub-classification into AHR 1 (neutrophilic capillary involvement) and AHR 2 (arterial fibrinoid necrosis).