Background. Although C4d deposition in peritubular capillaries has been identified as a strong risk factor for subsequent renal allograft loss, the optimal cutoff for the fraction of peritubular capillaries needed to establish a positive stain in formalin-fixed, paraffin-embedded material has not been defined systematically. The objective of this study was to establish the threshold for positive staining that best predicts renal outcome in renal biopsies in a multicenter study in which local and central pathologic conditions were compared. Methods. Unstained renal biopsy slides were obtained from 296 patients. The percentage of peritubular capillaries staining positively for C4d was detected by immunoperoxidase staining. Results. The percentage C4d deposition ranged from 0% to 90% with 44% (129/296) having a positive percentage of C4d staining. The median for positive cases was 25%. Local C4d+ results were reported qualitatively, with 28% recorded as positive for C4d. Using a centrally determined cutoff of 10%, tests for agreement of local and central C4d staining were fair (k0.40, 95% confidence interval 0.29-0.51). Raising the centrally determined cutoff to 25% or 50% did not change the k values (0.44 and 0.41, respectively). By Cox proportional hazards model, C4dpositivity (centrally determined assessment) using a cutoff of 10% was the strongest predictor of time to graft loss (hazard ratio 2.66, 95% confidence interval 1.68-4.21). Centrally determined C4d positivity correlated with Banff scores indicative of acute inflammation but not with scores indicative of fibrosis/atrophy or transplant glomerulopathy. Conclusions. Our findings indicate that C4d positivity, defined as more than or equal to 10% by immunoperoxidase, is a strong predictor of graft loss. Copyright © 2010 by Lippincott Williams & Wilkins.