BACKGROUND The association of early-onset posttransplantation hyperglycemia with long-term renal allograft survival is unknown. METHODS Seventy-one (SD 9) days after transplantation, 1410 first-time kidney transplant recipients without diabetes underwent an oral glucose tolerance test and were observed until primary outcome (graft loss) or December 31, 2008 (median [range], 6.0 years [0.3-13.8 years]). We used multivariable Cox regression analysis adjusted for age, gender, body mass index, creatinine level, donor age, preemptive transplantation, deceased donor, early rejection, and early cytomegalovirus infection to estimate hazard ratios for overall and death-censored allograft survival. RESULTS A total of 392 (28%) recipients experienced graft failure, and 235 (60%) were induced by death. Each 1 mmol/L increase in 2-hr plasma glucose (2hPG) was associated with 7% and 3% increased risk of unadjusted and adjusted overall graft failure (hazard ratio [95% confidence interval], 1.07 [1.04-1.10] and 1.03 [1.00-1.07]). Fasting plasma glucose was associated with unadjusted but not adjusted overall graft failure (1.09 [1.01-1.18] and 1.07 [0.98-1.17]). Neither 2hPG nor fasting plasma glucose was associated with death-censored graft loss (P=0.578 and P=0.896). Compared with recipients with normal glucose tolerance, recipients with posttransplantation diabetes mellitus showed a tendency toward increased overall multiadjusted graft failure (1.30 [0.98-1.73]). This was not observed in patients with impaired fasting glucose or impaired glucose tolerance. CONCLUSIONS In this study, 2hPG was associated with overall graft failure but not death-censored graft failure. The link between 2hPG and graft failure may be explained by the association with mortality.