Wide variations in tacrolimus levels have been identified as a risk factor for inferior kidney allograft survival but past studies have not properly accounted for the dynamic nature of drug exposure over time. Here we evaluated whether time-varying exposure to tacrolimus increases the risk of long-term adverse outcomes in a retrospective cohort study in adult kidney transplant recipients on tacrolimus-based immunosuppression. Time-dependent Cox proportional hazards models were used to examine the association between the standard deviation of tacrolimus levels (TacSD) starting at 1-year post-transplant and the composite end point of late allograft rejection, transplant glomerulopathy, or total graft loss (including death). Among 356 patients, there was a significant 27% increase in the adjusted hazard of the composite end point for every 1-unit increase in TacSD (hazard ratio 1.27 (95% confidence interval 1.03, 1.56)). There was also a graded increase in the relative hazard for the composite end point by TacSD threshold (hazard ratios 1.33, 1.50, 1.84, and 2.56 for TacSD 1.5, 2, 2.5, and 3, respectively). The results were similar for total graft loss and the composite end point excluding death. Thus, increased time-dependent TacSD may be an independent risk factor for adverse kidney transplant outcomes. TacSD may serve as a monitoring tool to identify high-risk patients. Whether interventions to decrease TacSD will improve outcomes requires further study.