OBJECTIVES: Corticosteroids used as immunosuppressants in preventing acute rejection (AR) and graft loss (GL) following kidney transplantation are associated with potentially serious adverse events. Evidence from randomized controlled trials suggested that patients on a tacrolimus-based immunosuppressant regimen were able to withdraw from steroids without increasing their risk of these events. However, the trials were of short-term duration. The objective was to measure the long-term tradeoff between potential graft loss and severe cardiovascular and other adverse events, for a heterogeneous-risk population, to determine the optimal timing of steroid withdrawal.  METHODS: A discrete event simulation was developed that included the following events: AR, GL, myocardial infarction (MI), and ten other adverse outcomes. Data from the United States Renal Data System were used to estimate event-specific parametric regressions, which accounted for steroid-sparing regimen (avoidance, 7-day withdrawal, 6-month withdrawal, 12-month withdrawal, and maintenance) as well as patients’ demographics, immunologic risks, and comorbidities. Regression estimates were then used to derive individual Weibull distributions, which, in turn, were used to guide patients in the model from event to event for 20 years. RESULTS: “Base case” patients (white, without cardiovascular disease or CVD) in 6-month and 12-month steroid withdrawal groups were significantly less likely to experience MI (9.6-9.8% vs 12.2%, p<0.01), new onset diabetes (NODM) (37.2%-42.4% vs 46.4%, p<0.01), bacterial infection (51.7%-57.6% vs 67.4%, p<0.01), and cardiac-related death (24.5%-25.7% vs 28.8%, p<0.01), compared to steroid maintenance. Compared to base case patients, those with CVD and African Americans were significantly (p<0.01) more likely to have GL and NODM across the steroid-sparing regimens. CONCLUSIONS: Strategies of