OBJECTIVES: While advances in immunosuppression have resulted in substantial improvements in short-term renal allograft survival, improvements in longer-term graft survival rates have been less marked. This may be due, in part, to antibody-mediated rejection driven by formation of de novo donor-specific antibodies (dnDSA). Because of its association with non-adherence, dnDSA has emerged as an important factor in longer-term graft loss. Long-term data on adherence, dnDSA formation, and graft failure were used to model clinical outcomes and costs associated with prolonged-release (PR) versus immediate-release (IR) tacrolimus (TAC) in renal transplant recipients in Canada. METHODS: A decision tree developed to capture differences in adherence between IR-TAC and PR-TAC was combined with a five-state Markov model of dnDSA formation, graft failure, and patient survival. Transition probabilities were determined by a series of Weibull, logistic, and least squares regression models. Adherence, quality of life, patient and graft survival, and drug costs were derived from Canada-specific sources. Analyses were run over a 25-year time horizon. Costs were reported in 2016 Canadian dollars, inflated where necessary. RESULTS: The proportion of patients experiencing dnDSA was reduced from 22.1% with IR-TAC to 20.5% with PR-TAC, reflecting a 7.2% relative reduction in dnDSA, and a number needed to treat of 63 to avoid dnDSA onset. In patients experiencing dnDSA, mean time to onset increased by 0.24 years to 7.9 years with PR-TAC relative to IR-TACAt a willingness-to-pay threshold of CAD 50,000 per quality-adjusted life year gained, PR-TAC would be considered cost-effective at a price 13.5% higher than that of IR-TAC. CONCLUSIONS: Based on modern clinical data on the incidence of dnDSA in adherent versus non-adherent patients, improved adherence associated with PR-TAC would delay the onset and reduce the incidence of dnDSA, and PR-TAC would remain cost-effective at a 13.5% higher per-milligram price when compared to IR-TAC in Canada.