Markov Model for Selection of Aortic Valve Replacement Versus Transcatheter Aortic Valve Implantation (Without Replacement) in High-Risk Patients

Abstract

Comparisons between transcatheter aortic valve implantation without replacement (TAVI) and tissue aortic valve replacement (AVR) in clinical trials might not reflect the outcomes in standard clinical practice. This could have important implications for the relative cost-effectiveness of these alternatives for management of severe aortic stenosis in high-risk patients for whom surgery is an option. The mean and variance of risks, transition probabilities, utilities, and cost of TAVI, AVR, and medical management derived from observational studies were entered into a Markov model that examined the progression of patients between relevant health states. The outcomes and cost were derived from 10,000 simulations. Sensitivity analyses were based on variations in the likelihood of mortality, stroke, and other commonly observed outcomes. Both TAVI and AVR were cost-effective compared to medical management. In the reference case (age 80 years, the perioperative TAVI and AVR mortality was 6.9% vs 9.8%, and annual mortality was 21% vs 24%), the utility of TAVI was greater than that of AVR (1.78 vs 1.72 quality-adjusted life years) and the lifetime cost of TAVI exceeded that of AVR ($59,503 vs $56,339). The incremental cost-effectiveness ratio was $52,773/quality-adjusted life years. Threshold analyses showed that variation in the probabilities of perioperative and annual mortality after AVR and after TAVI and annual stroke after TAVI were important determinants of the favored strategy. Sensitivity analyses defined the thresholds at which TAVI or AVR was the preferred strategy with regard to health outcomes and cost. In conclusion, TAVI satisfies current metrics of cost-effectiveness relative to AVR and might provide net health benefits at acceptable cost for selected high-risk patients among whom AVR is the current procedure of choice.