BACKGROUND- Traditional management of severe aortic stenosis (AS) is based on delay in aortic valve replacement (AVR) until the development of symptoms. Surgery for asymptomatic AS has been proposed to reduce the small risk of sudden death before AVR and avoid heart failure (HF) after AVR. Because a trial to compare these options is unlikely, we developed a Markov model to inform the choice between immediate surgery and watchful waiting in asymptomatic AS. METHODS AND RESULTS- We defined health states as preoperative, postoperative, postcomplication, and death. We calculated the implications of watchful waiting, tissue and mechanical AVR-based on risks, transitions, utilities, and cost derived from literature review. Further analyses evaluated situations thought to favor immediate surgery and watchful waiting. Sensitivity analyses were based on the likelihood of preoperative death and HF in follow-up. In the reference case (age, 65 years; post-AVR utility, 0.9; annualized pre-AVR mortality, 1%; and post-AVR HF, 11.3%), the utility of watchful waiting was superior to that of immediate mechanical or tissue AVR (quality-adjusted life-years, 7.4 versus 5.3 versus 5.3, respectively), and the cost was less than immediate surgery. Sensitivity analyses showed immediate surgery was not likely to be more effective regardless of the yearly probability of post-AVR HF in the watchful waiting group (range, 0% to 80%). Immediate surgery was likely to be effective when pre-AVR annual mortality reached 13%. CONCLUSIONS- Immediate surgery in asymptomatic severe AS does not improve outcomes unless risk of sudden death pre-AVR and HF after AVR are higher than currently reported.