To our knowledge, REPRISE III is the first large randomized comparison of 2 different transcatheter aortic valve replacement platforms: the mechanically expanded Lotus valve (Boston Scientific) and self-expanding CoreValve (Medtronic). To evaluate outcomes of Lotus vs CoreValve after 2 years. A total of 912 patients with high/extreme risk and severe, symptomatic aortic stenosis enrolled between September 22, 2014, and December 24, 2015, were randomized 2:1 to receive Lotus (607 [66.6%]) or CoreValve (305 [33.4%] at 55 centers in North America, Europe, and Australia. The first 2-year visit occurred on October 17, 2016, and the last was conducted on April 12, 2018. Clinical and echocardiographic assessments are complete through 2 years and will continue annually through 5 years. All-cause mortality and all-cause mortality or disabling stroke at 2 years. Other clinical factors included overall stroke, disabling stroke, repeated procedures, rehospitalization, valve thrombosis, and pacemaker implantation. Echocardiographic analyses included effective orifice area, mean gradient, and paravalvular leaks (PVLs). Of 912 participants, the mean (SD) age was 82.8 (7.3) years, 465 (51%) were women, and the mean (SD) Society of Thoracic Surgeons predicted risk of mortality was 6.8% (4.0%). At 2 years, all-cause death was 21.3% with Lotus vs 22.5% with CoreValve (hazard ratio [HR], 0.94; 95% CI, 0.69-1.26; P = .67) and all-cause mortality or disabling stroke was 22.8% with Lotus and 27.0% with CoreValve (HR, 0.81; 95% CI, 0.61-1.07; P = .14). Overall stroke was 8.4% vs 11.4% (HR, 0.75; 95% CI, 0.48-1.17; P = .21); disabling stroke was more frequent with CoreValve vs Lotus (4.7% Lotus vs 8.6% CoreValve; HR, 0.53; 95% CI, 0.31-0.93; P = .02). More Lotus patients received a new permanent pacemaker (41.7% vs 26.1%; HR, 1.87; 95% CI, 1.41-2.49; P < .01) or had a valve thrombosis (3.0% vs 0.0%; P < .01) compared with CoreValve. More patients who received CoreValve experienced a repeated procedure (0.6% Lotus vs 2.9% CoreValve; HR, 0.19; 95% CI, 0.05-0.70; P < .01), valve migration (0.0% vs 0.7%; P = .05), or embolization (0.0% vs 2.0%; P < .01) than Lotus. Valve areas remained significantly larger and the mean gradient was lower with CoreValve than Lotus (valve area, mean [SD]: Lotus, 1.53 [0.49] cm2 vs CoreValve, 1.76 [0.51] cm2; P < .01; valve gradient, mean [SD]: Lotus, 13.0 [6.7] mm Hg vs 8.1 [3.7] mm Hg; P < .01). Moderate or greater PVL was more frequent with CoreValve (0.3% Lotus vs 3.8% CoreValve; P < .01) at 2 years. Larger improvements in New York Heart Association (NYHA) functional class were observed with Lotus compared with CoreValve (improved by ≥1 NYHA class: Lotus, 338 of 402 [84.1%] vs CoreValve, 143 of 189 [75.7%]; P = .01; improved by ≥2 NYHA classes: 122 of 402 [37.3%] vs 65 of 305 [21.3%]). After 2 years, all-cause mortality rates, mortality or disabling stroke were similar between Lotus and CoreValve. Disabling stroke, functional class, valve migration, and PVL favored the Lotus arm whereas valve hemodynamics, thrombosis, and new pacemaker implantation favored the CoreValve arm. clinicaltrials.gov Identifier: NCT02202434.