© 2017 American College of Cardiology Foundation Background Limited data exist regarding transcatheter mitral valve replacement (TMVR) for patients with failed mitral valve replacement and repair. Objectives This study sought to evaluate the outcomes of TMVR in patients with failed mitral bioprosthetic valves (valve-in-valve [ViV]) and annuloplasty rings (valve-in-ring [ViR]). Methods From the TMVR multicenter registry, procedural and clinical outcomes of mitral ViV and ViR were compared according to Mitral Valve Academic Research Consortium criteria. Results A total of 248 patients with mean Society of Thoracic Surgeons score of 8.9 ± 6.8% underwent TMVR. Transseptal access and the balloon-expandable valve were used in 33.1% and 89.9%, respectively. Compared with 176 patients undergoing ViV, 72 patients undergoing ViR had lower left ventricular ejection fraction (45.6 ± 17.4% vs. 55.3 ± 11.1%; p < 0.001). Overall technical and device success rates were acceptable, at 92.3% and 85.5%, respectively. However, compared with the ViV group, the ViR group had lower technical success (83.3% vs. 96.0%; p = 0.001) due to more frequent second valve implantation (11.1% vs. 2.8%; p = 0.008), and lower device success (76.4% vs. 89.2%; p = 0.009) due to more frequent reintervention (16.7% vs. 7.4%; p = 0.03). Mean mitral valve gradients were similar between groups (6.4 ± 2.3 mm Hg vs. 5.8 ± 2.7 mm Hg; p = 0.17), whereas the ViR group had more frequent post-procedural mitral regurgitation moderate or higher (19.4% vs. 6.8%; p = 0.003). Furthermore, the ViR group had more frequent life-threatening bleeding (8.3% vs. 2.3%; p = 0.03), acute kidney injury (11.1% vs. 4.0%; p = 0.03), and subsequent lower procedural success (58.3% vs. 79.5%; p = 0.001). The 1-year all-cause mortality rate was significantly higher in the ViR group compared with the ViV group (28.7% vs. 12.6%; log-rank test, p = 0.01). On multivariable analysis, failed annuloplasty ring was independently associated with all-cause mortality (hazard ratio: 2.70; 95% confidence interval: 1.34 to 5.43; p = 0.005). Conclusions The TMVR procedure provided acceptable outcomes in high-risk patients with degenerated bioprostheses or failed annuloplasty rings, but mitral ViR was associated with higher rates of procedural complications and mid-term mortality compared with mitral ViV.