Patient and Surgeon Predictors of Mitral and Tricuspid Valve Repair for Infective Endocarditis

Mitral repair (MVr) is superior to replacement for degenerative disease; however, its benefit is less established for endocarditis. We report outcomes of repair or replacement for mitral/tricuspid endocarditis and identify predictors of MVr. Patients undergoing first-time surgery for mitral (n = 260) or tricuspid (n = 71) endocarditis between 1992 to 2018 were identified. Patients with aortic endocarditis were excluded. Primary outcome was all-cause mortality and secondary outcome was MVr. Patients were stratified into active and treated endocarditis separately for mitral and tricuspid groups. Predictors of MVr were assessed through multivariable logistic regression and adjusted likelihood of MVr through marginal effects estimates. A mitral specialist was defined by performing ≥25 annual degenerative MVr. Among 331 patients, 70% (181/260) of those with mitral valve endocarditis and 52% (37/71) of those with tricuspid endocarditis underwent repair. The MVr group compared with replacement had a higher proportion of elective acuity and less diabetes, hypertension, active endocarditis, cardiogenic shock, and dialysis. Estimated 5-year survival did not differ between repair versus replacement for active mitral (68 ± 14% vs 60 ± 14%, P = 0.34) or tricuspid endocarditis (60 ± 17% vs 61 ± 19%, P = 0.67), but was superior after repair for treated mitral endocarditis (86 ± 7% vs 51 ± 24%, P = 0.014). Independent predictors of mortality included dialysis for active and treated mitral endocarditis, and mitral replacement (vs MVr) for treated mitral endocarditis. The likelihood of MVr was 82 ± 5% for mitral specialists and 47 ± 9% for non-specialists (P < 0.001). MVr for endocarditis should be pursued, if feasible. Importantly, achieving MVr was driven not only by patient factors, but also surgeon experience.