Mitral Valve Repair or Replacement for Ischemic Mitral Regurgitation? The Italian Study on the Treatment of Ischemic Mitral Regurgitation (ISTIMIR)

OBJECTIVE(S): It is still uncertain whether mitral valve (MV) replacement is really inferior to mitral valve repair (MVR) for the treatment of chronic ischemic mitral regurgitation (CIMR). This multicentric study is aimed to give a contribution to answer this question. METHODS: Among 1,067 patients with CIMR and impaired left ventricular (LV) function (ejection fraction <40%) operated on at 13 Italian Institutions between 1996 and 2011, 298 (27.9%) underwent mitral valve (MV) replacement whereas 769 (72.1%)had a MVR. Propensity scores (PS) were calculated by a non-parsimonious multivariable logistic regression and 244 pairs of patients were successfully matched using calipers of width 0.2 SDs of the logit of the PS. The post-matching median standardized difference was 0.031143 (Interquartile Range [IQR] 0.015661–0.075236) and in none of the covariates it exceeded 10%. RESULTS: Early death were 3.3% (n = 8) in MVR vs. 5.3% (n = 13) in MV replacement (p = 0.32). During a median 46.5-month follow up (IQR 26–69) thirty-six patients (14.7%) undergoing repair and 41 (16.8%) in the replacement group died (p = 0.51). Eight-year survival were 81.6 ± 2.8 and 79.6 ± 4.8 (stratifi ed log-rank test 0.42) whereas freedom from valverelated death were 99.1 ± 4.6 and 99.2 ± 6.4 (p = 0.84 )in the repair and replacement group, respectively. Recurrence of MR (≥2+) was observed in 61 patients (25%) vs. 4(1.6%) in repair and replacement, respectively (p < 0.001). Freedom from reoperation were 80 ± 4.1 vs. 64.3 ± 4.3 (p < 0.001).Freedom from valve-related complications was 85.5 ± 5.2 after repair and 87.8 ± 4.6 after MVR (p = 0.88). Left ventricular ejection fraction did not signifi cantly improve and it was comparable in the two Groups at follow up control (36.9 ± 38.5, p = 0.66). At Cox proportional hazard regression models stratifi ed on the matched pairs MVR was a strong predictor of reoperation (HR 2.84 [95% CI 2.51– 3.26], p < 0.001). CONCLUSIONS: MV replacement is a suitable option for patients with CIMR and impaired LV function. It provides better results in terms of freedom from reoperation with comparable long-term survival and valverelated complication rates.