Redo surgery for noninfective isolated mitral valve disease: Initial outcome and further follow‐up compared to primary surgery

Introduction: Isolated redo‐mitral valve replacement (iMVR) is underreported and often mixed up with endocarditis in the present literature. The present study compares first with redo iMVR in noninfective mitral disease. Patients and Methods: A total of 3821 mitral valve procedures were analyzed. The study was restricted to isolated and noninfective mitral valve replacements done by sternotomy. Finally, 402 patients are included, consisting of 102 redo‐ and 300 first surgeries. The mean patient's age was 65.9 ± 10.4 years; the mean EuroSCORE II was 3.0 ± 2.2%. Median follow‐up was 221 days, ranging up to 9.9 years with a total of 367 patient‐years. Results: Redo's had higher EuroSCORE II (5.1 ± 2.9% vs. 2.3 ± 1.4%; p < .01), more atrial fibrillation (31.1% vs. 46.1%; p = .01), chronic obstructive pulmonary disease (7.3% vs. 17.6%; p = .05), coronary artery disease (7.3% vs. 17.6%; p = .03) and more frequently reduced ejection fraction < 30% (3.0% vs. 11.8%; p = .02). Main outcomes showed comparable 30‐days mortality (first: 4.1%, redo: 6.9%; p = .813). Postoperative morbidity of the redo's was associated with increased postoperative bleeding (p < .01) resulting in increased transfusions of packed red blood cells and fresh frozen plasma (each p < .01), more re‐explorations (p < .01) and longer primary intensive care unit stay (p < .01). Postoperative occurrence of stroke, respiratory or renal failure, and myocardial infarction as well as hospital stay differed not significantly. Estimated 5‐years survival was 65.5 ± 12.3% for all patients with no significant differences between the groups. Multivariate logistic regression respiratory failure as relevant for hospital (odds ratio [OR]: 12.3 [1.1–158]; p = .029) and stroke (OR: 4.8 [1.1–12.3]; p = .021) as relevant for long‐term mortality. Conclusion: iMVR for noninfective reasons is infrequent and rare. Compared to primary surgery, redo's suffer mainly from bleeding‐associated morbidity. This does not translate into pr