EComment. Repair versus mitral replacement in patients undergoing concomitant aortic valve replacement

Abstract

I congratulate Kim et al. on a very large series on double valve procedures with an impressively low in-hospitality mortality rate [1]. Their work raises a number of interesting questions: Firstly, 410 patients undergoing double valve procedures were excluded. Did all these patients have rheumatic mitral stenosis, or did another reason exist for their exclusion? The data on double valve replacement in patients with mitral stenosis may have helped to provide a control for mechanical valve related events and potentially a Kaplan-Meier survival comparison (which was not included in the manuscript) for the mitral repair group. Kaplan-Meier survival curves are univariate in nature. Did a multivariate Cox model procedure have the same outcomes as the inverse probability treatment weighted method? A simple Kaplan-Meier technique may be inaccurate, despite inverse probability treatment weighting, as significant differences in aortic tissue valve replacement and aortic graft replacement existed in their adjusted data. Of note, preoperative atrial fibrillation and ejection fraction also approached significance in the adjusted data. With nearly 25% of patients suffering postoperative mitral valve dysfunction at five years after repair surgery, I would tentatively interpret this as a need for caution for mitral valve repair in patients having concomitant mechanical aortic vale replacement surgery, or in patients with a life expectancy greater than five years. Patients undergoing tissue or mechanical aortic valve replacement surgery are by definition different patient populations, regardless of any mitral valve procedure performed [2]. Propensity matching and Cox multivariate analysis are unlikely to adjust for this selection bias. I would be interested in knowing the outcomes for patients undergoing double valve replacement versus aortic valve replacement plus mitral valve repair, for tissue and mechanical valve replacement subgroups. I thank Kim et al. for their thoughtful analysis of a high-risk difficult subgroup of patients. Conflict of interest: none declared.