Understanding the Impact of Comorbidity-Interaction in Patients Undergoing Transcatheter Edge-to-Edge Mitral Valve Repair on Outcomes

Abstract

ABSTRACTBACKGROUNDTranscatheter Edge-to-Edge Mitral Valve Repair (M-TEER) is an accepted procedure for high-risk surgical patients with degenerative and functional mitral regurgitation. Non-cardiovascular comorbidities (NCCs) are highly prevalent in patients undergoing M-TEER. Although the impact of mitral valve anatomy and cardiac comorbidities in determination of M-TEER outcomes has been studied, precise understanding of the effect of the burden of NCCs on patients undergoing M-TEER remains unclear for acute outcomes. Our objective was to identify the association of NCC comorbidity-interaction patterns in patients undergoing M-TEER on length of stay (LOS), cost of care, and in-hospital major adverse cardiovascular events (MACE).METHODS9 245 admissions from the Nationwide Readmission Database that underwent M-TEER between 2015 and 2018 were included in the study. Patients were categorized by the overall burden of non-cardiovascular comorbidities (0, 1, 2, and ≥ 3). NCC included chronic liver disease, chronic lung disease, obesity, diabetes mellitus, dementia, major depressive disorder, chronic anemia, chronic kidney disease including end-stage renal disease (ESRD) on dialysis, and malignancy. Logistic Regression and Machine Learning (ML) algorithms were used to assess associations between comorbidity burden and in-hospital MACE.RESULTSOut of 9 245 index admissions, in-hospital MACE was recorded in a total of 504 (5.3 %). Of these, the majority (30.4%) had one NCC (n = 2 861). Patients with at least three NCCs had the longest median LOS [3.0, IQR (1.0 – 11.0)] and highest median cost of hospital care [$47 275, IQR (34 175.8 – 71 149.4)]. The Gradient Boosting (GB) classifier performed the best in predicting MACE with an AUROC of 96 % (95% CI: 0.95 – 0.97). The top features of importance that predicted in-hospital MACE were admission type, number of NCCs, and age in descending order.CONCLUSIONSCalibrated GB classifier identified patients with three NCCs as the subset of admission having the highest probability of a positive MACE outcome.