Preoperative fragmented QRS as a predictor of postoperative left ventricular systolic dysfunction after isolated...

Despite preserved preoperative left ventricular ejection fraction (LVEF), many patients develop postoperative left ventricular (LV) systolic dysfunction after mitral valve surgery for chronic mitral regurgitation. Fragmented QRS (fQRS) on surface electrocardiography has been proposed as a simple marker of adverse myocardial substrate. This study evaluated the prognostic value of preoperative fQRS for predicting postoperative LV systolic dysfunction after isolated mitral valve replacement and developed an exploratory perioperative risk model.

This retrospective study included 279 consecutive patients undergoing elective isolated mitral valve replacement for chronic mitral regurgitation, with or without concomitant tricuspid annuloplasty, between January 2022 and June 2025. Fragmented QRS was assessed on preoperative 12-lead electrocardiography. Postoperative LV systolic dysfunction was defined as LVEF < 50% on early postoperative echocardiography. Multivariable logistic regression analyses, including an exploratory backward stepwise approach, were performed. Model discrimination and calibration were evaluated, and a nomogram was constructed for individualized risk estimation.

Fragmented QRS was present in 71 patients (25.4%). Postoperative LV systolic dysfunction occurred in 92 of 279 patients (33.0%) overall. Although preoperative LVEF was similar between groups, patients with fQRS had lower postoperative LVEF (p = 0.036) and a higher incidence of postoperative LV systolic dysfunction compared with those without fQRS (45.1% vs. 28.8%; p = 0.012). In the fully adjusted multivariable model, preoperative fQRS showed a directional association with postoperative LV systolic dysfunction, although this association narrowly missed conventional statistical significance (OR 1.90, 95% CI 0.99-3.64; p = 0.052). In the exploratory stepwise multivariable model developed for risk prediction, preoperative fQRS was associated with postoperative LV systolic dysfunction (OR 2.00, 95% CI 1.06-3.81; p = 0.033), together with longer aortic cross-clamp time (OR 1.015, 95% CI 1.010-1.020; p < 0.001) and chronic kidney disease (OR 3.06, 95% CI 1.11-8.56; p = 0.030). The exploratory model demonstrated acceptable discrimination (area under the ROC curve, 0.78). A nomogram was developed to facilitate individualized exploratory risk prediction.

Preoperative fQRS showed a directional association with postoperative LV systolic dysfunction after isolated mitral valve replacement, although this association narrowly missed conventional statistical significance in the fully adjusted multivariable model (OR 1.90, 95% CI 0.99-3.64; p = 0.052). When integrated with baseline ventricular function and perioperative variables, fQRS may contribute to exploratory perioperative risk stratification in selected patients. These findings require validation in larger prospective cohorts before clinical implementation.