Predictive value of QRS duration normalized to left ventricular volume for responding to cardiac resynchronization therapy in patients with mid-QRS duration enrolled in the multicenter registry

Abstract

Abstract Background For cardiac resynchronization therapy (CRT), patients with chronic heart failure (HF) and wide (>150msec) QRS duration (QRSd) received class I/A indication. However, its beneficial effect seemed to be limited for those with mid- (120–150msec) QRSd. Recent studies demonstrated that QRSd normalized to left ventricular end-diastolic volume (QRSd/ LVEDV) improved prediction of clinical outcome in patients with CRT. Therefore, we sought to investigate predictive value of QRSd/LVEDV for responding to CRT in patients with mid-QRSd. Methods This was retrospective multi-center observational cohort study. A total 506 consecutive patients who underwent CRT implantation in Tohoku University Hospital and National Cerebral and Cardiovascular Center were evaluated. Exclusion criteria were QRSd less than 120ms, upgrade procedures from other implanted non-CRT devices and bradycardia requiring pacing. We evaluated clinical variables, data of electrocardiogram and transthoracic echocardiography at baseline and 6 months after CRT implantation. Primary endpoint was a HF hospitalization after CRT implantation. Distribution of free from HF hospitalization during follow-up was calculated using Kaplan-Meier curves, and the effects of covariate on the time to endpoint were investigated using a Cox proportional hazards model. Results After 199 patients were excluded based on exclusion criterion, remaining 307 patients were included for the analysis. Mean age was 62±14 [SD] years, and 238 (77%) were male. Mean LVEF and LVEDV were 25±9% and 234±82ml, respectively, and 24% of patients had ischemic etiology of HF. During the median 948 days of follow-up, CRT patients with mid QRSd (n=126; 136±10msec), as compared with those with wide QRSd (n=181; 174±17msec), tended to have higher incidence of HF hospitalization (Wilcoxon p=0.03). Multivariate analysis showed that QRSd and QRSd/LVEDV were significant predictors for HF hospitalization in CRT patients with mid QRSd, and cut-off values (137msec of QRSd and 0.65 of QRSd/LVEDV), which was calculated by receiver operative curve analysis, was used for risk stratification. QRSd<137msec was significant negative predictors for HF hospitalization (p=0.005), and Mid-QRSd patients with QRSd≥137msec demonstrated equivalent clinical outcome with those with wide QRSd. Moreover, patients with QRSd/LVEDV≥0.65 tended to have lower incidence of HF hospitalization as compared with those without it among patients with QRSd<137msec (n=64, Figure). Conclusion The present study demonstrates that QRSd normalized to left ventricular end-diastolic volume (QRSd/ LVEDV) could be clinical value in predicting outcome in CRT patients with mid-QRSd. These findings indicate normalized QRSd reflects myocardial conduction properties and contribute to risk stratification. Funding Acknowledgement Type of funding sources: None.