Abstract Introduction Cardiac resynchronization therapy (CRT) reduces mortality and hospitalization for heart failure (HF) in selected patients with HF with reduced ejection fraction (HFrEF). However, there is a significant percentage of non-responders who are readmitted for HF after implantation. Several biomarkers and electrocardiographic findings have been suggested as prognostic markers in this group of patients. Fragmentation of the QRS complex (QRSf) has been associated with worse prognosis in several cardiac diseases, but evidence in patients undergoing CRT remains to be limited. Purpose To evaluate if the changes in QRSf status before and after CRT implantation have a prognostic role in patients undergoing CRT in terms of hospitalization for HF. Methods Consecutive patients undergoing CRT were prospectively included from October 2009 to December 2022. Demographic clinical, electrocardiographic and echocardiographic variables were recorded. Prospective follow-up variables were collected from electronic clinical records and telephone interview. QRSf complex was defined by the presence of different RSR′ patterns including an additional R-waves (R′) and notches of the R-wave and S-wave in two contiguous leads corresponding to a main coronary artery territory in a 12-lead surface electrocardiogram performed before and after CRT. Results A total of 244 patients (mean age 72 ± 10 years, 80% male) were included. The etiology of the LV systolic dysfunction was ischemic in 50.0%. Mean LVEF before implantation was 25 ± 9% and mean paced QRS width was 141 ± 23 ms. CRT-D was implanted in 176 patients (72.2%). At baseline, 150 patients had a QRSf. After CRT implantation, 69 (28.4%) patients experimented a change in QRSf status: 29 patients (11.9%) without prior QRSf developed QRSf, whereas 40 patients (16.5%) resolved QRSf after CRT. Median follow-up time was 36.0 months (interquartile range 18.2-75.7). At follow-up, 90 patients died and 91 presented at least one hospitalization for HF. Multivariate Cox regression survival analysis showed that development of QRSf after CRT implantation (HR 8.87; 95% CI 1.85-42.50) and the presence of QRSf before implantation with or without paced QRSf after CRT (HR 8.21; 95% CI 1.75-38.50 and HR 16.70; 95% CI 3.88-71.80 respectively) were independently associated with hospitalizations for HF. Univariate and multivariate Cox regression analyses for HF hospitalization are presented in Table 1. Figure 1 presents Kaplan-Meier curves for QRSf status and hospitalizations for HF. Conclusions According to our results, QRSf status before and after CRT have a significant prognostic role in terms of hospitalization for HF after CRT, especially in those who develop QRSf after CRT and those with baseline QRSf regardless QRSf status after CRT. The worse prognosis was observed in those with QRSf before and after CRT. Therefore, these patients should be followed more strictly because they are at higher risk of decompensation.Table 1.Cox Regression Analyses