QRS complex fragmentation before and after cardiac resynchronization therapy as a predictor of heart failure hospitalizations: differences between ischaemic and non-ischaemic aetiologies

Abstract Introduction Cardiac resynchronization therapy (CRT) reduces mortality and hospitalizations for heart failure (HF) in heart failure with reduced ejection fraction (HFrEF). However, the prognostic role of fragmented QRS (fQRS) in patients undergoing CRT concerning the ethology of cardiomyopathy remains unclear. Objective To assess the presence of fQRS before and after CRT and correlate it with the heart disease aetiology (ischaemic vs. non-ischaemic) and the occurrence of new HF hospitalizations. Methods Patients undergoing CRT between 2009 and 2022 were included. QRS fragmentation before and after CRT was defined by different patterns of RSR', additional R waves (R′), and notches in R and S waves in two contiguous leads. Four categories were distinguished based on the presence of fQRS or non-fragmented QRS (nfQRS) before and after implantation. Results A total of 244 patients were included (74±8 years; 92% male; 50% ischaemic aetiology; Table 1). After a median follow-up of 36 months, 92 patients were hospitalized for HF (51 ischaemic and 41 non-ischaemic, p=0.234). Multivariate Cox survival analysis stratified by the type of heart disease showed that fQRS before and after implantation was associated with HF hospitalization in the non-ischaemic group (Hazard Ratio [HR] 22.20; 95% Confidence Interval [CI] 2.74-179), while fQRS resolution or development after CRT did not associate a higher risk. The development, resolution, or persistence of fQRS post-CRT did not significantly increase risk in ischaemic patients (Table 2). Conclusions The persistence of fQRS after CRT in patients with HFrEF and non-ischaemic aetiology was associated with a higher risk of HF hospitalization, whereas those with pre-implant nfQRS or those who resolved fragmentation after CRT had a better prognosis.Table 1.Baseline characteristics  Table 2.Cox regression