Morphology-based selvester QRS score for phenotyping acute heart failure in patients with native left bundle branch...

Rapid differentiation of left ventricular (LV) systolic dysfunction from right heart failure (RHF) is essential in acute heart failure (AHF). In patients with native left bundle branch block (LBBB), conventional ECG parameters have limited diagnostic value. We evaluated whether the morphology-based Selvester QRS score (SQS) could distinguish heart failure phenotypes.

To compare SQS with conventional QRS metrics across AHF phenotypes and evaluate their diagnostic performance.

This retrospective single-center study included 318 consecutive patients hospitalized with AHF and native LBBB. Patients were classified as HFrEF (n=137), HFmrEF (n=82), HFpEF (n=56), or RHF (n=43). SQS was calculated using the 57-point LBBB-adapted Strauss-Selvester algorithm. ECGs were independently analyzed by two blinded cardiologists. Correlation, multivariable logistic regression, and ROC analyses were performed.

Mean age was 62±14 years, and 69.2% were male. QRS duration and amplitude did not differ among phenotypes. SQS was significantly higher in HFrEF than in HFmrEF, HFpEF, and RHF (median 46 vs. 40, 40, and 40; p<0.001). SQS correlated with LV and RV structural and functional parameters, including LVEF, TAPSE, LVEDD, LVESD, LV mass index, RVEDD, and E/e' (all p<0.001). SQS was the only clinically meaningful independent ECG predictor of HFrEF (QRS duration also reached significance with a paradoxical inverse OR, likely reflecting confounding); RVEDD and TAPSE independently predicted RHF. ROC analysis showed moderate discrimination for HFrEF (AUC=0.768) and RHF (AUC=0.710).

In AHF patients with native LBBB, SQS provides greater phenotypic discrimination than conventional QRS metrics, particularly for HFrEF. Lower SQS together with impaired RV parameters may facilitate early recognition of RHF. Prospective multicenter validation is warranted.