Electrocardiographic changes with left ventricular ejection fraction shifts in narrow QRS heart failure: A retrospective cohort study

Objectives: Heart failure (HF) with reduced ejection fraction is frequently accompanied by QRS prolongation; however, a substantial proportion of patients with significant systolic dysfunction maintain a normal QRS duration, limiting the utility of traditional depolarization-based electrocardiogram (ECG) markers for assessing disease severity. We sought to explore whether fluctuations in left ventricular ejection fraction (LVEF) may be associated with changes in corrected QT interval (QTc), QRS axis heart rate, and T-wave axis in HF patients with narrow QRS. Methods: We conducted an exploratory retrospective cohort study with patients identified from the University of California, Davis echocardiogram (echo) database between June 1, 2011 and June 30, 2025. Patients were included if they had one echo demonstrating LVEF = 30% and another demonstrating LVEF ≥ 50%, each paired with a 12-lead ECG obtained within ± 7 days. Patients with QRS duration > 100 ms, advanced atrioventricular block, or prior cardiac procedures visualized on echo were excluded. ECG parameters evaluated included QTc interval using Fridericia formula, QRS axis, ventricular heart rate, and T-wave axis. Linear mixed-effects models were used to assess within-patient changes in ECG parameters associated with LVEF status. The random-intercept model inherently accounts for all time-invariant subject characteristics such as sex by partitioning the between subject variance (τ00) from the within subject residual variance (σ2). Results: Of 2980 patients, 2908 were excluded due to ineligible echocardiogram or ECG findings, leaving only 72 included in the final analysis. Among these eligible patients, transitions between reduced and preserved LVEF were not associated with significant changes in QTc interval ( β = 13.19 ms, SE = 6.43, P = 0.0439) or QRS axis ( β = 1.06 degrees, SE = 4.15, P = 0.8). In contrast, ventricular heart rate ( β = -16.15 beats per minute, SE = 3.84, P = 7.3 × 10-5) and T-wave axis ( β = -32.65 degrees, SE = 10.16, P = 0.002) and were significantly higher during periods of reduced LVEF and decreased with recovery of systolic function. These preliminary findings raise the possibility that conventional ECG conduction metrics could remain relatively stable despite changes in ventricular performance, while chronotropic regulation and repolarization patterns may warrant further investigation as potentially more responsive markers of disease severity. Conclusion: Exploration of ECG changes in patients with HF and preserved QRS duration, revealed significant changes in ventricular heart rate and T-wave axis but not QTc interval or QRS axis when comparing LVEF states. These ECG features may represent candidate for noninvasive markers of HF severity when ventricular conduction is preserved, though larger confirmatory studies are needed.