Aims: This study aims to investigate whether information-theoretic measures of heart rate variability (HRV) can distinguish between ischemic (IHF) and non-ischemic (NIHF) etiologies in patients with heart failure with reduced ejection fraction (HFrEF), particularly under exercise stress. We also evaluated the feasibility of applying these metrics to short-term (ST) HRV for clinical and real-time monitoring purposes. Methods: Fifty-two HFrEF patients listed for heart transplantation (33 IHF, 19 NIHF) underwent continuous 12-lead ECG recording during rest and a graded cycling exercise test. RR intervals were extracted from ECG segments comprising 150 heartbeats at rest and 300 heartbeats during early and late exercise. We evaluated time-domain (MeanRR, SDRR), frequency-domain (LF/HF ratio) and nonlinear complexity measures, i.e. Conditional Entropy (CE) and Self-Entropy (SE), computed using a model-free k-nearest neighbor estimator. Group comparisons were performed using the Wilcoxon rank-sum test, with statistical significance set at p < 0.05. Results: Conventional HRV metrics showed expected changes with increased workload but failed to reveal differences between IHF and NIHF groups. In contrast, information-theoretic measures demonstrated significant etiology-specific. During the late phase of the exercise test, NIHF patients exhibited significantly higher CE compared to IHF patients, reflecting greater complexity and reduced determinism in heart rate dynamics. Conversely, SE was significantly lower in the NIHF group, indicating decreased regularity. These patterns were not statistically significant at rest and emerged in response to cardiovascular stress. Conclusion: Entropy-based HRV analysis from ST-ECG recordings may offer novel insights into the autonomic regulation differences between IHF and NIHF during exercise. These findings support the clinical value of information-theoretic metrics in characterize differences in cardiac dynamics in advanced heart failure, especially when assessed under stress conditions.