Optimal exercise prescription for patients with coronary heart disease across cardiac rehabilitation phases: a systematic review and network meta-analysis

BACKGROUND: Exercise-based cardiac rehabilitation is delivered across distinct clinical phases, yet the optimal exercise modality for improving cardiorespiratory fitness and functional capacity may differ by phase. This systematic review and network meta-analysis compared the relative effects of common exercise modalities on peak oxygen uptake (VO2peak) and 6-min walk distance (6MWD) in patients with coronary heart disease across rehabilitation phases. METHODS: English and Chinese databases were searched from inception to April 30, 2026. Randomized controlled trials enrolling adults with coronary heart disease and evaluating exercise interventions were eligible. Two reviewers independently screened studies, extracted data, and assessed risk of bias using the Cochrane risk-of-bias tool. A frequentist network meta-analysis synthesized direct and indirect evidence within each phase (phase I inpatient, phase II early outpatient, phase III long-term maintenance). Continuous outcomes were pooled as mean differences with 95% confidence intervals (VO2peak in mL·kg⁻¹·min⁻¹; 6MWD in meters). Global inconsistency testing guided the use of consistency vs. inconsistency models. Interventions were ranked using the surface under the cumulative ranking curve (SUCRA). RESULTS: Eighty-seven trials (7,241 participants) were included, covering aerobic exercise (AE), resistance training (RT), combined aerobic plus resistance training (AE + RT), high-intensity interval training (HIIT), traditional Chinese exercise (TCE; e.g., Tai Chi/qigong/Baduanjin), aerobic exercise combined with traditional Chinese exercise (AE + TCE), and conventional therapy (CT). Evidence was phase-imbalanced (9 phase I, 39 phase II, and 40 phase III trials); one study contributed outcomes to both Phase II and Phase III analyses. In phase I, evidence was comparatively limited; for VO2peak TCE and AE showed the most favorable probability of benefit vs. CT. Phase II provided the most informative network for VO2peak with RT and HIIT ranking highest, followed by AE + TCE. In phase III, for 6MWD, AE + RT ranked highest, with AE and TCE also showing favorable ranking; for VO2peak maintenance-phase signals favored RT- and combined-training-based approaches. CONCLUSIONS: The relative effectiveness of exercise modalities appears phase-dependent. These findings support phase-tailored exercise prescription in coronary heart disease, while emphasizing the need for additional head-to-head trials-particularly in early inpatient rehabilitation-to strengthen comparative evidence. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD420251156834, PROSPERO, identifier CRD420251156834.