Cardiorespiratory fitness gap during stable recovery after percutaneous coronary intervention relative to a contemporaneous angiographic reference group: a propensity score overlap-weighted cross-sectional study

Percutaneous coronary intervention (PCI) relieves epicardial coronary stenosis, but anatomical treatment does not necessarily indicate recovery of integrated exercise function. We examined whether patients during stable recovery after PCI had lower measured cardiorespiratory fitness (CRF) than a contemporaneous angiographic reference group and how much body composition and common clinical and metabolic factors accounted for the difference. This single-center retrospective cross-sectional study included 682 patients who underwent coronary angiography, cardiopulmonary exercise testing (CPET), and bioelectrical-impedance assessment: 209 in the reference group and 473 in the PCI group. Both groups completed the study assessments 4–6 weeks after the corresponding coronary angiography or planned PCI. Propensity score overlap weighting was used to compare absolute peak oxygen uptake (VO₂peak), body weight-normalized VO₂peak, and percent-predicted VO₂peak (pVO₂). Skeletal muscle mass and body fat mass or their height-indexed measures were aligned with the scale of each outcome. Robustness was assessed using an extended clinical model, cross-fitted reference-group mapping, and sensitivity analyses. The six propensity-score covariates were balanced after overlap weighting. In the primary extended model, the PCI group had lower absolute VO₂peak (difference, − 59.69 mL/min; 95% confidence interval [CI], − 109.65 to − 9.74; P = 0.019), body weight-normalized VO₂peak (− 0.83 mL·kg⁻¹·min⁻¹; 95% CI, − 1.54 to − 0.11; P = 0.023), and pVO₂ (− 4.28% points; 95% CI, − 7.13 to − 1.43; P = 0.003). In cross-fitted reference-group mapping, only the residual difference in pVO₂ remained statistically significant (− 4.16% points; 95% CI, − 7.45 to − 1.00; P = 0.012). Results were generally consistent across sensitivity analyses. Patients during stable recovery after PCI had a CRF gap relative to a contemporaneous angiographic reference group, with the pVO₂ difference showing the greatest consistency across analytical frameworks. These cross-sectional findings should not be interpreted as a causal effect of PCI, but they suggest that direct CPET can complement coronary anatomy, resting LVEF, and body-composition quantity in functional assessment.