Higher baseline fat mass index is associated with smaller improvement in cardiorespiratory fitness following cardiac rehabilitation

Background Cardiorespiratory fitness (CRF), quantified by peak oxygen uptake (VO₂peak), is a strong prognostic marker in cardiovascular disease. Although exercise-based cardiac rehabilitation (CR) improves VO₂peak, inter-individual response varies widely. Age, baseline CRF and exercise dose are established predictors of CRF improvement, whereas body mass index (BMI) shows no clear association. Although adiposity is inversely associated with CRF cross-sectionally, its influence on CRF improvement during CR remains uncertain. Purpose To examine whether baseline fat mass index (FMI) is associated with changes in VO₂peak normalised to fat-free mass (VO₂peak/FFM) following exercise-based cardiac rehabilitation and, secondarily, to evaluate the association of baseline fat-free mass index (FFMI) with this outcome. Methods This retrospective cohort study included adults with coronary artery disease or heart failure who participated in phase II exercise-based CR at a tertiary care centre between 2015 and 2024. Participants underwent symptom-limited cardiopulmonary exercise testing with gas exchange analysis before and after the programme. Body composition was assessed using bioelectrical impedance analysis prior to exercise testing. Baseline FMI and FFMI were expressed as sex-specific z-scores. The primary outcome was relative change in VO₂peak/FFM (%). Multivariable linear regression adjusted for age, sex, baseline VO₂peak/FFM, CR session number, testing interval, heart failure, hypertension, diabetes, dyslipidaemia and smoking. Restricted cubic spline modelling assessed non-linearity. Results Among 423 participants (mean age 60.2 ± 11.0 years; 86% male), the median number of supervised CR sessions was 35 and the median relative improvement in VO₂peak/FFM was 15.9% (interquartile range 4.3–30.9). In the fully adjusted model, each 1-standard deviation increase in FMI z-score was associated with a 5.05-percentage-point smaller improvement in VO₂peak/FFM (95% confidence interval −7.48 to −2.62; P < 0.001), whereas each 1-standard deviation increase in FFMI z-score was associated with a 2.26-percentage-point greater improvement (95% confidence interval 0.68 to 3.85; P = 0.005). Restricted cubic spline analysis demonstrated a significant non-linear association between FMI and VO₂peak/FFM improvement (P for non-linearity = 0.037), with greater gains at lower FMI levels, a plateau at intermediate levels, and progressively attenuated improvement at higher FMI levels (Figure 1). Conclusion Higher baseline FMI is independently and non-linearly associated with smaller improvement in VO₂peak/FFM following CR, whereas higher baseline FFMI is associated with greater improvement. Body composition assessment may provide additional information beyond BMI when evaluating variability in CRF response during cardiac rehabilitation.