Baseline peak VO₂ and treatment effect consistency of perioperative rehabilitation in cardiac valve surgery: an exploratory subgroup analysis of the PORT trial

Background In the PORT trial (Perioperative Rehabilitation in Patients Undergoing Elective Cardiac Valve Surgery: A Randomised Controlled Trial), a short-term multidomain perioperative rehabilitation programme was associated with a reduced incidence of postoperative pneumonia in patients undergoing elective cardiac valve surgery. Whether baseline peak oxygen consumption (peak VO₂), a marker of cardiorespiratory reserve, modifies the effect of this intervention remains uncertain. Objectives To evaluate whether the association between perioperative rehabilitation and postoperative outcomes differs according to baseline peak VO₂ (≤ 20 vs. > 20 mL·kg⁻¹·min⁻¹). Methods This exploratory subgroup analysis included 702 participants from the PORT trial who underwent preoperative cardiopulmonary exercise testing. Patients were stratified by baseline peak VO₂ and randomized to perioperative rehabilitation—comprising education, inspiratory muscle training, active cycle of breathing techniques, and early mobilization—or usual care. The primary endpoint was a composite of in-hospital all-cause mortality, postoperative pulmonary complications (PPCs), and prolonged hospitalization (> 7 days). Logistic and Cox regression models adjusted for prespecified covariates were used to estimate treatment effects within peak VO₂ strata, and treatment-by–peak VO₂ interactions were tested. Results Of 702 participants, 202 (29%) had peak VO₂ >20 and 500 (71%) had peak VO₂ ≤20 mL·kg⁻¹·min⁻¹. The composite primary endpoint did not differ significantly between rehabilitation and usual care in either stratum (interaction P = 0.62). Perioperative rehabilitation was associated with a lower incidence of postoperative pneumonia in both peak VO₂ >20 mL·kg⁻¹·min⁻¹ (adjusted OR 0.30; 95% CI 0.13–0.71) and ≤ 20 mL·kg⁻¹·min⁻¹ (adjusted OR 0.65; 95% CI 0.47–0.90), with no significant interaction between peak VO₂ category and treatment effect (P for interaction = 0.12). No significant effect modification was observed for other outcomes, including SICU stay, total hospitalization duration, or 3-month mortality. Conclusions In this exploratory analysis of the PORT trial, the reduction in postoperative pneumonia associated with short-term perioperative rehabilitation was consistent across strata of baseline peak VO₂. These findings suggest that the pneumonia benefit of respiratory-focused perioperative rehabilitation may not depend on baseline cardiorespiratory fitness, although adequately powered studies are required to definitively assess effect modification. Although the observed reduction in pneumonia was statistically significant and biologically plausible, it arose from a secondary exploratory analysis without adjustment for multiple comparisons and should therefore be interpreted with caution.