Residual body-mass dependence in $$\dot{\text{V}}$$O₂peak normalization across ratio, allometric and regression–residual methods

Abstract Conventional ratio scaling of peak oxygen uptake ( $$\dot{\text{V}}{\text{O}}_{{\text{2}}}$$ peak; mL kg⁻ 1 min⁻ 1 ) assumes proportionality to body mass but may still retain size dependence. We compared ratio, fixed-0.75 allometric, and sex-specific regression–residual normalization using 937 treadmill cardiopulmonary exercise tests from 814 participants aged 15–59 years (798 tests/693 males; 139 tests/121 females) in the public University of Málaga archive. $$\dot{\text{V}}{\text{O}}_{{\text{2}}}$$ peak was the highest eligible 30-s average. Participant-cluster bootstrapping and repeated participant-separated five-fold cross-validation accounted for repeated observations and evaluated internal predictive performance. Ratio-scaled $$\dot{\text{V}}{\text{O}}_{{\text{2}}}$$ peak remained inversely correlated with body mass (male r = − 0.43; female r = − 0.26). Fixed-0.75 scaling retained an inverse association in males (r = − 0.27) but not clearly in females (r = − 0.03). Estimated mass exponents were 0.42 (95% CI 0.33–0.51) and 0.71 (0.54–0.87), respectively. Median held-out residual–mass correlations for W-scores were − 0.01 in both sexes. The male exponent decreased with one test per participant (0.40) and with RER restrictions (0.39 at ≥ 1.05; 0.34 at ≥ 1.10). The smaller female sample limited precision. Thus, regression–residual scores achieved statistical adjustment within the source archive. Absent standardized supramaximal verification, directly measured body composition, balanced sex representation, and external validation, the findings do not establish physiological, clinical, or normative validity.