Comparative analyses of predictive equations for peak oxygen uptake in Japanese cardiac rehabilitation patients

A nonexercise prediction equation for cardiovascular disease (CVD) patients was developed using the Fitness Registry and the Importance of Exercise National Database (FRIEND), which reported improved accuracy in predicting peak oxygen uptake (VO 2 ) in CVD patients. However, racial/ethnic differences have been observed in the predicted peak VO 2 . The purpose of this study was to evaluate the usefulness of the FRIEND CVD cohort equation in predicting peak VO 2 in Japanese cardiac rehabilitation patients. We retrospectively enroled 362 Japanese cardiac rehabilitation patients (phase II–III) who underwent cardiopulmonary exercise testing. We compared the measured peak VO 2 with the predicted peak VO 2 calculated from three prediction equations: the FRIEND CVD cohort equation, the FRIEND healthy cohort equation, and the Japanese healthy population equation. The measured peak VO 2 was 20.9 ± 5.2 mL/kg/min. Bland‐Altman analysis revealed that the FRIEND CVD cohort equation had the lowest bias (−2.32 mL/kg/min), while the FRIEND healthy cohort (11.11 mL/kg/min) and the Japanese healthy population equations (3.14 mL/kg/min) showed higher biases. The FRIEND CVD cohort equation had the highest correlation coefficient ( r = 0.61), coefficient of determination ( R 2 = 0.34), and intraclass correlation coefficient (ICC = 0.47) with measured peak VO 2 compared to other equations. In Japanese cardiac rehabilitation patients, the FRIEND CVD cohort equation provided more accurate predictions of peak VO 2 compared to the equation developed for the Japanese healthy population. However, prediction accuracy remained moderate or lower, indicating limitations for clinical use and the need for CVD‐specific equations across different racial and ethnic groups.