Author(s): FUJITA, K., FUJITA, Y., ADACHI, T., MARUTANI, Y., YOKOYAMA, T., Institution: THE UNIVERSITY OF OSAKA, Country: JAPAN, Abstract-ID: 1945

INTRODUCTION:
The UKK Walk Test (UKKWT) is a widely used field test for estimating maximal oxygen uptake (VO₂max) based on walk completion time, heart rate at the end of the walk, body mass index (BMI), and age. Although the original prediction equation was developed in European populations, its accuracy may differ across ethnic groups. We previously revised the prediction equation for Japanese individuals and demonstrated improved model performance compared with the original equation proposed by Oja et al.; however, the predictive performance of the revised model has not been externally validated. Therefore, this study aimed to develop and externally validate a VO₂max prediction model derived from the UKK Walk Test in Japanese adults and to compare its performance with the original equation.
METHODS:
A total of 145 healthy Japanese adults capable of independent walking (73 men and 72 women; age 47.2 ± 16.6 years) participated in this study. Participants were stratified by sex, age, and measured VO₂max and randomly assigned to a training group (n = 73) and a validation group (n = 72). Standardized mean differences for these variables were all <0.1, indicating appropriate balance between groups. In the training group, a VO₂max prediction model was developed using stepwise multiple regression analysis, with measured VO₂max as the dependent variable and UKK Walk Test parameters as independent variables. Predictive performance was assessed using root mean square error (RMSE) and coefficient of variation (CV). External validation was conducted in the validation group, and model performance was compared with that of the original equation proposed by Oja et al. Data analysis was performed using SAS software (version 9.4; SAS Institute Inc., Cary, NC, USA).
RESULTS:
The final prediction model included walk completion time, heart rate at the end of the walk, body fat percentage, and age as predictors. In the training group, the model demonstrated an RMSE of 4.4 ml/kg/min and a CV of 11.7%, indicating improved performance compared with the original equation. When externally validated in the independent validation group, the model yielded an RMSE of 4.5 ml/kg/min and a CV of 12.0%, showing predictive performance comparable to that observed in the training group. In contrast, the original equation showed lower predictive accuracy (RMSE: 6.5 ml/kg/min; CV: 17.3%).
CONCLUSION:
A VO₂max prediction model derived from the UKK Walk Test was successfully developed and externally validated in Japanese adults, demonstrating improved predictive performance compared with the original equation. Incorporating body composition variables, such as body fat percentage, improved prediction accuracy. These findings support the use of population-specific calibration to improve the accuracy of field-based cardiorespiratory fitness assessment.