Non-invasive prediction of the first ventilatory threshold in Chinese patients with chronic heart failure for personalized exercise prescription

Purpose This study aimed to develop non-invasive prediction models for the first ventilatory threshold (VT 1 ) in patients with chronic heart failure (CHF), to address the inaccuracy of guideline-recommended, percentage-based exercise intensity prescriptions (40%–69% peak VO 2 , 55%–75% peak HR) and the limited accessibility of gold-standard cardiopulmonary exercise testing (CPET). Methods We analyzed 225 CHF patients who underwent standardized CPET. Using multivariate linear regression with ten-fold cross-validation, we developed prediction models for VT 1 oxygen consumption (VO 2 ) and heart rate (HR) based on readily available clinical parameters, including resting/peak exercise data, demographics, comorbidities, and medication. Model performance was evaluated using R 2 , root mean squared error (RMSE), mean absolute error (MAE), intraclass correlation coefficient (ICC), and Bland-Altman analysis. Results Traditional percentage-based methods were inadequate: 80% of patients achieved VO 2 -VT 1 at 60%–90% of peak VO 2 , and 75.6% reached HR-VT 1 at 70%–90% of peak HR. The VO 2 -VT 1 prediction model showed strong agreement ( R 2 = 0.65, RMSE = 1.42 mL/kg/min, ICC = 0.79). The HR-VT 1 model demonstrated moderate-to-strong agreement ( R 2 = 0.57, RMSE = 8.4 bpm, ICC = 0.71). Bland-Altman analysis indicated good agreement for both models (Within LoA: 95.1% and 95.6%). Conclusion CHF patients exhibit distinct exercise intensity patterns. Our validated models enable accurate, individualized estimation of VT 1 parameters using basic clinical and exercise test data, offering a practical alternative to full CPET for personalized exercise prescription in resource-limited settings.