Accurate prediction equations for ventilatory thresholds in cardiometabolic disease when gas exchange analysis is unavailable: development and validation

To develop and validate equations predicting heart rate (HR) at the first and second ventilatory thresholds (VTs) and an optimized range-adjusted prescription for patients with cardiometabolic disease (CMD). To compare their performance against guideline-based exercise intensity domains.

Cross-sectional study involving 2868 CMD patients from nine countries. HR predictive equations for first and second VTs (VT_1, VT₂) were developed using multivariate linear regression with 975 cycle-ergometer cardiopulmonary exercise tests (CPET). ‘Adjusted’ percentages of peak HR (%HR_peak) and HR reserve (%HRR) were derived from this group. External validation with 1893 CPET (cycle-ergometer or treadmill) assessed accuracy, agreement, and reliability against guideline-based %HR_peak and %HRR prescriptions using mean absolute percentage error (MAPE), Bland–Altman analyses, intraclass correlation coefficients (ICC). HR predictive equations (R²: 0.77 VT₁, 0.88 VT₂) and adjusted %HRR (VT₁: 42%, VT₂: 77%) were developed. External validation demonstrated superiority over widely used guideline-directed intensity domains for %HR_peak and %HRR. The new methods showed consistent performance across both VTs with lower MAPE (VT₁: 7.1%, VT₂: 5.0%), ‘good’ ICC for VT₁ (0.81, 0.82) and ‘excellent’ for VT₂ (0.93). Guideline-based exercise intensity domains had higher MAPE (VT₁: 6.8–21.3%, VT₂: 5.1–16.7%), ‘poor’ to ‘good’ ICC for VT₁, and ‘poor’ to ‘excellent’ for VT₂, indicating inconsistencies related to specific VTs across guidelines.

Developed and validated HR predictive equations and the optimized %HRR for CMD patients for determining VT₁ and VT₂ outperformed the guideline-based exercise intensity domains and showed ergometer interchangeability. They offer a superior alternative for prescribing moderate intensity exercise when CPET is unavailable.