External Validation and Local Adaptation of Estimated Pulse Wave Velocity Models in a Brazilian Population.

Abstract

Introduction: Estimated pulse wave velocity (ePWV) is a non-invasive and low-cost method that uses age and blood pressure to assess cardiovascular risk. However, as it is not a true measure of arterial mechanics, its accuracy in diverse cohorts, such as Brazilian populations, is uncertain due to inherent genetic and environmental differences.

Aim: to externally validate existing European ePWV equations and develop new, population-specific models for ePWV estimation in a large, admixed Brazilian cohort.

Methods: We analyzed data from 2122 Brazilian adults, assessing carotid-femoral pulse wave velocity (cfPWV), age, mean arterial pressure (MAP), and sex. Linear regression models were developed, incorporating all these variables as predictors. Model performance was evaluated using root mean square error (RMSE) and coefficient of determination (R²). Bland-Altman analyses assessed agreement between estimated and measured cfPWV.

Results: The newly developed equations demonstrated superior performance compared to existing European models. Our best-performing model (Equation 2) achieved an RMSE of 0.744 in individuals without cardiovascular risk factor, demonstrating superior performance to the model derived by the Arterial Stiffness Collaboration Group (ASCG) (RMSE: 0.879). Inclusion of sex as a predictor further improved model accuracy. Bland-Altman analyses revealed narrower limits of agreement for the new models. Notably, higher prediction errors were observed in subgroups underrepresented in the sample, such as individuals with very high cfPWV, advanced age, or elevated MAP.

Conclusions: Population-specific ePWV equations tailored to the Brazilian cohort provide more accurate estimations of arterial stiffness. This improved precision is clinically meaningful, allowing for better stratification of cardiovascular risk using a low-cost tool readily applicable in the public health system. These findings underscore the importance of developing and validating cardiovascular risk assessment tools within diverse populations to enhance predictive accuracy and clinical utility.