A Machine Learning Approach to Developing an Accurate Prediction of Maximal Heart Rate During Exercise Testing in Apparently Healthy Adults.

PURPOSE: Maximal heart rate (HR max ) continues to be an important measure of adequate effort during an exercise test. The aim of this study was to improve the accuracy of HR max prediction using a machine learning (ML) approach. METHODS: We used a sample from the Fitness Registry of the Importance of Exercise National Database, which included 17 325 apparently healthy individuals (81% males) who performed a maximal cardiopulmonary exercise test. Two standard formulas for HR max prediction were tested: Formula1 = 220 - age (yr), root-mean-squared error (RMSE) 21.9, relative root-mean-squared error (RRMSE) 1.1; and Formula2 = 209.3 - 0.72 × age (yr), RMSE 22.7 and RRMSE 1.1. For ML model prediction, we used age, weight, height, resting HR, and systolic and diastolic blood pressure. The following ML algorithms to predict HR max were applied: lasso regression (LR), neural networks (NN), support vector machine (SVM) and random forests (RF). An evaluation was performed using cross-validation and by computing the RMSE and RRMSE, Pearson correlation, and Bland-Altman plots. The best predictive model was explained with Shapley Additive Explanations (SHAP). RESULTS: The HR max for the cohort was 162 ± 20 bpm. All ML models improved HR max prediction and reduced RMSE and RRMSE compared with Formula1 (LR: 20.2%, NN: 20.4%, SVM: 22.2%, and RF: 24.7%). The predictions of all algorithms significantly correlated with HR max ( r = 0.49, 0.51, 0.54, 0.57, respectively; P < .001). Bland-Altman analysis demonstrated lower bias and 95% CI for all ML models in comparison with standard equations. The SHAP explanation showed a high impact of all selected variables. CONCLUSIONS: Machine learning, particularly the RF model, improved prediction of HR max using readily available measures. This approach should be considered for clinical application to refine HR max prediction.

The interface of hypothalamic-pituitary-adrenocortical axis and circulating brain natriuretic peptide in prediction of cardiopulmonary performance during physical stress.

Brain natriuretic peptide (NT-pro-BNP) was implicated in the regulation of hypothalamic-pituitary-adrenocortical (HPA) responses to psychological stressors. However, HPA axis activation in different physical stress models and its interface with NT-pro-BNP in the prediction of cardiopulmonary performance is unclear. Cardiopulmonary test on a treadmill was used to assess cardiopulmonary parameters in 16 elite male wrestlers (W), 21 water polo player (WP) and 20 sedentary age-matched subjects (C). Plasma levels of NT-pro-BNP, cortisol and adrenocorticotropic hormone (ACTH) were measured using immunoassay sandwich technique, radioimmunoassay and radioimmunometric techniques, respectively, 10min before test (1), at beginning (2), at maximal effort (3), at 3rdmin of recovery (4). In all groups, NT-pro-BNP decreased between 1 and 2; increased from 2 to 3; and remained unchanged until 4. ACTH increased from 1 to 4, whereas cortisol increased from 1 to 3 and stayed elevated at 4. In all groups together, ΔNT-pro-BNP2/1 predicted peak oxygen consumption (B=37.40, r=0.38, p=0.007); cortisol at 3 predicted heart rate increase between 2 and 3 (r=-0.38,B=-0.06, p=0.005); cortisol at 2 predicted peak carbon-dioxide output (B=2.27, r=0.35, p<0.001); ΔACTH3/2 predicted peak ventilatory equivalent for carbon-dioxide (B=0.03, r=0.33, p=0.003). The relation of cortisol at 1 with NT-pro-BNP at 1 and 3 was demonstrated using logistic function in all the participants together (for 1/cortisol at 1 B=63.40, 58.52; r=0.41, 0.34; p=0.003, 0.013, respectively). ΔNT-pro-BNP2/1 linearly correlated with ΔACTH4/3 in WP and W (r=-0.45, -0.48; p=0.04, 0.04, respectively). These results demonstrate for the first time that HPA axis and NT-pro-BNP interface in physical stress probably contribute to integrative regulation of cardiopulmonary performance.

Brain natriuretic peptide predicts forced vital capacity of the lungs, oxygen pulse and peak oxygen consumption in physiological condition.

Brain natriuretic peptide (NT-pro-BNP) is used as marker of cardiac and pulmonary diseases. However, the predictive value of circulating NT-pro-BNP for cardiac and pulmonary performance is unclear in physiological conditions. Standard echocardiography, tissue Doppler and forced spirometry at rest were used to assess cardiac parameters and forced vital capacity (FVC) in two groups of athletes (16 elite male wrestlers (W), 21 water polo player (WP)), as different stress adaptation models, and 20 sedentary subjects (C) matched for age. Cardiopulmonary test on treadmill (CPET), as acute stress model, was used to measure peak oxygen consumption (peak VO2), maximal heart rate (HRmax) and peak oxygen pulse (peak VO2/HR). NT-pro-BNP was measured by immunoassey sandwich technique 10min before the test - at rest, at the beginning of the test, at maximal effort, at third minute of recovery. FVC was higher in athletes and the highest in W (WP 5.60±0.29 l; W 6.57±1.00 l; C 5.41±0.29 l; p<0.01). Peak VO2 and peak VO2/HR were higher in athletes and the highest in WP. HRmax was not different among groups. In all groups, NT-pro-BNP decreased from rest to the beginning phase, increased in maximal effort and stayed unchanged in recovery. NT-pro-BNP was higher in C than W in all phases; WP had similar values as W and C. On multiple regression analysis, in all three groups together, ΔNT-pro-BNP from rest to the beginning phase independently predicted both peak VO2 and peak VO2/HR (r=0.38, 0.35; B=37.40, 0.19; p=0.007, 0.000, respectively). NT-pro-BNP at rest predicted HRmax (r=-0.32, B=-0.22, p=0.02). Maximal NT-pro-BNP predicted FVC (r=-0.22, B=-0.07, p=0.02). These results show noticeable predictive value of NT-pro-BNP for both cardiac and pulmonary performance in physiological conditions suggesting that NT-pro-BNP could be a common regulatory factor coordinating adaptation of heart and lungs to stress condition.