HR Max Prediction Based on Age, Body Composition, Fitness Level, Testing Modality and Sex in Physically Active Population

Maximal heart rate (HRmax) is associated mostly with age, but age alone explains the variance in HRmax to a limited degree and may not be adequate to predict HRmax in certain groups. The present study was carried out on 3374 healthy Caucasian, Polish men and women, clients of a sports clinic, mostly sportspeople, with a mean age of 36.57 years, body mass 74.54 kg, maximum oxygen uptake (VO₂max, ml^∗kg^-1 ^∗min^-1) 50.07. Cardiopulmonary exercise tests (CPET) were carried out on treadmills or cycle ergometers to evaluate HRmax and VO₂max. Linear, multiple linear, stepwise, Ridge and LASSO regression modeling were applied to establish the relationship between HRmax, age, fitness level, VO₂max, body mass, age, testing modality and body mass index (BMI). Mean HRmax predictions calculated with 5 previously published formulae were evaluated in subgroups created according to all variables. HRmax was univariately explained by a 202.5-0.53^∗age formula (<i>R</i> ² = 19.18). The weak relationship may be explained by the similar age with small standard deviation (SD). Multiple linear regression, stepwise and LASSO yielded an <i>R</i> ² of 0.224, while Ridge yielded <i>R</i> ² 0.20. Previously published formulae were less precise in the more outlying groups of the studied population, overestimating HRmax in older age groups and underestimating in younger. The 202.5-0.53^∗age formula developed in the present study was the best in the studied population, yielding lowest mean errors in most groups, suggesting it could be used in more active individuals. Tanaka's formula offers the second best overall prediction, while the 220-age formula yields remarkably high mean errors of up to 9 bpm. In conclusion, adding the studied variables in multiple regression models improves the accuracy of prediction only slightly over age alone and is unlikely to be useful in clinical practice.