Automatic Detection of Anaerobic Threshold based on the Heart Rate Response

Abstract

0294 Several methods have been applied for detecting anaerobic threshold (AT) based on the lactate blood concentration as well as the respiratory gas exchange. In 1982, Conconi and co-workers proposed a non-invasive method based on the breakpoint of heart rate (HR) which is equivalent to the AT. PURPOSE: To determine a computerized method to detect the AT based on the HR response in continuous treadmill testing. METHODS: Twenty-two apparently healthy male subjects (24±6 yrs; 174±5 cm; 74±9 kg; 10±4% of body fat) were submitted to a progressive (1.6mph .2min−1) and maximal treadmill testing (Inbrasport, Brazil). The HR was continuously monitored (Polar S610, Finland) and stored at 0.2 Hz. The HR vs. time was plotted and evaluated independently (double blind) by 3 experienced observers for detecting visually (VM) the HR breakpoint. Automatic method (AM) consisted in fit 2 lines by least square method. Regression lines were calculated for all possible divisions into two contiguous groups, and the pair of lines yielding the least-pooled residual sum of squares was chosen as representing the best fit. The intersection point of the two regression lines assigned AT. The agreement among observers (rv) and AM reliability (rA) were expressed by the intraclass correlation coefficient. Mean HR (MHR) at the AT workload and the mean HR achieved in the final 30-s of the AT workload (HR30) were compared by non-parametric Kolmogorov-Smirnov test for independent samples significant results adopted α = 0.05. RESULTS: The MHR and HR30 achieved by observers averaged to 176±10bpm and 177±10bpm, respectively, corresponding to 83±5% and 84±6% of the HR reserve. Values found by visual discrimination presented weak but significant intraclass correlation coefficients for MHR (rv = 0.67) and HR30(rv = 0.68), which was not different (p>0.1). The AM detected the breakpoints in all subjects, resulting in rA = 0.61 (MHR, p<0.05) and rA = 0.71 (HR30, p<0.05), compared to VM. CONCLUSION: Automatic method is a potential tool for detecting AT based on HR response, where the repeatability can be guaranteed independently of the observer experience. In conclusion, the AM presented high reliability, so it can be applied as well as the VM.