Sachio Usui

Non-invasive automatic determination of anaerobic threshold using double gas exchange parameters

A new technique for determining the anaerobic threshold (AT) in real-time during a progressive exercise test was proposed with special consideration to its safety. We applied the discriminant analysis based on the linear logistic regression model to two groups of gas exchange parameters at times just before and after the lactate (LA) threshold (determined by blood LA kinetics; the external criterion of AT) during the incremental exercise test in 21 young males. As a result, the selected variables for estimating AT accurately (the discriminant rate = 100%), were VE/VO2 and the relative change in VE. Our idea of the new technique for determining AT is based on the probability function (p(x)) derived from the discriminant analysis. At each time at which the gas exchange parameters are acquired during the exercise test, p(x) is calculated, and the decision whether the present exercise level is in the pre, just, or post-AT level is made by the p(x). The technique is able to stop the exercise test as quickly as possible after passing through the AT level. To evaluate the new technique, the data set of 66 females (aged 16-57) was applied as a form of simulation. It was shown that the exercise test could be stopped by the technique at a level of only 0.2-0.5 l/min in VO2 higher than AT. This advantage of the technique makes the exercise test safer for middle-aged or old men, and patients during cardiac rehabilitation.