A methodology to assess the accuracy of a portable metabolic system (VmaxST).

Abstract

Validity of a portable metabolic system (VmaxST) was investigated during gas exchanges simulations by a mechanical system (GESS) and during human exercise. Three tests were conducted while gas exchanges were measured continuously by VmaxST. Test 1 was composed of six simulations of gas exchanges during steady-state exercise (20 min at V̇E = 80 L.min-1). Test 2 was composed of seven simulations of gas exchanges during incremental exercise (V̇O(2) from 300 to 5600 mL.min-1). In the human trial, 11 subjects performed an incremental running exercise on a treadmill while gas exchanges were measured at the end of each stage with the Douglas bag method (DBM). Test 1 showed that the VmaxST measurements were stable, despite inaccurate measurements of gas concentrations at the start of the test. During test 2, the mean error (difference between measured and predicted value) and the upper and lower limits of agreement were -8.0%, -12.6%, and -3.4% for V̇O(2); -4.6%, -12.0%, and +2.8% for V̇CO(2); and -0.7%, -4.7%, and +3.3% for V̇E. During the human trial, no significant difference was shown between V̇O(2) measured by VmaxST and by DBM at any stage of exercise. The mean difference and the upper and lower limits of agreement between the VmaxST and the DBM measurements were -0.5%, -14.3%, and +13.3% for V̇O(2); -6.3%, -20.9%, and +8.3% for V̇CO(2); and -9.9%, -25.5%, and +5.7% for V̇E. The use of GESS showed that measurements of V̇O(2) by VmaxST could be biased in a standardized condition. In more realistic condition of use, this bias was lower but the accuracy of measurements was impaired.