Résumé
Muscle energetics has been largely and quantitatively investigated using (31)P MRS. Various methods have been used to estimate the corresponding rate of oxidative ATP synthesis (ATP(ox)); however, potential differences among methods have not been investigated. In this study, we aimed to compare the rates of ATP production and energy cost in two groups of subjects with different training status using four different methods: indirect method (method 1), ADP control model (method 2) and phosphate potential control model (method 3). Method 4 was a modified version of method 3 with the introduction of a correction factor allowing for similar values to be obtained for the end-exercise oxidative ATP synthesis rate inferred from exercise measurements and the initial recovery phosphocreatine resynthesis rate. Seven sedentary and seven endurance-trained subjects performed a dynamic standardised rest-exercise-recovery protocol. We quantified the rates of ATP(ox) and anaerobic ATP synthesis (ATP(ana)) using (31)P MRS data recorded at 1.5 T. The rates of ATP(ox) over the entire exercise session were independent of the method used, except for method 4 which provided significantly higher values in both groups (p < 0.01). In addition, methods 1-3 were cross-correlated, thereby confirming their statistical agreement. The rate of ATP(ana) was significantly higher with method 1 (p < 0.01) and lower with method 4 (p < 0.01). As a result of the higher rate of ATP(ox), EC (method 4) calculated over the entire exercise session was higher and initial EC (method 1) was lower in both groups compared with the other methods. We showed in this study that the rate of ATP(ox) was independent of the calculation method, as long as no corrections (method 4) were performed. In contrast, results related to the rates of ATP(ana) were strongly affected by the calculation method and, more exactly, by the estimation of protons generated by ATP(ox). Although the absolute EC values differed between the methods, within- or between-subject comparisons are still valid given the tight relationships between them.