31P MRS offers a unique view of muscle metabolism in vivo, but correct quantification is important. Inter-study correlation of estimates of [Pi] and [phosphocreatine (PCr)] in a number of published studies suggest that the main technical problem in calibrated 31P MRS studies is the measurement of PCr and Pi signal intensities, rather than absolute quantification of [ATP]. For comparison, we discuss the few published biopsy studies of calf muscle and a selection of the many studies of quadriceps muscle. The ATP concentration is close to the value that we obtained in calf muscle in our own study, presented here, on four healthy subjects, by localised 31P MRS using a surface coil incorporating an internal reference and calibrated using an external phantom. However, the freeze-clamp biopsy PCr concentration is approximately 20% lower than the value obtained by 31P MRS, consistent with PCr breakdown by creatine kinase during freezing. Finally, we illustrate some consequences of uncertainty in resting [PCr] for analysis of mitochondrial function from PCr kinetics using a published 31P MRS study of exercise and recovery: the lower the assumed resting [PCr], the lower the absolute rate of oxidative ATP synthesis estimated from the PCr resynthesis rate; in addition, the lower the assumed resting [PCr], or the higher the assumed [total creatine], the higher the apparent resting [ADP], and therefore the more sigmoid the relationship between the rate of oxidative ATP synthesis and [ADP]. Correct quantification of resting metabolite concentrations is crucially important for this sort of analysis. Our own results ([PCr] = 33 +/- 2 mM, [Pi] = 4.5 +/- 0.2 mM, and [ATP] = 8.2 +/- 0.4 mM; mean +/- SEM) are close to the overall mean values of the 10 published studies on calf muscle by 'calibrated' 31P MRS (as in the present work), and of [PCr] and [Pi] in a representative selection of 'uncalibrated' 31P MRS studies (i.e. from measured PCr/ATP and Pi/ATP ratios, assuming a literature value for [ATP]).
John Wiley & Sons, Ltd.