The influence of adipose tissue thickness (ATT) on near-IR spectroscopy (NIRS) measurements in vivo was studied in the human flexor digitorum superficialis muscle at rest and during sustained isometric handgrip exercise. NIRS was used for the quantitative measurement of muscle O(2) consumption (mV.O(2)) and forearm blood flow (FBF) in 78 healthy subjects. Skinfold thickness ranged from 1.4 to 8.9 mm within the group. Resting mV.O(2) was 0.11+/-0.04 ml of O(2).min(-1).100 g(-1), and FBF was 1.28+/-0.82 ml.min(-1).100 ml(-1). There was a negative correlation (r=-0.70, P< or =0.01), indicating a decrease in mV.O(2) with increasing ATT. mV.O(2) in the 10 leanest subjects appeared to be twice as high as that in the 10 subjects with the highest ATT. A poor correlation (r=0.29, P< or =0.01) was found between ATT and FBF. The gender difference that we found for mV.O(2) was due to the difference in ATT between female and male subjects. No correlation was found between maximum voluntary contraction and mV.O(2), nor between maximum voluntary contraction and ATT, indicating that the contraction force did not confound our results. These results show that ATT has a substantial confounding influence on in vivo NIRS measurements, and that it is essential to incorporate this factor into future NIRS muscle studies in order to justify comparisons between different groups. To facilitate such comparisons, upper and lower boundaries for normal values of mV.O(2) and FBF in relation to ATT are presented.