Blood flow to working skeletal muscle is frequently reduced in patients with heart failure or peripheral vascular disease. To determine if phosphorus nuclear magnetic resonance (NMR) can noninvasively detect such muscle underperfusion, gated phosphorus-31 NMR spectroscopy was used to compare muscle inorganic phosphate, phosphocreatine and pH during mild wrist flexion exercise at 0.2, 0.4 and 0.6 W in eight normal men, before and after reduction of forearm blood flow. Forearm flow was reduced by cuff inflation to a pressure determined by Doppler ultrasound to bring flow to 40 to 60% of control. Attention was focused on the inorganic phosphate to phosphocreatine (Pi/PCr) ratio and pH, two variables potentially sensitive to muscle oxygen delivery. At rest with normal flow, Pi/PCr averaged 0.12 +/- 0.03 and pH averaged 7.02 +/- 0.04. Exercise produced a progressive increase in Pi/PCr (0.2 W = 0.43 +/- 0.12; 0.4 W = 0.75 +/- 0.31; 0.6 W = 1.04 +/- 0.47) and a modest decrease in pH (0.2 W = 6.94 +/- 0.04; 0.4 W = 6.86 +/- 0.18; 0.6 W = 6.85 +/- 0.06). Flow reduction had no effect on Pi/PCr or pH at rest. In contrast, flow reduction during exercise was associated with higher Pi/PCr at all three work loads (0.2 W = 0.60 +/- 0.27; 0.4 W = 0.99 +/- 0.50; 0.6 W = 2.00 +/- 1.26 [all p less than 0.05 versus normal flow]) and lower pH (0.2 W = 6.78 +/- 0.12; 0.4 W = 6.69 +/- 0.23; 0.6 W = 6.65 +/- 0.30 [p less than 0.01 versus normal flow at 0.2 and 0.4 W; p = 0.05 at 0.6 W]).(ABSTRACT TRUNCATED AT 250 WORDS)