Background: After hospitalization for chest pain, women are more likely than men to have normal coronary angiograms. In such women, myocardial ischemia in the absence of clinically significant coronary-artery obstruction has long been suspected. Most methods for the detection of the metabolic effects of myocardial ischemia are highly invasive. Phosphorus-31 nuclear magnetic resonance (31P-NMR) spectroscopy is a noninvasive technique that can directly measure high-energy phosphates in the myocardium and identify metabolic evidence of ischemia. Methods: We enrolled 35 women who were hospitalized for chest pain but who had no angiographically significant coronary-artery obstructions and 12 age and weight-matched control women with no evidence of heart disease. Myocardial high-energy phosphates were measured with 31P-NMR spectroscopy at 1.5 tesla before, during, and after isometric handgrip exercise at a level that was 30 percent of the maximal voluntary grip strength. We measured the change in the ratio of phosphocreatine to ATP during exercise. Results: Seven (20 percent) of the 35 women with chest pain and no angiographically significant stenosis had decreases in the phosphocreatine: ATP ratio during handgrip that were more than 2 SD below the mean value in the control subjects without chest pain. There were no significant differences between the two groups with respect to hemodynamic variables at rest and during handgrip, risk factors for ischemic heart disease, findings on magnetic resonance imaging and radionuclide perfusion studies of the heart, or changes in brachial flow during the infusion of acetylcholine. Conclusions: Our results provide direct evidence of an abnormal metabolic response to handgrip exercise in at least some women with chest pain consistent with the occurrence of myocardial ischemia but no angiographically significant coronary stenoses. (C) 2000, Massachusetts Medical Society.