The goals of this study were to determine whether abnormalities in phosphorus metabolism could be noninvasively detected using phosphorus-31 nuclear magnetic resonance spectroscopy in patients with dilated cardiomyopathy and left ventricular hypertrophy, and whether these patient groups could be distinguished from each other based on parameters obtained using this technique. Seventeen patients and 14 control subjects were studied using nuclear magnetic resonance spectroscopy. Spectra were obtained from the human heart at rest using 3-dimensional spectroscopic imaging as a localization technique. Data were acquired over an average volume of 48 cc in 26.3 minutes using a 2 tesla imaging and spectroscopy unit. The ratio of phosphocreatine to adenosine triphosphate was 0.89 +/- 0.88 (mean +/- standard error) in normal subjects and did not differ significantly in patients with dilated cardiomyopathy or left ventricular hypertrophy. A prominent peak in the phosphodiester region was seen much more frequently in patients with dilated cardiomyopathy, resulting in significantly higher ratios of phosphodiester to phosphocreatine (1.28 +/- 0.35) and phosphodiester to adenosine triphosphate (0.79 +/- 0.18) in this group compared to normal subjects (0.33 +/- 0.08 and 0.29 +/- 0.08, respectively). However, the various patient groups could not be reliably distinguished from each other based on spectral patterns. These studies demonstrate the feasibility of performing phosphorus-31 nuclear magnetic resonance spectroscopic imaging in patients with myocardial disease. The initial results indicate that, under resting conditions, the ratio of phosphocreatine to adenosine triphosphate is not consistently altered in patients with severe global cardiomyopathies or hypertrophy. Phosphodiesters are elevated in some patients with dilated cardiomyopathy, a finding that may signify abnormal phospholipid metabolism in this condition.