Clinical phenotypes of persons with mitochondrial DNA (mtDNA) mutations vary considerably. Therefore, diagnosing mitochondrial myopathy (MM) patients can be challenging and warrants diagnostic guidelines. (31)phosphorous magnetic resonance spectroscopy ((31)P-MRS) have been included as a minor diagnostic criterion for MM but the diagnostic strength of this test has not been compared with that of other commonly used diagnostic procedures for MM. To investigate this, we studied seven patients with single, large-scale deletions-, nine with point mutations of mtDNA and 14 healthy subjects, who were investigated for the following: 1) (31)P-MRS of lower arm and leg muscles before and after exercise, 2) resting and peak-exercise induced increases of plasma lactate, 3) muscle morphology and -mitochondrial enzyme activity, 4) maximal oxygen uptake (VO(2max)), 5) venous oxygen desaturation during handgrip exercise and 6) a neurological examination. All MM patients had clinical symptoms of MM, > 2% ragged red fibers in muscle, and impaired oxygen desaturation during handgrip. Fourteen of 16 patients had impaired VO(2max), 10/16 had elevated resting plasma lactate, and 10/11 that were investigated had impaired citrate synthase-corrected complex I activity. Resting PCr/P(i) ratio and leg P(i) recovery were lower in MM patients vs. healthy subjects. PCr and ATP production after exercise were similar in patients and healthy subjects. Although the specificity for MM of some (31)P-MRS variables was as high as 100%, the sensitivity was low (0-63%) and the diagnostic strength of (31)P-MRS was inferior to the other diagnostic tests for MM. Thus, (31)P-MRS should not be a routine test for MM, but may be an important research tool.