We studied 19 symptomatic female carriers of the Duchenne muscular dystrophy (DMD) gene. Most of these dystrophinopathy patients had had an erroneous or ambiguous diagnosis prior to dystrophin immunofluorescence testing. We assessed clinical severity by a standardized protocol, measured X-chromosome inactivation patterns in blood and muscle DNA, and quantitated the dystrophin protein content of muscle. We found that patients could be separated into two groups: those showing equal numbers of normal and mutant dystrophin genes in peripheral blood DNA ("random" X-inactivation), and those showing preferential use of the mutant dystrophin gene ("skewed" X-inactivation). In the random X-inactivation carriers, the clinical phenotype ranged from asymptomatic to mild disability, the dystrophin content of muscle was > 60% of normal, and there were only minor histopathologic changes. In the skewed X-inactivation patients, clinical manifestations ranged from mild to severe, but the patients with mild disease were young (5 to 10 years old). The low levels of dystrophin (< 30% on average) and the severe symptoms of the older patients suggested a poor prognosis for those with skewed X-inactivation, and they all showed morphologic changes of dystrophy. The random inactivation patients showed evidence of biochemical "normalization," with higher dystrophin content in muscle than predicted by the number of normal dystrophin genes. Seventy-nine percent of skewed X-inactivation patients (11/14) showed genetic "normalization," with proportionally more dystrophin-positive nuclei in muscle than in blood. In 65% of the skewed X-inactivation patients, dystrophin was not produced by dystrophin-positive nuclei; an average of 20% of myofiber nuclei were genetically dystrophin-positive but did not produce stable dystrophin. Biochemical normalization seems to be the main mechanism for rescue of fibers from dystrophin deficiency in the random X-inactivation patients. In the skewed X-inactivation patients, genetic normalization is active, but production failure of dystrophin by dystrophin-normal nuclei may counteract any effect of biochemical normalization. In the skewed X-inactivation patients, the remodeling of the muscle through cycles of degeneration and regeneration led to threefold increase in the number of dystrophin-competent nuclei in muscle myofibers (3.3 +/- 4.6), while dystrophin content was on the average 1.5-fold less then expected (-1.54 +/- 3.38). Our results permit more accurate prognistic assessment of isolated female dystrophinopathy patients and provide important data with which to estimate the potential effect of gene delivery (gene therapy) in DMD.