The pathogeneses of parathyroid disease in patients with uremia and nonfamilial primary parathyroid hyperplasia are poorly understood. Because of multigland involvement, it has been assumed that these common diseases predominantly involve polyclonal (non-neoplastic) cellular proliferations, but an overall assessment of their clonality has not been done. We examined the clonality of these hyperplastic parathyroid tumors using X-chromosome inactivation analysis with the M27 beta (DXS255) DNA polymorphism and by searching for monoclonal allelic losses at M27 beta and at loci on chromosome band 11q13. Fully 7 of 11 informative hemodialysis patients (64%) with uremic refractory hyperparathyroidism harbored at least one monoclonal parathyroid tumor (with a minimum of 12 of their 19 available glands being monoclonal). Tumor monoclonality was demonstrable in 6 of 16 informative patients (38%) with primary parathyroid hyperplasia. Histopathologic categories of nodular versus generalized hyperplasia were not useful predictors of clonal status. These observations indicate that monoclonal parathyroid neoplasms are common in patients with uremic refractory hyperparathyroidism and also develop in a substantial group of patients with sporadic primary parathyroid hyperplasia, thereby changing our concept of the pathogenesis of these diseases. Neoplastic transformation of preexisting polyclonal hyperplasia, apparently due in large part to genes not yet implicated in parathyroid tumorigenesis and possibly including a novel X-chromosome tumor suppressor gene, is likely to play a central role in these disorders.