We performed a prospective study of 30 patients undergoing chronic hemodialysis to determine which of 6 generally available diagnostic procedures provided the most useful information for the assessment of bone disease in hemodialysis patients. The 6 procedures were: routine biochemical measurements, N-terminal parathyroid hormone (N-PTH), radiographic analysis of hands and clavicles, bone density determination by dual photon absorptiometry (DPA), deferoxamine stimulation test, and iliac crest bone biopsy. Serum N-PTH was elevated in 83% of patients but was not significantly associated with abnormalities of other biochemical parameters. No significant relationship was demonstrated between biochemical data and radiographic findings or between biochemical data and bone density by DPA. All patients with abnormal DPA had an elevation of N-PTH; therefore, DPA did not reveal any unsuspected disease. Bone biopsies were done in 20 patients and findings in each were consistent with uremic osteodystrophy, including osteitis fibrosa cystica in 11 patients and aluminum-associated bone disease in 2 patients. Six patients had mixed disease, and 1 patient had osteoporosis. Despite 11 positive deferoxamine tests, bone biopsy revealed aluminum deposition in only 7 of these patients, suggesting extraosseous aluminum accumulation in the remaining 4. Evaluation of the positive and negative predictive accuracies of DPA, x-ray analysis, N-PTH levels, and aluminum bone deposition revealed that normal DPA or x-ray findings do not exclude bone disease, that N-PTH level is a good marker for secondary hyperparathyroidism, and that a negative deferoxamine test excludes aluminum-associated bone disease. Discriminant analysis also reinforced these conclusions.(ABSTRACT TRUNCATED AT 250 WORDS)