In order to compare hemofiltration (HF) and hemodialysis (HD) in connection with the risk of aluminum overload and renal osteodystrophy, double bone biopsies after double tetracycline labeling and a desferrioxamine test were performed in 12 patients on HF and 15 patients on HD. The aluminum concentration was low ( < 0.6 μmol/l) both in the dialysate and the substitution fluid. The duration of treatment (about 2 years) and the cumulative doses of Al(OH)3 and CaCO3 were comparable in the two groups. None of the patients was taking 1α-OH-D. The aluminum balance during an HF run ranged from – 22 to + 1.8 μmol/l, the balance being positive only when the plasma aluminum was < 0.5 μmol/l. Basal plasma aluminum and its increase induced by desferrioxamine were comparable in the two groups. Bone aluminum content was also comparable, but was about 10 times higher than in 7 nonuremic controls. Bone aluminum content and plasma aluminum increase after desferrioxamine were correlated to the Al(OH)3 cumulative dose. None of the patients had florid osteomalacia with increased osteoid thickness, and only 1 in each group had traces of stainable aluminum. The mineralization front was decreased in 8 of 12 HF and in 9 of 14 HD patients, so that no difference was observed between the means of the two groups. The predominant histological bone picture of the patients was osteitis fibrosa which was present in 10 of 12 HF and in 13 of 15 HD patients. Mean osteoclast count and active resorption surface were comparable in the two groups, but was increased (5–10 times the mean of the controls). Conclusions: Aluminum overload exists to a comparable degree in HF and HD patients. It is dependent upon the cumulative dose of Al(OH)3, not upon the procedure per se. After 2 years of treatment, it is not severe enough to induce florid osteomalacia. At this stage, the main histological lesion is osteitis fibrosa induced by hyperparathyroidism. This latter has the same severty in HF a