Background. As the main mineral reservoir, bone acts as a calcium (Ca) and phosphate buffering system. Accordingly, phosphate removal by haemodialysis (HD) might be theoretically influenced by bone turnover, as well as by the interaction of regulatory molecules, such as PTH and osteoprotegerin (OPG). The present study investigated the relationship between these variables and phosphate removal by HD. Methods. Blood samples for serum Ca, phosphate, bicarbonate, intact PTH, PTH (1-84), bone alkaline phosphatase, tartrate-resistant acid phosphatase 5b, OPG and receptor activator of nuclear factor-kB ligand (RANKL) were obtained in 28 HD patients. Phosphate removal was measured by a continuous collection of the dialysate. Results. Pre-dialysis serum phosphate concentration is the critical factor in determining dialytic phosphate removal. However, multiple regression analysis reveals that phosphate removal is better explained by a combination of factors than by phosphate concentration alone. In this model, the PTH/OPG ratio is an additional positive factor, whereas age and vitamin D treatment are negative factors. Patients with pre-HD bicarbonate higher than 20 mEq/l had higher serum phosphate and, accordingly, higher phosphate removal; of interest, these individuals also have significant differences in RANKL/OPG. Mean (SD) OPG levels were significantly higher than that in the healthy population (16.2 (12.5) pmol/l; these values correlated with age (r ¼ 0.4, P<0.04). Mean serum RANKL (1.03 (1.02) pmol/l) was within the range of normal individuals.Conclusions. Dialytic phosphate removal has a crucial, direct relationship with pre-HD plasma phosphate levels. However, the phenomenon of phosphate removal is more precisely explained using a more complex relationship, defined by the interaction between serum phosphate, PTH/OPG, age and vitamin D administration. Serum RANKL levels are first reported in HD patients, and are not different from the normal population.