Intact PTH measurements between 1989-96 in a 116-patient (63 HD, 53 PD) dialysis unit were reviewed to evaluate the determinants of PTH. Prophylactic treatment included calcium carbonate and maximal alphacalcidol therapy. A forward stepwise multiple regression analysis showed that duration of dialysis, phosphate, albumin and chronic glomerulonephritis were independently positively correlated with PTH, and that ionized calcium, parathyroidectomy, age, diabetic nephropathy and systemic disease were independently negatively correlated. During the first five years of dialysis PTH rose from 124 ng/L (SD range 33-462) to 160 (63-402)* in HD patients but fell from 119 ng/L (33-423) to 88 (31-251)** in PD patients despite the less intensive treatment. PTH fell with increasing age (40-50 years 173 ng/L (52-575); > 60 years 100 (31-316)**) and hypoalbuminemia (< 400 micromol/L 68 ng/L (22-209); > 550 pmol/1138 (41-457)**). PD patients were generally older and increasingly malnourished; after correcting for these factors, no influence of dialysis modality on PTH could be seen. Low-calcium dialysate (1.25 mmol/L) was introduced for both dialysis groups in 1994. Consequent to this, aluminium consumption was virtually eliminated and consumption of alphacalcidol increased. PTH fell from 161 to 128 ng/L in HD patients but rose from 119 to 135 ng/L in PD patients. The incidence of parathyroidectomy fell from 4.3%/year to 1.5%/year*. CONCLUSIONS Malnourishment and increasing age reduce PTH secretion and are important determinants of hyperparathyroidism during maintenance dialysis. Adynamic bone disease is common in PD patients and associated with low PTH levels, and may be a consequence of malnourishment and involutional changes. The introduction of low-calcium dialysate reduced the incidence of parathyroidectomy. Control of hyperparathyroidism improved in HD but not PD patients.