On-line haemodiafiltration (HDF) is a technique which combines diffusion with elevated convection and uses pyrogen-free dialysate as a replacement fluid. The purpose of this study was to evaluate the difference between conventional HDF (1-3 l/h) and on-line HDF (6-12 l/h). The study included 37 patients, 25 males and 12 females. The mean age was 56.5 +/- 13 years and duration of dialysis was 62.7 +/- 49 months. Three patients dropped out for transplantation, three patients died and three failed to complete the study period. Initially all patients were on conventional HDF with high-flux membranes over the preceding 34 +/- 32 months. Treatment was performed with blood flow (QB) 402 +/- 41 ml/min, dialysis time (Td) 187 min, dialysate flow (QD) 654 +/- 126 ml/min and replacement fluid (Qi) 4.0 +/- 2 l/session. Patients were changed to on-line HDF with the same filtre and dialysis time, QD 679 +/- 38 ml/min (NS), QB 434 +/- 68 ml/min (P < 0.05) and post-dilutional replacement fluid 22.5 +/- 4.3 l/session (P < 0.001). We compared conventional HDF with on-line HDF over a period of 1 year. Dialysis adequacy was monitored according to standard clinical and biochemical criteria. Kinetic analysis of urea and beta2-micro-globulin (beta2m) was performed monthly. Tolerance was excellent and no pyrogenic reactions were observed. Pre-dialysis sodium increased 2 mEq/l during on-line HDF. Plasma potassium, pre- and post-dialysis bicarbonate, uric acid, phosphate, calcium, iPTH, albumin, total proteins, cholesterol and triglycerides remained stable. The mean plasma beta2m reduction ratio increased from 56.1 +/- 8.7% in conventional HDF to 71.1 +/- 9.1% in on-line HDF (P < 0.001). The pre-dialysis plasma beta2m decreased from 27.4 +/- 8.1 to 24.2 +/- 6.5 mg/l (P < 0.01). Mean Kt/V (Daugirdas 2nd generation) was 1.35 +/- 0.21 in conventional HDF compared with 1.56 +/- 0.29 in on-line HDF (P < 0.01), Kt/Vr (Kt/V taking into consideration post-dialysis urea rebound) 1.12 +/- 0.17 vs 1.26 +/- 0.20 (P < 0.01), BUN time average concentration (TAC) 44.4 +/- 9 vs 40.6 +/- 10 mg/dl (P < 0.05) and protein catabolic rate (PCR) 1.13 +/- 0.22 vs 1.13 +/- 0.24 g/kg (NS). There was a significant increase in haemoglobin (10.66 +/- 1.1 vs 11.4 +/- 1.5) and haematocrit (32.2 +/- 2.9 vs 34.0 +/- 4.4%), P < 0.05, during the on-line HDF period, which allowed a decrease in the erythropoietin doses (3861 +/- 2446 vs 3232 +/- 2492 UI/week), (P < 0.05). Better blood pressure control (MAP 103.8 +/- 15 vs 97.8 +/- 11 mmHg, P < 0.01) and a lower percentage of patients requiring antihypertensive drugs were also observed. The change from conventional HDF to on-line HDF results in increased convective removal and fluid replacement (18 l/session). During on-line HDF treatment, dialysis dose was increased for both small and large molecules with a decrease in uraemic toxicity level (TAC). On-line HDF provided a better correction of anaemia with lower dosages of erythropoietin. Finally, blood pressure was easily controlled.
