Albumin loss in on-line hemodiafiltration

Based on the increased hydraulic permeability of the new high permeability polyethersulfone membrane, DIAPES HF-800, we investigated the kinetics and handling of albumin in high volume on-line hemodiafiltration (HDF).

Seven patients on predilutional HDF were studied in two consecutive sessions. Blood flow rate and transmembrane pressure were continuously monitored. Spent dialysate was spilled at 20 ml/h every hour. Albumin was measured in blood and dialysate by immunonephelometry. Albumin and proteins adsorbed onto the dialyzer membrane were eluted after treatment with Triton X. Ultrafiltrates collected at 1 and 2 hours of treatment were pooled from different patients and incubated for 24 hours at 37 degrees C with bovine serum albumin (BSA). Total sulphydryl groups were evaluated using Ellmann's reagent [5, 5'-dithio-bis(2-nitrobenzoic acid)].

In all 7 patients, the total loss of albumin was 3.99 +/- 1.81 g, ranging between 1.09 and 6.82 g/session. Most albumin loss occurred in the first 60 min of pre-dilutional hemodiafiltration (1.92+0.83 g). There was no correlation between transmembrane pressure, urea clearance and the loss of albumin. Plasma water urea clearance values were stable over the treatment (234 +/- 14.3 ml/min). Plasma albumin concentration did not decrease during HDF sessions. Albumin adsorbed onto the dialyzers was 0.7 +/- 1.6 mg but the total amount of adsorbed proteins was much higher (130 + 90 mg). In addition, the ultrafiltrate collected during HDF sessions was able to induce oxidation of bovine serum albumin as measured by total protein sulfhydryl groups: bovine serum albumin incubated in the presence of ultrafiltrate collected at 1 hour had a sulfhydryl loss of 56.3 +/- 5.7% (p < 0.0001 vs control), and bovine serum albumin incubated with ultrafiltrate collected at 2 hours had a loss of 67.5 +/- 3.8% (p < 0.003 vs control).

The present study shows the high inter- and intra-patient variability of transmembrane passage of albumin in chronically uremic patients undergoing pre-dilutional HDF. Factors involved do not seem to be correlated to transmembrane pressure but rather to an interaction with the polymer surface. Albumin adsorption was minimal and was significantly lower than that of other plasma proteins. Albumin loss during HDF seemed to have no acute impact on plasma albumin. In addition, we demonstrated the presence of prooxidative compounds able to oxidize albumin, of which extracorporeal removal by HDF procedure could be beneficial for HD patients.