High-flux membranes are not permeable to hepatitis B virus-DNA.

Dr. G.M.Th. de Jong, Drechtsteden Ziekenhuis, Lokatie Refaja, PO Box 444, NL-3300 AK Dordrecht (The Netherlands) Dear Sir, High-flux membranes have been found to be permeable to HBsAg [1] in contrast to cellulose membranes [2]. Backfiltration from inoculated dialysate has been proposed as a possible route of transmission of hepatitis B virus (HBV) in dialysis patients [1]. As HBV-DNA, rather than HBsAg, is the marker of infectivi-ty, we have investigated whether AN69 membranes are permeable to HBV-DNA. Fifteen HBsAg-positive patients were investigated. Renal replacement therapy consisted of hemodialysis in 9 and hemofiltration in 6 patients. AN69 dialyzers were used. Blood was investigated for HBeAg, anti-HBe (ELISA, Abbott Laboratories, North Chicago, Ill.) and HBV-DNA [molecular hybridization according to the method described by Shimizu et al. [3]. Dialysate and ultrafiltrate were tested for HBsAg after 5-fold concentration, while HBVDNA was determined after phenol extraction from 2.5 ml as well as from 25 ml of fluid. All 9 hemodialysis patients were HBeAg-positive; HBV-DNA could be demonstrated in blood of 7 patients. Of the hemofiltration patients, 6 were HBeAg-positive and 1 anti-HBepositive; all were negative for HBV-DNA in blood. HBsAg was not found in the dialysate or ultrafiltrate of any patient, nor could HBV-DNA sequences be demonstrated in these fluids. The difference between cellulose and high-flux membranes is a greater permeability in the latter to larger molecules, notably in the range of 5,000-50,000 daltons [4]. As the molecular weight of HBsAg is 24,000 daltons [5], it is to be expected that it can be demonstrated in the dialysate during treatment with high-flux membranes but not with cellulose membranes. The molecular weight of the fulllength HBV-DNA can be estimated to be about 2,000,000 daltons, while an intact HBV virion (Dane particle) has a diameter of 42 nm. Therefore, it cannot pass into the dialysate. The absence of HBsAg in the dialysate or ultrafiltrate was unexpected, although others also failed to demonstrate this in all seroposi-tive patients [1]. The sieving coefficient for HBsAg must have been lower than was estimated by the molecular weight. In this study we could not demonstrate HBV-DNA sequences in dialysate, making infectivity unlikely. However, overt