Dialysis membrane-dependent removal of middle molecules during hemodiafiltration: the beta2-microglobulin/albumin relationship

Clin Nephrol. 2004 Jul;62(1):21-8. doi: 10.5414/cnp62021.


Aim: Current hemodialysis therapy modalities such as online hemodiafiltration (HDF) attempt to enhance solute removal over a wide molecular weight range through a combination of diffusion and convection. While the effects of variations of treatment modalities and conditions have been studied reasonably well, few studies have examined the efficacy of HDF to remove middle molecules in relation to the dialyzer and membrane characteristics. In this investigation, diverse high-flux dialyzers, covering a wide range of membrane permeabilities, were compared under identical in vivo conditions to assess their ability to eliminate larger uremic retention solutes (using beta2-microglobulin as a surrogate of middle molecules) without simultaneously causing excessive leakage of useful proteins such as albumin.

Patients and methods: In a prospective, crossover study, 3 ESRD patients were treated with 8 different brands of high-flux dialyzers at 4 different ultrafiltration (UF)/substitution flow rates (QS: 0, 30, 60, 90 ml/min) in post-dilution HDF mode. Thus, each patient underwent 32 treatment sessions, with a total of 96 treatment sessions conducted during the entire clinical study. Albumin and beta2-microglobulin levels were measured in both, dialysate and blood. Both, albumin and beta2-microglobulin elimination was dependent upon the permeability of the dialysis membrane as well as on the ultrafiltration/substitution flow rates applied.

Results: At the maximum UF rate of 90 ml/min, the total albumin loss (measured in the dialysate) ranged from 300 mg/4 h (for the FLX-15 GWS dialyzers) to 7,000 mg/4 h (for the BS-1.3U dialyzers). Up to 50% reduction of albumin occurred within the first 30 minutes of the dialysis treatment, and the leakage of albumin increased exponentially with increasing UF rates as well as increasing transmembrane pressure (TMP). The various dialyzers could be classified according to their UFR-dependent beta2-m reduction rates (RR), into low (< 50%; FLX-15 GWS, CT 150G), medium (50-70%; Polyflux 14 S, BLS 814SD, H4) and high (> 70%; BS-1.3U, APS 650, FX 60) removers of middle molecules. One dialyzer type (CT 150G) showed extremely low beta2-m RR and relatively high albumin losses. Most membranes, however, showed either low albumin leakage coupled with low beta2-m removal, or high beta2-m RR but at the expense of considerable albumin leakage. Only 2 membrane types approached the desired balance between high to medium beta2-m RR while simultaneously restricting the albumin leakage especially at higher filtration/substitution rates.

Conclusion: Our investigations demonstrate that not all dialysis membranes classified as "high-flux" are comparable in their ability to specifically and efficiently remove middle molecules, or curtail the unwanted excessive leakage of essential proteins from the patient's blood. Thus, the selection of appropriate high-flux dialyzers for specific patient requirements should be based more upon clinical evaluations and analyses rather than on product specifications alone.

Publication types

  • Clinical Trial
  • Comparative Study
  • Randomized Controlled Trial

MeSH terms

  • Albumins / analysis*
  • Cross-Over Studies
  • Female
  • Hemodiafiltration / methods*
  • Humans
  • Kidney Failure, Chronic / therapy*
  • Male
  • Membranes, Artificial*
  • Middle Aged
  • Prospective Studies
  • beta 2-Microglobulin / analysis*


  • Albumins
  • Membranes, Artificial
  • beta 2-Microglobulin