comparison of the new polyethersulfone high-flux membrane

8/4/2019 Comparison of the New Polyethersulfone High-flux Membrane

1/5

Nephrol Dial Transplant (2003) 18: 23822386

DOI: 10.1093/ndt/gfg410

Original Article

Comparison of the new polyethersulfone high-flux membrane

DIAPES

HF800 with conventional high-flux membranes duringon-line haemodiafiltration

Walter Samtleben1, Christina Dengler1, Birgit Reinhardt2, Annekatrin Nothdurft2 andHorst-Dieter Lemke2

1Department of Nephrology, University Hospital Munich-Grosshadern, Munich and 2Membrana Research,

Obernburg, Germany

AbstractBackground. Current modalities of renal replacementtherapy allow only a limited removal of larger, possiblytoxic molecules, which accumulate in uraemia.Recently, a haemodiafilter has been made availablewith the new, high-flux, polyethersulfone-basedmembrane DIAPES HF800. We performed a studyto compare DIAPES HF800 with two conventionalhigh-flux membranes in on-line haemodiafiltration(HDF), with respect to the removal properties forthe two marker proteins, 2-microglobulin (2m,11.8 kDa) and albumin (66.5 kDa).Methods. In a prospective, controlled study 10 stable

end-stage renal disease patients were randomly allo-cated to 30 sessions of post-dilutional on-line HDFwith three types of steam-sterilized membranes:DIAPES HF800, polysulfone and polyamide. Bloodflow rate was 250 ml/min and treatment time was240 min. Pre-treatment 2m and albumin plasmaconcentrations did not differ between the threegroups. The concentration of the two proteins wasdetermined before and after treatment in plasma aswell as in the continuously collected haemodiafiltrate.Results. Tolerance of all treatments was very good,without any side-effects for all filters. The mean plasmareduction rate of 2m was 77 1% for DIAPES

HF800 and polysulfone whereas it was 71 1% for

polyamide (P < 0.05). The mean 2m amount removedand found in the haemodiafiltrate per session was230 14mg for DIAPES HF800, 186 13mg forpolysulfone and 147 13 mg for polyamide (P < 0.05between each pair of membranes). The same rankingwas obtained for albumin removed and found inhaemodiafiltrate per session for the three membranes:5.7 0.4, 3.5 0.4 and 1.0 0.4 g, respectively.

Although DIAPES

HF800 showed the highestvalue for albumin in haemodiafiltrate the mean post-treatment plasma albumin was higher after thetreatment with DIAPES HF800 compared with theother membranes (P < 0.05).Conclusions. On-line HDF has shown to achieveplasma reduction rates for 2m of up to 77% for theDIAPES HF800 membrane and for polysulfone. Theamounts of2m and albumin in haemodiafiltrate weremuch higher for DIAPES HF800 than for the othertwo membranes indicating a greater permeability formolecules up to a molecular weight of 66.5 kDa. Thiscould, at least theoretically, offer the advantage also toremove uraemic toxins in the molecular weight rangeof albumin or of albumin-bound toxins. The futuremust show whether this will counterbalance the loss ofalbumin.

Keywords: albumin; 2-microglobulin; haemodia-filtration; high-flux membrane; polyethersulfone;post-dilution

Introduction

The introduction of a newly developed, highly perme-

able haemodialysis (HD) membrane arises interest intodays nephrological community. The question isoften considered as to whether a membrane withhigher permeability may fulfil an improved balancebetween a higher removal of uraemic toxins in themolecular weight range from urea to small proteinssuch as 2-microglobulin (2m) and a clinicallytolerable albumin loss. In recent years, the choice ofmembranes of either synthetically modified cellulosicor synthetic membranes has expanded to a largeextent. In parallel, a trend towards treatment modeswith higher plasma reduction rates has recently been

Correspondence and offprint requests to: Horst-Dieter Lemke,Forschungsstrasse, Membrana GmbH, D-63785 Obernburg,Germany. Email: [email protected]

Nephrol Dial Transplant 18(11)

ERAEDTA 2003; all rights reserved


2/5

witnessed [1,2]. The reason for such a trend may bederived from a revived interest in the field of uraemictoxins. Apart from being small, dialysable molecularsubstances, uraemic toxins may be bound to plasmaproteins such as albumin or can be proteins itself, like2m (11.8 kDa). The removal of plasma proteins up to

the 66.5 kDa range of albumin has been investigated byseveral authors employing different membranes anddialysis modalities [35]. The factors involved in theremoval of low molecular weight proteins dependnot only on membrane permeability, but also on thetransmembrane pressure applied during treatment [6]and on patient specific variables.

Recently, DIAPES HF800, a highly permeablemembrane made of polyethersulfone (PES) has beendeveloped. DIAPES HF800 exhibits an increasedhydraulic permeability compared with conventionalhigh-flux membranes. This leads to an in vitroultrafiltration rate of 80 ml/hmmHg for DIAPES

HF800 in the BLS819SD haemodiafilter (Bellco)compared with 55 ml/hmmHg for the polysulfonemembrane in the Hemoflow HF80S (FreseniusMedical Care). In this study, we evaluated the removalof 2m and albumin in patients undergoing post-dilutional on-line haemodiafiltration (HDF) usingDIAPES HF800 in comparison with two othersynthetic membranes.

Subjects and methods

Patients

The study was prospective, randomized and cross-over innature and the study protocol was approved by the local ethics

committee. Ten chronic uraemic patients on a regular thriceweekly HD programme were enrolled after giving informedconsent [seven males, three females; aged 51 13 years (range

3370 years); 42 37 months on dialysis (range 5118months)]. The patients were randomly treated once withDIAPES HF800 (Membrana GmbH, Wuppertal, Germany;BLS819SD, 1.8 m2, steam-sterilized, Bellco S.p.A.,Mirandola,

Italy), and once with polyamide (Polyflux 17S, Gambro,Lund, Sweden) and polysulfone (Hemoflow HF80S,Fresenius Medical Care, Bad Homburg, Germany), respec-tively, in a 4-h post-dilutional on-line HDF regime

(Gambro AK100 Ultra monitor). Treatments under investi-

gation were once a week after the long dialysis-free interval.Details of the HDF treatment parameters and of thehaemodiafilters are listed in Table 1. One patient had to be

admitted to the hospital after the second study treatmentfor reasons, which were not related to the study. As a

consequence, one treatment with DIAPES

HF800 could notbe performed.

Determination of albumin and 2m

In order to study the removal of2m and albumin with the

haemodiafiltrate, a T-connector was introduced into theeffluent haemodiafiltrate line allowing continuous collectionof haemodiafiltrate at a flow rate of 1 ml/min into a separate

beaker during the entire duration of the treatment (Figure 1).Samples were drawn from this haemodiafiltrate pool aftereach hour of treatment. Before sampling, the pool was mixed

well by a magnetic stirrer.

The mass of solute transferred to the haemofiltrate aftera given time t was calculated from:

mass of soluteQout * t * ct

where Qout is the outlet haemofiltrate flow rate, consisting ofthe dialysate flow rate Qdial/in, the substitution fluid flow rate

Qsub and the net ultrafiltration flow rate QUF as shown inFigure 1. ct is the concentration of the solute in the

cumulated haemodiafiltrate pool after time t.The plasma levels of 2m and albumin were measured

directly before (pre, t 0) and after (post, t 240min) thetreatment. 2m and albumin in haemodiafiltrate and in

plasma were measured by laser nephelometry (BN 100Analyzer, Dade-Behring Marburg GmbH, Marburg,

Germany) using N Latex 2-microglobulin and N Anti-serum to human albumin (Dade-Behring Marburg GmbH),

respectively. The 2m post-haemodiafilter concentration wascorrected for haemoconcentration according to Bergstromand Wehle [7].

Statistics

Differences between parameters were tested by multivariateANOVA and Duncans multiple range test. A P value of


3/5

Results

The absolute amount of 2m and of albumin detect-able at the end of the treatment in the cumulativehaemodiafiltrate pool as well as the pre- and post-treatment plasma values for albumin are shown inTable 2. Both total 2m and total albumin found in thehaemodiafiltrate decreased in the sequence DIAPES

HF800 > polysulfone > polyamide. There was a signi-ficant difference for both, 2m and albumin in the

haemodiafiltrate between the three membranes. Allother data, especially the respective pre-treatmentvalues, showed no significant differences between thethree membranes.

While the DIAPES HF800 haemodiafilter reducedthe plasma 2m level by 77 1% of the pre-dialysisvalue (identical to the reduction with polysulfone),polyamide achieved only a reduction rate of 71 1%(P < 0.05) as shown in Figure 2.

The amount of 2m in haemodiafiltrate corre-sponded with the amount of albumin in haemodiafil-trate. All membranes showed an increase of the

albumin concentration in haemodiafiltrate during thelast 2 h of treatment.

Pre-dialysis plasma albumin values and ultrafiltra-tion volumes were comparable for the three mem-branes, the latter despite the fact that in all casesthe ultrafiltration rates were set individually in orderto achieve the patients dry weights (see Table 1).Although DIAPES HF800 showed the highestalbumin loss into haemodiafiltrate, interestingly, thepost-dialysis plasma albumin level after the DIAPES

HF800 treatments was significantly higher comparedwith the other two haemodiafilters (Table 2).

Discussion

Treatments with DIAPES HF800 used in the HDFmode were tolerated very well without any side-effectsand did not differ in this regard from those withestablished membranes. Average blood pressure wasslightly increased before and normalized during treat-ments in all three groups as expected. Blood and

Fig. 1. HDF set-up and haemodiafiltrate collection scheme.

Table 2. 2M, albumin and blood pressure pre- and post-HDF (mean SEM)

DIAPES HF 800 Polysulfone Polyamide S

2m in haemodiafiltrate (mg) 23014a 18613a 14713a

Albumin in haemodiafiltrate (g) 5.70.4a 3.50.4a 1.00.4a

Plasma albumin concentration pre-HDF (g/l) 35.80.4 35.60.4 35.70.4Plasma albumin concentration post-HDF (g/l) 43.40.6b 40.90.5 41.20.5Systolic/diastolic blood pressure (mmHg) pre-HDF 1469/864 1427/785 1377/785Systolic/diastolic blood pressure (mmHg) post-HDF 1238/815 1349/794 13011/795

aSignificantly different from each other (P < 0.05).bSignificantly different from polysulfone and polyamide S (P < 0.05).

2384 W. Samtleben et al.


4/5

haemodiafiltrate flow rates, ultrafiltration and infusionvolumes, pre-treatment 2m and albumin concentra-tions were not significantly different between the threemembranes under investigation. Therefore, thesefactors did neither influence the removal of 2m norof albumin during treatment.

In conventional HD with high-flux polysulfone anda blood flow of 300 ml/min a 2m plasma reductionrate of$50% has been reported whereas on-line HDFin post-dilution mode with 100 ml/min substitutionfluid resulted in a reduction rate of 73% [8]. With asubstitution rate of 120 ml/min and a blood flow of350 ml/min 2m plasma reduction rate could not beraised further [8]. In our study, a higher 2m plasma

reduction rate was observed for both DIAPES

HF800 and polysulfone with a substitution fluidflow rate of only 60ml/min and a blood flow ofonly 250 ml/min. In another report on conventionalHDF (22 ml/min substitution flow rate, 402 ml/minblood flow rate) and for online HDF (120 ml/min,434 ml/min) using polysulfone as well as AN69

membranes a removal rate of 2m from plasma of 56and 71% was observed, respectively [2]. On the otherhand, during high-flux HD with a blood flow of250 ml/min using a variety of synthetic membranes(Asahi PAN DX-85, Nikkiso FLX-15GW, RenalSystems Primus 2000, Althin Medical Altra-Flux170G) lower amounts (between 80 and 170 mg) of

2m compared with our study (e.g. 230 14 mg withDIAPES HF 800) were measured in dialysate [9,10].

In uraemia, the correlation between hypoalbu-minaemia and higher morbidity and mortality is awell-documented fact [11]. In peritoneal dialysis with amean albumin loss of $4.26.6 g/day, patients withhigh peritoneal transport displayed lower serumalbumin levels and a lower survival compared withpatients with a lower peritoneal permeability [12].Serum albumin levels, however, are influenced byseveral different factors such as age, nutritionalstatus, inflammation, plasma volume expansion and

decreased protein synthesis, the latter possibly relatedto inflammation. Some studies make the peritoneal lossof albumin responsible for hypoalbuminaemia [12]. Onthe contrary, as shown recently, CAPD patients withnormal albumin, lose more albumin into dialysatecompared with hypoalbuminaemic patients. Reductionof plasma albumin was only observed in those patientshaving a reduced protein catabolic rate even withoutovert inflammation [13].

In our HDF study, albumin loss was lowest withpolyamide (1.0 0.7 g/session), intermediate withpolysulfone (3.5 1.4 g/session) and highest with theDIAPES HF800 membrane (5.7 1.4g/session). Tooperate the HDF process under stable conditions

during the entire session in all patients, we choosea blood flow rate of 250 ml/min and limited thefiltration/substitution volume to 60 ml/min. If themembranes, however, will be used at higher bloodflow rates and substitution volumes albumin loss willpresumably increase.

A previous study from an Italian group with twodifferent PMMA membrane-based filters, even in HDmode, which is prone to a reduced loss of albumincompared with HDF, revealed a higher loss of albumincompared with the three filters we have examined(Filtryzer BK-P, 5.9 g/session; BK-F, 7.4 g/session,Toray, Tokyo, Japan) [4]. In another study with theFiltryzer BK-F, the dialyser with the highest loss of

albumin reported to date, after 6 months of HD theplasma albumin level returned to 3.7 0.2g/dl afteran initial drop. In parallel, renal anaemia improved [5].It was hypothesized that enhanced removal of somecompounds of higher molecular weight (probablyhaving a molecular weight of$40 kDa), exerting aninhibitory effect on erythroid progenitors, could beresponsible [14].

Our study shows that with the three membranesDIAPES HF800, polysulfone and polyamide, 2mremoval into haemodiafiltrate was directly related toalbumin loss (Table 2). Thus, a higher loss of albumin

Fig. 2. Plasma concentrations and plasma reduction rates (RR) of 2m during post-dilutional on-line HDF for three different membranes(mean SEM). The plasma reduction rate for 2m of polyamide is significantly lower (P < 0.05) compared with the other two membranes(*). The post values of each membrane group are significantly different (P < 0.05) from pre values (**).

Comparison of DIAPES HF800 and conventional high-flux membranes during HDF 2385


5/5

seems to be unavoidable to remove more effectivelycertain uraemic toxins, e.g. proteins with a molecularweight higher than 2m [5]. In our study, a loss ofalbumin with DIAPES HF800 higher than withthe two other membranes was not associated with adecrease of post-treatment albumin concentration.Provided a sufficient dietary protein intake, a loss of6 g albumin/treatment in patients free of symptoms

of inflammation can probably be tolerated as itwill be compensated by enhanced albumin synthesis.However, these questions need to be addressed in aseparate, long-term study [15].

Taking into account the two aspects of excellent lowmolecular weight clearances and the high permeabilityfor low molecular weight proteins (indicated by thehigh 2m in haemodiafiltrate), a general conclusionregarding serum albumin levels, treatment conditionsand patients risk cannot be made at the present time.Although, DIAPES HF800 in HDF mode appears tooffer a valuable alternative to conventional membraneslong-term experience with this new membrane isneeded to show whether its higher removal capacityfor low molecular weight proteins is associated withadditional clinical benefit.

Conflict of interest statement. B. Reinhardt, A. Northdurft and H.-D.

Lemke were at the time of the study employed by Membrana GmbH,

the manufacturer of DIAPES HF800.

References

1. Canaud B, Bosc JY, Leray H et al. On-line haemodiafiltration:

state of the art. Nephrol Dial Transplant 1998; 13 [Suppl 5]:311

2. Maduell F, del Pozo C, Garcia H et al . Change from

conventional haemodiafiltration to on-line haemodiafiltration.

Nephrol Dial Transplant 1999; 14: 120212073. Kabanda A, Jadoul M, Pochet JM et al. Determinants of serum

concentrations of low molecular weight proteins in patients onmaintenance hemodialysis. Kidney Int 1994; 45: 16681696

4. Bonomini M, Fiederling B, Bucciarelli T, Manfrini V, Di Ilio C,Albertazzi A. A new polymethacrylate membrane for hemo-

dialysis. Int J Artif Organs 1996; 19: 232239

5. Buoncristiani U, Galli F, Benedetti S, Bianchi J, Floridi A,Canestrari F. Dramatic improvement of uremic anemia over

a 6-month treatment with protein-leaking dialysers. J Am SocNephrol 2000; 11: 259A (Abstract)

6. Colton CK, Henderson LW, Ford CA, Lysaght MJ. Kineticsof hemodiafiltration. I. In vitro transport characteristics of

a hollow-fiber blood ultrafilter. J Lab Clin Med 1975; 85:

355371

7. Bergstrom J, Wehle B. No change in corrected 2-micro-globulin concentration after Cuprophan hemodialysis. Lancet

1987; 1: 628629

8. Lornoy W, Becaus I, Billiouw JM, Van Malderen P,

DHaenens P. On-line haemodiafiltration. Remarkable removal

of 2-microglobulin. Long term clinical observation. Nephrol

Dial Transplant 2000; 15 [Suppl 1]: 4954

9. Sombolos K, Tsitamidou Z, Kyriazis G, Karagianni A,Kantaropoulou M, Progia E. Clinical evaluation of four

different high-flux hemodialysers under conventional conditionsin vivo. Am J Nephrol 1997; 17: 406413

10. Hoenich NA, Stamp S. Clinical performance of a newhigh-flux synthetic membrane. Am J Kidney Dis 2000; 36:

345352

11. Leavey SF, Strawderman RL, Jones CA, Port FK, Held PJ.Simple nutritional indicators as independent predictors ofmortality in hemodialysis patients. Am J Kidney Dis 1998, 31:

9971006

12. Churchill DN, Thorpe KE, Nolph KD, Keshaviah PR,

Oreopoulos DG, Page D for the Canada-USA (CANUSA)

Peritoneal Dialysis Study Group: increased peritonealmembrane transport is associated with decreased patient and

technique survival for continuous peritoneal dialysis patients.J Am Soc Nephrol 1998; 9: 12851292

13. Caravaca F, Arrobas M, Dominguez C. Serum albumin andother serum protein fractions in stable patients on peritoneal

dialysis. Perit. Dial Int 2000; 20: 70370714. Yamada S, Kataoka H, Kobayashi H, Ono T, Minakuchi J,

Kawano Y. Identification of an erythropoietic inhibitor from the

dialysate collected in the haemodialysis with PMMA membrane

(BK-F). Contrib Nephrol 1999; 125: 15917215. Krieter DH, Canaud B. High permeability of dialysis

membranes: what is the limit of albumin loss? Nephrol DialTransplant 2003; 18: 651654

Received for publication: 18.11.02Accepted in revised form: 13.6.03

2386 W. Samtleben et al.

comparison of the new polyethersulfone high-flux membrane

Documents