Journal of the American Society of Nephrology 1703

High-Flux Dialysis Membranes Improve Lipid Profile inChronic Hemodialysis Patient&’2Peter J. Blankestijn,3 Pieter F. Vos, Ton J. Rabelink, Herman J.M. van Rijn, Hans Jansen, and

Hem A. Koomans

P.J. Blankestijn. T.J. Rabelink. HA. Koomans. Depart-ment of Nephrology, University Hospital, Utrecht, TheNetherlands

H.J.M. van Rijn, Department of Clinical Chemistry,University Hospital, Utrecht, The Netherlands

P.F. Vos. Home Dialysis Foundation, Utrecht, TheNetherlands

H. Jansen, Department of Internal Medicine Ill andBiochemistry. Erasmus University. Rotterdam , theNetherlands

(J. Am. Soc. Nephrol. 1995; 5:1703-1708)

ABSTRACTIn a controlled prospective trial, the effect of a switch

from cellulose-based, low-flux dialysis membranes topolysulphone, high-flux membranes on lipid param-

eters was evaluated. Baseline values of lipid param-eters were identical in the study group and the controlgroup in which the dialysis membrane remainedunchanged. After 6 wk, total triglyceride, very low-

density lipoprotein (VLDL) triglyceride, and VLDL cho-

lesterol decreased, respectively, 28 ± 17 (P < 0.01), 38± 17 (P < 0.01), and 24 ± 21% (P < 0.05), and theproportion of total cholesterol that was high-densitylipoprotein cholesterol increased from 15 ± 5 to 18 ±5% (P < 0.05) in the high-flux polysulphone group,whereas these variables remained unchanged in thecontrol group. Low-density lipoprotein and total cho-lesterol as well as Kt/V, protein catabolic rate, para-thyroid hormone, albumin, and body weight did not

change. No change in Iipoprotein Iipase activity wasfound. In a second study, the effects of a singlehemodialysis session with high-flux polysulphone and

low-flux, cellulose-based membranes on lipid param-

eters and lipolytic activity were compared in a cross-over fashion. Treatment with both membranes re-

suIted in a significant decrease in plasma triglyceride,VLDL triglyceride, and VLDL cholesterol. Lipoproteinlipase activity increased during hemodialysis.

1 ReceIved June 28, 1994. Accepted November 30, 1994.

2 Part of this study has been presented at the 26th Annual Meeting of the

American Society of Nephrology, November 14-17, 1993, Boston, and publishedin abstract form In JASN (1993:4:334).

3 correspondence to Dr. P.J. Blankestlin, Room F03.226. Department of Nephrol-ogy, UnIversity Hospital, P.O. Box 85500, 3508 GA Utrecht, The Netherlands.

1044.6673/0509-1 703$03.00/0Journal of the American Society of Nephrologycopy�ght C 1995 by the American Society of Nephrology

Changes in lipid parameters and lipolytic activity

were identical during the two treatments. These dataindicate a favorable change in lipid parameters afterthe switch from low-flux, cellulose-based to high-flux,

polysulphone dialysis membranes that appeared notto be caused by an enhancement of lipoproteinlipase activity. Changes are of similar magnitude as

can be obtained with dietary and pharmacologic

treatments.

Key Words: Hemodialysis, high flux, lipid abnormallties

T he typical lipid profile of patients undergoing

chronic hemodlalysis treatment demonstrates a

moderate increase in triglyceride, normal cholesterol,

decreased high-density lipoprotein (HDL) cholesterol,and increased lipoprotein (a) ( 1 ). At present, it Is not

completely clear to what extent these abnormalities

contribute to the high Incidence of cardiovascular

morbidity and mortality in these patients. However, in

the normal population, this profile is associated

with an increased risk of cardiovascular disease, and

a decrease of triglyceride levels might reduce this

risk (2).

The concentrations of lipoproteins in renal failure

patients may be increased as a consequence of in-

creased synthesis, decreased catabolism, or a combi-

nation of both processes. There is some evidence of an

increased synthesis (3). However, most studies indi-

cate that decreased breakdown predominates ( 1),

which has been ascribed, at least in part, to reduced

lipid catabolism, secondary to reduced activity of lipo-

lytic enzymes, i.e. , lipoprotein lipase (LPL) and hepatic

lipase (HL) (4). LPL Is bound to the luminal surface of

the capillary endothellal cells, where hydrolysis takes

place. It can be released Into the bloodstream by

heparin, which causes an acute decrease in triglycer-

ide (4,5). This heparin-induced lipolysis is markedly

diminished In hemodlalysis patients (4,5). Suggested

underlying mechanisms of the reduced lipolytic activ-

ity include depletion of LPL stores by repeated admin-

istration of heparin (5), the existence of LPL inhibi-

tor(s) in uremic plasma (6), and increased levels of

apolipoprotein CIII (7), whIch is an inhibitor of lipolytic

activity. Also, hyperparathyroidism might play a role,

although the exact mechanism is unclear (8). Finally,

cytokines produced during dialysis might be able to

reduce LPL activity (9). The type of dialysate buffer,

that is, acetate or bicarbonate, does not seem to

influence lipid parameters (10), whereas dialysate glu-

cose concentration up to 1 1 mmol/L does not contrib-

ute to the dyslipidemla ( 1 1). Some data suggest that

Lipid Profile During High-Flux Dialysis

1704 Volume 5 - Number 9 - 1995

switching from regular to low-molecular-weight hepa-

rin might influence dyslipidemia beneficially (12).

Some recent data Indicate that patients undergoing

dialysis with high-flux membranes show lower predi-

alysis triglyceride as compared with more traditional

membranes (13-15). Seres et a!. (15) found that LPL

activity after high-flux hemodialysis was higher than

after dialysis with cellulose-based membranes and

suggested that this could be the cause of the triglyc-

eride-lowering effect. However, the interpretation of

these studies ( 13-15) Is not without difficulties, be-

cause confounding factors such as dialysis character-

istics, diet, medication (especially heparin), and levels

of parathyroid hormone were changed or not men-

tioned, and blood samples were taken either in a

nonfasting or in a fasting state. Therefore, we evalu-

ated in a prospective, controlled trial whether a switch

of low-flux, cuprophane to high-flux, polysulphone

membranes has a beneficial effect on lipid parame-

ters. Dialysis characteristics, heparin dosage, medica-

tion, and diet were not changed and were evaluated. In

addition, we compared the acute effects of a single

hemodialysis treatment with either membrane on LPL

activity and lipids.

PATIENTS AND METHODS

Patients

The total population of our institution comprises approxi-mately 1 20 chronic hemodialysis patients. Of this popula-tion, we included 28 consecutive, stable, long-term (>6months on hemodialysis) patients ( 10 male), who were onlow-flux, cuprophane membranes (Asahi AM 140 and 160Nova: Asahi Medical Co, Tokyo, Japan). All were white and

gave informed consent. Patients were excluded when theywere diabetic or were on medications that raise or lowerlipids. Bicarbonate dialysate containing 1 1 mmol/L glucosewas used throughout. The diet comprised 1 .0 to 1 .2 g/kgbody wt of protein, 30 to 35% of caloric Intake as fat, and 50to 55% as carbohydrates. Primary renal diagnosis was un-known in six patients, chronic pyelonephritis with or without

urolithiasis in seven, glomerulonephritis in five, Alport’sdisease in one, polycystic kidney disease in five, glomerulo-

sclerosis in two, and hypertension in two. The median age(range) was 63 (22 to 80) yr, duration on hemodialysistreatment was 24 (6 to 1 76) months, two patients weredialyzed twice, and the remainder were dialyzed thriceweekly. Mean ± SD duration of therapy per week was 1 1 ±

1 .4 h. Total amount of heparmn administered during dialysiswas 15.247 ± 3.462 U/wk. Regular heparmn was used. The28 patients were randomly divided in two groups. Before thestart of the study, it was ascertained that the two groupswere comparable with respect to lipid parameters by com-paring the mean values of cholesterol and triglyceride mea-surements, which were done routinely every 4 months. Four-teen patients (Group A) were studied before and after switchto high-flux membranes: in the remaining 14 patients (GroupB), the dialysis membrane remained unchanged.

Protocols

The two groups were studied according to the followingprotocol. After an overnight fast, blood was collected from thearteriovenous fistula before dialysis for lipid parameters and

parathyroid hormone. Subsequently, 50 U/kg body wt ofregular heparin was injected iv in the contralateral arm. After20 mlii, blood was collected from the arteriovenous fistula ina chilled tube for LPL. In the week before the blood collection,calculated Kt/V and protein catabolic rate were recorded.Residual renal function was taken into account. From thenext dialysis session onward, the patients in Group A weredialyzed with a F6OS polysulphone, high-flux dialyzer (Fre-senlus AG, Bad Homburg, Germany). In both groups, bloodand dialysate flow, length and number of weekly sessions,and total amount of heparmn used during dialysis, as well asmedication and diet, were not changed during the studyperiod. Six weeks after the first measurements of lipid pa-rameters, blood was collected by the use of the same proto-col. Dialysis membranes were used only once. Diet adher-ence was evaluated by the use of a diet diary at baseline andafter 6 wk.

In a second study, six chronic hemodialysis patients (3

men) were randomly selected from the previous study group.In random order, 1 wk apart, they were treated with thehigh-flux, polysulphone or the low-flux, cuprophane mem-brane after an overnight fast. They were not allowed to eat ordrink during the dialysis session. Blood was sampled before,during, and at the end of the dialysis. After blood samplingfor basal values, a bolus injection of 2,500 U of heparin wasgiven at the start of dialysis. Then, a continuous infusion ofheparin was instituted: the dose was adjusted according tothe results of periodic partial thromboplastin time determi-nations but was the same during both studies. All patientswere treated three times for 3.5 h a week.

Methods

Lipid parameters were measured as described previously(16). In brief, for the separation of lipoproteins, plasmasamples were subjected to a single ultracentrifugation stepat 4#{176}Cfor 20 h in a 50.3 Ti rotor (Beckman Instruments, PaloAlto, CA). Cholesterol and triglyceride were measured with afully automated Hitachi 7 1 7 analyzer from Boehringer

Mannheim GmbH (Mannheim, Germany) with the enzymaticapplications with reagents obtained from the same company.The phosphotungstlc acid-magnesium chloride precipitationmethod was used to measure HDL cholesterol together withthe cholesterol-cholesteroloxidase-phenol 4-aminophenazonemethod. Lipoprotein (a) was measured as described elsewhere(17).

LPL and HL were determined separately by an immuno-chemical method, essentially as described by Huttunen et a!.(18), with a gum acacia-stabilized l3Hltrioleoylglycerol emul-

sion. HL activity was determined as the salt-resistant lipasein the presence of 1 M NaC1. LPL activity was determinedafter the Inhibition of HL with a goat antibody raised againstHL purified from postheparin human plasma. Fasted ratserum was added as a source of apolipoprotein CII. A traceramount of I ‘4Cloleate was added to the substrate as aninternal standard for the extraction of l3Hjoleate liberatedfrom [3Hltrioleoylglycerol substrate. In each series of deter-minations, pooled plasmas with high and low LPL and HLwere included as a reference. The normal values of LPL are>70 U/L in men and >80 U/L in women and ofHL are >300U/L in men and >225 U/L in women.

Kt/V was calculated with the urea values determinedbefore and after hemodialysis (19). Data are presented asmean ± SD. Results of both studies were analyzed by atwo-way analysis of variance. If the variance ratio obtainedby the analysis of variance reached statistical significance,

75

50

25 Ia)C(a

.C

0a)Ca)0

a)

I-25

-50

-75

I**

LT��rr

Blankestijn et al,

Journal of the American Society of Nephrology 1705

the differences between the means of the observations were

analyzed at the 5 and 1% significance levels by the least

significance difference test.

RESULTS

The patients who were switched to the high-flux

membrane (Group A) and the control patients (GroupB) did not differ with regard to age, duration on

chronic hemodialysis therapy, frequency (in both

groups, one patient twice weekly and the remainder

thrice weekly), length of weekly sessions ( 10.0 ± 1.5

versus 10.8 ± 1 .4 h/wk), and total amount of heparin

administered during the hemodialysis session (14.835

± 2,230 versus 16.398 ± 3,900 U/week). The switch

did not cause any clinically detectable changes or side

effects, although the study was not specifically de-

signed to address this question. Lipid parameters at

baseline did not differ (Table 1 ). After 6 wk of treat-

ment with a high-flux membrane, fasting total triglyc-

eride, very low-density lipoprotein (VLDL) triglyceride,

and VLDL cholesterol were significantly decreased In

Group A, respectively, by 28 ± 17 (P < 0.01), 38 ± 17

(P < 0.01), and 24 ± 21% (P < 0.05)(Table 1; Figure 1).

Other lipid parameters did not change; however, the

contribution of HDL cholesterol to total cholesterol

increased from 15 ± 5 to 18 ± 5% (P < 0.05) in the

high-flux group. LPL and HL were low and showed no

change in Group A and were therefore not measured

in Group B (Table 1 ). Baseline Kt/V, protein catabolic

rate, parathyroid hormone, and body weight, as well

as albumin, were not significantly different and re-

maimed unchanged in both groups (Table 2). Residual

renal function did not change.

In the second study, we found that, during dialysis

sessions with the high-flux, polysulphone membrane

and with the low-flux, cuprophane membrane, there

was a clear effect on lipid parameters. Baseline values

were identical (Table 3). In all six patients, triglyceride

decreased within 1 h ofhemodialysis from 2.27 ± 0.47

TG VLDL-TG VLDL-chol

Figure 1 . Percent change in lipids. Open bars, high-fluxpolysulphone; hatched bars, low-flux, cellulose based. TG,triglyceride; chol, cholesterol. Comparison with value atweek 0: � = P < 0.05, ** = � < oo��

to 1 .33 ± 0.36 during high-flux polysulphone and

from 2.38 ± 0.88 to 1 .6 1 ± 0.7 1 mmol/L during

low-flux cellulose based (both P < 0.05) and remained

low during the remainder of the dialysis session (Table

3). VLDL triglyceride and VLDL cholesterol decreased

significantly during both treatments (Table 3). There

were no differences in the degrees of decrease in lipids

between the two membranes. Lipolytic activity was

undetectable before the start of hemodialysis, in-

creased within 1 h, and decreased toward the end

(Figure 2). Values at 1 , 2, and 3.5 h of hemodialysis

were, respectively, during high-flux polysulphone 30

± 10, 13 ± 6, and 9 ± 8 U/L and during low-flux

cuprophane 29 ± 12, 7 ± 2, and 6 ± 3 U/L. There was

no difference between the two study groups.

DISCUSSION

This is the first prospective, controlled trial examin-

ing the effect on lipid profile of a switch from low-flux,

TABLE 1 . Lipid profile before and after switch to dialysis with high-flux polysulphone membranea

ParameterHigh Flux Control

WeekO Week#{243} WeekO Week 6

Total TG (mmol/L) 2.58 ± 1.17 1.82 ± 0.59c 2.43 ± 0.96 2.48 ± 0.80VLDL TG (mmol/L) 1.88 ± 1.30 1.08 ± 0.44c 1.71 ± 0.96 1.71 ± 0.85HDL + LDL TG (mmol/L) 0.78 ± 0.22 0.74 ± 0.19 0.72 ± 0.22 0.77 ± 0.18Total chol (mmol/L) 5.18 ± 1 .1 1 4.71 ± 0.95 5.07 ± 0.89 5.20 ± 0.57VLDL chol (mmol/L) 1.45 ± 0.61 1 .04 ± 037b 39 ± 0.76 1.33 ± 0.53LDL chol (mmol/L) 2.98 ± 0.72 2.87 ± 0.73 2.84 ± 0.82 2.99 ± 0.59HDL chol (mmol/L) 0.75 ± 0.16 0.81 ± 0.16 0.85 ± 0.20 0.88 ± 0.19HDL/total chol (%) 15.3 ± 4.5 18.0 ± 50b 16.8 ± 3.7 16.8 ± 4.1Lp(a) (mgIL) 350 ± 243 340 ± 130 328 ± 242 412 ± 178

LPL (U/L) 83 ± 35 89 ± 25HL (U/L) 243 ± 66 250 ± 58

a Data are means ± SD. In the control group. the dialysis membrane remained unchanged. 1G. triglyceride; chol, cholesterol; Lp(a), lipoprotein A;

LDL, low-density lipoprotein.b p < 0.01 , comparison with value at Week 0.C p < 0.05, comparison with value at Week 0.

2

time (h)

Lipid Profile During High-Flux Dialysis

1706 Volume 5 ‘ Number 9 ‘ 1995

TABLE 2. Clinical and biochemical parameters before and after switch to dialysis with high-flux polysulphonemembranea

ParameterHigh Flux Control

Week 0 Week 6 Week 0 Week 6

BodyWt(kg) 65±13 65±13 64±13 64±10

PTH (pmol/L) 38 ± 24 37 ± 11 25 ± 15 24 ± 11

Albumin (gil) 33 ± 3.5 33 ± 3.3 33 ± 6.1 35 ± 4.4Kt/V (week) 2.73 ± 0.56 3.09 ± 0.62 2.95 ± 0.54 3.00 ± 0.52PCR (g/kg per day) 1.02 ± 0.12 0.98 ± 0.15 1.06 ± 0.28 1.04 ± 0.18Predlalysis urea (mmol/L) 28 ± 6.7 28 ± 7.3 28 ± 6.6 29 ± 2.5

0 Data are means ± SD. In the control group, the dialysis membrane remained unchanged. PTH, parathyroid hormone; P�R. protein catabolic rate.

TABLE 3. Effects of a single hemodialysis treatment on lipid parameters

ParameterHigh flux Control Difference in

Ch 95% CIange ( )Start End Change Start End Change

Total TG (mmol/L) 2.27 ± 0.47 1.75 ± 069b -0.52 ± 0.51 2.38 ± 0.88 1.95 ± 086b -0.43 ± 0.24 0.08 (-0.32-0.47)VLDL TG (mmol/L) 1.47 ± 0.39 0.93 ± 051b -0.53 ± 0.45 1.77 ± 0.85 1.13 ± 0.75C -0.63 ± 0.37 -0.10 (-0.42-0.23)HDL + LDL TG (mmol/L) 0.80 ± 0.20 0.82 ± 0.24 0.02 ± 0.09 0.62 ± 0.1 1 0.83 ± 017b 0.22 ± 0.16 0.20 (-0.07-0.47)Total chol (mmol/L) 5.22 ± 0.39 5.48 ± 0.52 0.27 ± 0.33 5.30 ± 1.01 5.58 ± 091b 0.28 ± 0.21 0.02 (-0.55-0.59)VLDL chol (mmol/L) 1.47 ± 0.35 0.88 ± 0.29’� -0.58 ± 0.23 1.87 ± 0.98 1.07 ± 0.87c -0.80 ± 0.28 -0.22 (-0.62-0.18)LDI chol (mmol/L) 2.90 ± 0.35 3.70 ± 0.48c 0.80 ± 0.28 2.65 ± 0.51 3.58 ± 0.38C 0.93 ± 0.35 0.13 (-0.26-0.52)HDL chol (mmol/L) 0.85 ± 0.16 0.90 ± 0.16 0.05 ± 0.09 0.78 ± 0.16 0.93 ± 021b o�5 � 0.13 0.10 (-0.26-0.52)

a Data are means ± SD. control, low-flux, cellulose based; 1G. triglycerIde; LDL. low-density lipoprotein; cI. confidence Interval.

b p < 0.05, comparison with value at start of hemodialysis.C p < 0.01 . comparison with value at start of hemodialysis.

.J

E.�,-

E

-J

0.-J

Figure 2. Effects of a single hemodialysis on LPL Data aremeans ± SD at baseline and after 1, 2, and 3.5 h (end) ofdialysis. Triangles, high-flux polysulphone; squares, low-flux,cellulose based.

cellulose-based membranes to high-flux, polysul-

phone membranes in patients on maintenance hemo-

dialysis. The baseline lipid abnormalities in our pa-

tients are similar to those described in the literature

( 1). Hemodialysis with the high-flux, polysulphone

membrane was associated with a decrease in total

triglyceride, VLDL triglyceride, and VLDL cholesterol

and an Increase In the proportion of total cholesterol

that was HDL cholesterol. An effect of confounding

factors such as dialysis characteristics, medication,

diet, change in parathyroid hormone levels, or heparin

dosage could be excluded, because these factors werenot different in the treatment group and the control

group and did not change throughout the study.

These results establish that hemodialysis with a high-

flux membrane has a favorable effect on lipid param-

eters.

Dumler et a!. ( 13) found a decrease in triglyceride

and cholesterol when their patients were switched

from conventional hemodialysis therapy with cellu-

lose-based membranes to hemodialysis with high-

flux, polysulphone membranes. However, because

this study was designed to evaluate the safety and

efficacy of short-time dialysis, several other dialysis

characteristics were changed as well, making it diffi-

cult to define the contribution of the dialysis mem-

brane. Josephson et a!. ( 14) found, in a cross-sec-

tional study, lower triglyceride and identical total

cholesterol and HDL cholesterol in 18 patients who

were treated with high-flux membranes (polysulphone

or cellulose triacetate) as compared with 1 6 patients

undergoing dialysis using low-flux, cellulose-based

membranes. However, in many oftheir patients, blood

was sampled while the patients were not in fasting

state. Finally, Seres et a!. ( 15) demonstrated lower

triglyceride and higher LPL in patients on high-flux,

polysulphone dialysers as compared with patients

Blankestijn et al,

Journal of the American Society of Nephrology 1707

dialyzed with cellulose-based membranes. However,

their study was mainly cross-sectional. Moreover, be-

cause blood samples were obtained immediately after

dialysis, the steady-state effects on lipids and LPL

were not evaluated in that study.

The change in lipid variables In this study could very

well fit with an enhancement of lipolytic activity (5).

However, we found that the decrease in predialysis

triglyceride did not coincide with an increase in pre-

dialysis lipolytic activity. Therefore, we studied in a

subset of patients the effects of a single dialysis on

lipid parameters and lipolytic activity. Both dialysis

treatments caused similar decrements in triglyceride,

VLDL triglyceride, and VLDL cholesterol and an incre-

ment in low-density lipoprotein, presumably the re-

sult of the increased VLDL catabolism during dialysis.

This triglyceride-lowering effect of hemodialysis was

also found by others (20,2 1) and is probably caused by

the increase in plasma LPL activity. LPL is bound to

the endothelium and released by heparin (5). Teraoka

et a!. (20) found that lipolytic activity increased and

triglyceride levels decreased during hemodialysis with

heparmn, but not when heparin was omitted. We also

found a transient increase in LPL during hemodialy-

sis, but importantly, this effect was similar with the

two membrane types. Therefore, our data make It

unlikely that the long-term triglyceride-lowering effect

of high-flux membranes Is due to LPL stimulation. A

possible explanation is offered in a recent report by

Otsubo et a!. (22). They showed that treatment with

high-flux cellulose triacetate membranes decreased

triglyceride and increased HDL cholesterol and that

these changes were associated with a favorable

change in the apolipoprotein CII/CIII ratio. In that

study, apolipoprotein CII remained unchanged,

whereas apolipoprotein CIII decreased during treat-

ment with the high-flux membrane. This change

might be of significance. LPL requires interaction with

apolipoprotein CII to be active (5). Indeed, patients

with an apolipoprotein CII deficiency have a critically

impaired LPL-mediated lipolysis, although their pos-

theparin LPL activity is normal when measured in vitro

in the presence of exogenous activators (23). Apoli-

poprotein CIII is a competitive Inhibitor of apolipopro-

tein CII (5) and Is increased In uremic patients (7).

Therefore, a decrease in the apolipoprotein Cli/Cill

ratio renders a decrease in the susceptibility of the

substrate lipoprotein to LPL action, without affecting

in vitro LPL measurements.

The two types of membranes used In this study,

namely, low-flux, cellulose-based and high-flux, poly-

sulphone, differ from each other at least in two as-

pects, that is, in biocompatibility and in permeability.

One could Imagine that both differences might be of

importance is establishing the change In lipid proffie.

Cuprophane instead of synthetic membranes induce

cytokine release, which can cause a reduction in LPL

activity (9). This study could not confirm different

effects on LPL activity. Alternatively, high-flux mem-

branes might be able to remove more effectively lipase-

inhibiting factors (22). This possibility is supported by

the observation that high-flux membranes cause a

decrease in apolipoprotein CIII (22). It seems most

likely that the increased permeability of the present

high-flux membranes and not the fact that they weremade of polysulphone determines the lipid-lowering

effect, because Otsubo et a!. (22) reported similar

lipid-lowering results using another type of high-flux

membrane.

The exact contribution of lipid abnormalities in

determining cardiovascular morbidity and mortality

in end-stage renal failure patients is unknown at

present. Both hypertriglyceridemia and low HDL cho-

lesterol are recognized as risk factors for coronary

heart disease (24,25) and lowering low-density 11-

poprotein cholesterol together with lowering triglycer-

ide and Increasing HDL cholesterol reduces the coro-

nary risk in the nonuremic population (26). No such

data exist for end-stage renal failure patients, and It

remains to be established whether it is beneficial to

control the uremic lipid abnormalities, by whatever

means. Recently, It has been advocated that hyper-

triglyceridemia (triglyceride >2.26 mmol/L [200 mgI

dLfl, when in the presence ofcoexisting coronary heart

disease, diabetes mellitus, chronic renal failure, or

low levels of HDL cholesterol, should be treated (2).

One or more of these conditions usually exist in

hemodialysis patients. Studies in uremic patients

have documented the efficacy of dietary modulations

(27-29) or drugs such as clofibrate, gemfibrozil, and

bezafibrate (30-32). Of note, the presenily described

switch in dialysis membrane is equally effective as

dietary modifications (27-29) and almost as effective

as fibrates (30-32). The latter, however, can have

serious side effects in renal failure patients (33,34),

which limits their clinical use. This study indicates

that the choice of dialysis membrane offers an oppor-

tunity to reduce the dyslipidemla without adverse

effects and without problems with the compliance of

the patients.

ACKNOWLEDGMENTST.J. Rabelink is sponsored by a fellowship of the Royal Dutch Acad-

emy of Sciences (KNAW).

REFERENCES

1 . Attman P0, Samuelsson 0, Alaupovic P: Lipoproteinmetabolism in renal failure. Am J Kidney Dis 1993:21:573-592.

2. International Committee for the Evaluation of Hyper-triglyceridemia as a Vascular Risk Factor: The hyper-triglyceridemias: risk and management. Am J Cardiol199 1 :68LSuppl 3]: 1A-42A.

3. Chan MK, Varghese Z, Persaud JW, Baillod RA, Moor-head JF: Hyperlipidemia in patients on maintenancehemo- and peritoneal dialysis: the relative pathogeneticroles of triglyceride production and triglyceride removal.Clin Nephrol 1982:17:183-190.

4. Chan MK, Persaud J, Varghese Z, Moorhead J: Patho-genetic roles of post-heparmn lipases in lipid abnormall-ties in hemodialysis. Kidney mt 1984:25:812-818.

5. Flier JS, Underhifi LH: Lipoprotein lipase. A multifunc-tional enzyme relevant to common metabolic diseases. NEngl J Med 1989:320:1060-1068.

Lipid Profile During High-Flux Dialysis

1708 Volume 5 ‘ Number 9 ‘ 1995

6. Murase T, Cattran DC, Rubenstein B, Steiner G: Inhibi-tion of lipoprotein lipase by uremic plasma, a possiblecause of hypertriglyceridemia. Metabolism 1975:24:1279-1286.

7. Statrans I, Felts JM, Zacherle B: Apoprotein compositionof plasma lipoproteins in uremic patients in hemodialy-sis. Clin Chim Acta 1979:93:135-43.

8. Akmal M, Kasim SE, Soliman AR, Massry SG: Excessparathyroid hormone adversely affects lipid metabolismin chronic renal failure. Kidney Int 1990:37:854-858.

9. Dinarello CA: Cytokines: Agents provocateurs in hemo-dialysis. Kidney mt 1992:41:683-694.

10. Diamond SM, Henrich WL: Acetate dialysate versusbicarbonate dialysate: A continuing controversy. Am JKidney Dis 1987:9:3-11.

1 1 . Rosborough DC, Van Stone JC: Dialysate glucose. Se-mm Dial 1993:6:260-263.

12. Deuber HJ, Schulz W: Reduced lipid concentrationsduring four years of dialysis with low molecular weightheparin. Kidney Int 1991:40:496-500.

13. Dumler F, Stalla K, Mohini R, Zasuwa G, Levin NW:Clinical experience with short time hemodialysis. Am JKidney Dis 1992:19:49-56.

14. Josephson MA, Fellner 5K, Dasgupta A: Improved lipidprofiles in patients undergoing high-flux hemodialysis.Am J Kidney Dis 1992:20:361-366.

15. Seres DS, Strain GW, Hashim SA, Goldberg Li, LevinNW: Improvement of plasma lipoprotein profiles duringhigh-flux dialysis. J Am Soc Nephrol 1993:3:1409-1415.

16. Gimpel JA, Van Rijn HJM, Kortlandt W: Lipoproteins,lipids and apolipoproteins as risk factors for angio-graphically assessed coronary artery disease. Clin ChemEnzym Comms 1992:5:63-68.

17. Leerink CB, Gimpel JA, Kortlandt W, Bouma BN, vanRijn HJM: Kinetic analysis oflipoprotein (a) inhibition ofplasminogen activation by tissue plasminogen activatorin vitro. Fibrinolysis 1991:5:233-238.

18. Huttunen JK, Ehnholm C, Kekki M, Nikkila EA: Post-heparmn plasma lipoprotein lipase and hepatic lipase innormal subjects and in patients with hypertriglyceridae-mia: Correlations to sex, age and various parameters oftriglyceride metabolism. Clin Sci Mol Med 1976:50:249-260.

19. Dauglrdas JT: Chronic hemodialysis prescription: Aurea kinetic approach. In: Daugirdas JT, Ing TS, Eds.Handbook ofDialysis. 2nd Ed. Boston: Little, Brown andCompany. 1994:92-120.

20. Teraoka J, Matsui N, Nakagawa 5, Takeuchi Y: The roleofheparmn in the changes oflipid patterns during a singledialysis. Clin Nephrol 1982:17:96-99.

2 1 . Wessel-Aas T, Christophersen B: Systemic heparmniza-tion of uremic patients and its effects on blood lipids and

in vivo toxicity ofthe plasma. Clin Nephrol 1982; 18:135-140.

22. Otsubo Y, Uchida Y, Yasumoto Y, Yamashita W, ArimaT: Apo CIII is a potent factor on lipid abnormalities inhemodialysis patients [Abstract]. J Arm Soc Nephrol1993:4:375.

23. Breckenbridge WC, Little JA, Steiner G, Chow A,Poapst M: Hypertriglyceridemia associated with defi-ciency of apolipoprotein CII. N Engl J Med 1978:298:1265-1273.

24. Fontbonne A, Echwege E, Cambien F, et at.: Hypertrig-lyceridemia as a risk factor for coronary heart diseasemortality in subjects with impaired glucose tolerance ordiabetes. Results from the 1 1 -year follow-up of the ParisProspective Study. Diabetologia 1989:32:300-304.

25. Gordon T, Castelli WP, Hjortland MC, Kannel WB, Daw-ber TR: High density lipoprotein as a protective factoragainst coronary heart disease: The Framingham Study.Am J Med 1977:62:707-714.

26. Flick MH, Elo 0, Haapa K, et at.: Helsinki Heart Study.Primary prevention trial with gemfibrozil in middle-agedmen with dyslipidemia. N Engl J Med 1987:317:1237-1245.

27. Sanfelippo ML, Swenson RS, Reaven GM: Response ofplasma triglycerides to dietary change in patients onhemodialysis. Kidney Int 1978:14:180-186.

28. Gokal R, Mann JI, Ledingham DM: Dietary treatment ofhyperlipidemia in chronic hemodialysis patients. Am JClin Nutr 1978:31:1915-1918.

29. Cattran DC, Steiner G, Fenton SSA, Ampil M: Dialysishyperlipidemia: response to dietary manipulations. ClinNephrol 1980:13:177-182.

30. Goldberg AP, Applebaum-Bowden DM, Bierman EL, etat.: Increase in lipoprotein lipase during clofibrate treat-ment of hypertriglyceridemla in patients on hemodialy-sis. N Engl J Med 1979:301:1073-1076.

3 1 . Pasternack A, Vfinttinen T, Solakivi T, Kuusi T, Korte T:Normalisation oflipoprotein lipase and hepatic lipase bygemfibrozil results in correction of lipoprotein abnormal-ities in chronic renal failure. Chin Nephrol 1987:27:163-168.

32. Grutzmacher P, Scheuermann EH, Siede W, et at.: Lipidlowering treatment with bezafibrate in patients onchronic haemodialysis: pharmacokinetics and effects.Klin Wochenschr 1986:64:910-916.

33. Pierides AM, Alvarez-Ude F, Kerr DNS: Clofibrate-induced muscle damage in patients with chronic renalfailure. Lancet 1975:2:1279-1272.

34. Kijima Y, Sasaoka T, Kanayama M, Kubota 5: Untowardeffects of clofibrate in hemodialyzed patients. N Engl JMed 1977:296:515.