Improved Lipid Profiles in Patients Undergoing High-Flux Hemodialysis

Improved Lipid Profiles in Patients Undergoing High-Flux Hemodialysis M.A. Josephson, MD, S.K. Fellner, MD, and A. Dasgupta, PhD • Hyperlipidemia is one of many atherogenic risk factors encountered by patients undergoing chronic hemodialysis (HD). We have studied lipid profiles in these patients and have found less hypertriglyceridemia in those undergoing high-flux HD than those receiving traditional HD. Mean ± SEM triglyceride level was 1.62 ± 0.15 mmol/L (143.3 ± 13.6 mg/dL) in high-flux dialysis patients, 2.39 ± 0.27 mmol/L (211.6 ± 24.1 mg/dL) in conventional dialysis patients, and 1.55 ± 0.13 mmol/L (137.1 ± 11.5 mg/dL) in normal age- and sex-matched controls. In addition, we found that in patients undergoing high-flux HD, females had higher high-density Iipoprotein2 (HDL2) levels (0.62 ± 0.03 mmol/L [23.8 ± 1.3 mg/dL]) than males (0.33 ± 0.04 mmol/L [12.9 ± 1.7 mg/dL]) (P < 0.01). The mechanism(s) responsible for divergent lipid profiles in subsets of HD patients deserves further investigation. Whether reductions of hypertriglyceridemia and/or increases of HDL2 will diminish the incidence of cardiovascular disease in dialysis patients is unknown. © 1992 by the National Kidney Foundation, Inc. INDEX WORDS: Hemodialysis; hyperlipidemia; hypertriglyceridemia; atherosclerosis, high-density lipoprotein2 cholesterol.

O

VER THE PAST two decades, the relationship between dialysis and lipid abnormalities has aroused much interest. A 1974 retrospective analysis by Lindner et al of 39 chronic hemodialysis (HD) patients suggested that atherosclerotic complications in these patients was markedly greater than in age-matched normal and hypertensive individuals. 1 The proposal that long-term HD is associated with an accelerated rate of atherosclerosis was not uniformly accepted or confirmed. 2 Nicholls et al presented evidence suggesting that atherosclerosis exists prior to the initiation of HD and becomes clinically apparent during dialysis therapy. 3 Whether a relationship exists between the HD procedure and atherosclerosis remains unclear. Patients undergoing dialysis experience excessive cardiovascular morbidity and mortality compared to their nonuremic counterparts. In 1988, 44.5% of dialysis patients died of cardiac causes, and 6.3% of cerebral vascular accidents4 compared with 35% and 6.9%, respectively, in the general US population. 5 Numerous atherogenic risks including hypertension, diabetes mellitus, abnormalities of calcium and phosphorous metabolism, platelet dysfunction, and disordered lipid metabolism are present in many dialysis patients. 6,7 The hypertriglyceridemia, common in chronic HD patients, is reflected in the high triglyceride content of high density lipoproteins (HDLs) and low-density lipoproteins (LDLs).6,8.11 This hypertriglyceridemia has been ascribed to diminished lipoprotein lipase and hepatic lipase, as well as to increased triglyceride synthesis. Although total serum cholesterol may

be normal, HDL cholesterol is significantly reduced in these patients. 6,12,13 HDL2 (the cardioprotective portion of HDL) is reduced in patients undergoing conventional HD.14 Their typical lipid profile demonstrates hypertriglyceridemia, a reduced HDL cholesterol fraction, and an elevated very-low-density lipoprotein (VLDL) cholesterol fraction in the setting of normal total cholesterol, resembling Frederickson's type IV hyperlipoproteinemia. 6,15,16 The observation that many of our patients undergoing highflux HD had near-normal triglyceride levels prompted investigation oflipids in these patients, as well as those using traditional cellulose acetate membranes and healthy controls. MATERIALS AND METHODS

Patient Selection We randomly selected 18 patients undergoing high-flux HD and 16 patients undergoing conventional HD who had been on their respective modality for at least 8 months (Table I). At our institution, high-flux HD is performed using polysulfone (F-80 Fresenius, Oberursel, Germany) or cellulose triacetate (CT 190G, Baxter Healthcare, Deerfield, IL) membranes with blood flow rates of 400 to 450 mL/min; conventional HD uses cellulose acetate (CAl 10, Baxter Healthcare) membranes and blood flow rates of 250 to 300 mL/min. Bicarbonate dialysate is used throughout. There was

From the Departments of Medicine and Pathology, University of Chicago, Pritzker School of Medicine, Chicago, IL. Received December 23, 1991; accepted in revised form June 23, 1992. Address reprint requests to Michelle Josephson, MD, Section of Nephrology, MC 5100, University of Chicago Hospital, Chicago, IL 60637. © 1992 by the National Kidney Foundation, Inc. 0272·6386/92/2004·0007$3.00/0

American Journal of Kidney Diseases, Vol XX, No 4 (October), 1992: pp 361-366

361

362

JOSEPHSON, FELLNER, AND DASGUPTA Table 1. Patient Characteristics

N Male Race Age (yr) Years of HD Membrane Dialysate Blood flows (mL/min) HTN PCKD FSGS SLE PSGN HERED OM MEMB CONGEN IVDA

Controls

Conventional HD

High-Flux HD

18

16 8 15 B, 1 W 46 ± 6 5 ± 4.5 15 CA, 1 C Bicarbonate

18

9 16 B, 2 W 47 ± 10 NA NA NA NA NA NA NA NA NA NA NA NA NA NA

250-300 5 0 1 3 0 1 4 0

9 18 B 47 ± 14 6± 4 16CT,2P Bicarbonate 400-450 7 2 1 3 1 2 1 0 0

Abbreviations: HTN, hypertension; PCKD, polycystic kidney disease; FSGS, focal segmental glomerular sclerosis; SLE, systemic lupus erythematosus; PSGN, post streptococcal glomerulonephritis; HERED, hereditary nephritis; OM, diabetes mellitus; MEMB, membranous; CONGEN, congenital nephritis; IVDA, intravenous drug abuse; CT, cellulose triacetate; P, polysulphone; CA, cellulose acetate; C, cuprophane; B, black subjects; W, white subjects; NA, not applicable.

no difference in years of HD between high-flux and conventional HD patients (6 ± 4 and 5 ± 4.5 years, respectively). None of the patients was receiving either thyroid replacement or lipid-lowering agents. One patient undergoing high-flux and two patients undergoing standard HD were receiving J3-blockers. One patient undergoing high-flux HD took prednisone (15 mg/d). The standard prescribed daily diet was I to 1.2 g/ kg protein, 2 g sodium, 2 g potassium, and I g phosphorous. Two patients undergoing high flux and one standard HD were instructed to lower their cholesterol intake as well. Another two patients using high flux and one standard HD were given commercial nutritional supplements.

Data Collection

cholesterol-4-ene-3-one. During these reactions hydrogen peroxide was generated and reacted with Leuc Dye (Eastman Kodak, Rochester, NY) in a colorometric reaction directly proportional to the cholesterol concentration of the serum sample.

HDL Cholesterol HDL levels were determined after precipitation of VLDL and LDL in serum samples with I % dextran sulfate (molecular weight, 50,000) in 0.5 mmol of MgCh. The HDL cholesterol levels were measured using an Ektachem-700 autoanalyzer and dry reagent slides (Eastman Kodak, Rochester, NY).

HDL2 Cholesterol HDL2 was precipitated from the HDL fraction with 1% dextran sulfate in 1.5 mmol MgCh. The concentration of HDL3 was measured in the supernatant. The concentration of HDL2 was calculated as the difference between HDL and HDL3.

LDL Cholesterol LDL cholesterol levels were calculated by Friedewald's formula: LDL cholesterol = Total Cholesterol - (HDL Cholesterol + Triglycerides/5).

Triglycerides Lipase was used to generate glycerol from triglycerides. Glycerol was then converted to dihydroxyacetone and hydrogen peroxide by a coupled enzyme system (adenosine triphosphate [A TPj/glycerol kinase followed by glycerophosphate oxidase). Hydrogen peroxide reacted with Leuc Dye in a colorometric reaction directly proportional to the triglyceride concentration of the serum sample.

Urea Kinetic Modeling The fractional dialyzer urea clearance (KT IV) is performed monthly on all patients undergoing HD at the University of Chicago dialysis center. Our goal is to maintain the patient's KT /V > 1.1 to 1.2. The values presented in this study represent the measured KT/V. For this report, a minimum of four monthly values (based on calculated volumes) were averaged to obtain a representative KT/V for each patient.

Protein Catabolic Rate Protein catabolic rate (PCR), or total grams of protein catabolized daily, was calculated using Keshaviah's mapY

We recorded age, gender, etiology of end-stage renal disease, presence of other medical pathology, length oftime on dialysis, medications, calculated and estimated KT/V, dialysate, HD membrane, blood flow rates, interdialytic weight gains, the heparin doses administered during dialysis, and calculated protein catabolic rate. Immediately before a dialysis treatment, serum samples were obtained for total cholesterol, HDL cholesterol, HDL3 cholesterol, and triglycerides. An additional group of samples was obtained from fasting patients.

Statistical Analysis

Total Cholesterol

Triglyceride levels in high-flux dialysis patients (1.62 ± 0.15 mmol/L [143.3 ± 13.6 mg/dL]) were not different from normal controls (1.55 ± 0.13

Cholesterol ester was hydrolyzed by cholesterol ester hydrolase, then cholesterol oxidase oxidized the cholesterol to

Student's unpaired two-tailed t test was used to compare results between groups. Linear regression analysis was performed to look for an association between two variables. Data are presented as mean ± SEM. Statistical significance is reported at the 0.05 and 0.0 I levels.

RESULTS

363

LIPID PROFILES IN HIGH-FLUX HEMODIALYSIS

o •

3.0 ~ ~

~

ra

CO~i1OlS CO~V£NTIOSALHD

HIGH,f1.VX HD

S 2.0

~ ~

~E

a

~

Fasting

Rlndom

Fig 1. Triglyceride levels in controls, conventional HD patients, and high-flux HD patients. Triglyceride measurements in random and fasting high-flux HD patients did not significantly differ from controls. Triglycerides were elevated in both random and fasting conventional HD patients compared with random and fasting controls (P < 0.01 and 0.05, respectively) and high-flux HD patients (P < 0.05 and 0.01, respectively).

mmol/L [137.1 ± 11.5 mg/dL]), but were elevated in patients undergoing conventional HD (2 .39 ± 0.27 mmol/L [211.6 ± 24.1 mg/dL)) (P < 0.01) (Fig 1, Table 2). A similar pattern was found in the fasting state. The mean triglyceride level in three fasting conventional HD patients (2.61 ± 0.57 mmol/L [231.0 ± 50.1 mg/dLD was significantly higher than the level in II fasting high-flux HD (1.26 ± 0.10 [111.5 ± 8.6 mg/dL)) and eight control patients (1.31 ± 0.20 mmol/L [116.1 ± 18.1 mg/dL)), P < 0.01 and P < 0.05, respectively. Total cholesterol was 4.26 ± 0.20 mmol/L (164.9 ± 7.9 mg/dL) in those undergoing standard HD, 4.50 ± 0.23 mmol/L (173.9 ± 8.8 mg/ dL) in high-flux HD, and 4.53 ± 0.24 mmol/L (175.3 ± 9.1 mg/dL) in healthy controls (Table 2). Consistent with prior reports of normal cholesterol levels in HD patients, there was no sig-

nificant difference in cholesterol measurements among the three groups. HDL was higher in individuals receiving high-flux HD (1.23 ± 0.06 mmol/L [47.7 ± 2.5 mg/dLD than conventional HD (1.12 ± 0.06 mmol/L [43.4 ± 2.5 mgjdLD. Both high-flux and standard HD patients had significantly lower HDL content than healthy controls (1.42 ± 0.06 mmol/L [55.0 ± 2.4 mg/ dLD, P < 0.05 and P < 0.01, respectively. The mean HDL2 level observed in standard HD patients (0.39 ± 0.04 mmol/L [15.2 ± 1.5 mg/dLD was significantly lower (P < 0.05) than in controls (0.52 ± 0.04 mmol/L [20.3 ± 1.7 mg/dL]). The HDL2 level of the high-flux group (0.48 ± 0.05 mmol/L [18.5 ± 2.1 mgjdLD did not differ from control (Fig 2A, Table 2). Females had higher HDL2 levels in all groups than males, but only in patients undergoing high-flux HD was this difference significant (Fig 2B). HDL2 was 0.33 ± 0.04 mmol/L (12.9 ± 1.7 mgjdL) in male highflux HD patients and 0.62 ± 0.03 mmol/L (23.8 ± 1.3 mg/dL) in females (P < 0.01). The differences in lipid levels between conventional and high-flux HD patients persisted when blacks alone were compared (Table 3). KT/V in high-flux patients (1.03 ± 0.03) was not different from that of conventional HD users (0.9 ± 0.06), P = 0.06. The mean dose of heparin administered to the high-flux (4,828.1 ± 641.9 U) and conventional (5,708.3 ± 857.9 U) HD groups did not differ significantly. A correlation between heparin dose and triglyceride level was not demonstrated by linear regression analysis (r = 0.3, P = 0.08). Diet assessed both by history taken by our clinical dietitian and by determination of peR and serum albumin did not differ

Table 2. Lipid Profiles of Patients and Controls Parameters

N Cholesterol (Mean ± SEM) HDL (Mean ± SEM) HDL2 (Mean ± SEM) Triglycerides (Mean ± SEM)

Controls (A)

4 .53 (175.3 1.42 (55.0 0.52 (20.3 1.55 (137.1

18 ± 0.24 mmol/L ± 9.1 mg/dL) ± 0.06 mmol/L ± 2.4 mg/dL) ± 0.04 mmol/L ± 1.7 mg/dL) ± 0.13 mmol/L ± 11.5 mgjdL)

B v A * .P < 0.01 ; B v At. P < 0.05; C v A :t. P < 0.05.

Conventional HD (8)

4.26 (164.9 1.12 (43.4 0.39 (15.2 2.39 (211.6

± ± ± ± ± ± ± ±

16 0.20 mmoljL 7 .9 mg/dL) 0.06 mmol/L* 2.5 mgjdL)* 0.04 mmoljLt 1.5 mgjdL)t 0.27 mmol/L* 24.1 mg/dL)*

High-Flux HD (C)

4.50 (173.9 1.23 (47.7 0.48 (18.5 1.62 (143.3

± ± ± ± ± ± ± ±

18 0.23 mmolfL 8.8 mgjdL) 0.06 mmol/L:t 2.5 mgjdL):t 0.05 mmoljL 2.1 mg/dL) 0.15 mmoljL 13.6 mg/dL)

JOSEPHSON, FELLNER, AND DASGUPTA

364

10

a

®

•

o

CO!
41. .. IIICNInO

o .....

;

0

~ ~I ~

(J

HDL

Co ...w,n,'onll

COftltOI

HDL 2

H, •• dl,I,111

Hl gh. Flu. H, •• dl ,I,.I.

Fig 2. (A) HDL and HDL2 cholesterol levels in controls, conventional HD patients, and high-flux HD patients. Those undergoing conventional or high-flux HD had significantly lower HDL content than controls (P < 0.01 and 0.05, respectively). HDL2 was lower in conventional HD patients than controls (P < 0.05). HDL2 levels of the high-flux subjects did not differ from controls. (B) Comparison of HDL2 cholesterol in male and female controls, conventional HD patients, and high-flux HD patients. HD~ levels in male and female controls and conventional HD patients was not significantly different. In subjects undergoing high-flux HD, females had significantly increased HDL2 levels compared with males (P < 0.01).

between the two groups. peR was 1.2 ± 0.1 g/ kg/d in the conventional dialysis group and 1.1 ± 0.2 g/kg/d in the others. Serum albumin was 3.73 ± .07 g/dL in the standard HD subjects and 3.76 ± .06 in the high-flux group. DISCUSSION

We have confirmed that patients undergoing standard HD have elevated triglyceride levels, as well as diminished total HDL and HDL2 levels. We have shown that patients using high-flux HD had normal triglyceride and normal HDL2 levels, but the latter may be gender-specific to females. Previous investigators have noted a HD HDL2 gender difference,14,18 but recognition of normal HDL2 in female high-flux patients is a new finding. We have compared lipid levels in inner city, predominately black, conventional and high-flux HD patients. Most studies demonstrating hypertriglyceridemia in HD patients examined white

subjects. Goldberg et aI compared lipid profiles in black and white HD patients and found higher triglyceride and lower HDL cholesterol levels in white than black males. 18 Racial differences in triglycerides have also been noted in nonuremic populations. 19 Whether white high-flux patients have less hypertriglyceridemia remains to be demonstrated. Overall differences in lipid profiles did not appear due to either "better" dialysis therapy as reflected in KT/V or nutritional status. Furthermore the age, duration of dialysis, race, and etiology of end-stage renal disease did not differ significantly between the two dialysis groups. There were two diabetics patients included in the high-flux group (both female) and four in the standard HD group (three females). Both highflux diabetic patients used insulin whereas only one of the conventional HD patients did. Highflux HD requires rapid blood flow rates. Only those patients with adequate access can be treated

Table 3. Lipid Profiles of Black HD Patients Parameters

N Cholesterol (Mean ± SEM) HDL (Mean ± SEM) HDL2 (Mean ± SEM) Triglycerides (Mean ± SEM)

Conventional HD

4.31 (166.6 1.10 (42.7 0.38 (14.8 2.40 (212.2

15 ± 0.21 mmoljL ± 8.3 mgjdL) ± 00 . 7 mmoljL ± 2 .6 mgjdL) ± 0.04 mmoljL ± 1.67 mgjdL) ± 02 . 9 mmolJL ± 25.8 mgjdL)

High-Flux HD

4.50 (173.9 1.23 (47.7 0.48 (18.5 1.62 (143.3

± ± ± ± ± ± ± ±

18 0.23 mmolJL 8.8 mgjdL) 0.06 mmoljL 2 .5mgjdL) 0.05 mmoljL 2 .1mgjdL) 0.15 mmoljL 13.6 mgjdL)

PValue

NS NS NS

<0.05

365

LIPID PROFILES IN HIGH-FLUX HEMODIALYSIS

with high-flux HD. Because patients with atherosclerosis are less likely to have access which tolerate rapid blood flows, one might invoke a selection bias. However, the majority of study patients were later switched to high-flux HD when machines became available. The mechanisms responsible for lower triglyceride levels in high-flux patients and higher HDL2 in female high-flux patients remain to be elucidated. Proposed mechanisms contributing to elevated triglyceride levels in conventional HD patients have included decreased triglyceride removal and increased triglyceride production. The former may be due to diminished lipoprotein lipase and hepatic lipase activity, depressed lecithin-cholesterol-acyl transferase activity, and impaired {3-oxidation offree fatty acids. The latter may result from dietary carbohydrate, glucose, and acetate in the dialysate, increased lipolysis, and increased levels of circulating insulin. 6 •20 There is no reason to assume that the mechanisms responsible for hypertriglyceridemia would occur in conventionally dialyzed patients and not in high-flux patients. Is there something specific about high-flux HD that may influence one or

more mechanisms responsible for hypertriglyceridemia? High-flux HD differs from conventional HD in its greater biocompatibility, the shorter exposure time of patients to dialyzer membranes, and the removal of large middle molecules. 21 ,22 Perhaps high-flux membranes eliminate a lipoprotein lipase inhibitor present in uremic plasma. 23 Whether any of these differences is responsible for our findings is uncertain and deserves further study. We have shown that patients undergoing highflux HD have less hypertriglyceridemia than patients undergoing conventional HD. One limitation of our report is that it is cross-sectional. We plan to examine the lipid patterns of patients as they switch from one HD modality to another. Studies are needed to investigate the mechanisms behind our findings and to determine whether reduction of hypertriglyceridemia will alter the incidence of atherosclerosis and cardiovascular disease in this patient population. ACKNOWLEDGMENT We would like to thank Christine L. Wolgemuth MS, RD, and Elaine Moragne MS, RD, for assistance in evaluating the dietary prescriptions and compliance.

REFERENCES I. Lindner A, Charra B, Sherrard DJ, et al: Accelerated

atherosclerosis in prolonged maintenance hemodialysis. N Engl J Med 290:697-701,1974 2. Burke JF Jr, Francos GC, Moore LL, et al: Accelerated atherosclerosis in chronic-dialysis patients-Another look. Nephron 21:181-185,1978 3. Nicholls AJ, Catto GRD, Edward N, et al: Accelerated atherosclerosis in long-term dialysis and renal-transplant patients: Fact or fiction? Lancet 1:276-278, 1980 4. US Renal Data System: USRDS 1990 Annual Data Report. Bethesda MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, August 1990 5. National Center for Health Statistics: Health, United States, 1990. Hyattsville, MD, Public Health Service, 1991 6. Chan MK, Varghese Z, Moorhead JF: Lipid abnormalities in uremia dialysis and transplantation. Kidney Int 19:625-637, 1981 7. Green D, Stone NJ, Krumlovsky FA: Putative atherogenic factors in patients with chronic renal failure. Prog Cardiovasc Dis 26:133-144, 1983 8. Bagdade JD, Porte D Jr, Bierman EL: Hypertriglyceridemia, a metabolic consequence of chronic renal failure. N Engl J Med 279:181-185, 1968 9. Bagdade J, Casaretto A, Albers J: Effects of chronic uremia, hemodialysis, and renal transplantation on plasma lipids and lipoproteins in man. J Lab Clin Med 87:37-48, 1976 10. Grutzmacher P, Marz W, Peschke B, et al: Lipoproteins

and apolipoproteins during the progression of chronic renal disease. Nephron 50:103-111, 1988 II. Cramp DG, Tickner DJ, Beale DJ, et al: Plasma triglyceride secretion and metabolism in chronic renal failure. Clin Chim Acta 76:237-241, 1977 12. Nicholls AJ, Cummings AM, Catto GRD, et al: Lipid relationships in dialysis and renal transplant patients. Q J Med 198:149-160,1981 13. Brunzell JD, Albers JJ, Haas LB, et al: Prevalence of serum lipid abnormalities in chronic hemodialysis. Metabolism 26:903-910, 1977 14. Rubies-Prat J, Espinel E, Joven J, et al: High-density lipoprotein cholesterol subfractions in chronic uremia. Am J Kidney Dis 9:60-65, 1987 15. Cramp DG, Tickner TR, Varghese Z, et al: Plasma lipoprotein patterns in patients receiving dialysis therapy for chronic renal failure. Clin Chim Acta 76:233-236, 1977 16. Heuck C-C, Eberhard R: Hyperiipoproteinemia in renal insufficiency. Nephron 25:1-7, 1980 17. Keshaviah P, Davis-Pollack R, Luhring D, et al: A Practical Guide to Rapid High Efficiency Dialysis. Minneapolis, MN, Regional Kidney Disease Program, 1987 18. Goldberg AP, Harter HR, Patsch W, et al: Racial differences in plasma high-density lipoproteins in patients receiving hemodialysis. N Engl J Med 308: 1245-1252, 1983 19. Tyroler HA, Hames CG, Krishan I, et al: Black-white differences in serum lipids and lipoproteins in Evans County. Prev Med 4:541-549, 1975

366 20. Appel G: Lipid abnormalities in renal disease. Kidney Int 39:169-183, 1991 21. Keshaviah P, Collins A: Rapid high-efficiency bicarbonate hemodialysis. Trans Am Soc Artif Intern Organs 32: 17-23, 1986 22. Rockel A, Abdelhamid S, Fliegel P, et al: Elimination

JOSEPHSON, FELLNER, AND DASGUPTA of low molecular weight proteins with high flux membranes. Contrib Nephrol 46:69-74, 1985 23. Murase T, Cattran DC, Rubenstein B, et al: Inhibition of lipoprotein lipase by uremic plasma, a possible cause of hypertriglyceridemia. Metabolism 24: 1279-1286, 1975