Atherogenic lipoprotein changes in the absence of hyperlipidemia in patients with chronic renal failure treated by hemodialysis

Transcription

1 /97/$ Elsevier Science Ireland Ltd. All rights reserved. PII S (97) Atherosclerosis 131 (1997) Atherogenic lipoprotein changes in the absence of hyperlipidemia in patients with chronic renal failure treated by hemodialysis Tetsuo Shoji a, Yoshiki Nishizawa a, *, Takahiko Kawagishi a, Masaaki Tanaka a, Koichi Kawasaki a, Tsutomu Tabata b, Takashi Inoue b, Hirotoshi Morii a a Second Department of Internal Medicine, Osaka City Uni ersity Medical School, 1-5-7, Asahi-machi, Abeno-ku, Osaka 545, Japan b Di ision of Internal Medicine, Inoue Hospital, 16-17, Enoki-cho, Suita, Osaka 564, Japan Received 1 November 1996; received in revised form 12 February 1997; accepted 3 March 1997 Abstract We compared plasma lipid and lipoprotein parameters between 210 chronic renal failure patients treated by hemodialysis and 223 age- and sex-matched healthy control subjects to examine whether atherogenic lipoprotein changes were present in hemodialysis patients in the absence of hyperlipidemia. The hemodialysis group showed higher levels of plasma triglycerides, very low density lipoprotein (VLDL) cholesterol, and intermediate density lipoprotein (IDL) cholesterol and a lower level of high density lipoprotein (HDL) cholesterol. Low density lipoprotein (LDL) cholesterol of the hemodialysis group was not elevated but their LDL was significantly more triglyceride-enriched than that of controls. Subjects were then divided into five categories according to their plasma triglyceride levels at an interval of 50 mg/dl, and comparison was made between the two groups in the same range of plasma triglycerides. Hemodialysis patients again showed higher levels of VLDL- and IDL-cholesterol, and lower levels of HDL-cholesterol than the control group even in the plasma triglycerides-matched comparisons. Similarly, higher VLDLand IDL-cholesterol levels in hemodialysis patients were significant in plasma total cholesterol-matched subgroup comparisons. Multiple regression analysis indicated that the relationship between plasma lipid concentrations and individual lipoprotein levels were substantially altered in uremic state. The 95th percentile level of IDL-cholesterol in the nonuremic controls was 15 mg/dl, and 45% of hemodialysis patients exceeded this level. Decreased HDL-cholesterol levels 35 mg/dl were seen in 6% of the control and 38% of the hemodialysis group. Elevated IDL-cholesterol and decreased HDL-cholesterol were persistently found in hemodialysis patients with normal lipid levels. It is concluded that hemodialysis patients exhibited more atherogenic lipoprotein profile than nonuremic subjects with comparable levels of plasma triglycerides and total cholesterol. Especially, increased IDLand decreased HDL-cholesterol levels in hemodialysis patients persisted even at very low levels of plasma lipids. Since elevated IDL and decreased HDL-cholesterol are implicated in the progression of atherosclerosis, these findings are of clinical importance in the diagnosis of lipoprotein disorder in chronic renal failure Elsevier Science Ireland Ltd. Keywords: Lipoprotein; Cholesterol; Triglycerides; Hemodialysis; Uremia 1. Introduction Plasma lipoprotein disorder is an important factor in the pathogenesis of atherosclerosis [1,2]. Previous studies showed that an elevated level of low density lipoprotein (LDL) is a strong risk factor of atherosclerosis * Corresponding author. Tel.: ; fax: [1], while increased high density lipoprotein (HDL) is a protective factor [3,4]. In addition, angiographic studies in the general population [5 11] emphasize the roles of triglyceride-rich lipoproteins including subfractions of very low density lipoproteins (VLDL) and intermediate density lipoprotein (IDL) in the progression of coronary artery disease. In fact, these triglyceride-rich lipoproteins are present in human atherosclerotic plaque [12].

2 230 T. Shoji et al. / Atherosclerosis 131 (1997) In patients with chronic renal failure treated by hemodialysis, atherosclerosis is accerelated [13,14]. We revealed that intima-media thickness of carotid and femoral arteries is greater in hemodialysis patients than age- and gender-matched healthy subjects [15]. Hemodialysis paitents often show lipid and lipoprotein abnormalities a condition called uremic dyslipidemia which is characterized by hypertriglyceridemia, elevated VLDL and IDL, and lowered HDL-cholesterol. These lipoprotein abnormalities may be an important factor for accelerated atherosclerosis in hemodialysis patients [16]. In routine clinical chemistry, plasma total cholesterol and triglycerides are measured for screening lipoprotein abnormalities. However, plasma total cholesterol is not usually elevated in hemodialysis patients [16,17]. Ron et al. [18] reported accumulation of remnant lipoproteins in normocholesterolemic patients with uremia. Therefore potentially atherogenic lipoprotein abnormalities may be present behind the apparently normal plasma lipid levels in these patients. In the present study, we measured plasma lipids and the four major lipoprotein classes in 210 hemodialysis patients and 223 age- and sex-matched healthy control subjects to examine whether individual lipoprotein abnormalities can be recognized by measuring plasma total cholesterol and triglycerides. The results indicated that hemodialysis patients exhibited more atherogenic lipoprotein profile than nonuremic subjects having comparable levels of plasma triglycerides and total cholesterol, and that potentially atherogenic lipoprotein changes were not rare in the hemodialysis population even in the absence of apparent hyperlipidemia. These findings are of clinical importance in the diagnosis of lipoprotein disorder in chronic renal failure. 2. Subjects and methods 2.1. Subjects The subjects were 210 nondiabetic hemodialysis patients and 223 age- and sex-matched healthy control subjects. Their clinical characteristics were shown in Table 1. The hemodialysis patients received 3 5 h of regular hemodialysis, three times a week in morning sessions using bicarbonate dialysate. The patients were taking antihypertensive drugs, active vitamin D, phosphate binder calcium carbonate, but no hypolipidemic agents at the time of this study. They received recombinant human erythropoietin injections at a dosage of U/week for renal anemia. Control subjects were selected from participants of a local health check program. Exclusion criteria were overt proteinuria, liver dysfunction as defined by serum alanine aminotransferase (ALT) greater than 50 IU, fasting blood glucose greater than 140 mg/dl, treated hypertension, treated hyperlipidemia and treated diabetes mellitus. Because we intended to compare lipid and lipoprotein parameters between groups, no exclusion criteria was made regarding plasma lipid levels to avoid selection bias Lipid and lipoprotein measurements Plasma lipoprotein fractions were measured by Bronzert and Brewer method [19] with modification. For adjusting plasma density, we prepared three kinds of KBr solutions of d=1.006 g/ml (solution A), d= g/ml (solution B) and d=1.120 g/ml (solution C). Plasma from one subject was put into three Hitachi 0.5PC tubes (tubes A, B and C), 200 l plasma per tube. Equal volumes of KBr density solution were added to each tube and vortexed, so that the densities in tubes A, B and C were 1.006, and g/ml, respectively. Samples were spun at rpm ( g) for 3 h at 4 C in a Hitachi ultracentrifuge model CS120F with a Hitachi rotor model RP100AT2. Top and bottom layers were clearly separated by a colorless intermediate layer. The bottom layer plus part of the clear middle layer (200 l) was carefully aspirated and transfered to a small plastic tube by a microsyringe with a fine needle, or by a Pipetman with a long and narrow tip. The remaining top layer plus part of the clear layer was vortexed well and transfered to a small plastic tube. Separation of lipoproteins by this method was ascertained by polyacrylamide gel electrophoresis (PAGE) [20] using commercially available precast cyrindrical cm gels (Lipophor, JOKO, Tokyo). Prestained original plasma and top and bottom fractions in tubes A, B and C were subjected to PAGE. Densitometric patterns of the gels indicated that the top fraction of tubes A, B and C contained VLDL, VLDL+IDL and VLDL+IDL+LDL respectively, and that the bottom fractions of tubes A, B and C contained IDL+LDL+HDL, LDL+HDL and HDL, respectively. Cholesterol concentration in Table 1 Characterization of the subjects Control Hemodialysis P-value Age (years) NS Gender (M/F) 88/135 86/124 NS* BMI (kg/m 2 ) Current smokers (%) * Duration of hemo dialysis (years) Plasma total cholesterol (mg/dl) Plasma triglycerides (mg/dl) Values are expressed as mean S.E., P-values by Student s t-test or by * 2.

3 T. Shoji et al. / Atherosclerosis 131 (1997) VLDL was directly measured in the top fraction of tube A. Cholesterol in IDL was calculated by subtracting cholesterol in the top of tube A (VLDL) from that in the top of tube B (VLDL+IDL). Cholesterol in LDL was calculated by subtracting cholesterol in the top of tube B (VLDL+IDL) from that in the top of tube C (VLDL+IDL+LDL). Cholesterol in HDL was measured as cholesterol in the bottom of tube C. Triglycerides in the four lipoprotein fractions were similarly determined by subtraction. This one-step micro-ultracentrifugation method was validated by comparing with the standard sequential preparative ultracentrifugation by Havel et al. [21]. For this purpose, cholesterol concentrations in VLDL, IDL, LDL and HDL from 50 plasma samples were measured by the two methods. The results by the two methods agreed well, giving correlation coefficients of , slopes of , and intercepts of mg/dl in simple regression analyses. Total cholesterol and triglycerides in lipoprotein-containing plasma fractions were assayed by enzymatical methods using commercially available kits (Cholesterol E-test by cholesterol oxidase-daos method; Triglycerides E-test by glycerol 3 phosphate oxidase-daos method, Wako, Japan) using standards included in the kits. For each assay, 20 l of sample was mixed with the reagents, incubated for 5 min at 37 C, and absorbance at 600 nm was read by a Hitachi spectrophotometer model 2000A Statistics Data were summarized as mean standard error (S.E.). Difference in prevalence was evaluated by 2 test. Comparison of mean values between the two groups was performed by Student s t-test. Correlation between two variables was evaluated by simple regression analysis. Comparison of two regression slopes was performed according to Ichihara [22]. Multiple regression was employed to evaluate independent associations among variables. P-values less than 0.05 were taken to be significant. 3. Results As shown in Table 1, plasma total cholesterol in hemodialysis patients was not elevated but rather decreased as compared with that of control subjects. Table 2 gives mean levels of plasma lipoproteins in hemodialysis and control subjects. Hemodialysis patients showed significantly higher cholesterol levels in VLDL and IDL, whereas their cholesterol levels in plasma, LDL, and HDL were significantly lower than the control subjects. Increases of triglycerides in plasma, IDL, and LDL in hemodialysis patients were Table 2 Lipoprotein parameters in the control and hemodialysis populations Control Hemodialysis P-value Cholesterol (mg/dl) VLDL IDL LDL HDL Triglycerides (mg/dl) VLDL IDL LDL HDL NS Cholesterol/triglycerides ratio (mg/mg) VLDL IDL LDL HDL Values are expressed as mean S.E. P-values by Student s t-test. significant. Increased VLDL-triglycerides were at a borderline significance. Regarding cholesterol/triglycerides ratio in lipoprotein fractions, the hemodialysis group showed an increase in VLDL and decrease in IDL, LDL and HDL. Frequency distributions of lipoprotein-cholesterol levels were compared between hemodialysis and control subjects (Fig. 1). Distributions of VLDL- and IDLcholesterol of hemodialysis patients were shifted to the right and those of LDL- and HDL-cholesterol to the left compared with those of the control subjects. Elevated IDL-cholesterol greater than 15 mg/dl was seen in only 5% of the control group, whereas it was found in 45% of hemodialysis patients. Low HDL-cholesterol (less than 35 mg/dl) was seen in 6% of the control group, while it was found in 38% of hemodialysis patients. Subjects were divided into five categories according to their plasma triglyceride levels at intervals of 50 mg/dl, and lipoprotein parameters were compared between the two groups within the same range of plasma triglycerides (Fig. 2). VLDL- and IDL-cholesterol increased as plasma triglycerides increased in both hemodialysis and control groups. In the same triglyceride category, however, both VLDL- and IDL-cholesterol were significantly higher in hemodialysis patients than those of the controls. LDL-cholesterol did not show consistent relationship with plasma triglycerides, and it was significantly lower in hemodialysis patients than the controls when plasma triglycerides were lower than 200 mg/dl. HDL-cholesterol decreased as plasma triglycerides increased in the hemodialysis and control groups. Within the same triglyceride category, however, HDL-cholesterol was significantly lower in hemodialysis patients than in the control subjects.

4 232 T. Shoji et al. / Atherosclerosis 131 (1997) Similarly, the subjects were divided into five categories according to their plasma total cholesterol levels to compare lipoprotein parameters between the hemodialysis and control groups within the same range of plasma total cholesterol (Fig. 3). In such comparisons, hemodialysis patients consistently showed higher IDLcholesterol and lower LDL-cholesterol levels than the control subjects. HDL-cholesterol in the hemodialysis group was significantly lower that of the control group in the range of total cholesterol 161 mg/dl. Multiple regression analyses were performed to examine the impact of uremia on individual lipoprotein levels, independent of plasma total cholesterol, triglycerides, age and gender. Combination of uremia and other four variables predicted VLDL-cholesterol levels to a significant extent (Table 3). VLDL-cholesterol showed a significant association with uremia, although it had a much closer association with plasma triglycerides. Simple regression analysis between plasma triglycerides (X) and VLDL-cholesterol (Y) indicated that regression slopes in the hemodialysis group were Fig. 2. Comparisons of individual lipoprotein levels between hemodialysis and control groups in the same ranges of plasma triglyceride. Subjects were divided into five categories according to their plasma triglyceride concentrations and cholesterol levels in each lipoprotein fraction were compared between hemodialysis and control groups. Mean S.E. * P 0.05, ** P 0.01, *** P versus control group by Student s t-test. Fig. 1. Frequency distribution of lipoprotein-cholesterol measurements. Individual lipoprotein cholesterol levels were determined by the one-step micro-ultracentrifugation and frequency distribution was compared between hemodialysis and control groups. significantly greater than that in the control group (0.257 in hemodialysis versus in the control, P=0.001). IDL-cholesterol was significantly affected by uremia and the impact of uremia on IDL-cholesterol was approximately twice greater than that of triglycerides or total cholesterol. Female gender showed a modest but significant increasing effect on IDL cholesterol. LDL-cholesterol was most closely associated with total cholesterol levels. Plasma triglycerides, age and gender were also significant correlates of LDL-cholesterol, whereas uremia did not show significant association with LDL-cholesterol. HDLcholesterol was most closely and inversely associated with plasma triglycerides. Total cholesterol and female gender showed increasing effects on HDL-cholesterol, whereas uremia had a significant decreasing effect on HDL-cholesterol.

5 T. Shoji et al. / Atherosclerosis 131 (1997) To know how low plasma lipids should be in order to normalize lipoprotein cholesterol levels in hemodialysis patients, we calculated prevalence of elevated VLDL-, IDL-, and LDL-cholesterol and decreased HDL-cholesterol levels employing various criteria for plasma total cholesterol and triglycerides (Table 4). Here, VLDLcholesterol of 30 and 40 mg/dl corresponded to plasma triglycerides of 150 and 200 mg/dl [23]. IDL-cholesterol of 15 and 20 mg/dl were the 95th and 99th percentile levels, respectively of the 223 healthy controls. When normolipidemia is defined as plasma total cholesterol 200 mg/dl and triglycerides 200 mg/dl [23], prevalence of VLDL-cholesterol 40 mg/dl and LDLcholesterol 130 mg/dl were 3 and 0% in normolipidemic healthy controls. These low prevalences could be achieved also in hemodialysis patients having plasma total cholesterol 200 mg/dl and triglycerides Fig. 3. Comparisons of individual lipoprotein levels between hemodialysis and control groups in the same ranges of plasma total cholesterol. Subjects were divided into five categories according to their plasma total cholesterol concentrations and cholesterol levels in each lipoprotein fraction were compared between hemodialysis and control groups. Mean S.E. * P 0.05, ** P 0.01, *** P versus control group by Student s t-test mg/dl. In contrast, elevated IDL-cholesterol and decreased HDL-cholesterol levels were persistently found in hemodialysis patients even at very low lipid levels. For example, in hemodialysis patients with plasma total cholesterol 160 mg/dl and triglycerides 100 mg/dl, IDL-cholesterol 15 mg/dl and HDLcholesterol 35 mg/dl were still seen in 12 and 31%, respectively. In these hemodialysis patients, the mean level of IDL-cholesterol was mg/dl (mean S.E.), still significantly higher than that of the total control subjects. 4. Discussion Plasma total cholesterol and triglycerides are measured for screening lipoprotein disorders, especially in the risk assesment of coronary artery disease. Hypercholesterolemia is usually a reflection of elevated LDL and hypertriglyceridemia is a marker of an increased VLDL level in the general population. In some disease conditions, however, significant changes of individual lipoprotein levels may be present in the absence of apparent hyperlipidemia. Elevated IDL was reported in normolipidemic patients with non-insulin-dependent diabetes [24]. The present study has shown that hemodialysis patients have a variety of potentially atherogenic lipoprotein changes, especially elevated IDL- and decreased HDL-cholesterol levels, even in the absence of apparent hyperlipidemia. IDL-cholesterol was increased in the hemodialysis population as previously shown in smaller studies by others [18,25] and by ourselves [26,27]. The present study could further demonstrate markedly high prevalence of elevated IDL in the hemodialysis population. IDL-cholesterol of 15 mg/dl was the 95th percentile value in the 223 nonuremic control subjects. As many as 95 out of 210 hemodialysis patients exceeded the 95th percentile level. Because an elevated IDL level is a strong correlate of the extent and progression of angiographically assessed coronary atherosclerosis [5 11], the increased IDL-cholesterol in the hemodialysis population would be an atherogenic lipoprotein change. However, such a risk can not readily be recognized by merely measuring plasma cholesterol and triglycerides. Ron et al. [18] reported that IDL-cholesterol was elevated in uremic patients independently of the presence of hypercholesterolemia. It was also the case with this study, and we showed that IDL-cholesterol was elevated in hemodialysis patients even when triglycerides were not elevated. These findings do not indicate that IDL-cholesterol was not related to plasma lipid levels, because IDL-cholesterol increased as a function of plasma triglycerides and total cholesterol concentrations. Multiple regression analysis also indicated that IDL-cholesterol was independently assocated with

6 234 T. Shoji et al. / Atherosclerosis 131 (1997) Table 3 Multiple regression analyses of factors determining VLDL-, IDL-, LDL- and HDL-cholesterol levels Dependent variables Independent variables Intercept R 2 TC TG Age Gender Urernia VLDL-Chol b **** 0.787**** 0.046** (NS) 0.364**** 0.914**** IDL-Chol b **** 0.314**** (NS) 0.066* 0.624**** 0.539**** LDL-Chol b **** 0.231**** 0.038* 0.068**** (NS) 0.875**** HDL-Chol b **** 0.551**** (NS) 0.072* 0.201**** 0.590**** The values in the table indicate partial regression coefficients (b) and standard regression coefficients ( ). TC, total cholesterol; TG, triglycerides; R 2, multiple coefficient of determination. * P 0.05, ** P 0.01, *** P 0.001, **** P Dummy variables were used for gender (0 for female, 1 for male) and for uremia (0 for not uremic, 1 for uremic). plasma total cholesterol and triglycerides. Importantly, the impact of uremia on IDL-cholesterol was much greater than those of these lipids levels. HDL cholesterol of hemodialysis patients was lower than that of the nonuremic controls, in good agreement with earlier studies [26,28,29]. Because HDL-cholesterol is in an inverse correlation with plasma triglycerides [3,4], hypertriglyceridemia is an important factor explaining decreased HDL-cholesterol levels in hemodialysis patients. In this study, HDL-cholesterol decreased as plasma triglycerides increased in both hemodialysis and control populations. However, HDL-cholesterol was still lower in hemodialysis patients even in triglyceride-matched subgroup comparisons. This indicates that lower HDL-cholesterol in hemodialysis patients was not due simply to hypertriglyceridemia. Prevalence of decreased HDL-cholesterol ( 35 mg/dl) remained high when analyzed in hemodialysis patients with plasma triglycerides ( 100 mg/dl). Furthermore, multiple regression analysis indicated a significant negative impact of uremia on HDL-cholesterol that was independent of plasma lipids, age and gender. We previously showed that 51% of the variation in HDL-cholesterol in chronic renal failure patients was accounted for by triglycerides, LPL, HTGL and LCAT [26]. Therefore, the impact of uremia on HDL-cholesterol was mediated presumably by changes in these enzyme activities. Whatever the mechanism is, it is clearly shown that when plasma total cholesterol and triglyceride levels were comparable, hemodialysis patients had a lower HDL-cholesterol, and therefore greater risk for atherosclerosis, than the general population. Hemodialysis patients had higher VLDL-cholesterol and a higher cholesterol/triglyceride ratio in VLDL than the control subjects. Cholesterol-enrichment of uremic VLDL suggests relative increase of cholesterolrich VLDL subpopulations or -VLDL-like particles in the fraction of the patients. Because the atherogenicity of such lipoproteins is shown in recent angiographic studies [5 11], VLDL of hemodialysis patients would be potentially more atherogenic than that of nonuremic subjects. The simple regression slope between plasma triglycerides (X-axis) and VLDLcholesterol (Y-axis) was in the control group. This figure is quite close to that by Friedewald et al. [30] who calculated VLDL-cholesterol as 20% of plasma triglycerides in mg/dl in nonuremic subjects. In contrast, the regression slope was significantly greater in hemodialysis (slope=0.257, P=0.001) than the control groups. These figures would help estimating VLDL-cholesterol levels more precisely from routine lipid measurements in hemodialysis patients. LDL-cholesterol was not increased but rather decreased in hemodialysis patients. This is in good agreement with earlier studies by others [28] and by ourselves [26,29]. In contrast to the effects of uremia on VLDL, IDL and HDL levels, uremia did not show significant impact on LDL-cholesterol levels independently of plasma lipids, gender and age in a multiple regression analysis. The clinical significance of the low LDLcholesterol in hemodialysis patients is unknown at present. Although LDL-cholesterol level was decreased, it was at the expence of enrichment of LDL with triglycerides. Triglyceride-enriched LDL may be small and dense [31] and therefore more atherogenic [32]. In our study, a certain level of total cholesterol or triglycerides reflected different levels of individual lipoproteins between the hemodialysis and control populations. This suggests that the same lipid levels could represent different degrees of atherogenic risk in different populations. How can we recognize potentially atherogenic lipoprotein changes in hemodialysis patients by routine clinical tests? HDL-cholesterol levels can be directly measured by precipitation methods without ultracentrifugation. In contrast, measurement

7 T. Shoji et al. / Atherosclerosis 131 (1997) Table 4 Various criteria of normolipidemia and prevalence (%) of abnormal lipoprotein cholesterol levels in hemodialysis patients Control subjects Hemodialysis patient Plasma TC (mg/dl) and 200 Total healthy control (%) Total hemodialysis Plasma TG (mg/dl) VLDL-cholesterol 30 mg/dl mg/dl IDL-cholesterol 15 mg/dl mg/dl LDL-cholesterol 150 mg/dl mg/dl HDL-cholesterol 40 mg/dl mg/dl The table shows prevalences (%) of elevated VLDL-, IDL-, LDL- and HDL-cholesterol in the control subjects and the hemodialysis patients with normolipidemia as defined by different criteria for plasma total cholesterol and triglyceride concentrations. TC, plasma total cholesterol; TG, triglycerides. of IDL-cholesterol requires ultracentrifugation. Sequential floatation method [21] is a standard technique, but it is time-consuming. Although we employed a rapid micro ultracentrifugation technique, it may not be simple enough for routine assays. It is possible to estimate IDL-cholesterol by using the multiple regression model including plasma total cholesterol, triglycerides, age, gender and presence or absence of uremia, although such an estimation is not sufficiently accurate. We need more practical methods to detect elevated IDL, because probably 37% of normolipidemic hemodialysis patients will have raised atherogenic lipoprotein, IDL. We previously reported that hemodialysis patients had greater intima-media thickness of carotid and femoral arteries than age- and gender-matched healthy subjects [15]. In the present study, we showed that elevated IDL-cholesterol and lowered HDL-cholesterol levels are present even in the absence of hyperlipidemia in the hemodialysis population and these changes presumably contribute to the advanced atherosclerosis in this population. In chronic renal failure, hepatic triglyceride lipase (HTGL) activity is markedly reduced [26,29], and the impairment of this enzyme activity is closely associated with both the increased IDL-cholesterol [33] and the decreased HDL-cholesterol levels [26] in uremic patients. HTGL level is shown to be adversely affected by hypocalcemia and secondary hyperparathyroidism in renal failure [26]. Therefore, these endocrinological alterations due to renal failure could account, at least in part, for the unique and potentially atherogenic lipoprotein profile and advanced atherosclerosis in hemodialysis patients. In support of this, elevation of parathyroid hormone level has been shown to be an independent risk factor for the arterial wall thickening in hemodialysis patients [15]. In conclusion, the relationship between plasma lipids and lipoprotein levels was altered in chronic renal failure, and potentially atherogenic lipoprotein changes were present even in normolipidemic hemodialysis patients. Results in the present study call for more careful interpretation of plasma lipid measurements in hemodialysis patients. References [1] Kannel, WB, Castelli, WP, Gordon, T, McNamara PM. Serum cholesterol, lipoproteins and risk of coronary heart disease. The Framimgham Study. Ann Intern Med 1971;74:1 12. [2] Stamler J, Wentworth D, Neaton JD. Is the relationship between serum cholesterol and risk of premature death from coronary heart disease continuous and graded? Findings in primary screenees of the Multiple Risk Factor Intervention Trial (MRFIT). JAMA 1986;256: [3] Miller GJ, Miller NE. Plasma high-density lipoprotein concentration and development of ischemic heart disease. Lancet 1975;1: [4] Gordon T, Castelli WP, Hjortland MC, Dawber TR. High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study. Am J Med 1977;62: [5] Tatami R, Mabuchi H, Ueda R, Haba T, Kametani T, Ito S, Koizumi J, Ohta M, Miyamoto S, Nakayama A, Kanaya H, Oiwake H, Genda A, Takeda R. Intermediate-density lipoprotein and cholesterol rich very low density lipoprotein in angiographically determined coronary artery disease. Circulation 1981;64: [6] Kameda K, Matsuzawa Y, Kubo M, Ishikawa K, Maejima L, Yamamura T, Yamamoto A, Tarui S. Increased frequency of

8 236 T. Shoji et al. / Atherosclerosis 131 (1997) lipoprotein disorders similar to type III hyperlipoproteinemia in survivors of myocardial infarction in Japan. Atherosclerosis 1984;51: [7] Krauss RM, Lindgren FT, Williams PT, Kelsey SF, Brensike J, Vranizan K, Detre KM, Levy R. Intermediate-density lipoproteins and progression of coronary disease in hypercholesterolaemic men. Lancet 1987;2: [8] Watts GF, Mandelia S, Bruns JNH, Slavin BM, Coitart DJ, Lewis B. Independent associations between plasma lipoprotein subfraction levels and the course of coronary artery disease in the St Thomas Atherosclerosis regression study (STARS). Metabolism 1993;42: [9] Phillips NR, Waters D, Havel RJ. Plasma lipoproteins and progression of coronary artery disease evaluated by angiography and clinical events. Circulation 1993;88: [10] Drexel H, Amann FW, Beran J, Rentsch K, Candinas R, Muntwyler J, Luethy A, Gasser T, Follath F. Plasma triglycerides and three lipoprotein cholesterol fractions are independent predictors of the extent of coronary atherosclerosis. Circulation 1994;90: [11] Rapp JH, Lespine A, Hamilton RL, Colyvas N, Cahmeton AH, Tweedie-Hardman J, Kotite L Kunitake ST, Havel RJ, Kane JP. Triglyceride-rich lipoproteins isolated by selected-affinity antiapolipoprotein B immunosorption from human atherosclerotic plaque. Arterioscler Thromb 1994;14: [12] Mack WJ, Krauss RM, Hodis HN. Lipoprotein subclasses in the Monitored Atherosclerosis regression study (MARS): Treatment effects and relation to coronary angiographic progression. Arterioscler Thromb Vasc Biol 1996;16: [13] Lindner A, Charra B, Sherrard DJ, Scriboer BH. Accelerated atherosclerosis in prolonged maintenance hemodialysis. N Engl J Med 1974;290: [14] Ma KW, Greene EL, Raij L. Cardiovascular risk factors in the chronic renal failure and hemodialysis population. Am J Kidney Dis 1992;19: [15] Kawagishi T, Nishizawa Y, Konishi T, Kawasaki K, Emoto M, Shoji T, Tabata T, Inoue T, Morii H. High-resolution B-mode ultrasonography in evaluation of atherosclerosis in uremia. Kidney Int 1995;48: [16] Attman P-O, Samuelsson O, Alaupovic P. Lipoprotein metabolism and renal failure. Am J Kidney Dis 1993;21: [17] Bagdade JD, Porte D Jr, Bierman EL. Hypertriglyceridemia: A metabolic consequence of chronic renal failure. N Engl J Med 1968;279: [18] Ron D, Oren L, Aviram M, Better OS, Brook JG. Accumulation of lipoprotein remnants in patients with chronic renal failure. Atherosclerosis 1983;46: [19] Bronzert TJ, Brewer HB, Jr. New micromethod for measuring cholesterol in plasma lipoprotein fractions. Clin Chem 1977;23: [20] Muniz N. Measurement of plasma lipoproteins by electrophoresis on polyacrylamide gel. Clin Chem 1977;23: [21] Havel RJ, Eder HA, Bragdon JH. The distribution and chemical composition of ultracentrifuially separated lipoproteins in human serum. J Clin Invest 1955;34: [22] Ichihara K. Statistics for Bioscience. Practical Technique and Theory. Tokyo: Nankodo Co. Ltd, 1990, pp [23] The expert panel: Summary of the second report of the National cholesterol education program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (adult treatment panel II). JAMA 1993;269: [24] Kasama T, Yohino G, Iwatani L, Iwai M, Hatanaka H, Kazumi, Oimomi M, Baba S. Increased cholesterol concentration in intermediate density lipoprotein fraction of normolipidemic noninsulin-dependent diabetics. Atherosclerosis 1987;63: [25] Nestel PJ, Fidge NH, Tan MH. Increased lipoprotein remnant formation in chronic renal failure. N Engl J Med 1982;307: [26] Shoji T, Nishizawa Y, Nishitani H, Yamakawa M, Morii H. Impaired high density lipoprotein metabolism in uremic patients. Kidney Int 1992;41: [27] Nishizawa Y, Shoji T, Emoto M, Kawasaki K, Konishi T, Tabata T, Inoue T, Morii H. Reduction of intermediate density lipoprotein by pravastatin in hemo- and peritoneal dialysis patients. Clin Nephrol 1995;43: [28] Attman P-O, Alaupovic P. Lipid abnormalities in chronic renal insufficiency. Kidney Int 1991;39(Suppl):S16 S23. [29] Shoji T, Nishizawa Y, Nishitani H, Yamakawa M, Morii H. Roles of hypoalbuminemia and lipoprotein lipase on hyperlipoproteinemia in continuous ambulatory peritoneal dialysis. Metabolism 1991;40: [30] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18: [31] Coresh J, Kwiterovich PO Jr, Smith HH, Bachorik PS. Association of plasma triglyceride concentration and LDL particle diameter, density and chemical composition with premature coronary artery disease in men and women. J Lipid Res 1993;34: [32] Austin MA, Breslow JL, Hannekens CH, Buring JE, Willett WC, Krauss RM. Low density lipoprotein subclass patterns and risk of myocardial infarction. J Am Med Assoc 1988;260: [33] Shoji T, Nishizawa Y, Kawagishi T, Morii H. Secondary hyperparathyroidism, increased IDL and atherosclerosis in hemodialysis patients. Nephron (in press)..