Women have a larger and less atherogenic low density lipoprotein particle size than men

atherosclerosis

ELSEVTER. Atherosclerosis 119 (1996) 181-190

Women have a larger and less atherogenic low density lipoprotein particle size than men

Matti Nikkilg*“, Timo Pitkgjgrvl ‘b, Timo Koivula”, Tiina SolakivFd, Terho Lehtimiikicsd, Pekka Laippala”, Hannu Jokela”, Erkki LehtomHkib, Kaija Sepp2,

Pekka Sillanaukeef

‘Department of Internal Medicine, City Hospital of’ Tampere, Hatanpiiii, 33100 Tampere IO, Finland bCommunity Health Centre of‘ Tampere, Tampere, Finland

‘Department of Clinical Chemistry, Tampere University Hospital, Tampere, Finland ‘Medical School, University of Tampere, Tampere, Finland

‘School of Public Health, University of Tampere, Tampere, Finland ‘Biomedical Research Center, ALKO Ltd, Finland

Received 28 December 1995; revision received 23 June 1995; accepted 29 June 1995

Abstract

Some epidemiological studies have shown that serum total cholesterol increases with age, especially in women. On the other hand, the risk of coronary artery disease is smaller in women than in men. Earlier studies have shown that a small dense low density lipoprotein (LDL) is more atherogenic than a large LDL. We studied LDL size and apolipoprotein E (apo E) phenotypes in premenopausal and postmenopausal women and in men at the same age. In this study 342 subjects participating in a health screening study were examined. There were four subgroups: 40-year-old men (n = 85), 40-year-old women (n = 80), 70-year old men (n = 88) and 70-year-old women (n = 89). In the present study LDL size was larger (P < 0.01) in women (26.39 + 0.07 nm) than in men (25.95 + 0.07 nm). We found that LDL size correlated highly positively (r = 0.606; P < 0.001) with serum high density lipoprotein (HDL) concentration and inversely with serum triglyceride concentration (r = - 0.627; P < 0.001). Measuring serum HDL cholesterol and triglycerides iu health screening studies gives information indirectly about LDL size and its atherogenicity. Apo E phenotype was not significantly associated with serum triglycerides, but was associated with LDL size, LDL cholesterol, total cholesterol and HDL cholesterol. In our sample LDL size decreased and LDL cholesterol and total cholesterol increased according to the most prevalent apo E phenotypes in the order E2/3, E3/3, E3/4 and E4/4. Subjects with phenotype apo E 4/4 had the smallest LDL size (25.70 + 0.19 nm), the highest total cholesteral (6.53 &- 0.35 mmol/l) and the lowest HDL cholesterol values (1.28 f 0.04 mmol/l). We conclude that there was a significant interaction between sex and age in serum total cholesterol which was highest in older women. However, their LDL size was larger and their LDL is less atherogenic. Apo E phenotype had a significant influence on LDL size.

Keywords: Low density lipoprotein; Apolipoprotein E; LDL size; Coronary artery disease

* Corresponding author

0021-91.50/96/$09.50 0 1996 Elsevier Science Ireland Ltd. All rights reserved SSDI 0021-9150(95)05645-D

182 M. Nikkilii et al. / Atherosclerosis 119 (1996) 181-190

1. Introduction

Many epidemiological studies have shown that moderately elevated serum total cholesterol is not a cardiovascular risk factor for elderly women [1,2], while in middle-aged men there is a strong correlation between increased serum total cholesterol and a high coronary artery disease (CAD) mortality [3]. It has been shown that low density lipoprotein (LDL) cholesterol correlates with CAD even better than total cholesterol [4]. The diameter of LDL varies considerably from one individual to another. The prevalence of small dense LDL is higher in men than in women [5]. These particles are atherogenic and associated with an increased triglyceride level [5]. Hyper- triglyceridemia would appear to be an important risk factor for CAD and correlates inversely with LDL size [6]. Austin et al. identified three LDL phenotypes: (1) pattern A, characterized by larger LDL ( > 25.5 nm); (2) pattern B, in which small dense LDL particles predominate (peak diameter < 25.5 nm); and (3) an intermediate pattern [7]. Pattern B is accompanied by higher levels of serum triglycerides and lower levels of high density lipoprotein (HDL) cholesterol.

Some authors have reported that increased numbers of small dense LDL particles, found in hyperapobetalipoproteinemia (hyperapo B), were strongly associated with premature CAD [8]. Many studies show a high prevalence of small dense LDL among patients with premature CAD [5,6]. However, it is unclear whether LDL particle characteristics are independent risk factors for CAD or results of other atherogenic lipid abnormalities. CAD is associated with both hyper- triglyceridemia and the small dense LDL particles VI.

Apolipoprotein E (apo E) is found in chylomi- crons, very low density lipoprotein (VLDL) remnants and HDL particles [9]. Three common alleles (~2, ~3, ~4) determine six apo E phenotypes (212, 213, 214, 3/3, 314, 4/4). The apo E isoforms have been shown to influence concentrations of total cholesterol apolipoprotein B and LDL cholesterol [lo]. Individuals with the ~2 allele have a decreased production of LDL from VLDL and increased catabolism of LDL [l 11. The opposite effects are seen in those with allele ~4 [12].

The purpose of our study was to compare lipid profile and LDL size between 40- and 70-year-old men and women with regard to the influence of apo E phenotype.

2. Subjects and methods

The study group, enrolled consecutively in al- phabetical order comprised 342 subjects, 40 and 70 years of age, from 5200 subjects taking part in an annual health screening survey in the City of Tampere (176 000 inhabitants). All invited subjects (n = 342) took part in this substudy. There were four subgroups: 85 men and 80 women 40 years of age, 88 men and 89 women 70 years of age. The age of 40 years was chosen for the premenopausal group of women (all were premenopausal) and 70 years for the postmenopausal group of women. The male groups were the same age as the female groups. Subjects treated with lipid lowering drugs were excluded.

Histories of hypertension (treated hypertension or untreated stibjects with diastolic blood pressure > 100 mmHg in two measurements), diabetes, CAD, smoking, medication and family history of cardiovascular disease were recorded by question- naire. Height and weight were measured and the body mass index (BMI) calculated. Demographic findings of the subjects are given in Table 1.

Blood samples were taken in tubes containing EDTA anticoagulant after 12 h fast and 15 min rest. Serum total cholesterol and triglycerides were measured by enzymatic methods (Nycotest, Nycomed AS, Oslo, Norway) using the Monarch 2000 (Monarch, Lexington, IL) analyzer. Seronorm lipid was used (Nycomed AS) as standard. HDL cholesterol concentrations were deter- mined, as was total cholesterol, by assay of supernatants remaining after precipitation of LDL and VLDL with polyethylene glycol reagent [ 131. Non-denaturing gradient gel electrophoresis, which separates LDL particles by size, was used to identify distinct subpopulations of LDL particles [14]. For determination of LDL size, electrophoresis of whole plasma was carried out using ready made 2%16% gradient gels (Pharmacia, Uppsala, Sweden). A mixture of globular proteins of known diameter was run on each gel to cali-

M. Nikkilii et al. I Atherosclerosis 119 (1996) 181-190

Table 1 Demographic findings and frequencies of selected diseases in different age groups (n = 342)

183

40-year-old 70-year-old

BMI” (kg/m2) Hypertension (%) Use of betablockers or diuretics (%) Smoking (%) Stroke (w,) Diabetes mellitus (%) CAD (%) Family history of CVD (%) Alcohol consumption” (drinks/week) Diastolic blood pressure” (mmHg)

Men (n = 85)

25.7 50.4 1.2 1.0

36.5 0.0 0.0 0.0

36.5 8.2 0.7 &

83.5 kO.8

Women (n = 80) Men (n = 88) Women (n = 89)

24.5 f 0.6 26.6 F 0.4 26.0 & 0.5

2.5 29.6 31.8 0.0 37.5 45.5

37.5 13.6 6.7 0.0 3.4 1.1 2.5 8.0 10.2 0.0 20.5 13.6

40.0 40.0 50.6 3.4 * 0.4 5.8 & 0.9 0.9 f 0.2

79.8 k 0.9 88.0 + 0.9 87.5 f I.0

BMI, body mass index; CAD, coronary artery disease; CVD cardiovascular disease. Sex differences were not significant when 40-year-old men and 70-year-old men were compared separately with the women of the same age. “Values are mean f S.E.M.

brate for particle size. The standards included thyroglobulin dimer (23.6 nm), thyroglobulin (17 nm) and ferritin (12.2 nm). After a 24 h electrophoresis at 125 V the gels were stained with Oil Red 0 at 60°C for 24 h and scanned with a laser densitometer (LKB, Ultroscan 2202) connected to an integrator (Hewlett-Packard, 3390A). Some subjects had several bands, but all were classified by the most intensively stained subfraction. A control plasma sample stored at - 70°C was included in every run. The inter-assay coefficient of variation was < 5%.

Apo E phenotyping was performed from delipidated plasma using isoelectric focusing (IEF), cysteamine treatment and immunoblotting, as previously described [15]. In brief, 10 ~1 plasma was delipidated with ethanol diethyl ether 3:l (v/v) for 24 h at - 20°C. The final precipitate was dissolved in 200 ~1 0.1 mmol/l Tris-HCL, containing 1% sodium decyl sulfate and 1 ~1 /3- merkaptoethanol. Removal of sialic acid was achieved by incubating 10 ~1 plasma with 20 pl(2 mU/pl) neuramidase for 1 h at 37°C and then samples were delipidated and dissolved as /3- merkaptoethanol-treated samples. IEF was carried out for 3.5 h at 750 V on 5% polyacrylamide gels. After IE:F the proteins were transferred to nitrocellulose membranes using an LKB 2117- 2500 electroblotting unit. After transfer the nitro-

cellulose membranes were double antibody stained in 1% gelatin-Tris-buffered saline (TBS) buffer, by using as a first antibody polyclonal rabbit anti-human apo E antibody (1: 1000) and as a second antibody IgG- horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody in dilution 1:2000. ApoE isoforms were then visual- ized in a substrate solution containing 4-chloro-l- naphtol reagent in a cold methanol, TBS and hydrogen peroxide. The recognition of apo E phenotypes in the blots was based on comparison with IEF patterns from known apo E phenotypes. The rare phenotypes (E2/2, E2/3 and E4/4) were verified by another focusing using parallel /?- merkaptoethanol-, cysteamine- and neuramidase- treated samples. The reliability of the method was confirmed by phenotyping 20 samples at the Na- tional Public Health Institute in Helsinki.

The major factors examined in the study were age, sex, and apo E. We also studied the effect of the betablockers. The possible confounders were smoking, BMI, alcohol consumption and diastolic blood pressure. These were taken as covariates into the models. Statistical analysis was based on the multi-way analysis of covariance, where the confounding factors were taken into the model as covariates and their effect adjusted from the statistical quantities, i.e. the means and standard deviations displayed were adjusted with respect to

184 M. Nikkilii et al. 1 Atherosclerosis 119 (1996) 181-190

Table 2 The concentrations of total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides, the ratio of HDL to total cholesterol and LDL size (mean + S.E.M.) adjusted with respect to body mass index, smoking, diastolic blood pressure and alcohol consumption in subjects (n = 342) participating in a health screening program


Men (n = 85) Women (n = 80) Men (n = 88) Women (n = 89)

Total cholesterol (mmol/l) 5.54*0.11 5.02 * 0. I 1 6.13kO.11 6.63 + 0.11 LDL cholesterol (mmol/l) 3.63 If: 0. I4 2.86kO.13 4.22 k 0.14 4.46 + 0.21 HDL cholesterol (mmol/l) I.19 +0.04 I .69 f 0.04 I .24 + 0.04 1.49 * 0.04 Triglycerides (mmol/l) 1.42 f 0.08 1.03 k 0.08 I .40 + 0.07 I .42 + 0.08 HDL/Total cholesterol (%) 22.5 + 0.91 34.7 * 0.91 20.7 + 0.90 22.9 + 0.94 LDL size (nm) 25.87 k 0.07 26.40 i 0.07 26.03 i 0.07 26.38 k 0.07

Values are mean + S.E.M. Interaction between age and sex was significant for all the concentrations of lipids except LDL size. LDL size was significant only in relation to sex.

the covariates when adequate. When selecting variables in multiple regression analysis, we ap- plied the best subsets regression utilizing Mallows’ C,, criterion. Here we fitted all the possible models to the data and then using the Mallows’ C, we selected the best model. This criterion takes into account both the fit of the model and the number independent variables used. We also calculated the correlation coefficients. The cross-tabulations were analyzed using the chi-square statistic and here also an odds ratio was used with 95% confidence intervals. The computation was carried out on a VAX/VMS computer using BMDP Statisti- cal Software (1993 version).

Informed consent was obtained from all subjects after the nature and purpose of the study had been explained. The study protocol was ap- proved by the Ethics Committee of the Commu- nity Health Centre of the City of Tampere.

3. Results

We used analysis of covariance (ANCOVA) to analyze the lipid concentrations. In Table 2 adjusted means and S.E.M. are presented. The inter- pretation is based on this table and on the respective ANCOVA analysis, the results of which are not presented.

There was a significant interaction between sex and age in serum total cholesterol, which was highest in older women with a level of 6.63 + 0.11 mmol/l (mean f S.E.M.) when compared

with other subgroups (Table 2). Similar results can be seen in LDL cholesterol values. A significant interaction between sex and age was also found in serum HDL cholesterol, which was highest in 40-year-old women (1.69 ) 0.04 mmol/l), decreasing with age to 1.49 f 0.04 mmol/l in 70-year-old women.

The HDL cholesterol value was lowest in 40- year-old men (1.19 + 0.04 mmol/l), with a mini- mal age effect. In contrast, in women serum HDL cholesterol decreased significantly with age. Fur- thermore, the means of HDL cholesterol were higher in women than in men of both age groups. The significant interaction between age and sex in the analysis of triglycerides was due to the increase with age in women, while the means of triglycerides in men remained the same in both age groups, at corresponding levels found in the older women. The interaction between sex and age was significant also for the ratio of HDL cholesterol to total cholesterol, although the means decreased with age in both sexes. Espe- cially in younger women the mean of the ratio of HDL cholesterol to total cholesterol was very high (34.7 f 0.91%) compared with men, and also the decrease with age in women was more marked than in men.

Serum total cholesterol was 2 6.5 mmol/l in 52.8% of the 70-year-old females, but in only 6.3% of the 40-year-old females (P < 0.001). Serum HDL cholesterol was < 1.0 mmol/l in 5.0% of the 40-year old women, but in 16.5% of

M. Nikkilii et al. I Atherosclerosis 119 (1996) 181-190

Table 3 The distribution (%I) of serum lipid values, LDL size and allele ~4 of apo E phenotype in different age groups (n = 342)

185


Men (n = 85) Women (n = 80) Men (n = 88) Women 01 = 89)

Total cholesterol

2 6.5 mmol:l 16.5 6.3 35.2 52.8** LDL cholestercl > 4.5 mmol/l 16.5 6.3 35.2 50.6* HDL cholesterol < 1.0 mmol/l 16.5 5.0* 21.6 7.9* Triglycerides 2 2.0 mmolil 18.8 2.5*’ 18.2 22.5 HDL/total cholesterol < 20% 47.1 10.8*** 53.4 41.6 LDL size < 25.50 nm 27.1 10.0** 19.3 9.0

Apo E phenotype allele c4 19.4 16.7 18.8 17.0

*P < 0.05. **P < 0.01, ***P < 0.001 when 40-year-old subjects and 70-year-old subjects were compared separately.

the 40-year-old men (P < 0.05) (Table 3). The corresponding sex difference was similar in the 70-year-old groups. The proportion of the lowest values ( < 20.0%) of the ratio of HDL cholesterol to total cholesterol was smallest in 40-year-old women, the difference being significant (P < 0.001) compared with other groups.

LDL size was greater (26.39 + 0.07 nm) in women than in men (25.95 + 0.07 nm) (P < 0.01). As to the diameter of the LDL particle, the only significant factors were gender and apo E phenotype. The values were largest (26.40 f 0.07 nm) in 40-year-old women and almost as large in 70-year-old women (26.38 &- 0.07 nm) compared with middle-aged men (25.87 + 0.07 nm) and older men (26.03 + 0.07 nm).

The small dense LDL particle ( < 25.5 nm) was found more frequently in 40-year-old-men (27.1%) than in other groups (Table 3). Apo E phenotype had an influence on LDL size, as shown in Table 4. LDL size was lowest in subjects with apo E phenotype E4/4 and increased in the order E3/4 < E3/3 < E2/3. Apo E phenotype was not significantly associated with serum triglycerides, but it had an influence on serum total, LDL and HDL cholesterol (Table 4). In 70-year-old subjects CAD was present in 20.5% of men and 13.6% of women. The proportion of large LDL

particles ( > 25.5 nm) was very small (3.0%) in 70-year-old subjects with CAD compared with subjects without (24.0%).

The distribution of cardiovascular diseases and the use of betablockers and diuretics in the groups divided according to LDL size are given in Table 5. Here we also calculated odds ratios, but according to their confidence intervals no significant results were found. We would like to emphasize that because of the low frequencies the power of the analysis is not satisfactory. There were no significant differences between groups, but the trend was obvious in the case of CAD and family history of cardiovascular disease among 70-year- old men. The distribution of different apo E phenotypes was very similar in men and women (Table 6). The frequency of allele c4 was highest in 40-year-old men, but the differences were not significant.

Correlation coefficients between LDL size and some demographic and laboratory findings were calculated within the whole study population and in 70-year-olds alone. LDL size correlated with BMI positively and with alcohol consumption negatively, but not in 70-year-olds alone. There was a very poor correlation between total cholesterol and LDL size, but a strong positive correlation (v = 0.606; P < 0.001) between LDL size

186

Table 4

M. Nikkilii et al. / Atherosclerosis 119 (1996) 181-190

The size of LDL, the concentrations of total cholesterol, LDL cholesterol, HDL cholesterol and triglycerides (mean + S.E.M.) in different apo E phenotype groups adjusted with respect to body mass index, smoking, diastolic blood pressure and alcohol consumption in the whole study population

Frequencies LDL size (nm) Total chol (mmol/l) LDL chol (mmol/l) HDL chol (mmol/l) Triglycerides (mmol/l)

Apo E2/3 (n = 24) 26.35 +0.13 5.60 + 0.24 3.29 f 0.20 I .50 + 0.04 1.82 + 0.21 Apo E3/3 (n = 198) 26.19 +0.04 5.77 + 0.08 3.29 + 0.20 1.40 * 0.04 I .40 + 0.07 Apo E3/4 (n = 89) 26.02 + 0.06 5.93 * 0.12 3.87 i 0.11 1.35 & 0.04 1.24 i 0.05 Apo E4/4 (n = 10) 25.70 k 0.19 6.53 & 0.35 4.36 k 0.23 1.28 + 0.04 1.28 kO.14

Phenotype apo E 2/4 was excluded because the group size was small and allele ~2 and ~4 have opposite effects on serum lipids (n = 311) The differencies were significant between groups in LDL size, total, LDL and HDL cholesterol.

and HDL cholesterol, and a strong negative correlation (r= - 0.627; P < 0.001) between LDL size and triglycerides. Diastolic blood pressure in 70-year-old men correlated negatively with LDL size (r = - 0.309; P < O.Ol), but there was only a weak negative correlation in the whole study population. CAD, diabetes, stroke, the use of betablockers or diuretics, smoking and family history of cardiovascular diseases did not influence LDL size in the cross-tabulations. LDL size was analyzed using analysis of covariance where we regarded betablockers, sex, age and apo E as factors and BMI, alcohol, smoking, diastolic blood pressure as covariates. Because of the low frequencies we were forced to make the analysis using a two factors structure successively. The power of this analysis is questionable, but the only trend we found was the very slight interaction between betablockers and sex (P = 0.13), which indicated that among females (especially those without betablockers) LDL size was higher.

There was a trend for LDL size to become smaller in groups drinking more alcohol: the higher the alcohol consumption, the smaller was LDL size, the differences not, however, being significant. Teetollers and subjects using below 40 g alcohol per week had significantly larger LDL size than groups using over 40 g alcohol per day, but the differences were not significant when adjusted with sex and age. When applying the best subset regression analysis to the prediction of the variability of LDL size separately in women and men, the best models were somewhat different. In women 65% of the variability in LDL size could be accounted for by alterations in serum triglyce-

ride, HDL cholesterol and glucose concentrations. In men the best model with significant contribu- tion included serum triglyceride and HDL concentrations, diastolic and systolic blood pressure and alcohol consumption. Together these variables explained 47% of the LDL size variability.

4. Discussion

In this study total cholesterol values increased with age in both sexes, but more markedly in women. The same trend has been shown in earlier studies [16,17], but without correction for confounding factors. Such results are not usually taken into consideration in clinical practice.

Some epidemiologieal studies have shown that the age-related rise in serum total cholesterol in women is not accompanied by an increase in CAD risk [16]. CAD is many times more frequent in men than in women. This sex difference is most remarkable in Finland, where middle-aged men die ten times more frequently of CAD than middle-aged women, according to official statistics on causes of deaths. In men elevated total cholesterol correlates highly with CAD mortality [3], but not in women [18]. In some studies elevated total cholesterol appeared not to be a risk factor for elderly women [l]. According to some authors, the increase in serum total cholesterol of women might be an adaptive process during ageing, nec- essary to maintain the physical or chemical characteristics of the cell membrane [19].

Total cholesterol levels rise after the menopause, while HDL cholesterol declines slightly [20]. Also, in the present study total

Table 5

hf. Nikkilij et al. I Aiherosclerosis 119 (1996) ISI- 190 187

The distribution (‘Xb) of selected diseases in 70-year-old subjects (II = 187) divided into two groups according to LDL size, when cut-off-point is 25.5 nm

LDL size 5 25.5 nm LDL size > 25.5 nm

Men (n = 17) Women (n = 8) Men (n = 71) Women (II = 8 I)

CAD 23.5 25.0 19.7 12.4 Stroke 0.0 0.0 4.2 1.2

Hypertension 35.3 25.0 28.2 32. I Betablockers!diuretics 35.3 62.5 38.0 43.8 Diabetes mellitus 17.6 25.0 5.6 8.6 Family history of CVD I I.8 0.0 8.5 9.9

CAD. coronary artery disease: CVD, cardiovascular disease. The ditferences were not statistically significant,

cholesterol was very high in postmenopausal women (70-year-old) compared with other groups, whereas HDL cholesterol was lower in other groups than in premenopausal women (40-year- old). As girls pass through puberty, HDL cholesterol levels do not fall as they do in boys. The relationship between circulating sex hormone levels and the Ioccurrence of CAD was studied in a group of men undergoing coronary angiography. No significant differences were found for serum concentrations of estradiol or testosterone between patients with CAD and controls [21]. The importance of HDL cholesterol concentration has been shown in several large epidemiological studies [22-251. These surveys indicated that a low concentration of HDL cholesterol is a common antecedent to clinical CAD.

In the Framingham study the prevalence of small dense LDL particles was significantly asso-

Table 6 The percentage of ditferent apo E phenotypes in the whole study population (n = 332) and separately in men (n = 163) and women (n := 169)

All Men Women

Apo E2,2 Apo E2;3 Apo E3,‘3 Apo E2,4 Apo E3,4 Apo E4:4

0.0 0.0 0.0 7.3 6.4 7.9

59.8 58.6 60.7 3.0 I.2 5.0

26.9 30. I 23.9 3.0 3.7 2.5

100.0 100.0 100.0

ciated with increased triglyceride levels [5]. There is evidence that triclycerides constitute an independent predictive risk factor for CAD in individuals with low serum concentrations of HDL cholesterol [2,26]. In our study serum triglycerides appeared to be an important determinant of LDL size and HDL metabolism, just as described in earlier studies [27].

In this and earlier studies the presence of small dense LDL is associated with a lipid profile that includes higher levels of plasma triglycerides and lower levels of HDL cholesterol. For this reason small dense LDL has been designated the atherogenic lipoprotein phenotype. The syndrome of insulin resistance (syndrome X) has features of a very similar dyslipidemia [28]. This and the association of small dense LDL with diabetes suggest that LDL particle diameter may be related to hyperinsulinemia and insulin resistance. Central obesity dyslipidemia, hypertension, glucose intolerance and hyperinsulinemia have been found to add independently to the likelihood of having small dense LDL. A recent paper has demonstrated that diet could induce LDL size relating to ,apo E phenotype [29]. These results indicated ‘that reduced dietary fat lowers levels of large, buoy- ant LDL particles by an ,apo E-dependent mech- anism. In our study normal diet with low fat vs’. high fat did not influence LDL size.

In our study small and” dense LDL with a cut-off point of 25.5 nm was most common (27.1%) in 40-year-old men and most infrequent

188 M. Nikkilii et al. / Atherosclerosis 119 (1996) 181-190

(9.1%) in 70-year-old women. There was no difference in LDL size between the two female age groups, but LDL size was significantly larger in women than in men in both age groups. In an earlier study menopausal status had an influence on LDL size: 13% of premenopausal and 49% of postmenopausal women had small dense LDL [30]. We could not find any difference between premenopausal and postmenopausal women in LDL size. There was no CAD in 40-year-old subjects, but this condition was a common finding in 70-year-old subjects (20.5% in men and 13.6% in women). LDL particle size was negatively associated with elevated triglyceride levels and positively with decreased HDL cholesterol level. The proportion of large LDL particles was lower in CAD patients (3%) than in controls (24%). The concentrations of total cholesterol and LDL cholesterol correlated only weakly to the size of LDL. Some authors have shown that betablockers could influence lipid profile and lead to a serious underestimation of the prevalence of LDL abnormalities and an overestimation of HDL abnormalities in patients with CAD [31]. Betablock- ers have been previously described as having a favorable effect of LDL composition and oxidizability [32]. In this study we found only a very slight interaction between betablockers and sex. LDL size was higher among females not using betablockers than in those using them. Because of small numbers of subgroups it was not possible to analyze with all confounders at the same time.

Earlier studies have shown that the apo E phenotype is not an independent risk factor for CAD. The apo E polymorphism influences lipoprotein levels and, possibly, thereby indirectly increases the risk of CAD [33]. Previous studies have sug- gested that apo E polymorphism may influence the development of atherosclerosis through its role in the catabolism of apolipoprotein B containing lipoproteins. Apo E is found in chylomi- crons, VLDL remnants and HDL particles [9]. Allele ~4 has been shown in one recent paper to be a significant risk factor after adjustment for traditional coronary risk factors and lipids [34].

The distribution of apo E phenotypes found in this study is comparable to that found in the general Caucasian population. The homozygous

E3 phenotype is the most common in the general population. There was a significant decrease for LDL size and a significant increase for LDL cholesterol concentration from apo E2/3 to apo E4/4 (Table 4). It seems that especially in apo E phenotype 4/4, serum cholesterol level is enhanced and LDL size reduced. In earlier studies there are controversial results on LDL size among the different apo E phenotypes: in the Framingham Study apo E phenotype was associated with LDL size [35], but in another study it was not [36]. One previous work demonstrates an association between apo E alleles and triglycerides [34]. It is quite possible in the present study that the small numbers of subgroups divided according to apo E phenotype do not have enough statistical power to prove this point, but the trend is similar.

In conclusion, women have larger and less atherogenic LDL than men regardless of menopausal status. Elderly women have the highest cholesterol level combined with large LDL size. Small LDL size correlates very highly with decreased serum HDL cholesterol and increased serum triglycerides, which are increased risk factors for premature CAD. In health screening studies serum HDL cholesterol and triglycerides should be measured, which indirectly give information about LDL size and its atherogenicity. The group of phenotype apo E4/4 showed the highest total cholesterol, the lowest HDL cholesterol and the smallest LDL size, which are all atherogenic characteristics. LDL size decreased and total cholesterol increased according to the most prevalent apo E phenotypes in the order E2/3, E3/3,E3/4 and E4/4.

Acknowledgments

This work was supported by the Yrjii Jahnson Foundation, the Medical Research Fund of Tam- pere University Hospital. During the study Terho Lehtimaki received a fellowship from the Emil Aaltonen Foundation.

References

[I] Forette B, Tortrat D, Wolmark Y: Cholesterol as risk factor for mortality in elderly women. Lancet 1989;i:868.

M. Nikkilii er al. i Atherosclerosis 119 (I 996) 181~ 190 I89

[2] Nikkila M, Pitkajarvi T, Koivula T, Lehtomaki E, Jokela H. Age-related changes and sex differences in serum lipid concentrations of health-screening subjects. Cardiovasc Risk Factors, 1991;1:204.

[3] Multiple Risk Factor Intervention Trial Research Group. Multiple Risk Factor Intervention Trial: risk factor changes and mortality results. J Am Med Assoc 1982;248: 1464.

[4] Romm PA, Green CE, Reagan K, Rackley CE. Relation of serum lipoprotein cholesterol levels to presence and severity of angiographic coronary artery disease. Am J Cardiol 1991;67:479.

[5] 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;34:1687.

[6] Campos H, Genest JJ, Blijlevens E. Low density lipoprotein particle size and coronary artery disease. Arte- riosclerosis, Thromb 1992; 12: 187.

[7] Austin MA, King MC, Vranizian MA, Krauss RM. Atherogenic lipoprotein phenotype. A proposed genetic marker fclr coronary heart disease risk. Circulation 1990;82:495.

(81 Kwiterovic:h PO, Coresh J, Bachorik PS. The prevalence of hyperapobetalipoproteinemia and other lipoprotein phenotypes in premature coronary atherosclerosis. Am J Cardiol 1993;71:631.

[9] Mahley RW, lnnerarity TL. Lipoprotein receptors and cholesterol homeostasis. Biochem Biophys Acta, 1983;737:1’?7.

[IO] Ehnholm C, Lukka M, Kuusi T, Nikkila E, Utermann G. Apolipoprotein E polymorphism in the Finnish population: gene frequencies and relation to lipoprotein concentrations. J Lipid Res 1986;27:227.

[Ill Uterman G. Apolipoprotein E polymorphism in health and disease. Am Heart J 1987;I 134:433.

[I21 Davignon 1, Gregg RE, Sing CF. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis 1988;8:1.

[13] Lippi U, Graziani MS, Manzato F. Assay of cholesterol in high density lipoprotein in hypertriglyceridemic sera. Clin Chem 1986;32:2118.

[14] Nichols AV, Krauss RM, Musliner TA. Non-denaturing polyacrylamide gradient gel electrophoresis. Methods En- zymol 1986;128:417.

[I51 Lehtimlki T, Moilanen T, Viikari J. Akerblom H, Ehn- holm C, Rijnnemaa T, Marniemim J, Dahlen G, Nikkari T. Apolipoprotein E phenotypes in Finnish youths: a cross-sectional and 6-year follow-up study. J Lipid Res 1990;3 I :48’7.

[l6] Heiss G, Tamir I, Davis C, Tyroler HA, Rifkind BM Schonfeld G, Jacobs D, Frantz ID Jr. Lipoprotein- cholesterol distributions in selected North American pop- ulations: The Lipid Reserch Clinic’s Program Prevalence Study. Circulation 1980;61:302~315.

[17] Mann Jl, Lewis B, Shepherd J, Winder AF, Fenster S Rose L, Morgan B. Blood lipid concentrations and other

cardiovascular risk factors: distribution, prevalence and detection in Britain. Br Med J 1988;296:1702.

[l8] Simons LA. Interrelation of lipids and lipoproteins with coronary artery disease in 19 countries. Am J Cardiol 1986;57:5G.

[I91 Oliver MF. Might treatment of hypercholesterolaemia increase non-cardiac mortality? Lancet, 1991;337: 1529.

[20] La Rosa JC. Lipids and cardiovascular disease: do the findings and therapy apply equally to men and women? Women’s Health Issues, 1992;2:102.

[2l] Hauner H, Stangl K, Burger K, Busch U, Blumer H Pfeiffer EF. Sex hormone concentrations in men with angiographically assessed coronary artery disease ~ relationship to obesity and body fat distribution. Klin Wochenschr, 1991;69:664.

[22] Miller NE, Thelle DS. F&de OH, Mjos OD. The Tromsii Heart Study. High density lipoprotein and coronary heart disease: a case control study. Lancet 1977;i:965.

[23] Manninen V, Tenkanen L, Koskinen P. Huttunen J MHnttHri M, Heinonen 0, Frick H. Joint effects of serum triglyceride and LDL cholesterol and HDL cholesterol concentrations on coronary heart risk in the Helsinki Heart Study. Circulation 1992;85:37.

[24] Gordon DJ, Probstfield JL, Garrison RJ, Neaton JB, Castelli WP, Knoke JD, Jacobs DR, Bangdiwala S. Ty- roler AH. High density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation 1989;79:8.

[25] Assmann G, Schulte H, Funke H, von Eckardstein A Seedorf U. The prospective cardiovascular Munster (PROCAM) study: identification of high-risk individuals for myocardial infarction and the role of high-density lipoprotein. In: Miller NE. (ed). High Density Lipo- proteins and Atherosclerosis II. Amsterdam: Excerpta Medica, ICS 826, 1989;51- 65.

[26] Castelli WP. Garrison RJ, Wilson PWF, Abbot ED Kalousdian S, Kannel WB. Incidence of coronary heart disease and lipoproteins levels. The Framingham Study. J Am Med Assoc 1986;256:2835.

[27] McNamara JR, Campos H, Ordovas J, Peterson J. Wilson PWF, Schaefer EJ. Effect of gender, age, and lipid status on low density lipoprotein subfraction distribution. Results from the Framingham Offspring Study. Arteriosclerosis 1987;7:483.

[28] Selby JV, Austin MA, Newman B, Zhang D, Quesen- berry CP Jr. Mayer EJ, Krauss RM: LDL subclass phenotypes and the insulin resistance syndrome in women. Circulation 1993;88:381.

[29] Dreon DM, Fernstrom HA, Miller B, Krauss RM. Apolipoprotein E isoform phenotype and LDL subclass response to a reduced-fat diet. Arteriosclerosis Thromb Vast Biol 1995;15:105.

[30] Austin MA, King M-C, Vranizan KM, Newman B. Krauss RM. Inheritance of low density lipoprotein subclass patterns: results of complex segregation analysis. Am 3 Hum Genet 1988;43:838.

190 M. Nikkih? et al. 1 Atherosclerosis 119 (1996) 181-190

[31] Genest J Jr, McNamara JR, Ordovas JM, Jenner JL Silberman SR, Andersson KM, Wilson PWF, Salem DN, Schaefer EJ. Lipoprotein cholesterol, apolipoprotein Al and B lipoprotein(a) abnormalities in men with premature coronary artery disease. J Am Co11 Cardiol 1992;19:792.

[32] Croft KD, Dimmit SB, Moulton C, Beilin LJ. Low density lipoprotein composition and oxidizability in coronary disease - apparent favourable effect of beta block- ers. Atherosclerosis 1992;97:123.

[34] Wilson PWF, Myers RH, Larson MG, Ordovas JM, Wolf PA, Schaefer EJ. Apolipoprotein E alleles, dyslipidemia and coronary heart disease. J Am Med Assoc 1994;272: 1666.

[35] Schaefer EJ, Lamon-Fava S, Johnson S, Ordovas JM Schaefer MM, Castelh WP, Wilson PWF. Effects of gender and menopausal status on the association of apolipoprotein E phenotype with plasma lipoprotein levels. Results from the Framingham Offspring Study. Arte- riosclerosis Thromb 1994;14:1105.

[33] Stuyt PM, Brennikmeier BJ, Demacker PN. Apolipo- [36] Zhao SP, Verhoeven J, Vink J, Hollaar L, van der Laarse protein E phenotypes, serum lipoproteins and apolipo- A, de Knijff P, van Hooft FM. Relationship between proteins in angiographically assessed coronary heart apolipoprotein E and low density lipoprotein particle size. disease. Stand J Clin Lab 1991;51:425. Atherosclerosis 1993;102:147.

Documents

Women have a larger and less atherogenic low density lipoprotein particle size than men