Interactions Between Lifestyle-Related Factors and the ApoE Polymorphism on Plasma Lipids and Apolipoproteins
The EARS Study
Abstract To elucidate how the apolipoprotein (apo) E polymorphism and modifiable factors interact in explaining plasma lipid and apolipoprotein levels, we studied 1448 young adults (18 to 26 years old), participating in the European Atherosclerosis Research Study (EARS). Venous blood was collected after an overnight fast. Modifiable factors, eg, body mass index (BMI), waist-to-hip ratio (WHR), tobacco and alcohol consumption, and physical activity, were determined by using standardized protocols. Associations of modifiable factors with apoE levels were homogeneous across apoE phenotypes. In contrast, correlations of BMI with total cholesterol and apoB levels, as well as correlations between WHR and apoB, were significantly (P<.05 to P<.01) stronger in E2 carriers than in subjects with other phenotypes. Total cholesterol and apoB levels were comparable in E2 carriers in the upper tertile of BMI or WHR to those in E3/3 subjects, suggesting that the lowering effect of the E2 allele was no longer present. The inverse association between the plasma cholesteryl linoleate-to-oleate ratio, a marker for the dietary polyunsaturated-to-saturated fatty acid ratio, and triglycerides was also stronger in E2 carriers (−0.33 versus −0.17 in E3/3 and −0.24 in E4 carriers). Associations with other modifiable factors were notably consistent across apoE phenotypes. Gender and modifiable factors explained three times more (31%) of the interindividual variation in apoB levels in E2 carriers than in E3/3 subjects (9%) or E4 carriers (14%), mainly due to a larger variance explained by BMI. Our results suggest that the apoE polymorphism acts in a relatively uniform manner, independently of lifestyle. However, the associations of adiposity to total cholesterol and apoB levels appear to be stronger in apoE2 carriers.
- Received February 8, 1996.
- Accepted October 12, 1996.
Apolipoprotein E is a structural component of VLDL and HDL. The protein is polymorphic, with three common isoforms found in the general population, coded for by three codominant alleles (ε2, ε3, ε4). Numerous studies have shown that the apoE polymorphism is associated with plasma total and LDL cholesterol, as well as with plasma apoE and apoB concentrations. In the population at large, the ε2 allele is associated with lowered levels of apoB, total cholesterol, and LDL cholesterol, while the opposite is true for the ε4 allele (see References 1 and 21 2 ). A meta-analysis combining the data of 45 population samples clearly demonstrated that the ε2 and ε4 alleles are both associated with elevated concentrations of triglycerides.3
ApoE allele frequencies vary widely across populations around the world,4 5 and even across relatively close populations such as the European.5 6 7 Notwithstanding these variations, the allele effects on lipid levels are remarkably consistent across populations.4 6 7 This consistency suggests that the apoE polymorphism acts in a relatively uniform manner, despite differences in genetic background and environment. However, this does not preclude a more subtle modulation of apoE effects by modifiable factors, which would not be well accounted for by interpopulation comparisons. Such a modulation is suggested, in particular, by intervention studies showing that the response of plasma lipids to dietary change is not uniform across apoE phenotypes.8 9 10 11 12
Very few studies have addressed the issue of the interaction between modifiable factors and apoE phenotype effects on lipids. One of the reasons might be that very large sample sizes are required for having an acceptable power of detecting such interactions. It has been suggested that apoE genotypes might modify the relationship of measures of obesity and fat distribution,13 14 smoking and alcohol consumption,15 and physical activity16 to lipids.
EARS is a large multicenter study of biological, lifestyle, and genetic risk factors for coronary heart disease, carried out in young adults from 11 countries throughout Europe. In an earlier paper,7 we described the association of the apoE polymorphism with lipids and apolipoproteins. Associations with plasma total and LDL cholesterol, triglyceride, apoB, and apoE levels were consistent with the now well-identified effects of ε2 and ε4 alleles on these traits. These effects exhibited a great consistency among the different European populations, although there was a clear-cut North-to-South opposite gradient in the ε2 and ε4 allele frequencies. The large number of subjects participating in EARS allowed us to further elucidate whether the apoE locus interacts with environmental factors.
Therefore, the aim of the present study was to investigate whether the effects of modifiable factors, eg, obesity, fat distribution, dietary fat composition, smoking, alcohol consumption, and physical activity on plasma total cholesterol, triglyceride, apoB, and apoE levels were modulated by the apoE polymorphism.
A detailed description of EARS is given elsewhere.17 Briefly, 1994 male and female students, aged between 18 and 26 years, from 14 university populations in 11 European countries have been studied. Students whose father had a verified myocardial infarction before the age of 55 were recruited and represent cases (n=682). Two age- and sex-matched control subjects were recruited by computer selection from the same university population (n=1312). Students were grouped into five regions on the basis of geography, language, and age-standardized mortality rates18 : Finland (Oulu and Helsinki), Great Britain (Glasgow and Bristol), Northern Europe (Göteborg, Aarhus, and Hamburg), Middle Europe (Ghent, Innsbruck, and Zurich), and Southern Europe (Bordeaux, Barcelona, Reus, and Naples).
Venous blood was collected after an overnight fast. Height, weight, and waist and hip circumferences were measured, and BMI (weight in kilograms divided by height in meters squared) and WHR were calculated. Details of lifestyle, eg, smoking habits, alcohol consumption, and physical activity were determined with standardized questionnaires and protocols.17
Cholesterol and triglyceride concentrations were measured according to the Lipid Research Clinic’s Manual of Laboratory Operations, standardized according to the Centers for Disease Control and Prevention, Atlanta, Ga. ApoB levels were measured by immunonephelometry on a Behring BNA nephelometer. ApoE levels were measured by ELISA according to published procedures.19 ApoE phenotyping was performed by isoelectric focusing of delipidated plasma followed by immunoblotting.20 21
The composition of cholesteryl esters in plasma was determined by reversed-phase high-performance liquid chromatography as described previously.22 Four major components were determined: cholesteryl palmitate (16:0), oleate (18:1), linoleate (18:2), and arachidonate (20:4). The L/O ratio was calculated as a marker for the P/S ratio.23
Only subjects for whom all lipid and modifiable factors and the apoE phenotype were available (n=1795) were included in statistical analyses. Additionally, women taking oral contraceptives (n=321) were excluded because of the large effect on lipid parameters studied. Since very few subjects had the E2/2 (n=12) or E4/4 phenotype (n=31), regrouping of the subjects into three groups was performed: carriers of the ε2 allele (E2/2 and E3/2 phenotypes), subjects with the E3/3 phenotype, and E4 carriers (E4/3 and E4/4 phenotypes). Subjects with the E4/2 phenotype (n=26) could not be assigned to any of the groups and were therefore excluded, leaving 1448 subjects. All analyses were carried out using the Statistical Analysis System (SAS, version 6.09, SAS Institute, Cary, NC).
Although in EARS a large number of lipids and apolipoproteins were measured, we decided to focus only on those traits for which there was no controversy about the influence of the apoE polymorphism, to limit the possibility of finding spurious interactions. Given the strong correlation between LDL cholesterol and apoB levels, LDL cholesterol was omitted because it was not directly measured but assessed by Friedewald’s formula. Since in our earlier paper7 apoE allele effects had been shown to be very homogeneous across regions and among cases and controls, we analyzed pooled data with adjustment for region and case/control status. An additional adjustment was performed for age and, depending on the analysis, gender.
Triglycerides and apoE levels were log transformed to improve normality for statistical testing.
Phenotype-specific associations of continuous modifiable factors with lipid and apolipoprotein levels were determined by partial Pearson correlation coefficients. For physical activity, Spearman’s correlation coefficients were determined. The homogeneity of associations of modifiable variables with lipid and apolipoprotein levels across apoE phenotypes was tested by analysis of variance, including E2 and E4/lifestyle interaction terms in the model. The E3/3 phenotype was taken as the reference category.
Finally, multivariate regression analysis was conducted in each apoE phenotype group, with lipid and apolipoprotein levels successively taken as the dependent variable, and modifiable factors and gender as independent variables. In each apoE phenotype group, the proportion of variance (R2) attributable to gender and all modifiable factors combined was calculated as the ratio of the sum of squares due to these factors to the age-, region-, and case/control status–adjusted total sum of squares.
Associations of the apoE polymorphism with total cholesterol, triglyceride, apoE, and apoB levels were as expected (Table 1⇓). The apoE polymorphism was not associated with any of the modifiable factors studied, eg, indices for obesity and fat distribution, tobacco and alcohol consumption, physical activity, and the L/O ratio, used as a marker of the P/S ratio of the diet (Table 1⇓). Ranges in modifiable factors were slightly smaller in E2 carriers than in E3/3 subjects or E4 carriers (data not shown).
Correlations of plasma lipid and apolipoprotein levels with modifiable factors according to apoE phenotype are shown in Table 2⇓. The apoE polymorphism did not alter correlations between modifiable factors and apoE levels. In contrast, correlations of BMI and WHR with total cholesterol and apoB levels were stronger in subjects with the ε2 allele than in subjects with the E3/3 phenotype, whereas E4 carriers did not differ from E3/3 subjects. For the correlation between WHR and total cholesterol levels, however, the interaction term did not reach statistical significance. The correlation between tobacco consumption and apoB levels was also higher in E2 carriers (P=.053). The ε2 allele modified the association between triglyceride concentrations and the L/O ratio in a similar way, increasing the inverse correlation between these two variables. All associations with alcohol consumption and physical activity were homogeneous among the apoE phenotype groups (Table 2⇓).
The stronger association of BMI and WHR with apoB levels in E2 carriers was demonstrated in both men and women (significance of E2 interaction terms, P<.05 in both genders). Three-way interaction terms with gender were not statistically significant. The correlations of apoB with BMI were 0.48 and 0.39 in female and male E2 carriers, respectively, and the correlations with WHR were 0.27 and 0.16, respectively. By contrast, the interaction between the ε2 allele and BMI and WHR on total cholesterol, as well as the interaction between the ε2 allele and the L/O ratio on triglyceride concentrations was significant only in women (P<.01). In male subjects the correlations were quite similar among the three apoE phenotype groups. However, in neither case did the three-way interaction term with gender reach significance. Interaction effects did not differ significantly according to case/control status and region.
The stronger correlations of BMI and WHR with total cholesterol and apoB levels in E2 carriers suggested that an increase in these modifiable factors resulted in a larger rise in the levels in these subjects than in those with other phenotypes. A similar conclusion can be drawn for the relationship between the L/O ratio and plasma triglyceride concentrations. To further elucidate these interactions, we determined mean total cholesterol, apoB, and triglyceride levels according to gender-specific tertiles of BMI, WHR, and L/O ratio after stratification by apoE phenotype (Figs 1 through 3⇓⇓⇓). The lowering effect of the ε2 allele on total cholesterol and apoB levels was much less pronounced in the upper tertiles of BMI and WHR, so much that the levels in E2 carriers belonging to the upper tertile of BMI were comparable to those in E3/3 subjects (Figs 1⇓ and 2⇓).
Plasma triglyceride concentrations decreased according to tertiles of the L/O ratio (Fig 3⇑). Triglyceride concentrations were most elevated in E2 carriers in the lowest tertile, suggesting that the E2 allele exhibits its triglyceride-raising effect mainly when a diet high in saturated and low in polyunsaturated fat is consumed.
In multivariate regression analysis, the interactions demonstrated in univariate analyses remained statistically significant. In E2 carriers, 31.4% of the interindividual variation in apoB levels could be explained by gender, BMI, WHR, tobacco and alcohol consumption, physical activity, and the L/O ratio (Table 3⇓). In E3/3 subjects, this proportion was only 9.2%, a proportion quite similar to that observed in E4 carriers (13.6%). The higher R2 in E2 carriers was mostly explained by a larger effect of BMI on apoB in this phenotype group. For total cholesterol levels, congruent results were found. In contrast, the proportion of apoE and triglyceride variance explained by gender and modifiable factors was fairly similar in the three phenotype groups.
One of the major challenges of genetic epidemiology is to try to elucidate how environment and genes interact in determining individual susceptibility to multifactorial diseases. Most of the susceptibility genes to multifactorial diseases are frequent polymorphisms that, taken in the population at large, have a rather low impact at the individual level. However, in specific subgroups, genetic effects might amplify effects of lifestyle factors. One example is the possible modulation by alcohol intake of the cholesteryl ester transfer protein gene effect on HDL and the risk of myocardial infarction.24 Another example is the interaction between smoking habits and polymorphisms of the β-fibrinogen gene on plasma fibrinogen.25 26 The identification of such gene-environment interactions is crucial, since besides providing us a better understanding of the mechanisms involved in gene regulation, it may help to focus intervention strategies on target subgroups of the population.
While the effects of the apoE polymorphism on lipids have been extensively studied in various populations, only few studies have investigated interactions with environmental factors, and never in such a large sample as the present one. In the present study, associations between modifiable factors (eg, BMI, WHR, tobacco and alcohol consumption, and physical activity) and plasma apoE levels, studied in young adults from 11 countries throughout Europe, were nonsignificant and notably similar across apoE phenotypes. In contrast, a stronger effect of BMI and WHR on total cholesterol and apoB levels was demonstrated in E2 carriers than in those with the E3/3 phenotype. Multivariate analysis indicated that the modifiable factors explained about three times more of the interindividual variance in total cholesterol and apoB levels in E2 carriers than in other subjects, but this was mainly due to the larger contribution of BMI to the variability of these levels. Results for LDL cholesterol paralleled those found for apoB levels, but these results were not shown, because LDL cholesterol levels were calculated and not measured directly.
The fact that BMI emerged among all the factors studied is not unexpected, since the apoE polymorphism primarily affects lipid metabolism, and adiposity is a major metabolic factor. On the other hand, it is possible that other lifestyle factors, such as smoking, may not have yet exhibited their full effect on lipids, since subjects participating in EARS were relatively young (18 to 26 years). Although the interaction term did not reach statistical significance, we demonstrated that the association between tobacco consumption and apoB levels was also stronger in E2 carriers. When studying older subjects, who have longer lifetime risk-factor exposure, differences according to phenotypes may become more pronounced.
The deviation of the E2 carriers from the other apoE phenotype groups is in accordance with results of Reilly et al13 showing that the heterogeneity of regression of several lipids and apolipoproteins to concomitants was mostly due to differences between the ε32 and ε33 genotypes. In their study, associations of triglyceride, total cholesterol, and HDL cholesterol levels with weight and WHR were stronger in women with the E3/2 than in women with the E3/3 phenotype. In men, on the contrary, associations of WHR with the same lipid parameters were weaker in subjects with the E3/2 phenotype. In contrast to our results, associations with apoB levels were not different between phenotypes. Some of the interactions in the present study, eg, the interaction of the apoε2 allele with BMI and WHR on total cholesterol levels and with the L/O ratio on triglyceride concentrations, were also restricted to women. These results suggest that sex-specific factors (eg, hormonal factors) act as important regulators on these complex metabolic pathways.
A large study in children aged 8 to 16 years also demonstrated stronger correlations between adiposity and apoB levels in E2 carriers.27 In a small study in obese women, both the ε2 and ε4 allele altered the relationships between body fatness indices and plasma lipoproteins, but in contrast to the present study, no correlations were found between adiposity and apoB levels in E3/2 subjects.14
The stronger correlation between adiposity measures demonstrated in our study and in the study of Srinivasan et al27 suggests that weight loss, aimed at lowering apoB or LDL cholesterol levels, might be more effective in subjects carrying the ε2 allele. A study of Muls et al,28 however, demonstrated no differences in the effect of weight loss on lipid levels according to apoE phenotype. Results of another study suggested that weight gain was associated with a larger increase in triglyceride and β-lipoprotein concentrations not in E2 carriers but in E4 carriers.29
Several experimental studies demonstrated higher cholesterol responses to a dietary regimen reducing the amount of dietary fat in subjects with the E4/3 or E4/4 phenotype,8 9 10 11 12 while others failed to do so.30 31 It was also suggested that the apoE polymorphism did not have any major effect on the response of lipid levels to increased dietary cholesterol.32 In our study we found that the association of the L/O ratio, a marker for the dietary P/S ratio,23 with plasma triglyceride concentrations was more marked in E2 carriers. A recent observational study published by Marshall et al33 demonstrated that the association between dietary cholesterol and plasma LDL cholesterol was strongest in E2 carriers. These results are at variance with the results from experimental studies. When showing modulation of dietary responses by the apoE polymorphism, it is rather the ε4 allele that deviates. This discrepancy might, on the one hand, reflect differences between the effect of normal dietary fatty acid intake and the effect of a lipid-lowering diet. Lipid concentrations are more variable after a change in dietary saturated fat or cholesterol.32 The observational data presented here may better represent the effects of long-term dietary adaptation. On the other hand, plasma cholesteryl esters only partly reflect the fatty acid composition of the diet.23 34 35 EARS II, recently carried out with a similar design as the study described here and including oral glucose and fat tolerance tests, will allow us to study more precisely the effect of the apoE polymorphism on dietary responses.
The E2 isoprotein has defective receptor-binding affinity.36 Differences in binding affinity of the apoE isoforms for the remnant (apoB/E) receptor and the LDL receptor will result in differences in in vivo clearance rates and may therefore underlie the reported differences in (apo)lipoprotein levels according to apoE genotypes.36 37 Obesity and abdominal fat accumulation result in a higher VLDL secretion and consequently higher LDL cholesterol levels.38 Despite the upregulation of the LDL receptor in E2 carriers, the diminished receptor-binding capacity of the E2 isoform might result in a slower clearance of excess VLDL secreted and therefore result in a stronger rise in LDL particles with increasing adiposity in E2 carriers. This might explain the stronger correlations with total cholesterol apoB levels demonstrated in E2 carriers than in individuals with other apoE phenotypes. On the other hand, the BMI/E2 interaction could reflect a gene-gene interaction with some other gene involved in lipolysis. The lipoprotein lipase gene is mentioned as a candidate gene for obesity, and it has been suggested that the Asn291→Ser mutation of the lipoprotein lipase gene might interact with the ε2 allele to predispose to hyperlipidemia.39 In the same line of evidence, the postheparin plasma lipoprotein lipase activity has been shown to be related to plasma triglyceride and apoB levels only in E2 carriers.40
In conclusion, this large study among healthy European students showed that the apoE polymorphism did not modify effects of modifiable factors on plasma apoE concentrations and had little influence on the effects of triglyceride concentrations. Therefore, the apoE isoform seems to act in a relatively uniform manner, independently of lifestyle. However, the association of adiposity with total cholesterol and apoB levels appears to be altered in apoE2 carriers. The identification of gene-environment interactions may help to focus intervention strategies on target subgroups in the population.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|EARS||=||European Atherosclerosis Research Study|
|L/O ratio||=||plasma cholesteryl linoleate-to-oleate ratio|
|P/S ratio||=||dietary polyunsaturated-to-saturated fatty acid ratio|
EARS Project Leader
J. Shepherd, Glasgow, UK.
EARS Project Management Group
F. Cambien, Paris, France; G. De Backer, Ghent, Belgium; M.M. Galteau, Nancy, France; D. St J. O’Reilly, Glasgow, UK; M. Rosseneu, Brugge, Belgium; and L. Wilhelmsen, Göteborg, Sweden.
EC COMAC-Epidemiology Liaison Officer
T. Sorensen, Copenhagen, Denmark.
The EARS Group, Collaborating Centers, and their Associated Investigators
Austria. Recruitment Center and Laboratory: H.J. Menzel, C. Sandholzer, C. Duba, H.G. Kraft, Institute for Medical Biology and Genetics, University of Innsbruck.
Belgium. Recruitment Center: G. De Backer, S. De Henauw, D. De Bacquer, A. Bael, Department of Hygiene and Social Medicine, State University of Ghent. Laboratory: M. Rosseneu, N. Vinaimont, Department of Clinical Chemistry, University Hospital St Jan, Brugge.
Denmark. Recruitment Center and Laboratory: C. Gerdes, O. Faergeman, L.U. Gerdes, I.C. Klausen, Medical Department I, Aarhus Amtssygehus.
Finland. Recruitment Center and Laboratory: C. Ehnholm, National Public Health Institute, Helsinki; R. Elovaino, J. Peräsalo, The Finnish Student Health Service. Recruitment Center: A. Kesaniemi, Department of Internal Medicine, University of Oulu; P. Palomaa, The Finnish Student Health Service.
France. EARS Data Center: F. Cambien, L. Tiret, R. Agher, V. Nicaud, R. Rakotovao, INSERM U 258, Unité de Recherche d’Epidémiologie Cardiovasculaire, Hôpital Broussais, Paris. EARS Central Laboratory: M.M. Galteau, S.M. Visvikis, Center de Médecine Préventive, Nancy. Laboratory: J.C. Fruchart, J.M. Bard, P. Lebel, Service de Recherche sur les Lipoprotéines et L’Athérosclérose (SERLIA), INSERM U 325, Institut Pasteur, Lille. Laboratory: L. Bara; Laboratoire de Thrombose Expérimentale, Paris. Recruitment Center: C. Bady, J. Beylot, A. Lindoulsi, L. Tiret, UFR de Santé Publique, Bordeaux.
Germany. Recruitment Center and Laboratory: U. Beisiegel, A. Jorge, M. Papanikolaou, Medizinische Klinik Universitätskrankenhaus, Hamburg.
Italy. Recruitment Center: E. Farinaro, Community Medicine, Institute of Hygiene and Preventive Medicine; F. De Lorenzo, C. Cortese, M. Liguori; Institute of Internal Medicine and Metabolic Disease, University of Naples–Frederico II.
The Netherlands. Laboratory: L.M. Havekes, P. de Knijff, IVVO-TNO Health Research, Gaubius Institute, Leiden.
Spain. Recruitment Center: S. Sans, T. Puig, Programma CRONICAT, Hospital Sant Pau, Barcelona. Recruitment Center and Laboratory: M. Heras, A.E. La Ville, P.R. Turner, M. Masana, Unitat Recerca Lipids, Universitat Barcelona, Reus.
Sweden. Recruitment Center: L. Wilhelmsen, S. Johansson, I. Wallin, Department of Medicine, Ostra Hospital, University of Göteborg.
Switzerland. Recruitment Center: F. Gutzwiller, B. Marti, M. Knobloch, P. Anliker, Institute of Social and Preventive Medicine, University of Zurich.
United Kingdom. Recruitment Center: D. Stansbie, H. Denton, S. Plumridge, Department of Chemical Pathology, Bristol, Royal Infirmary. Recruitment Center and Laboratory: J. Shepherd, D. St J. O’Reilly, M.J. Murphy, G. Lindsay, Institute of Biochemistry, Royal Infirmary, Glasgow. Laboratory: S. Humphries, P. Talmud, S. Ye, University College London School of Medicine.
EARS was supported by EC Concerted Action MRH4-COMAC Epidemiology. We are grateful to Viviane Nicaud for assistance in data analysis.
Davignon J, Gregg RE, Sing CF. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis. 1988;8:1-21.
Mahley RW. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science. 1988;240:622-630.
Dallongeville J, Lussier-Cacan S, Davignon J. Modulation of plasma triglyceride levels by apoE phenotype: a meta-analysis. J Lipid Res. 1992;33:447-454.
Luc G, Bard J-M, Arveiler D, Evans A, Cambou J-P, Bingham A, Amouyel P, Schaffer P, Ruidavets J-B, Cambien F, Fruchart J-C, Ducimetière P. Impact of apolipoprotein E polymorphism on lipoproteins and risk of myocardial infarction: the ECTIM Study. Arterioscler Thromb. 1994;14:1412-1419.
Tiret L, de Knijff P, Menzel H-J, Ehnholm C, Nicaud V, Havekes LM for the EARS Group. ApoE polymorphism and predisposition to coronary heart disease in youths of different European populations: the EARS study. Arterioscler Thromb. 1994;14:1617-1624.
Tikkanen MJ, Huttunen JK, Ehnholm C, Pietinen P. Apolipoprotein E4 homozygosity predisposes to serum cholesterol elevation during high-fat diet. Arteriosclerosis. 1990;10:285-288.
Lopez-Miranda J, Ordovas JM, Mata P, Lichtenstein AH, Clevidence B, Judd JT, Schaefer EJ. Effect of apolipoprotein E phenotype on diet-induced lowering of plasma low density lipoprotein cholesterol. J Lipid Res. 1994;35:1965-1975.
Dreon DM, Fernstrom HA, Miller B, Krauss RM. Apolipoprotein E isoform phenotype and LDL subclass response to a reduced-fat diet. Arterioscler Thromb Vasc Biol. 1995;15:105-111.
Miettinen TA, Gylling H, Vanhanen H, Ollus A. Cholesterol absorption, elimination, and synthesis related to LDL kinetics during varying fat intake in men with different apoprotein E phenotypes. Arterioscler Thromb. 1992;12:1044-1052.
Pouliot M-C, Després J-P, Moorjani S, Lupien PJ, Tremblay A, Bouchard C. Apolipoprotein E polymorphism alters the association between body fatness and plasma lipoproteins in women. J Lipid Res. 1990;31:1023-1029.
Robertson FW, Cumming AM. Effects of apoprotein E polymorphism on serum lipoprotein concentration. Arteriosclerosis. 1985;5:283-292.
Taimela S, Lehtimäki T, Porkka KVK, Räsänen L, Viikari JSA. The effect of physical activity on serum total and low-density lipoprotein cholesterol concentrations varies with apolipoprotein E phenotype in male children and young adults: the Cardiovascular Risk in Young Finns Study. Metabolism. 1996;45:797-803.
The EARS Group. The European Atherosclerosis Research Study (EARS): design and objectives. Int J Epidemiol. 1994;23:465-471.
World Health Organization. Health Statistics Annual. Geneva, Switzerland: World Health Organization; 1989.
Bury J, Vercaemst R, Rosseneu M, Belpaire F. Quantification of apolipoprotein E by enzyme-linked immunosorbent assay. Clin Chem. 1986;32:265-270.
Havekes LM, de Knijff P, Beisiegel U, Havinga J, Smit J, Smit M, Klasen E. A rapid micromethod for apolipoprotein E phenotyping directly in serum. J Lipid Res. 1987;28:455-463.
Fumeron F, Betoulle D, Luc G, Behague I, Ricard S, Poirier O, Jemaa R, Evans A, Arveiler D, Marques-Vidal P, Bard JM, Fruchart JC, Ducimetière P, Apfelbaum M, Cambien F. Alcohol intake modulates the effect of a polymorphism of the cholesteryl ester transfer protein gene on plasma high density lipoprotein and the risk of myocardial infarction. J Clin Invest. 1995;96:1664-1671.
Behague I, Poirier O, Nicaud V, Evans A, Arveiler D, Luc G, Cambou JP, Scarabin PY, Bara L, Green F, Cambien F. Beta fibrinogen gene polymorphisms are associated with plasma fibrinogen and coronary artery disease. Circulation. 1996;93:440-449.
Muls E, Kempen K, Vansant G, Cobbaert C, Saris W. The effects of weight loss and apolipoprotein E polymorphism on serum lipids, apolipoprotein AI and B, and lipoprotein(a). Int J Obes. 1993;17:711-716.
Marshall JA, Kamboh MI, Bessesen DH, Hoag S, Hamman RF, Ferrell RE. Associations between dietary factors and serum lipids by apolipoprotein E polymorphism. Am J Clin Nutr. 1996;63:87-95.
Cambien F, Warnet J-M, Vernier V, Ducimitière P, Jacqueson A, Flament C, Orssaud G, Richard J-C, Claude J-R. An epidemiologic appraisal of the associations between the fatty acids esterifying serum cholesterol and some cardiovascular risk factors in middle-aged men. Am J Epidemiol. 1988;127:75-86.
Gregg RE, Brewer HB Jr. The role of apolipoprotein E and lipoprotein receptors in modulating in vivo metabolism of apolipoprotein B–containing lipoproteins in humans. Clin Chem. 1988;34:B28–B32.
Després J-P. Obesity and lipid metabolism. Curr Opin Lipidol. 1991;2:5-15.
Zhang H, Reymer PWA, Liu MS, Forsythe IJ, Groenemeyer BE, Frohlich J, Brunzell JD, Kastelein JJP, Hayden MR, Ma Y. Patients with apoE3 deficiency (E2/2, E3/2, and E4/2) who manifest with hyperlipidemia have increased frequency of an Asn 291→Ser mutation in the human LPL gene. Arterioscler Thromb Vasc Biol. 1995;15:1695-1703.