Apolipoprotein E4 phenotype increases non-fasting serum triglyceride concentration in infants the STRIP study

Transcription

1 Atherosclerosis 152 (2000) Apolipoprotein E4 phenotype increases non-fasting serum triglyceride concentration in infants the STRIP study Anne Tammi a, *, Tapani Rönnemaa b, Jorma Viikari b, Eero Jokinen c, Helena Lapinleimu d, Christian Ehnholm e, Olli Simell d a Cardiorespiratory Research Unit, Uni ersity of Turku, Kiinamyllynkatu 10, FIN Turku, Finland b Department of Medicine, Uni ersity of Turku, Kiinamyllynkatu 4 8, FIN Turku, Finland c Hospital for Children and Adolescents, Uni ersity of Helsinki, Stenbäckinkatu II, FIN Helsinki, Finland d Department of Pediatrics, Uni ersity of Turku, Kiinamyllynkatu 4 8, FIN Turku, Finland e National Public Health Institute, P.B. 450, FIN Helsinki, Finland Received 25 January 1999; received in revised form 28 September 1999; accepted 15 October 1999 Abstract As genetically determined apolipoprotein E (apo E) phenotypes influence serum cholesterol concentration, we analysed whether serum triglyceride values are also affected by the apo E phenotypes in infants. Non-fasting serum triglyceride values were measured in 7- and 13-month-old participants in the STRIP project, a randomised, prospective trial aimed at reducing children s exposure to known atherosclerosis risk factors (n=1062). The mean S.D. non-fasting serum triglyceride concentrations in 7-month-old infants with apo E4/4 (n=36), E3/4 (n=209), E3/3 (n=412), and E2/3 (n=66) were , , , and mmol/l, respectively. Triglyceride concentrations were higher in infants with apo E4/4 or3/4 than in those with apo E3/3 (P-value for difference 0.01 and 0.009, respectively). The apo E phenotype similarly influenced non-fasting serum triglyceride concentrations at the age of 13 months. The differences in serum triglyceride values in apo infants (apo E3/4 and 4/4 infants combined) and apo E4 infants (apo E2/3 and 3/3 infants combined) occurred independently of the relative weight of the infant, milk type used at 7 months of age (breast milk or formula), and time elapsed from the previous meal. To conclude, apo E phenotypes regulate non-fasting serum triglyceride values in healthy infants. Apo E3/4 and apo E4/4 predispose infants to higher values than apo E3/3 phenotype, suggesting that the 4 allele may increase atherosclerosis risk also via it s effect on postprandial triglyceride metabolism Elsevier Science Ireland Ltd. All rights reserved. Keywords: Triglycerides; Apo E; Infants; Breast milk; Cholesterol 1. Introduction Apolipoprotein E (apo E) is a constituent of triglyceride rich lipoproteins (i.e. chylomicrons, very low density lipoproteins and their remnants) and some subclasses of high density lipoproteins (HDL) [1]. The gene for human apo E is polymorphic; the three common alleles 2, 3, and 4 encode the isoproteins E2, E3, and E4, respectively. The homozygous phenotypes E2/2, E3/3, and E4/4 and the heterozygous phenotypes E2/3, E2/4, and E3/4 thus predominate [2]. * Corresponding author. Tel.: ; fax: address: anne.tammi@utu.fi (A. Tammi). Apo E acts as a ligand for low density lipoprotein (LDL) and remnant receptors, thus playing an important role in the metabolism of cholesterol and triglyceride rich lipoproteins [1]. Apo E also participates in catabolism of chylomicrons together with lipoprotein lipase (LPL), independently of the LDL receptor [3] and is involved in the conversion of intermediate density lipoprotein to LDL [4]. The association between the apo E phenotypes and serum total and LDL cholesterol concentrations is well established. On average, subjects with apo E2 have lower and those with apo E4 higher serum total and LDL cholesterol concentration than those with apo E3 [5 8]. The association between the apo E phenotypes and serum triglyceride values has remained less clear /00/$ - see front matter 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S (99)

2 136 A. Tammi et al. / Atherosclerosis 152 (2000) The prevalence of 2 allele is increased in individuals with hypertriglyceridemia [9,10] and clearance of triglyceride rich particles is prolonged also in normolipidemic individuals with one or two 2 alleles [11,12]. E2/2 phenotype is an obligatory, but not the only, prerequisite for the development of human type III hyperlipidemia, characterised by accumulation of remnants of triglyceride rich lipoprotein particles in plasma [13]. However, this rare condition alone fails to explain the 2-hypertriglyceridemia association. A few studies suggest that 4 carriers also have higher serum triglyceride values than 3 homozygotes [14 18], and the meta-analysis by Dallongeville and coworkers shows that subjects with apo E2 phenotype and apo E4 heterozygotes have higher serum triglyceride concentrations than E3 homozygotes [19]. To analyse the effects of the apo E phenotypes on serum triglyceride values in breast-fed and recently weaned infants, serum triglyceride values and apo E phenotypes were determined in a large population of 7- and 13-month-old infants in Finland, a country with an exceptionally high prevalence of the 4 allele. 2. Methods 2.1. Subjects This study comprises a part of the participants of the STRIP project (Special Turku Coronary Risk Factor Intervention Project), which is a randomised, prospective trial aimed at decreasing exposure of children to known environmental atherosclerosis risk factors. The ongoing project was launched in Turku, Finland, in Details of the study design have been published [20]. In brief, 1062 infants were randomised to intervention (n=540) and control (n= 522) groups at the age of 7 months. This study comprises those infants from whom blood was successfully drawn at the ages of 7 months (n=745) and 13 months (n=859). Blood samples from 671 infants were available at both ages Counselling The intervention families visited the counselling team (a paediatrician and a dietician) at infant s ages of 7, 8, 10 and 13 months. Infant s fat intake was not restricted, but saturated fatty acids were suggested to be replaced by polyunsaturated and monounsaturated fatty acids to approach a ratio of polyunsaturated to monounsaturated to saturated fatty acids of 1:1:1. Infant foods low in saturated fat and cholesterol were thus recommended. A daily cholesterol intake of less than 200 mg was also proposed. Solid foods were introduced to all infants at 3 6 months of age; all infants received breast milk or formula to the age of 1 year. The intervention infants then had skim milk as their primary milk source. The intervention parents were advised to add two to three teaspoonfuls (10 15 g) vegetable oil or soft margarine into the infant s daily diet to confirm an adequate supply of energy and fat (30 35% of total energy). The control families visited the same counselling team at the infant s ages of 7 and 13 months. They received no individualised dietary counselling about the amount or quality of fat in the child s diet. Solid foods were introduced to control infants at 3 6 months of age. The control infants consumed breast milk or formula also until the age of 1 year but changed then to milk with at least 1.9% fat, as was counselled at the well baby clinics in Finland at that time Ethics The STRIP project has been approved by the Joint Commission on Ethics of the Turku University and the Turku University Central Hospital Biochemical determinations and weight measurement Non-fasting blood samples were drawn between 08:00 and 17:00 h under cutaneous anaesthesia (Emla, Astra, Södertälje, Sweden) from an antecubital vein. The mean time interval between the infant s previous meal and sampling was h at 7 months of age and h at 13 months of age; the time interval was similar among carriers of different apo E phenotypes. Serum was separated by low-speed centrifugation (3400 g, for 12 min) after clotting at room temperature and was stored for 1 month at 25 C. Serum triglyceride concentration was determined with a fully enzymatic colorimetric GPO-PAP method (Boehringer, Mannheim, Germany) using a Kone CD analyser. Serum HDL cholesterol concentration was measured using a fully enzymatic method after precipitation of LDL and VLDL with dextran sulphate as described [21]. Apo E phenotypes were determined using isoelectric focusing and immunoblotting of delipidated serum [7]. All analyses were performed in the laboratory of the Research and Development Unit of Social Insurance Institution in Turku. The weight of the infants was measured to the nearest 0.01 kg with a baby scale (Seca 725, Hamburg, Germany) and expressed as relative weight, i.e. as deviation in percentages from the mean weight of healthy Finnish children of the same height and sex [22].

3 A. Tammi et al. / Atherosclerosis 152 (2000) Statistical analysis The results are expressed as means S.D. Serum triglyceride concentrations were log-transformed for statistical analyses because of skewness of the data. Differences between two groups were tested with a two-sample t-test. One- or two-way analysis of variance was used in analyses of overall differences between the apo E phenotypes. In some analyses, subjects with apo E3/4 and 4/4 phenotypes were analysed together as the apo group. Similarly, in some analyses apo E2/3 and apo E3/3 children were combined to form the apo E4 group. Pearson s correlation coefficient (r) was calculated for serum triglyceride values at the two age points and for correlation between concentrations of serum HDL cholesterol and triglycerides. P-values 0.05 were considered significant. Statistical analyses were performed using SAS release 6.12 program package (SAS Institute, Cary, NC). 3. Results The apo E allele frequencies in the groups of infants at the age of 7 and 13 months were identical ( 2, 0.06; 3, 0.74; and 4, 0.20) (Table 1). Only four and six infants had apo E2/2 phenotype and only 18 and 22 infants had apo E2/4 phenotype at the ages of 7 and 13 months, respectively. The non-fasting serum triglyceride concentration (mean S.D.) in the apo E2/2 infants was and mmol/l at the ages of 7 and 13 months, respectively. Similarly, the mean concentration in the apo E2/4 infants was and mmol/l, respectively. Because of the small number of infants with these apo E phenotypes, they were excluded from further analyses. The mean time interval between infant s previous meal and drawing of the blood sample did not correlate with measured serum triglyceride concentrations at the ages of 7 and 13 months (r= 0.002, P=0.96 and r= 0.01, P=0.80, respectively). Further, the mean time interval and serum triglyceride values showed no correlation in the apo infants (r=0.02, P=0.81 for infants aged 7 months and r= 0.05, P=0.43 for infants aged 13 months) and in the apo E4 infants (r 0.01, P=0.94 for infants aged 7 months and r= 0.06, P=0.17 for infants aged 13 months). Also, no correlation was observed when infants with the four main apo E phenotypes were studied separately at these age points (data not shown). The mean non-fasting serum triglyceride concentration was higher at the age of 7 than at 13 months ( vs mmol/l). In those 671 infants whose samples were successfully measured at both age points, serum triglyceride concentration decreased by mmol/l (P=0.0001) during the 6-month follow-up. The serum triglyceride concentrations at 7 and 13 months of age correlated (r=0.21, P=0.0001) in these 671 children despite the fact that the time interval between the previous meal and drawing the blood sample varied randomly at both ages. Girls had higher mean non-fasting serum triglyceride concentration than boys at the age of 7 months ( vs mmol/l, P=0.02) as well as at 13 months ( vs mmol/l, P=0.0002). As serum triglyceride values and apo E phenotypes associated similarly in the girls and boys, the genders were combined in further analyses. The apo infants had higher non-fasting serum triglyceride concentration than the apo E4 infants both at 7 and 13 months of age. Stratification of serum triglyceride values according to apo E phenotypes Table 1 Characteristics of the study subjects at the ages of 7 and 13 months E2/3 E3/3 E3/4 E4/4 E4 N Proportion of study infants (%) a Females, n Male, n Relative weight (%) b N Proportion of study infants (%) a Females, n Males, n Relative weight (%) b a Infants with apo E2/2 or apo E2/4 phenotype are excluded. b Values are means S.D. Relative weight is expressed as deviation in percentage from the mean weight of healthy Finnish children of same height and sex. For the overall differences in relative weight between the four different apo E phenotypes at the ages of 7 and 13 months P=0.11 and 0.51, respectively (one-way analysis of variance).

4 138 A. Tammi et al. / Atherosclerosis 152 (2000) Table 2 Non-fasting serum triglyceride concentration (mmol/l) in 7- and 13-month-old infants stratified according to apo E phenotype a E2/3 E3/3 E3/4 E4/4 E4 Mean S.D Median (range) 1.60 ( ) 1.40 ( ) 1.63 ( ) 1.70 ( ) 1.70 ( ) 1.40 ( ) P-value 0.42 b b 0.01 b c Mean S.D Median (range) 1.10 ( ) 1.30 ( ) 1.40 ( ) 1.70 ( ) 1.40 ( ) 1.20 ( ) P-value 0.39 b 0.02 b b c a group is apo E3/4 and 4/4 infants combined, E4 group is E2/3 and E3/3 infants combined. b P-values refer to comparisons between the values of the apo E3/3 infants and infants with other phenotypes (one-way analysis of variance). c P-values refer to comparisons between the values of the infants and E4 infants (two-sample t-test). Table 3 Non-fasting serum triglyceride concentration (mmol/l) in and E4 infants stratified according to milk type used at the age of 7 months and stratified according to randomised group in the STRIP project (intervention or control) at the age of 13 months a Breast milk only Breast milk+formula Formula only E4 E4 E4 N Mean S.D (range) ( ) ( ) ( ) ( ) ( ) ( ) Intervention Control E4 E4 N Mean S.D (range) ( ) ( ) ( ) ( ) a At the age of 7 months, apo E4 phenotype associated with serum triglyceride values (P=0.01), whereas milk type showed no such association with triglyceride values (P=0.34). For the interaction term (milk phenotype), P=0.95 (two-way analysis of variance). At the age of 13 months, apo E4 phenotype associated with serum triglyceride values (P=0.005), whereas the group (intervention or control) showed no such association with serum triglyceride values (P=0.29). For the interaction term (intervention phenotype), P=0.40 (two-way analysis of variance). showed that the infants with apo E4/4 had the highest serum triglyceride concentration, and the triglyceride values declined in the order E4/4 E3/4 E3/3 (Table 2). of age, multivariate analysis with serum triglyceride concentration as the dependent variable and apo E4 phenotype ( or E4 ) and milk type (breast milk or formula) as the independent variable showed that the apo E4 type independently associated with non-fasting serum triglyceride concentration, whereas milk type showed no such association (Table 3). Consequently, data on infants fed solely breast milk, breast milk and formula, or solely formula were combined for further analyses. Non-fasting serum triglyceride values in infants aged 13 months showed no association with the group to which they were randomised (intervention group or control group) in a covariance analysis with triglyceride values as a dependent variable and treatment group and apo E4 type as covariates. Meanwhile, the apo E4 phenotype associated with serum triglyceride values (Table 3). Consequently, data on the intervention and the control infants were combined for further analyses. The relative weights of infants with different apo E phenotypes were closely similar (Table 1). When the infants were divided according to relative weight tertiles, the E4 infants in the heaviest tertile had the highest mean serum triglyceride concentration at the age of 7 months (Table 4). Apo E4 phenotype influenced serum triglyceride values in a multivariate analysis including apo E4 phenotype and relative weight tertile. The relative weight tertile also had an effect on serum triglyceride values at the age of 7 months, but not at 13 months. The multivariate analysis showed no weight apo E phenotype interaction.

5 A. Tammi et al. / Atherosclerosis 152 (2000) The 4 allele also had an effect on serum HDL cholesterol concentration in these infants. Serum HDL cholesterol concentration was lower in the infants with 4 allele than in those without the allele at 7 months of age ( vs mmol/l, respectively; P=0.03) and again at 13 months of age ( vs mmol/l; P=0.003). Concentrations of serum HDL cholesterol and triglycerides showed a moderate correlation in the apo infants (r= 0.34, P=0.0001) and in the apo E4 infants (r= 0.37, P=0.001) at the age of 7 months. Similar correlations were found at the age of 13 months (r= 0.35, P= for the apo E4 infants and r= 0.31, P= for the apo E4 infants). We have earlier shown that in the whole STRIP study material, the mean concentration of serum total cholesterol is higher in infants with apo phenotype than in infants with apo E4 phenotype [8]. Similarly, in the present substudy in the infants from whom serum triglyceride value was available, the concentration of serum total cholesterol was higher in the apo infants than in the apo E4 infants ( vs mmol/l, P=0.0001) at 7 months of age and at 13 months of age ( vs mmol/l, P=0.0001). 4. Discussion The main finding of this study is that non-fasting serum triglyceride concentration is higher in the infants with the 4 allele than in those without the allele at the ages of 7 and 13 months. Furthermore, the infants homozygous for 4 had higher mean concentrations than the infants heterozygous for 4 and the infants without the 4 allele showed the lowest concentrations. In accordance with our data, a recent study in adults suggests that postprandial increases in serum triglyceride values are higher in subjects carrying the 4 allele than in subjects with other alleles [23]. Also, clearance of intestinal and hepatogenous triglyceride rich lipoprotein remnants is impaired in normolipidemic young men with the apo E4/3 phenotype [24]. A few other studies suggest that subjects carrying apo E4 may have increased fasting serum triglyceride concentration [14 18]. In the meta-analysis by Dallongeville and co-workers, 45 population samples from 17 countries suggest that plasma triglyceride values are increased in children and adults carrying the 2 allele or the 3/4 combination [19]. In that meta-analysis, serum triglyceride values in apo E4 homozygotes were similar to those in apo E3 homozygotes. However, this finding may have been an artefact caused by the small number of subjects carrying the E4/4 phenotype (1.6% of those studied). Meanwhile, the multinational EARS study suggests, that the 2 and 4 alleles equally increase the plasma triglyceride concentration [14]. Even though a few population studies support an association of the 2 allele with hypertriglyceridemia, our study shows no such association. For full expression of hypertriglyceridemia in subjects with the 2 allele, additional factors known to promote hypertriglyceridemia are probably required, and these factors (obesity, impaired glucose metabolism, or abuse of alcohol) were not features of our study infants. Another explanation for the absence of the 2-hypertriglyceridemia association in our study might again be that only 9% of the infants analysed carried the 2 allele as the infants with apo E2/2 and E2/4 phenotypes were excluded from statistical analyses. Table 4 Non-fasting serum triglyceride concentration (mmol/l) in 7- and 13-month-old apo and E4 infants stratified according to relative weight tertiles a E4 I tertile II tertile III tertile I tertile II tertile III tertile N Mean S.D (range) ( ) ( ) ( ) ( ) ( ) ( ) N Mean S.D (range) ( ) ( ) ( ) ( ) ( ) ( ) a Infants in the first (I) relative weight tertile were the lightest and those in the third (III) tertile, the heaviest. For difference in serum triglyceride values in relative weight tertiles in 7-month-old infants, P=0.13 ( infants) and P=0.04 (E4 infants). For difference in serum triglyceride values in relative weight tertiles in 13-month-old infants, P=0.22 ( infants) and 0.48 (E4 infants) (two-way analysis of variance). The apo E4 phenotype influenced serum triglyceride values (P=0.002 at 7 months of age and P= at 13 months of age). The relative weight tertile also had an effect on serum triglyceride values at the age of 7 months (P=0.01), but not at the age of 13 months (P=0.13). For the interaction term (weight phenotype) P=0.55 at 7 months of age and P=0.62 at 13 months of age (two-way analysis of variance).

6 140 A. Tammi et al. / Atherosclerosis 152 (2000) The slower clearance of chylomicron and VLDL remnants by the 2 carriers than by 3/3 homozygotes might well explain their high serum triglyceride concentration. But what could be the mechanism for our finding that the 4 allele strongly increases non-fasting serum triglyceride values in infants? As apo E and LPL co-ordinately enhance binding and uptake of chylomicrons by human hepatocytes via LDL receptor related protein in vitro [3], the possibility exists that the apo E4 molecule might function poorly in this reaction and thus disturb co-ordinate chylomicron catabolism. Different polymorphic forms of apo E may indeed be able to modulate postheparin LPL activity, and, consequently, plasma lipoprotein profile [25]. In support of this hypothesis, apo E4 positive men failed to show inverse correlation between plasma triglyceride concentration and postheparin LPL activity and positive correlation between plasma HDL cholesterol concentration and LPL activity while those parameters correlated well in apo E2 positive men. Furthermore, in the Stanislas cohort of 1101 healthy subjects, apo E and LPL simultaneously modulated fasting serum triglyceride values, 4 allele being associated with high values [18]. Serum triglyceride and serum HDL cholesterol concentrations show a strong inverse correlation in adults and children [26,27]. In this study, the apo infants, who had the highest serum triglyceride concentrations also had the lowest serum HDL cholesterol concentrations, and serum concentrations of HDL cholesterol and triglycerides showed a moderate correlation. A tendency towards low serum HDL cholesterol concentration in 4 carriers has been reported earlier in the STRIP project [8] and in other studies [14,19]. Because serum samples in our study were non-fasting, the triglyceride values represent the triglycerides synthesised in the liver and those derived from ingested fat. Still, the interval between infant s regular meal and drawing of the blood sample was of equal length among the apo infants and the apo E4 infants. However, a strong linear correlation exists between the fasting triglyceride concentration and postprandial triglyceride response and, thus, the fasting serum triglyceride level is an important determinant of postprandial lipemia [28,29]. A proper oral fat loading test with specified fat load meal and a standardised blood drawing protocol with a number of tests was not feasible because the study subjects were infants. Despite these shortcomings we observed a clear association between apo E4 phenotype and high serum triglyceride values in healthy infants at 7 and 13 months of age. This association would probably have been even stronger if we had measured fasting triglyceride values or used a standardised fat load protocol. Obesity predisposes adults to high serum triglyceride concentrations. The E4 infants in the heaviest weight tertile in this study had the highest serum triglyceride concentrations. In multivariate analysis the relative weight of infants showed no or only a modest association with serum triglyceride values, whereas infant s apo E4 phenotype (apo or apo E4 ) had a significant effect on serum triglyceride values suggesting that apo phenotype associates with high serum triglyceride values independently of relative body weight. In summary, this study shows that the inherited apo E phenotypes influence non-fasting serum triglyceride values in healthy infants. The 4 allele associates with high serum triglyceride concentration, suggesting that apo phenotype may increase atherosclerosis risk not only via high LDL cholesterol concentration but also via its effect on postprandial triglyceride metabolism. Acknowledgements This study was supported by grants from Foundation for Pediatric Research, Finland, Medical Council of the Academy of Finland, Mannerheim League for Child Welfare, Sigrid Juselius Foundation, Yrjö Jahnsson Foundation, Finnish Cardiac Research Foundation, Juho Vainio Foundation, and Turku University Hospital Research Fund. References [1] Mahley RW, Innerarity TL, Rall SC Jr, Weisgraber KH. Plasma lipoproteins: apolipoprotein structure and function. J Lipid Res 1984;25: [2] Zannis VI, Just PW, Breslow JL. Human apolipoprotein E subclasses are genetically determined. Am J Hum Genet 1981;33: [3] Mann WA, Meyer N, Weber W, Rinninger H, Greten H, Beisiegel U. Apolipoprotein E and lipoprotein lipase co-ordinately enhance binding and uptake of chylomicrons by human hepatocytes. Eur J Clin Invest 1995;25: [4] Ehnholm C, Mahley RW, Chappel DA, Weisgraber KH, Ludwig E, Witztum JL. Role of apolipoprotein E in the lipolytic conversion of -very low density lipoproteins to low density lipoproteins in type III hyperlipoproteinemia. Proc Natl Acad Sci 1984;81: [5] Utermann G, Pruin N, Steinmetz A. Polymorphism of apolipoprotein E. III. Effect of a single polymorphic gene locus on plasma lipid levels in man. Clin Genet 1979;15: [6] Ehnholm C, Lukka M, Kuusi T, Nikkilä E, Utermann G. Apolipoprotein E polymorphism in the Finnish population: gene frequencies and relation to lipoprotein concentrations. J Lipid Res 1986;27: [7] Lehtimäki T, Moilanen T, Viikari J, A kerblom HK, Ehnholm C, Rönnemaa T, et al. Apolipoprotein E phenotypes in Finnish youths: a cross-sectional and 6-year follow-up study. J Lipid Res 1990;31: [8] Lapinleimu H, Viikari J, Rönnemaa T, Välimäki I, Tuominen J, Marniemi J, et al. Apolipoprotein E polymorphism and serum lipids in a randomized prospective trial of an infant diet with reduced saturated fat and cholesterol. Pediatrics 1996;98:

7 A. Tammi et al. / Atherosclerosis 152 (2000) [9] Assmann G, Schmitz G, Menzel H-J, Schulte H. Apolipoprotein E polymorphism and hyperlipidemia. Clin Chem 1984;30: [10] Eto M, Watanabe K, Ishii K. Reciprocal effects of apolipoprotein E alleles ( 2 and 4) on plasma lipid levels in normolipidemic subjects. Clin Genet 1986;29: [11] Brenninkmeijer BJ, Stuyt PMJ, Demacker PNM, Stalenhoef AFH, van t Laar A. Catabolism of chylomicron remnants in normolipidemic subjects in relation to the apoprotein phenotype. J Lipid Res 1987;28: [12] Weintraub MS, Eisenberg S, Breslow JL. Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoprotein E. J Clin Invest 1987;80: [13] Davignon J, Gregg RE, Sing CH. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis 1988;8:1 21. [14] Tiret L, de Knijff P, Menzel H-J, Ehnholm C, Nicaud V, Havekes LM. ApoE polymorphism and predisposition to coronary heart disease in youths of different European populations: the EARS study. Arterioscler Thromb 1994;14: [15] Valdez R, Howard BW, Stern MP, Haffner SM. Apolipoprotein E polymorphism and insulin levels in a biethnic population. Diabetes Care 1995;18: [16] Haffner SM, Stern MP, Miettinen H, Robbins D, Howard BW. Apolipoprotein E polymorphism and LDL size in a biethnic population. Arterioscler Thromb Vasc Biol 1996;16: [17] Srinivasan SR, Ehnholm C, Wattigney WA, Bao W, Berenson GS. The relation of apolipoprotein E polymorphism to multiple cardiovascular risk in children: the Bogalusa Heart Study. Atherosclerosis 1996;123: [18] Salah D, Bohnet K, Gueguen R, Siest G, Visvikis S. Combined effects of lipoprotein lipase and apolipoprotein E polymorphisms on lipid and lipoprotein levels in the Stanislas cohort. J Lipid Res 1997;38: [19] Dallongeville J, Lussier-Cacan S, Davignon J. Modulation of plasma triglyceride levels by apo E phenotype: a meta-analysis. J Lipid Res 1992;33: [20] Lapinleimu H, Viikari J, Jokinen E, Salo P, Routi T, Leino A, Rönnemaa T, Seppänen R, Välimäki I, Simell O. Prospective randomized trial in 1062 infants of diet low in saturated fat and cholesterol. Lancet 1995;345: [21] Niinikoski H, Viikari J, Rönnemaa T, Lapinleimu H, Jokinen E, Salo P, Seppänen R, Leino A, Tuominen J, Välimäki I, Simell O. Prospective randomized trial of low-saturated-fat, low-cholesterol diet during the first 3 years of life. The STRIP baby project. Circulation 1996;94: [22] Sorva R, Perheentupa J, Tolppanen E-M. A novel format for a growth chart. Acta Paediatr Scand 1984;73: [23] Dart A, Sherrad B, Simpson H. Influence of apo E phenotype on postprandial triglyceride and glucose responses in subjects with and without coronary heart disease. Atherosclerosis 1997;130: [24] Bergeron N, Havel RJ. Prolonged postprandial responses of lipids and apolipoproteins in triglyceride-rich lipoproteins of individuals expressing an apolipoprotein epsilon 4 allele. J Clin Invest 1996;97: [25] St Amand J, Moorjani S, Lupien PJ, Prud homme D, Després J-P. The relation of plasma triglyceride, apolipoprotein B, and high-density lipoprotein cholesterol to postheparin lipoprotein lipase activity is dependent on apolipoprotein E polymorphism. Metabolism 1996;45: [26] Davies CE, Gordon D, LaRosa J, Wood PDS, Halperin M. Correlations of plasma high-density lipoprotein cholesterol levels with other plasma lipid and lipoprotein concentrations. Circulation 1980;62(Suppl 4): [27] Freedman DS, Srinivasan SR, Webber LS, Berenson GS. Divergent levels of high density lipoprotein cholesterol and apolipoprotein A-I in children: the Bogalusa heart study. Arteriosclerosis 1987;4: [28] Brown AJ, Roberts DCK. The effect of fasting triacylglyceride concentration and apolipoprotein E polymorphism on postprandial lipemia. Arterioscler Thromb 1991;11: [29] Reznik Y, Pousse P, Herrou M, Morello R, Mahoudeau J, Drosdowsky MA, et al. Postprandial lipoprotein metabolism in normotriglyceridemic non-insulin-dependent diabetes patients: Influence of apolipoprotein E polymorphism. Metabolism 1995;45: