Influence of apolipoprotein E polymorphism on the tracking of childhood levels of serum lipids and apolipoproteins over a 6-year period. The Bogalusa Heart Study

ATHEROSCLEROSIS Atherosclerosis127 (I 996) 73-79

ELSEVIER

Influence of apolipoprotein E polymorphism on the tracking of childhood levels of serum lipids and apolipoproteins over a 6-year period. The Bogalusa Heart Study Sathanur R.. SrinivasaP, Christian Ehnholmb, Wendy A. Wattigney”, Gerald S. Berenso@* “Tulane Center for Cardiovascular

.Health, Tulane School of Public Health and Tropical Medicine, 1501 Canal Street, 14th Floor, New Orleans, LA 70112-2824, USA bThe National Public Health Institute, Helsinki, Finland

Received6 February 1996;revised29 April 1996;accepted1 July 1996

Abstract The influence of apolipoprotein (apo) E polymorphism on the tracking of serum lipoprotein variables over a 6-year follow-up period was examined in 442 individuals aged 5- 15 years at baseline. The apo E phenotype-specific differences in total cholesterol and low-density lipoprotein (LDL) cholesterol levels persisted in the study cohort at baseline and follow-up examinations. However, the correlations of baseline versus follow-up levels of total cholesterol and LDL cholesterol varied according to the apo E phenotype group, with the apo E2 group, carrying E2/2 and E3/2 phenotypes, showing highest correlation for these variables (r = 0.73-0.74) and the apo E4 carrying E3/4 and E4/4 phenotypes the lowest (r = 0.48-0.59). The tracking correlation for LDL cholesterol in the apo E2 group was different from that of the other phenotype groups (P < 0.05). In terms of persistence in ranks over time, of the individuals who were in the highest quartile of LDL cholesterol at baseline none of those in the apo E2 group, 63% of those in the apo E3 group carrying E3/3 phenotype, and 60% of those in the apo I?.4group maintained this high rank at follow-up; corresponding values for persistence in ranking at the lowest quartile over time were 82% for the apo E2 group, 57% for the apo E3 group, and 33% for the apo E4 group. Further, in a multiple regression model, apo E phenotype was retained as a predictor variable only in the case of LDL cholesterol. Thus, apo E polymorphism influences not only the level of LDL cholesterol in childhood, but also its tracking at least over a 6-year period. Keywords:

Apo E polymorphism; Lipoproteins; Tracking; Childhood risk factors

-

1. Introduction Elevated levels of low-density lipoprotein (LDL) cholesterol and decreased levels of high-density lipoprotein (HDL) cholesterol are predictive of coronary artery disease (CAD) in adults [l]. Initial stages of atherosclerosis have been shown to relate strongly to adverse levels of lipoproteins in youth [2,3]. Further, it has been demonstrated recently that serum total cholesterol levels measured in young adults are associated with subsequent CAD in middle age [4]. It is becoming

* Correspondingauthor.

increasingly clear that childhood levels of serum lipoproteins persist (track) over time, and are somewhat predictive of adulthood values [5-81. For example, approximately 50% of children who had serum total cholesterol or LDL cholesterol in the top quintiles or quartiles remained so 12- 15 years later as young adults. Of importance, genetic factors are considered to play a role in the tracking of risk factor variables [9]. That apolipoprotein (apo) E genotype influences serum lipoprotein levels is now well recognized [lo]. In humans, there are three common allelic variants (e2, e3, and e4) at the structural locus for apo E [11,12]. These three alleles produce three homozygous phenotypes (E2/2, E3/3, and E4/4) and three heterozygous pheno-

0021-9150/96/%15.00 0 1996 Elsevier Science Ireland Ltd. All rights reserved PII SOO21-9150(96)05937-O

74

S.R. Srinivasan et al. / Atherosclerosis

types (E3/2, E4/3, and E4/2). The e2 allele is associated with lower and the e4 with higher total cholesterol, LDL cholesterol, and apo B levels than is the e3 allele [13-161. Although previous studies have shown no significant effect of this gene on longitudinal changes of serum lipoprotein in both children and adults [17], information is lacking in terms of tracking. Recently, the Bogalusa Heart Study reported the relative apo E allele frequencies and the impact of this polymorphism on serum lipoprotein levels in a biracial (black-white) population of children [ 181. The objective of the present study was to examine the influence of apo E polymorphism on the tracking of serum lipoprotein levels in a cohort from the above study who participated in multiple cross-sectional surveys over a period of 6 years. 2. Methods 2.1. Population

The Bogalusa Heart Study is a long-term epidemiologic study of early natural history of cardiovascular diseasein a biracial (65% white, 35% black) population of children and young adults [19]. In 1984-85, 2666 children in grades 3 through 12 were examined, representing 85% of all eligible individuals. A random subsample (n = 892) of children aged 8- 17 years was selected for apo E phenotyping to provide equal race and sex distribution [18]. Six hundred sixty-six individuals who were phenotyped for apo E participated in an earlier (1981-82) survey (year one) of children. Four hundred forty-two of these 5- to 15-year-olds who participated later as 1l- to 21-year-olds in a survey of children and young adults (year six) formed the study cohort for this report. 2.2. General examination

Essentially the same protocols were used in all surveys [19]. All participants were instructed to fast for 12 h prior to venipuncture, with compliance ascertained by an interview. Height and weight were measured to calculate the ponderal index (kg/m3), a measure of adiposity. 2.3. Serum lipid and lipoprotein analyses

Cholesterol and triglyceride levels were measured chemically with a Technicon Auto Analyzer II (Technicon Instrument, Tarrytown, NY) according to protocols developed by the Lipid Research Clinics Program until 1986 [20]. Since then, these variables were determined enzymatically on an Abbott VP instrument (Abbott Laboratories, North Chicago, IL). Both chemical

127 (1996) 73- 79

Table 1 Baseline serum lipid and lipoprotein levels (mean k SE) of individuals present or not present at follow-up. The Bogalusa Heart Study Variable” Cholesterol (mg/dl) Total VLDL LDL HDL Triglycerides Ow/dl)

Present at follow-up (n = 442)

Not present at follow-up (n = 224)

162.5f 7.0 f 94.3 f 61.3 f 61.2 +

160.8+_1.8 5.9 & 0.4 92.8 + 1.7 62.0 1 1.3 56.6 * 1.9

1.3 0.3 1.2 0.9 1.3

a.Age-, race-, and sex-adjusted.

and enzymatic procedures met the performance requirements of the Lipid Standardization Program of the Centers for Disease Control and Prevention, Atlanta, GA, which is monitoring the accuracy of measurements of total cholesterol, triglycerides, and HDL cholesterol concentrations. Serum lipoprotein cholesterols were measured by a combination of heparin-calcium precipitation and agar-agarose gel electrophoresis procedures

WI. 2.4. Apo E phenotyping

Serum samples that were kept frozen at - 7°C after lipid and lipoprotein analyses were sent to Helsinki, Finland (Dr. Ehnholm) for apo E phenotyping. Phenotyping was performed directly in serum using a modification [22] of the method of Havekes et al. [23], which is based on isoelectric focusing of delipidated serum followed by immunoblotting using rabbit antihuman apo E antiserum. Table 2 Mean + SE values of baseline demographic and serum lipoprotein variables in the study cohort by apo E phenotype group. The Bogalusa Heart Study Variable

Apo E phenotype group E2 (n = 32)

Age (years) Race Black/white (%)b Sex Male/female (%) Cholesterol (mg/dl) TotaP VLDL LDLb HDL Triglycerides (mg/dl)

9.1 + 0.5

E3 (n = 268) E4 (n = 142) 9.6 + 0.2

9.5 * 0.2

88/12

54146

65135

47153

49151

42158

152.6+ 4.8 5.4 & 1.1 77.8 + 4.5” 69.2 f 3.5” 63.0 f. 4.3

161.9+ 6.8 + 93.0 f 62.1 k 58.1 +

1.7 0.4 1.5 1.2 1.6

168.4+ 2.3” 7.0 * 0.5 101.0* 2.1” 60.5 f 1.6 64.1 + 2.2

a,bVariations among phenotype groups; a = P < 0.01; b = P < .OOOl; c age-, race-, and sex-adjusted; ,I different from E3 group, PC 0.05.

S.R. Srinivasan et al. / Atherosclerosis

127 (1996) 73- 79

Longitudinal Changes in Serum Lipoprotein Variables of the Study Cohort by Apo E Phenotype Group. The Bogalusa Heart Study 180

110~

LDL Cholesterol Total Cholesterol .,.~_~.~~~ ^,,I),_,,M-“----- 100._ _ _~< .j. ;. .( 170

,, _L, Ih_m*_--"

,_,

--' ; ~I,

5 5 E

90 160-

s 5E

10 6s

8

60

6

5!i Year

4 1

Year6

Year1

Year6

Fig. 1. Mean levels c’f serum lipoprotein variables in the study cohort (initial age: 5- I5 years) by apo E phenotype group and year of examination. The Bogalusa Heart Study.

2.5. Statistical analysis

Because sample size of homozygous phenotypes E2/2 (n = 3) and E4/4 (n = 15) in the study cohort was too small, individuals were classified into the following three phenotype groups as in previous studies [24]: (1) apo E2 group (n = 32; 88% black) carrying either the E2/2 or E3/2 phenotype; (2) apo E3 group (n = 268; 54% black) carrying the most common E3/3 phenotype; (3) apo E4 group (n = 142; 65% black) carrying either the E4/3 or E4;4 phenotype. Individuals who had the E4/2 phenotype (n = 22) were not included in the selection of study cohort because they carry an allele that is common to either the apo E2 group or apo E4 group. Statistical analysis was performed using the SAS software package [25]. Baseline (year one) serum lipid and lipoprotein levels were compared between individuals who were present at follow-up (year 6) examination and those who were lost to follow up using analysis of covariance to adjust for age, race and sex. Similar analysis was performed to examine differences in serum lipid and lipoprotein levels among the three apo E phenotype groups in the study cohort. Post hoc comparisons on the adjusted means were made by t-tests. Change (year 6-year 1) in levels of serum lipoprotein variables were likewise com.pared among the three apo E phenotype groups. Spearman’s rank correlation coefficients, a measure of tracking, were calculated between year 1 and year 6 levels of lipoprotein variables by apo E phenotype groups. Apo E-specific differences in the degree of tracking for each lipoprotein variable was assessedby testing the statistical significance of the difference be-

tween correlation coefficients after applying Fisher’s Z-transformation. The variance of the Z-transformed correlation coefficient is l/(n - 3) [26]. The persistence of high or low levels of lipoprotein variables was evaluated by age-, race-, and sex-specific quartiles, according to apo E phenotype groups. Significant predictors of follow-up levels of lipoprotein variables were explored using a stepwise regression procedure, with the significance level for entry and staying specified at 0.15 and 0.05, respectively. The independent variables consisted of baseline values of the same lipoprotein variables, race (0 = white, 1 = black), sex (0 = male, 1 = female), age, ponderal index (kgjm3), the change in ponderal index (year 6 - year l), and apo E phenotype (1 = E2, 2 = E3, 3 = E4).

3. Results 3.1. Lipid and lipoprotein levels

The study cohort (n = 442; 60% black) re-examined in year 6 represented 66%)of those identified at baseline (year 1). In order to examine the representativeness of the study cohort in terms of serum lipid and lipoprotein levels, their baseline values were compared to the baseline values of those who were not present at the followup survey six years later (Table 1). The age-, race-, and sex-adjusted baseline values of lipoprotein variables did not differ significantly between the two groups. Further, baseline levels of lipoprotein variables in the study cohort showed expected race differences, with blacks having higher total cholesterol (P < 0.02) and HDL

S.R. Srinivasan et al. / Atherosclerosis 127 (1996) 73-79

16

Correlation of Year 1 versus Year 6 Serum Lipoprotein Variables in the Study Cohort by Apo E Phenotype Group. The Bogalusa Heart Study

E 2 g 0.6 t 8 s ‘Z m m 0.4 t 8 5 kE 0.2 0 cn”

0

TC

TG

VLDL-C

LDL-C

HDL-C

Fig. 2. Correlation between year 1 and year 6 levels of serum lipoprotein variables in the study cohort by apo E phenotype group. The Bogalusa Heart Study.

cholesterol (P < O.OOOl),and lower triglycerides (P < 0.003) and very low density lipoprotein (VLDL) cholesterol (P < 0.05) than their white counterparts (data not shown). The mean baseline values of demographic and serum lipoprotein variables for the study cohort are given in Table 2 by apo E phenotype group. The age, race, and sex distributions of the study cohort indicated that blacks and whites were not evenly distributed among the apo E phenotype groups. Lipoprotein data on blacks and whites were combined for presentation, because there were only four whites in the apo E2 phenotype group. The age-, race-, and sex-adjusted mean levels of serum total cholesterol and LDL cholesterol varied among the three phenotype groups (P < 0.01-0.0001). The apo E2 phenotype group had lower values of LDL cholesterol compared to the apo E3 phenotype group (P < 0.05). On the other hand, the apo E4 phenotype group showed an opposite trend for total cholesterol and LDL cholesterol (P < 0.05). The above apo E phenotype-specific difference in total cholesterol and LDL cholesterol levels persisted in the study cohort at the follow-up examination 6 years later (Fig. 1). The follow-up levels of VLDL cholesterol and HDL cholesterol continued to show no apo E phenotype-specific difference. Changes in total cholesterol and LDL cholesterol over time by apo E phenotype groups were examined in terms of distribution of percent change (either positive or negative) of these variables between year 6 and year h 1 within each group. The results showed similarity among the phenotype groups (data not shown). The mean percent change between year 6 and year 1 levels of these lipoprotein variables did not vary among the phenotype groups (phenotype effect: total cholesterol, P = 0.93; LDL cholesterol, P = 0.75).

3.2. Tracking of lipoprotein variables

Tracking of levels of serum lipoprotein variables over time was examined by apo E phenotype group in terms of Spear-man’srank correlation between year 1 and year 6 values (Fig. 2). The correlation by apo E phenotype group was significant (P < 0.01-0.0001) for all the lipoprotein variables, except for triglycerides and VLDL cholesterol in the apo E2 phenotype group. As noted in earlier reports [5-81, the correlations were of highest magnitude for total cholesterol and LDL cholesterol. Both of these variables showed a trend in the magnitude of the correlation, according to the apo E phenotype group. Like total cholesterol, LDL cholesterol displayed the highest correlation (I = 0.74) in the apo E2 phenotype group, and the lowest correlation (Y= 0.48) in the apo E4 phenotype group. Further, the correlation coefficient for LDL cholesterol in the apo E2 phenotype group was significantly (P < 0.05) different from that of the other two phenotype groups. Tracking was further examined in terms of persistence in rankings at lowest or highest quartile of lipoprotein cholesterols by apo E phenotype group from baseline to follow-up after 6 years (Fig. 3). If there was no greater tracking than expected by chance, only 25% of those originally in a given quartile would remain in that ranking. The influence of apo E phenotype on persistence in ranks over time was evident only for LDL cholesterol. At the highest quartile of LDL cholesterol, none of those in the apo E2 phenotype group, 63% of those in the apo E3 phenotype group, and 60% of those in the apo E4 phenotype group showed persistence in ranks over time. On the other hand, at the lowest quartile of LDL cholesterol 82% of those in the apo E2 phenotype group, 57% of those in

S.R. Srinivasan rt al. )/Atherosclerosis

127 (1996) 71- 79

Persistence of Serum Lipoprotein Levels in the Study Cohort over 6 years by Apo E Phenotype Group. The Bogalusa Heart Study VLDL Cholesterol

LDL Cholesterol

HDL Cholesterol

loo----

Yr. 1 Quartile Yr. 6

: Lowest

Quartile:

Lowest

Highest

Lowest

Highest

Lowest

Highest

Highest

Lowest

Highest

Lowest

Highest

Fig. 3. Percent. of the study cohort remaining in extreme quartiles of serum lipoprotein variables after 6 y by apo E phenotype group. The Bo>alusa Heart Study

the apo E3 phenotype group, and 33% of those in the apo E4 phenotype group showed persistence in ranks over time. 3.3. Predictor variables

A stepwise multiple regression procedure was used to identify selected predictor variables that were related independently to follow-up levels of serum lipids and lipoproteins (Table 3). As in previous reports [5-81, the best predictor of follow-up level was the baseline level. Although change in adiposity, age, race, and sex contributed to the prediction of follow-up levels to varying degrees, apo E phenotype was retained as a predictor variable only in the case of LDL cholesterol. As expected, the prediction was best for total cholesterol and LDL cholesterol with 49% variability explained.

4. Discussion

The present study demonstrates that childhood levels of a single measurement of serum total cholesterol and LDL cholesterol levels are reasonably predictive of subsequent levels, as previously reported [5-8,27,28], and tracking of LDL cholesterol is influenced by apo E polymorphism. Further, the association of apo E polymorphism with serum total cholesterol and LDL cholesterol levels noted in our population study cohort is consistent with that of other studies showing higher values with phenotypes E4/4 and E4/3 and lower values with phenotypes E3/2 and E2/2 compared to those of the most common phenotype E3/3 [13- 161.

It is also apparent from the current study that the apo E phenotype-specific difference in total cholesterol and LDL cholesterol levels persist over time. This is consistent with the earlier finding that the effect of the apo E allele on lipoprotein levels is persistent in all age groups [17,18,22]. It has been reported that the apo E allele per se has no appreciable effect on the longitudinal changes of lipoprotein levels [ 171. Further, it has been proposed that the apo E gene acts as a ‘level’ gene, having a direct influence on serum lipoprotein levels 129,301. Although the apo E phenotype-specific difference in lipoprotein levels were maintained over a 6-year followup in the current study cohort, the correlations of year 1 versus year 6 levels of total cholesterol and LDL cholesterol varied according to the apo E phenotype. Individuals carrying the e2 allele showed the highest correlation for these two variables, while those carrying the e4 allele showed the lowest, reflecting the influence of apo E polymorphism on the phenomenon of tracking. Further, in a multiple regression model, apo E phenotype was retained as a predictor variable only in the case of LDL cholesterol. Additional studies in other populations are needed to confirm the present findings. It is of interest that at the highest quartile of LDL cholesterol distribution none of the e2 individuals and 60% of the e4 individuals showed persistence in rank over time. In contrast, at the lowest quartile of the distribution 82% of the e2 individuals and 33% of the e4 individuals displayed persistence in ranks over time. Thus, it appears that at low levels of LDL cholesterol the tendency to remain so over time may be high for those carrying the e2 allele and low for those carrying the e4 allele. This is consistent with the fact that the

S.R. Srinivasan et al. 1Atherosclerosis 127 (1996) 73- 79

78

Table 3 Predictors of follow-up levels of serum lipoprotein variables in the study cohort. The Bogalusa Heart Study Total cholesterol

Triglycerides

VLDL cholesterol

LDL cholesterol

HDL cholesterol

Baseline value Sex A wt/ht’

Baseline value A wt/ht3 Race” wt/ht’

Baseline value A wt/hts Race”

R2 = 0.49

R2 = 0.27

R2 = 0.20

Baseline value A wt/ht3 Age Sex Apo E phenotype R* = 0.49

Baseline value Age” Race Sex A wt/h+ R* = 0.32

Predictor variables are listed in the order they entered the model. a Negative association. Abbreviations: wt, weight; ht, height.

average effect of the e2 allele is to lower LDL cholesterol levels and the average effect of the e4 allele is to raise LDL cholesterol levels [10,15]. However, at high levels of LDL cholesterol, tracking was not remarkably high in e4 individuals. In fact, it remained similar to that of e3 individuals. The biological basis for such divergence in tracking of LDL cholesterol levels among the carriers of different apo E alleles is not clear. Available evidence provides some clues regarding the mechanism(s) by which different isoforms of apo E modulate serum LDL cholesterol levels [3 l-371. It has been shown that apo E2 binds poorly to high affinity lipoprotein receptors and their ligands [31,32]. A combination of reduced intestinal cholesterol absorption [33] and impaired apo E-mediated uptake of remnant particles [34-361 in e2 individuals vs. e3 or e4 individuals may decrease the delivery of dietary cholesterol to the liver. This may result in the up-regulation of hepatic LDL receptors, enhanced LDL catabolism, and as a consequence,decreasein serum LDL cholesterol levels. This implies that e2 individuals, unlike e3 or e4 individuals, may be relatively resistant to changes in LDL levels over time, and hence may track better. Regression toward the mean phenomenon is considered an important determinant of the change in lipoprotein variables over time [8,38], while baseline levels are by far the most important predictor of follow-up levels [5-81. The present study shows that percent change in total cholesterol and LDL cholesterol over time did not vary significantly among the apo E phenotype groups, suggesting that regression toward the mean effect may not vary among these groups. Further, it has been reported that conclusions regarding the prediction of high-risk follow-up levels based on baseline values are not affected by controlling for the regression toward the mean term [39]. Therefore, conclusions regarding the influence of apo E phenotype on tracking of total cholesterol and LDL cholesterol over time is not likely to be affected by regression toward the mean. In conclusion, apo E polymorphism influences not only the level of LDL cholesterol in childhood but also

its persistence over time. Whether the influence of apo E polymorphism on tracking of LDL cholesterol levels observed in the present study continues long-term is not known. Further, the role of the interaction between environmental variability and apo E polymorphism in determining tracking of LDL cholesterol levels needs to be elucidated.

Acknowledgements The Bogalusa Heart Study is a joint effort of many people whose cooperation is gratefully acknowledged. We are especially grateful to the study participants, their families, and the Bogalusa School System for making this study possible. We thank the field and laboratory staff.

References VI Kannel WB, CasteIIi WP, Gordon T. Cholesterol in the predic-

tion of atherosclerotic disease: new perspectives based on the Framingham Study. Ann Intern Med 1979;90:85. PI Newman WP, Freedman DS, Voors AW et al. Relation of serum lipoprotein levels and systolic blood pressure to early atherosclerosis: The Bogalusa Heart Study. N Engl J Med 1986;314:138. [31 Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group, Relationship of atherosclerosis in young men to serum lipoprotein cholesterol concentrations and smoking. J Am Med Assoc 1990;264:3018. [41 Klag MJ, Ford DE, Mead LA et al. Serum cholesterol in young men and subsequent cardiovascular disease. N Engl J Med 1993;328:313. I51Lauer RM, Clark WR. Use of cholesterol measurements in childhood for the prediction of adult hypercholesterolemia: The Muscatine Study. J Am Med Assoc 1990;264:3034. PI Webber LS, Srinivasan SR, Wattigney WA, Berenson GS. Tracking of serum lipids and lipoproteins from childhood to adulthood: The Bogalusa Heart Study. Am J Epidemiol 1991;133:884. [71 Mellies MJ, Laskarzewski PM, Glueck CJ. Tracking of highand low-density-lipoprotein cholesterol from childhood to young adulthood in a single large kindred with familial hypercholesterolemia. Metabolism 1985;34:747.

S. R. Srinioasan et al. I) Athrrosclerosis

[8] Porkka KVK, Viikari JSA, Taimela S, Dahl M, Akerblom HK. Tracking and predictiveness of serum lipid and lipoprotein measurements in childhood: A 12-year follow-up. The Cardiovascular Risk in Young Finns Study. Am J Epidemiol 1994;140:1096. [9] Harris CL, Sing CF, Clark WR, Schrott HG, Multivariate models for human genetic analysis: aggregation, coaggregation, and tracking elf systolic blood pressure and weight. Am J Hum Genet l983;35:1196. [IO] Davignon J, Gregg RE, Sing CF. Apolipoprotein E polymorphism and atherosclerosis. Arteriosclerosis 1988;8:1. [It] Utermann G, Hees M, Steinmetz A. Polymorphism of apolipoprotein E and occurrence of dysbetalipoproteinemia in man. Nature 1977;269:604. [12] Zannis VI, Just PW, Breslow JL. Human apolipoprotein E isoprotein subclasses are genetically determined. Am J Hum Genet ‘1981;33:11. [13] Utermann A, Pruin N, Steinmetz A. Polymorphism of apolipoprotein E. 111. Effect of a fsingle polymorphic gene locus on plasma lipid levels in man. Clin Genet 1979;15:63. [14] Ehnholm C. Lukka M, Kuusi T et al. Apolipoprotein E polymorphism in the Finnish population: Gene frequencies and relation to lipoprotein concentrations. J Lipid Res 1986;27:227. [15] Boerwinkle El, Visvikis S, Welsh D et al. The use of measured genotype information in the analysis of quantitative phenotypes in man. II. The role of the apolipoprotein E polymorphism in determining levels. variability, and covariability of cholesterol, betalipoprotein, and triglycerides in a sample of unrelated individuals. Am J Med Genet 1987;27:567. [I61 Xhignesse M, Lussier-Cacan S, Sing CF et al. Influence of common vanants of apolipoprotein E on measures of lipid metabolism in a sample selected for health. Arterioscler Thromb 1991;11:1100. [17] GuegusenR, ‘Jisvikis S, Steinmetz J, Siest G, Boerwinkle E. An analysis of genotype effects and their interactions by using the apolipoprotein E polymorphism and longitudinal data. Am J Hum Genet 1989:45:793. [18] Sriniva.san SR, Ehnholm C, Wattigney W, Berenson GS. Apolipoprotein E polymorphism and its association with serum lipoprotein concentrations in black versus white children. The Bogalusa He;lrt Study. Membolism 1993:42:381, 1191Berenson GS, McMahan CA, Voors AW et al. Cardiovascular Risk Factors in Children. The Early NatUral History of Atherosclerosis and Essential Hypertension. New York, NY: Oxford University Press, 1980, pp. l-450. [20] Lipid Reseamh Clinics Program: Manual of Laboratory Operations, Vol. 1. Bethesda MD: National Institutes of Health. DHEW Publication No. (NIH) 1974;755628. [21] Srinivasan SR, Berenson GS. Serum lipoprotein in children and methods for study. In: Lewis, LA. (ed). CRC Handbook of Electrophoresis, Vol. 3: Lipoprotein Methodology and Human Studies, Boca Raton, FL: CRC, 1983; pp. 1855206. [22] Lehtimaki T, Moilanen T, Viikari J et al. Apolipoprotein E phenotypes n Finnish youths: A cross-sectional and 6-year follow-up study. J Lipid Rcs 1990;31:487.

127 (1996) 7.3- 79

19

v31 Havekes L, de Knijeff P, Beisiegel U et al. A rapid micro-

method for apolipoprotein E phenotyping directly in serum. J Lipid Res 1987;28:455. ~241 Srinivasan SR, Ehnholm C, Wattigney W, Berenson GS. Relationship between obesity and serum lipoproteins in children with different apolipoprotein E phenotype: The Bogaiusa Heart Study. Metabolism 1994;43:470. [W Snedecor GW. Cochran WG. Statistical Methods, 7th ed., Ames. Iowa: Iowa State University Press, 1980: pp. 185-188. WI SASSTAT User’s Guide, Version 6. Fourth Edition, Volumes I and 2, Cary, NC: SAS Institute, Inc., 1989. 1271Gillum RF, Taylor HL, Brozek J et al. Blood lipids in young men followed 32 years. J Chronic Dis 1982;35:635. WI Orchard TJ, Donahue RP, Kuller LA et al. Cholesterol screening in childhood: Does it predict adult hypercholesterolemia? The Beaver County experience. J Pediatr 1983;103:687. m de Knijeff P. Boomsma DI, de Wit E et al. The effect of the apolipoprotein E phenotype on plasma lipids is not influenced by environmental variability: results of a Dutch twin study. Hum Genet 1993:91:268. 1301Magnus P, Berg K, Borrensen AL, Nance WE. Apparent influences of marker genotype on variation in serum cholesterol in monozygotic twins. Clin Genet 1981;19:67. [311 Weisgraber KH, lnnerarity TL, Mahley RW. Abnormal lipoprotein receptor-binding activity of the human apolipoprotein due to cysteine-arginine interchange at a single site. J Biol Chem 1982:257:2518. ~321 Brewer HB Jr, Zech LA, Gregg RE et al. Type III hyperlipoproteinemia: Diagnosis, molecular defects, pathology and treatment. Ann Intern Med 1983:98:623. [331 Kesaniemi AY, Ehnholm C, Miettinen TA. Intestinal cholesterol absorption efficiency in man is related to apolipoprotein E phenotype. J Clin Invest 1987:80:578. [341 Weintraub MS, Eisenberg S, Breslow JL. Dietary fat clearance in normal subjects is regulated by genetic variation in apolipoproteins. Eur J Clin Invest 1987;80:1571. [351 Have1 RJ. Chao Y, Windler EE et al. lsoprotein specificity in the hepatic uptake of apolipoprotein E and the pathogenesis of dysbetalipoproteinemia. Proc Nat1 Acad Sci USA 1980:77: 4349. [361 Brenninkmeijer BJ, Stuyt PH, Demacker PN et al. Catabolism of chylomicron remnants in normolipidemic subjects in relation to the apolipoprotein E phenotype. J Lipid Res 1987;28:361. 1371Bergeron N. Have1 RJ. Prolonged postprandial responses of lipids and apolipoproteins in triglyceride-rich lipoproteins of individuals expressing an apolipoprotein e allele. J Clin Invest 1996:97:65. [381 Criqui MH. Frankville DD. Barrett-Connor E, Klauber MR, Holdbrook MJ, Turner JD. Change and correlates of change in high and low density lipoprotein choelsterol after 6 years: A prospective study. Am J Epidemiol 1983;118:52. [391 Gillman MW, Cook NR, Rosner B et al. Identifying children at high risk for the development of essential hypertension. J Pediatr 1993:122:837.