Apolipoprotein E polymorphisms, dietary fat and fibre ...

Clinical research

European Heart Journal (2007) 28, 2930–2936 doi:10.1093/eurheartj/ehm482

Prevention and epidemiology

Apolipoprotein E polymorphisms, dietary fat and fibre, and serum lipids: the EPIC Norfolk study Kelvin Wu1, Richard Bowman1, Ailsa A. Welch2, Robert N. Luben2, Nick Wareham2,3, Kay-Tee Khaw2, and Sheila A. Bingham1,2* 1

MRC Dunn Human Nutrition Unit, Diet & Cancer Group, Hills Road, Cambridge CB2 2XY, UK; 2EPIC Norfolk, Department of Public Health and Primary Care, University of Cambridge, Cambridge, UK; and 3MRC Epidemiology Unit, Cambridge, UK Received 20 March 2007; revised 18 September 2007; accepted 28 September 2007; online publish-ahead-of-print 2 November 2007

KEYWORDS

Aims To investigate whether blood lipid response to dietary fat and fibre vary according to the apolipoprotein E (APOE) gene locus. Methods and results Regression analysis of intake of dietary fat and lipid fractions according to APOE gene loci was assessed by Pyrosequencing and validated with restriction fragment length polymorphism in 22 915 participants of the Norfolk arm of the European Prospective Investigation of Cancer. There were significant (P , 0.001) differences in serum lipids according to genotype, highest total and lowdensity lipoprotein (LDL) cholesterol, and lowest high-density lipoprotein and triglycerides in 14/14 individuals. There were positive associations between total and saturated fat and serum total and LDL cholesterol, and significant inverse associations (P , 0.001) between polyunsaturated fat and dietary fibre and lipid fractions overall. Associations were in the same direction for 12, 13, and 14 expressing individuals with no significant interactions between diet and genotype group on blood lipids, except in the 3% individuals expressing 12/14 (P , 0.05) in whom the associations were doubled. Conclusion In this largest study to date, ApoE gene loci status does not confer exemption from population targets to reduce dietary saturated fat and increase dietary fibre in order to reduce blood lipids and risk of coronary heart disease.

Introduction Dietary factors such as saturated fatty acids are one of the major determinants of serum cholesterol,1,2 which in turn is a risk marker of coronary heart disease (CHD). However, it has also been demonstrated that the serum lipid and lipoprotein response to dietary changes show wide variation between individuals.3,4 One reason for the variation could be genetic factors, of which apolipoprotein E (APOE) is a candidate. APOE has three major isoforms encoded by three (12, 13, and 14) alleles from two single nucleotide polymorphisms altering amino acids at sites 112 and 158.5–7 The APOE allele distributions in most Caucasian populations show similar patterns. The most common allele 13 (Cys112; Arg158) has a frequency between 0.70 and 0.85, 14 (Arg112; Arg158) is less frequent between 0.10 and 0.20, and the rarest 12 (Cys112; Cys158) between 0.05 and 0.10.8 APOE is primarily involved in the cellular uptake of lipoproteins through ligand-receptor interactions with the low-density lipoprotein (LDL) receptor and chylomicron remnant receptors.9 Through this interaction, APOE mediates the uptake and metabolism of lipoproteins and is

* Corresponding author. Tel: þ41 1223252760; fax: þ41 1223252765. E-mail address: [email protected]

thought to be a major determinant of blood lipid levels in humans.9,10 The properties of the two APOE binding domains differ according to the three isoforms, leading to functional differences.11,12 It has been well established that the presence of the APOE 14 allele is associated with increased levels of total and LDL serum cholesterol, whereas the presence of the APOE 12 allele is associated with the reverse effect.10 However, the effect of the APOE polymorphisms on the response of serum lipid levels to the intake of dietary constituents is unclear. Some intervention studies of up to 210 individuals have reported that those expressing the 14 variant are more likely to respond to increased dietary intake of fat and cholesterol, with elevation of both total and LDL serum cholesterol levels.13–17 Other studies failed to observe the same effect.18–20 Existing population based studies are also relatively small, comprising less than 1000 individuals, and the results are also inconsistent.19–24 The reported inconsistencies in results can be attributed to lack of power from small sample sizes in both these types of studies particularly in the APOE 14 group.25 We have therefore investigated relations between serum lipids and the consumption of dietary saturated, monounsaturated, and polyunsaturated fat in APOE genotypes in 22 915 free-living participants of the European Prospective

Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2007. For permissions please email: [email protected]

Downloaded from http://eurheartj.oxfordjournals.org/ by guest on May 29, 2013

Apolipoprotein E; Fat; Fibre; EPIC; Serum lipids

APOE polymorphisms, diet, and serum lipids

Investigation of Cancer Norfolk study (EPIC – Norfolk).26 In addition we have also investigated associations between genotype and intake of dietary fibre, which also mediates serum lipid levels.24

Methods Protocol

Genotype determination DNA for genotyping was extracted from an extra 9 mL blood in EDTA collected at the second health check. For those individuals who did not give blood at the second health check, DNA for genotyping was extracted from remnant red blood cell samples and buffy coats collected at the first health check.26 APOE genotype was assessed using Pyrosequencing. PCR was performed to obtain an amplicon of 185 bp covering the area of interest, amino acids 112 and 158, with primers 50 -GGG CGC GGA CAT GGA GG-30 and a 50 biotin labelled primer 50 -Biotin-CCC CGG CCT GGT ACA CTG-30 designed by Pyrosequencing Assay Design Software (Pyrosequencing AB, Sweden). A 50 mL PCR mixture was prepared with 1PCR Buffer, 2 mM MgCl2, 5% DMSO, 0.125 mM of each dNTP with 75% of dGTP replaced by dITP, 10 pmol of each primer, two units of Taq Gold and 12 ng of DNA. The annealing temperature was 628C and PCR was performed for 45 cycles in a DNA Thermal Cycler (PTC-225; MJ Research, Inc., Watertown, MA, USA). The detailed Pyrosequencing sample preparation procedure has been described elsewhere.30–33 The amplicon was denatured to isolate single stranded DNA on the beads then immersed in 15 pmol of the two sequencing primers in annealing buffer (Pyrosequencing, AB, Sweden). The two sequencing primers were 50 -GAC ATG GAG GAC GTG-30 for site 112 and 50 -CCG ATG ACC TGC AGA-30 for site 158 designed using PSQ Primer design software (Biotage). The Pyrosequencing (Pyrosequencing, AB, Sweden) machine was prepared as recommended by the manufacturer and the samples were loaded into the machine. The dispensation order for the machine was: G C T G A G C T C G C G T.

Validation and reproducibility One hundred and thirty five samples were genotyped using RFLP (restriction fragment length polymorphism)34 and the Pyrosequencer to check for reproducibility, and were fully concordant.

Dietary data A food frequency questionnaire (FFQ) consisting of 131 items was sent to all participants prior to the first health check. An in-house computer program (Compositional Analyses from Frequency Estimates), with an associated database, was developed for data entry, analysis, and conversion to nutrient intakes using databases published by the Royal Society of Chemistry.35 Participants also completed 7 day food diaries, but due to the time and resources required for coding, only data from 6 416 were available at the time of this analysis.27 The direction of change between diet and blood lipids was the same for both sets of data in these 6 416 individuals but FFQ data was used to gain sufficient power for the analysis by separate genotype.

Statistics The genotype data was combined with data from the first health check in the main EPIC database. The combined database of 24 513 individuals with genotype was then issued for statistical analysis. After excluding those subjects with missing or incomplete data on diet or lipids from the first health check, complete data were available for analysis of 22 915 participants. To increase statistical power, the six APOE genotype groups (12/ 12, 12/13, 13/13, 13/14, 14/14, and 12/14) were classified into four groups for analysis. The 13/13 genotype has long been regarded as the wild type due to the high population frequency of this group.10 It has also been well documented that the 12 allele lowers and the 14 allele increases serum total and LDL cholesterol.36 To estimate the genotype effects on serum lipid response to diet, the different deviations from the wild type were classified together, 12 expressing (12/12, 12/13), 14 expressing (13/14, 14/14) and 13/ 13. Although one study showed that the impact of the 12 allele on serum lipids to be greater than the 14 allele,37 other studies have not investigated this. Therefore the 12/14 genotype was considered as a separate group. The differences between the means and standard deviation (S.D.) were assessed using ANOVA. Categorical significant differences were accessed with the x2 test. Regressions between diet and blood lipids were adjusted for sex, age, body mass index (BMI), smoking status (current, past, and never), exercise index, use of lipid lowering drugs, total energy intake, and for non-fat components of energy intake, alcohol, and carbohydrate. Regression coefficients (b) and the standard errors (S.E.) were normalized to show the change of serum lipid for every S.D. change of dietary constituent intake. The results were expressed as two-tailed test of significance (P-value) and the 95% confidence intervals (95% C.I.). The Bonferroni correction for multiple tests was used to demarcate significant regressions to be used for further analysis by genotype.38 Each serum lipid was tested with six dietary constituents therefore the demarcating value for significance was P , 0.008. Using this demarcation, associations between polyunsaturated and saturated fat and total and LDL cholesterol and fibre and triglycerides and HDL cholesterol were investigated further. As there was little difference between results using total fat or fat as percentage of energy, the convention of analysis as percentage energy for the fat components is reported here.

Results APOE genotype and allele distribution Genotype frequencies for the 22 915 samples were: 12/12 0.6%; 12/13 12.4%; 12/14 2.6%; 13/13 58.6%; 13/14 23.5%; 14/14 2.3%. The allele distribution of 8.1% 12, 76.55% 13 and 15.35% 14 was similar to that reported for Caucasian populations in the United Kingdom, of 8% 12, 77% 13 and 15% 14.10


EPIC Norfolk is a prospective cohort of men and women recruited at age 45–75 years.26,27 In 1993 a total of 35 Norfolk medical practices agreed to participate and invited 77 630 individuals to take part in the study, of whom 30 445 completed a Health and Lifestyle questionnaire prior to a health check. Twenty five thousand six hundred and thirty nine attended this first health check from 1993 to 1997 when blood and urine samples, and data on height, weight, respiratory function, activity, anthropometry and blood pressure, were collected by trained nurses.26–28 All 30 445 participants have been followed up for their health status and as a part of follow up they were invited back for a second health check at the beginning of 1998.26 Fifteen thousand seven hundred and eighty three individuals attended, of whom 15 025 had attended the first health check. Permission for the study was obtained from The Norfolk and Norwich Hospital Ethics Committee. Forty two millilitre of non-fasting blood was collected at both health checks in EDTA, citrated and plain monovettes, and separated into plasma, serum, buffy coats and red blood cells as described previously.23 Serum lipids were analysed for total cholesterol, high-density lipoprotein (HDL) cholesterol and triglyceride on an RA-1000 (Bayer Diagnostics, Basingstoke). LDL cholesterol was calculated using the Friedewald formula;29 when serum TG exceeded 4 mmol/L, LDL cholesterol was not calculated.

2931

2932

APOE genotype and coronary heart disease risk factors

Associations between dietary constituents and serum lipids Table 2 shows that there were highly significant associations between serum lipids with nearly all the tested dietary constituents. Mean intakes in the 22 915 participants were as follows: total fat 32.45 (5.83)% of energy intake, monounsaturated fat 11.40 (2.34), polyunsaturated fat 6.14 (1.98), saturated fat 12.47 (3.31)% of energy intake, cholesterol 0.28 (0.12) g/day and fibre 18.50 (6.56) g/day. Results were in a similar direction but somewhat stronger in females compared with males (data not shown). Fat and monounsaturated fat and cholesterol intake were positively associated with total cholesterol and LDL cholesterol, whereas polyunsaturated fat and fibre were inversely associated. Fibre was also inversely associated with triglycerides (P , 0.001) and monounsaturated fat positively associated with triglycerides (P , 0.001). Associations with P , 0.008 were analysed further by genotype and assessed for differences in the four expressing genotype groups. Table 3 shows that most associations in Table 2 remained significant in the 13/13 group. Associations were in the same direction for 12 and 14 groups with no evidence for differences between coefficients (Z ¼ .p0.05). b coefficients were of greater magnitude in the 12/14 group, with significant differences between the 13/13, 14 expressing and the 12/14 group for total serum cholesterol and total fat (% energy) (Z¼2.37, 2.29 respectively, P , 0.05) and also between the 13/13 and the 12/14 group for total serum cholesterol and monounsaturated fat (% energy) (Z ¼ 2.99, P , 0.05). There were no other significant diet lipid differences between other genotype groups. There were significant (P , 0.014) but small differences in mean intakes as percentage energy of total fat and saturated fat between the different genotype groups that was mainly

attributable to the 137 participants who were 12/12, and a small difference (P ¼ 0.012) in cholesterol but no differences in dietary fibre, or monounsaturated fat as a percentage of energy (data not shown).

Discussion There were significant differences in blood lipids in mean serum total and LDL cholesterol between the 12, 13/13, and 14 genotype groups, as has been shown previously.10,39 However the present study with approximately 23 000 participants was much larger than previous ones and thus more individuals in the rarer groups of 14/14, 12/12 and of 12/14 were investigated. The mean total, LDL and HDL cholesterol of the 12/14 group was consistently intermediate between the mean levels of the 12/13 and 13/13 groups, suggesting that the up regulatory effect of the 12 genotype is more acute than the down regulatory effect of the 14 genotype, supporting the suggestion that 12 allele has phenotypic dominance.37 The present study is cross-sectional with consequent limitations concerning the effect of change on blood lipids and therefore risk of CHD. However, it is notable that highly significant associations between diet and serum lipid fractions in this large cross-sectional prospective study were seen in the directions predicted from carefully controlled intervention studies on blood lipids.1,2 The positive association between monounsaturated fat and serum total and LDL cholesterol was unexpected, as some studies suggest monounsaturated fat has no effect or a negative effect on serum LDL cholesterol.1,40,41 However, positive associations between oleic acid and LDL and total cholesterol (and with total and saturated fat) have been shown in a previous but smaller cross-sectional analysis in women in the Framingham study.42 As predicted from the Keys and Hegsted equations, the magnitude of the inverse association with polyunsaturated fat on serum LDL cholesterol was approximately half that of the positive effect of saturated fat.1,2 However, every 3% of energy as saturated fat increased LDL cholesterol by only about 1.52% (b ¼ 0.06, Table 2, of the average 3.96 mmol/L, Table 1). From the Keys equation, a change of 5% would be expected.2 This smaller effect than predicted is likely due to known attenuation in dietary assessment introduced by the use of a FFQ rather than more detailed methods.43 Insufficient dietary diaries were available for analysis of lipid nutrient associations by genotype, but the directions of change were similar using both methods. Despite a number of intervention and association studies, there is controversy as to whether the association of ApoE phenotypes with serum lipids is related to dietary factors.13–25 Inconsistencies in results have been attributed to differences in study design and small sample sizes leading to lack of statistical power, particularly in the 14 groups.25 In this study of nearly 23 000 individuals, all genotype groups increased serum LDL cholesterol in response to dietary saturated fat, with highly significant b coefficients for every genotype group, but there were no significant difference in the magnitude of the associations according to genotype. The magnitude of responses for fat were up to six-fold greater in the 12/14 groups and there was a significant difference (P , 0.05) between the regression coefficients for total and monounsaturated fat when compared


Mean serum lipids and CHD risk factors in different genotype groups were very similar in the male and female populations. There was no significant genotype distribution difference between the two sexes (x2 P ¼ 0.73), so basic statistics with the sexes combined for different genotype groups are shown Table 1. Sex, age, BMI, systolic blood pressure, smoking status, use of hypertension lowering drugs, and physical activity index were not significantly different between the different genotype groups. The percentages of individuals prescribed lipid lowering drugs at baseline was low, 1.6% overall, with a significant difference between the genotype groups, lowest in the individuals carrying the 12 allele, and highest for individuals carrying the 14 allele (P , 0.001). Apart from the 12/14 group, mean serum total and LDL cholesterol concentrations were lowest in the 12 expressing groups (P , 0.001) and highest in the 14 expressing groups (P , 0.001). The mean serum HDL cholesterol was highest in the 12 expressing groups and lowest in the 14 expressing group (P , 0.001). The mean serum triglyceride was lowest in the 13 group and highest in the 12 group (P , 0.001). In the 12/14 group the mean total, LDL and HDL cholesterol was intermediate between groups 12/13 and 13/13 (Table 1 and Figure 1).

K. Wu et al.

All (n ¼ 22 915) Sex Age – Mean (SD) BMI (kg/m2) – Mean (SD) SBP (mmHg) – Mean (SD) Chol (mmol/L) – Mean (SD) LDL (mmol/L) – Mean (SD) HDL (mmol/L) – Mean (SD) Tg (mmol/L) – Mean (SD) Smoking (%)

Physical activity index

Hypotensive drugs Hypolipidaemic drugs

M (%) F (%)

Current Past Never 1 (%) 2 (%) 3 (%) 4 (%) Yes (%) Yes (%)

12/12 (n ¼ 137, 0.6%)

13/12 (n ¼ 2842, 12.4%)

13/13 (n ¼ 13 428, 56.6%)

14/13 (n ¼ 5385, 23.5%)

14/14 (n ¼ 527, 2.3%)

12/14 (n ¼ 596, 2.6%)

P-value

45.85 54.15 58.71 (9.29) 26.31 (3.85)

46.76 53.24 58.83 (9.65) 25.96 (3.43)

46.40 53.62 58.73 (9.33) 26.45 (3.96)

45.88 54.12 58.72 (9.30) 26.32 (3.82)

45.23 54.77 58.72 (9.30) 26.27 (3.85)

48.40 51.60 58.14 (9.07) 26.04 (4.03)

45.69 54.31 58.26 (9.16) 26.25 (3.74)

0.733 0.733 0.571 0.127

135.42 (18.32)

138.14 (19.61)

135.25 (17.98)

135.58 (18.35)

135.18 (18.40)

134.93 (17.51)

134.61 (18.91)

0.235

6.18 (1.17)

5.77 (1.44)

5.74 (1.10)

6.19 (1.14)

6.36 (1.17)

6.59 (1.32)

6.05 (1.23)

,0.001

3.96 (1.03)

3.24 (1.31)

3.44 (0.94)

3.99 (1.00)

4.17 (1.05)

4.38 (1.14)

3.74 (1.00)

,0.001

1.42 (0.43)

1.50 (0.55)

1.46 (0.43)

1.42 (0.43)

1.39 (0.41)

1.34 (0.39)

1.43 (0.44)

,0.001

1.82 (1.11)

2.30 (1.33)

1.86 (1.12)

1.77 (1.05)

1.86 (1.14)

2.01 (1.57)

1.98 (1.33)

,0.001

11.54 42.52 45.93 30.27 28.62 22.79 18.32 18.4 1.6

10.87 46.38 42.75 30.22 27.34 24.46 17.99 25.0 1.1

11.55 42.36 46.09 30.36 28.34 22.42 18.88 18.2 0.5

11.81 42.36 45.83 30.41 28.48 22.60 18.51 17.9 1.4

11.06 42.85 46.09 29.94 29.14 23.29 17.63 19.1 2.3

9.34 42.55 48.11 29.08 27.84 22.16 20.92 18.7 2.7

11.92 43.16 44.93 30.83 29.39 24.28 15.50 21.8 2.0


Table 1 The number (n) and percentage (%) in each genotype group and mean and standard deviation (S.D.) of some coronary heart disease risk factors and serum lipids

0.778 0.778 0.778 0.774 0.774 0.774 0.774 0.203 ,0.001

BMI, Body mass index; SBP, Systolic blood pressure; Chol, Cholesterol; LDL, Low-density lipoprotein; HDL, High-density lipoprotein; Tg, Triglycerides.

2933


2934

K. Wu et al.

Figure 1

Mean serum lipid concentration variations between the different genotype groups as shown numerically in Table 1.

Serum lipid

Cholesterol (mmol/L)

LDL cholesterol (mmol/L)

HDL cholesterol (mmol/L)

Triglycerides (mmol/L)

Dietary constituent

Total fat (g/day) Monounsaturated fat (g/day) Polyunsaturated fat (g/day) Saturated fat (g/day) Cholesterol (g/day) Fibre (g/day) Total fat (g/day) Monounsaturated fat (g/day) Polyunsaturated fat (g/day) Saturated fat (g/day) Cholesterol (g/day) Fibre (g/day) Total fat (g/day) Monounsaturated fat (g/day) Polyunsaturated fat (g/day) Saturated fat (g/day) Cholesterol (g/day) Fibre (g/day) Total fat (g/day) Monounsaturated fat (g/day) Polyunsaturated fat (g/day) Saturated fat (g/day) Cholesterol (g/day) Fibre (g/day)

b

0.139 0.108 20.069 0.118 0.047 20.033 0.123 0.101 20.054 0.095 0.045 20.034 20.001 20.027 20.009 0.016 0.002 0.019 0.023 0.070 20.010 0.002 0.009 20.041

S.E.

0.037 0.027 0.011 0.017 0.014 0.010 0.034 0.025 0.010 0.016 0.013 0.009 0.013 0.009 0.004 0.006 0.005 0.003 0.034 0.025 0.010 0.016 0.013 0.009

P-value

,0.001 ,0.001 ,0.001 ,0.001 0.001 0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.928 0.003 0.024 0.007 0.607 ,0.001 0.500 0.005 0.328 0.899 0.490 ,0.001

95% Confidence interval Lower

Upper

0.066 0.054 20.091 0.084 0.020 20.052 0.056 0.052 20.074 0.064 0.020 20.052 20.026 20.045 20.016 0.004 20.007 0.012 20.044 0.021 20.030 20.029 20.016 20.059

0.212 0.161 20.047 0.152 0.075 20.014 0.190 0.150 20.034 0.126 0.070 20.016 0.024 20.009 20.001 0.028 0.012 0.025 0.090 0.120 0.010 0.033 0.034 20.023

The sample size n ¼ 22 915. All analyses adjusted for sex, age, BMI, smoking status, exercise index, hypolipidaemic drugs, total energy intake, alcohol intake, and carbohydrate intake.

with those of the 13/13 individuals. Due to the comparatively small size of the 12/14 group, it is possible that this difference occurred by chance. Fibre was inversely associated with all blood lipid fractions, except for a positive association with serum HDL. The inverse associations between fibre and serum total and LDL cholesterol are supported by numerous studies that show fibre intake to consistently reduce serum total

and LDL cholesterol concentrations.44 In intervention studies, fibre has not previously been shown to reduce serum triglyceride levels.44 However, in the Framingham study of women, there were cross-sectional inverse associations between fibre and serum triglycerides.45 The effect of fibre on cholesterol is attributed to fermentation of fibre polysaccharides in the large bowel and the consequent absorption of the short chain fatty acid propionate.46,47 As


Table 2 Regression coefficient (b), standard error (S.E.), P-value, and 95% confidence interval for standardized dietary constituents vs. serum lipids

this effect is independent of ApoE, the lack of differences according to genotype on the inverse effect of fibre on blood levels is as expected. Despite the large size of this study, the largest to date, there was little evidence that, apart from the small (3%) of individuals who are of 12/14 genotype, different APOE genotypes respond differently to increased dietary saturated or total fat intake. ApoE gene loci status does not confer exemption from targets to reduce dietary saturated fat and increase dietary fibre in order to reduce blood lipids and risk of CHD in populations as a whole.

0.305 (0.105, 0.506), 0.003 0.334 (0.159, 0.510), ,0.001 20.094 (20.200, 0.013), 0.084 0.209 (0.066, 0.352), 0.004 20.016 (20.152, 0.120), 0.818 0.194 (0.025, 0.025), 0.025 0.158 (0.004, 0.311), 0.004 20.065 (20.154, 0.023), 0.147 0.162 (0.043, 0.282), 0.008 0.011 (20.104, 0.127), 0.845 20.003 (20.062, 0.056), 0.917 20.024 (20.058, 0.010), 0.168 20.003 (20.050, 0.043), 0.883 0.005 (20.039, 0.049), 0.830 0.374 (0.188, 0.560), ,0.001 20.100 (20.244, 0.044), 0.172

12/14 (n ¼ 596)

2935

Tg

HDL

LDL

Tot fat en (%), Total fat energy (%); MUFA en (%), Monounsaturated fat energy (%); PUFA en (%), Polyunsaturated fat energy (%); Sat fat en (%), Saturated fat energy (%).

0.062 (0.005, 0.119), 0.033 0.046 (20.005, 0.096), 0.075 20.055 (20.086, 20.024), 0.001 0.091 (0.050, 0.131), ,0.001 20.046 (20.085, 20.008), 0.018 0.053 (0.001, 0.106), 0.050 0.035 (20.011, 0.081), 0.139 20.053 (20.082, 20.024), ,0.001 0.085 (0.048, 0.122), ,0.001 20.044 (20.080, 20.009), 0.014 20.013 (20.029, 0.003), 0.120 20.007 (20.017, 0.003), 0.177 0.001 (20.012, 0.014), 0.863 0.021 (0.009, 0.033), 0.001 0.034 (20.015, 0.083), 0.177 20.051 (20.088, 20.014), 0.007 0.059 (0.022, 0.095), 0.002 0.061 (0.029, 0.093), ,0.001 20.044 (20.064, 20.024), ,0.001 0.068 (0.042, 0.094), ,0.001 20.036 (20.060, 20.011), 0.005 0.045 (0.012, 0.079), 0.007 0.055 (0.026, 0.084), ,0.001 20.032 (20.051, 20.014), ,0.001 0.049 (0.026, 0.072), ,0.001 20.037 (20.059, 20.015), 0.001 20.011 (20.022, 0.001), 0.066 20.006 (20.013, 0.001), 0.122 0.012 (0.003, 0.021), 0.009 0.021 (0.012, 0.030), ,0.001 0.036 (0.008, 0.065), 0.013 20.036 (20.058, 20.014), 0.001 0.047 (20.028, 0.123), 0.126 0.012 (20.055, 0.078), 0.729 20.051 (20.094, 20.007), 0.022 0.080 (0.027, 0.134), 0.0003 20.025 (20.076, 0.026), 0.330 0.058 (20.009, 0.125), 0.090 0.034 (20.025, 0.093), 0.255 20.042 (20.081, 20.003), 0.034 0.075 (0.027, 0.123), 0.002 20.027 (20.072, 0.018), 0.243 20.022 (20.046, 0.002), 0.068 20.002 (20.017, 0.014), 0.843 0.013 (20.007, 0.032), 0.195 0.010 (20.008, 0.029), 0.266 20.020 (20.085, 0.044), 0.542 20.010 (20.059, 0.039), 0.691 Tot fat en (%) MUFA en (%) PUFA en (%) Sat fat en (%) Fibre (g/day) Tot fat en (%) MUFA en (%) PUFA en (%) Sat fat en (%) Fibre (g/day) MUFA en (%) PUFA en (%) Sat fat en (%) Fibre (g/day) MUFA en (%) Fibre (g/day) Chol

12 expressing (n ¼ 2979)

13/13 (n ¼ 13 428)

14 expressing (n ¼ 5912)

Conflict of interest: none declared.

References 1. Hegsted DM, Ausman LM, Johnson JA, Dallal GE. Dietary fat and serum lipids: an evaluation of the experimental data. Am J Clin Nutr 1993; 57:875–883. 2. Keys A, Anderson JT, Grande F. Prediction of serum-cholesterol responses of man to changes in fats in the diet. Lancet 1957;273:959–966. 3. Beynen AC, Katan MB, Van Zutphen LF. Hypo- and hyperresponders: individual differences in the response of serum cholesterol concentration to changes in diet. Adv Lipid Res 1987;22:115–171. 4. Katan MB, Beynen AC, de Vries JHM, Nobels A. Existence of consistent hypo- and hyperresponders to dietary cholesterol in man. Am J Epidemiol 1986;123:221–234. 5. Utermann G, Langenbeck U, Beisiegel U, Weber W. Genetics of the apolipoprotein E system in man. Am J Hum Genet 1980;32:339–347. 6. Utermann G, Steinmetz A, Weber W. Genetic control of human apolipoprotein E polymorphism: comparison of one- and two-dimensional techniques of isoprotein analysis. Hum Genet 1982;60:344–351. 7. Zannis VI, Breslow JL. Human very low density lipoprotein apolipoprotein E isoprotein polymorphism is explained by genetic variation and posttranslational modification. Biochemistry 1981;20:1033–1041. 8. Loktionov A, Vorster H, O’Neill IK, Nell T, Bingham SA, Runswick SA, Cummings JH. Apolipoprotein E and methylenetetrahydrofolate reductase genetic polymorphisms in relation to other risk factors for cardiovascular disease in UK Caucasians and Black South Africans. Atherosclerosis 1999;145:125–135. 9. Mahley RW. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science 1988;240:622–630. 10. Mahley RW, Rall SC Jr. Apolipoprotein E: far more than a lipid transport protein. Annu Rev Genomics Hum Genet 2000;1:507–537. 11. Weisgraber KH. Apolipoprotein E: structure-function relationships. Adv Protein Chem 1994;45:249–302. 12. Weisgraber KH, Innerarity TL, Rall SC Jr, Mahley RW. Apolipoprotein E: receptor binding properties. Adv Exp Med Biol 1985;183:159–171. 13. 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. 14. Gylling H, Miettinen TA. Cholesterol absorption and synthesis related to low density lipoprotein metabolism during varying cholesterol intake in men with different apoE phenotypes. J Lipid Res 1992;33:1361–1371. 15. Sarkkinen E, Korhonen M, Erkkila A, Ebeling T, Uusitupa M. Effect of apolipoprotein E polymorphism on serum lipid response to the separate modification of dietary fat and dietary cholesterol. Am J Clin Nutr 1998;68:1215–1222. 16. 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. 17. Weggemans RM, Zock PL, Ordovas JM, Pedro-Botet J, Katan MB. Apoprotein E genotype and the response of serum cholesterol to dietary fat, cholesterol and cafestol. Atherosclerosis 2001;154:547–555. 18. Pasagian-Macaulay A, Aston CE, Ferrell RE, McAllister AE, Wing RR, Kuller LH. A dietary and behavioral intervention designed to lower coronary heart disease. Risk factors are unaffected by variation at the APOE gene locus. Atherosclerosis 1997;132:221–227. 19. Savolainen MJ, Rantala M, Kervinen K, Jarvi L, Suvanto K, Rantala T, Kesaniemi YA. Magnitude of dietary effects on plasma cholesterol concentration: role of sex and apolipoprotein E phenotype. Atherosclerosis 1991;86:145–152.


Table 3 Standardized regression coefficients (b) and 95% confidence intervals and P-values for the change in serum lipids for every SD change in dietary constituents


2936

33. Ronaghi M, Karamohamed S, Pettersson B, Uhlen M, Nyren P. Real-time DNA sequencing using detection of pyrophosphate release. Anal Biochem 1996;242:84–89. 34. Loktionov A, Bingham SA, Vorster H, Jerling JC, Runswick SA, Cummings JH. Apolipoprotein E genotype modulates the effect of black tea drinking on blood lipids and blood coagulation factors: a pilot study. Br J Nutr 1998;79:133–139. 35. Welch AA, Luben R, Khaw KT, Bingham SA. The CAFE computer program for nutritional analysis of the EPIC-Norfolk food frequency questionnaire and identification of extreme nutrient values. J Hum Nutr Diet 2005;18: 99–116. 36. Sing CF, Davignon J. Role of the apolipoprotein E polymorphism in determining normal plasma lipid and lipoprotein variation. Am J Hum Genet 1985;37:268–285. 37. Bohnet K, Regis-Bailly A, Vincent-Viry M, Schlenck A, Gueguen R, Siest G, Visvikis S. Apolipoprotein E genotype epsilon 4/epsilon 2 in the STANISLAS Cohort Study – dominance of the epsilon 2 allele? Ann Hum Genet 1996; 60:509–516. 38. Ottenbacher KJ. Quantitative evaluation of multiplicity in epidemiology and public health research. Am J Epidemiol 1998;147:615–619. 39. 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. 40. Kris-Etherton PM. A new role for diet in reducing the incidence of cardiovascular disease: evidence from recent studies. Curr Atheroscler Rep 1999;1:185–187. 41. Lapointe A, Balk EM, Lichtenstein AH. Gender differences in plasma lipid response to dietary fat. Nutr Rev 2006;64:234–249. 42. Millen BE, Franz MM, Quatromoni PA, Gagnon DR, Sonnenberg LM, Ordovas JM, Wilson PW, Schaefer EJ, Cupples LA. Diet and plasma lipids in women 1. Macronutrients and plasma total and LDL cholesterol: the Framingham Nutrition Studies. J Clin Epidemiol 1996;49:657–663. 43. Bingham SA, Luben R, Welch A, Wareham N, Khaw KT, Day N. Are imprecise methods obscuring a relation between fat and breast cancer? Lancet 2003;362:212–214. 44. Truswell AS. Dietary fibre and plasma lipids. Eur J Clin Nutr 1995; 49(Suppl. 3):S105–S109. 45. Sonnenberg LM, Quatromoni PA, Gagnon DR, Cupples LA, Franz MM, Ordovas JM, Wilson PW, Schaefer EJ, Millen BE. Diet and plasma lipids in women. II. Macronutrients and plasma triglycerides, high-density lipoprotein, and the ratio of total to high-density lipoprotein cholesterol in women: the Framingham Nutrition Studies. J Clin Epidemiol 1996;49: 665–672. 46. Boila RJ, Salomons MO, Milligan LP, Aherne FX. The Effect of Dietary Propionic-Acid on Cholesterol-Synthesis in Swine. Nutr Rep Int 1981;23: 1113–1121. 47. Venter CS, Vorster HH, Cummings JH. Effects of dietary propionate on carbohydrate and lipid metabolism in healthy volunteers. Am J Gastroenterol 1990;85:549–553.


20. Salminen M, Lehtima ¨ki T, Fan Y, Vahlberg T, Kivela ¨Sirkka- Liisa. Apolipoprotein E polymorphism and changes in serum lipids during a familybased counselling intervention. Public Health Nutr 2006;9:859–865. ˚ kerblom HK, Ehnholm C, Ro 21. Lehtima ¨ki T, Moilanen T, Porkka K, A ¨nnemaa T, Ra ¨sa ¨nen L, Laippala P, Viikari J. Association of serum lipids with apolipoprotein E phenotype and linoleic acid content of serum cholesteryl esters in the cardiovascular risk in young Finns study. Nutr Metab Cardiovasc Dis 1993;3:61–65. ˚ kerblom HK, Ro 22. Lehtima ¨ki T, Moilanen T, Porkka K, A ¨nnemaa T, Ra ¨sa ¨nen L, Viikari J, Nikkari T. Association between serum lipids and apolipoprotein E phenotype is influenced by diet in a population-based sample of freeliving children and young adults: the cardiovascular risk in young Finns study. J Lipid Res 1995;36:653–661. 23. Loktionov A, Scollen S, McKeown N, Bingham SA. Gene-nutrient interactions: dietary behaviour associated with high coronary heart disease risk particularly affects serum LDL cholesterol in apolipoprotein E epsilon4-carrying free-living individuals. Br J Nutr 2000;84: 885–890. 24. 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;68:87–95. 25. Masson LF, McNeill G, Avenell A. Genetic variation and the lipid response to dietary intervention: a systematic review. Am J Clin Nutr 2003;77: 1098–1111. 26. Day N, Oakes S, Luben R, Khaw KT, Bingham S, Welch A, Wareham N. EPIC-Norfolk: study design and characteristics of the cohort. European Prospective Investigation of Cancer. Br J Cancer 1999;80(Suppl. 1):95–103. 27. Bingham SA, Welch AA, McTaggart A, Mulligan AA, Runswick SA, Luben R, Oakes S, Khaw KT, Wareham N, Day NE. Nutritional methods in the European Prospective Investigation of Cancer in Norfolk. Public Health Nutr 2001;4:847–858. 28. Wareham NJ, Jakes RW, Rennie KL, Schuit J, Mitchell J, Hennings S, Day NE. Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Public Health Nutr 2003;6:407–413. 29. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18:499–502. 30. Ahmadian A, Gharizadeh B, Gustafsson AC, Sterky F, Nyren P, Uhlen M, Lundeberg J. Single-nucleotide polymorphism analysis by pyrosequencing. Anal Biochem 2000;280:103–110. 31. Alderborn A, Kristofferson A, Hammerling U. Determination of singlenucleotide polymorphisms by real-time pyrophosphate DNA sequencing. Genome Res 2000;10:1249–1258. 32. Nyren P, Karamohamed S, Ronaghi M. Detection of single-base changes using a bioluminometric primer extension assay. Anal Biochem 1997; 244:367–373.

K. Wu et al.