Combined effects of dietary fat and birth weight on serum cholesterol concentrations: the Hertfordshire Cohort Study 1 3

Transcription

1 Combined effects of dietary fat and birth weight on serum cholesterol concentrations: the Hertfordshire Cohort Study 1 3 Sian M Robinson, Sue F Batelaan, Holly E Syddall, Avan Aihie Sayer, Elaine M Dennison, Helen J Martin, David J Barker, Cyrus Cooper, and the Hertfordshire Cohort Study Group ABSTRACT Background: Blood cholesterol responses to the manipulation of dietary fat vary widely between persons. Although epidemiologic evidence suggests that prenatal growth and nutrition influence adult cholesterol homeostasis, whether prenatal growth modifies the association between dietary fat intake and serum cholesterol concentration in adults is unknown. Objective: The aim was to examine the relation between fat intake and serum cholesterol concentrations in men and women whose birth weights were known. Design: We studied a cohort of men and women aged y. Diet was assessed with a food-frequency questionnaire. Total, HDL-, and LDL-cholesterol concentrations and the ratio of HDL to LDL cholesterol were measured in fasting blood samples from 574 men and 562 women who did not have coronary heart disease. Results: Total and saturated fat intakes were not associated with serum cholesterol concentrations in men or women. However, subdivision by birth weight showed associations in men but not in women. High intakes of total and saturated fat were associated with reduced HDL-cholesterol concentrations in men with birth weights 3.2 kg (7 lb) but not in men with higher birth weights. Similar effects on the HDL-to-LDL cholesterol ratio were observed (P for interaction 0.02 for total fat and 0.01 for saturated fat). When 32 men taking cholesterol-lowering medication were excluded, the interactions were strengthened (P and 0.006, respectively). Conclusion: The adverse effects of high intakes of total and saturated fat on serum cholesterol concentrations in men may be confined to those with lower birth weights. Am J Clin Nutr 2006;84: KEY WORDS variability Cholesterol, fat intake, birth weight, individual INTRODUCTION Elevated serum cholesterol concentrations, which are known to predict coronary heart disease, are affected by the amount and composition of fat in the diet (1). However, blood lipid responses to the manipulation of dietary fat vary significantly between persons (2, 3). Although genetic variation contributes to this heterogeneity, much of the variation in serum cholesterol concentrations between persons remains unexplained (3). Epidemiologic studies have shown that persons with low birth weights are at an increased risk of coronary heart disease in adult life (4 9). This observation may be explained, in part, by an association between low birth weight and altered cholesterol metabolism. Disturbed intrauterine growth is associated with immediate changes in cholesterol concentrations, and follow-up studies have shown that serum cholesterol concentrations track from the age of 6 mo into adult life (10). One mechanism that may explain the link between prenatal growth and adult disease is a permanent change in gene expression in response to the early environment (11); it is possible that the variability in individual cholesterol responses to manipulation of dietary fat could originate through changes in gene expression that occurred during early development when growth was restricted. Persons with low birth weights may thus respond to high fat intakes in ways that make them more vulnerable to coronary heart disease. Direct experimental evidence supports this hypothesis (12). In guinea pigs, modest restriction of the maternal diet during pregnancy led to an altered response in the adult offspring to a high cholesterol diet. In male offspring whose mothers diets were restricted during pregnancy, high cholesterol feeding for a 6-wk period led to higher plasma total cholesterol concentrations than were found in male offspring whose mothers had been given unlimited food. When the offspring were divided into those with birth weights above or below the median, plasma total cholesterol concentrations before cholesterol feeding did not differ in the 2 groups. After cholesterol feeding, however, the concentrations were 31% higher in the group with lower birth weights. No similar effects were observed in the female offspring. This led to the conclusion that maternal undernutrition and altered prenatal growth permanently change cholesterol homeostasis in males. To investigate whether these effects are evident in humans, we examined the association between intakes of total and saturated fat and serum cholesterol concentrations in a cohort of 574 men and 562 women whose birth weights were recorded. 1 From the Medical Research Council Epidemiology Resource Centre (SMR, SFB, HES, AAS, EMD, HJM, and CC) and the Developmental Origins of Adult Health and Disease Centre (DJPB), University of Southampton, Southampton, United Kingdom. 2 Supported by the Medical Research Council, United Kingdom. 3 Reprints not available. Address correspondence to SM Robinson, MRC Epidemiology Resource Centre, University of Southampton, Southampton General Hospital, Southampton SO16 6YD, United Kingdom. smr@mrc.soton.ac.uk. Received October 11, Accepted for publication March 8, Am J Clin Nutr 2006;84: Printed in USA American Society for Nutrition 237

2 238 ROBINSON ET AL SUBJECTS AND METHODS Design and study population From 1911 to 1948, midwives made detailed records, which included birth weight, on all infants born in the county of Hertfordshire, United Kingdom. These records have been described previously (4, 13). In 1998, 1760 men and 1447 women born between 1931 and 1939 and still living in East Hertfordshire were traced. General Practitioners gave permission for us to write to 1397 (79%) of the men and 1364 (94%) of the women. Seven hundred sixty-eight (55%) of the men and 714 (52%) of the women agreed to be interviewed at home. The mean ( SD) birth weights of the persons who were interviewed were similar to those who were not interviewed: compared with kg for the men and compared with kg for the women. Gestational age at birth was not recorded. During the interview, diet was assessed by a trained research nurse and the details of the subject s medical and social background were obtained. The subjects were then invited to attend a clinic after fasting overnight, where height and weight were measured and blood samples were taken. Complete dietary data were available for 737 men and 675 women who attended the clinic. Coronary heart disease was defined as the presence of 1of the following: angina, as defined by the Rose-World Health Organization chest pain questionnaire (14); electrocardiography Minnesota codes 1 1, 1 2 (Q and QS codes) (15); or a history of surgery for coronary revascularization, either coronary artery bypass graft or coronary angioplasty. One hundred fourteen men and 77 women had coronary heart disease and were excluded from the analysis, which left 623 men and 598 women for the analysis. The study was approved by the East Hertfordshire Local Research Ethics Committee. All subjects gave written informed consent. Dietary assessment Diet was assessed by using an administered food-frequency questionnaire (FFQ) that was based on the European Prospective Investigation of Cancer questionnaire (16). The FFQ included 129 foods and food groups and was used to assess the average frequency of consumption of the listed foods over the 3-mo period before the home interview. The frequencies of other foods that were not listed on the FFQ were also recorded if consumed 1 time/wk. Standard portion sizes were allocated to each food (17) apart from milk and sugar for which daily quantities consumed were recorded. For all spreading fats and cooking oils used, individual brand names were collected so that the profiles of saturated, monounsaturated and polyunsaturated fat in these products could be estimated. Nutrient intakes were calculated by multiplying the frequency of consumption of a portion of each food by its nutrient content (18). Dietary interviews took place between 1998 and Serum cholesterol measurements Serum total cholesterol and HDL cholesterol were measured with standard enzymatic methods (19 21). LDL-cholesterol concentrations were derived with the Friedewald-Fredrickson formula (22). Fasting serum HDL- and LDL-cholesterol concentrations were available for 574 of the men and 562 of the women who attended clinic and who had no coronary heart disease. TABLE 1 Characteristics of the men and women 1 Statistical analysis Variables were summarized with the use of means and SDs or frequencies and proportions, as appropriate. Multiple linear regression was used to explore the main and interactive effects of birth weight and dietary fat intakes on cholesterol concentrations. Sex-specific SD scores for birth weight and each dietary characteristic were used as predictor variables in the regression models. The distributions of total and saturated fat intakes, body mass index [BMI; calculated as weight (in kg)/height 2 (in m)], HDL cholesterol, and the ratio of HDL to LDL cholesterol were positively skewed and were log e transformed for statistical analysis. Accordingly, the means and SDs for these variables are geometric. Analyses were based on continuously distributed variables throughout; sex-specific thirds of birth weights and adult fat intakes are used in the tables. Analyses were conducted with the STATA statistical software package version 8 (Stata Corp, College Station, TX). RESULTS Men (n 574) Women (n 562) Birth weight (kg) BMI (kg/m 2 ) Alcohol use 4 Nondrinker [n (%)] 21 (3.7) 108 (19.2) Light drinker [n (%)] 286 (49.8) 366 (65.1) Moderate drinker [n (%)] 125 (21.8) 68 (12.1) Heavy drinker [n (%)] 142 (24.7) 20 (3.6) Daily energy intake (kcal) Daily fat intake (g) Percentage calories from fat (%) Daily saturated fat intake (g) Percentage calories from saturated fat (%) 2 Total cholesterol (mmol/l) HDL cholesterol (mmol/l) LDL cholesterol (mmol/l) HDL:LDL cholesterol No difference in BMI was observed between men and women (P 0.40, t test), but all other variables were significantly different (P 0.001; t test for continuous variables, chi-square test for proportion of nondrinkers, and Mann-Whitney U test for number of alcohol units. 2 Values are arithmetic x SD. 3 Values are geometric x SD. 4 Categorized according to weekly consumption of alcohol units. Light drinkers: 10 units for men, 7 units for women; moderate drinkers: units for men, 8 14 units for women; heavy drinkers: 22 units for men, 15 units for women. Subjects The age of the men and women ranged from 59 to 71 y [mean ( SD) age: y for the men, y for the women]. Fifty-eight percent of the men and 61% of the women were in the manual social classes. The mean birth weight, current BMI, weekly alcohol consumption, daily energy and fat intakes, and serum cholesterol concentrations of the men and women studied are shown in Table 1.

3 BIRTH WEIGHT, DIETARY FAT, AND SERUM CHOLESTEROL 239 TABLE 2 Correlations between birth weight, dietary fat intakes, and serum cholesterol concentrations in 574 men and 562 women 1 Total cholesterol LDL cholesterol HDL cholesterol HDL:LDL cholesterol mmol/l mmol/l mmol/l Men Birth weight (kg) 0.01 ( 0.07, 0.10) 0.01 ( 0.07, 0.09) 0.06 ( 0.03, 0.14) 0.03 ( 0.05, 0.11) Fat intake (g) 0.00 ( 0.08, 0.09) 0.01 ( 0.09, 0.07) 0.02 ( 0.06, 0.10) 0.02 ( 0.06, 0.10) Saturated fat intake (g) 0.00 ( 0.08, 0.08) 0.00 ( 0.08, 0.08) 0.02 ( 0.06, 0.10) 0.02 ( 0.06, 0.10) Women Birth weight (kg) 0.04 ( 0.12, 0.04) 0.04 ( 0.12, 0.04) 0.03 ( 0.05, 0.11) 0.05 ( 0.04, 0.12) Fat intake (g) 0.05 ( 0.13, 0.04) 0.05 ( 0.13, 0.03) 0.04 ( 0.05, 0.12) 0.06 ( 0.03, 0.13) Saturated fat intake (g) 0.03 ( 0.11, 0.06) 0.02 ( 0.10, 0.06) 0.00 ( 0.08, 0.09) 0.01 ( 0.07, 0.09) 1 Unadjusted Pearson correlation coefficients (95% CIs). No correlation was statistically significant. Univariate associations Total, HDL, and LDL cholesterol and the ratio of HDL- to LDL-cholesterol concentrations were not significantly related to birth weight or to total fat or saturated fat intakes (Table 2). Adjustment for the effects of current BMI, age, and alcohol intake showed a small positive association between birth weight and serum HDL-cholesterol concentration in men (r 0.09, P 0.04), but did not influence the lack of association between birth weight and total cholesterol, LDL cholesterol, and the ratio of HDL- to LDL-cholesterol concentration in men or between fat intakes and any of the cholesterol outcomes considered in men and women. Interactive effects The associations between total and saturated fat intakes and serum cholesterol concentrations were examined in men and women of different birth weights. As shown in Table 3, the men and women were subdivided into 9 groups according to their birth weight and fat intake. In the men in the lowest third of birth weight, ie, 3.2 kg (7 lb), higher intakes of total and saturated fat were associated with lower HDL-cholesterol concentrations. In men in the highest third of birth weight, ie, 3.65 kg (8 lb), these trends were reversed (P for interaction 0.02 for total fat and 0.03 for saturated fat intakes, and P 0.09 and 0.08, respectively, after adjustment for the effects of BMI, age, and alcohol intake). This interaction was not seen in the women (Table 3). We tested the heterogeneity of the interactive effects of fat intake and birth weight on serum HDL-cholesterol concentrations between the men and women. The 3-way interactions were not statistically significant. In a model where men and women were pooled and in which the 3 main effects (fat intake, birth weight, and sex) and a 2-factor interaction term for fat and birth TABLE 3 Serum HDL-cholesterol concentrations according to thirds of current fat intake and thirds of birth weight ( 3.2, 3.2 and 3.65, or 3.65 kg) in 574 men and 562 women 1 Serum HDL cholesterol Men Women 3.2 kg 3.2 and 3.65 kg 3.65 kg 3.2 kg 3.2 and 3.65 kg 3.65 kg mmol/l mmol/l Thirds of daily total fat intake 77 g, Men; 60 g, women [54] [59] [56] [73] [60] [51] 77 and 97 g, Men; 60 and [63] [67] [76] [79] [80] [36] g, women 97 g, Men; 78 g, women [62] [80] [57] [73] [69] [41] P for interaction P for interaction, adjusted Thirds of daily saturated fat intake 28.3 g, Men; 20.9 g, women [49] [65] [54] [75] [60] [49] 28.3 and 36.9 g, Men; 20.9 and [69] [61] [82] [73] [83] [39] 28.6 g, women 36.9 g, Men; 28.6 g, women [61] [80] [53] [77] [66] [40] P for interaction P for interaction, adjusted Values are geometric x SD; n in brackets. 2 Obtained from multiple linear regression models with fat intakes and birth weight analyzed as continuously distributed variables. 3 Adjusted for the effects of current BMI, age, and alcohol intake. In a pooled model for men and women, in which the 3 main effects (fat intake, birth weight, and sex) and a 2-way interaction term for fat and birth weight were included to predict serum HDL concentration, the interaction term was statistically significant (P for interaction in total fat intake model; P for interaction in saturated fat intake model).

4 240 ROBINSON ET AL TABLE 4 Serum LDL-cholesterol concentrations according to thirds of current fat intake and thirds of birth weight 3.2, 3.2 and 3.65, or 3.65 kg in 574 men and 562 women 1 Serum LDL cholesterol Men Women 3.2 kg 3.2 and 3.65 kg 3.65 kg 3.2 kg 3.2 and 3.65 kg 3.65 kg Thirds of daily total fat intake 77 g, Men; 60 g, women [54] [59] [56] [73] [60] [51] 77 and 97 g, Men; 60 and [63] [67] [76] [79] [80] [36] 78 g, women 97 g, Men; 78 g, women [62] [80] [57] [73] [69] [41] P for interaction P for interaction, adjusted Thirds of daily saturated fat intake 28.3 g, Men; 20.9 g, women [49] [65] [54] [75] [60] [49] 28.3 and 36.9 g, Men; [69] [61] [82] [73] [83] [39] 20.9 and 28.6 g, women 36.9 g, Men; 28.6 g, women [61] [80] [53] [77] [66] [40] P for interaction P for interaction, adjusted Values are arithmetic x SD; n in brackets. 2 Obtained from multiple linear regression models with fat intakes and birth weight analyzed as continuously distributed variables. 3 Adjusted for the effects of current BMI, age, and alcohol intake. weight were included to predict serum HDL-cholesterol concentrations, the interaction term was statistically significant (P for interaction in the total fat intake model, P for interaction in the saturated fat intake model). Although the men and women could have been pooled for subsequent analyses, we considered men and women separately because of the evidence of sex-specific differences in the early life programming of cholesterol metabolism (12). Serum LDL-cholesterol concentrations in the men and women are shown in Table 4 according to birth weight and total and saturated fat intakes. Unlike the effects on HDL-cholesterol concentration, there were no clear differences in the association between total or saturated fat intake and LDL-cholesterol concentration in men or women of different birth weights. No interactive effects of birth weight and fat intake on serum total cholesterol concentrations were observed in either men or women (data not shown). The combined effects of current fat intakes and birth weight on the ratio of HDL- to LDL-cholesterol concentration are shown in Figure 1. In the men with birth weights in the lowest third of the distribution [ie, 3.2 kg (7lb)], higher intakes of total and saturated fat were associated with lower HDL-to-LDL cholesterol ratios. In the men with birth weights in the highest third [ie, 3.65 kg (8lb)], these trends were reversed (P for interaction 0.02 for total fat and 0.01 for saturated fat; adjusted P 0.05 and 0.02, respectively). Exclusion of persons taking cholesterol-lowering medication Thirty-two of the men were taking either statin or fibrate medication to reduce their serum cholesterol concentrations. HDLand LDL-cholesterol concentrations and the ratio of HDL to LDL cholesterol in men of different birth weights, after exclusion of men taking cholesterol-lowering medications, according to total and saturated fat intakes are shown in Table 5. The interactions between the effects of dietary fat and birth weight on HDL and on the ratio of HDL- to LDL-cholesterol concentrations were strengthened after excluding these men. Their exclusion also showed a tendency for serum LDL-cholesterol concentrations to increase with increasing saturated fat intakes in the men whose birth weights were 3.2 kg (7lb), although the test for interaction was not significant. Exclusion of 42 women who were taking cholesterol-lowering medication from the analysis did not show significant interactions between fat intake and birth weight on any of the cholesterol outcomes considered (data not shown). DISCUSSION We found that total and saturated fat intakes have different associations with serum HDL-cholesterol concentrations and with the ratio of HDL to LDL cholesterol in men with different birth weights. The influence of the interactions between birth weight and adult fat intake on adult HDL cholesterol and the HDL-to-LDL cholesterol ratio were strengthened after exclusion of men taking cholesterol-lowering medication. No comparable effects were seen in women. These findings parallel the interactions observed in guinea pigs, where plasma total cholesterol responses to a 6-wk period of cholesterol feeding differed according to birth weight in males but not in females (12). Study population The men and women we studied came from a large cohort of men and women who were born between 1931 and 1939 in the county of Hertfordshire, United Kingdom (13). At that time, the birth weights of all babies born in the county were recorded. The criteria for selection for the present study were that the men and women were living in the Eastern part of the county and were willing to take part in the study. Fifty-four percent of the men and women approached agreed to be interviewed at home, and their birth weights were comparable with those of the men and women

5 BIRTH WEIGHT, DIETARY FAT, AND SERUM CHOLESTEROL 241 FIGURE 1. Mean (95% CI) serum HDL-to-LDL cholesterol ratios in 574 men and 562 women by thirds of daily fat intake and birth weight. Total fat intake: o, low intake ( 77 g for the men, 60 g for the women);, medium intake ( 77 and 97 g for the men, 60 and 78 g for the women); and s, high intake ( 97 g for the men, 78 g for the women). Saturated fat intake: o, low intake ( 28.3 g for the men, 20.9 g for the women);, medium intake ( 28.3 and 36.9 g for the men, 20.9 and 28.6 g for the women); and s, high intake ( 36.9 g for the men, 28.6 g for the women). P values for interaction were obtained from multiple linear regression models with fat intakes and birth weights analyzed as continuously distributed variables. In a pooled model for men and women in which the 3 main effects (fat intake, birth weight, and sex) and a 2-factor interaction term for fat and birth weight were included to predict the HDL-to-LDL cholesterol ratio, the interaction term was statistically significant (P for interaction in the total fat intake model and 0.05 in the saturated fat intake model). who chose not to participate in the study. The energy intakes of the subjects were consistent with the requirements of older adults (23), although compared with the last national cohort of older men and women studied in , the percentage energy contributed by total and saturated fat was lower (24). In contrast with early ecological studies, but similar to other cross-sectional surveys, we did not find simple univariate associations between dietary fat intake and serum cholesterol concentrations (25 27). We assessed diet using an administered FFQ. Although there is concern that FFQs are prone to measurement error (28), random misclassification of a person s fat intakes would result in an attenuation of associations, and it is very unlikely that measurement error could explain the interactive effects of diet and birth weight on serum cholesterol concentration we observed. Prenatal growth and adult cholesterol metabolism Although associations between weight at birth and adult serum cholesterol concentrations have been described before, they are not consistent across studies (29, 30). For example, our finding of a positive association between birth weight and serum HDLcholesterol concentration in men, after adjustment for current BMI, agrees with other reports (29); however, in the present study, we did not find this association in the women or with LDLor total cholesterol concentrations in men or women. It has been suggested that part of the heterogeneity in the relation between birth weight and adult cholesterol concentrations is due to stronger associations reported in smaller studies and in studies of cholesterol concentrations in infants (30). Greater insight into the relation between prenatal growth and adult cholesterol metabolism may come from examination of proportions rather than size at birth. In a study conducted in adults in Sheffield, United Kingdom, whose abdominal circumference at birth had been recorded, small circumference predicted raised total and LDLcholesterol concentrations (31). Associations with other measures of body size at birth, including birth weight, were weak. Because abdominal circumference at birth reflects liver size, and the liver has a major role in the regulation of cholesterol homeostasis (32), an inference is that impaired liver growth in utero resets cholesterol homeostasis toward a more atherogenic profile. In the study conducted in guinea pigs, adult liver weight tended to be reduced in the low-birth-weight offspring and was associated with a lower plasma HDL-cholesterol concentration and higher LDL-cholesterol concentration after cholesterol feeding in the males but not females (12).

6 242 ROBINSON ET AL TABLE 5 Serum HDL-cholesterol concentrations and HDL-to-LDL cholesterol ratios according to thirds of current fat intakes and thirds of birth weight ( 3.2, 3.2 and 3.65, or 3.65 kg) in 542 men who were not taking cholesterol-lowering medication 1 HDL cholesterol LDL cholesterol HDL:LDL cholesterol 3.2 kg 3.2 and 3.65 kg 3.65 kg 3.2 kg 3.2 and 3.65 kg 3.65 kg 3.2 kg 3.2 and 3.65 kg 3.65 kg mmol/l mmol/l Thirds of daily total fat intake 77 g [46] [55] [50] [46] [55] [50] [46] [55] [50] 77 and 97 g [57] [65] [73] [57] [65] [73] [57] [65] [73] 97 g [62] [77] [57] [62] [77] [57] [62] [77] [57] P for interaction P for interaction, adjusted 3 Daily saturated fat intake 28.3 g [41] [60] [48] [41] [60] [48] [41] [60] [48] 28.3 and 36.9 g [63] [60] [79] [63] [60] [79] [63] [60] [79] 36.9 g [51] [77] [53] [61] [77] [53] [61] [77] [53] P for interaction P for interaction, adjusted 3 1 Values are arithmetic x SD for LDL cholesterol and geometric x SD for HDL cholesterol and HDL:LDL cholesterol; n in brackets. 2 Obtained from multiple linear regression models with fat intakes and birth weight analyzed as continuously distributed variables. 3 Adjusted for the effects of current BMI, age, and alcohol intake.

7 BIRTH WEIGHT, DIETARY FAT, AND SERUM CHOLESTEROL 243 Interactive influences on adult cholesterol metabolism Our data suggest that adult cholesterol responses to dietary fat are conditional on birth weight, and hence prenatal growth, in men. This agrees with the experimental findings from the study conducted in guinea pigs, in which the cholesterol response to a cholesterol challenge was increased in low-birth-weight male offspring (12). This interactive effect of birth weight and postnatal diet on adult cholesterol concentrations was only evident after the challenge of a high-cholesterol diet, because basal cholesterol concentrations were similar in low- and high-birthweight offspring. If combined effects of prenatal growth and postnatal diet influence adult cholesterol metabolism, this would suggest that the most atherogenic adult cholesterol profiles in men would be found in populations where restricted prenatal growth is common and where adult diets are high in total and saturated fat clearly of concern for populations in nutrition transition (33). Effects of different types of dietary fat We found no significant differences of the interaction between total or saturated fat intakes and birth weight on serum cholesterol concentrations, although we expected to see stronger effects with saturated fat intakes. One explanation may be that our ability to estimate saturated and unsaturated fat intakes separately was limited. Although we obtained brand information for all spreading fats and cooking oils used by the men and women studied, these products only account for 20% of the saturated fat intakes of older UK adults (24). Equal contributions to saturated fat intakes are made by cereal products and meat products, but we may not have had sufficiently detailed composition information to allow an accurate assessment of the separate fat fractions. For this reason, we did not include monounsaturated or polyunsaturated fat intakes in our analyses. We did not assess trans-fat intakes because our composition database was incomplete. Sex differences in adult cholesterol metabolism The interactive effects of prenatal growth and adult diet on serum cholesterol concentration were only seen in the men. Although our study may have insufficient statistical power to detect 3-way interactive effects of birth weight, fat intake, and sex on serum cholesterol concentrations, the finding agrees with a growing body of observational and experimental evidence of sex differences in the relation between prenatal growth and adult cholesterol concentrations. In a follow-up study conducted in Filipino adolescents whose mothers had been studied during pregnancy, low maternal energy status was associated with more atherogenic lipid profiles in the male adolescents than in the female adolescents, as indicated by LDL-, HDL-, and total cholesterol concentrations (34). These associations were independent of the adolescents energy and fat intakes and their adiposity. Because in the female offspring, higher maternal energy status was associated with more atherogenic profiles, the authors suggested that there were sex differences in the relation between fetal nutrition and postnatal lipid metabolism. Such sex differences in cholesterol metabolism may be expected, because men and women also differ in their cholesterol response to a reduction in fat intake (35). Combined effects of early and later experience on adult disease Other studies have established that responses to adverse influences in adult life are conditioned by prenatal growth. In the American Nurses Health Study, the highest risk of coronary heart disease was found in women who had low birth weight and high BMI as adults (36). In men from Helsinki, Finland, low income was associated with increased rates of coronary heart disease, as would be expected (37). The association, however, was confined to men who were thin at birth, which was defined by a low ponderal index (birth weight/length 3 ). These findings are consistent with the ideas of Dubos (38), who wrote that the effects of the physical and social environments cannot be understood without knowledge of individual history. The present study suggests that responses to dietary fat are also conditioned by prenatal growth, which could help to explain the inconsistent associations observed between birth weight and serum cholesterol concentration (29, 30) and in the cross-sectional relations between dietary fat intake and serum cholesterol concentration (25 27). The analysis of the effects of dietary fat on serum cholesterol concentrations therefore needs to take birth weight into account; conversely, the analysis of the effects of birth weight on serum cholesterol concentrations has to take current diet into account. If replicated, our findings could have implications for public health. Lower birth weight would define a group of men who are at increased risk of coronary heart disease and who may benefit from a reduction in their dietary fat intake. We thank the men and women who took part in the study, the Hertfordshire Cohort Study research nurses who collected the data, and Clare Statham who assisted with data management. The Hertfordshire Cohort Study Group includes David Phillips, Sue Craighead, Clare Watson, Vanessa Cox. EMD, AAS, HES, DJPB, and CC designed the study. The study was supervised by EMD, AAS, HES, HJM, and the Hertfordshire Cohort Study Group. SMR and SFB were responsible for the dietary data. The data were analyzed by HES. All authors contributed to interpretation of the data. SMR and DJPB wrote the manuscript. None of the authors had a conflict of interest. REFERENCES 1. Hu FB, Manson JE, Willett WC. Types of dietary fat and risk of coronary heart disease: a critical review. J Am Coll Nutr 2001;20: Schaefer EJ, Lamon-Fava S, Ausman LM, et al. Individual variability in lipoprotein cholesterol response to National Cholesterol Education Program Step 2 diets. Am J Clin Nutr 1997;65: Masson LF, McNeill G, Avenell A. Genetic variation and the lipid response to dietary intervention: a systematic review. Am J Clin Nutr 2003;77: Barker DJP, Osmond C, Winter PD, Margetts B, Simmonds SJ. Weight in infancy and death from ischaemic heart disease. Lancet 1989;2: Osmond C, Barker DJP, Winter PD, Fall CHD, Simmonds SJ. Early growth and death from cardiovascular disease in women. BMJ 1993; 307: Frankel S, Elwood P, Sweetnam P, Yarnell J, Davey Smith G. Birth weight, body mass index in middle age, and incident coronary heart disease. Lancet 1996;348: Rich-Edwards JW, Stampfer MJ, Manson JE, et al. Birth weight and risk of cardiovascular disease in a cohort of women followed up since BMJ 1997;315: Leon DA, Lithell HO, Vagero D, et al. Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of Swedish men and women born BMJ 1998;317: Eriksson JG, Forsen T, Tuomilehto J, Osmond C, Barker DJP. Early growth and coronary heart disease in later life: longitudinal study. BMJ 2001;322: Labarthe DR, Eissa M, Varas C. Childhood precursors of high blood

8 244 ROBINSON ET AL pressure and elevated cholesterol. Annu Rev Public Health 1991;12: Gluckman PD, Hanson MA. Living with the past: evolution, development and patterns of disease. Science 305: Kind KL, Clifton PM, Katsman AI, Tsiounis M, Robinson JS, Owens JA. Restricted fetal growth and the response to dietary cholesterol in the guinea pig. Am J Physiol 1999;277:R Syddall HE, Aihie Sayer A, Dennison EM, Martin HJ, Barker DJP, Cooper C. Cohort profile: the Hertfordshire cohort study. Int J Epidemiol 2005;34: Rose GA, Blackburn H. Cardiovascular Survey Methods. Geneva, Switzerland: World Health Organization, 1968 (World Health Organisation Monograph Series 56). 15. Prineas RJ, Crow RS, Blackburn H. The Minnesota code manual of electrocardiographic findings: standards and procedures for measurement and classification. Boston, MA: John Wright, Bingham SA, Gill C, Welch A, et al. Comparison of dietary assessment methods in nutritional epidemiology: weighed records v. 24h recalls, food-frequency questionnaires and estimated-diet records. Br J Nutr 1994;72: Ministry of Agriculture Fisheries and Food. Food portion sizes. London, United Kingdom:HMSO, Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT. McCance and Widdowson s the composition of foods. 5th ed (and Supplements). Cambridge, United Kingdom: The Royal Society of Chemistry, Allain CC, Poon LS, Chan CSG, Richmond W, Fu PC. Enzymatic determination of serum total cholesterol. Clin Chem 1974;20: Lie RF, Schmitz JM, Pierre KJ, Gochman N. Cholesterol oxidase-based determination by continuous flow analysis of total and free cholesterol in serum. Clin Chem 1976;22: Lopes-Virella MF, Stone P, Ellis S, Colwell JA. Cholesterol determination in high density lipoproteins separated by three different methods. Clin Chem 1977;23: 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: Department of Health. Dietary reference values for food energy and nutrients for the United Kingdom. Report on health and social subjects 41. London, United Kingdom: HMSO, Finch S, Doyle W, Lowe C, et al. National Diet and Nutrition Survey: people aged 65 years and over. London, United Kingdom: HMSO, Bolton Smith C, Woodward M, Smith WCS, Tunstall Pedoe H. Dietary and non-dietary predictors of serum total and HDL-cholesterol in men and women: results from the Scottish Heart Health Study. Int J Epidemiol 1991;20: Sonnenberg LM, Posner BM, Belanger AJ, Cupples LA, D Agostino RB. Dietary predictors of serum cholesterol in men: the Framingham cohort population. J Clin Epidemiol 1992;45: Mozaffarian D, Rimm EB, Herrington DM. Dietary fats, carbohydrate and the progression of coronary atherosclerosis in postmenopausal women. Am J Clin Nutr 2004;80: 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: Lauren L, Jarvelin MR, Elliott P, et al. Relationship between birthweight and blood lipid concentrations in later life: evidence from the existing literature. Int J Epidemiol 2003;32: Huxley R, Owen CG, Whincup PH, Cook DG, Colman S, Collins R. Birth weight and subsequent cholesterol levels: exploration of the fetal origins hypothesis. JAMA 2004;292: Barker DJP, Martyn CN, Osmond C, Hales CN, Fall CHD. Growth in utero and serum cholesterol concentrations in adult life. BMJ 1993;307: Spady DK, Woolett LA, Dietschy JM. Regulation of LDL-cholesterol levels by dietary cholesterol and fatty acids. Annu Rev Nutr 1993;13: Popkin BM. The nutrition transition: an overview of world patterns of change. Nutr Rev 2004;62:S Kuzawa CW, Adair LS. Lipid profiles in adolescent Filipinos: relation to birth weight and maternal energy status during pregnancy. Am J Clin Nutr 2003;77: Weggemans RM, Zock PL, Urgert R, Katan MB. Differences between men and women in the response of serum cholesterol to dietary changes. Eur J Clin Invest 1999;29: Rich-Edwards JW, Kleinman K, Michels KB, et al. Longitudinal study of birth weight and adult body mass index in predicting risk of coronary heart disease and stroke in women. BMJ 2005;330: Barker DJP, Forsen T, Uutela A, Osmond C, Eriksson JG. Size at birth and resilience to the effects of poor living conditions in adult life: longitudinal study. BMJ 2001;323: Dubos R. The mirage of health. Piscataway, NJ: Rutgers University Press, 1987.