点击显示 收起
1 From the Departments of Pediatrics and Medicine, Columbia University, New York.
2 Supported by grants HD 32195, HL 56984, and RR 00645 from the National Institutes of Health, Bethesda, MD. 3 Address reprint requests to L Berglund, Department of Medicine, Room PH 10-305, Columbia University, 630 West 168th Street, New York, NY 10032. E-mail: berglun{at}cudept.cis.columbia.edu.
ABSTRACT
Background: Predictors of postprandial lipemia have not been explored in children.
Objective: Our objective was to determine whether the postprandial triacylglycerol response is associated with low HDL-cholesterol and high fasting triacylglycerol concentrations and family history of early-onset ischemic heart disease (IHD) in children.
Design: We administered a standardized fat load (52.5 g fat/m2) to 60 children (mean age: 14.0 y), 20 with and 40 without a family history of early-onset IHD, and to 29 mothers, all recruited from families enrolled in the Columbia University Biomarkers Study. Plasma lipid and retinyl palmitate concentrations were measured in the fasting state and 3, 6, and 8 h after the oral fat load.
Results: In children, postprandial lipemia, as indicated by the incremental area under the triacylglycerol response curve, was associated with elevated fasting triacylglycerol concentrations (1.13 mmol/L; P < 0.01), with low fasting HDL-cholesterol concentrations (0.91 mmol/L; P < 0.01), and with the combination of low HDL-cholesterol and high triacylglycerol concentrations (P < 0.05). Family history of IHD, baseline LDL-cholesterol concentration, and apolipoprotein E genotype were not associated with the postprandial triacylglycerol or retinyl palmitate response. The mothers had fasting triacylglycerol concentrations similar to those of their children but a more prolonged response with higher triacylglycerol concentrations at 6 and 8 h (P < 0.01 and P < 0.05, respectively).
Conclusions: In children, a delayed postprandial triacylglycerol response to a fat load is associated with the combination of high fasting triacylglycerol and low HDL-cholesterol concentrations. Predictors of postprandial triacylglycerol concentrations may be similar in children and adults.
Key Words: Atherosclerosis • postprandial period • triacylglycerol • children • coronary heart disease • HDL cholesterol • risk factors • ischemic heart disease • Columbia University Biomarkers Study
INTRODUCTION
The arterial lesions of atherosclerosis have their origins in childhood (1–4). Although many different factors may influence the development of atherosclerosis, data in adults suggest that impaired postprandial lipoprotein metabolism may contribute to, or be a marker of, the development and progression of this disease (5, 6). Patients with ischemic heart disease (IHD) have delayed clearance of triacylglycerol-rich lipoproteins after consumption of a high-fat meal compared with control subjects (6–10). Prolonged circulation of chylomicron remnants and VLDLs increases the exposure of the vascular bed to the atherogenic contents of these particles (11, 12). In support of the atherogenic effect of such exposure, in vitro studies have shown that cholesteryl ester–enriched triacylglycerol-rich lipoprotein remnants are taken up by monocyte-derived macrophages in the arterial wall and contribute to atheromatous plaque formation (13–15).
Studies in adults have shown associations of delayed postprandial fat clearance with abnormalities in lipoprotein profile. Delayed clearance is seen in conjunction with hypertriacylglycerolemia, low concentrations of HDL cholesterol (primarily the antiatherogenic HDL2 subfraction), and the presence of small, dense LDL particles, ie, a pattern typical of insulin resistance (16–19). Adults carrying the apolipoprotein (apo) E2 allele also have relatively slow postprandial fat clearance (20). These relations have not been assessed in children, in whom mean fasting triacylglycerol concentrations are generally lower and mean HDL-cholesterol concentrations higher than in adults (21).
Whether delayed postprandial fat clearance in young adults is related to family history of early-onset IHD has been debated. Two studies have addressed postprandial clearance in young adult men. Uiterwaal et al (22) showed more prolonged postprandial hypertriacylglycerolemia in 80 young men aged 15–30 y whose fathers had angiographic evidence of early coronary artery disease than in age-matched men whose fathers had no lesions on coronary angiography. However, Slyper et al (23) found no evidence of altered chylomicron remnant clearance or elevated postprandial triacylglycerol response in young men of similar age recruited from families positive for IHD compared with young men with no family history of IHD.
To explore predictors of postprandial triacylglycerol response to a fat load in children, we recruited a subsample of children and their mothers from the families with or without a family history of early-onset IHD in the Columbia University Biomarkers Study. The aims of the postprandial study were 1) to determine whether delayed postprandial fat response in children is related to fasting triacylglycerol concentrations, HDL- and LDL-cholesterol concentrations, or apo E genotype; 2) to compare the postprandial fat response in children with that of their mothers; and 3) to investigate whether children from families with a history of early-onset IHD had elevated or prolonged postprandial lipemia compared with children from families without such a history.
SUBJECTS AND METHODS
Subjects
The children participating in this study were recruited from families enrolled in the Columbia University Biomarkers Study, a 5-y observational cohort study initiated to identify biological markers in children from families with a history of premature atherosclerosis. Families with or without a history of early-onset IHD were recruited from clinics for high-risk cardiovascular disease and pediatric practices at the New York Presbyterian Hospital, lists of cardiac disease patients generated through the Presbyterian Hospital Clinical Information System, private cardiology practices, low-risk settings such as general pediatric practices, and the general population through fliers posted within the medical center. Medical histories were obtained through individualized interviews conducted in English or Spanish with use of structured questionnaires. Family history of early-onset IHD was classified as positive if one or both parents had an early onset of clinical IHD (55 y of age for men and 65 y of age for women). Early-onset IHD was indicated by a history of myocardial infarction, coronary artery bypass surgery, coronary angioplasty, sudden death documented in the patient's medical record, or coronary arteriographic documentation of 50% narrowing of the luminal diameter of 1 major epicardial coronary artery. Family history was classified as indeterminant if family members were unsure of their medical history or if there was a history of early-onset IHD in 1 of the grandparents. Family history was categorized as negative if the self-reported medical history was negative for early-onset IHD in both parents and all 4 grandparents.
Families with 1 healthy child 4–21 y of age were eligible for participation in the Biomarkers Study. At the time of recruitment for the postprandial phase of the study, 335 families had been enrolled in the overall study, and all families with children between the ages of 10 and 21 y were invited to participate. Children <10 y of age were excluded from postprandial studies because of the length of time subjects were required to fast before (12 h) and after (6–8 h) the fat load and the need for repeated blood draws at each postprandial time point. Additionally, children were excluded if they had diabetes mellitus, had kidney or thyroid disease, were taking any lipid-altering medications, or had a fasting triacylglycerol concentration >2.26 mmol/L (>200 mg/dL). Participation of 1 parent and 1 child from each family was required unless the child was >18 y of age. Two hundred forty-two children met the inclusion criteria. Of these eligible families, 60 children and 29 mothers representing 29 families participated in the postprandial study. Informed consent was obtained from all participating families. The study was approved by the Institutional Review Board at the Columbia Presbyterian Center of the New York Presbyterian Hospital.
Postprandial lipemia protocol
Subjects were admitted as outpatients to the Irving Center for Clinical Research at the Columbia Presbyterian Center of the New York Presbyterian Hospital on the morning of the study. All subjects were instructed to fast for 12 h before admission. Vital signs and anthropometric measurements were obtained, a history was taken, and a physical examination was performed. Fasting blood samples were obtained for measurement of baseline lipid, retinyl palmitate, and lipoprotein concentrations.
The subjects were then given a fat formula, described in detail below, to consume within 15 min. Formula containers were rinsed twice with distilled water to ensure that all the content was consumed. Sugar- and caffeine-free soda, water, and sugar-free gum were the only foods allowed during the protocol before the last blood draw. Subjects were kept as inactive as possible and were not permitted to walk around during the protocol. To make the protocol less challenging for the participating children, we initially obtained blood samples at 3 and 6 h after consumption of the oral fat load. After completion of the first recruitment phase (n = 30 children), preliminary analysis showed that mean postprandial triacylglycerol and retinyl palmitate concentrations at 6 h were not significantly below those at 3 h. The children tolerated the protocol well. We therefore extended the protocol for the remaining 30 children and their parents to include a blood sample collection at 8 h.
Fat formula
The oral fat load contained ice cream, heavy cream (Tuscan Heavy Cream, graded; Tuscan Dairy, Union, NJ), safflower oil, chocolate or strawberry syrup (Nestlé, Glendale, CA), and a powdered whey protein source (Promod; Ross Laboratories, Columbus, OH). The nutrient composition of the formula per m2 body surface area was 52.5 g fat (75% of total energy, of which 26 g was saturated fat), 24 g carbohydrate (15% of total energy), 16 g protein (10% of total energy), and 150 mg cholesterol; it provided 2647 kJ (632 kcal)/m2. The formula was prepared with a dairy digestive supplement (Lactaid; McN-PPC, Inc, Fort Washington, PA) 24 h before administration. Aqueous retinyl palmitate (50000 IU, or 1g/m2 body surface area) was added (Vitadrel-Tropfen; Pharma-Wernigerode, Wernigerode, Germany) on the day of the study. On the morning of the day of the test, the weight and height of each participant was obtained, the body surface area was calculated, and the appropriate amount of the formula was administered.
Anthropometric measurements and laboratory analyses
Height was measured to the nearest 1.0 cm by using a rigid stadiometer. Weight of subjects in light clothing was measured to the nearest 0.1 kg by using a calibrated triple beam-balance scale. Body mass index (BMI) was calculated as the weight (kg) divided by the height (m) squared (kg/m2). Body surface area was calculated according to the equation of Dubois (24). Circumferences at the waist (the narrowest part of the torso) and hips (the maximum extension of the buttocks) were measured with a tape measure (25).
Blood for serum and plasma analyses was drawn into sterile tubes, the latter containing EDTA. Serum samples were left at room temperature for 30 min, whereas the plasma samples were immediately placed on ice and kept at 4°C until centrifugation. Both serum and plasma were separated from cells by centrifugation at 30000 x g for 30 min at 4°C. Plasma to be used for retinyl palmitate analysis was protected from light and stored under nitrogen. Samples were stored at -80°C until assayed.
Cholesterol and triacylglycerol concentrations were analyzed by using standard enzymatic procedures (Boehringer Mannheim, Mannheim, Germany) with a Hitachi 704 automated spectrophotometer (Hitachi, Tokyo). HDL-cholesterol concentrations were measured after precipitation of plasma apo B–containing lipoproteins with dextran sulfate–magnesium chloride (26). LDL-cholesterol concentrations were determined by using the Friedewald equation (27). Plasma retinyl palmitate concentrations were determined by reversed-phase HPLC essentially as described previously (10, 28). Apo E genotyping was performed by using genomic DNA as described by Hixson and Vernier (29).
Statistical methods
Because data were collected over a 2-y period, trends over time for triacylglycerol and retinyl palmitate data were assessed by repeated-measures analysis of variance (ANOVA). No secular trends were noted in the mean baseline or postprandial triacylglycerol or retinyl palmitate concentrations from 1996 to 1997 for either children or adults. Data from both years were therefore combined for further analysis.
Children were grouped on the basis of their triacylglycerol concentrations (normal or high, dichotomized at 1.13 mmol/L, or 100 mg/dL), LDL-cholesterol concentrations (normal or high, dichotomized at 3.36 mmol/L, or 130 mg/dL), HDL-cholesterol concentrations (low or normal, dichotomized at 0.91 mmol/L, or 35 mg/dL), and apo E genotype status (apo E2, E3, or E4 carriers, defined as E3/2, E3/3, and E4/3 + 4/4, respectively). Cutoff concentrations for LDL-cholesterol groups were based on those suggested by the Expert Panel of the National Cholesterol Education Program (21). Because there is no consensus on appropriate triacylglycerol or HDL-cholesterol concentrations for children, cutoff concentrations for triacylglycerol and HDL groups were based on Lipids Research Clinics data reflecting the 90th percentile for triacylglycerol and the 5th percentile for HDL-cholesterol concentration (30).
Fasting and postprandial triacylglycerol and retinyl palmitate values were transformed to their natural logarithms before statistical analyses to normalize skewed distributions. Nontransformed values are shown in the tables for descriptive purposes, but statistical testing and resulting P values were based on transformed distributions. A two-factor ANOVA with interaction was performed on ethnicity. Mean differences between groups in age, BMI, waist-to-hip ratio, lipid, and lipoprotein data were evaluated by Student's t test. To quantify the magnitude of the postprandial response in triacylglycerol and retinyl palmitate, the area under the response curve (AUC) was calculated. The formula used to compute the AUC is based on the trapezoidal rule and measures AUC above the fasting (baseline) measurement (10). Because of the limited number of subjects at the 8-h time point, we truncated the AUC formula to fit the 0-, 3-, and 6-h time points only (for both children and adults). All values were combined by using nontransformed values, and AUC distributions were transformed as above for statistical analyses of triacylglycerol and retinyl palmitate distributions.
Analysis of covariance (ANCOVA) was used to adjust lipid, lipoprotein, and triacylglycerol and retinyl palmitate AUCs for 1) ethnic differences after a comparison of offspring with or without a family history of early-onset IHD and of high and normal LDL-cholesterol groups and 2) waist-to-hip ratio differences after a comparison of high and normal triacylglycerol groups and of children with and without a family history of early-onset IHD. Separate one-factor ANOVAs were performed on the HDL-cholesterol groups (normal HDL cholesterol and low HDL cholesterol with or without high triacylglycerol) to compare age, BMI, and waist-to-hip ratio.
In addition, a repeated-measures ANOVA was used to determine differences in fasting and postprandial triacylglycerol and retinyl palmitate concentrations between HDL-cholesterol and triacylglycerol groups and between time points. Significant interactions were tested with separate one-factor ANOVAs at each time point with appropriate post hoc analysis. Pearson correlation coefficients were computed to determine relations between postprandial AUCs and continuous variables (age, BMI, waist-to-hip ratio, and HDL-, LDL-cholesterol, and fasting triacylglycerol concentrations). Bivariate analysis was used to determine associations between postprandial AUC and categorical variables (ethnicity, sex, and family history). From these analyses, predictor variables for postprandial AUC were selected and simultaneously assessed by multiple linear regression analysis. Statistical analyses were performed by using SAS software (version 6.12; SAS Institute, Cary, NC).
RESULTS
As shown in Table 1, demographic data and fasting lipid measurements for the 60 children participating in the postprandial study did not differ significantly from those of the other 242 children 10 y of age in the Biomarkers Study, except for triacylglycerol concentrations. The mean triacylglycerol concentration was lower in the participating children because the triacylglycerol cutoff concentration for recruitment was lower (<2.26 mmol/L, or <200 mg/dL) in the postprandial study than in the overall Biomarkers Study (<4.52 mmol/L, or <400 mg/dL).
View this table:
TABLE 1. Characteristics of children assessed for postprandial lipemia compared with all children (10 y of age) in the Columbia University Biomarkers Study, New York, 1994–19971
There were no significant differences in age, BMI, waist-to-hip ratio, or fasting lipoprotein concentrations between boys (n = 27) and girls (n = 33) in the postprandial study (Table 2). Furthermore, the postprandial response of both triacylglycerol and retinyl palmitate did not differ significantly by sex. When the children were divided according to ethnicity, the only significant difference was that the mean fasting LDL-cholesterol concentration was higher in the non-Hispanic white children. This agrees with the finding that a family history of early-onset IHD was more common among the non-Hispanic white than among the white children (see below). However, mean age, BMI, waist-to-hip ratio, fasting and postprandial triacylglycerol and retinyl palmitate concentrations, and HDL-cholesterol concentrations did not differ significantly between the ethnic groups. Additionally, there was no significant interaction between sex and ethnicity for any of the demographic, anthropometric, or lipid variables measured.
View this table:
TABLE 2. Subject characteristics and fasting and postprandial triacylglycerol and retinyl palmitate concentrations of study children by sex and ethnicity1
Comparisons of postprandial responses in children were performed in relation to fasting triacylglycerol, HDL- and LDL-cholesterol concentrations, and apo E allele status (Table 3). For each of these variables, children were categorized into 2 groups. Mean age, BMI waist-to-hip ratio, and proportion of individuals with a positive family history of early-onset IHD did not differ significantly by category of triacylglycerol, HDL-cholesterol or LDL-cholesterol concentration, or apo E allele status among the children in the sample (data not shown), except that children with higher fasting triacylglycerol concentrations (
View this table:
TABLE 3. Fasting and postprandial triacylglycerol and retinyl palmitate concentrations of children by HDL cholesterol (HDLC), LDL cholesterol (LDLC), and apolipoprotein E (apo E) allele group1
As shown in Table 3, children with low HDL-cholesterol concentrations (n = 20) had significantly higher postprandial triacylglycerol concentrations at all 3 time points than did children with higher HDL-cholesterol concentrations (n = 40). After adjustment for the baseline triacylglycerol concentration for each individual, the incremental AUC for the postprandial triacylglycerol response remained significantly greater in the low HDL-cholesterol group (Table 3). The mean retinyl palmitate concentrations did not differ significantly between these 2 HDL-cholesterol groups over the 8-h postprandial period, and the overall AUCs for retinyl palmitate for the 2 groups were not significantly different.
In contrast with results for the triacylglycerol and HDL-cholesterol groups, we did not find any significant difference in either baseline or postprandial triacylglycerol concentrations between children with normal and those with high LDL-cholesterol concentrations (Table 3). The postprandial retinyl palmitate response also did not differ significantly between the 2 LDL-cholesterol groups. Because there was a significant ethnic difference in the normal compared with the high LDL-cholesterol group (non-Hispanic white/Hispanic: 9/33 compared with 11/7, respectively; P < 0.05), we evaluated the response to fat load after adjustment for ethnicity. However, no significant difference was found.
We also evaluated the postprandial response in relation to apo E allele carrier status. Because there were only 2 apo E2 carriers (both carrying the apo E3/2 genotype), we compared the response for apo E3 carriers (E3/3; n = 44) and apo E4 carriers (E4/3 and 4/4; n = 14). As seen in Table 3, there were no significant differences in the postprandial response of either triacylglycerol or retinyl palmitate between the apo E groups. There was a nonsignificant trend for higher baseline triacylglycerol as well as higher triacylglycerol and retinyl palmitate concentrations 3 h postprandially in children with the apo E4 compared with the apo E3 allele.
In adults, a low HDL-cholesterol concentration, either in the absence or presence of a high triacylglycerol concentration, is a risk factor for IHD (31). Delayed clearance of triacylglycerol-rich lipoproteins after a high-fat meal has been reported in adults with high fasting triacylglycerol concentrations in combination with low HDL-cholesterol concentrations (32). To study whether the postprandial response would differ in children with low HDL-cholesterol concentrations, depending on whether or not they had elevated fasting triacylglycerol concentrations, we divided the 20 children in our study with low HDL cholesterol (<0.91 mmol/L, or <35 mg/dL) into those with high fasting triacylglycerol (1.47 mmol/L; n = 8) (HT/LHDL) and normal fasting triacylglycerol (<1.47 mmol/L; n = 12) (NT/LHDL). There was no significant difference between groups in age, BMI, or waist-to-hip ratio (data not shown). A significant interaction between postprandial triacylglycerol and time was found for the groups (P < 0.01). Follow-up post hoc analysis showed that (Figure 1), subjects with HT/LHDL had significantly higher 3- and 6-h triacylglycerol concentrations than did subjects with NT/LHDL (3 h: 3.54 ± 0.19 compared with 2.22 ± 0.27 mmol/L; 6 h: 3.26 ± 0.39 compared with 1.88 ± 0.25 mmol/L; P < 0.01 at both time points). The children with NT/LHDL tended to have a higher postprandial triacylglycerol response than did children with normal HDL-cholesterol concentrations, although the difference was not significant (Figure 1). Additionally, despite the limited number of children in each low HDL-cholesterol subgroup, the difference in postprandial triacylglycerol AUC between the subjects with HT/LHDL and those with NT/LHDL was nearly significant (7.40 ± 1.21 compared with 5.20 ± 1.09 mmol•h/L, respectively; P = 0.07). These observations suggest that fasting triacylglycerol concentration modulates postprandial response in children with low HDL-cholesterol concentrations, as it does in adults. No significant differences were observed between groups with respect to plasma retinyl palmitate concentrations (Figure 1) or the retinyl palmitate AUC (data not shown).
FIGURE 1. . Comparison of mean (±SE) baseline (0 h) and postprandial (3, 6, and 8 h) triacylglycerol responses and postprandial retinyl palmitate responses in children with low HDL-cholesterol, high triacylglycerol concentrations (; <0.91 mmol/L and 1.47 mmol/L, respectively; n = 8); low HDL-cholesterol, normal triacylglycerol concentrations (; <0.91 mmol/L and <1.47 mmol/L, respectively; n = 12); and normal HDL-cholesterol concentrations (; 0.91 mmol/L; n = 40). *Significantly different from other 2 groups, P < 0.001; **significantly different from those with low HDL-cholesterol and normal triacylglycerol concentrations, P < 0.01; ***significantly different from those with low HDL-cholesterol and high triacylglycerol concentrations, P < 0.001.
A unique feature of this study was the possibility to directly compare the postprandial triacylglycerol and retinyl palmitate responses of parents with those of their children. In this comparison, 6 children were excluded because postprandial data were not available for their parents. Furthermore, 8 of the mothers in this study were taking potentially lipid-altering medications, including thyroid medication, oral contraceptives, and oral estrogen. Women taking medication and their children were excluded from the child-mother comparisons. As shown in Table 4, mean BMI, waist-to-hip ratio, fasting triacylglycerol, and LDL- and HDL-cholesterol concentrations did not differ significantly between the children and mothers included in this analysis. Of note, however, mean triacylglycerol concentrations 6 and 8 h postprandially, as well as 8-h retinyl palmitate concentrations, were significantly greater in the mothers than in the children (Figure 2). Additionally, mean differences in triacylglycerol at 8 h was nearly significant (P = 0.06) even after adjustment for the baseline triacylglycerol concentration for each individual, suggesting a more sluggish triacylglycerol response to the oral fat load in mothers than in their children. Because only mothers were included in this analysis, we also compared the results for mothers and daughters (n = 22). Significant differences were found for triacylglycerols at 6 and 8 h and for retinyl palmitate at 8 h (data not shown).
View this table:
TABLE 4. Characteristics and fasting lipid concentrations of study children compared with their mothers1
FIGURE 2. . Comparison of mean (±SE) baseline and postprandial triacylglycerol responses and postprandial retinyl palmitate responses in children (; n = 38) compared with their mothers (; n = 21). *, **Significantly different from children: *P < 0.05, **P < 0.01. n = 19 children and 10 mothers at 8 h.
Mothers (not taking lipid-altering medications) were also grouped by the same criteria as children, ie, high and low triacylglycerol, low and normal HDL cholesterol, normal and high LDL cholesterol, and apo E genotype. Mothers with high triacylglycerol concentrations (1.13 mmol/L; n = 13) or low HDL-cholesterol concentrations (0.91 mmol/L; n = 6) had significantly higher fasting, 3-h, and 6-h triacylglycerol concentrations than did mothers with lower triacylglycerol or higher HDL-cholesterol concentrations, but these differences did not remain significant after correction for baseline triacylglycerol concentrations (data not shown). There were no significant differences in retinyl palmitate concentrations between mothers in the low compared with the normal triacylglycerol or HDL-cholesterol groups (data not shown). Furthermore, no differences were observed in postprandial triacylglycerol or retinyl palmitate concentrations when mothers were grouped by normal or high LDL-cholesterol concentrations or by the presence of the apo E3 compared with the apo E4 allele (data not shown).
Finally, we classified the children's postprandial response data in relation to whether there was a family history of early-onset IHD. As seen in Table 5, there was a greater proportion of Hispanic children than non-Hispanic white children among the offspring of families without a history of early-onset IHD, likely reflecting the differential recruitment of non-Hispanic white children from high-risk settings. Additionally, waist-to-hip ratio was significantly greater in children from families with than without a history early-onset IHD. There were no significant differences in mean age or BMI between the 2 groups of children (Table 5). Furthermore, there were no significant differences in mean fasting or postprandial triacylglycerol or postprandial retinyl palmitate concentrations between children with or without a family history of early-onset IHD, except for the 3-h triacylglycerol concentration, which was higher in children without a family history of early-onset IHD. This difference remained significant after adjustment for baseline triacylglycerol concentrations, ethnicity, and waist-to-hip ratio (ANCOVA). However, there was no difference between the children with and without a family history of early-onset IHD, in the overall postprandial response as measured by the incremental AUC, for either triacylglycerol or retinyl palmitate.
View this table:
TABLE 5. Subject characteristics and fasting and postprandial triacylglycerol and retinyl palmitate concentrations of children with (IHD+) and without (IHD-) a family history of early-onset ischemic heart disease1
The results of the multivariate analyses with AUCs for postprandial triacylglycerol and retinyl palmitate as the dependent variables are shown in Table 6. The baseline triacylglycerol concentration was the only significant predictor of the postprandial triacylglycerol AUC (P = 0.05). None of the predictor variables was significantly associated with the postprandial retinyl palmitate AUC.
View this table:
TABLE 6. Multiple linear regression coefficients relating area under the postprandial triacylglycerol and retinyl palmitate curves (AUCs) to baseline triacylglycerol, family history of early-onset ischemic heart disease (IHD), HDL cholesterol, LDL cholesterol, apolipoprotein E (apo) allele status, ethnicity, and sex1
DISCUSSION
Postprandial lipemia in children, as indicated by the triacylglycerol AUC, was significantly associated with elevated fasting triacylglycerol concentrations (1.13 mmol/L), low HDL-cholesterol concentrations (0.91 mmol/L), and the combination of low HDL-cholesterol and high fasting triacylglycerol concentrations. These predictors were not associated with the postprandial retinyl palmitate response, a better measure of chylomicron remnant clearance. Children had lower triacylglycerol concentrations at 6 and 8 h and lower retinyl palmitate concentrations at 8 h than did their mothers, indicating a more sluggish clearance mechanism for triacylglycerol-rich lipoproteins in women than in children. However, we cannot currently exclude the possibility that different fat absorption patterns could also contribute to this pattern. A family history of early-onset IHD, LDL-cholesterol concentration, and apo E allelic pattern were not associated with postprandial triacylglycerol or retinyl palmitate response in the children in our study.
The importance of the postprandial lipoprotein pattern in the development of atherosclerosis is supported by several experimental findings. Results from animal and tissue culture studies suggest that remnant lipoproteins, whether intestinal or hepatic in origin, can enter the arterial vascular wall and contribute to lipid accumulation (13–15, 33, 34). Additionally, human studies have supported a relation between postprandial lipemia and coronary heart disease in adults, the triacylglycerol response being more pronounced in patients with established IHD (6–10). Clearance of chylomicrons, as assessed by plasma retinyl palmitate concentrations, was also shown to be delayed in adults with IHD compared with that in healthy subjects (8, 9, 35).
Extensive data now exist showing the clustering of metabolic cardiovascular risk factors in both adults and children. The Collaborative Lipid Research Clinics Prevalence Study was among the first to show an association between obesity, low HDL-cholesterol, and high triacylglycerol concentrations in childhood (36). Over an 8-y period, the Bogalusa Heart Study showed that children with persistently low HDL cholesterol, high triacylglycerol, and high BMI had a significantly increased prevalence of dyslipidemia as adults (37). In adults, delayed postprandial clearance is associated with higher fasting triacylglycerol concentrations (6). Fasting triacylglycerol concentrations are generally lower in children than in adults (28, 30). In view of the evidence for a role of delayed postprandial clearance as a risk factor for IHD, the question arises whether delayed postprandial clearance is also present at an early age in children, and if so, which metabolic characteristics predict abnormal clearance in children. We found significantly greater postprandial lipemia in children with elevated fasting triacylglycerol concentrations. Postprandial response was also greater in children with low than in those with normal HDL-cholesterol concentrations. In contrast, LDL-cholesterol concentrations did not predict postprandial response. Similar findings for triacylglycerol and HDL cholesterol were shown in adults but not previously in children (7–9, 16, 17). The relation between HDL-cholesterol and triacylglycerol concentrations has been attributed in adults to the metabolic interaction between HDL particles and triacylglycerol-rich lipoproteins. The concentration of circulating triacylglycerol in plasma has been shown to be a major driving force for the exchange of cholesteryl ester in HDL for triacylglycerols in chylomicrons and VLDL via the cholesterol-ester transfer protein (38, 39). Higher total triacylglycerol concentrations lead to increased core-lipid exchange and consequently may lower HDL-cholesterol concentrations (40).
There is evidence in adults that isolated low HDL-cholesterol concentrations, even in the absence of high triacylglycerol concentrations, are associated with an increased risk of IHD (31). We found that children with low HDL-cholesterol and high fasting triacylglycerol concentrations had a 42% higher triacylglycerol AUC than did those with low HDL-cholesterol and lower fasting triacylglycerol concentrations. Cohen and Grundy (41) found a similar postprandial effect in adult men with low HDL cholesterol in that subjects with a high fasting triacylglycerol concentration had significantly greater postprandial triacylglycerolemia than did those with normal fasting triacylglycerol concentrations. The high triacylglycerol, low HDL-cholesterol lipid profile is common in adults with insulin resistance (19), and recent studies in adults showed an impaired postprandial response in subjects with intraabdominal fat accumulation (42, 43). Our study underscores the finding that a profile of low HDL-cholesterol and high triacylglycerol concentrations is associated with a more pronounced postprandial response than is a low HDL-cholesterol profile in the absence of a high triacylglycerol concentration, and that the combination of low HDL cholesterol and high triacylglycerol concentrations with an elevated postprandial response can be detected in childhood.
We found no relation between fasting triacylglycerol concentration and postprandial retinyl palmitate concentration, a marker of chylomicrons and chylomicron remnants (10). These negative findings should be interpreted with caution because of the relatively small number of subjects in our study. Although 2 separate steps, lipolysis and receptor uptake, are required for retinyl palmitate clearance, only lipolysis is required for triacylglycerol clearance. Our result may suggest that the lipolysis step is primarily affected by fasting triacylglycerol and HDL-cholesterol concentrations, whereas LDL-receptor activity, a major factor in lipoprotein remnant uptake, is less affected. Alternatively, endogenous VLDL and its remnants (which would not carry retinyl palmitate) may contribute to the increased triacylglycerol response to a fat challenge in individuals with low HDL cholesterol and high triacylglycerol, or, a possible transfer of retinyl palmitate from triacylglycerol-rich to cholesterol-rich lipoproteins could contribute to our negative findings.
The apo E2 allele has been reported to be associated with an increased postprandial response in adults (20). There is controversy regarding the effect on postprandial clearance of triacylglycerol-rich lipoproteins of the apo E4 allele compared with the apo E3 allele. Bergeron and Havel (44) found impaired clearance of chylomicron and VLDL remnants in normolipidemic men with the apo E4/3 compared with the apo E3/3 genotype. Conversely, Boerwinkle et al (45) found no significant effects of the apo E4 allele on postprandial triacylglycerol response after adjusting for baseline triacylglycerol concentrations, although an effect on retinyl palmitate clearance was found. We did not find any difference in postprandial triacylglycerol or retinyl palmitate clearance in carriers of the apo E4 allele compared with those with the apo E3/3 genotype. However, there were few apo E2 carriers in our study, which limited our ability to address whether carrying the apo E2 allele affected postprandial lipid response in childhood. Further studies are needed to address the effect of the apo E2 allele on postprandial lipoprotein metabolism in children.
A strength of the present study was the simultaneous collection of postprandial lipid data for mothers and children from the same families. We found a significant difference in postprandial response between children and their mothers. Other studies in adults found that over a wide range, age may affect the postprandial response, clearance being slower with increasing age (21, 46, 47). Several mechanisms could contribute to this finding, among them being decreased activity of lipoprotein lipase, differences in lipoprotein receptor activity, and differences in the interaction between circulating triacylglycerols and adipose tissue uptake (20, 47, 48).
Previous studies have yielded conflicting results on the association between postprandial lipid response and family history of IHD (22, 23). The results of our study differed from the studies of Uiterwaal et al (22) and Slyper et al (23) in that we included both girls and boys and younger children (mean age: 14 y). Uiterwaal et al studied 80 adult sons of fathers with early IHD and 55 adult sons of fathers without IHD, and postprandial hypertriacylglycerolemia was elevated during the 6–10-h interval in the subjects with a history of IHD. Slyper et al found no difference in postprandial response between 60 young males (aged 15–45 y) with a positive family history of IHD and 41 men of similar age with a negative family history of IHD.
In conclusion, our results show that a profile of low HDL cholesterol and high triacylglycerol concentration is associated with impaired postprandial triacylglycerol response in children. The early detection of this dyslipidemic pattern in children could provide a basis for corrective measures to prevent the advent of early atherosclerosis in adulthood.
ACKNOWLEDGMENTS
We thank Minnie Myers, Mae Huang, Nelson Fontanez, and Roseanne Zott for technical assistance, and we are grateful to the nursing staff and the nutrition unit of the Irving Center for Clinical Research for their assistance.
REFERENCES