Postprandial Triglyceride Response in Young Adult Men and Familial Risk for Coronary Atherosclerosis

Methods

Participants

Men with very severe coronary artery disease (patients), defined as more than 70% occlusion in at least three major coronary vessels, were selected from coronary angiography databases of cardiology departments of the Zuiderziekenhuis Rotterdam (1988 to 1991), the University Hospital Rotterdam (1988 to 1989), the Refaja Ziekenhuis in Dordrecht (1990 to 1991), and the Antonius Ziekenhuis in Nieuwegein (1992), all hospitals situated in the Netherlands. Simultaneously, a reference group of men [controls] was selected who at coronary angiography had no or, at most, only minor lesions, defined as 20% stenosis or less in all coronary vessels. Further, participants were selected according to the following additional criteria: 1) age between 45 and 65 years; 2) blood pressure not exceeding 160/100 mm Hg; 3) absence of liver disease, diabetes mellitus, thyroid disease, and renal disease; 4) first coronary angiography within 2 years before examination for our study; and 5) first consultation of a physician for cardiac symptoms within 5 years before the examination for our study.

Eligible participants were sent a letter asking whether they had a son 15 to 30 years of age and, if so, whether the son was willing to participate in the study. These sons (identified by their fathers) received a separate letter inviting them to have an oral lipid loading test. Participants also received a short questionnaire about smoking habits, alcohol intake, physical activity, and fat intake.

We screened medical files of 629 patients whose coronary angiographic data met the criteria. Of these patients, 19 had died, 46 had diabetes mellitus, 6 had renal disease, 2 had thyroid disease, 183 had either no son or sons outside the required age range, 58 could not be contacted, 17 had no contact with their children, 78 had a cardiac history exceeding 5 years, and 63 could not be invited for other reasons (hospitalization, other serious diseases). Of 157 families (fathers and sons) who met all the criteria, 55 (either father or son) refused to participate in the study (response, 65%), leaving 102 fathers and 139 sons. The latter had oral lipid loading tests. The study protocol was approved by the medical ethics committees of the Zuiderziekenhuis Rotterdam and the University Hospital Rotterdam. Informed consent forms were obtained from all participants in the study.

Baseline Measurements

Fathers were asked to visit the hospital at 9:00 a.m. after fasting for at least 12 hours. Fathers responded to a questionnaire about the number of first-degree relatives who had had myocardial infarctions and about medication use at the time of the examination (medication was taken to the hospital where the examination took place). Systolic blood pressure and diastolic blood pressure were measured using a random zero sphygmomanometer (Hawksley, Lancing, United Kingdom). Fasting serum blood samples were drawn by antecubital venipuncture for measurement of levels of triglycerides, total cholesterol, low-density lipoprotein (LDL) cholesterol, and HDL cholesterol (and its subfractions HDL₂ and HDL₃). Height and weight were measured without shoes and without heavy clothing.

The sons were invited to come to the hospital, after the same period of fasting, on a separate day to have an oral lipid loading test. For sons, questionnaires were used to obtain data about use of medication, fat intake, alcohol intake, and smoking habits, referring to a 1-month period before the examination for this study. Daily total fat intake was calculated from an 81-item semiquantitative food frequency questionnaire by using a computerized food-composition table [18]. In the sons, blood samples were taken by antecubital venipuncture for measurement of baseline levels of serum triglycerides, total cholesterol, LDL cholesterol, HDL cholesterol (and its subfractions HDL₂ and HDL₃), apoprotein A-1, apoprotein A-2, apoprotein B, and retinyl palmitate concentrations. This baseline measurement of lipid levels was taken as the starting point (t₀) for the oral lipid loading test. In all of the sons, apolipoprotein E was phenotyped.

Oral Lipid Loading Test

Sons of patients and sons of controls came to the hospital at 7:45 a.m. after an overnight fasting of 12 hours. Height and weight were measured first to calculate body surface area. Five minutes after the venipuncture for obtaining baseline lipid levels (at t₀), all participants received a liquid lipid load, which consisted of a mixture of dairy cream (40% fat), egg yolk, milk powder, and retinyl palmitate (in aqueous solution) [15]. Participants received the lipid load in a dose based on their individual body surface area (77.5 g fat, 0.5 g cholesterol, and 27 000 IU of retinyl palmitate per square meter of body surface area). The mixture was consumed within 15 minutes.

The participants received an antecubital venous catheter (Venflon Viggo AB, Helsingborg, Sweden), which was kept open during the test period by means of disposable obturators (Venflon). Through this catheter, blood samples were drawn at 2 (t₂), 4 (t₄), 5 (t₅), 6 (t₆), 7 (t₇), 8 (t₈), 10 (t₁₀), and 12 (t₁₂) hours after starting consumption of the oral lipid load. Total cholesterol, triglyceride, and retinyl palmitate concentrations were determined in serum isolated from these samples. During the 12-hour period, other sources of calories were withheld from participants. Because postprandial exercise has been reported to decrease postprandial lipemia [19], participants stayed in the hospital and were asked to refrain from heavy physical activity during the test period.

Laboratory Analyses

Serum total cholesterol levels were measured using an automated enzymatic method (Boehringer Mannheim, Mannheim, Germany) CHOD-PAP reagent kit [20]. Levels of HDL cholesterol and LDL cholesterol were measured by the same method after precipitation. For HDL cholesterol, the phosphotungstate method according to Burstein [21], with a minor modification as described by Grove [22], was used. For LDL cholesterol, precipitation was carried out with polyvinylsulfate (Boehringer Mannheim). Throughout the entire study period, results of total cholesterol and HDL cholesterol determinations were within limits of the quality control program of the World Health Organization Regional Lipid Reference Centre (Prague, Czechoslovakia).

Levels of apoprotein A-1 and B were assayed using an automated immunoturbidimetric method (Kone Diagnostics, Espoo, Finland). Levels of apoprotein A-2 were determined by radial immunodiffusion against specific antiserum (Boehringer Mannheim, Germany) according to Cheung and Albers [23], with slight modifications. All automated analyses were carried out on the Kone Specific Analyzer (Kone Instruments) using frozen ( −20°C) serum samples. High-density lipoprotein₂ and HDL₃ in serum were assayed as described by Gidez and colleagues [24] with slight modifications. High-density lipoprotein₂ and HDL₃ were separated using stepwise precipitation of apoprotein B containing lipoproteins with heparin/Mn²⁺ and HDL₂ with dextran sulfate. Apolipoprotein E phenotyping was done by isoelectric focusing of delipidated serum followed by immunoblotting, using apolipoprotein E antiserum as first antibodies [25]. Retinyl palmitate analyses were done as described previously by Groot and colleagues [15].

Statistical Analysis

Means ± SDs were calculated for baseline characteristics of all family members. The total cholesterol/HDL cholesterol and LDL cholesterol/HDL cholesterol ratios were calculated for fathers and sons, and the apoprotein A-1/apoprotein B ratio was calculated for sons. Baseline differences between patients and controls as well as their respective sons were evaluated using the two group t-test or the Mann-Whitney U test when appropriate. Differences in postprandial responses between sons of patients and sons of controls with regard to triglycerides and retinyl palmitate were evaluated with an age-adjusted repeated-measures analysis to determine if group differences at each postprandial measurement time differed significantly from group differences at any other measurement time. Subsequently, linear regression analysis was used to compare mean group postprandial responses. Adjustments were made for age, HDL₂ cholesterol, HDL₃ cholesterol, apolipoprotein E phenotype, and daily total fat intake. This was done at each postprandial measurement time (t₂ to t₁₂) and for areas under the curve for several time (≥ 2 hour) intervals. In the analyses, individual postprandial measurements and areas under the curve were evaluated after subtracting the initial individual values (t₀) for triglyceride and retinyl palmitate levels from all respective postprandial measurements, yielding the net postprandial change.

Results

Two sons of patients did not tolerate the oral lipid because of gastrointestinal problems and were excluded from the study. All other participants tolerated the fat load well and did not have gastrointestinal symptoms or steatorrhea during or after the test. Two (other than already excluded) participants had baseline triglyceride values of 2.52 and 2.45 mmol/L, respectively. The first participant had six measurements of postprandial retinyl palmitate that were more than 3 to 5 standard deviations above the mean, and the latter participant had five measurements of postprandial triglyceride that were more than 3 to 5 standard deviations above the mean. Because of the marked effect of these outliers on the precision of results obtained in the group as a whole, these two participants were excluded from analyses.

Baseline characteristics are given for both groups of fathers and sons in Table 1. No statistically significant differences were noted between patients and controls. Sons of patients were slightly older and had a lower total fat intake. All patients (n = 59) had had coronary angiography for evaluation of symptoms of chest pain. For controls, indications for coronary angiography were symptoms of chest pain (n = 30), evaluation of cardiac valves (n = 7), suspected cardiomyopathy (n = 5), and ventricular septum defect (n = 1). At the time of coronary angiography, 19 (32.2%) patients and 11 (25.6%) controls were treated for hypertension. The average time elapsed between date of the coronary angiography and date of the examination was 13.2 ± 0.9 months (mean ± SE) for patients and 13.2 ± 1.1 months for controls.

Baseline serum measurements after an overnight fast are given for both groups in Table 2. As expected, patients had higher serum levels of total cholesterol, LDL cholesterol, and triglycerides and had higher total cholesterol/HDL cholesterol and of LDL cholesterol/HDL cholesterol ratios than did controls. Levels of HDL cholesterol and its subfractions tended to be slightly lower in patients than in controls (Table 2). Compared with sons of controls, sons of patients had somewhat less favorable lipoprotein and apoprotein profiles (not statistically significant). Sons of patients had significantly decreased levels of HDL₃ cholesterol (P = 0.016) and apoprotein A-2 (P = 0.001) when compared with controls. The frequencies of apolipoprotein E phenotypes did not differ between sons of patients and sons of controls. Apoprotein E2 alleles were present in 8 sons of controls and 6 sons of patients, 1 of whom was of the E2/E2 phenotype. This participant was not an outlier with regard either to baseline lipid profile or postprandial data.

The mean fat load ingested by sons of patients was 399.4 ± 3.2 g (mean ± SE) and by sons of controls was 395.1 ± 3.8 g. None of the sons in both groups had previous or current diseases or received medication known to have an effect on lipoprotein metabolism.

Figure 1 shows age-adjusted mean postprandial curves for levels of triglyceride, retinyl palmitate, and total cholesterol. Figure 1 (left panel) shows mean postprandial triglyceride curves for sons of patients and sons of controls. In both groups, oral lipid loading led to markedly similar increases in triglyceride concentrations, maximum levels appeared between 4 to 6 hours postprandially. After 12 hours, triglyceride concentrations had decreased below initial levels. The age-adjusted repeated-measures analysis showed an interaction effect between group and time (chi-square test = 19.2; P = 0.014), indicating that the two curves were significantly different from each other.

In Figure 1 (middle panel) the mean postprandial retinyl palmitate curves are shown. Retinyl palmitate concentrations increased from virtually zero in both groups to maximum concentrations between 5 and 6 hours postprandially. Thereafter, retinyl palmitate concentrations decreased to levels at t₁₂ that were approximately tenfold the initial levels at t₀. The repeated-measures analysis for these curves indicated no significant difference (chi-square test = 6.42; P = 0.60).

In Figure 1 (right panel), curves are shown for postprandial total cholesterol responses. Sons of patients had higher mean total cholesterol levels at t₀ than did sons of controls, and this difference remained remarkably constant during the lipid loading test, showing no significant response to the intervention. The age-adjusted difference in mean cholesterol levels (mean of all cholesterol measurements) was 2.3 ± 5.87 mg/dL (mean ± SE; P = 0.70).

A slight difference in age was noted between sons of patients and sons of controls. Age has been shown to have an effect on postprandial lipoprotein changes [26] and postprandial retinyl ester response [27], and therefore the differences in postprandial response between sons of patients and sons of controls were adjusted for age. Table 3 shows the age-adjusted differences between net postprandial changes in triglyceride and retinyl palmitate responses between sons of patients and sons of controls. Triglyceride response differences showed a maximum at t₅ and reached significance at times t₈, t₁₀, and t₁₂. With respect to retinyl palmitate, sons of patients and sons of controls showed a slight difference in response, increasing to a maximum at t₅ to t₆. None of the retinyl palmitate differences was found to be statistically significant.

In Figure 2A, age-adjusted areas under the curves for 2-hour time intervals are given for postprandial triglyceride responses in sons of patients and sons of controls. Peak areas for both groups were found in the interval t₄ to t₆. Differences were mainly found in intervals t₆ to t₈, t₈ to t₁₀, and t₁₀ to t₁₂. In the latter interval, the triglyceride response of sons of controls decreased below t₀ (baseline) levels to a greater extent than in sons of patients. Figure 2B shows total area under the curves for 10-, 8-, 6-, and 4-hour time intervals up to t₁₂. For all intervals, areas under the curve differed significantly between groups. Figures 2C and 2D show similar data for retinyl palmitate responses in the groups. For 2-hour intervals and larger intervals, areas under the curves for sons of patients tended to be higher, although differences did not reach statistical significance.

The results for postprandial retinyl palmitate levels as well as triglyceride levels did not materially change after adjustment for baseline HDL₂ cholesterol levels, HDL₃ cholesterol levels, and apolipoprotein E phenotype. Total fat intake was decreased in sons of patients compared with sons of controls (P = 0.012; Table 1). Adjustment for this difference in postprandial analyses did not change the results, with adjusted group differences in triglyceride levels at t₈ of 0.32 mmol/L (95% CI, 0.04 to 0.59), at t₁₀ of 0.2 mmol/L (CI, 0.047 to 0.36), and at t₁₂ of 0.1 mmol/L (CI, 0.002 to 0.26). All sons were asked whether they had changed their lifestyles with regard to smoking habits, physical activity, and fat intake since their fathers had had cardiac examination. None of the group differences was statistically significant. Also, a similar analysis of sons who reported to live separately from their fathers at the time of examination (45 sons of patients [56%] and 24 sons of controls [42%]) yielded no significant group differences with respect to lifestyle changes. An analysis of postprandial data in these groups showed similar results, with sons of patients having higher triglyceride levels at t₈ (difference, 0.47 mmol/L; CI, −0.018 to 0.95) and t₁₀ (difference, 0.29 mmol/L; CI, 0.015 to 0.57). When analyses were done for participants with normal levels of triglycerides only (serum triglyceride ≤ 2.28 mmol/L), thereby excluding 13 participants, no major changes in the results were found.

Discussion

Familial aggregation of risk factors for coronary artery disease is well established, and the offspring of patients with very severe coronary artery disease are at increased risk for this disease [28, 29]. We assessed whether this increased familial risk was related to postprandial metabolic handling of an oral fat load in healthy young male adults whose fathers had severe coronary artery disease. Baseline measurements showed that sons of patients and sons of controls had similar levels of lipids and apolipoproteins, only levels of HDL₃ cholesterol and apoprotein A-2 were significantly decreased in sons of patients.

After oral lipid loading, however, prolonged postprandial hypertriglyceridemia was observed in sons of patients, amounting to statistically significant group differences of 0.35 mmol/L after 8 hours, 0.21 mmol/L after 10 hours, and 0.13 mmol/L after 12 hours. Differences in postprandial triglyceride levels between groups were mainly observed in the declining part of the curves, which suggests a delayed clearance of triglyceride from plasma. Data on postprandial retinyl palmitate responses also suggest similar differences for postprandial triglyceride levels with regard to the direction of the difference and the postprandial period in which differences appeared. However, these latter differences were not statistically significant.

A marked contrast between patients with coronary artery disease and controls with regard to angiographic findings was chosen to minimize the chance of misclassification of patients and also because of the anticipated subtle differences between both groups of offspring. Considering the efforts asked on a strictly voluntary basis from fathers, and particularly their offspring, the response rate of 65% in this study is satisfactory. However, it is possible that postprandial lipoprotein metabolism in these young healthy participants was in some way differentially related to the response in both groups and, although unlikely, it is possible that selection mechanisms have affected the findings. Because contraceptive steroids have an effect on postprandial lipid metabolism in young healthy women [30], we restricted the study to sons.

Chylomicron metabolism is thought to occur essentially in two successive steps. First, these particles interact with lipoprotein lipase, an enzyme that is located on capillary endothelial surfaces of various tissues, including heart, skeletal muscle, and adipose tissue [31]. This lipase catalyzes the hydrolysis of triglycerides, and this results in uptake of free fatty acids by tissues [32]. Thereafter, resulting chylomicron remnant particles are taken up relatively quickly and efficiently [33] by parenchymal cells in the liver. The importance of this process, with regard to atherogenesis, is that any hampering in this course of events might lead to higher plasma levels of atherogenic triglyceride-rich lipoproteins and their remnants.

In principle, the oral lipid loading test cannot distinguish between mechanisms involved in chylomicron entry into and clearance of chylomicrons and their remnants from the circulation. In our study, no attempt was made to assess differences between intestinal lipid absorption in the two groups of young adult men. In a study of postprandial lipoprotein metabolism, Weintraub and colleagues [13] showed that gastrointestinal transit time was similar in normal participants and participants with different types of hyperlipoproteinemias. Previous studies [14-16] showed that mean triglyceride and mean retinyl palmitate response curves are similar during the first 4 hours postprandially when comparing patients and controls, which we also found in our study in sons of such patients. This suggests that differences in fat absorption or in intestinal secretion, if any, are not likely to explain group differences during the second half of the test period.

Retinyl palmitate is a useful marker of the metabolism of intestinally derived triglyceride-rich lipoproteins and their remnants because it is incorporated into the core of these lipoproteins [34] and is retained within them until their irreversible clearance from plasma by the liver [35]. Although considerable postprandial exchange between lipoproteins of total retinyl esters has been reported in humans [36], little exchange of retinyl palmitate between intestinally derived lipoproteins and other lipoproteins has been shown [37, 38]. In our study, retinyl palmitate responses showed considerable interindividual variation in both groups, which may explain why differences between group responses did not reach statistical significance. An alternative explanation might be that the postprandial retinyl ester response is greater in older persons than in younger persons [27], thus making it more difficult to detect group differences in the young. Further, postprandial hypertriglyceridemia may partly be induced by endogenous very-low-density lipoprotein and its remnants [39], which do not contain retinyl palmitate. Finally, significant group differences in levels of retinyl palmitate may emerge after the 12-hour measuring period used in the present study and, thus, may have been missed [14].

In our study, data were analyzed after subtracting individual fasting triglyceride levels and retinyl palmitate levels from respective subsequent measurements. Differences in body height and weight were considered by standardizing the oral lipid load to body surface area as a function of body height and weight. Possible confounding effects of age, HDL (₂) cholesterol levels, HDL₃ cholesterol levels, and apolipoprotein E phenotype were adjusted for in the analysis. Apolipoprotein E polymorphism has recently been shown to have an effect on postprandial retinyl palmitate concentrations [40]. Adjustment for these factors did not materially change results with regard to group differences.

The hypothesis that postprandial lipoprotein metabolism plays an essential role in atherogenesis was first postulated by Zilversmit [3]. This hypothesis was addressed in recent studies in humans [14-16] in which patients with coronary artery disease were compared with controls with respect to postprandial lipemia; the study found that patients had higher postprandial triglyceride and retinyl palmitate responses. In another study, Simons and colleagues [41] found that 4 hours postprandially the apoprotein B48/apoprotein B100 ratio, used as a relative measure of chylomicrons and chylomicron remnants in plasma, was higher in patients with coronary artery disease than in controls. In addition, a positive correlation between the postprandial triglyceride response to an oral lipid load and carotid artery wall thickness has been reported [42]. These studies suggest that prolonged presence of triglyceride-rich lipoproteins in plasma may indeed play a role in atherogenesis.

The design of our investigation enabled us to examine a group of young adult men at high familial risk for developing clinical manifestations of coronary artery disease much later in life. One advantage of the design of this study is that sons of these patients are not exposed to treatments such as medication that could alter triglyceride metabolism, as might occur when patients themselves are studied. However, if patients have been given advice about lifestyle, this may in turn have led to lifestyle changes in their sons. In both groups of sons, most reported not to have changed their levels of physical activity, smoking behavior, and fat intake since their fathers were examined for cardiac symptoms; approximately equal proportions reported to eat less fat. Additionally, an analysis of sons living separately from their fathers, who are assumed to be less influenced by lifestyle advice given to their fathers, yielded similar results with statistically significant differences in the declining part of the triglyceride curves, despite the considerably decreased number of participants. Although no major differences were noted with regard to reported fat intake, daily total fat intake as calculated from the food frequency questionnaire was significantly decreased in sons of patients than in sons of controls. Long- and short-term effects of fat intake on postprandial lipoprotein levels have been reported [43], but in our study, adjustment for daily total fat intake did not change postprandial results. Thus, we believe that changes in physical activity, smoking behavior, and fat intake are not likely to explain postprandial differences found between both groups of sons.

Our data show that healthy young sons whose fathers have established coronary artery disease have prolonged postprandial hypertriglyceridemia. This finding suggests that changes in postprandial lipid metabolism are associated with familial risk for coronary atherosclerosis.