Clearance of chylomicron remnants in normolipidaemic patients with coronary artery disease: case control study over three yearsBMJ 1996; 312 doi: https://doi.org/10.1136/bmj.312.7036.935 (Published 13 April 1996) Cite this as: BMJ 1996;312:935
- Moshe S Weintraub, head, department of internal medicine Ca,
- Itamar Grosskopf, instructor, department of medicine Ca,
- Toby Rassin, instructor, department of medicine Ea,
- Hylton Miller, head, coronary catheterisation unita,
- Gideon Charach, instructor, department of medicine Ca,
- Heschi H Rotmensch, senior lecturer, department of medicine Ca,
- Meir Liron, head, department of medicine Ea,
- Ardon Rubinstein, head, unit of lipid metabolisma,
- Adrian Iaina, head, department of nephrologya
- a Tel Aviv-Sourasky Medical Center and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
- Correspondence and requests for reprints to: Dr Weintraub.
- Accepted 12 January 1996
Objective: To test the hypothesis that subjects who clear chylomicron remnants slowly from plasma may be at higher risk of coronary artery disease than indicated by their fasting plasma lipid concentrations.
Design: Case control study over three years.
Setting: An 800 bed general municipal hospital.
Subjects: 85 normolipidaemic patients with coronary artery disease selected prospectively and matched with 85 normolipidaemic subjects with normal coronary arteries on angiography.
Interventions: All subjects were given a vitamin A fat loading test which specifically labels intestinal lipoproteins with retinyl palmitate.
Main outcome measure: Postprandial lipoprotein metabolism.
Results: The area below the chylomicron remnant retinyl palmitate curve was significantly increased in the coronary artery disease group as compared with the controls (mean 23.4 (SD 15.0) v 15.3 (8.9) µmol/l.h; 95% confidence interval of difference 4.37 to 11.82).
Conclusions: Normolipidaemic patients with coronary artery disease had significantly higher concentrations of chylomicron remnants in plasma than normolipidaemic subjects with normal coronary vessels. This may explain the mechanism underlying the susceptibility to atherosclerosis of coronary artery disease patients with normal fasting lipid values. As diet and drugs can ameliorate the accumulation of postprandial lipoproteins in plasma, the concentration of chylomicron remnants should be measured in patients at high risk of coronary artery disease.
The atherogenic effect of postprandial lipoproteins is hypothesised to be inversely related to their metabolic capacity
Patients with coronary artery disease have decreased metabolic capacity of chylomicron remnants as compared with subjects with normal coronary arteries
An ensuing postprandial dyslipidaemia may exist in patients with coronary artery disease who have normal fasting lipid values
A vitamin A fat loading test may be useful in identifying these patients
Measurements in fasting subjects show that plasma lipid and lipoprotein abnormalities are associated with coronary and peripheral atherosclerosis.1 2 3 4 However, most of our lives are spent in the postprandial state, during which vessel walls are exposed to triglyceride rich lipoproteins—namely, chylomicrons and chylomicron remnants. Animal and human studies5 6 7 8 9 10 suggest that these lipoproteins may be atherogenic. They are metabolised on the endothelial surface of arteries and their cholesterol becomes incorporated into the artery wall, where it may stimulate the formation of atherosclerotic lesions.11 12 13 14 15 The accumulation of postprandial lipoproteins in plasma due to slowed catabolism may therefore enhance atherosclerosis in otherwise normolipidaemic subjects.
This picture is not detected by examining fasting state plasma samples only. Zilversmit was the first to publish this hypothesis,5 which was followed by many confirmatory studies. Most, however, were small and provided only indirect support by showing correlations between postprandial lipoprotein metabolism and other well known risk factors for coronary artery disease.16 17 18 19 20 Only one study found that coronary artery disease patients with normal fasting lipid values had higher concentrations of postprandial lipoproteins than patients without coronary artery disease.21 That study, however, included only 20 patients.
There are two main difficulties in testing Zilversmit's hypothesis. The first is in differentiating postprandial lipoproteins from endogenous very low density lipoproteins, which have similar physical and chemical properties. The second is the need to study a large number of coronary artery disease patients with normal fasting lipid concentrations and an age and sex matched control group. We overcame the first difficulty by using the vitamin A fat loading test, which specifically labels postprandial lipids with retinyl palmitate.22 23 24 25 26 27 28 This method reliably detects both chylomicron and chylomicron remnant metabolism.16 18 19 20 29 With respect to the size of the study we selected 85 normolipidaemic subjects with radiologically normal coronary arteries (controls) and matched them for age and sex with 85 normolipidaemic patients with advanced coronary artery disease. We investigated all participants for metabolic behaviour of postprandial lipoproteins.
Subjects and methods
The 85 patients with coronary artery disease and 85 controls were selected prospectively from the cardiology department of this hospital during February 1990 and April 1993. All patients and controls had normal fasting lipid and low density lipoprotein cholesterol concentrations. All had undergone coronary angiography because of chest pain suspected to be of coronary origin and all had stable cardiovascular and metabolic variables and were free of congestive heart failure, diabetes mellitus, chronic renal failure, and endocrine disease. No women patient was taking contraceptive or hormone replacement drugs.
Selection criteria with regard to lipid and lipoprotein concentrations were: total cholesterol less than 6.2 mmol/l, triglycerides less than 2.0 mmol/l, and low density lipoprotein cholesterol less than 4.1 mmol/l. High density lipoprotein cholesterol concentrations were not used as a criterion for exclusion. Coronary vessels with less than 20% stenosis were defined as normal. Patients in the coronary artery disease group had two or three vessel disease with at least 80% occlusion.
MATCHING CRITERIA AND METHOD OF SELECTION
During the three years we selected 85 subjects with normal coronary vessels and 192 patients with coronary artery disease who fulfilled the study criteria. In most cases we successfully matched the two groups according to the following criteria with priority in descending order: sex, age, body surface area, hypertension, drugs being taken, and cigarette smoking. Hypertension was defined as systolic blood pressure >150 mm Hg and diastolic pressure >90 mm Hg treated for at least two years. Cigarette smoking was defined as consumption of more than 10 cigarettes daily for at least five years. All the participants claimed to have stopped smoking during the study. The two groups were matched for treatment with β blockers (treatment equivalent 60-160 mg propranolol daily), angiotensin converting enzyme inhibitors (treatment equivalent 12.5-50.0 mg captopril daily), and calcium inhibitors (verapamil 120-240 mg and nifedipine 30-60 mg daily) because of the known effects of these drugs on lipid metabolism. Patients taking diuretics or very high doses of the above drugs were excluded.
Finding normal coronary vessels on angiography is fairly rare. Whenever such patients were detected they were asked to continue the drugs they were taking for suspected angina until they could have a vitamin A fat loading test. This was never less than one week after the angiography. Thereafter, a patient with proved coronary artery disease who had undergone coronary angiography and who fulfilled the inclusion criteria was matched as closely as possible with a non-coronary artery disease patient and also given the vitamin A fat loading test. Of the 192 patients with coronary artery disease, 85 were so matched to controls during the three years. All participants in the study gave fully informed consent.
VITAMIN A FAT LOADING TEST
The vitamin A fat loading test was performed as described.19 After a 12 hour overnight fast participants were given a fatty meal plus 60000 units of aqueous vitamin A/m2 body surface area. The meal contained 50g fat/m2 body surface area and 145 mg cholesterol/MJ 65% of the energy being taken as fat, 20% as carbohydrate, and 15% as protein. Blood samples were drawn before the meal, hourly thereafter for eight hours, and at 10 hours. Subjects tolerated the meal well and none had diarrhoea or other symptoms of malabsorption.
ANALYSIS OF SAMPLES
Venous blood was drawn from the forearm and transferred to a tube containing sodium EDTA. Samples were centrifuged immediately at 1500 g for 15 minutes and 1 ml plasma stored wrapped in foil at -20°C for retinyl ester assay. A further 0.5 ml was stored at 4°C for triglyceride determinations. An aliquot of 2.5 ml plasma was subjected to preparative ultracentrifugation to separate chylomicron fraction (>1000 Sf units) from non-chylomicron fraction (<1000 Sf units).29 30 31 32
RETINYL ESTER ASSAY AND LIPID AND LIPOPROTEIN DETERMINATIONS
The assays were carried out in subdued light with high performance liquid chromatography grade solvents as described.16 18 19 20 Cholesterol and triglyceride concentrations were measured enzymatically by using the reagents cholesterol 236991 and triglyceride 126012 (Boehringer Mannheim, Indianapolis, United States). High density lipoprotein cholesterol was measured after precipitation of whole plasma with dextran sulphate magnesium.
The significance of differences between the groups in clinical characteristics and fasting plasma lipid and lipoprotein concentrations was determined by χ2 test for discrete variables and a two sample t test for continuous variables. The significance of metabolic responses after the fatty meal was analysed by assuming the same behaviour of the curves of plasma triglyceride, chylomicron retinyl palmitate, and non-chylomicron retinyl palmitate values in the two groups. Areas under the curves were calculated for each patient and the Mann-Whitney test used to detect differences between the groups.33 A non-parametric test was used because data were skewed. Spearman's test was used to evaluate the correlation between chylomicron retinyl palmitate and non-chylomicron plasma retinyl palmitate concentrations, triglyceride responses, and fasting lipid and lipoprotein concentrations, which were not normally distributed.
Stepwise multiple regression analysis was performed to adjust the metabolic response after the fatty meal for the difference in intake of nitrates and aspirin between the coronary artery disease patients and the controls. Analysis of covariance was performed to adjust the metabolic response for the difference in baseline triglyceride and high density lipoprotein cholesterol concentrations between the groups.
Table 1 presents the clinical features of the coronary artery disease patients and controls. The groups were similar in age, body surface area, male to female ratio, fasting total cholesterol and low density lipoprotein cholesterol concentrations, cigarette smoking, and prevalence of hypertension. The coronary artery disease group was significantly different from the controls in fasting plasma triglyceride (mean 1.6 (SD 0.3) v 1.4 (0.4) mmol/l; P<0.001) and high density lipoprotein cholesterol concentrations (1.1 (0.2) v 1.4 (0.4) mmol/l; P<0.0001), though no patient was hypertriglyceridaemic. Significantly more coronary artery disease patients than controls were taking aspirin (77 v 43; P<0.0001) and nitrates (46 v 21; P<0.0001). Thirty two coronary artery disease patients had two vessel disease and 53 three vessel disease.
RESPONSES TO VITAMIN A FAT LOADING TEST
Plasma, chylomicron, and non-chylomicron retinyl palmitate and plasma triglyceride responses to the vitamin A fat loading test were determined for each participant. Figures 1 and 2 give the mean results in each group. Table 2 shows the mean areas below the retinyl palmitate and triglyceride curves of the different fractions in the two groups. The area below the plasma retinyl palmitate curve was significantly larger in the coronary artery disease patients than in the controls (mean 65.3 (SD 28.3) v 55.0 (27.5) µmol/l.h; P=0.017). This increase in postprandial lipoprotein concentration in the coronary artery disease group was due mainly to the non-chylomicron fraction, whose area under the curve was highly significantly larger than in the controls (23.4 (15.0) v 15.3 (8.9) µmol/l.h; P<0.001).
There was also evidence of higher postprandial plasma triglyceride concentrations in the coronary artery disease patients than in the controls. The areas below the plasma triglyceride concentration curves in the two groups were 23.6 (SD 7.2) and 21.1 (9.2) mmol/l.h respectively (P=0.055). No significant differences were found in the chylomicron fractions between the two groups. Adjustment for baseline differences in triglyceride and high density lipoprotein cholesterol concentrations showed a non-significant difference in area below the plasma retinyl palmitate curve between the coronary artery disease patients and the controls (P=0.43 after adjustment for triglyceride concentrations; P=0.51 after adjustment for high density lipoprotein cholesterol concentrations). However, the significant difference in area below the non-chylomicron retinyl palmitate curve between the coronary artery disease patients and the controls was independent of baseline differences in triglyceride (P<0.001) and high density lipoprotein cholesterol concentrations (P<0.001).
To determine factors that might play a part in exogenous fat metabolism and atherosclerosis we sought correlations between chylomicron and non-chylomicron retinyl palmitate and plasma triglyceride responses and fasting lipid and lipoprotein concentrations (table 3). Significant positive correlations were found between fasting triglyceride concentrations and chylomicron as well as chylomicron remnant fractions in all participants (r=0.50 and r=0.42 respectively; P<0.001). Negative correlations were found between fasting high density lipoprotein cholesterol concentrations and chylomicron as well as chylomicron remnant retinyl palmitate areas and areas below plasma triglyceride curves in all participants (r=-0.31, r=-0.32, and r=-0.43 respectively; P<0.001)
The significant difference in area below the non-chylomicron retinyl palmitate curve between the coronary artery disease patients and the controls (P<0.0001) was not explained by the different intake of nitrates (P=0.078) or aspirin (P=0.075).
This study shows clearly that coronary artery disease patients with normal fasting lipid and low density lipoprotein cholesterol concentrations have significantly higher concentrations of chylomicron remnants in plasma than non-coronary artery disease patients with normal fasting lipid values. The findings therefore strongly support Zilversmit's hypothesis that atherogenesis is a postprandial phenomenon.5
Postprandial lipoprotein metabolism is believed to occur in two stages. Initially, the chylomicrons produced in the intestines which carry exogenous lipids interact with lipoprotein lipase in extrahepatic tissues. This results in triglyceride hydrolysis and the delivery of free fatty acids to the tissues.11 34 35 After most of the triglycerides are hydrolysed remnant particles are formed,36 37 which are removed from the circulation by hepatocyte receptors that recognise apolipoprotein E.38 39 Though considerable post-prandial exchange between lipoproteins of total retinyl esters has been reported in humans,40 little exchange of retinyl palmitate has been shown between intestine derived lipoproteins and other lipoproteins.23 Also, most of the exchange of retinyl esters was found in the low density lipoprotein fraction. Our study groups had similar low density lipoprotein concentrations. Hence this exchange would not affect the difference found in the non-chylomicron fractions of the two groups.
Karpe et al reported a delayed appearance of retinyl palmitate compared with the patterns of increase of plasma triglycerides and apolipoprotein B-48.41 This finding questions the validity of retinyl palmitate labelling of chylomicrons. In hundreds of tests performed by us16 18 19 20 29 and in many other studies21 22 23 24 25 26 27 28 plasma triglyceride and retinyl palmitate concentrations peaked simultaneously. This disparity may be explained by the different methods used. Karpe et al used exclusively soybean oil in their fat load meal whereas we and others used dairy cream. Poly-unsaturated fat delays the absorption of vitamin A.42
We think that the vitamin A fat loading method as used by us in the past eight years is an efficient tool for studying postprandial lipoprotein metabolism. Results with this method indicate that coronary artery disease patients clear chylomicron remnants at a significantly slower rate than do subjects with normal coronary vessels. Chylomicron remnants accumulated in the circulation may be particularly atherogenic. In type III hyperlipoproteinaemia accumulation of these fat particles due to delayed clearance is associated with severe premature atherosclerotic disease.6 8 Thus our findings may explain the mechanism underlying the susceptibility to atherosclerosis in the subgroup of coronary artery disease patients with normal fasting lipid values
EFFICIENCY OF CATABOLIC MECHANISM
Though fasting lipid and lipoprotein concentrations were normal in our two groups of patients, the coronary artery disease group had significantly higher concentrations of fasting triglycerides and lower concentrations of high density lipoprotein cholesterol. This disparity, however, had no influence on non-chylomicron metabolism. All participants showed significant positive correlations between fasting triglyceride concentrations and postprandial lipoprotein areas under the curve (P<0.001) and significant negative correlations between fasting high density lipoprotein cholesterol concentrations and postprandial lipoprotein concentrations. These observations were reported in most earlier studies on postprandial lipoprotein metabolism and accord with current concepts on the relations between the metabolic rates of these lipoproteins, the fasting concentrations of triglycerides, and the surface remnant transfer to the high density lipoprotein system.43 44
Could the difference in fasting triglyceride and high density lipoprotein concentrations rather than the difference in remnant clearance in the two groups be the determining factor in coronary artery disease? The catabolic pathways of all the triglyceride rich lipoproteins, endogenous very low density lipoproteins, intermediate density lipoproteins, and exogenous chylomicrons and chylomicron remnants are identical. They are lipolysed by the same enzyme—namely, lipoprotein lipase—and taken up by the same receptors. It seems that this catabolic mechanism may be less efficient in patients with coronary artery disease. Measuring only fasting lipid and lipoprotein values will not detect this “defect.” However, in the postprandial state, when the catabolic pathways are confronted by an increased mass of exogenous triglycerides and lipoproteins, this defect is augmented and can be detected by the fat loading test.
Thus in patients with coronary artery disease high normal fasting plasma triglyceride concentrations, increased concentrations of chylomicron remnants, and low concentrations of high density lipoprotein cholesterol are due to the same inefficient catabolic mechanism. Borderline concentrations of fasting triglycerides and high density lipoprotein cholesterol should therefore raise suspicion of a possible accumulation of postprandial lipoproteins, especially in high risk patients. In these cases a vitamin A fat loading test can be done. Though the test is cumbersome and time consuming, it may be simplified. We find that testing one or two blood samples drawn seven or eight hours or both after a fat load may detect a defect in chylomicron remnant removal. Testing a single blood sample drawn at five or six hours will detect defects in chylomicron lipolysis.
In conclusion our study shows that there is a substantial accumulation of chylomicron remnants in patients with premature coronary heart disease and normal fasting lipid and low density lipoprotein cholesterol concentrations. It seems that the arteries of these patients are exposed to high concentrations of chylomicron remnants for prolonged periods. This may have a role in the aetiology of the disease, as first suggested by Zilversmit.5 The accumulation of chylomicron remnants can be ameliorated dramatically by diet or fibric acid derivatives.29 The vitamin A fat loading test as described or a simplified version may be useful for identifying these patients.
We thank Professor Jan L Breslow, of the Rockefeller University, New York, for help in planning the study and support in developing the methods for studying postprandial lipoprotein metabolism.
Funding The United States-Israel Binational Science Foundation (grant 87-00084).
Conflict of interest None.