- Email: [email protected]
A longitudinal study of the biological variability of plasma lipoproteins in healthy young adults
75
Atherosclerosis, 34 (1979) 75-81 @ Elsevier/North-Holland Scientific
Publishers,
Ltd.
A LONGITUDINAL STUDY OF THE BIOLOGICAL VARIABILITY PLASMA LIPOPROTEINS IN HEALTHY YOUNG ADULTS
O.D. MJ@S *, S.N. RAO, L. BJ@RU, T. HENDEN, N.E. MILLER
D.S. THELLE,
OF
O.H. F(bRDE and
Institutes of Medical Biology and Community Medicine, University of Tromsd, Tromsl, Norway; and Department of Chemical Pathology and Metabolic Disorders, St Thomas’ Hospital Medical School, London, England (Received 31 May, 1979) (Revised, received 14 June, 1979) (Accepted 15 June, 1979)
Summary The fasting serum concentrations of total cholesterol, low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), total triglyceride and apoprotein Al were measured at intervals of 12-18 weeks for 60 weeks in 17 male and 11 female healthy young adults in order to assess the variability of these risk factors for coronary disease. No statistically significant seasonal changes were detected in any variable in either sex, although a progressive rise in apoprotein Al concentration was observed. The coefficients of variation for random fluctuations with time were in the rank order: total cholesterol < HDL-C < apoprotein Al < LDL-C < triglyceride. These differences were attributable to biological, rather than to methodological, factors. Within subjects, HDL cholesterol concentration varied inversely with triglyceride concentration and directly with apoprotein Al concentration. The marked differences which exist in the biological variability of lipid risk factors for atherosclerosis need to be taken into account when making comparisons in epidemiological studies of the predictive powers of single onentry measurements for future disease. Fluctuations of HDL-C with time appear to be related in part to variations in triglyceride-rich lipoprotein metabolism. Key words:
Apoprotein Al - Cholesterol protein - Low density lipoprotein
* Requests for reprints to: Professor Troms4. Norway.
Coronary heart disease - High density - Triglyceride
O.D. Mjgs. Institute of Medical Biology,
University
of ‘horns&
lipo-
9000
76
Introduction Susceptibility to coronary heart disease (CHD) has long been recognised to be positively correlated with the plasma total cholesterol concentration [ 1,2], and in some communities also with the plasma triglyceride concentration [ 21. Recent studies have confirmed that the correlation between CHD and plasma cholesterol reflects their underlying relationships to the low density lipoprotein (LDL) concentration [3,4]. On the other hand, the correlation between CHD and plasma triglyceride’ is at least partly indirect, reflecting their separate inverse relationships to the plasma high density lipoprotein (HDL) cholesterol level [4,5]. Low concentrations of apoprotein Al, the major protein component of HDL, appear also to be predictive of future CHD [6]. The recognition that CHD risk shows independent relationships to different plasma lipoproteins has led to comparisons of their predictive powers, in an attempt to assess their relative importance in the pathogenesis of coronary atherosclerosis [4-6]. The predictive power of a single measurement of any lipid risk factor, however, will depend not only on its precise relationship to the disease process (e.g. whether causal or non-causal), but also on its biological variability and the precision with which the measurement is made. We have therefore studied the variability of several lipid risk factors for CHD in healthy young males and females in a longitudinal study of 60 weeks duration. In view of the relative paucity of information concerning the determinants of HDL cholesterol concentration, its relationships to the triglyceride, LDL cholesterol and apoprotein Al concentrations as a function of time were also examined. Methods Subjects Seventeen male and 11 female medical students were studied. Ages were 22-29 years (mean 25 years), body weights 48-77 kg (mean 65 kg) and relative body weights 1.80-2.54 g/cm* (mean 2.12 g/cm*). All were healthy and free of vascular disease. Three had a paternal history of myocardial infarction at age 54-64 years. All but 3 (females) were non-smokers. Clinical procedures The subjects were requested to adhere to their usual living habits for the duration of the study. Each attended for venepuncture, after an overnight fast, on five occasions: during the 6th, 21st, 36th .and 48th weeks of 197’7 and the 14th week of 1978. Venepuncture of females was performed without regard to the menstrual cycle. Blood was allowed to clot at room temperature, and the serum deep frozen (-20°C) until analysis. Laboratory procedures At the end of the study period, the serum samples were thawed and.assayed for total cholesterol and triglyceride by enzymatic procedures, as previously
77
described [ 51. Serum HDL cholesterol was measured after precipitation of VLDL and LDL with heparin and Mn2+ [5]. The LDL cholesterol concentration was estimated from these values using the formula of Friedewald et al.
[71. Serum apoprotein Al concentration was measured by electroimmunoassay, using a modification of the method of Curry et al. [8] (1.5% agarose in 50 mM sodium barbitone buffer, pH 8.6; 8 V/cm for 16 h). Monospecific antiserum to human apoprotein Al was produced in rabbits according to Albers et al. [ 91. Within-batch coefficients of variation of replicate measurements made on the same serum samples were similar for the 4 assay procedures: 2.5% for total cholesterol, 2.7% for HDL cholesterol, 2.2% for total triglyceride, and 3.0% for apoprotein Al.
Statistics The presence or absence of seasonal variation in each of the 5 variables was assessed by Bayesian growth curve analysis [lo], using the pooled data from all 28 subjects at all time points. The relationships between HDL cholesterol and the other variables as a function of time were explored by linear regression analysis. Triglyceride concentrations were converted to their logarithms prior to analysis, in view of their skewed distribution [ 111. Results The mean values and standard deviations of each variable at each time point in males are given in Table 1. Results for females appear in Table 2. Bayesian analysis confirmed the visual impression that there were no statistically significant seasonal variations. In males apoprotein Al concentration showed a significant (P < 0.05) upward trend throughout the study period. A similar trend was observed in females, but this was not statistically significant.
TIZBLE 1 FASTING SERUM TOTAL CHOLESTEROL, TOTAL TRIGLYCERIDE, HDL CHOLESTEROL, LDL CHOLESTEROL AND APOPROTEIN Al CONCENTRATIONS IN 17 MALE SUBJECTS DURING A PERIOD OF 60 WEEKS. RESULTS ARE GIVEN AS: MEAN (SD) Measurement
Week of the year 6
21
36
48
14
Total cholesterol
4.65 (0.79)
4.71 (0.73)
4.74 (0.79)
4.65 (0.67)
4.54 (0.73)
(mmolfi) Total triglyceride
0.78 (0.34)
0.85 (0.34)
0.84 (0.37)
0.99 (0.51)
0.84 (0.36)
(mmoI/D HDL cholesterol
1.48 (0.32)
1.60 (0.35)
1.47 (0.27)
1.53 (0.30)
1.56 (0.34)
(mmoI/D LDL cholesterol
2.80 (0.62)
2.75 (0.59)
2.81 (0.59)
2.67 (0.56)
2.53 (0.47)
(mmol/D Apoprotein Al
119
129
133
142
144
(miz/dD
(17.7)
(18.7)
(18.8)
(18.5)
(20.6)
TABLE 2 FASTING SERUM TOTAL CHOLESTEROL, TOTAL TRIGLYCERIDE, HDL CHOLESTEROL, LDL CHOLESTEROL AND APOPROTEIN Al CONCENTRATIONS IN 11 FEMALE SUBJECTS DURING A PERIOD OF 60 WEEKS. RESULTS ARE GIVEN AS: MEAN (SD) Measurements
Week of the year 6
21
36
40
14
Total cholesterol
4.96 (0.91)
5.04 (0.96)
5.03 (0.95)
4.67 (0.96)
4.80 (0.75)
(mmoI/I) Total triglyceride
0.69 (0.13)
0.79 (0.17)
0.83 (0.33)
0.75 (0.17)
0.65 (0.17)
(mmoI/l) HDL cholesterol
1.62 (0.44)
1.73 (0.31)
1.69 (0.36)
1.70 (0.40)
1.77 (0.38)
(mmoI/l) LDL cholesterol
2.91 (0.59)
3.04 (0.86)
2.92 (0.68)
2.65 (0.65)
2.60 (0.69)
(mmoI/l) Apoprotein
128
131
135
145
143
Al
(9.7)
(12.6)
(10.4)
(15.3)
(17.7)
(mg/dI)
TABLE 3 MEAN WITHIN-SUBJECT COEFFICIENTS OF VARIATION FOR FASTING SERUM TOTAL CHOLESTEROL, HDL CHOLESTEROL, LDL CHOLESTEROL, TOTAL TRIGLYCERIDE AND APOPROTEIN Al CONCENTRATIONS IN 28 SUBJECTS OVER A 60-WEEK PERIOD
Males Females
Total cholesterol
HDL cholesterol
LDL cholesterol
Total triglyceride
Apoprotein Al
7.80 6.70
0.42 9.42
10.4 11.8
21.4 30.4
8.33 10.2
TABLE 4 PELATIONSHIPS OF THE SERUM HDL CHOLESTEROL CONCENTRATION TO THE SERUM TOTAL TRIGLYCERIDE, LDL CHOLESTEROL AND APOPROTEIN Al CONCENTRATIONS IN 28 SUBJECTS DURING A 60-WEEK PERIOD Triglyceride
LDL cholesterol
Apoprotein Al
4.19 NS
+0.25 NS
+0.62
-0.22 <0.05
+0.02 NS
+0.43
Between subjects (i.e. cross-sectional association ;: Within subjects (i.e. longitudinal association r P
The mean within-subject coefficients of variation for each of the measurements are presented in Table 3. In both sexes variation with time was least in total cholesterol and greatest in total triglyceride. Linear regression analysis of the pooled data from both sexes showed that within subjects HDL cholesterol varied inversely with plasma triglyceride concentration and directly with apoprotein Al concentration (Table 4).
79
Discussion There are two alternative approaches to the study of changes in plasma lipoprotein concentrations in vivo over prolonged periods of time: to make serial measurements during the course of the study period; or to store samples of plasma or serum until the last has been collected, and then perform all of the laboratory assays together. The first method suffers from the potential disadvantage that undetected changes in the laboratory procedures might occur with time. The second approach, on the other hand, is complicated by the possibility of changes taking place in the serum samples during storage. Since there was already evidence that the lipoprotein lipid measurements required for the present study can be reliably made after storage of serum at -20°C [ 5,121, this approach was adopted. The results obtained provide the first longitudinal data on serum lipoproteins in healthy young adults. Previous longitudinal studies in this field have been restricted to measurements of the plasma total cholesterol or triglyceride concentrations [11,13-161, or have involved subjects participating in a coronary risk factor intervention trial [17]. Although Nikkila [18] made serial measurements of HDL cholesterol concentration in 9 subjects, this was for 5-6 months only. The absence of any seasonal variation in serum total cholesterol or triglyceride in the present data contrasts with some [11,15,16], but concurs with other [ 13,141, longitudinal observations made in other communities. Van Gent et al. [ 191 concluded from cross-sectional data that there was a rise in HDL cholesterol concentration during the summer in 1000 subjects aged 40 years living in The Netherlands. The absence of a similar trend in Troms4 might reflect an ethnic difference, or the climatic differences under which the studies were conducted. Comparison of these two results is complicated, however, by the crosssectional nature of the Dutch study, in which the seasonal trends might have reflected differences in the compositions of the population samples studied at different times, rather than being representative of changes occurring in the community as a whole. The progressive rise recorded in apoprotein Al concentration in males is of uncertain causation. Possible explanations include a genuine age-related effect, and an undetected alteration in living habits (e.g. physical activity). The additional possibility of a storage artefact cannot be completely discounted, however, and further information is needed on this topic. The demonstration that the various lipid risk-factors for coronary disease show a wide range of biological variability (Table 3) complicates the interpretation of recent prospective studies of CHD, which have utilised single on-entry measurements of plasma lipoproteins [4,5]. A single measurement of a relatively constant factor (e.g. HDL cholesterol) in an individual will tend to be a more accurate estimate of its true mean value in that subject than will a single measurement of a very variable factor (e.g. triglyceride). For this reason the predictive power of a lipid risk-factor for future CHD will depend not only on the nature of its relationship to the atherogenic process (e.g. whether causal or non-causal), but also on its inherent biological variability (and, in laboratories where this differs for different risk-factors, the precision with which it is
80
measured). Such factors clearly merit consideration when epidemiological correlations are used as a basis for aetiological concepts of atherogenesis. The small changes which did occur in HDL cholesterol concentration with time were correlated positively with those in apoprotein Al concentration and negatively with those in triglyceride concentration. This finding supports other evidence, obtained from cross-sectional [9] and metabolic [ 201 studies, that the status of triglyceride transport is a significant determinant of HDL cholesterol concentration in man. Further examination of such temporal relationships may provide greater insight into the regulation of HDL metabolism. The present data were obtained from a group of subjects (young students) whose life-style differs from that of the majority of the community. To assess whether or not the results are representative of the population as a whole, it would be necessary to repeat the study on a random sample of individuals. It should also be emphasised that the participants were a.ll normolipidaemic, and that the results do not, therefore, exclude the possibility of seasonal changes in plasma lipoproteins occurring in hyperlipidaemic subjects. Acknowledgements The authors thank Ms. S. Chinn, Dr. S. Darby and Mr. R. Morris for assistance with the statistical analyses. References 1 Medalie, J.H., Kahn. H.A.. Neufeld, H.N., Riss. E. and Goldbourt, U., Five year myocardial infarction incidence. II. Association of single variables to age and birthplace. J. Chron. Dls.. 26 (1973) 329. 2 Carlson, L.A. and B&tiger. L.E., Ischaemic heart disease in relation to fasting values plasma triglyceride and cholesterol. Lancet, i (1972) 865. 3 Rhoads, G.G., Gulbrandsen. C.L. and Kagan. A., Serum lipoproteins and coronary heart disease in a population study of Hawaii Japanese men. New Engl. J. Med., 294 (1976) 293. 4 Gordon, T., Castelli. W.P.. Hjortland, M.C.. Kannel, W.B. and Dawber. T.R.. High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study. Am. J. Med., 62 (1977) 707. 5 Miller, N.E.. F&de, O.H., Thelle. D.S. and Mjds. O.D.. The Tromsd Heart Study. High density llpoprotein and coronary heart disease: a prospective case-control study. Lancet, i (1977) 965. 6 Ishikawa. T., Fidge, N., Thelle. D.S.. F$rde, O.H. and Miller, N.E.. The Tromsd Heart Study: Serum apolipoprotein Al concentration in relation to future coronary heart disease. Eur. J. Clln. Invest., 8 (1978) 179. 7 Friedewalde. W.T.. Levy, R.I. and Fredrickson, D.S., Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem.. 18 (1972) 499. 8 Curru, M.D., Alaupovic. P. and Suenram. CA., Determination of apolipoprotein A and its constitutive AI and AH polypeptides by separate electroimmunoassays. Clin. Chem.. 22 (1976) 315. 9 Albers, J.J., Wahl, P.W., Cabana, V.G.. Hazzard. W.R. and Hoover, J.J., Quantitation of apolipoprotein Al of human plasma high density lipoprotein. Metabolism, 25 (1976) 633. 10 Fearn, T.. A Bayesian approach to growth curves. Biometrika, 62 (1975) 89. 11 Thelle. D.S.. Fdrde. O.H.. Try, K. and Lehmann. E.H., The Troms# Heart Study. Methods and main results of the cross-sectional study. Acta Med. Scand., 200 (1976) 107. 12 Cooper. G.R.. Lipid Standardization Laboratory, Centre for Disease Control, Atlanta, Ga., U.S.A. Pers. comm. 1977. 13 Miller. N.E., Clifton-Bllgh, P.. Nestel. P.J. and Whyte, H.M., Controlled clinical trial of a new bile acid-sequestering resin colestipol. in the treatment of hypercholesterolaemia. Med. J. Aust.. 1 (1973) 1223. 14 Persson. B.. Seasonal variation of lipoprotein llpase activity in human subcutaneous adipose tissue. Clin. Sci. Mol. Med., 47 (1974) 631.
81 15
Fuller, J.H.. Grainger. S.L.. Jsxrett, R.J. and Keen, H.. Possible healthy population. Clln. Chim. Acta. 52 (1974) 305.
16
Thomas, C.B.. Hollies. H.W.D. and Eisenberg, F.F.. Observations on seasonal variations in total serum cholesterol level among healthy young prisoners. Ann. Intern. Med., 64 (1961) 413. Hulley. S.B.. Cohen, R. and Widdowson, G.. Plasma high density lipoprotein cholesterol level. Influ-
17
seasonal
variation
of plasma lipids in a
ence of risk factor intervention. J. Am. Med. Ass., 238 (1977) 2269. Nikkti, E.A.. Metabolic and endocrine control of plasma high density lipoprotein concentration. In: A.M. Gotto, N.E. Miller and M.F. Oliver (Eds.). High Density Lipoproteins and Atherosclerosis. Elsevier, Amsterdam, 1978, p. 177. 19 Van Gent, CM., Van der Voort, H. and Hessel, W.W., High density lipoprotein cholesterol monthly variation and association with cardiovascular risk factors in 1000 forty-year-old Dutch citizens. Clin. Chim. Acta, 88 (1978) 155. 20 Nicoll. A., Miller, N.E. and Lewis, B., High density lipoprotein metabolism. Adv. Lipid Res. (in press).
18
Atherosclerosis, 34 (1979) 75-81 @ Elsevier/North-Holland Scientific
Publishers,
Ltd.
A LONGITUDINAL STUDY OF THE BIOLOGICAL VARIABILITY PLASMA LIPOPROTEINS IN HEALTHY YOUNG ADULTS
O.D. MJ@S *, S.N. RAO, L. BJ@RU, T. HENDEN, N.E. MILLER
D.S. THELLE,
OF
O.H. F(bRDE and
Institutes of Medical Biology and Community Medicine, University of Tromsd, Tromsl, Norway; and Department of Chemical Pathology and Metabolic Disorders, St Thomas’ Hospital Medical School, London, England (Received 31 May, 1979) (Revised, received 14 June, 1979) (Accepted 15 June, 1979)
Summary The fasting serum concentrations of total cholesterol, low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), total triglyceride and apoprotein Al were measured at intervals of 12-18 weeks for 60 weeks in 17 male and 11 female healthy young adults in order to assess the variability of these risk factors for coronary disease. No statistically significant seasonal changes were detected in any variable in either sex, although a progressive rise in apoprotein Al concentration was observed. The coefficients of variation for random fluctuations with time were in the rank order: total cholesterol < HDL-C < apoprotein Al < LDL-C < triglyceride. These differences were attributable to biological, rather than to methodological, factors. Within subjects, HDL cholesterol concentration varied inversely with triglyceride concentration and directly with apoprotein Al concentration. The marked differences which exist in the biological variability of lipid risk factors for atherosclerosis need to be taken into account when making comparisons in epidemiological studies of the predictive powers of single onentry measurements for future disease. Fluctuations of HDL-C with time appear to be related in part to variations in triglyceride-rich lipoprotein metabolism. Key words:
Apoprotein Al - Cholesterol protein - Low density lipoprotein
* Requests for reprints to: Professor Troms4. Norway.
Coronary heart disease - High density - Triglyceride
O.D. Mjgs. Institute of Medical Biology,
University
of ‘horns&
lipo-
9000
76
Introduction Susceptibility to coronary heart disease (CHD) has long been recognised to be positively correlated with the plasma total cholesterol concentration [ 1,2], and in some communities also with the plasma triglyceride concentration [ 21. Recent studies have confirmed that the correlation between CHD and plasma cholesterol reflects their underlying relationships to the low density lipoprotein (LDL) concentration [3,4]. On the other hand, the correlation between CHD and plasma triglyceride’ is at least partly indirect, reflecting their separate inverse relationships to the plasma high density lipoprotein (HDL) cholesterol level [4,5]. Low concentrations of apoprotein Al, the major protein component of HDL, appear also to be predictive of future CHD [6]. The recognition that CHD risk shows independent relationships to different plasma lipoproteins has led to comparisons of their predictive powers, in an attempt to assess their relative importance in the pathogenesis of coronary atherosclerosis [4-6]. The predictive power of a single measurement of any lipid risk factor, however, will depend not only on its precise relationship to the disease process (e.g. whether causal or non-causal), but also on its biological variability and the precision with which the measurement is made. We have therefore studied the variability of several lipid risk factors for CHD in healthy young males and females in a longitudinal study of 60 weeks duration. In view of the relative paucity of information concerning the determinants of HDL cholesterol concentration, its relationships to the triglyceride, LDL cholesterol and apoprotein Al concentrations as a function of time were also examined. Methods Subjects Seventeen male and 11 female medical students were studied. Ages were 22-29 years (mean 25 years), body weights 48-77 kg (mean 65 kg) and relative body weights 1.80-2.54 g/cm* (mean 2.12 g/cm*). All were healthy and free of vascular disease. Three had a paternal history of myocardial infarction at age 54-64 years. All but 3 (females) were non-smokers. Clinical procedures The subjects were requested to adhere to their usual living habits for the duration of the study. Each attended for venepuncture, after an overnight fast, on five occasions: during the 6th, 21st, 36th .and 48th weeks of 197’7 and the 14th week of 1978. Venepuncture of females was performed without regard to the menstrual cycle. Blood was allowed to clot at room temperature, and the serum deep frozen (-20°C) until analysis. Laboratory procedures At the end of the study period, the serum samples were thawed and.assayed for total cholesterol and triglyceride by enzymatic procedures, as previously
77
described [ 51. Serum HDL cholesterol was measured after precipitation of VLDL and LDL with heparin and Mn2+ [5]. The LDL cholesterol concentration was estimated from these values using the formula of Friedewald et al.
[71. Serum apoprotein Al concentration was measured by electroimmunoassay, using a modification of the method of Curry et al. [8] (1.5% agarose in 50 mM sodium barbitone buffer, pH 8.6; 8 V/cm for 16 h). Monospecific antiserum to human apoprotein Al was produced in rabbits according to Albers et al. [ 91. Within-batch coefficients of variation of replicate measurements made on the same serum samples were similar for the 4 assay procedures: 2.5% for total cholesterol, 2.7% for HDL cholesterol, 2.2% for total triglyceride, and 3.0% for apoprotein Al.
Statistics The presence or absence of seasonal variation in each of the 5 variables was assessed by Bayesian growth curve analysis [lo], using the pooled data from all 28 subjects at all time points. The relationships between HDL cholesterol and the other variables as a function of time were explored by linear regression analysis. Triglyceride concentrations were converted to their logarithms prior to analysis, in view of their skewed distribution [ 111. Results The mean values and standard deviations of each variable at each time point in males are given in Table 1. Results for females appear in Table 2. Bayesian analysis confirmed the visual impression that there were no statistically significant seasonal variations. In males apoprotein Al concentration showed a significant (P < 0.05) upward trend throughout the study period. A similar trend was observed in females, but this was not statistically significant.
TIZBLE 1 FASTING SERUM TOTAL CHOLESTEROL, TOTAL TRIGLYCERIDE, HDL CHOLESTEROL, LDL CHOLESTEROL AND APOPROTEIN Al CONCENTRATIONS IN 17 MALE SUBJECTS DURING A PERIOD OF 60 WEEKS. RESULTS ARE GIVEN AS: MEAN (SD) Measurement
Week of the year 6
21
36
48
14
Total cholesterol
4.65 (0.79)
4.71 (0.73)
4.74 (0.79)
4.65 (0.67)
4.54 (0.73)
(mmolfi) Total triglyceride
0.78 (0.34)
0.85 (0.34)
0.84 (0.37)
0.99 (0.51)
0.84 (0.36)
(mmoI/D HDL cholesterol
1.48 (0.32)
1.60 (0.35)
1.47 (0.27)
1.53 (0.30)
1.56 (0.34)
(mmoI/D LDL cholesterol
2.80 (0.62)
2.75 (0.59)
2.81 (0.59)
2.67 (0.56)
2.53 (0.47)
(mmol/D Apoprotein Al
119
129
133
142
144
(miz/dD
(17.7)
(18.7)
(18.8)
(18.5)
(20.6)
TABLE 2 FASTING SERUM TOTAL CHOLESTEROL, TOTAL TRIGLYCERIDE, HDL CHOLESTEROL, LDL CHOLESTEROL AND APOPROTEIN Al CONCENTRATIONS IN 11 FEMALE SUBJECTS DURING A PERIOD OF 60 WEEKS. RESULTS ARE GIVEN AS: MEAN (SD) Measurements
Week of the year 6
21
36
40
14
Total cholesterol
4.96 (0.91)
5.04 (0.96)
5.03 (0.95)
4.67 (0.96)
4.80 (0.75)
(mmoI/I) Total triglyceride
0.69 (0.13)
0.79 (0.17)
0.83 (0.33)
0.75 (0.17)
0.65 (0.17)
(mmoI/l) HDL cholesterol
1.62 (0.44)
1.73 (0.31)
1.69 (0.36)
1.70 (0.40)
1.77 (0.38)
(mmoI/l) LDL cholesterol
2.91 (0.59)
3.04 (0.86)
2.92 (0.68)
2.65 (0.65)
2.60 (0.69)
(mmoI/l) Apoprotein
128
131
135
145
143
Al
(9.7)
(12.6)
(10.4)
(15.3)
(17.7)
(mg/dI)
TABLE 3 MEAN WITHIN-SUBJECT COEFFICIENTS OF VARIATION FOR FASTING SERUM TOTAL CHOLESTEROL, HDL CHOLESTEROL, LDL CHOLESTEROL, TOTAL TRIGLYCERIDE AND APOPROTEIN Al CONCENTRATIONS IN 28 SUBJECTS OVER A 60-WEEK PERIOD
Males Females
Total cholesterol
HDL cholesterol
LDL cholesterol
Total triglyceride
Apoprotein Al
7.80 6.70
0.42 9.42
10.4 11.8
21.4 30.4
8.33 10.2
TABLE 4 PELATIONSHIPS OF THE SERUM HDL CHOLESTEROL CONCENTRATION TO THE SERUM TOTAL TRIGLYCERIDE, LDL CHOLESTEROL AND APOPROTEIN Al CONCENTRATIONS IN 28 SUBJECTS DURING A 60-WEEK PERIOD Triglyceride
LDL cholesterol
Apoprotein Al
4.19 NS
+0.25 NS
+0.62
-0.22 <0.05
+0.02 NS
+0.43
Between subjects (i.e. cross-sectional association ;: Within subjects (i.e. longitudinal association r P
The mean within-subject coefficients of variation for each of the measurements are presented in Table 3. In both sexes variation with time was least in total cholesterol and greatest in total triglyceride. Linear regression analysis of the pooled data from both sexes showed that within subjects HDL cholesterol varied inversely with plasma triglyceride concentration and directly with apoprotein Al concentration (Table 4).
79
Discussion There are two alternative approaches to the study of changes in plasma lipoprotein concentrations in vivo over prolonged periods of time: to make serial measurements during the course of the study period; or to store samples of plasma or serum until the last has been collected, and then perform all of the laboratory assays together. The first method suffers from the potential disadvantage that undetected changes in the laboratory procedures might occur with time. The second approach, on the other hand, is complicated by the possibility of changes taking place in the serum samples during storage. Since there was already evidence that the lipoprotein lipid measurements required for the present study can be reliably made after storage of serum at -20°C [ 5,121, this approach was adopted. The results obtained provide the first longitudinal data on serum lipoproteins in healthy young adults. Previous longitudinal studies in this field have been restricted to measurements of the plasma total cholesterol or triglyceride concentrations [11,13-161, or have involved subjects participating in a coronary risk factor intervention trial [17]. Although Nikkila [18] made serial measurements of HDL cholesterol concentration in 9 subjects, this was for 5-6 months only. The absence of any seasonal variation in serum total cholesterol or triglyceride in the present data contrasts with some [11,15,16], but concurs with other [ 13,141, longitudinal observations made in other communities. Van Gent et al. [ 191 concluded from cross-sectional data that there was a rise in HDL cholesterol concentration during the summer in 1000 subjects aged 40 years living in The Netherlands. The absence of a similar trend in Troms4 might reflect an ethnic difference, or the climatic differences under which the studies were conducted. Comparison of these two results is complicated, however, by the crosssectional nature of the Dutch study, in which the seasonal trends might have reflected differences in the compositions of the population samples studied at different times, rather than being representative of changes occurring in the community as a whole. The progressive rise recorded in apoprotein Al concentration in males is of uncertain causation. Possible explanations include a genuine age-related effect, and an undetected alteration in living habits (e.g. physical activity). The additional possibility of a storage artefact cannot be completely discounted, however, and further information is needed on this topic. The demonstration that the various lipid risk-factors for coronary disease show a wide range of biological variability (Table 3) complicates the interpretation of recent prospective studies of CHD, which have utilised single on-entry measurements of plasma lipoproteins [4,5]. A single measurement of a relatively constant factor (e.g. HDL cholesterol) in an individual will tend to be a more accurate estimate of its true mean value in that subject than will a single measurement of a very variable factor (e.g. triglyceride). For this reason the predictive power of a lipid risk-factor for future CHD will depend not only on the nature of its relationship to the atherogenic process (e.g. whether causal or non-causal), but also on its inherent biological variability (and, in laboratories where this differs for different risk-factors, the precision with which it is
80
measured). Such factors clearly merit consideration when epidemiological correlations are used as a basis for aetiological concepts of atherogenesis. The small changes which did occur in HDL cholesterol concentration with time were correlated positively with those in apoprotein Al concentration and negatively with those in triglyceride concentration. This finding supports other evidence, obtained from cross-sectional [9] and metabolic [ 201 studies, that the status of triglyceride transport is a significant determinant of HDL cholesterol concentration in man. Further examination of such temporal relationships may provide greater insight into the regulation of HDL metabolism. The present data were obtained from a group of subjects (young students) whose life-style differs from that of the majority of the community. To assess whether or not the results are representative of the population as a whole, it would be necessary to repeat the study on a random sample of individuals. It should also be emphasised that the participants were a.ll normolipidaemic, and that the results do not, therefore, exclude the possibility of seasonal changes in plasma lipoproteins occurring in hyperlipidaemic subjects. Acknowledgements The authors thank Ms. S. Chinn, Dr. S. Darby and Mr. R. Morris for assistance with the statistical analyses. References 1 Medalie, J.H., Kahn. H.A.. Neufeld, H.N., Riss. E. and Goldbourt, U., Five year myocardial infarction incidence. II. Association of single variables to age and birthplace. J. Chron. Dls.. 26 (1973) 329. 2 Carlson, L.A. and B&tiger. L.E., Ischaemic heart disease in relation to fasting values plasma triglyceride and cholesterol. Lancet, i (1972) 865. 3 Rhoads, G.G., Gulbrandsen. C.L. and Kagan. A., Serum lipoproteins and coronary heart disease in a population study of Hawaii Japanese men. New Engl. J. Med., 294 (1976) 293. 4 Gordon, T., Castelli. W.P.. Hjortland, M.C.. Kannel, W.B. and Dawber. T.R.. High density lipoprotein as a protective factor against coronary heart disease. The Framingham Study. Am. J. Med., 62 (1977) 707. 5 Miller, N.E.. F&de, O.H., Thelle. D.S. and Mjds. O.D.. The Tromsd Heart Study. High density llpoprotein and coronary heart disease: a prospective case-control study. Lancet, i (1977) 965. 6 Ishikawa. T., Fidge, N., Thelle. D.S.. F$rde, O.H. and Miller, N.E.. The Tromsd Heart Study: Serum apolipoprotein Al concentration in relation to future coronary heart disease. Eur. J. Clln. Invest., 8 (1978) 179. 7 Friedewalde. W.T.. Levy, R.I. and Fredrickson, D.S., Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin. Chem.. 18 (1972) 499. 8 Curru, M.D., Alaupovic. P. and Suenram. CA., Determination of apolipoprotein A and its constitutive AI and AH polypeptides by separate electroimmunoassays. Clin. Chem.. 22 (1976) 315. 9 Albers, J.J., Wahl, P.W., Cabana, V.G.. Hazzard. W.R. and Hoover, J.J., Quantitation of apolipoprotein Al of human plasma high density lipoprotein. Metabolism, 25 (1976) 633. 10 Fearn, T.. A Bayesian approach to growth curves. Biometrika, 62 (1975) 89. 11 Thelle. D.S.. Fdrde. O.H.. Try, K. and Lehmann. E.H., The Troms# Heart Study. Methods and main results of the cross-sectional study. Acta Med. Scand., 200 (1976) 107. 12 Cooper. G.R.. Lipid Standardization Laboratory, Centre for Disease Control, Atlanta, Ga., U.S.A. Pers. comm. 1977. 13 Miller. N.E., Clifton-Bllgh, P.. Nestel. P.J. and Whyte, H.M., Controlled clinical trial of a new bile acid-sequestering resin colestipol. in the treatment of hypercholesterolaemia. Med. J. Aust.. 1 (1973) 1223. 14 Persson. B.. Seasonal variation of lipoprotein llpase activity in human subcutaneous adipose tissue. Clin. Sci. Mol. Med., 47 (1974) 631.
81 15
Fuller, J.H.. Grainger. S.L.. Jsxrett, R.J. and Keen, H.. Possible healthy population. Clln. Chim. Acta. 52 (1974) 305.
16
Thomas, C.B.. Hollies. H.W.D. and Eisenberg, F.F.. Observations on seasonal variations in total serum cholesterol level among healthy young prisoners. Ann. Intern. Med., 64 (1961) 413. Hulley. S.B.. Cohen, R. and Widdowson, G.. Plasma high density lipoprotein cholesterol level. Influ-
17
seasonal
variation
of plasma lipids in a
ence of risk factor intervention. J. Am. Med. Ass., 238 (1977) 2269. Nikkti, E.A.. Metabolic and endocrine control of plasma high density lipoprotein concentration. In: A.M. Gotto, N.E. Miller and M.F. Oliver (Eds.). High Density Lipoproteins and Atherosclerosis. Elsevier, Amsterdam, 1978, p. 177. 19 Van Gent, CM., Van der Voort, H. and Hessel, W.W., High density lipoprotein cholesterol monthly variation and association with cardiovascular risk factors in 1000 forty-year-old Dutch citizens. Clin. Chim. Acta, 88 (1978) 155. 20 Nicoll. A., Miller, N.E. and Lewis, B., High density lipoprotein metabolism. Adv. Lipid Res. (in press).
18