- Email: [email protected]
HDL cholesterol levels in patients with myocardial infarction and their families
Atherosclerosis, Elsevier
21
59 (1986) 21-29
ATH 03709
HDL Cholesterol Levels in Patients with Myocardial Infarction and their Families Daniel Pometta*,
Alan Suenram ** , Esmdil Sheybani,
Department of Medicine, Diabetes Division, Hapita
Bernard
Grab and Richard
James
Cantonal Uniuersitaire, CH 121 I Geneva 4 (Switzerland)
(Received 20 December, 1984) (Revised, received 30 July, 1985) (Accepted 2 August, 1985)
Summary The present study using a quartile distribution of myocardial infarction patients demonstrated that the first-degree relatives of the myocardial infarction patients with the lowest HDL cholesterol have similarly the lowest HDL cholesterol. Low HDL cholesterol among these relatives was not secondary to increased VLDL triglycerides, as it persisted when subjects with hyper VLDL triglycerides were excluded. Familial low HDL cholesterol could not be attributed to known environmental factors as their levels did not differ significantly between the groups compaired. There was a significant correlation between HDL cholesterol levels of the parents and that of their younger offspring. The correlation was not significant with the offspring aged 20 and over. It appeared that there was a familial trend in low HDL cholesterol levels, more apparent among the young offspring than among the adult offspring, who may possibly not share any more the parental environment for factors liable to influence HDL cholesterol. This finding is compatible with a hereditary trait.
Key words:
Families infarction
HDL
cholesterol
-
HDL
Introduction Clustering of coronary occurs in certain families.
heart disease (CHD) The influence of lipo-
Supported by grants 808.083 of the Swiss National Foundation. * To whom reprint requests should be sent. ** Present address: Department of Pathology, University of Texas, Health Science Center 84, 7703 Floyd Curl Drive, San Antonio, TX 782, U.S.A. 0021-9150/86/$03.50
0 1986 Elsevier Science Publishers
Ireland,
environmental
factors
-- HDL
heredity
-
Myocardial
proteins has been extensively investigated and the hereditary pattern and pathogenic mechanisms have been well established in certain diseases like familial hypercholesterolemia [l]. HDL cholesterol is low in CHD patients [2,3] and there is a negative association of HDL cholestef-ol level with the subsequent incidence of coronary heart disease [4]. Little is known on HDL cholesterol content in relation to the clustering of coronary heart disease in certain families. Familial hyper-cY-lipoproteinemia has a protective Ltd.
22 effect against the occurrence of coronary heart disease [5] and familial hypo-a-lipoproteinemia appears to increase cardio-vascular morbidity [6]. Studies of the correlation between HDL cholesterol of children and parent demonstrate a familial aggregation of HDL cholesterol [7-91. Recent studies on first-degree relatives of patients who suffered acute myocardial infarction have pointed out that adult male [lO,ll] and female [12] close relatives of CHD patients had low HDL cholesterol. Similarly low HDL cholesterol has been observed in prepubertal sons and in pubertal sons and daughters of CHD patients [13] suggesting that in families in which cardio-vascular disease was present low HDL cholesterol may occur in the absence of overt cardio-vascular disease and could play a role in the further development of CHD. It was suggested [12] that in the absence of environmental factors known to lower HDL cholesterol content, a hereditary trait might be responsible for the low HDL cholesterol in CHD patients and their relatives. However, when considering the individual HDL cholesterol levels of probands and of their relatives, it was apparent that they were widely scattered from low up to, sometimes, high HDL cholesterol levels. To investigate further the familial pattern of HDL cholesterol distribution, patients who suffered acute myocardial infarction were divided in 4 groups defined by the quartile limits of their HDL cholesterol distribution. The first-degree relatives of these patients were then allocated in 4 different groups according to the quartile to which the proband did belong (for example individuals in group 1 were the relatives of the MI patients with HDL cholesterol below the first quartile limit). The aim of the present study was to compare first-degree relatives thus differentiated by the HDL cholesterol class level of their parental proband. To quantitate the parent-offspring HDL cholesterol relationship further, without using preset categories (quartile strata), the parent-offspring correlation coefficients were calculated foi father and mother with their offspring.
Subjects and Methods The probands were patients suffering from acute myocardial infarction and hospitalized in the Department of Medicine. They were considered for the study if they were less than 55 years old for the men and 70 years for the women and if they had 2 or more first-degree relatives who could be investigated [10,12]. The difference in age selection of males and females was chosen because of the scarcity of myocardial infarction in females below the age of 55 years. The patients were examined at the time of acute myocardial infarction. The allocation to the 4 quartiles groups was made on the lipid levels measured more than 3 months after discharge from hospital. The first-degree relatives, aged 20 and over, of myocardial infarction patients were then distributed, male and female separately, in groups according to the quartile distribution of their myocardial infarction proband. Groups 1, 2, 3 and 4 corresponded, respectively, to the first-degree relatives of myocardial infarction patients of quartiles I-IV. The female and male spouses of the probands and their daughters and sons below the age of 20 years were also included in the study, in order to evaluate the correlations between HDL cholesterol levels in parents and offspring and to study the influence of environmental and hereditary factors. All subjects underwent medical examination. A medical history was obtained and the subjects were questioned concerning their habits of cigarette smoking, alcohol consumption and physical activity both at work and at leisure. The alcohol consumption was expressed as liters of wine (10’) per week. None of the subjects were on lipid-lowering drugs. The influence of the body weight was assessed using the body weight index (the weight in kg divided by height in cm minus 100). Venous blood was obtained after an overnight fast. Serum lipoproteins were separated by quantitative preparative ultracentrifugation [lo]. Cholesterol [ 141 and triglycerides were measured ]I51. An attempt was made to assess quantitatively the possible relationship of HDL cholesterol between parents and offspring using the statistical methods of [16]. One-way analysis of variance was
23 performed [17] to compare the natural variation of offspring in the same sibship with the added variation arising from genetic and environmental conditions peculiar to sibship. Intra-class correlations were also derived from these analyses of variance [17]. To help in discriminating the genetic role from the environmental influence, the data collected for parent pairs were submitted to the same statistical analysis. Pearson’s coefficients of correlation [17] were calculated to measure the degree of association of HDL cholesterol level between parents and their offspring. In order to evaluate the influence of common household environment [ 161, intra-class correlations and parent-child correlations were compaired between “young” sibship (individual under 20 years) and “adult” sibship (20 years and over), more likely to have left the common household. Because of the age independency of HDL cholesterol after puberty [18,19], no adjustment was made for this variable. The analysis was also performed with values adjusted for sex, but the results are not reported here, as they did not modify the conclusions from the unadjusted values. However, for the parent-intra-class correlation, adjustment for sex was required as the within-pair variability was evidently affected by the well-established sex difference in HDL cholesterol level. Adjustment was also made for differences in CHD condition (proband; non proband) within parent pairs. The logarithm transformation was applied to values of serum triglycerides, HDL and VLDL cholesterol and triglycerides in order to normalise their distribution. Conventional t, x2 and F statistical tests were used to assess respectively the significance of mean proportion differences, proportion differences and variance ratios [17]. The statistical significance was fixed at 5% probability level. Results Probands One hundred seventeen consecutively hospitalised probands (76 men and 41 women) who suffered acute myocardial infarction entered the study more than 3 months after the event. Their HDL cholesterol (mmol/l) was lower than
the levels observed in the male (n = 389) and female (n = 168) control population of the same age range (1.13 + 0.278, P < 0.001 vs 1.38 k 0.542 for the men and 1.25 k 0.298, P < 0.001 vs 1.57 + 0.347 for the women). The HDL cholesterol limits for the quartile I-IV were 0.49-0.88, 0.93-1.11, 1.14-1.27 and 1.29-1.86 mmol/l for the male probands. The quartile limits for the female probands were, respectively, 0.70-1.01, 1.06-1.17, 1.22-1.50 and 1.53-1.97 mmol/l. The upper limits of the third quartile for men and women correspond to the 45th and 47th percentile distribution of the normal Geneva population, Only 4 patients had HDL cholesterol levels above the 80th percentile of whom only 1 at the 93rd percentile. First-degree relatives of the myocardial infarction probands There were 375 first-degree relatives of myocardial infarction patients who were examined. Of them 86 (44 girls and 42 boys) were children less than 20 years and 289 (133 male and 156 female) were first-degree relatives 20 years and older. In addition 67 spouses were examined. The first-degree relatives aged 20 and over were divided into 4 groups according to the HDL cholesterol quartile of the corresponding myocardial infarction proband. Relatives in groups 1, 2, 3 and 4 were related to the probands of HDL cholesterol quartiles I-IV, respectively. The number of male relatives in groups l-4 were 35, 33, 28 and 37 (Table l), respectively, while the number of females in the corresponding groups were 58, 43, 36 and 19 (Tables 2 and 3). Mule adult first-degree
relatives
Serum lipids The male adult first-degree relatives of the myocardial infarction probands with the lowest HDL cholesterol had higher serum cholesterol (5.85 mmol/l; P -C 0.05) and triglycerides (1.57 mmol/l; P < 0.05) than the relatives of myocardial infarction patients with the highest HDL cholesterol (5.25 and 1.13 mmol/l, respectively) (Table 1).
24 TABLE
1
LIPOPROTEIN LIPIDS (MMOL/L) OF MALE ADULT FIRST-DEGREE RELATIVES IN RELATION TO THE HDL CHOLESTEROL LEVEL OF THE PROBANDS. Groups 1 (n = 35) Serum Cholesterol SD Triglycerides SD log Antilog
a 2 (n = 33)
3 (n = 28)
4 (n = 37)
log
5.85 1.19 0.196 0.287 I.57
5.46 0.96 0.090 0.294 1.23
5.59 1.27 0.086 0.227 1.22
5.25 * 1.30 0.054 0.275 1.13 *
Cholesterol log SD log Antilog Triglycerides log SD log Antilog
- 0.230 0.333 0.59 -0.390 0.359 0.91
- 0.337 0.343 0.46 -0.158 0.383 0.69
0.309 0.312 0.49 0.158 0.307 0.69
- 0.406 0.329 0.39 * - 0.237 0.340 0.58 *
3.47 0.88 0.33 0.12
3.16 0.80 0.29 0.10
3.34 0.93 0.28 0.08
3.06 0.93 0.28 0.11
VLDL
LDL
Cholesterol SD Triglycerides SD HDL
Cholesterol log SD log Antilog Triglycerides SD
0.028 0.097 1.07 0.15 0.05
0.077 0.109 1.19’ 0.13 0.07
0.082 0.095 1.21* 0.13 0.04
0.095 0.120 1.25 * 0.13 0.05
* Statistically different from group 1, P < 0.05. a Subjects in groups 1-4 were first-degree relatives of patients of HDL cholesterol quartile groups I-IV, respectively.
Lipoprotein lipids The male relatives (group 1) of the myocardial infarction probands with the lowest HDL cholesterol similarly had lower HDL cholesterol (1.07 mmol/l; P < 0.05) than other relatives (Table 1). Higher VLDL cholesterol (0.59 vs 0.39 mmol/l, P -C 0.05) and VLDL triglycerides (0.91 vs 0.58 mmol/l, P < 0.05) were measured in the relatives of CHD patients with the lowest HDL cholesterol compared to the relatives (group 4) of probands with the highest HDL cholesterol. LDL cholesterol and triglycerides were not significantly different between the relatives of groups 1 and 4.
Body weight and physical activity The mean body weight index in group 4 was 0.998 + 0.141 and did not differ from group 1 (1.064 f 0.177). Percentage distribution of male relatives according to physical activity levels, both at work and at leisure, did not differ significantly between groups 1 and 4 (Table 4). Alcohol intake and cigarette smoking Among the male first-degree relatives of probands with low HDL cholesterol (group l), 94% were regular alcohol consumers compared to 78% among the male relatives of probands with high HDL cholesterol (group 4). Among the alcohol
TABLE
2
LIPOPROTEIN LIPIDS (MMOL/L) OF FEMALE ADULT FIRST-DEGREE RELATIVES IN RELATION TO THE HDL CHOLESTEROL LEVEL OF THE PROBANDS
Groups
a
1 n = 58 Serum Cholesterol SD Triglycerides SD log Antilog
log
2 n = 43
3 n = 36
4 n=19
5.69 * 1.03 0.016 0.223 1.04
5.76 * 1.30 - 0.009 0.229 0.98
5.52 1.26 - 0.066 0.198 0.86
5.17 0.89 - 0.055 0.201 0.88
0.419 0.291 0.38 0.311 0.341 0.49
- 0.404 0.280 0.39 - 0.313 0.296 0.49
- 6.526 0.270 0.30 - 0.448 0.276 0.36
- 0.503 0.243 0.31 - 0.435 0.328 0.37
VLDL
Cholesterol log SD log Antilog Triglycerides log SD log Antilog LDL
Cholesterol SD Triglycerides SD
3.46 * 0.78 0.30 0.09
3.49 * 1.15 0.29 0.12
3.22 1.06 0.30 0.12
2.99 0.68 0.28 0.10
0.139 0.092 1.38 0.16 0.04
0.128 0.113 1.34 0.13 0.05
0.174 0.092 1.49 0.15 0.05
0.162 0.104 1.45 0.15 0.05
HDL
Cholesterol log SD log Antilog Triglycerides SD
* Statistically different from group 4, P i 0.05. a Subjects in groups 1-4 were first-degree relatives of patients of HDL cholesterol quartile groups I-IV, respectively.
25 TABLE
3
LIPOPROTEIN LIPIDS (MMOL/L) OF FEMALE ADULT FIRST-DEGREE RELATIVES IN RELATION TO THE HDL CHOLESTEROL LEVEL OF THE PROBANDS Groups a 1+2 n=lOl
3+4 n = 55
Serum
Cholesterol SD Triglycerides SD log Antilog
5.12 1.15 0.005 0.225 1 .Ol
5.40 1.15 - 0.063 0.198 0.87
-0.413 0.285 0.39 * -0.313 0.321 0.49 *
- 0.523 0.259 0.30 -0.444 0.292 0.36
LDL Cholesterol SD Triglycerides SD
3.41* 0.95 0.30 0.10
3.14 0.95 0.29 0.11
HDL Cholesterol log SD log Antilog Triglycerides SD
0.134 0.101 1.36 * 0.15 0.06
0.170 0.095 1.48 0.15 0.05
log
VLDL Cholesterol log SD log Antilog Triglycerides log SD log Antilog
Female adult first-degree relatives When the female first-degree relatives were allocated to 4 groups according to the HDL cholesterol quartiles of their corresponding myocardial infarction probands, it appeared that only 19 female relatives were in group 4 in comparison to 58 in group 1 (Table 2). HDL cholesterol tended to be lower in groups 1 and 2 than in groups 3 and 4, but the differences were not significant. In contrast, serum and LDL cholesterol were significantly higher in groups 1 and 2 than 4 (Table 2). In order to overcome the drawback of the small number of subjects in group 4, it was decided to compare the combined groups 3 and 4 with the combined groups 1 and 2 (Table 3). There were 101 female relatives in combined groups 1 and 2 and 55 in combined groups 3 and 4.
* Statistically different from groups 3 + 4, P i 0.05. a Subjects in groups l-4 were first-degree relatives of patients of HDL cholesterol quartile groups I-IV, respectively.
TABLE
4
PHYSICAL ACTIVITY AT WORK PER CENT OF SUBJECTS WITH ERATE ACTIVITY Physical activity(%)
At work Heavy Moderate or slight At leisure Heavy Moderate or slight
Male relatives (‘%) Group
1
Group
AND AT LEISURE: HEAVY OR MOD-
Female relatives (W) 4
Groups 1+2
consumers of group 1, 91% drank more than 1 1 wine (lO”)/week. Only 70% of the alcohol consumers of group 4 drank a comparable amount of alcohol. The percentage of smokers was 60% in group 1, who smoked daily an average of 20 f 8 cigarettes as compared to 40% in group 4, who smoked 25 * 16 cigarettes daily. These differences in alcohol and tobacco consumption were not significant at the 5% probability level.
Groups 3+4
17
19
3
2
83
81
97
98
11
16
1
2
89
84
99
98
Serum lipids Serum cholesterol and triglycerides concentration were not different for the female first-degree relatives (groups 1 + 2) of the coronary probands with lower HDL cholesterol compared to the female relatives (groups 3 + 4) of the probands with higher HDL cholesterol. Lipoprotein lipids HDL cholesterol was decreased in the female relatives of the probands with lower HDL cholesterol (groups 1 + 2) compared to relatives of probands with higher HDL cholesterol (1.36 vs 1.48 mmol/l, P -c 0.05) (Table 3). The female relatives of the probands with lower HDL cholesterol had higher LDL cholesterol concentration, higher VLDL cholesterol and VLDL triglycerides concentrations than the relatives of the probands with high HDL cholesterol (Table 3).
26
Body weight and physical activity The mean body weight index in (1.08 * 0.209, P < 0.01) was different 3 + 4 (0.99 & 0.184). There was no physical activity, both at work and at 4).
TABLE
groups 1 + 2 from groups difference in leisure (Table
Cigarette smoking and alcohol intake Thirty-one % of the relatives in groups 1 + 2 were smoking an average of 19 k 13 cigarettes/day against 18% in groups 3 + 4 who smoked an average of 20 + 12 cigarettes/day. There were 46% of alcohol consumers (1.8 l/week) in groups 1 + 2 against 49% alcohol consumers (1.5 l/week) in groups 3 + 4. These differences in alcohol and tobacco consumption were not significant at the 5% probability level.
6
ANALYSIS OF VARIANCE AND INTRA-CLASS CORRELATION FOR HDL CHOLESTEROL LEVEL IN SIBLINGS BY AGE CATEGORY AND THEIR PARENTS Population
Pediatric sibship Adult sibship Parent pairs
Intra-class correlation
Individual
variance
Residual
Added
Total
r
P
0.00430 0.00983 0.01046
0.00223 0.00259 -0.00010
0.00653 0.01242 0.01036
0.342 0.209 -0.010
0.01 0.06 NS
siblings by age category and in their parents is shown in Table 6. The results are consistent with the above-mentioned values of the Pearson product-moment correlation. Discussion
Correlations between HDL cholesterol levels in parents and their offspring The correlation coefficients between the HDL cholesterol levels of the parents and their offspring were calculated separately for the adult offspring (age 20 years and over) and for the younger offspring (below the age of 20) (Table 5). The HDL cholesterol correlation coefficients between the young offspring and their father (0.363; P < 0.01) and mother (0.344; P < 0.02) were highly significant. In contrast, when the adult offspring were considered, the correlation coefficients were reduced by more than half and did not reach the 5% probability level of statistical significance. There was no correlation between the HDL cholesterol of both parents. The analysis of variance and intra-class correlation for HDL cholesterol level in
TABLE
5
CORRELATION COEFFICIENTS FOR HDL CHOLESTEROL BETWEEN PARENTS AND THEIR OFFSPRINGS AND INTRA-CLASS CORRELATION FOR HDL CHOLESTEROL IN PARENT PAIRS - 0.010 father-other /0.344 0.363 **\ children < 20 years * P < 0.02 ** P < 0.01
*
- 0.010 father-other /0.163 0.148\ adult offspring 2 20 years
Plasma lipoproteins play a role in the pathogenesis of cardio-vascular disease and HDL are of particular importance [4]. In previous studies [lo-121 we have demonstrated that first-degree relatives of myocardial infarction patients have lower HDL, cholesterol than the control population. The present results demonstrate a familial resemblance between the HDL cholesterol levels observed in the myocardial infarction patients and those measured in their relatives. Both MI patients and their relatives had low HDL cholesterol but the relatives of myocardial infarction patients with the lowest HDL cholesterol similarly had lower HDL cholesterol than the remaining relatives, even if the mean HDL cholesterol levels (Tables l-3) of the latter were below the 50th percentile limit of the normal male (1.32 mmol/l) and female (1.53 mmol/l) population. As the familial resemblance can be related both to genetic and culture-environmental factors [20], attempts should be made to differentiate their respective influence. The most important cultural and environmental factors known to affect HDL cholesterol levels are cigarette smoking 121-231, alcohol consumption [22-251, physical activity [21,25] and obesity [26]. Low HDL cholesterol in male (group 1) and female (groups 1 + 2) relatives of MI patients with
27 the lowest HDL cholesterol was not associated with significant differences in any of the tested socio-cultural factors, such as intensity of physical activity (Table 4) drinking and smoking habits. Low HDL cholesterol in certain families can therefore not be explained by such environmental factors. The influence of diet has not been evaluated, but in a Geneva school survey, we were unable to find a difference in HDL cholesterol and apolipoprotein A between children of Swiss and Italian origin [27]. They represent the main ethnic groups with the major differences in food habits. Therefore dietary habits are unlikely to play a significant role in the population under study. In the absence of known environmental factors responsible for the low HDL cholesterol levels, the possibility of a familial trait, such as obesity [19] or increased VLDL [4], should be considered as a possible cause of the low HDL cholesterol observed in certain families. The relatives of the patients with low HDL cholesterol tended to have higher body weight indices than the others. However, the difference (7%) was not significant for the men and did not exceed 8% for the women, a moderate but significant excess (P < 0.01). If differences in body weight play a role on HDL, they cannot explain, especially in men, the familial trend in HDL cholesterol levels. As with the low HDL cholesterol probands (quartile I), their relatives had higher VLDL triglyceride and cholesterol values, than those observed in the relatives with the highest HDL cholesterol levels. But the mean VLDL triglyceride were still in the normal range. It was not likely that the low HDL cholesterol was secondary to elevated VLDL. However, to rule out such a possibility, the relatives with VLDL triglycerides above 1.80 mmol/l (which represents the 95th percentile limit for the men, age 20-29 years) were excluded during a subsequent comparison. HDL cholesterol was still significantly lower (Table 7) in the relatives of MI patients with the lowest HDL cholesterol level and is therefore not related to increased VLDL triglycerides. To analyse further the HDL cholesterol pattern within families, the correlations between HDL cholesterol of parents and their offspring were
calculated by considering separately the offspring below (“paediatric”) and above (“adults”) age 20, the former being more likely to share the same parental household than the latter. This allowed evaluation of the influence of environment. For the paediatric offspring, both father-child and mother-child correlations were significant. For the adult offspring, the father-child and mother-child correlations were not significant and were less than half the correlation observed in paediatric offspring. These data are in agreement with the observation of Morrison et al. [16]. The loss of common environment shared by parents and adult offspring is a plausible explanation for the observed decrease in correlation in adult offspring. An influence of genetic origin on the remaining positive correlation cannot however, be excluded. The comparison of within and between sibship variances (Table 6) is more informative in this respect. Although the total individual variance was almost twice as high for the adult siblings than for the paediatric siblings, it is interesting to note that the added variance (between-sibship variance), which could arise from genetic or environmental causes, was almost the same for both age categories of sibling. The difference in total variance was clearly due to the difference in the individual residual variances (within-sibship variance); the lowest value (0.00430) was observed for the paediatric siblings still sharing the same household environment with their parents, while the withinvariance for adult sibship (0.00983) was practically equal to the within-variance (0.01046) for parent pairs. TABLE
1
HDL CHOLESTEROL AFTER EXCLUSION TRIGLYCERIDE
OF FIRST-DEGREE OF SUBJECTS WITH
Men
RELATIVES HIGH VLDL
Women
Group 1Group 4Groups n = 21 n = 33 n =91
I+ 2Groups 3 + 4 ” = 54
HDL Cholesterol log 0.053 SD log 0.079 Antilog (mmol/l)l.I3 *
0.114 0.106 1.30
0.140 0.094 1.38*
0.173 0.093 I .49
* Statistically different from group 4 for men and from group 3 + 4 for women (P < 0.05).
28 The intra-class correlation, being the ratio of the added variance to the total variance, was affected in the adult sibship by the larger residual variance of this sibling age group. It was clearly significant for the paediatric sibship (P < 0.01) but only marginally significant for the adult sibship (0.05 < P < 0.06). As expected there was no added variance between parent pairs and consequently no intra-class correlation among them. This finding and the similarity (i) of the residual variances among adult sibship and parent pairs, and (ii) of the added variances among paediatric and adult sibships is compatible with the hypothesis of a genetic contribution in the determination of the HDL cholesterol levels in families of myocardial infarction patients. Further research is needed to confirm the present findings. Acknowledgments We wish to thank Mrs. B. Kalix, F. Ruinard, and F. Ferracin, technicians, and Mrs. M. Jaillot, nurse, for their help during the study. We are grateful to Miss F. Kaempfen for her secretarial assistance in preparing the manuscript. References 1 Brown,
2
3
4
5
6 7
MS., Kovanen, P.T. and Goldstein, J.L., Regulation of plasma cholesterol by lipoprotein receptors, Science, 212 (1981) 628. Micheli, H. Pometta, D. and Bloch, A., Cholesterol et triglycerides des lipoprotdnes seriques isolees par ultracentrifugation Relation avec l’atheroscltrose coronarienne, Schweiz. Med. Wschr., 104 (1974) 9417. Miller, G.J. and Miller, N.E., Plasma high-density lipoprotein concentration and development of ischaemic heart-disease, Lancet, i (1975) 16. 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, Amer. J. Med., 26 (1977) 707. Glueck, C.J., Gartside, P., Fallat, R.W., Slelski, J. and Steiner, P.M., Longevity syndromes - familial hypobetaand familial hyperalphalipoproteinemia, J. Lab. Clin. North Med., 88 (1976) 941. Malloy, M.J. and Kane, J.P., Hypolipidemia, Med. Clin. North Amer., 66 (1982) 469. Morrison, J.A., Kelly, K. and Horvitz, R., Patient-offspring and sibling lipid and lipoprotein associations during and after sharing of household environments - The Princeton School District Family Study, Metabolism, 131 (1982) 158.
8 Garrison, R.J., Castelli, W.P., Feinleib, M., Kannel, W.B., Havlik, R.J., Padgett, S.J. and McNamara, P.M., The association of total cholesterol, triglycerides, and plasma lipoprotein cholesterol levels in first degree relatives and spouse pairs, Amer. J. Expidemiol., 110 (1979) 313. 9 Sosenko, J.M., Breslow, J.L., Ellison, R.C. and Miettinen, O., Familial aggregation of total cholesterol, high density lipoprotein cholesterol and total triglyceride levels in plasma, Amer. J. Epidemol., 112 (1980) 656. 10 Micheli, H., Pometta, D., Jomot, C. and Scherrer, J.R., High density lipoprotein in male relatives of coronary patients, Atherosclerosis, 32 (1979) 269. 11 Pometta, D., Micheli, H., Suenram, A. and Jornot, C., HDL et lipoprotdnes dans les familles de malades coronariennes, Schweiz. Med. Wschr., 111 (1981) 117. 12 Pometta, D., Micheli, H., Suenram, A. and Jomot, C., HDL lipids in female close relatives of coronary heart disease patients Environmental and genetic influences, Atherosclerosis, 34 (1979) 419. I., 13 Pometta, D., Micheli, H., Raymond, L. and Oberhaensli, Decreased HDL cholesterol in prepubertal and pubertal children of CHD patients, Atherosclerosis, 36 (1980) 101. J.T. and Keys, A., Cholesterol in serum and 14 Anderson, lipoprotein fractions, Clin. Chem., 2 (1956) 145. manual micro-method for de15 Soloni, E.G., Simplified termination of serum triglycerides, Clin. Chem., 17 (1971) 529. J.A., Khoury, P., Laskarzewski, P.M., Mellies, 16 Morrison, M.J. Heinemeyer, R. and Glueck, C.J., Familial associations of lipids and lipoproteins in families of hypercholesterolemic probands, Arteriosclerosis, 2 (1982) 151. 17 Snedecor, G.W. and Cochran, W.G., Statistical Methods, 6th edition, Iowa State University Press, Ames, IA, 1978. 18 Lewis, B., Chait, A., Sigurdsson, G., Mancini, M., Farinaro. E., Oriente, P., Carlson, L.A., Ericsson, M., Micheli, H. and Pometta, D., Serum lipoproteins in four European Communities - Quantitative comparison, Eur. J. Clin. Invest., 8 (1978) 165. 19 Hasstedt, S.J., Albers, J.J., Cheung, M.C., Jorde, L.B., Wilson, D.E., Edwards, C.O., Cannon, W.N., Ash, K.O. and Williams, R.R., The inheritance of high density lipoprotein cholesterol and apolipoproteins A-I and A-II, Atherosclerosis, 51 (1984) 21. 20 Debray, Q. and Lalouel, J.M., Definition des facteurs d’environment par rapport au facteur gtnetique - Leur contribution a la ressemblance familiale, J. Genet. Hum.. 29 (1981) 505. 21 Enger, S.C., Herbjomsen, K., Erikssen, K. and Fretland, A., High density lipoproteins (HDL) and physical activity The influence of physical exercise, ages and smoking on HDL cholesterol and the HDL/total cholesterol ratio, Stand. J. Clin. Lab. Invest., 37 (1977) 251. 22 Morrison, J.A., Mellies, M., De Groot, J., Khoury, P., Gartside, P.S. and Glueck, C.J., Cigarette smoking, alcohol intake, and oral contraceptive - Relationships to lipids and lipoproteins in adolescent school-children, Metabolism, 28 (1979) 1166. 23 Heyden, S., Heiss, G., Manegold, C., Tyroler, H.A., Hames, C.G., Bartel, A.G. and Copper, G., The combined effect of
29 smoking and coffee drinking on LDL and HDL cholesterol, Circulation, 60 (1979) 22. 24 Castelli, W.P., Doyle, J.T., Gordon, T., Hames, C.G., Hjortland, M.C., Hulley, S.B., Kagan, A. and Zukel, W.J., Alcohol and blood lipids The cooperative lipoprotein phenotyping study, Lancet, ii (1977) 153. 25 Willett. W., Hennekens, C.H., Siegel, A.J., Adner, M.M. and Castelli, W.P.. Alcohol consumption and high density lipoprotein cholesterol in marathon runners, N. Engl. J. Med., 303 (1980) 1159.
26 Pometta, D. and Micheli, H., Atherosclerose coronarienne et lipoproteines seriques. Schweiz. Med. Wschr., 109 (1979) 1926. 27 Oberhansli, I., Pometta, D., Micheli, H., Raymond, L. and Suenram. A., Lipid, lipoprotein and apo-A and apo-B lipoprotein distribution in Italian and Swiss school children The Geneva Survey, Pediat. Res., 280 (1982) 1563.
21
59 (1986) 21-29
ATH 03709
HDL Cholesterol Levels in Patients with Myocardial Infarction and their Families Daniel Pometta*,
Alan Suenram ** , Esmdil Sheybani,
Department of Medicine, Diabetes Division, Hapita
Bernard
Grab and Richard
James
Cantonal Uniuersitaire, CH 121 I Geneva 4 (Switzerland)
(Received 20 December, 1984) (Revised, received 30 July, 1985) (Accepted 2 August, 1985)
Summary The present study using a quartile distribution of myocardial infarction patients demonstrated that the first-degree relatives of the myocardial infarction patients with the lowest HDL cholesterol have similarly the lowest HDL cholesterol. Low HDL cholesterol among these relatives was not secondary to increased VLDL triglycerides, as it persisted when subjects with hyper VLDL triglycerides were excluded. Familial low HDL cholesterol could not be attributed to known environmental factors as their levels did not differ significantly between the groups compaired. There was a significant correlation between HDL cholesterol levels of the parents and that of their younger offspring. The correlation was not significant with the offspring aged 20 and over. It appeared that there was a familial trend in low HDL cholesterol levels, more apparent among the young offspring than among the adult offspring, who may possibly not share any more the parental environment for factors liable to influence HDL cholesterol. This finding is compatible with a hereditary trait.
Key words:
Families infarction
HDL
cholesterol
-
HDL
Introduction Clustering of coronary occurs in certain families.
heart disease (CHD) The influence of lipo-
Supported by grants 808.083 of the Swiss National Foundation. * To whom reprint requests should be sent. ** Present address: Department of Pathology, University of Texas, Health Science Center 84, 7703 Floyd Curl Drive, San Antonio, TX 782, U.S.A. 0021-9150/86/$03.50
0 1986 Elsevier Science Publishers
Ireland,
environmental
factors
-- HDL
heredity
-
Myocardial
proteins has been extensively investigated and the hereditary pattern and pathogenic mechanisms have been well established in certain diseases like familial hypercholesterolemia [l]. HDL cholesterol is low in CHD patients [2,3] and there is a negative association of HDL cholestef-ol level with the subsequent incidence of coronary heart disease [4]. Little is known on HDL cholesterol content in relation to the clustering of coronary heart disease in certain families. Familial hyper-cY-lipoproteinemia has a protective Ltd.
22 effect against the occurrence of coronary heart disease [5] and familial hypo-a-lipoproteinemia appears to increase cardio-vascular morbidity [6]. Studies of the correlation between HDL cholesterol of children and parent demonstrate a familial aggregation of HDL cholesterol [7-91. Recent studies on first-degree relatives of patients who suffered acute myocardial infarction have pointed out that adult male [lO,ll] and female [12] close relatives of CHD patients had low HDL cholesterol. Similarly low HDL cholesterol has been observed in prepubertal sons and in pubertal sons and daughters of CHD patients [13] suggesting that in families in which cardio-vascular disease was present low HDL cholesterol may occur in the absence of overt cardio-vascular disease and could play a role in the further development of CHD. It was suggested [12] that in the absence of environmental factors known to lower HDL cholesterol content, a hereditary trait might be responsible for the low HDL cholesterol in CHD patients and their relatives. However, when considering the individual HDL cholesterol levels of probands and of their relatives, it was apparent that they were widely scattered from low up to, sometimes, high HDL cholesterol levels. To investigate further the familial pattern of HDL cholesterol distribution, patients who suffered acute myocardial infarction were divided in 4 groups defined by the quartile limits of their HDL cholesterol distribution. The first-degree relatives of these patients were then allocated in 4 different groups according to the quartile to which the proband did belong (for example individuals in group 1 were the relatives of the MI patients with HDL cholesterol below the first quartile limit). The aim of the present study was to compare first-degree relatives thus differentiated by the HDL cholesterol class level of their parental proband. To quantitate the parent-offspring HDL cholesterol relationship further, without using preset categories (quartile strata), the parent-offspring correlation coefficients were calculated foi father and mother with their offspring.
Subjects and Methods The probands were patients suffering from acute myocardial infarction and hospitalized in the Department of Medicine. They were considered for the study if they were less than 55 years old for the men and 70 years for the women and if they had 2 or more first-degree relatives who could be investigated [10,12]. The difference in age selection of males and females was chosen because of the scarcity of myocardial infarction in females below the age of 55 years. The patients were examined at the time of acute myocardial infarction. The allocation to the 4 quartiles groups was made on the lipid levels measured more than 3 months after discharge from hospital. The first-degree relatives, aged 20 and over, of myocardial infarction patients were then distributed, male and female separately, in groups according to the quartile distribution of their myocardial infarction proband. Groups 1, 2, 3 and 4 corresponded, respectively, to the first-degree relatives of myocardial infarction patients of quartiles I-IV. The female and male spouses of the probands and their daughters and sons below the age of 20 years were also included in the study, in order to evaluate the correlations between HDL cholesterol levels in parents and offspring and to study the influence of environmental and hereditary factors. All subjects underwent medical examination. A medical history was obtained and the subjects were questioned concerning their habits of cigarette smoking, alcohol consumption and physical activity both at work and at leisure. The alcohol consumption was expressed as liters of wine (10’) per week. None of the subjects were on lipid-lowering drugs. The influence of the body weight was assessed using the body weight index (the weight in kg divided by height in cm minus 100). Venous blood was obtained after an overnight fast. Serum lipoproteins were separated by quantitative preparative ultracentrifugation [lo]. Cholesterol [ 141 and triglycerides were measured ]I51. An attempt was made to assess quantitatively the possible relationship of HDL cholesterol between parents and offspring using the statistical methods of [16]. One-way analysis of variance was
23 performed [17] to compare the natural variation of offspring in the same sibship with the added variation arising from genetic and environmental conditions peculiar to sibship. Intra-class correlations were also derived from these analyses of variance [17]. To help in discriminating the genetic role from the environmental influence, the data collected for parent pairs were submitted to the same statistical analysis. Pearson’s coefficients of correlation [17] were calculated to measure the degree of association of HDL cholesterol level between parents and their offspring. In order to evaluate the influence of common household environment [ 161, intra-class correlations and parent-child correlations were compaired between “young” sibship (individual under 20 years) and “adult” sibship (20 years and over), more likely to have left the common household. Because of the age independency of HDL cholesterol after puberty [18,19], no adjustment was made for this variable. The analysis was also performed with values adjusted for sex, but the results are not reported here, as they did not modify the conclusions from the unadjusted values. However, for the parent-intra-class correlation, adjustment for sex was required as the within-pair variability was evidently affected by the well-established sex difference in HDL cholesterol level. Adjustment was also made for differences in CHD condition (proband; non proband) within parent pairs. The logarithm transformation was applied to values of serum triglycerides, HDL and VLDL cholesterol and triglycerides in order to normalise their distribution. Conventional t, x2 and F statistical tests were used to assess respectively the significance of mean proportion differences, proportion differences and variance ratios [17]. The statistical significance was fixed at 5% probability level. Results Probands One hundred seventeen consecutively hospitalised probands (76 men and 41 women) who suffered acute myocardial infarction entered the study more than 3 months after the event. Their HDL cholesterol (mmol/l) was lower than
the levels observed in the male (n = 389) and female (n = 168) control population of the same age range (1.13 + 0.278, P < 0.001 vs 1.38 k 0.542 for the men and 1.25 k 0.298, P < 0.001 vs 1.57 + 0.347 for the women). The HDL cholesterol limits for the quartile I-IV were 0.49-0.88, 0.93-1.11, 1.14-1.27 and 1.29-1.86 mmol/l for the male probands. The quartile limits for the female probands were, respectively, 0.70-1.01, 1.06-1.17, 1.22-1.50 and 1.53-1.97 mmol/l. The upper limits of the third quartile for men and women correspond to the 45th and 47th percentile distribution of the normal Geneva population, Only 4 patients had HDL cholesterol levels above the 80th percentile of whom only 1 at the 93rd percentile. First-degree relatives of the myocardial infarction probands There were 375 first-degree relatives of myocardial infarction patients who were examined. Of them 86 (44 girls and 42 boys) were children less than 20 years and 289 (133 male and 156 female) were first-degree relatives 20 years and older. In addition 67 spouses were examined. The first-degree relatives aged 20 and over were divided into 4 groups according to the HDL cholesterol quartile of the corresponding myocardial infarction proband. Relatives in groups 1, 2, 3 and 4 were related to the probands of HDL cholesterol quartiles I-IV, respectively. The number of male relatives in groups l-4 were 35, 33, 28 and 37 (Table l), respectively, while the number of females in the corresponding groups were 58, 43, 36 and 19 (Tables 2 and 3). Mule adult first-degree
relatives
Serum lipids The male adult first-degree relatives of the myocardial infarction probands with the lowest HDL cholesterol had higher serum cholesterol (5.85 mmol/l; P -C 0.05) and triglycerides (1.57 mmol/l; P < 0.05) than the relatives of myocardial infarction patients with the highest HDL cholesterol (5.25 and 1.13 mmol/l, respectively) (Table 1).
24 TABLE
1
LIPOPROTEIN LIPIDS (MMOL/L) OF MALE ADULT FIRST-DEGREE RELATIVES IN RELATION TO THE HDL CHOLESTEROL LEVEL OF THE PROBANDS. Groups 1 (n = 35) Serum Cholesterol SD Triglycerides SD log Antilog
a 2 (n = 33)
3 (n = 28)
4 (n = 37)
log
5.85 1.19 0.196 0.287 I.57
5.46 0.96 0.090 0.294 1.23
5.59 1.27 0.086 0.227 1.22
5.25 * 1.30 0.054 0.275 1.13 *
Cholesterol log SD log Antilog Triglycerides log SD log Antilog
- 0.230 0.333 0.59 -0.390 0.359 0.91
- 0.337 0.343 0.46 -0.158 0.383 0.69
0.309 0.312 0.49 0.158 0.307 0.69
- 0.406 0.329 0.39 * - 0.237 0.340 0.58 *
3.47 0.88 0.33 0.12
3.16 0.80 0.29 0.10
3.34 0.93 0.28 0.08
3.06 0.93 0.28 0.11
VLDL
LDL
Cholesterol SD Triglycerides SD HDL
Cholesterol log SD log Antilog Triglycerides SD
0.028 0.097 1.07 0.15 0.05
0.077 0.109 1.19’ 0.13 0.07
0.082 0.095 1.21* 0.13 0.04
0.095 0.120 1.25 * 0.13 0.05
* Statistically different from group 1, P < 0.05. a Subjects in groups 1-4 were first-degree relatives of patients of HDL cholesterol quartile groups I-IV, respectively.
Lipoprotein lipids The male relatives (group 1) of the myocardial infarction probands with the lowest HDL cholesterol similarly had lower HDL cholesterol (1.07 mmol/l; P < 0.05) than other relatives (Table 1). Higher VLDL cholesterol (0.59 vs 0.39 mmol/l, P -C 0.05) and VLDL triglycerides (0.91 vs 0.58 mmol/l, P < 0.05) were measured in the relatives of CHD patients with the lowest HDL cholesterol compared to the relatives (group 4) of probands with the highest HDL cholesterol. LDL cholesterol and triglycerides were not significantly different between the relatives of groups 1 and 4.
Body weight and physical activity The mean body weight index in group 4 was 0.998 + 0.141 and did not differ from group 1 (1.064 f 0.177). Percentage distribution of male relatives according to physical activity levels, both at work and at leisure, did not differ significantly between groups 1 and 4 (Table 4). Alcohol intake and cigarette smoking Among the male first-degree relatives of probands with low HDL cholesterol (group l), 94% were regular alcohol consumers compared to 78% among the male relatives of probands with high HDL cholesterol (group 4). Among the alcohol
TABLE
2
LIPOPROTEIN LIPIDS (MMOL/L) OF FEMALE ADULT FIRST-DEGREE RELATIVES IN RELATION TO THE HDL CHOLESTEROL LEVEL OF THE PROBANDS
Groups
a
1 n = 58 Serum Cholesterol SD Triglycerides SD log Antilog
log
2 n = 43
3 n = 36
4 n=19
5.69 * 1.03 0.016 0.223 1.04
5.76 * 1.30 - 0.009 0.229 0.98
5.52 1.26 - 0.066 0.198 0.86
5.17 0.89 - 0.055 0.201 0.88
0.419 0.291 0.38 0.311 0.341 0.49
- 0.404 0.280 0.39 - 0.313 0.296 0.49
- 6.526 0.270 0.30 - 0.448 0.276 0.36
- 0.503 0.243 0.31 - 0.435 0.328 0.37
VLDL
Cholesterol log SD log Antilog Triglycerides log SD log Antilog LDL
Cholesterol SD Triglycerides SD
3.46 * 0.78 0.30 0.09
3.49 * 1.15 0.29 0.12
3.22 1.06 0.30 0.12
2.99 0.68 0.28 0.10
0.139 0.092 1.38 0.16 0.04
0.128 0.113 1.34 0.13 0.05
0.174 0.092 1.49 0.15 0.05
0.162 0.104 1.45 0.15 0.05
HDL
Cholesterol log SD log Antilog Triglycerides SD
* Statistically different from group 4, P i 0.05. a Subjects in groups 1-4 were first-degree relatives of patients of HDL cholesterol quartile groups I-IV, respectively.
25 TABLE
3
LIPOPROTEIN LIPIDS (MMOL/L) OF FEMALE ADULT FIRST-DEGREE RELATIVES IN RELATION TO THE HDL CHOLESTEROL LEVEL OF THE PROBANDS Groups a 1+2 n=lOl
3+4 n = 55
Serum
Cholesterol SD Triglycerides SD log Antilog
5.12 1.15 0.005 0.225 1 .Ol
5.40 1.15 - 0.063 0.198 0.87
-0.413 0.285 0.39 * -0.313 0.321 0.49 *
- 0.523 0.259 0.30 -0.444 0.292 0.36
LDL Cholesterol SD Triglycerides SD
3.41* 0.95 0.30 0.10
3.14 0.95 0.29 0.11
HDL Cholesterol log SD log Antilog Triglycerides SD
0.134 0.101 1.36 * 0.15 0.06
0.170 0.095 1.48 0.15 0.05
log
VLDL Cholesterol log SD log Antilog Triglycerides log SD log Antilog
Female adult first-degree relatives When the female first-degree relatives were allocated to 4 groups according to the HDL cholesterol quartiles of their corresponding myocardial infarction probands, it appeared that only 19 female relatives were in group 4 in comparison to 58 in group 1 (Table 2). HDL cholesterol tended to be lower in groups 1 and 2 than in groups 3 and 4, but the differences were not significant. In contrast, serum and LDL cholesterol were significantly higher in groups 1 and 2 than 4 (Table 2). In order to overcome the drawback of the small number of subjects in group 4, it was decided to compare the combined groups 3 and 4 with the combined groups 1 and 2 (Table 3). There were 101 female relatives in combined groups 1 and 2 and 55 in combined groups 3 and 4.
* Statistically different from groups 3 + 4, P i 0.05. a Subjects in groups l-4 were first-degree relatives of patients of HDL cholesterol quartile groups I-IV, respectively.
TABLE
4
PHYSICAL ACTIVITY AT WORK PER CENT OF SUBJECTS WITH ERATE ACTIVITY Physical activity(%)
At work Heavy Moderate or slight At leisure Heavy Moderate or slight
Male relatives (‘%) Group
1
Group
AND AT LEISURE: HEAVY OR MOD-
Female relatives (W) 4
Groups 1+2
consumers of group 1, 91% drank more than 1 1 wine (lO”)/week. Only 70% of the alcohol consumers of group 4 drank a comparable amount of alcohol. The percentage of smokers was 60% in group 1, who smoked daily an average of 20 f 8 cigarettes as compared to 40% in group 4, who smoked 25 * 16 cigarettes daily. These differences in alcohol and tobacco consumption were not significant at the 5% probability level.
Groups 3+4
17
19
3
2
83
81
97
98
11
16
1
2
89
84
99
98
Serum lipids Serum cholesterol and triglycerides concentration were not different for the female first-degree relatives (groups 1 + 2) of the coronary probands with lower HDL cholesterol compared to the female relatives (groups 3 + 4) of the probands with higher HDL cholesterol. Lipoprotein lipids HDL cholesterol was decreased in the female relatives of the probands with lower HDL cholesterol (groups 1 + 2) compared to relatives of probands with higher HDL cholesterol (1.36 vs 1.48 mmol/l, P -c 0.05) (Table 3). The female relatives of the probands with lower HDL cholesterol had higher LDL cholesterol concentration, higher VLDL cholesterol and VLDL triglycerides concentrations than the relatives of the probands with high HDL cholesterol (Table 3).
26
Body weight and physical activity The mean body weight index in (1.08 * 0.209, P < 0.01) was different 3 + 4 (0.99 & 0.184). There was no physical activity, both at work and at 4).
TABLE
groups 1 + 2 from groups difference in leisure (Table
Cigarette smoking and alcohol intake Thirty-one % of the relatives in groups 1 + 2 were smoking an average of 19 k 13 cigarettes/day against 18% in groups 3 + 4 who smoked an average of 20 + 12 cigarettes/day. There were 46% of alcohol consumers (1.8 l/week) in groups 1 + 2 against 49% alcohol consumers (1.5 l/week) in groups 3 + 4. These differences in alcohol and tobacco consumption were not significant at the 5% probability level.
6
ANALYSIS OF VARIANCE AND INTRA-CLASS CORRELATION FOR HDL CHOLESTEROL LEVEL IN SIBLINGS BY AGE CATEGORY AND THEIR PARENTS Population
Pediatric sibship Adult sibship Parent pairs
Intra-class correlation
Individual
variance
Residual
Added
Total
r
P
0.00430 0.00983 0.01046
0.00223 0.00259 -0.00010
0.00653 0.01242 0.01036
0.342 0.209 -0.010
0.01 0.06 NS
siblings by age category and in their parents is shown in Table 6. The results are consistent with the above-mentioned values of the Pearson product-moment correlation. Discussion
Correlations between HDL cholesterol levels in parents and their offspring The correlation coefficients between the HDL cholesterol levels of the parents and their offspring were calculated separately for the adult offspring (age 20 years and over) and for the younger offspring (below the age of 20) (Table 5). The HDL cholesterol correlation coefficients between the young offspring and their father (0.363; P < 0.01) and mother (0.344; P < 0.02) were highly significant. In contrast, when the adult offspring were considered, the correlation coefficients were reduced by more than half and did not reach the 5% probability level of statistical significance. There was no correlation between the HDL cholesterol of both parents. The analysis of variance and intra-class correlation for HDL cholesterol level in
TABLE
5
CORRELATION COEFFICIENTS FOR HDL CHOLESTEROL BETWEEN PARENTS AND THEIR OFFSPRINGS AND INTRA-CLASS CORRELATION FOR HDL CHOLESTEROL IN PARENT PAIRS - 0.010 father-other /0.344 0.363 **\ children < 20 years * P < 0.02 ** P < 0.01
*
- 0.010 father-other /0.163 0.148\ adult offspring 2 20 years
Plasma lipoproteins play a role in the pathogenesis of cardio-vascular disease and HDL are of particular importance [4]. In previous studies [lo-121 we have demonstrated that first-degree relatives of myocardial infarction patients have lower HDL, cholesterol than the control population. The present results demonstrate a familial resemblance between the HDL cholesterol levels observed in the myocardial infarction patients and those measured in their relatives. Both MI patients and their relatives had low HDL cholesterol but the relatives of myocardial infarction patients with the lowest HDL cholesterol similarly had lower HDL cholesterol than the remaining relatives, even if the mean HDL cholesterol levels (Tables l-3) of the latter were below the 50th percentile limit of the normal male (1.32 mmol/l) and female (1.53 mmol/l) population. As the familial resemblance can be related both to genetic and culture-environmental factors [20], attempts should be made to differentiate their respective influence. The most important cultural and environmental factors known to affect HDL cholesterol levels are cigarette smoking 121-231, alcohol consumption [22-251, physical activity [21,25] and obesity [26]. Low HDL cholesterol in male (group 1) and female (groups 1 + 2) relatives of MI patients with
27 the lowest HDL cholesterol was not associated with significant differences in any of the tested socio-cultural factors, such as intensity of physical activity (Table 4) drinking and smoking habits. Low HDL cholesterol in certain families can therefore not be explained by such environmental factors. The influence of diet has not been evaluated, but in a Geneva school survey, we were unable to find a difference in HDL cholesterol and apolipoprotein A between children of Swiss and Italian origin [27]. They represent the main ethnic groups with the major differences in food habits. Therefore dietary habits are unlikely to play a significant role in the population under study. In the absence of known environmental factors responsible for the low HDL cholesterol levels, the possibility of a familial trait, such as obesity [19] or increased VLDL [4], should be considered as a possible cause of the low HDL cholesterol observed in certain families. The relatives of the patients with low HDL cholesterol tended to have higher body weight indices than the others. However, the difference (7%) was not significant for the men and did not exceed 8% for the women, a moderate but significant excess (P < 0.01). If differences in body weight play a role on HDL, they cannot explain, especially in men, the familial trend in HDL cholesterol levels. As with the low HDL cholesterol probands (quartile I), their relatives had higher VLDL triglyceride and cholesterol values, than those observed in the relatives with the highest HDL cholesterol levels. But the mean VLDL triglyceride were still in the normal range. It was not likely that the low HDL cholesterol was secondary to elevated VLDL. However, to rule out such a possibility, the relatives with VLDL triglycerides above 1.80 mmol/l (which represents the 95th percentile limit for the men, age 20-29 years) were excluded during a subsequent comparison. HDL cholesterol was still significantly lower (Table 7) in the relatives of MI patients with the lowest HDL cholesterol level and is therefore not related to increased VLDL triglycerides. To analyse further the HDL cholesterol pattern within families, the correlations between HDL cholesterol of parents and their offspring were
calculated by considering separately the offspring below (“paediatric”) and above (“adults”) age 20, the former being more likely to share the same parental household than the latter. This allowed evaluation of the influence of environment. For the paediatric offspring, both father-child and mother-child correlations were significant. For the adult offspring, the father-child and mother-child correlations were not significant and were less than half the correlation observed in paediatric offspring. These data are in agreement with the observation of Morrison et al. [16]. The loss of common environment shared by parents and adult offspring is a plausible explanation for the observed decrease in correlation in adult offspring. An influence of genetic origin on the remaining positive correlation cannot however, be excluded. The comparison of within and between sibship variances (Table 6) is more informative in this respect. Although the total individual variance was almost twice as high for the adult siblings than for the paediatric siblings, it is interesting to note that the added variance (between-sibship variance), which could arise from genetic or environmental causes, was almost the same for both age categories of sibling. The difference in total variance was clearly due to the difference in the individual residual variances (within-sibship variance); the lowest value (0.00430) was observed for the paediatric siblings still sharing the same household environment with their parents, while the withinvariance for adult sibship (0.00983) was practically equal to the within-variance (0.01046) for parent pairs. TABLE
1
HDL CHOLESTEROL AFTER EXCLUSION TRIGLYCERIDE
OF FIRST-DEGREE OF SUBJECTS WITH
Men
RELATIVES HIGH VLDL
Women
Group 1Group 4Groups n = 21 n = 33 n =91
I+ 2Groups 3 + 4 ” = 54
HDL Cholesterol log 0.053 SD log 0.079 Antilog (mmol/l)l.I3 *
0.114 0.106 1.30
0.140 0.094 1.38*
0.173 0.093 I .49
* Statistically different from group 4 for men and from group 3 + 4 for women (P < 0.05).
28 The intra-class correlation, being the ratio of the added variance to the total variance, was affected in the adult sibship by the larger residual variance of this sibling age group. It was clearly significant for the paediatric sibship (P < 0.01) but only marginally significant for the adult sibship (0.05 < P < 0.06). As expected there was no added variance between parent pairs and consequently no intra-class correlation among them. This finding and the similarity (i) of the residual variances among adult sibship and parent pairs, and (ii) of the added variances among paediatric and adult sibships is compatible with the hypothesis of a genetic contribution in the determination of the HDL cholesterol levels in families of myocardial infarction patients. Further research is needed to confirm the present findings. Acknowledgments We wish to thank Mrs. B. Kalix, F. Ruinard, and F. Ferracin, technicians, and Mrs. M. Jaillot, nurse, for their help during the study. We are grateful to Miss F. Kaempfen for her secretarial assistance in preparing the manuscript. References 1 Brown,
2
3
4
5
6 7
MS., Kovanen, P.T. and Goldstein, J.L., Regulation of plasma cholesterol by lipoprotein receptors, Science, 212 (1981) 628. Micheli, H. Pometta, D. and Bloch, A., Cholesterol et triglycerides des lipoprotdnes seriques isolees par ultracentrifugation Relation avec l’atheroscltrose coronarienne, Schweiz. Med. Wschr., 104 (1974) 9417. Miller, G.J. and Miller, N.E., Plasma high-density lipoprotein concentration and development of ischaemic heart-disease, Lancet, i (1975) 16. 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, Amer. J. Med., 26 (1977) 707. Glueck, C.J., Gartside, P., Fallat, R.W., Slelski, J. and Steiner, P.M., Longevity syndromes - familial hypobetaand familial hyperalphalipoproteinemia, J. Lab. Clin. North Med., 88 (1976) 941. Malloy, M.J. and Kane, J.P., Hypolipidemia, Med. Clin. North Amer., 66 (1982) 469. Morrison, J.A., Kelly, K. and Horvitz, R., Patient-offspring and sibling lipid and lipoprotein associations during and after sharing of household environments - The Princeton School District Family Study, Metabolism, 131 (1982) 158.
8 Garrison, R.J., Castelli, W.P., Feinleib, M., Kannel, W.B., Havlik, R.J., Padgett, S.J. and McNamara, P.M., The association of total cholesterol, triglycerides, and plasma lipoprotein cholesterol levels in first degree relatives and spouse pairs, Amer. J. Expidemiol., 110 (1979) 313. 9 Sosenko, J.M., Breslow, J.L., Ellison, R.C. and Miettinen, O., Familial aggregation of total cholesterol, high density lipoprotein cholesterol and total triglyceride levels in plasma, Amer. J. Epidemol., 112 (1980) 656. 10 Micheli, H., Pometta, D., Jomot, C. and Scherrer, J.R., High density lipoprotein in male relatives of coronary patients, Atherosclerosis, 32 (1979) 269. 11 Pometta, D., Micheli, H., Suenram, A. and Jornot, C., HDL et lipoprotdnes dans les familles de malades coronariennes, Schweiz. Med. Wschr., 111 (1981) 117. 12 Pometta, D., Micheli, H., Suenram, A. and Jomot, C., HDL lipids in female close relatives of coronary heart disease patients Environmental and genetic influences, Atherosclerosis, 34 (1979) 419. I., 13 Pometta, D., Micheli, H., Raymond, L. and Oberhaensli, Decreased HDL cholesterol in prepubertal and pubertal children of CHD patients, Atherosclerosis, 36 (1980) 101. J.T. and Keys, A., Cholesterol in serum and 14 Anderson, lipoprotein fractions, Clin. Chem., 2 (1956) 145. manual micro-method for de15 Soloni, E.G., Simplified termination of serum triglycerides, Clin. Chem., 17 (1971) 529. J.A., Khoury, P., Laskarzewski, P.M., Mellies, 16 Morrison, M.J. Heinemeyer, R. and Glueck, C.J., Familial associations of lipids and lipoproteins in families of hypercholesterolemic probands, Arteriosclerosis, 2 (1982) 151. 17 Snedecor, G.W. and Cochran, W.G., Statistical Methods, 6th edition, Iowa State University Press, Ames, IA, 1978. 18 Lewis, B., Chait, A., Sigurdsson, G., Mancini, M., Farinaro. E., Oriente, P., Carlson, L.A., Ericsson, M., Micheli, H. and Pometta, D., Serum lipoproteins in four European Communities - Quantitative comparison, Eur. J. Clin. Invest., 8 (1978) 165. 19 Hasstedt, S.J., Albers, J.J., Cheung, M.C., Jorde, L.B., Wilson, D.E., Edwards, C.O., Cannon, W.N., Ash, K.O. and Williams, R.R., The inheritance of high density lipoprotein cholesterol and apolipoproteins A-I and A-II, Atherosclerosis, 51 (1984) 21. 20 Debray, Q. and Lalouel, J.M., Definition des facteurs d’environment par rapport au facteur gtnetique - Leur contribution a la ressemblance familiale, J. Genet. Hum.. 29 (1981) 505. 21 Enger, S.C., Herbjomsen, K., Erikssen, K. and Fretland, A., High density lipoproteins (HDL) and physical activity The influence of physical exercise, ages and smoking on HDL cholesterol and the HDL/total cholesterol ratio, Stand. J. Clin. Lab. Invest., 37 (1977) 251. 22 Morrison, J.A., Mellies, M., De Groot, J., Khoury, P., Gartside, P.S. and Glueck, C.J., Cigarette smoking, alcohol intake, and oral contraceptive - Relationships to lipids and lipoproteins in adolescent school-children, Metabolism, 28 (1979) 1166. 23 Heyden, S., Heiss, G., Manegold, C., Tyroler, H.A., Hames, C.G., Bartel, A.G. and Copper, G., The combined effect of
29 smoking and coffee drinking on LDL and HDL cholesterol, Circulation, 60 (1979) 22. 24 Castelli, W.P., Doyle, J.T., Gordon, T., Hames, C.G., Hjortland, M.C., Hulley, S.B., Kagan, A. and Zukel, W.J., Alcohol and blood lipids The cooperative lipoprotein phenotyping study, Lancet, ii (1977) 153. 25 Willett. W., Hennekens, C.H., Siegel, A.J., Adner, M.M. and Castelli, W.P.. Alcohol consumption and high density lipoprotein cholesterol in marathon runners, N. Engl. J. Med., 303 (1980) 1159.
26 Pometta, D. and Micheli, H., Atherosclerose coronarienne et lipoproteines seriques. Schweiz. Med. Wschr., 109 (1979) 1926. 27 Oberhansli, I., Pometta, D., Micheli, H., Raymond, L. and Suenram. A., Lipid, lipoprotein and apo-A and apo-B lipoprotein distribution in Italian and Swiss school children The Geneva Survey, Pediat. Res., 280 (1982) 1563.