© 1996 American Heart Association, Inc.
Articles |
Poland and United States Collaborative Study on Cardiovascular Epidemiology
A Comparison of HDL Cholesterol and Its Subfractions in PopulationsCovered by the United States Atherosclerosis Risk in Communities Study and thePol-MONICA Project
yna Broda
ska
From the Department of CVD Epidemiology and Prevention (G.B., S.L.R.),
Stefan Cardinal Wyszyski National Institute of Cardiology, Warsaw,
Poland; the School of Public Health (C.E.D.), University of North Carolina,
Chapel Hill; the Unit of Clinical Epidemiology (A.P., E.B.), School of Public
Health, Collegium Medicum, Jagiellonian University, Kraków, Poland; the
School of Public Health (O.D.W.), University of Alabama, Birmingham; the
School of Public Health (A.R.F.), University of Minnesota, Minneapolis; and
the School of Public Health (M.S.), Johns Hopkins University, Baltimore, Md.
Correspondence to Sandra Irving, Collaborative Studies Coordinating Center, Suite 203, 137 E Franklin St, Chapel Hill, NC 27514. E-mail uccshi.cscc@mhs.unc.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMethods Results Discussion Appendix References |
|---|
Abstract HDL cholesterol (HDL-C) levels are inversely related to coronary heart disease (CHD) risk, and HDL-C distributions vary among countries. Poland is one of the few developed countries in which CHD rates are increasing at the same time that US rates have been falling, but whether these differences are explained by differences in risk factors such as HDL-C has not been determined. To examine this possibility, levels of HDL-C and its subfractions were compared in US and Polish urban and rural men and women aged 45 to 64 years. Age-adjusted HDL-C means were 0.20 mmol/L higher in urban Polish men and 0.37 mmol/L higher in rural Polish men than in their US counterparts (P<.0001); means in urban Polish women were 0.06 mmol/L higher (P<.05) and in rural Polish women 0.09 mmol/L higher (P<.001) than in their US counterparts. Adjustment for age, education, alcohol intake, smoking, BMI, heart rate, and menopause status (in women) had little effect on differences. Means of HDL2 and HDL3 levels showed similar between-country differences, although differences were minimal for HDL2 in urban men and women, and HDL3 means did not differ between rural women. BMI was inversely related to HDL-C and both subfractions in all gender-country-site strata (P<.001), and alcohol was directly related to HDL-C (P<.001) in all strata except Polish women. Cigarette smoking was negatively related to HDL-C and both subfractions in all US samples except HDL2 in urban men, whereas in Polish samples, significant associations were found only in urban women for HDL-C and in rural and urban women for HDL3. Age, heart rate, and education showed inconsistent or no association with HDL-C and its subfractions in either country. This profile of HDL-C and its subfractions in Polish samples contrasts sharply with the opposite trend in CHD mortality rates, which suggests either that other risk factors may account for the trends or that the relationship between HDL-C and CHD may differ between the two countries.
Key Words: cardiovascular disease • HDL cholesterol • Poland • United States
|
Introduction |
|---|
|
TopAbstractIntroductionMethods Results Discussion Appendix References |
|---|
In Poland, CHD mortality rates continue to increase, whereas in the United States a decrease has been observed since the early 1960s. The analysis by Uemura and Pisa1 shows that from 1970 through 1985 the CHD mortality rates in populations aged 30 to 69 years decreased in the United States by 49% in men and 48% in women. During the same time in Poland, CHD mortality increased by 72% and 59%, respectively. In 1985 the CHD mortality rates in men were similar in the United States and Poland (235 and 230 per 100 000 population), and in women the CHD mortality rate was even higher in the United States (80 per 100 000 population) than in Poland (54 per 100 000 population).1 These contrasting trends enhanced the motivation of Poland and the National Heart, Lung, and Blood Institute of the United States to undertake joint scientific studies in CVD epidemiology. The aim of these studies is to evaluate CHD risk, risk factor levels, and trends within both countries and their potential contribution to differences in the overall mortality between the two countries. There is also interest in the variations of risk within and between countries according to different economic, geographic, and occupational characteristics. Among many examined CHD risk factors, HDL-C level distribution is of great interest in both countries.
Data from prospective and prevalence studies indicate that a low level of HDL-C is associated with increased CHD risk.2 3 4 5 However, HDL-C includes two major subfractions: lipid-enriched HDL2, which has antiatherogenic properties, and protein-enriched HDL3, whose role in the antiatherogenic process is less clear.3 6 7 Several factors related to CHD are also associated with changes in HDL-C levels. Higher HDL-C levels are associated with exercise, moderate alcohol consumption, and exogenous hormone use in females, whereas lower HDL-C levels are associated with progestogen use in males, obesity, high carbohydrate intake, adult diabetes, hypertriglyceridemia, low physical activity, and cigarette smoking.8 9 10 11 Factors influencing particular HDL-C subfraction concentrations are less well known.
This article compares the concentrations and distributions of HDL-C and its subfractions in samples from population surveys of US and Polish populations and assesses the relationships between CHD risk factors and HDL-C, HDL2, and HDL3 within and between the two countries. The between-country comparisons were based on populations with similar demographic and occupational characteristics.
|
Methods |
|---|
|
TopAbstractIntroductionMethods Results Discussion Appendix References |
|---|
Study Populations and Data Collections
Data from the first examination of the US ARIC study and the second screening of the MONICA project in Poland (Pol-MONICA–Warsaw and Kraków) are the basis for this report. Study designs, populations, methods, and standardization procedures have been described.12 13 Briefly, US data were collected between 1987 and 1989 for participants aged 45 to 64 years from probability samples from four communities (Forsyth County, North Carolina; Jackson, Miss; suburbs of Minneapolis, Minn; and Washington County, Maryland). Only data from white men and women from the Minneapolis and Washington County communities (henceforth labeled as urban and rural, respectively, for the United States, even though Washington County is a semirural area) were included in the present analysis. Data from the two other ARIC sites were not used because there were no comparable Polish data. Polish data were collected between 1987 and 1988 and included men and women aged 35 to 64 years residing in Tarnobrzeg Province in southeastern Poland (Pol-MONICA Kraków, henceforth called Polish rural) and two residential districts of Warsaw (Pol-MONICA Warsaw, henceforth called Polish urban). The present report includes only men and women aged 45 to 64 years.
Data collection between and within countries was standardized through the use of common methodology and similar training and local certification procedures of data collection personnel. Data on age, education, smoking, antihypertensive medication use, antihyperlipidemic medication use, and hormone use by women were obtained by using a questionnaire. The Polish questionnaire for commonly collected data was adapted from translations of questionnaires of several US studies and the World Health Organization MONICA project and tested for use in Poland. The cut point used for educational comparisons was 12 years, which is equivalent to completion of secondary school in Poland and high school in the United States. Alcohol consumption was estimated by a dietary questionnaire; in US centers the participants were asked about their usual weekly consumption of wine, beer, or hard liquor, and in Poland they were asked about consumption during the past week. Daily consumption was estimated by computing the weekly consumption of ethanol and dividing by seven. Women were classified as postmenopausal if they reported no menstrual period in the prior 2 years in the United States and in Poland if they do not currently have menstrual periods. Weight was measured on balanced beam scales for participants without shoes and heavy outer garments. Standing height was measured to the nearest centimeter following weight measurement. HR in Poland was measured in a sitting position after 5 minutes of rest by counting a radial pulse in a 15-second interval and multiplying the value by four. In the United States, HR was measured in the supine position during a standard 12-lead resting electrocardiogram examination.
The subjects fasted overnight for at least 12 hours. In both US and Polish centers blood samples were collected with EDTA as the anticoagulant. Plasma TC determinations were performed by using the direct Liebermann-Burchard method in both Polish centers, whereas in the ARIC centers plasma TC was measured by using an enzymatic method with Boehringer Mannheim kits. In Poland HDL-C and its subfractions were analyzed by using the two-step precipitation technique of Gidez et al.14 HDL-C was evaluated in the supernatant after precipitation with heparin-MnCl2 of apoB-containing lipoproteins. The supernatant was then reprecipitated with dextran sulfate, which resulted in the precipitation of HDL2; the supernatant containing HDL3 was then analyzed for cholesterol content. HDL2-C was calculated as the difference between total HDL-C and HDL3-C. In US centers the precipitation method of Warnick et al15 was used for HDL-C, HDL2, and HDL3 estimations. HDL-C was separated by precipitating LDL and VLDL with dextran sulfate and magnesium ions. HDL3 was isolated by precipitating the HDL2 from total HDL-C by increasing the concentrations of these reagents. The amount of HDL2-C was determined by subtracting the subfractions of HDL3-C from HDL-C. External and internal quality control procedures were used to monitor the quality of the laboratory analyses. The Polish laboratories successfully participated in the World Health Organization Lipid Reference Program. The US centers and the Kraków center successfully participated in the Lipid Standardization Program of the Centers for Disease Control and Prevention (CDC); the Kraków center was responsible for the additional standardization of the Warsaw center via CDC Lipid Standardization Program methods.
Since this article focuses on HDL-C, HDL2, and HDL3 concentrations among persons without conditions that are likely to influence lipid concentrations, the following persons were excluded from analysis: those who had fasted for less than 12 hours prior to examination (United States, 242; Poland, 59); those taking antihyperlipidemic or antihypertensive drugs (United States, 3227; Poland, 299); and women taking hormone medication (United States, 537; Poland, 40). In addition, subjects were excluded who had incomplete or invalid data for any of the variables used in the analyses. A total of 4036 participants were excluded in the United States (of 7908 examined) and 412 participants in Poland (of 1836 examined).
Statistical Methods
Mean values of HDL-C, HDL2,
and
HDL3 were first examined by gender, country, and site.
Differences in adjusted means of HDL-C and its subfractions between
countries were estimated by an ANCOVA model for each site with country
as a variable in the model. Model A adjusted for age only; model B
used age, educational level, BMI, smoking, HR, alcohol consumption, and
menopausal status (in women). Multiple regression analysis was
used to assess the association of selected independent variables
with HDL-C, HDL2, and HDL3 in
each country and area (urban and rural) for men and women separately.
The independent variables included in the models were age,
education (ED) level (ED1=1 for <12 years of school, ED2=1
for 12
years of school, and ED1=ED2=0 for >12 years of school),
alcohol
consumption (weekly alcohol consumption was converted into grams of
ethanol per day with the assumptions that in Poland 1000 mL of beer
contains 31.6 g ethanol, 100 mL of wine contains 9.48 g ethanol, and
100 mL of hard liquor contains 31.6 g ethanol and that in the United
States a 12-oz [355 mL] bottle of beer contains 13.2 g ethanol, a
4-oz [118 mL] glass of wine contains 10.8 g ethanol, and a 1.5-oz
[44-mL] shot of hard liquor contains 15.1 g ethanol), number of
cigarettes smoked per day (exsmokers and nonsmokers were coded as 0),
BMI (body weight in kilograms divided by height in meters squared), and
HR (bpm). The Statistical Analysis System16 was
used for all analyses.
|
Results |
|---|
|
TopAbstractIntroductionMethods Results Discussion Appendix References |
|---|
Descriptive Statistics
This study included 3872 men and women screened in the United States and 1424 subjects screened in Poland. These subjects are not representative of the populations of each country, and interpretation of the results should be limited to the present subjects rather than generalized to the total urban or rural populations of either country. Inferences to national populations should be avoided, but internal relationships at locations within samples are valuable and meaningful. Sample sizes and descriptive statistics for variables used in this article are shown in Table 1
. US participants were more educated, reported a higher
alcohol intake, had a slower HR, and, except for rural women, reported
a lower proportion of current smokers. BMI in men was higher in US
samples; in women it was higher in Polish samples. TC means were higher
in Polish women than in their US counterparts, but in men the
differences between Polish and US samples were not significant.
Proportions of postmenopausal women in both countries were very high
(55% to 73%); the proportion was higher in the Polish urban sample
(73%) than in the US urban sample (55%) but showed little difference
in rural areas (United States, 70%; Poland, 68%). Within countries,
regional differences were greater in Polish than US samples, mostly for
educational level and percentage of current smokers among women.
|
Mean
values and cumulative distributions of total
HDL-C, HDL2, and HDL3 are
presented in Table 2 and Figs 1 and
2. For both countries the unadjusted,
age-adjusted, and multivariable-adjusted means of HDL-C
and its subfractions (especially HDL2) were higher in women
than in men. Unadjusted and age-adjusted means were identical to
two decimal places.
|
|
), 95% and confidence intervals (CI) for the means of
HDL-C (top), HDL2 (middle), and HDL3
(bottom) by site in US and Polish men aged 45-64 years.
|
Substantial intercountry differences were mainly
evident in men. Mean
unadjusted (Fig 1
) and age-adjusted (Table 2)
values of HDL-C,
HDL2, and HDL3 were higher in
Poland than in the United States. The intercountry differences for men
were 0.20 mmol/L for HDL-C, 0.04 mmol/L for
HDL2, and 0.16 mmol/L for HDL3
for urban samples and 0.37, 0.17, and 0.20 mmol/L for rural samples,
respectively (Fig 1 and Table 2). Women were
similar to men, as the
unadjusted (Fig 2) and age-adjusted (Table 2)
HDL-C values were
higher in urban and rural areas in Polish samples than in their US
counterparts; however, the differences were smaller than in men (0.06
in urban and 0.09 mmol/L in rural areas). The intercountry differences
among women in HDL2 and HDL3 means were
less consistent than in men; only HDL2 in
rural areas and HDL3 in urban areas were significant (both
higher in Polish women).
Adjustment for age, education, alcohol
consumption, number of
cigarettes smoked, BMI, HR, and menopause status (in women) had little
effect on the means of HDL-C, HDL2, and
HDL3. After adjustment, these means were significantly
higher in all Polish subjects than in their US counterparts except for
means of HDL3 for rural women, in whom this difference was
not significant (Table 2). The prevalence of low HDL-C (<1
mmol/L) was
about two to six times greater for US than for Polish men (urban, 35%
versus 16%; rural, 42% versus 7%) (Fig 1). For US women (Fig
2) it
was three to four times greater than for Polish women (urban, 8%
versus 3%; rural, 12% versus 3%).
Regression Models
The associations of HDL-C and its
subfractions with age,
education, smoking, HR, BMI, and alcohol consumption are shown in
Tables 3 through 5, which
show results of multiple regression analysis by country,
gender, and site. To facilitate the comparison between the
variables in each model, the coefficients are shown as partial
regression coefficients expressing the predicted changes in HDL-C,
HDL2, and HDL3 levels based on
the following changes: age (5 years), alcohol intake (15 g/d), smoking
status (20 cigarettes/d), BMI (3 kg/m2), and HR (10
bpm).
HDL-C (Table 3)
Men
Age was
significantly and positively associated with HDL level
only in the US rural sample. Lower education was positively related to
HDL in both Polish samples, with a much stronger association in the
rural than in the urban sample; low-educated Polish men in urban
and rural areas had 0.13 and 0.33 mmol/L higher HDL-C levels,
respectively, than subjects with higher education. In the United States
a significant but rather weak association of education with HDL-C was
found only in the urban sample and, in contrast to Polish samples,
lower education was inversely related to HDL-C. Cigarette smoking was
inversely related to HDL-C in all samples, but it reached significance
only in the US samples, showing a decrease in HDL-C level of about
0.08, to 0.09 mmol/L, for every 20 cigarettes/d smoked. Alcohol
consumption was positively and significantly related to HDL in all four
samples, and the strength of the association was similar across
samples. With every 15 g/d of ethanol consumption, HDL-C increased
about 0.08, to 0.14 mmol/L. BMI was negatively related to HDL-C in all
four samples. An increase of 3 kg/m2 in BMI decreased HDL-C
about 0.08, to 0.1 mmol/L. HR showed a significant negative but rather
weak association with HDL-C in the US urban sample only. The models
explained only about 14% to 19% of HDL variability in both
countries.
|
Women
Age was not significantly
associated with HDL-C in any of the four
samples. Lower education level was significantly and inversely related
to HDL-C in US urban women only. Cigarette smoking was significantly
and negatively associated with HDL in both US samples and in Polish
urban women. This association in women was much stronger than in men,
showing a decrease in HDL-C of about 0.13, to 0.20 mmol/L, for every 20
cigarettes/d smoked. Alcohol was positively correlated with HDL in
women, although the association was significant in US samples only; the
strength of the association was approximately twofold higher than in
men. BMI was negatively associated with HDL-C in all four samples; the
strength of the association was similar to that for men. HR was not
significant in any of the four samples. The models explained only about
10% to 20% of HDL-C variability and was slightly better fitted to US
than Polish samples.
HDL2 (Table 4)
Men
Age was weakly significant and positively associated with
HDL2 in the US rural sample only. Lower education
level and HDL2 were negatively related in the US
urban sample and positively related in the Polish rural sample.
Cigarette smoking was negatively and significantly associated with
HDL2 in the rural US sample only. Alcohol was
significantly and positively related to HDL2 in all
four samples; the strength of the associations was higher in the Polish
samples. BMI was significantly and negatively related to
HDL2 in all four samples; the strength of the
association was similar across the samples. HR showed significant but
rather weak inverse association with HDL2 in US
urban men only. The models accounted for only 7% to 12% of
HDL2 variability (slightly more in Polish than in US
samples).
|
Women
Age was not significantly related
to HDL2 in
any of the four samples. Education was significant only in the US urban
sample, with an inverse association of lower education and
HDL2 level. Cigarette smoking was significantly and
inversely related to HDL2 only in the US subjects
and was not significant among their Polish counterparts. Alcohol
consumption was also significant only in the US samples, with a
positive association with HDL2. BMI was negatively
associated with HDL2 in all four samples. HR showed
a significant inverse but rather weak association with
HDL2 in urban US women only. The models explained
about 11% to 16% of HDL2 variability in US samples
and 4% to 7% of HDL2 variability in Polish
samples.
HDL3 (Table 5)
Men
Age and HR were not significantly related to HDL3 in
any of the four samples. Education level was significant only in the
Polish samples, with a positive association of lower education and
HDL3. Cigarette smoking was inversely and significantly
related to HDL3 in US urban and rural men but not in their
Polish counterparts. All four samples had a significant association
between HDL3 and alcohol consumption (positively) and BMI
(negatively); the strength of the association was similar across the
samples. The models explained slightly more HDL3
variability in the US (16% to 18%) than in the Polish (12% to 14%)
samples.
|
Women
Education and HR were not
significantly related to
HDL3 in any of the four samples. A significant but rather
weak positive association of age with HDL3 was found only
in urban US women. Cigarette smoking and BMI were significantly and
negatively related to HDL3 in all four samples; the
strength of the associations was similar across the samples. Alcohol
was significantly and positively related to HDL3 in both US
samples. The models accounted for more HDL3 variability in
the US (11% to 15%) than in the Polish (6% to 8%) samples.
In summary, alcohol was positively related to HDL and its subfractions for men in both sites and countries but only for women in the United States. BMI was inversely related to HDL and its subfractions in both countries, sites, and genders. Cigarette smoking was inversely related to HDL in both US sites and genders and to Polish urban women. The relation of other variables to HDL and its subfractions varied for country, site, and gender.
|
Discussion |
|---|
|
TopAbstractIntroductionMethods Results Discussion Appendix References |
|---|
These results raise interesting questions best addressed in a prospective study of the determinants of CVD morbidity and mortality in two countries with strikingly different CHD mortality trends. A systematic reexamination and quantitative analysis of four large prospective American studies (the Framingham Heart Study, the Lipid Research Clinics Prevalence Mortality Follow-up Study, the Coronary Primary Prevention Trial, and the Multiple Risk Factor Intervention Trial) identified HDL-C as an independent and strong protective factor for CHD in US populations. A 1-mg/dL (0.026-mmol/L) increment in HDL-C was associated with a CHD risk decrement of 2% to 3%. These findings were observed in men and women over 6 to 10 years of follow-up.4 In contrast, the first report from the British Regional Heart Study (BRHS) suggested that HDL-C was not a significant risk factor for CHD after adjustment for other risk factors.17 When BRHS results were reviewed with American studies and the BRHS data were analyzed by using a similar statistical model as in analyses of US studies (TC was included as a covariate), a consistent and significant inverse relation of HDL-C and CHD was apparent.4 The second report from BRHS also confirmed these associations18 during 7 years of follow-up. Comparable data are not available for Poland; however, preliminary results from a 7-year follow-up of the Warsaw Pol-MONICA population cohort provide hints of a similar negative association between HDL-C and CHD mortality (unpublished data for the years 1984 through 1990) even though the mean HDL-C concentration is rather high in the Warsaw population.
There is, therefore, clear evidence that a low level of HDL-C is directly related to CHD risk within populations,19 20 21 but no uniform evidence is available as to whether this is due to one or both of its subfractions. A widely held view is that any benefit of high HDL-C is mainly or solely associated with the HDL2 subfraction.19 21 Two studies,20 22 however, have reported that HDL3 but not HDL2 is negatively associated with arteriosclerosis.
The main aim of this study was to report the major differences in HDL-C, HDL2, and HDL3 concentrations and relate them to behavioral and demographic factors in samples from two countries with opposite trends in CHD mortality. Our observations were restricted to samples of the same race, gender, age, and regional structure and to comparable methods of collecting and scoring most of the independent variables. Moreover, the opportunity to fit the same statistical models to the separate data sets allowed us to make direct comparison of the functional relationships (the multivariate-adjusted regression coefficients) of the HDL-C correlates among populations from two countries.
Substantial differences in the concentrations of HDL-C and its subfractions in the US and Polish population samples reported here were found mainly in men. Mean HDL-C and its subfractions were higher in Polish urban and rural men; these differences persisted after adjustment for lifestyle and social variables. Especially large intercountry differences were observed for HDL-C and HDL2 in the rural areas, which may be partly related to the higher physical activity in the work of rural Polish men (due to the rather low mechanization of Polish agriculture). In women, between-country differences were similar to those in men but much weaker; HDL3 in rural areas was not significant. The proportion of subjects with an HDL-C level <1 mmol/L was three to five times higher in US men and women than in their Polish counterparts. It must be noted that for the data in this report, the United States and Poland did not use identical methods of HDL-C measurement. Experiments reported by Warnick et al23 suggest that although the HDL-C values obtained by various precipitation methods are highly correlated, there are systematic differences between methods, and care must be taken in comparing results obtained by two different precipitation techniques. In the analyses done by Warnick et al,23 the plasma HDL-C values obtained by using the dextran sulfate–Mg2+ method were consistently 0.05 to 0.1 mmol/L lower than those obtained by using the heparin-Mn2+ method, so it may be concluded that real differences between the two countries were smaller than those presented here.
HDL-C concentration, which was consistently lower in men than in women in all groups, showed almost no association with age. The restricted age range (45 to 64 years) may be the reason for this. Gender differences in HDL-C were predominantly a consequence of HDL2 differences. The gender differences in HDL-C may result either directly, from physiological differences, or indirectly, from differences in the prevalence or effects of other factors that influence lipoprotein levels. Because gender differences in HDL-C persisted after controlling for age, education, alcohol intake, smoking status, BMI, HR, and menopause status (in women), our study supports the first hypothesis. This conclusion is also supported by other studies24 25 ; however, Godsland et al,26 who reviewed several studies on plasma lipids and atherosclerosis, found that gender differences in HDL-C in communities appeared to be minimal when CHD morbidity in a population was a less important health problem. They stated, therefore, that gender differences in lipoprotein risk factors cannot be ascribed solely to intrinsic physiological differences between men and women. They concluded that the interaction of gender with other factors, perhaps related to SES, might be invoked to explain the gender differences in lipoprotein levels that are observed in societies with moderate-to-high CHD incidence, such as the United States and Poland.
The multiple regression analysis showed that while there were similarities in the relationships of lifestyle and sociodemographic variables to HDL-C and its subfractions for both the US and Polish samples, there were also important differences. BMI was the strongest predictor of HDL-C and both of its subfractions in all US and Polish samples regardless of gender and region. The regression coefficient suggests a similar magnitude of association. An inverse relationship between HDL-C and BMI has been reported by several investigators,8 10 11 27 whereas some have found, contrary to our findings, that BMI was related strongly to HDL2 but not HDL3.8 11 27 Obesity is often associated with resistance to insulin-mediated glucose uptake and hyperinsulinemia. Peripheral insulin resistance leads, among other consequences, to a decrease of HDL-C concentration and an increase of triglyceride and LDL-C concentrations through the mechanism of enhanced lipolysis and increased release of free fatty acids from adipose tissue.28 Obesity, which is also associated with hypertension, type II diabetes, and dyslipidemia, is a common problem in industrialized societies; in the United States, obesity is a major contributor to a continued high prevalence of CHD and premature death.29 In our study, in both US and Polish samples, the means of BMI were above desirable values,30 which indicates that a substantial fraction of these subjects were overweight or obese.
Alcohol intake is a lifestyle factor often thought to offer health benefits as well as hazards. Among the putative health benefits, there is the well-established positive correlation between alcohol intake and HDL-C31 32 whereby alcohol may reduce CHD risk. However, relatively little information is available on the relationship of HDL-C subfractions to alcohol intake. In our study, alcohol consumption was associated positively with HDL-C and both subfractions in all US subjects and in Polish men. The associations were stronger in Polish than in US men, but alcohol consumption was higher among US subjects. This may indicate the underreporting or inaccuracy in evaluation of alcohol consumption of Polish men. Another explanation may be different drinking habits in the two countries. In Poland soft, everyday drinks low in alcohol are not popular. Polish men usually drink strong alcohol (vodka) in a relatively high amount on one occasion (200 g and more), but frequency of drinking is probably lower than in the United States. The lack of any significant association of alcohol consumption with HDL-C and its subfractions in Polish women is probably due to the very low alcohol consumption by women in Poland (less than 1.0 g/d). Our results corroborate some studies,8 32 but others have reported that alcohol consumption is more strongly related to HDL3 than HDL2.10 27 Alcohol reduces hepatic lipase activity, which is responsible for the removal of HDL-C particles from the circulation, and this mechanism could be one of the explanations of an increase of HDL-C in people with moderate alcohol consumption.32
Smoking is associated with lower levels of HDL-C in many8 33 34 but not all35 studies. The observed lower levels of HDL-C in some studies were primarily due to a reduction in HDL3,33 while in one study smoking was associated with a decrease in both HDL-C subfractions.34 The decreased HDL-C in chronic smokers is postulated to be a result of lowered synthesis of HDL-C by an inhibition of triglyceride-rich VLDL and chylomicron catabolism.36 In our study, smoking was strongly and negatively related to HDL-C and to both its subfractions mainly in US men and women. In Poland, this association was much weaker and significant in urban women for HDL-C and HDL3 and in rural women only for HDL3. It is worth noting that in Polish men, among whom the prevalence of smoking was the highest, we found no significant association between smoking and HDL-C or its subfractions. The lack of a significant relationship between smoking and HDL in Polish men may be partly explained by a much higher alcohol consumption and lower BMI among Polish smokers than nonsmokers as well as by lower education among smokers, which in turn is connected with higher physical activity at work.37
Increased HR has been reported to be a risk factor for
CHD.38 HR may influence atherosclerosis
directly because of local hemodynamic effects on
vessels.39 Another possibility is that HR is not causally
related to CHD but is associated with CHD through factors such as
exercise, smoking, blood pressure, and lipid abnormalities. The
cross-sectional Tromso Study,40 which involved more
than 19 000 men and women, reported that HR was positively correlated
with LDL cholesterol and triglyceride levels
and negatively with HDL-C. Mechanisms for these relations are not
clear. Activation of the autonomic nervous system by many factors,
including stress and smoking, may be one of the explanations. With
regard to HDL-C, the predominant 
-1 receptor stimulation associated
with enhanced sympathetic activity may induce changes in
precapillary sphincter tone and decrease delivery of
triglyceride-rich lipoproteins to
endothelial lipase, resulting in decreased VLDL
catabolism and HDL-C synthesis.41 Another explanation is
that HR can reflect physical and cardiorespiratory
fitness42 (physically active people have lower HR than
sedentary individuals). Regular moderate physical exercise increases
HDL-C and HDL2 concentration by increases of dietary
fat tolerance, insulin sensitivity, and lipoprotein lipase
activity.43 In our analysis HR was a weak
predictor of HDL and its subfractions in US samples and was not
significant in Polish samples, although means of HR were much higher in
Polish than in US samples. The differences in HR means between US and
Polish samples are probably in part a result of differences in methods
of measurement of HR (from pulse in Poland and from resting
electrocardiogram in the United States) and a higher
prevalence of smoking and a higher blood pressure level among Polish
subjects. It is difficult to say to what extent work- and
leisure-time physical activity differ between US and Polish
subjects because of a lack of comparable physical activity evaluation
in the US ARIC and Pol-MONICA Studies.
Education level showed an inconsistent association with HDL-C and its subfractions in both countries. In Poland, a lower education level was significantly associated with higher levels of HDL-C and its subfractions in men (except HDL2 in the urban area) but not in women; in the United States a significant relationship was found for HDL-C and HDL2 in men and women from urban areas, but this association was in an opposite direction than in Polish samples. It is unlikely that education itself influences HDL-C level; rather, it is affected by other lifestyle factors connected with education, such as physical activity, nutritional habits, and alcohol consumption.43 44 45 Education is an important marker for SES. International comparisons show that in developing countries cardiovascular risk increases with increasing SES.46 It has been postulated that the initial effect of industrialization and affluence on lifestyle is an increase in CHD risk factors caused by unfavorable changes in the health behaviors of the society that are characterized by an intake of a diet rich in fat and calories, obesity, and physical inactivity.46 47 After some period of higher SES and higher CHD risk, health consciousness appears to beneficially affect diet, physical activity, and smoking habits so that the relationship between SES and CHD risk becomes inverse. This has happened in the United States, where the decline in CHD risk factors and mortality that has been observed for many years1 47 is most pronounced in the highest SES groups. Because SES transformation began much later in Poland than in the United States, Poland can be considered to be a developing country, which probably partly explains the increasing CHD mortality trend observed during the years that preceded this study.1 The opposite associations between HDL and education observed in the present study between the United States and Poland may be partially a result of inequalities in the level of SES development in the two countries.
Menopausal status is considered to be a risk factor for CHD, and
many studies have focused on lipid changes in postmenopausal
women.48 Our study included women aged 45 to 64 years, a
majority of whom (
70% in Poland and 50% to 70% in the United
States) were postmenopausal; most of the remainder were
perimenopausal (data not shown). We therefore decided not to enter
menopausal status into the regression models.
In conclusion, among the subjects in the present study the best predictors of HDL-C and its subfractions were BMI, smoking, and alcohol consumption. The composite effect of major risk factors such as age, smoking, alcohol consumption, BMI, HR, and education explained only a small fraction of HDL-C, HDL2, and HDL3 variability in both countries, and this fraction of explained variability was generally a little higher among US subjects. The remainder of the variability may be attributed to measurement errors, intraindividual variation of risk factors, and other factors not included in our analyses, such as physical activity, nutritional habits, and genetic factors.43 44 45 In Polish subjects levels of HDL-C, HDL2, and HDL3, both unadjusted and adjusted for age and lifestyle factors, were higher than in US subjects. These differences contrast sharply with rising CHD rates in Poland and suggest either that other risk factors account for this trend or that the relationship between HDL-C and CHD risk may differ between the two countries. Many researchers have indicated that whereas median serum cholesterol levels correlate well with fatal ischemic heart disease rates internationally, no such relation exists in countries in which HDL concentrations are low and ischemic heart disease is not prevalent.49 50 51 52 More research is required to further define the association between HDL-C and CHD in terms of possible synergistic and antagonistic effects among HDL-C and other risk factors.
|
Selected Abbreviations and Acronyms |
|---|
|
|
Acknowledgments |
|---|
This study was supported by a Polish Governmental Grant, contract 11.6, by the State Committee for Scientific Research, contract Nos. 0253/S4/92/02 and 4-1474-9101; by the National Heart, Lung, and Blood Institute, contract Nos. No1-HV-1-2243, No1-HV-1-08112, No1-HC-55020; and by US-PL II Fund Maria Sktodowska-Curie No. Mz/HHS-94-184. The authors express their sincere thanks to Sandy Irving for her editorial help, Dr Teri Manolio for her valuable comments, Diane Williams and Dr Yuan Li Shen for all computations performed, Kim Ray for assistance with the figures and editing, and Pat Coley for typing the manuscript.
|
Appendix |
|---|
|
TopAbstractIntroductionMethods Results Discussion Appendix References |
|---|
Pol-MONICA Acknowledgment List
Stefan Cardinal Wyszy
ski National Institute of
Cardiology, Warsaw-Maria Polakowska, Anna
Ku
miska, Henryka Wagrowska, and Pawel Kurijata
School of Public Health, Collegium Medicum, Jagiellonian
University, Kraków-Malgorzata Kwa
niak, Malgorzata
Malczewska, Iwona Morawska, and Witold Rostworowski
ARIC
Acknowledgment List
University of Minnesota, Minn-Caryl
De Young, Jaci
Dion, Lowell Bedquist, and Chris Bunkins
The Johns Hopkins University, Baltimore, Md-Patricia M. Crowley, Lilly Downs, Sunny Harrell, and Patricia Hawbaker
The Methodist Hospital, Atherosclerosis Clinical Laboratory, Houston, Tex-Doris J. Epps, Charles E. Rhodes, and Selma M. Soyal
University of North Carolina, Chapel Hill (Coordinating Center)-Chris Coffey, John H. Crouch, Thomas G. Goodwin, and Dick Hayes
Received July 20, 1995; accepted November 28, 1995.
|
References |
|---|
|
TopAbstractIntroductionMethods Results Discussion Appendix References |
|---|
1. Uemura K, Pisa Z. Trends in cardiovascular disease mortality in industrialized countries since 1950. World Health Stat Q. 1989;41:155-178.
2. Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. High density lipoprotein as a protective factor against coronary heart disease: the Framingham Study. Am J Med. 1977;62:707-714. [Medline] [Order article via Infotrieve]
3. Miller NE, Hammett F, Saltissi S, Rao S, van Zeller H, Coltart J, Lewis B. Relationship of angiographically defined coronary artery disease to plasma lipoprotein subfractions and apolipoproteins. Br Med J. 1981;282:1741-1744.
4.
Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD,
Castelli WP, Knoke JD, Jacobs DR Jr, Bangdiwala S, Tyroler HA.
High-density lipoprotein cholesterol and
cardiovascular disease: four prospective American
studies. Circulation. 1989;79:8-15.
5. Heiss G, Johnson NJ, Reiland S, Davis CE, Tyroler HA. The epidemiology of plasma high-density lipoprotein cholesterol levels: the Lipid Research Clinics Program Prevalence Study. Circulation. 1980;62(suppl IV):IV-116-IV-136.
6. Tall AR. Plasma high density lipoproteins: Metabolism and relationship to atherogenesis. J Clin Invest. 1990;86:379-384. Review.
7. Shepherd J, Packard CJ, Stewart JM, Vallance BD, Lawrie TD, Morgan HG. The relationship between cholesterol content and subfraction distribution of plasma high density lipoproteins. Clin Chim Acta. 1980;101:57-62. [Medline] [Order article via Infotrieve]
8.
Steenkamp HJ, Jooste PL, Spiler-Benade AJ, Langenhoven
ML, Rossouw JE. Relationship between high-density
lipoprotein subfractions and coronary risk factors in a rural
white population.
Arteriosclerosis. 1990;10:1026-1031.
9. Levy RI, Rifkind BM. The structure, function and metabolism of high-density lipoproteins: a status report. Circulation. 1980;62(suppl IV):IV-4-IV-8.
10.
Haffner SM, Applebaum-Bowden D, Wahl PW, Hoover JJ,
Warnick GR, Albers JJ, Hazzard WR. Epidemiological correlates of
high-density lipoprotein subfractions, apolipoproteins A-I, A-II,
and D and lecithin cholesterol acyltransferase: effects of
smoking, alcohol, and adiposity.
Arteriosclerosis. 1985;5:169-177.
11. Albrink MJ, Krauss RM, Lindgren FT, Von der Groeben J, Pan S, Wood PD. Intercorrelations among plasma high density lipoprotein, obesity and triglycerides in a normal population. Lipids. 1980;15:668-676. [Medline] [Order article via Infotrieve]
12.
The ARIC Investigators. The
Atherosclerosis Risk in Communities (ARIC) Study:
design and objectives. Am J Epidemiol. 1989;129:687-702.
13. World Health Organization. MONICA Manual. 2nd ed. Geneva, Switzerland: World Health Organization; November 1990.
14. Gidez LI, Miller GJ, Burstein M, Slagles S, Eder HA. Separation and quantitation of subclasses of human high density lipoprotein subclasses by a simple precipitation procedure. J Lipid Res. 1982;23:1206-1223. [Abstract]
15. Warnick GR, Benderson JM, Albers JJ. Quantitation of high-density lipoprotein subclasses after separation by dextran sulfate and Mg+2. Clin Chem. 1982;28:1574. Abstract.
16. SAS Software. Version 6.03. Cary, NC: SAS Institute Inc; 1988.
17. Pocock SJ, Shaper AG, Phillips AN, Walker H, Whitehead TP. High-density lipoprotein is not a major risk factor for ischaemic heart disease in British men. Br Med J. 1986;292:515-519.
18. Pocock SJ, Shaper AG, Phillips AN. Concentration of high density lipoprotein cholesterol, triglycerides, and total cholesterol in ischaemic heart disease. Br Med J. 1989;298:998-1002.
19. Stampfer MJ, Sacks MF, Salvini S, Willett WC, Hennekens CH. A prospective study of cholesterol, apolipoproteins, and the risk of myocardial infarction. N Engl J Med. 1991;325:373-381. [Abstract]
20.
Levy RL, Brensike JF, Epstein SE, Kelsey SF, Passamani
ER, Richardson JM, Loft IK, Stone NJ, Aldrich RF, Battaglini JW,
Moriarty DJ, Fisher ML, Friedman L, Friedewald W, Detre KM. The
influence of changes in lipid values induced by cholestyramine and diet
on progression of coronary artery disease: results of the NHLBI
Type II Coronary Intervention Study.
Circulation. 1984;69:325-337.
21. Musliner TA, Krauss RM. Lipoprotein subspecies and risk of coronary disease. Clin Chem. 1988;34(suppl):B78-B83.
22. Kempen HJ, van Gent CM, Buytenhek R, Buis B. Association of cholesterol concentrations in low-density lipoprotein, high-density lipoprotein, and high-density lipoprotein subfractions, and of apolipoproteins AI and AII, with coronary stenosis and left ventricular function. J Lab Clin Med. 1987;109:19-26. [Medline] [Order article via Infotrieve]
23.
Warnick GR, Cheung MC, Albers JJ. Comparison of
current methods for high-density lipoprotein
cholesterol quantitation. Clin Chem. 1979;25:596-604.
24. Vodak PA, Wood PD, Haskell WL, Williams PT. HDL-cholesterol and other plasma lipid and lipoprotein concentrations in middle aged male and female tennis players. Metabolism. 1980;29:745-752. [Medline] [Order article via Infotrieve]
25. Nichols AV. Human serum lipoproteins and their inter-relationship. Adv Biol Med Phys. 1967;11:109-158. [Medline] [Order article via Infotrieve]
26. Godsland IF, Wynn V, Crook D, Miller NE. Sex, plasma lipoproteins, and atherosclerosis: prevailing assumptions and outstanding questions. Am Heart J. 1987;114:1467-1503. [Medline] [Order article via Infotrieve]
27. Diehl AK, Fuller JH, Mattock MB, Salter AM, el-Gohari R, Keen H. The relationship of high density lipoprotein subfractions to alcohol consumption, other lifestyle factors, and coronary heart disease. Atherosclerosis. 1988;69:145-153. [Medline] [Order article via Infotrieve]
28. Kaplan NM. Obesity, insulin and hypertension. Cardiovasc Risk Factors. 1994;3:133-139.
29.
Stamler J. Epidemic obesity in the United
States. Arch Intern Med. 1993;153:1040-1044.
30.
Pyörälä K, De Backer G, Graham I,
Poole-Wison P, Wood D. Prevention of coronary heart
disease in clinical practice: recommendations of the Task Force of the
European Society of Cardiology, European
Atherosclerosis Society and European Society of
Hypertension. Eur Heart J. 1994;15:1300-1331.
31.
Miller NE, Bolton CH, Hayes TM, Bainton D, Yarnell JW,
Baker IA, Sweetnam PM. Associations of alcohol consumption with
plasma high density lipoprotein cholesterol and its major
subfractions: the Caerphilly and Speedwell Collaborative Heart Disease
Studies. J Epidemiol Community Health. 1988;42:220-225.
32. Taskihen MR, Valimaki M, Nikkila EA, Kussi T, Ehnholm C, Ylikahn R. High density lipoprotein subfractions and post heparin plasma lipases in alcoholic men before and after ethanol withdrawal. Metabolism. 1982;31:1168-1174. [Medline] [Order article via Infotrieve]
33. Gnasso A, Haberbosch W, Schettler G, Schmitz G, Augustin J. Acute influence of smoking on plasma lipoproteins. Klin Wochenschr. 1984;62:36-42.
34. Swank AM, Fell RD. Effects of acute smoking and exercise on high density lipoprotein-cholesterol and subfractions in black female smokers. Metabolism. 1990;39:343-348. [Medline] [Order article via Infotrieve]
35. Stamford VA, Matter S, Fell RD. Cigarette smoking, exercise and high density lipoprotein cholesterol. Atherosclerosis. 1984;52:73-83. [Medline] [Order article via Infotrieve]
36. Elkeles RS, Khan SR, Chowdhury O, Swallow MB. Effects of smoking on oral fat tolerance and high density lipoprotein cholesterol. Clin Sci. 1983;65:669-672.[Medline] [Order article via Infotrieve]
37. Grzonkowski ST, Piotrowski W. Alcohol consumption and cigarette smoking and their associations with coronary heart disease risk factors [in Polish]. Zdrowie Publ. 1992;2:88-93.
38. Kannel WB, Kannei C, Paffenbarger RS Jr, Cupples IA. Heart rate and cardiovascular mortality: the Framingham Study. Am Heart J. 1987;113:1489-1494. [Medline] [Order article via Infotrieve]
39.
Beere PA, Glagov S, Zarins CK. Retarding effect
of lowered heart rate on coronary
atherosclerosis. Science. 1984;226:180-182.
40.
Bonaa KH, Arnesen E. Association between heart
rate and atherogenic blood lipid fractions in a population: the Tromso
Study. Circulation. 1992;86:394-405.
41. Sacks FM, Dzau VJ. Adrenergic effects on plasma lipoprotein metabolism: speculation on mechanisms of action. Am J Med. 1986;80(suppl 2A):71-81.
42. American College of Sports Medicine. The recommended quantity and quality of exercise for developing and maintaining cardiorespiratory and muscular fitness in healthy adults. Med Sci Sports Exerc. 1990;22:265-274. [Medline] [Order article via Infotrieve]
43. Despress JP. Impact of exercise on plasma lipids: reducing the risk for coronary heart disease. Can Lipidol Rev. 1994;15:12-19.
44. Tall AR. Metabolic and genetic control of HDL cholesterol levels. J Int Med. 1992;231:661-668. [Medline] [Order article via Infotrieve]
45. Margolis S, Dobs AS. Nutritional management of plasma lipid disorders. J Am Coll Nutr. 1989;8(suppl 1):335-455.
46. Kaplan BA, Kell JE. Socioeconomic factors and cardiovascular disease: a review of the literature. Circulation. 1993;88(pt 1):1973-1998.
47. Luepker RV, Rosamond WD, Murphy R, Sprafka JM, Folsom AR, McGovern BG, Blackburn H. Socioeconomic status and coronary heart disease risk factor trends: the Minnesota Heart Survey. Circulation. 1993;88(pt 1):2172-2179.
48.
Hjortland MC, McNamara PM, Kannel WB. Some
atherogenic concomitants of menopause: the Framingham Study.
Am J Epidemiol. 1976;103:304-311.
49. Watkins LO, Neaton JD, Kuller LH, for the MRFIT Research Group. Racial differences in high-density lipoprotein cholesterol and coronary heart disease incidence in the usual-care group of the Multiple Risk Factor Intervention Trial. Am J Cardiol. 1986;57:538-545. [Medline] [Order article via Infotrieve]
50.
Kesteloot H, Huang DX, Yank XS, Claes J, Rosseneu M,
Geboers J, Jossens JV. Serum lipids in the People's Republic of
China: comparison of Western and Eastern populations.
Arteriosclerosis. 1985;5:427-433.
51. Keys A, Karvonen MJ, Fidanza F. Serum cholesterol studies in Finland. Lancet. 1958;2:175-178. [Medline] [Order article via Infotrieve]
52. Lewis B, Chait A, Sigurdsson G. Serum lipoproteins in four European communities: a quantitative comparison. Eur J Clin Invest. 1978;8:165-173.[Medline] [Order article via Infotrieve]
This article has been cited by other articles:
 |
|
 |
|
  A. Britton and M. McKee The relation between alcohol and cardiovascular disease in Eastern Europe: explaining the paradox J Epidemiol Community Health, May 1, 2000; 54(5): 328 - 332. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||