American Journal of Epidemiology
Copyright O 1996 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol 143, No 11
Printed In U.SA.
Sex Difference in High Density Lipoprotein Cholesterol in Six Countries
C. E. Davis,1 D. H. Williams,1 R. G. Oganov,2 S.-C. Tao,3 S. L. Rywik,4 Y. Stein,5 and J. A. Little6
It is known that women have higher levels of high density lipoprotein (HDL) cholesterol than men. The
authors examined the association between HDL cholesterol and biologic sex in 8,631 women and 10,690 men
aged 45-54 years from six countries studied between 1972 and 1989. The variation in the sex difference for
HDL cholesterol was significant; the smallest difference (0.06 mmol/liter) was seen in China and the largest
(0.40 mmol/liter) in Canada. Adjustment for differences in body mass index, smoking, alcohol use, and heart
rate reduced but did not eliminate the variability. The sex difference in HDL cholesterol levels, usually assumed
to be due to biologic factors, differs across cultures and may be related to environmental factors. Am J
Epidemiol 1996;143:1100-6.
alcohol drinking; body mass index; lipoproteins, HDL cholesterol; sex factors; smoking
Elevated levels of high density lipoprotein (HDL)
cholesterol have been shown to be associated with a
reduced risk of coronary heart disease in many epidemiologic studies (1-4). It has also been observed in
many studies that women have higher HDL cholesterol levels than do men, and this has been hypothesized to be one of the reasons women have a lower
incidence of coronary disease (1, 5-9). It is thus of
interest to estimate the mean difference in HDL cholesterol between men and women and to determine
whether this difference is similar in different ethnic
and cultural settings. It is also important to determine
whether differences in factors such as obesity, alcohol
consumption, and cigarette smoking, which are known
to be correlated with HDL cholesterol levels, can
explain the sex difference in HDL cholesterol.
Received for publication May 25,1995, and in final form February
22, 1996.
Abbreviations: ARIC, Atherosclerosis Risk in Communities; HDL,
high density lipoprotein; LRC, Lipid Research Clinics.
1
Department of Biostatistics, School of Public Health, University
of North Carolina, Chapel Hill, NC.
2
National Center for Preventive Medicine of the Russian Federation, Moscow, Russia.
3
Cardiovascular Institute and Fu Wai Hospital, Chinese Academy of Medical Sciences, Beijing, China.
•* Department of Cardiovascular Disease Epidemiology and
Prevention, Cardinal Wyszynski National Institute of Cardiology,
Warsaw, Poland.
s
Hadassah Medical Organization, The Hebrew University, Jerusalem, Israel.
8
LJpid Research Clinic, St. Michael's Hospital, Toronto, Ontario,
Canada
Reprint requests to Dr. C. E. Davis, Department of Biostatistics,
Collaborative Studies Coordinating Center, Suite 203,137 E. Franklin Street, Chapel Hill, NC 27514.
We report here the estimated sex differences in
HDL cholesterol levels in 23 samples from population-based epidemiologic studies conducted in six different countries. Our specific purposes were: 1) to
estimate the sex difference in HDL cholesterol; 2) to
determine whether this difference was similar in magnitude across a variety of ethnic and cultural populations; 3) to estimate the association of a set of correlates with HDL cholesterol in different cultures; and
4) to adjust the estimated sex difference for differences
in these correlates.
MATERIALS AND METHODS
This report includes data from population-based
studies conducted in Canada (one population), China
(four populations), Israel (one population), Poland
(two populations), Russia (two populations), and the
United States (13 populations). Data collection for the
US Lipid Research Clinics (LRC), the Atherosclerosis
Risk in Communities (ARIC) Study, and the Canadian
and Israeli LRC studies was sponsored by the US
National Heart, Lung, and Blood Institute. Data collection in the People's Republic of China, the Russian
Federation (formerly the Soviet Union), and Poland
was sponsored by grants from the respective governments and by bilateral agreements between the respective governments and the US government. The Department of Biostatistics of the University of North
Carolina, under contracts from the National Heart,
Lung, and Blood Institute, served as the coordinating
center for all of the studies.
This analysis was restricted to men and women aged
45-54 years. Other exclusions included participants
1100
Sex Difference in HDL Cholesterol
known not to have fasted for at least 12 hours, pregnant women, and persons with missing data for HDL
cholesterol. Table 1 describes the populations studied
by country, and gives the dates of the studies and the
sample sizes that were used in this analysis. Study
designs, populations studied, methods used, and standardization procedures have been described in detail
elsewhere (10-18). Brief descriptions are given here.
ARIC data were collected in 1987-1989 from probability samples of four US communities. Data were
collected in various ethnic groups, of which only white
and black men and women were selected for this
analysis. The Polish data were obtained from random
samples of an urban population and a rural population
1101
screened in 1983-1984. In China, samples were selected from two regions and in two settings within
each region. The two regions were Beijing in the north
of China and Guangzhou in the south. In each region,
samples of urban workers and rural farmers were
obtained. Each LRC study screened participants selected from separate, well-defined populations in the
United States, Canada, Israel, and Russia. All 13 clinics (nine US, one Canadian, one Israeli, and two
Russian) followed a standardized protocol. The North
American and Israeli clinics were screened twice between 1972 and 1975. A randomly chosen 15 percent
of all participants from the first screening were invited
to the second screening, along with a selected group of
TABLE 1. Population* and •ample sizes for 23 samples from Canada, China, Israel, Poland, Russia, and the United States,
1972-1989
Population
Location, years, and ref no
Canada (1972-1976) (15)
China (1987-1988) (14)
Beijing—urban
Beijing—rural
Guangzhou—urban
Guangzhou—rural
Israel (1976-1980) (15)
Employees of Canadian Bed Telephone Company, Dofasco Steel Foundry, Simpson's
department store, and Eaton's department store
Workers from the Capital Iron and Steel complex and additional women from a
specified residential district
Men and women from farm brigades in the Shijingshan Agricultural District
Employed or retired workers from Guangzhou Shipyard Company
Men and women working In agricultural villages In the Dashl township of Panyu
County.
Parents of men and women aged 17-18 years reskfing In Jerusalem and examined for
Induction Into the Israeli armed forces.
Poland (1983-1984) (12)
Cracow
Warsaw
Residents of Tamobrzeg Province.
Residents of the Praga North and Praga South Districts of Warsaw
Russia (1972-1982) (16)
Moscow
St Petersburg
Men residing in Oktyabrskfl District: women reskJng In Krasnopresenkil District
Men and women residing In PetrogradsWI District
United States
ARIC* Study (1987-1989) (10)
Blacks (total)
Whites (total)
Jackson
Forsyth County
Blacks
Whites
Minneapolis
Washington County
UpM Research CBnlcs
(1972-1977) (15)
Baltimore, Maryland
Cincinnati, Ohio
Houston, Texas
Iowa
La Jofla, California
Minnesota
Oklahoma
Palo Alto, California
Seattle, Washington
Black residents of Jackson, Mississippi
White and black residents of Forsyth County, North Carolina
White residents of the northwestern suburbs of Minneapolis, Minnesota
White residents of Washington County, Maryland
Adult members of the Columbia Medical Plan.
Parents of students In grades 1, 3, 5, 7, 9,11, and 12 enrolled In the Princeton, Ohio,
School District
Parents of sophomore high school students In selected schools In the Houston
Independent School District
Adults residing In Cedar County, Iowa, and certain rural townships of neighboring
counties on July 1,1973.
Permanent residents of Rancho Bernardo, California, as of August 2,1972.
Residents of four census tracts In Richfield, Minnesota
Residents of five rural Oklahoma counties.
Stanford University employees.
Employees of Pacific Northwest Bell Telephone Company.
TOTAL (all six countries)
• ARIC, Atherosclerosis Risk in Communities
Am J Epidemiol
Vol. 143, No. 11, 1996
Sample size
Men
Woman
131
81
1,688
1,800
690
401
356
642
387
443
241
328
262
211
851
426
425
928
504
424
4,036
2,094
1,942
696
191
505
738
2,450
653
1,288
3,123
1,169
83
743
917
790
119
876
1,175
1,072
536
30
504
18
44
40
34
31
88
38
67
62
77
100
92
54
82
62
64
61
10.690
8,631
1102
Davis et al.
participants with elevated cholesterol or triglycerides.
Only the random sample data from the second screening were used in this analysis. The First Prevalence
Russian LRC Study in 1975-1977 sampled only men,
and those data were used in this analysis. The Second
Prevalence Russian LRC Study in 1978-1982 sampled both men and women, and only the women were
used in this analysis.
Data on age, smoking, alcohol consumption, and
hormone use by women were obtained using a standard questionnaire. Weight was measured on a balanced scale while the participant stood without
shoes and heavy outer garments. Height was measured in the standing position. Body mass index was
computed as weight (kg) divided by height (m)
squared. In all studies but the ARIC Study, heart
rate was measured over a 15-second interval on the
right radial artery and then multiplied by 4 for a
1-minute estimate. In the ARIC Study, 1-minute
heart rate was measured while the participant lay in
a supine position during a standard 12-lead electrocardiographic examination.
Lipid laboratories in each country were standardized
according to the Centers for Disease Control program.
Plasma lipid determinations were made in all of the
samples except those from China and ARIC, where
serum was used. Serum samples were adjusted to be
equivalent to plasma samples using a standard formula
(19). In Poland, the US LRC studies, Canada, Russia,
and Israel, HDL cholesterol was evaluated in the supernatant after precipitation of other cholesterol fractions with heparin and manganese (Mn2 + ) (20). In the
ARIC Study and China, HDL cholesterol was measured using a mixture of dextran sulfate, an analog of
heparin, and magnesium (Mg2 + ) (21).
Since different laboratories were used in the studies,
comparison of HDL cholesterol levels between males
and females was done within each sampled population
and the estimates were combined across samples. For
descriptive purposes, age-adjusted differences between men and women were computed using direct
standardization and the World Health Organization
standard. Totals for countries with more than one
sample were obtained by weighting the average inversely proportional to the variance of the withinsample variance of the sex difference. Mixed linear
models were used to estimate the association between
sex and HDL cholesterol within and across sampled
populations. In these models, the sampled population
(23 levels) was considered a random effect and all
other variables (e.g., age, sex, body mass index) were
considered fixed effects. PROC MIXED in SAS (22)
was used to compute data in the mixed models.
Since HDL cholesterol levels differed by ethnicity
within the US samples, the ARIC data were divided
into five groups, three white (Forsyth County, North
Carolina; Minnesota; and Washington County, Maryland) and two black (Forsyth County and Jackson,
Mississippi). Results are presented separately for
ARIC blacks and whites.
RESULTS
Table 2 gives age-adjusted mean HDL cholesterol
levels by sex, as well as the 95 percent confidence
interval for the sex difference, in each sampled population. The mixed linear model analysis indicated that
the variation among countries was far greater than the
variation within countries; therefore, we will concentrate here on differences among countries.
By comparison with women in the other countries,
the Chinese and Israeli women had lower mean HDL
cholesterol levels. Mean HDL cholesterol levels were
similar among women in the other sampled populations. There was more variation in mean HDL cholesterol among men, with the US white, Canadian, and
Israeli men having lower mean levels than men from
the other countries.
The sex differences are plotted in figure 1 and vary
markedly by country, from a low of 0.06 mmol/liter in
China to approximately 0.40 mmol/liter in the North
American white populations (Canada, ARIC whites,
and the US LRC studies). The linear model analysis
indicated that the differences among the countries
were statistically significant (p < 0.0001).
A linear model was used to estimate the associations
of smoking, alcohol consumption, body mass index,
heart rate, and hormone use (females only) with HDL
cholesterol. Alcohol consumption data were not available for the Israeli sample, so it was not included in
this model. The interaction tests in the model indicated
that the strength of the associations between smoking,
alcohol, and heart rate and HDL cholesterol differed in
women and men, while the association between body
mass index and HDL cholesterol was similar in men
and women. Table 3 gives the estimated differences
in HDL cholesterol associated with fixed differences
in these variables, as estimated from these crosssectional data. It is noteworthy that cigarette smoking
and alcohol consumption were estimated to have much
stronger effects on HDL cholesterol in women than in
men. The association of HDL cholesterol with use of
exogenous hormones was sizable. In our samples, few
women from Poland, Russia, or China were taking
hormones at the time of the examination; thus, this effect was largely confined to the North American samples.
Figure 2 shows the sex differences in HDL cholesterol after adjustment for age, smoking, alcohol conAm J Epidemiol
Vol. 143, No. 11, 1996
Sex Difference in HDL Cholesterol
1103
TABLE 2. Age-adjusted mean high density llpoprotetn cholesterol levels (mmol/IHer), by sex, and 95% confidence Intervals for
sex differences (women minus men) In sample populations from Canada, China, Israel, Poland, Russia, and the United States,
1972-1989
Man
Women
1 /watbw
LOCaDOfi
No
Mean
No.
Mean
Sax
difference
ep*
oc
1.16
0.40
0.05
0.30-0.51
1.33
1.28
1.30
1 38
1.41
0.06
0.08
0.08
0.00
0.06
0.01
0.02
0.02
0.03
0.03
0.04-0.09
0.04-0.12
0.04-0.11
-0.06-0.07
0.00-0.11
QCq/
pi*
81
1.56
131
1,800
387
642
328
443
.39
36
38
1.39
1.46
1,688
401
690
241
356
Jerusalem, Israel
211
.35
262
1.04
0.31
0.03
0.25-0.37
Poland
Cracow
Warsaw
928
504
424
1.56
1.57
.54
851
426
425
1.44
1.52
1.36
0.11
0.06
0.18
0.02
0.02
0.03
0.08-0.15
0.01-0.10
0.13-0.23
Russia
Moscow
St Petersburg
696
191
505
.49
.41
.52
4,036
2,094
1,942
1.38
1.33
1.42
009
0.08
0.09
0.02
0.03
0.02
0.06-0 12
0.02-0 13
0.05-0.13
United States
ARIC* Study (whites)
Forsyth County, North Carolina
Minnesota
Washington County, Maryland
3,123
876
1,175
1,072
.51
.53
.57
44
2,450
743
917
790
1.12
1.10
1 17
1.08
0.39
0.43
0.40
036
0 01
0.02
0.02
0.02
0.37-0.41
0.39-0.47
0.37-0.43
0.33-0.39
ARIC Study (blacks)
Forsyth County, North Carolina
Jackson, Mississippi
1,288
119
1,169
.53
51
54
736
83
653
1.33
1 23
1 34
0.21
0.28
0.20
0.02
0.02
0.17-0 25
0 16-0.40
0.16-0.24
504
18
40
31
92
54
82
62
64
61
.57
59
.48
.35
.48
536
30
44
34
86
1.16
1.26
1.15
0.97
1 19
93
.65
.38
.72
.47
36
67
62
77
100
1.24
1.23
0.39
0.33
0.33
0.38
0.28
0.69
0.42
040
0.47
034
0.02
0.09
0.07
0.10
0.05
0.09
0.06
0.06
0.06
0.06
0 34-0.43
0.15-0.51
0 20-0 46
0 19-0.58
0 19-0.38
0.52-0.87
0 30-0.54
0.28-0.51
0.35-0.60
0.22-0.46
Toronto, Ontario, Canada
China
Baling—rural
Beijing—urban
Guangzhou—rural
Guangzhou—urban
UpW Research Clinics
Baltimore, Maryland
Cincinnati, Ohio
Houston, Texas
Iowa
La JoBa, California
Minnesota
Oklahoma
Palo Alto, California
Seattle, Washington
1
0.99
1.25
1.13
006
SE, standard error; Cl, confidence interval; ARIC, Atherosclerosis Risk In Communities.
TABLE 3. Estimated differences In high density llpoproteln
(HDL) cholesterol associated wtth differences in other factors,
by sex, in sample populations from Canada, China, Israel,
Poland, Russia, and the United States combined, 1972-1989
HDL cholesterol difference
Variable
Men
Current smoker (yes)
Alcohol intake (20 g/day)
Heart rate (20 beats/minute)
Body mass irtdext (3 units)
Hormone use (yes)
-0.05
0.06
0.01
-0.07
(mmoVnter)
Woman
-0.15*
0.18*
-0.03*
-0.07
0.16
* Difference between men and women was statistically significant at p < 0.05.
t Weight (kgVhelght1 (m2).
Am J Epidemiol
Vol. 143, No. 11, 1996
sumption, heart rate, body mass index, and hormone
use. In these models, the sex X correlate interaction
terms were used so that adjustment took account of the
differences in association noted above. Although variation in the HDL cholesterol sex difference among the
countries was reduced by this adjustment (range: 0.37
mmol/liter in Canada to 0.15 mmol/liter in China), the
difference among countries remained statistically significant (p < 0.0001).
To estimate the sex-specific covariate-adjusted
mean value for each country, we conducted separate
linear analyses for each sex which included the covariates listed above. The covariate adjustment had
very little effect on the variation among countries in
males. The range among the males from table 1 (ig-
1104
Davis etal.
06
Canada
0.6ARIC-W
US-LRC
Israel
0.4-
ARIC-B
t
Poland
Russia
0.1-
China
o-l
FIGURE 1 . Age-adjusted sex difference in high density Iipoprotein (HDL) cholesterol (women minus men) in 23 population samples from six
countries (Canada, China, Israel, Poland, Russia, and the United States), 1972-1989. Honzontal lines are point estimates; solid bars are 95
percent confidence Intervals ARIC-B, Atherosclerosis Risk In Communities (blacks); ARIC-W, Atherosclerosis Risk in Communities (whites);
US-LRC, US Upld Research Clinics.
noring the Israeli sample, since it was not included in
the covariate-adjusted model) was 0.32 mmol/liter
(Poland minus US ARIC whites). However, for females, the range increased from 0.18 mmol/liter (US
LRC minus China; table 1) to 0.29 mmol/liter (Poland
minus China) after covariate adjustment In general,
the covariate adjustment decreased the estimated
means for the North American populations, with the
largest change occurring in the mean value for US
black females (0.9 mmol/liter).
Lastly, if we ignored the heterogeneity among countries, the overall estimate of the sex difference in HDL
cholesterol was 0.27 mmol/liter (95 percent confidence interval 0.25-0.28).
DISCUSSION
This analysis has demonstrated that the mean difference in HDL cholesterol between females and
males varies across a wide variety of populations and
o.e
0.5-
Canada
0.4-
ARIC-B
US-LRC
ARJC-W
I
—•-
|
I
Russla
Poland
China
0.1-
FIGURE 2. Covariate-adjusted sex difference in high density Iipoprotein (HDL) cholesterol (women minus men) in 23 population samples
from six countries (Canada, China, Israel, Poland, Russia, and the United States), 1972-1989. Data were adjusted for age, body mass index,
smoking status, alcohol intake, centered heart rate, and hormone use. Horizontal lines are point estimates; solid bars are 95 percent
confidence intervals ARIC-B, Atherosclerosis Risk in Communrties (blacks); ARIC-W, Atherosclerosis Risk in Communities (whites); US-LRC,
US Upid Research Clinics.
Am J Epidemiol
Vol. 143, No. 11, 1996
Sex Difference in HDL Cholesterol
cultures. Moreover, the magnitude of this difference is
greatly influenced by environmental factors such as
cigarette smoking, alcohol consumption, body mass,
and exogenous hormone use. The association of alcohol consumption and smoking with HDL cholesterol
level was found to be stronger in women than in men.
Although differences in these factors explained some
of the variation among populations, considerable variation in the estimated sex difference remained after
adjustment.
These results indicate that alcohol consumption, cigarette smoking, obesity, heart rate, and exogenous
hormone use clearly influence the magnitude of the
difference in HDL cholesterol between men and
women. Thus, not all of the observed sex difference in
HDL cholesterol can be attributed to intrinsic physiologic or endogenous hormonal differences.
There are at least four possible explanations for the
fact that the correlates of HDL cholesterol only partially explained the differences among the countries.
First, genetic differences among the populations may
affect the magnitude of the sex difference. Second,
there may be other environmental correlates of HDL
cholesterol not measured in these studies which would
explain the differences. Third, variability in the measurement of the correlates could lead to underadjustment of the associations. If a covariate has reliability R
(:£l), adjustment removes 1007? percent of the initial
confounding effect (23). For example, the reliability of
heart rate has been found to be 0.60 in INTERSALT
(24). Thus, the covariate adjustment is approximately
60 percent complete. Since all of the covariates were
subject to variability, adjustment was likely not to
have completely adjusted for possible confounding.
Fourth, although laboratory variation is a possible
explanation, it is not likely to explain the sex difference, since our comparisons were done within-center.
Laboratory variation could be a possible explanation
for the differences in absolute levels of HDL cholesterol seen across populations. We suspect that all four
of these possible explanations operated to cause the
differences among countries, but we have no empirical
method for estimating the magnitude of their relative
effects.
It has been hypothesized that the sex difference in
HDL cholesterol may partially explain the excess coronary heart disease morbidity and mortality seen in
males. It will be interesting to compare the differences
in morbidity and mortality among the countries and
populations sampled and to correlate sex differences in
disease incidence with sex differences in factors such
as HDL cholesterol. Preliminary data from Chinese
national health statistics have shown that the male:
female ratio of age-standardized coronary disease
Am J Epidemiol
Vol. 143, No. 11, 1996
1105
mortality rates is distinctly lower (1.4: 1) in Beijing
and Guangzhou than in the United States and other
countries with higher overall coronary disease mortality (25). Morbidity and mortality data are currently
being collected in several of these populations.
ACKNOWLEDGMENTS
This work was supported in part by the US National
Heart, Lung, and Blood Institute under contract N01-HV08112 with the University of North Carolina at Chapel Hill.
REFERENCES
1. Gordon DJ, Probstfield JL, Garrison RJ, et al High-density
hpoprotein cholesterol and cardiovascular disease: four prospective American studies. Circulation 1989;79:8-15.
2. Castelli WP. Epidemiology of coronary heart disease: The
Framingham Study. Am J Med 1984;76:4-12.
3. Gordon T, Castelli WP, Hjortland MC, et al. High density
lipoprotein as a protective factor against coronary heart disease The Framingham Study Am J Med 1977;62:707-714.
4. Castelli WP, Garrison RJ, Wilson PW, et al. Incidence of
coronary heart disease and lipoprotein cholesterol levels: The
Framingham Study JAMA 1986;2562835-8.
5. Eaker ED, Thorn T, Castelli WP Coronary heart disease in
women. Soz Praventivmed 1988,33:10-16.
6 McNamara JR, Campos H, Ordovas JM, et al. Effect of
gender, age, and lipid status on low density lipoprotein subfraction distribution, results from the Framingham Offspring
Study. Arteriosclerosis 1987;7:483-9O.
7. Lerner DJ, Kannel WB. Patterns of coronary heart disease
morbidity and mortality in the sexes: a 26-year follow-up of
the Framingham population. Am Heart J 1986; 111.383—90.
8. Crouse JR. Gender, lipoproteins, diet, and cardiovascular risk:
sauce for the goose may not be sauce for the gander. Lancet
1989:1:318-20.
9. Kinnunen PKJ, Virtonen JA, Vianio P. Lipoprotein lipase and
hepatic endothelial lipase: their roles in plasma hpoprotein
metabolism. Atherosclerotic Rev 1983;11:65-105.
10. The AR1C Investigators The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. Am J Epidemiology 1989;129:687-702.
11. WHO MONICA Project MONICA manual, revised edition.
Geneva, Switzerland. Cardiovascular Diseases Unit, World
Health Organization, 1990.
12. Rywik S, Sznajd J, Przestalska-Malkin H, et al. Monitoring
trends in cardiovascular disease incidence and mortality and
their determinants: "Pol-Monica" longitudinal study. I. Methodological principles of the study. (In Polish with English
abstract). Przeg Lek 1985;42:250-5.
13. WHO MONICA Project Principal Investigators. The World
Health Organization MONICA Project (Monitoring Trends
and Determinants in Cardiovascular Disease): a major international collaboration. J Clin Epidemiol 1988;41:105-14.
14. People's Republic of China-United States Cardiovascular and
Cardiopulmonary Epidemiology Research Group. An epidemiological study of cardiovascular and cardiopulmonary disease risk factors in four populations in the People's Republic
of China. Baseline report from the P.R.C.-U.S.A. Collaborative Study. Circulation 1992,85:1083-96
15. Lipid Research Clinics Program. Protocol of the Lipid Research Clinics Program Prevalence Study. Chapel Hill, NC:
LRC Central Patient Registry and Coordinating Center, De-
1106
16.
17.
18.
19.
Davis et al.
partment of Biostatistics, University of North Carolina at
Chapel HiU, 1974.
Levy RI, Klimov AN, Williams OD. Epidemiology studies:
overview. Atherosclerosis Rev 1988; 17:93-102.
Deed AD, Konstantinov W , Shestov DB. Population descriptions and methodology for US-USSR Collaboration in Area 1
(Pathogenesis of Atherosclerosis): Second Prevalence Study
details for Moscow and Leningrad Lipid Research Clinics.
Atherosclerosis Rev 1988:17:103-9.
US-USSR Steering Committee for Problem Area 1: The
Pathogenesis of Atherosclerosis. Collaborative US-USSR
study on the prevalence of dyslipoproteinemias and ischemic
heart disease in American and Soviet populations. Am J
Cardiol 1977;40:260-8. (Also, in Russian. Ter Arkh 1977;4:
26-37).
Laboratory Methods Committee of the Lipid Research Clinics
Program. Cholesterol and triglyceride concentrations in
serum/plasma pairs. Chn Chem 1977;23:60-3.
20. Albers JJ, Wamick GR, Johnson N, et al. Quality control of
plasma high-density lipoprotein cholesterol measurement
methods. The Lipid Research Clinics Program Prevalence
Study. Circulation 1980;62:IV9-18.
21. Warnick GR, Benderson JM, Albers JJ. Dextran sulfateMg2+ precipitation procedure for quantitation of highdensity-lipoprotem cholesterol. Clin Chem 1982;28:1379-88.
22. SAS Institute, Inc. SAS/STAT software: changes and enhancements, release 6.07. Cary, NC: SAS Institute, Inc, 1992:
287-366. (SAS technical report P-229).
23. Cochran WG. Planning and analysis of observational studies.
New York, NY. John Wiley and Sons, Inc, Publishers, 1983.
24. Dyer AR, Shipley M, Elliott P. Urinary electrolyte excretion
in 24 hours and blood pressure in the INTERSALT Study. I.
Estimates of reliability. Am J Epidemiol 1994;139927-39.
25. Tao S-C, Huang ZD, Wu XG, et al. CHD and its risk factors
in the People's Republic of China. Int J Epidemiol 1989;
18(suppl l):S159-63.
Am J Epidemiol
Vol. 143, No. 11, 1996