American Journal of Epidemiology
Copyright © 1998 by The Johns Hopkins University School of Hygiene and Public Health
All rights reserved
Vol. 148, No. 8
Printed in U.S.A.
African American-White Differences in Lipids, Lipoproteins, and
Apolipoproteins, by Educational Attainment, among Middle-aged Adults:
The Atherosclerosis Risk in Communities Study
Patricia A. Metcalf,1 A. Richey Sharrett,2 Aaron R. Folsom,3 Bruce B. Duncan,4 Wolfgang Patsch,5
Richard G. Hutchinson,6 Moyses Szklo,7 C. E. Davis,1 and H. A. Tyroler4
Measures of socioeconomic status have been shown to be related positively to levels of high density
lipoprotein (HDL) cholesterol in white men and women and negatively in African American men. However, there
is little information regarding the association between educational attainment and HDL fractions or apolipoproteins. The authors examined these associations in 9,407 white and 2,664 African American men and
women aged 45-64 years who participated in the Atherosclerosis Risk in Communities Study baseline survey,
and they found racial differences. A positive association for HDL cholesterol, its fractions HDI-a and HDL3
cholesterol, and its associated apolipoprotein A-l was found in white men and white women, but a negative
association was found in African American men, and there was no association in African American women. In
whites, there was also an inverse association of low density lipoprotein (LDL) cholesterol and apolipoprotein
B with educational attainment. With the exception of African American men, advanced education was
associated with a more favorable cardiovascular lipid profile, which was strongest in white women. Racial
differences in total cholesterol (women only), plasma triglycerides, LDL cholesterol, apolipoprotein B (women
only), HDL cholesterol, HDl_2 and HDLg cholesterol, and apolipoprotein A-l were reduced at higher levels of
educational attainment. Apart from triglycerides in men and HDL3 cholesterol in women, these African
American-white lipid differences associated with educational attainment remained statistically significant after
multivariable adjustment for lifestyle factors. Lipoprotein(a) showed no association with educational attainment. These findings confirm African American-white differences in lipids, lipoproteins, and apolipoproteins
across levels of educational attainment that were not explained by conventional nondietary lifestyle variables.
Understanding these differences associated with educational attainment will assist in identifying measures
aimed at prevention of cardiovascular disease. Am J Epidemiol 1998; 148:750-60.
apoliproteins; education; lipids; lipoproteins
The associations between race and high density lipoprotein (HDL) cholesterol levels and other factors
have received much attention; although African Amer-
icans have a greater incidence of hypertension, the
rates of coronary heart disease among African American men are no higher than those among white men
but are equal to or higher in African American women
as compared with white women (1-3). Therefore, it
has been postulated that protective factors may exist in
African American men to account for their lower-thanexpected mortality rate. One factor may be HDL cholesterol, which is approximately 20 percent (7-11 mg/
dl) higher in African American men than in white men
(4-10). Similarly, plasma triglyceride levels are consistently lower in African Americans than in whites
(8-11). However, a potentially adverse factor among
African Americans is their twofold higher level of
lipoprotein(a) as compared with the level in whites
(12), although the pathogenicity of lipoprotein(a) may
differ in African Americans.
Low socioeconomic status is an independent risk
factor for coronary heart disease (13) and is characteristic of many African Americans (14). Socioeco-
Received for publication February 28, 1997, and accepted for
publication December 4, 1997.
Abbreviations: ARIC, Atherosclerosis Risk in Communities; HDL,
high density lipoprotein; LDL, low density lipoprotein.
1
Collaborative Studies Coordinating Center, University of North
Carolina, Chapel Hill, NC.
2
ARIC Project Office, Bethesda, MD.
3
Division of Epidemiology, School of Public Health, University of
Minnesota, Minneapolis, MN.
4
Department of Epidemiology, School of Public Health, University
of North Carolina, Chapel Hill, NC.
5
Atherosclerosis Clinical Laboratory, Methodist Hospital,
Houston, TX.
6
Department of Medicine, University of Mississippi Medical
Center, Jackson, MS.
7
School of Hygiene and Public Health, The Johns Hopkins
University, Baltimore, MD.
Reprint requests to Dr. C. E. Davis, Collaborative Studies Coordinating Center, Suite 203, 137 East Franklin Street, University of
North Carolina, Chapel Hill, NC 27514.
750
Racial Differences in Lipid Levels by Educational Attainment
nomic status is an important predictor of overall health
status and the quality of health care received (4). The
association between socioeconomic status and HDL
cholesterol levels differs by race and is reported to be
positive among whites and African American women
but negative among African American men (6, 7, 10).
We examined the associations of educational attainment with levels of plasma total cholesterol, triglycerides, low density lipoprotein (LDL) cholesterol and
its associated apolipoprotein B, HDL cholesterol,
HDL2 and HDL3 cholesterol, and the major HDL
cholesterol structural apolipoprotein A-I and lipoprotein(a) in participants of the baseline examination of
the Atherosclerosis Risk in Communities Study
(ARIC). These associations were studied to determine
whether there had been any changes in patterns observed previously and to extend these observations to
lipid fractions not examined previously. Because lipid
levels are influenced by several health behaviors, such
as physical activity, obesity, cigarette smoking, and
alcohol intake (15), and by hormone therapy in women
(16), we also examined the foregoing associations
after adjusting for these behavioral factors.
MATERIALS AND METHODS
The study population comprised 15,792 men and
women aged 45-64 years who participated in the
ARIC Study. Participants were recruited between
1987 and 1989 from four communities: Forsyth
County, North Carolina; Jackson, Mississippi; the
northwestern suburbs of Minneapolis, Minnesota; and
Washington County, Maryland. The response rates
were 66, 46, 67, and 65 percent, respectively. All participants from Jackson and 14 percent of the participants
from Forsyth County were African American; almost all
of the others were white. Information on the complete
study design, sampling strategy, and examination techniques has been published elsewhere (17).
Participants were excluded from analysis for the
following reasons: 53 were not African American or
white, 112 were not within the stated age range, 1,151
had not fasted for 12 hours, 1,227 had diabetes mellitus, 435 were on lipid-lowering medications, 16 had
missing information on education, and 727 had missing information on yearly family income. Therefore,
12,071 participants remained.
Lipid measurements were determined on plasma
from blood collected after a 12-hour fast into tubes
containing ethylenediaminetetraacetic acid. Aliquots
were stored at — 70°C, shipped on dry ice to the ARIC
Central Lipid Laboratory at weekly intervals, and analyzed within approximately 6 weeks of receipt.
Concentrations of cholesterol and triglycerides were
determined on a Cobas-Bio centrifugal analyzer
Am J Epidemiol
Vol. 148, No. 8, 1998
751
(Hoffman-La Roche, Basel, Switzerland) using enzymatic procedures (18, 19) and commercial reagents
(Boehringer Mannheim Biochemical, Indianapolis,
Indiana). HDL cholesterol was measured after precipitation with magnesium chloride and dextran sulfate of
lipoproteins containing apolipoprotein B. HDL3 cholesterol was measured after further precipitation with
magnesium chloride and dextran sulfate (20). HDL2
cholesterol was calculated as the difference between
HDL cholesterol and HDL3 cholesterol. LDL cholesterol was calculated according to the Friedewald (21)
formula in participants with plasma triglyceride levels
of <400 mg/dl.
Apolipoprotein A-I and apolipoprotein B levels
were determined by radioimmunoassay (22, 23). Lipoprotein(a) concentrations were measured by an
enzyme-linked immunoassay (24) and included apolipoprotein(a) and associated apolipoprotein B.
Plasma pools of cholesterol, triglycerides, HDL
cholesterol, and HDL3 cholesterol from the Centers
for Disease Control and Prevention were used for
internal quality control and calibrator materials following Lipid Research Clinics Program protocols
(25). Coefficients of laboratory variation, in percentages, were as follows: triglycerides, 2.7; total cholesterol, 2.5; HDL cholesterol, 3.7; HDL3 cholesterol,
8.2; apolipoprotein A-I, 9.0; LDL cholesterol, 5.2;
apolipoprotein B, 9.0; and lipoprotein(a), 9.0. External
control was defined as successful participation in the
Lipid Standardization Program of the Centers for Disease Control and Prevention. The overall variability of
measurement (i.e., due to storage, shipping, sample
processing, transcription, and analysis) was determined by using samples taken from the blood collection tubes of a subset of participants. The samples
were stored at each center for another week and then
sent to the laboratory in a blinded fashion. Blinded
replicate coefficients of variation for triglycerides, total cholesterol, HDL cholesterol, LDL cholesterol,
apolipoprotein A-I, and lipoprotein(a) were 6.7, 3.0,
4.4, 5.3, 11.0, and 16.9 percent, respectively.
Race, years of schooling, yearly family income,
pack-years of cigarette smoking, current alcohol intake, and hormone therapy were assessed by questionnaire. In these participants, alcohol intake was previously shown to be associated with higher HDL
cholesterol levels in all race-sex groups after adjustment for body mass index, physical activity, smoking,
and age (15). Hormone therapy was classified as those
women currently taking estrogen or estrogen and progestin medications. Years of schooling was categorized as basic (not a high school graduate), intermediate (a high school graduate and/or vocational
schooling), or advanced (at least 1 year of college).
752
Metcalf et al.
Yearly family income was grouped as follows:
<$8,000, $8,000-<$16,000, $16,000-<$35,000,
and ^$35,000. Participation in leisure-time sports activities was assessed using a standardized questionnaire (26); scores ranged from one (low) to five (high).
The sports activity score is a combined index of intensity and duration/frequency of exercise, which has
been shown to be associated with higher HDL cholesterol levels in all race-sex groups after adjustment for
body mass index, alcohol intake, smoking, and age.
However, in this study the association was not statistically significant in African Americans (15).
Body mass index was calculated as weight (kg)/
height (m) 2 . To assess whether there were racial differences in lipid concentrations across various levels
of educational attainment, analyses of covariance were
used, with lipid levels as the dependent variable and
incorporation of categorical variables for race and
education and its corresponding interaction term. To
provide some measure of the strength of the associations with levels of education in each race and sex
group, we calculated partial correlation coefficients
between the ordinal spectrum of education or income
and lipid levels. Values for/? were calculated using the
chi-square test for categorical variables and Student's
t test for continuous variables. Tests of differences in
mean lipid levels across educational strata were calculated using multivariate analysis of covariance and
including the relevant interaction term after adjustment for the variables shown in tables 4 and 5.
The African American men and women were slightly
younger than the white men and women. Both level of
educational attainment and yearly family income were
higher among whites than among African Americans.
The level of participation in leisure sports and the
number of pack-years of cigarette smoking were also
higher among whites. In women, mean body mass
index was substantially higher and mean alcohol intake was much lower among African Americans than
among whites. More white women than African
American women reported using hormone therapy.
Age-adjusted mean levels of plasma lipids, lipoproteins, and apolipoproteins are shown in table 2. The
most striking racial differences were found in levels of
plasma lipoprotein(a), which were approximately
twice as high in African Americans as in whites.
Plasma triglyceride levels were substantially lower in
African Americans than in whites. Mean levels of
HDL cholesterol, its component fractions HDL2 and
HDL3 cholesterol, and its associated apolipoprotein
A-I were 11-39 percent higher in African American
men than in white men but only 0-5 percent higher in
women of both races. Racial differences in LDL cholesterol and apolipoprotein B levels were small.
In table 3, age, body mass index, alcohol intake,
leisure sports activity score, pack-years of cigarette
smoking, and hormone therapy are shown according to
educational attainment. The number of mean packyears of cigarette smoking was 38-45 percent lower
among participants with an advanced education as
compared with those with a basic education. In whites,
mean alcohol intake was 38 percent higher in men and
more than twice as high in women with an advanced
education as compared with those with a basic educa-
RESULTS
Table 1 summarizes the characteristics of the participants examined during the ARIC baseline survey.
TABLE 1.
Selected characteristics of men and women aged 45-64 years examined during the ARIC* baseline survey, 1987-1989
Women
Men
White
(n = 4,409)
African American
(n= 1.029)
Mean
Age (years)
53.6
Body mass index (kg/m2)
27.1
Cigarette smoking (packyears)
388.3
Alcohol intake (g/day)
10.7
Sports activity score
2.3
Educationt
Basic
Intermediate
Advanced
Yearly family income
<S8,000
S8,000-<$16,000
$16,000-<$35,000
a$35,000
Hormone therapy
SE*
%
Mean
SE
%
African American
(n= 1,635)
P
value
Mean
SE
%
White
(n = 4,998)
Mean
SE
%
P
value
0.18
0.13
54.6
27.2
0.08
0.06
<0.001
0.485
52.8
30.4
0.14
0.14
53.7
26.3
0.08
0.08
< 0.001
< 0.001
15.63
0.58
0.03
467.7
10.5
2.7
7.43
0.28
0.01
<0.001 141.6
1.7
0.758
<0.001
2.1
8.27
0.18
0.02
227.2
3.6
2.4
4.69
0.11
0.01
< 0.001
< 0.001
<0.001
43.1
26.0
30.9
17.0
39.3
43.7
18.6
22.2
35.1
24.2
1.6
6.3
31.6
60.5
<0.001
<0.001
36.8
30.5
32.7
14.7
51.1
34.2
30.1
25.4
29.7
15.2
15.2
3.7
10.6
35.2
50.5
21.0
<0.001
<0.001
< 0.001
* ARIC, Atherosclerosis Risk in Communities; SE, standard error.
t Basic, < high school graduate; intermediate, high school graduate and/or vocational schooling; advanced, >1 year of college.
Am J Epidemiol
Vol. 148, No. 8, 1998
Racial Differences in Lipid Levels by Educational Attainment
753
TABLE 2. Age-adjusted mean plasma lipid, lipoprotein, and apolipoprotein levels (mg/dl) of men and women aged 45-64 years
examined during the ARICf baseline survey, 1987-1989
African American (A)
( n = 1 ,029)
Total cholesterol
Triglyce rides
LDLf cholesterol
Apolipoprotein B
HDLt cholesterol
HDL2 cholesterol
HDL3 cholesterol
Apolipoprotein A-l
Upoprotein(a)$
Men
White,(W)
Women
(n = 4,409)
Mean
SEt
Mean
SE
211.0
112.8
137.6
92.6
51.5
14.1
37.4
132.8
147.6
1.22
2.67
. 1.15
0.86
0.42
0.20
0.31
0.82
2.96
210.5
141.5
139.8
95.2
43.1
10.2
32.9
119.3
75.8
0.59
1.29
0.56
0.42
0.20
0.10
0.15
0.39
1.42
African American
Difference:
(A-W)/
W (%)
Mean
0.2
-20.3***
-1.6
-2.7**
19.4***
38.6***
13.5***
11.4***
94.6***
216.4
102.5
136.9
90.5
59.2
18.2
41.0
145.5
168.4
(n= 1,635)
SE
1.02
1.63
< 0.98
0.72
0.42
0.23
0.26
0.78
2.60
White
(n = 4,998)
Mean
SE
Difference:
(A-W)/
W(%)
216.5
121.0
134.1
90.5
58.3
17.3
41.0
143.0
85.3
0.58
0.93
0.56
0.40
0.24
0.13
0.15
0.45
1.48
-15.2***
2.1*
0.0
1.5
5.1***
0.0
1.7**
97.3***
0.1
* p < 0.05; ** p < 0.01; *** p < 0.001.
f ARIC, Atherosclerosis Risk in Communities; SE, standard error; LDL, low density lipoprotein; HDL, high density lipoprotein.
i In mg/liter.
tion. In contrast, mean alcohol intake was 36 percent
lower in African American women with an advanced
education as compared with those with a basic education, but little difference in mean alcohol intake by
level of education was observed in African American
men. Mean body mass indexes were substantially
lower in both African American and white women
with an advanced education as compared with those
with a basic education, and use of hormone therapy
was associated with higher educational attainment in
both racial groups. There was little difference in mean
body mass index across levels of education among
white men, but mean body mass index rose with level
of education among African American men.
Participation in leisure sports activities also varied
by level of education and was more marked among
whites. This score was 20 percent higher for white
men, 12 percent higher for white women, 14 percent
higher for African American men, and 5 percent
higher for African American women with an advanced
education as compared with those with a basic education (table 3). Lower educational attainment was also
associated with higher age. Similar but less striking
associations were observed with yearly family income
(data not shown).
Age-adjusted mean plasma lipid, lipoprotein, and
apolipoprotein levels showed similar patterns across
levels of educational attainment (table 4) and yearly
family income (data not shown). In contrast, lipoprotein(a) showed no consistent associations with education (or income). Whites with higher levels of education fairly consistently had a slightly more favorable
lipid profile, with lower levels of triglycerides, LDL
cholesterol, and apolipoprotein B and higher levels of
HDL cholesterol, its fractions HDL2 and HDL3 cholesterol, and apolipoprotein A-I. These associations
Am J Epidemiol
Vol. 148, No. 8, 1998
were more marked among white women than among
white men.
Although similar lipid associations were observed in
African American women, they were much weaker
than those in white men and women. In contrast,
African American men with a higher level of education had a less favorable lipid profile with significantly
lower levels of HDL cholesterol and its fractions
HDL2 and HDL3 cholesterol. Overall, associations
were stronger between lipids and educational attainment than between lipids and yearly family income in
whites but slightly weaker, although not statistically
significant, in African Americans (data not shown).
Partial correlation coefficients across educational
strata ranged from —0.09 for HDL cholesterol in African American men to 0.07 for HDL cholesterol in
white men and from —0.12 for triglycerides in white
women to 0.15 for HDL and HDL2 cholesterol in
white women. Similarly, partial correlations with income ranged from -0.15 for HDL2 cholesterol in
African American men to 0.04 for HDL2 cholesterol
and apolipoprotein A-I in white men and from -0.08
for triglycerides in white women to 0.06 for HDL cholesterol in African American women (data not shown).
The associations between educational attainment
and mean age-adjusted lipid differences in African
American and white men and women are shown in
table 5 (model 1) together with significance levels for
race X education interaction terms. Significant interaction terms indicate that the associations of lipids
with educational attainment were different in African
Americans and whites. Because of the influences of
hormone therapy in women and of body mass index,
smoking, alcohol intake, and sports activity on plasma
lipids and lipoproteins, these associations were also
examined after adjusting for these variables (model 2).
TABLE 3.
Selected characteristics,* according to educational attainment,t of men and women aged 45-64 years examined during the ARIC$ baseline survey, 1987-1989
Women
Men
Basic
Mean
Number
African American
White
Age (years)
African American
White
443
Mean
SE
SE
P
value§
Mean
Intermediate
SE
Mean
602
733
318
1,927
268
1,73"I
751
Mean
Basic
SE
Advanced
Mean
SE
P
value§
537
1,711
496
i,554
0.27
0.20
52.8
54.3
0.34
0.13
52.0
54.0
0.31
0.13
<0.001
<0.001
54.4
55.6
0.22
0.21
52.5
53.7
0.24
0.11
51.4
53.0
0.24
0.13
<0.001
<0.001
Body mass index (kg/m2)
African American
White
26.7
27.3
0.22
0.14
27.3
27.3
0.28
0.09
27.6
27.2
0.26
0.09
0.021
0.579
31.5
28.0
0.26
0.20
30.4
26.4
0.29
0.10
29.0
25.5
0.28
0.13
<0.001
<0.001
Alcohol intake (g/day)
African American
White
10.1
8.2
1.07
0.64
12.7
10.5
1.37
0.43
9.7
11.3
1.25
0.40
0.209
<0.001
2.2
1.8
0.25
0.29
1.2
3.5
0.27
0.15
1.4
4.6
0.26
0.19
0.015
<0.001
2.1
2.4
0.04
0.03
2.3
2.6
0.05
0.02
2.5
2.9
0.04
0.02
<0.001
<0.001
2.1
2.3
0.03
0.03
2.1
2.4
0.03
0.02
2.2
2.6
0.03
0.02
0.048
<0.001
488.8
632.4
22.67
17.87
369.4
502.2
267.0
373.1
26.45
11.11
<0.001
<0.001
187.3
307.8
11.30
12.80
125.0
227.5
12.51
6.84
105.8
192.3
11.91
8.35
<0.001
<0.001
Cigarette smoking (pack-years)
African American
White
Vn J Epi demiol
Vo
CD
CO
CO
SEt
Advanced
55.4
56.9
Sports activity score
African American
White
p
Intermediate
Hormone therapy (%)
African American
White
29.11
11.72
10.7
16.0
* Values are means unless otherwise noted,
t Basic, < high school graduate; intermediate, high school graduate and/or vocational schooling; advanced, >1 year of college.
i ARIC, Atherosclerosis Risk in Communities; SE, standard
s
error,
§ p values for test of no differences between mean levels across socioeconomic strata.
16.3
21.2
20.3
24.2
<0.001
<0.001
I"
5'
TABLE 4. Age-adjusted mean plasma lipid, lipoprotein, and apolipoprotein levels (mg/dl) and age-adjusted partial correlation coefficients (r), according to educational
attainment,t of men and women aged 45-64 years examined during the ARICj baseline survey, 1987-1989
Men
Basic
00
CO
CO
00
Women
Advanced
Basic
r
Mean
SEt
Mean
SE
Mean
SE
Total cholesterol
African American
White
209.5
213.5
1.86
1.45
214.3
210.6
2.40
0.94
210.3
209.2
2.21
0.89
Triglycerides
African American
White
104.4
143.4
4.06
3.16
124.6
142.9
5.23
2.05
112.2
139.4
LDI4 cholesterol
African American
White
135.9
143.3
1.75
1.37
139.2
140.7
2.27
0.89
Apolipoprotein B
African American
White
92.2
97.5
1.31
1.02
93.9
95.7
HDI4 cholesterol
African American
White
53.1
42.3
0.64
0.50
HDL2 cholesterol
African American
White
15.0
9.8
HDL, cholesterol
African American
White
Intermediate
Advanced
Mean
SE*
Mean
SE
Mean
SE
0.01
-0.04*
216.9
220.9
1.67
1.53
215.5
218.4
1.84
0.86
216.4
211.9
1.78
0.99
-0.01
-0.08***
4.81
1.95
0.04
-0.02
101.5
135.0
2.68
2.44
101.2
123.9
2.95
1.30
104.8
110.7
2.85
1.59
0.01
-0.12***
138.6
137.7
2.07
0.84
0.03
-0.06***
137.9
140.2
1.67
1.47
136.1
136.1
1.78
0.78
136.4
128.7
1.72
0.96
-0.02
-0.10***
1.69
0.66
92.1
93.9
1.55
0.63
0.00
-0.05**
90.8
95.1
1.15
1.05
90.4
91.8
1.26
0.56
90.1
86.5
1.22
0.68
-0.01
-0.11***
51.5
42.3
0.82
0.32
49.2
44.2
0.76
0.31
-0.09**
0.07***
58.7
53.7
0.69
0.63
59.3
57.6
0.76
0.33
59.8
61.3
0.74
0.41
0.02
0.15***
0.31
0.24
13.7
9.9
0.39
0.15
13.3
10.6
0.36
0.15
-0.08**
0.06***
17.9
14.8
0.38
0.35
18.4
16.8
0.42
0.18
18.3
19.0
0.41
0.23
0.02
015***
38.1
32.5
0.47
0.37
37.8
32.5
0.61
0.24
35.9
33.6
0.56
0.23
-0.08*
0.05***
40.8
38.8
0.43
0.39
40.8
40.8
0.47
0.21
41.5
42.3
0.46
0.25
0.01
0.11***
Apolipoprotein A-l
African American
White
134.5
118.4
1.24
0.97
134.0
118.2
1.60
0.63
129.6
120.6
1.47
0.60
-0.06*
0.04**
145.1
135.0
1.29
1.17
144.1
142.8
1.42
0.62
147.3
146.7
1.37
0.76
0.01
LJpoprotein(a)§
African American
White
150.1
75.6
4.53
3.48
143.3
75.9
5.76
2.28
147.9
75.9
5.36
2.15
0.00
0.00
170.7
82.0
4.27
3.89
168.0
84.9
4.73
2.07
166.1
87.4
4.57
2.53
-O.01
0.02
CO
p
Intermediate
Q
w.
D
w
i
o
CD
CO
•g
ai.
CO
* p < 0.05; ** p < 0.01; ••* p < 0.001. p values for test of differences across educational strata.
t Basic, < high school graduate; intermediate, high school graduate and/or vocational schooling; advanced, £1 year of college,
i ARIC, Atherosclerosis Risk in Communities; SE, standard error; LDL, low density lipoprotein; HDL, high density lipoprotein.
§ In mg/Iiter.
0.12***
2
o
a
o'
i.
i
D
CD
756
TABLE 5. Mean African American (A)-white (W) lipid differences (mg/dl), according to educational attainment, * in men and women aged 45-64 years examined during
the ARICf baseline survey, 1987-1989
Women
Men
Model 1 *
A-W
Total cholesterol
Basic
Intermediate
Advanced
Triglycerides
Basic
Intermediate
Advanced
5;
1
rf
p
CO
CO
CO
00
(A-W)/
P
valueH
A-W
(A-W)/
CD
Model 1 *
Model 2§
P
valueH
A-W
(A-W)/
1
Model 2§
P
valued
A-W
(A-W)/
W (%)
0)
P
valued
-4.0
3.7
1.1
-1.9
1.8
0.5
0.0726
-4.0
3.5
2.2
-1.9
1.6
1.1
0.0539
-4.0
-2.8
4.4
-1.8
-1.3
2.1
0.0083
-6.5
-3.1
2.7
-2.9
-1.4
1.3
0.0055
-39.1
-15.4
-27.2
-27.3
-10.8
-19.5
0.0077
-37.1
-16.8
-30.8
-26.4
-11.8
-21.7
0.0599
-33.5
-22.7
-6.0
-24.8
-18.3
-5.4
0.0001
-415
-31.7
-15.3
-31.6
-25.2
-13.1
0.0001
LDLf cholesterol
Basic
Intermediate
Advanced
-7.5
-1.5
0.9
-5.2
-1.0
0.7 '
0.0227
-6.5
-1.4
1.6
-4.6
-1.0
1.2
0.0141
-2.3
0.1
7.7
-1.7
0.1
6.0
0.0013
-4.6
-2.7
3.1
-4.1
-2.0
2.3
0.0069
Apolipoproteln B
Basic
Intermediate
Advanced
-5.4
-1.8
-1.8
-5.5
-1.8
-1.9
0.2170
-4.7
-1.8
-1.5
-4.9
-1.9
-1.6
0.3100
-4.4
-1.4
3.6
-4.6
-1.5
4.1
0.0005
-6.7
-3.2
0.8
-7.2
-3.5
0.9
0.0018
HDLt cholesterol
Basic
Intermediate
Advanced
10.9
9.2
5.0
25.7
21.7
11.3
0.0001
9.5
8.9
6.1
22.0
21.0
14.0
0.0034
5.0
1.6
-1.5
9.4
2.8
-2.4
0.0001
7.6
5.9
3.2
13.5
10.4
5.5
0.0042
HDL2 cholesterol
Basic
Intermediate
Advanced
5.2
3.8
2.6
53.1
38.6
24.5
0.0001
4.7
3.8
3.0
47.0
38.6
28.3
0.0035
3.0
1.6
-0.7
20.4
9.6
-3.5
0.0001
4.2
3.5
1.4
26.3
21.3
8.1
0.0002
HDL3 cholesterol
Basic
Intermediate
Advanced
5.6
5.4
2.4
17.4
16.6
7.2
0.0002
4.8
5.1
3.2
14.4
15.7
9.6
0.0451
2.0
0.0
-0.8
5.1
-0.2
-1.9
0.0012
3.4
2.4
1.8
8.4
6.0
4.4
0.3282
Apolipoproteln A-l
Basic
Intermediate
Advanced
16.1
15.7
9.0
13.6
13.3
7.5
0.0022
14.1
15.5
10.4
11.7
13.1
8.7
0.0582
10.1
1.3
0.7
7.5
0.9
0.5
0.0001
13.6
6.8
7.6
9.9
4.8
5.4
0.0256
Upoprotein(a)#
Basic
Intermediate
Advanced
74.4
67.4
72.0
98.4
88.8
94.9
0.6996
75.6
69.0
72.6
96.1
91.1
96.4
0.9416
88.7
83.1
78.7
108.2
97.9
90.1
0.4369
84.9
81.6
78.0
104.7
96.2
88.2
0.6864
* Basic, < high school graduate; Intermediate, high school graduate and/or vocational schooling; advanced, 2:1 year of college,
t ARIC, Atherosclerosis Risk in Communities; LOL, low density lipoprotein; HDL, high density lipoprotein.
t Adjusted for age.
§ Adjusted for age, body mass Index, pack-years of cigarette smoking, alcohol intake, and sports activity score in men and women and for hormone therapy In women only.
H For race x education Interaction.
# In mg/liter.
Racial Differences in Lipid Levels by Educational Attainment
Level of education had little influence on African
American-white differences in lipoprotein(a) or total
cholesterol (in men only). For other lipids, the levels in
African Americans and whites were more similar at
higher levels of education and yearly family income
than at lower levels.
In men and women, statistically significant ageadjusted race X education interaction terms (model 1)
were found for HDL cholesterol, its fractions HDL2
and HDL3 cholesterol, and its associated apolipoprotein A-I, reflecting the reduced differences in lipid
levels between African Americans and whites at
higher levels of education and yearly family income.
After further adjustment for body mass index, smoking, alcohol intake, sports activity score, and hormone
therapy (model 2), the race X education interaction
terms for apolipoprotein A-I in men and HDL3 cholesterol in women were no longer statistically significant. Racial differences in triglycerides were also reduced at higher levels of educational attainment and
income but were no longer significant in men after
multivariable adjustment. Differences in levels of
LDL cholesterol in men and women, and total cholesterol and apolipoprotein B in women, were reduced
significantly at higher levels of educational attainment, and they remained significant after multivariable adjustment.
Similar age-adjusted models built for yearly family
income showed generally weaker associations than
those for educational attainment that were no longer
significant after multivariable adjustment (data not
shown). Because only 12 percent of the African Americans were from Forsyth County, the differences in
lipid levels across levels of education might reflect the
lipid levels in African Americans from Jackson. However, age- and lifestyle-adjusted lipid levels in African
American men and women from Jackson and Forsyth
showed no significant differences between levels of
educational attainment (data not shown).
DISCUSSION
African American-white differences in lipid levels
Findings from this study (table 2) confirm reports of
1) higher levels of HDL cholesterol (4-10, 27) and
similar levels of total cholesterol (8-10) and LDL
cholesterol (3, 6) in African American men as compared with white men and 2) higher levels of apolipoprotein A-I (3, 7, 28) and lipoprotein(a) (12) and
lower levels of plasma triglycerides (6, 8-11) in African American men and women as compared with
white men and women. Similarly, lower LDL cholesterol levels in African American women as compared
with white women have been reported (8,9). Although
Am J Epidemiol
Vol. 148, No. 8, 1998
757
the higher levels of HDL cholesterol in African American men as compared with white men may have a
genetic basis (1, 3-6, 8, 9), the discrepancy between
the differences in African American and white men
found in this and many other studies (4-10, 27),
versus the lower HDL cholesterol levels reported in
relatively affluent African American men and women
as compared with white men and women in the Framingham Minority Study (11), suggests that environmental, lifestyle, or social factors might more be important determinants of HDL cholesterol levels.
In contrast, LDL cholesterol levels were higher in
white men than in African American men (8, 9), and
African American women had higher levels of HDL
cholesterol (8), total cholesterol, and apolipoprotein B
than white women did (3). However, to our knowledge
our findings on HDL2 and HDL3 cholesterol have not
been reported previously, but they are consistent with
our findings on HDL cholesterol. Apart from lipoprotein(a) in African Americans, these findings suggest
that potentially modifiable lifestyle factors influence
plasma lipid levels.
Sex differences within racial groups
In our study, the associations found between education level and total, LDL, HDL, HDL2, and HDL3
cholesterol and apolipoprotein A-I and B (table 4) are
consistent with data from the Lipid Research Clinics
Program Prevalence Study (29) and the Second National Health and Nutrition Examination Survey (7). In
contrast, Hames et al. (2) reported that in 1960, there
was a positive association between total cholesterol
and education level in white men and women. However, this was before the dramatic decrease in coronary
heart disease mortality that occurred during the late
1960s (4,13). Similarly, Freedman et al. (10) observed
no association between HDL cholesterol and socioeconomic status among white men.
An inverse association between total cholesterol and
educational attainment in whites was also reported in
the Troms0 Heart Study (30) but not in three Chicago,
Illinois, epidemiologic studies (31). However, inverse
associations of LDL cholesterol and its associated
apolipoprotein B with educational attainment in whites
(table 4) do not appear to have been reported previously. Although statistically significant, the racial differences were relatively small.
In agreement with our findings, the Princeton
School District Study (32) reported an inverse association between levels of plasma triglycerides and education in women. An inverse association between
HDL cholesterol and socioeconomic status among African American men was reported previously (5-8,
10) and is consistent with the results from the Evans
758
Metcalf et al.
County Heart Study (2), the Charleston Heart Study
(33), and the Multiple Risk Factor Intervention Trial
(6) regarding the relative protectiveness of coronary
heart disease in African American men and the finding
that mortality among African American women is equal
to or greater than that among white women (1-4).
In contrast to our findings, a positive association
was observed between HDL cholesterol and educational attainment in African American women in the
Framingham Minority Study (11) and also in the Minnesota Heart Survey (5), where it was attributed to a
high frequency of obesity at lower educational levels.
In contrast, an inverse relation was reported among
poor African American women hospitalized in Chicago (34). Thus, the relation between HDL cholesterol
and educational attainment in African American
women appears inconsistent and provides further evidence of potentially modifiable factors.
African American-white differences in
educational attainment
Among African American and white men, the differences in HDL cholesterol levels associated with
educational attainment observed in our study were
reported with various indicators of socioeconomic status by the Multiple Risk Factor Intervention Trial (6),
the Second National Health and Nutrition Examination
Survey (7), and the Vietnam Experience Study (10). In
these studies, the higher HDL cholesterol levels in
African American men were reported to drop from 11
mg/dl to 3-7 mg/dl (7, 10) with increasing levels of
socioeconomic status. Although the Minnesota Heart
Survey (5) observed a similar relation between HDL
cholesterol concentrations across educational levels, it
did not reach statistical significance. A 1 mg/dl increase in HDL cholesterol is associated with a 2 and 3
percent lower risk of coronary heart disease in men
and women, respectively, and a decrease in coronary
heart disease mortality of 3.7-4.7 percent (35).
There were also statistically significant African
American-white differences associated with education
and HDL 2 cholesterol, HDL3 cholesterol, and apolipoprotein A-I in men and women, mirroring the patterns observed for total HDL cholesterol. However, for
men, the apolipoprotein A-I racial difference associated with educational attainment was no longer statistically significant after adjusting for body mass index,
smoking, alcohol intake, and sports activity score.
Similarly, after adjusting for these variables and for
hormone therapy, the apolipoprotein A-I racial difference in women was weaker. This finding suggests that
these statistically significant racial differences are explained by lifestyle factors.
In our study, racial differences in plasma triglycer-
ides associated with educational attainment were not
statistically significant in men after adjusting for body
mass index, smoking, alcohol intake, and sports activity but remained significant in women: a negative
association in white women and no association in
African American women. To our knowledge, these
observations have not been reported previously. Similarly, lipoprotein(a) levels in African Americans and
whites were not influenced by educational attainment,
another relation that does not appear to have been
documented.
Inconsistencies in the results of African Americanwhite comparisons of lipids, lipoproteins, and apolipoproteins between studies suggest the existence of
potentially modifiable factors in addition to genetic
factors. Included may be variations in job-related
physical activity, unmeasured lifestyle or cultural factors, or regional differences in lifestyle, such as consumption of sucrose, cereal fiber, fruits, and vegetables (36, 37). Dietary habits may also be important, as
African American men in this study consumed higher
amounts of dietary cholesterol than did white men at
all levels of education (38).
Limitations
A limitation of this study is the low overall level of
response (60 percent), which raises the possibility of
selection bias and generalizability of the results to the
nonrespondent population. Another potential for selection bias could arise from exclusion of participants
who had not fasted for at least 12 hours. However, this
bias is likely to be small, as 53 percent of the subjects
in this group had diabetes mellitus. We excluded these
people because abnormalities in lipid and lipoprotein
concentrations are common in these patients (39). Participants taking medications that primarily lower cholesterol concentrations were also excluded to avoid
confounding. A further limitation is that we did not
adjust for dietary intake.
Education as a marker of socioeconomic status
Educational attainment ascertains the socioeconomic status of an individual as a measure of the
influence of prestige, power, access to resources, lifestyle behaviors that influence health, and psychosocial
stresses, all of which could potentially affect plasma
lipid levels. Educational attainment, as a surrogate
single indicator of socioeconomic status, which applies to the individual, is cumulative and cannot decrease once a certain level is achieved. Although educational level could increase, it was extremely
unlikely in our study of people aged 45-64 years.
Kitagawa and Hauser (40) noted that education provides a better global measure of lifestyle and conAm J Epidemiol
Vol. 148, No. 8, 1998
Racial Differences in Lipid Levels by Educational Attainment
sumption patterns than do other socioeconomic indicators. In our study, African American-white lipid
differences were stronger for educational attainment
than for yearly family income.
Educational level is related to receipt and evaluation
of information and to motivation to use the resources
available. However, the validity of using educational
attainment level as an index of socioeconomic status
and as a predictor of health and illness is more likely
to be dissimilar for African Americans and whites.
This residual confounding caused by African
American-white inequalities in education and income
is a cause for concern. Among people in our study
aged 45-64 years, the quality of education was lower
among African Americans, and African Americans
were generally paid less than whites in an equivalent
occupation. Moreover, there were African Americanwhite differences in the level of occupation attained
given similar education. Another limitation was the
categorization of education. For example, it was probable that a substantially higher proportion of whites
had completed well over 1 year of college.
Socioeconomic status and coronary
heart disease
The inverse association observed between socioeconomic status and levels of HDL cholesterol and its
fractions in African American men raises a question
concerning the association between socioeconomic
status and coronary heart disease in these men. One
study in Charleston, South Carolina (33), reported a
strong inverse association between socioeconomic status and incidence of coronary heart disease in African
American men, but rates were based on a small number of peer-nominated men of high socioeconomic
status. Other multiracial studies have found weak associations between socioeconomic status and risk of
coronary heart disease among African Americans (41,
42) and an inverse association among white men (13).
There is considerable geographic variation in coronary heart disease mortality in the United States. The
majority of African American participants in our study
were from Jackson, Mississippi, and they have one of
the lowest rates of coronary heart disease mortality
(43). Therefore, the African American differences in
the associations between lipids and educational attainment reported here may reflect regional differences
related to lifestyle characteristics unique to the southeastern United States.
tein A-I levels and lower plasma triglyceride levels.
Our findings confirm African American-white differences in lipids, lipoproteins, and apolipoproteins, but
not lipoprotein(a), across levels of educational attainment that were generally not explained by conventional nondietary lifestyle variables. Except in African
American men, higher education was associated with a
more favorable cardiovascular lipid profile, which
were strongest in white women. Understanding these
differences associated with educational attainment will
assist in identifying measures aimed at prevention of
cardiovascular disease.
ACKNOWLEDGMENTS
Support was provided by National Heart, Lung, and
Blood Institute contracts N01-HC-55015, N01-HC-55016,
N01-HC-55018, NOl-HC-55019, N01-HC-55020, N01HC-55021, and N01-HC-55022. This research was carried
out by Patricia A. Metcalf during the tenure of an Overseas
Research Fellowship of the Health Research Council of
New Zealand. Bruce B. Duncan was supported by a fellowship
from the Brazilian Ministry of Education (CAPES, Coordenacao de Aperfeicoamento de Pessaol de Nivel Superior).
The authors thank the following persons from the ARIC
field centers: Phyllis Johnson, Marilyn Knowles, and
Catherine Paton of the University of North Carolina at
Chapel Hill, North Carolina; Jeanette Bensen, Kay Burke,
Wilhelmenia Cheeks, and Revitha Cook of the University of
North Carolina, Forsyth County, North Carolina; Betty
Warren, Dorothy Washington, Mattye Watson, and Nancy
Wilson of the University of Mississippi Medical Center,
Jackson, Mississippi; Irene Keske, Nancy MacLennan,
Sandy Mechels, and Gail Murton of the University of Minnesota, Minneapolis, Minnesota; Rodney Palmer, Serena
Bell, Joyce Chabot, and Carol Christman of The Johns
Hopkins University, Baltimore, Maryland; Valerie Stinson,
Pam Pfile, Hogan Pham, and Teri Trevino of the University
of Texas Medical School, Houston, Texas; Wanda R.
Alexander, Doris J. Harper, Charles E. Rhodes, and Selma
M. Soyal of the Methodist Hospital, Atherosclerosis Clinical Laboratory, Houston, Texas; Nancy Bourne, Charlene
Kearney-Cash, Kelli Collins, and Celilah Cook of the
Bowman-Gray School of Medicine, Ultrasound Reading
Center, Winston-Salem, North Carolina; Debbie RubinWilliams, W. Brian Stewart, Chimmon Walter, and Louis
Wijnberg of the Collaborative Studies Coordinating Center
of the University of North Carolina at Chapel Hill, Chapel
Hill, North Carolina.
REFERENCES
Conclusions
African Americans as compared with whites had a
more protective pattern of lipids with higher mean
HDL, HDL2, and HDL3 cholesterol and apolipoproAm J Epidemiol
759
Vol. 148, No. 8, 1998
1. Gillum RF, Liu KC. Coronary heart disease mortality in
United States blacks, 1940-1978: trends and unanswered
questions. Am Heart J 1984;106:729-32.
2. Hames CG, Rose K, Knowles M, et al. Black-white compar-
760
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Metcalf et al.
isons of 20-year coronary heart disease mortality in the Evans
County Heart Study. Cardiology 1993;82:122-36.
Harris-Hooker S, Sanford GL. Lipids, lipoproteins and coronary heart disease in minority populations. Atherosclerosis
1994; 108(suppl):S83-S 104.
Glueck CJ, Gartside P, Laskarzewski PM, et al. High-density
lipoprotein cholesterol in blacks and whites: potential ramifications for coronary heart disease. Am Heart J 1984; 108:
815-26.
Sprafka JM, Norsted SW, Folsom AR, et al. Life-style factors
do not explain racial differences in high-density lipoprotein
cholesterol: the Minnesota Heart Survey. Epidemiology 1992;
3:156-63.
Watkins LO, Neaton JD, Kuller LH. 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;57:538-45.
Linn S, Fulwood R, Rifkind B, et al. High density lipoprotein
cholesterol levels among US adults by selected demographic
and socioeconomic variables: the Second National Health and
Nutrition Examination Survey 1976-1980. Am J Epidemiol
1989;129:281-94.
Tyroler HA, Glueck CJ, Christensen B, et al. Plasma highdensity lipoprotein cholesterol comparisons in black and white
populations. The Lipid Research Clinics Program Prevalence
Study. Circulation 1980;62 (suppl):IV99-IV107.
Morrison JA, Khoury P, Mellies M, et al. Lipid and lipoprotein distributions in black adults: the Cincinnati Lipid Research Clinic's Princeton School Study. JAMA 1981,245:
939-42.
Freedman DS, Strogatz DS, Eaker E, et al. Differences between black and white men in correlates of high density
lipoprotein cholesterol. Am J Epidemiol 1990;132:656-69.
Wilson PWF, Savage DD, Castelli WP, et al. HDL-cholesterol
in a sample of black adults: the Framingham Minority Study.
Metabolism 1983;32:328-32.
Schreiner PJ, Morrisett JD, Sharrett AR, et al. Lp(a) as a risk
factor for preclinical atherosclerosis. Arterioscler Thromb
Vase Biol 1993;13:826-33.
Tyroler HA, Wing S, Knowles MG. Increasing inequality in
coronary heart disease mortality in relation to educational
achievement. Profile of places of residence, United States,
1962 to 1987. Ann Epidemiol 1993;3(suppl):S51-S54.
Howard BV, Le N-A, Belcher JD, et al. Concentrations of
Lp(a) in black and white young adults: relations to risk factors
for cardiovascular disease. Ann Epidemiol 1994;4:341-50.
Patsch W, Sharrett AR, Sorlie PD, et al. The relation of high
density lipoprotein and its subfractions to apolipoprotein A-I
and fasting triglycerides: the role of environmental factors.
The Atherosclerosis Risk in Communities (ARIC) Study.
Am J Epidemiol 1992;136:546-57.
Nabulsi AA, Folsom AR, White A, et al. Association of
hormone-replacement therapy with various cardiovascular risk
factors in postmenopausal women. N Engl J Med 1993;328:
1069-75.
The ARIC Investigators. The Atherosclerosis Risk in Communities (ARIC) Study: design and objectives. Am J Epidemiol 1989;129:687-702.
Seidel J, Hagele E, Zeigerhorn J, et al. Reagent for the
enzymatic determination of serum total cholesterol with improved lipolytic efficiency. Clin Chem 1983;29:1075-80.
Nagele U, Hagele EO, Sauer G, et al. Reagent for the enzymatic determination of serum total triglycerides with improved lipolytic efficiency. J Clin Chem Clin Biochem 1984;
22:165-74.
Patsch W, Brown SA, Morrisett JD, et al. A dual-precipitation
method evaluated for measurement of cholesterol in highdensity lipoprotein subfractions HDL 2 and HDL3 in human
plasma. Clin Chem 1989;35:265-70.
Friedewald WT, Levy RI, Fredrickson DS. Estimation of the
concentration of low-density lipoprotein in plasma, without
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
the use of the preparative ultracentrifuge. Clin Chem 1972;
18:499-502.
Brown SA, Rhodes CE, Dunn K, et al. Effect of blood collection and processing on radioimmunoassay results for apolipoprotein A-I in plasma. Clin Chem 1988;34:920-4.
Brown SA, Epps DF, Dunn JK, et al. Effect of blood collection and processing on radioimmunoassay results for apolipoprotein B in plasma. Clin Chem 1990;36:1662-6.
Gaubatz JW, Cushing GL, Morrisett JD. Quantitation, isolation, and characterization of human lipoprotein (a). Methods
Enzymol 1986; 129:167-86.
Lipid Research Clinic Program. Manual of laboratory
operations: lipid and lipoprotein analyses. Bethesda, MD:
National Institutes of Health, 1982.
Baecke JAH, Burema J, Frijters JER. A short questionnaire for
the measurement of habitual physical activity in epidemiologic studies. Am J Clin Nutr 1982;36:936-42.
Sprafka JM, Folsom AR, Burke GL, et al. Prevalence of cardiovascular disease risk factors in blacks and whites: the Minnesota
Heart Survey. Am J Public Health 1988;78:1546-9.
Tyroler HA, Heiss G, Schonfeld G, et al. Apolipoprotein A-I,
A-II and C-II in black and white residents of Evans County.
Circulation 1980;62:249-54.
Heiss G, Haskell W, Mowrey R, et al. Plasma high-density
lipoprotein cholesterol and socioeconomic status. The Lipid
Research Clinics Program Prevalence Study. Circulation
1980;62(suppl):IV108-IV136.
Jacobsen BK, Thelle DS. Risk factors for coronary heart
disease and level of education: the Troms0 Heart Study. Am J
Epidemiol 1988;127:923-32.
Liu K, Cedres LB, Stamler J, et al. Relationship of education
to major risk factors and death from coronary heart disease,
cardiovascular diseases and all causes. Findings of three Chicago epidemiologic studies. Circulation 1982;66:1308-14.
Khoury PR, Morrison JA, Laskarzewski P, et al. Relationships
of education and occupation to coronary heart disease risk
factors in schoolchildren and adults: the Princeton School
District Study. Am J Epidemiol 1981;113:378-95.
Keil JE, Loadholt CB, Weinrich MC, et al. Incidence of
coronary heart disease in blacks in Charleston, South Carolina.
Am Heart J 1984;108:779-86.
Ford E, Cooper R, Simmons B, et al. Sex differences in high
density lipoprotein cholesterol in urban blacks. Am J Epidemiol 1988; 127:753-61.
Gordon DJ, Probstfield JL, Garrison RJ, et al. High-density
lipoprotein cholesterol and cardiovascular disease: four prospective American studies. Circulation 1989;79:8—15.
Marmot MG, Adelstein AM, Robinson N, et al. Changing
social-class distribution of heart disease. Br Med J 1978;2:
1109-12.
Ernst N, Fisher M, Smith W, et al. The association of plasma
high-density lipoprotein cholesterol with dietary intake and
alcohol consumption. The Lipid Research Clinics Program
Prevalence Study. Circulation 1980;62 (suppl):IV41-IV52.
Shimakawa T, Sorlie P, Carpenter MA, et al. Dietary intake
patterns and sociodemographic factors in the Atherosclerosis
Risk in Communities Study. Prev Med 1994;23:769-80.
Ginsberg HN. Lipoprotein physiology in nondiabetic and diabetic states. Relationship to atherogenesis. Diabetes Care
1991;14:839-55.
Kitagawa EM, Hauser PM. Differential mortality in the United
States: a study in socioeconomic epidemiology. Cambridge,
MA: Harvard University Press, 1973.
Kraus JF, Borhani NO, Franti CE. Socioeconomic status,
ethnicity, and risk of coronary heart disease. Am J Epidemiol
1980; 111:407-14.
Keil JE, Sutherland SE, Knapp RG, et al. Mortality rates and
risk factors for coronary heart disease in black as compared
with white men and women. N Engl J Med 1993,329:73-8.
Leaverton PE, Feinleib M, Thorn T. Coronary heart disease
mortality rates in United States blacks, 1968-1978: interstate
variation. Am Heart J 1984; 108:732-7.
Am J Epidemiol
Vol. 148, No. 8, 1998