© 1999 American Heart Association, Inc.
Original Contributions |
The Relationship Between Risk Factor Levels and Presence of Coronary Artery Calcification is Dependent on Apolipoprotein E Genotype
From the Department of Human Genetics, University of Michigan, Ann Arbor, Mich (S.L.R.K., M.B.H., C.F.S.) and the Department of Human Genetics, University of Pittsburgh, Pittsburgh, Penn (R.E.F.).
Correspondence to Dr. Sharon L. R. Kardia, Department of Human Genetics, 4708 Medical Science Building II, University of Michigan Medical School, Ann Arbor, MI 48109-0618. E-mail skardia{at}umich.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Abstract—An important research question in the study of the genetics of coronary artery disease (CAD) is whether information about genetic variation will improve our ability to predict CAD beyond established risk factors. This question is especially relevant to the goal of identifying young, asymptomatic adults with coronary atherosclerosis who would benefit most from interventions to reduce risk. Coronary artery calcification (CAC) detected by electron-beam computed tomography is a relatively new method for detecting coronary atherosclerosis in asymptomatic individuals that has been shown to be a more accurate indicator of coronary atherosclerosis in asymptomatic individuals than other noninvasive techniques. In a study of asymptomatic women (n=169) and men (n=160) between the ages of 20 and 59 representative of the Rochester, Minnesota population, we used logistic regression to ask whether the most common Apolipoprotein (Apo) E genotypes (
3/2, 3/3, and
4/3) predict the presence of CAC. The addition of
information about ApoE genotypes to logistic
models containing each separate risk factor did not improve prediction
of CAC (P>0.10 in both women and men). However, there
was significant evidence (P<0.10) that associations
between variation in the probability of having CAC and variation in
body mass index, plasma total cholesterol, and
plasma ApoB in men and body mass index, plasma
triglycerides, plasma ApoA1, and plasma ApoE in women were
dependent on ApoE genotype. Thus, variation in
the gene coding for ApoE may play a role in determining the
contribution of established risk factors to risk of CAC.
Key Words: atherosclerosis • risk factors • genetics • calcium • computed tomography
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Coronary artery disease (CAD) is the primary cause of mortality in the United States today.1 CAD is primarily detectable from clinical phenotypes such as myocardial infarction, angina, or sudden death due to coronary artery occlusion. Given the severity of these first disease symptoms, a major research goal has been to identify asymptomatic individuals with increased risk of coronary atherosclerosis so that treatment can be provided before they reach the clinical horizon. However, early diagnosis of asymptomatic individuals has been difficult because most noninvasive methods of coronary atherosclerosis detection (such as the treadmill exercise test or thallium scintigraphy) have low sensitivity in younger, asymptomatic individuals.2 It has been known for many years that calcium deposits in coronary arteries are associated with atherosclerotic plaques.3 4 5 Several studies4 6 7 have found strong associations between CAD and the presence of coronary artery calcification (CAC) identified by autopsy, by computed tomography (CT), and by fluoroscopy. Recently, technologies have been developed to accurately detect CAC with electron beam CT. Recent studies8 9 have shown that CAC is a more sensitive marker for coronary atherosclerosis than other noninvasive techniques and that presence of CAC predicts future CAD morbidity and mortality in asymptomatic and symptomatic10 adults.
During the past 10 years, genetics has become a central focus of research into the causes and predictors of common human diseases like CAD. One research focus has been on gaining knowledge about the genetic causes of common diseases that will direct our attention to the intermediate biochemical and physiological traits that may be used to detect those with increased risk or may serve as targets for early intervention to prevent disease. A second focus has been on identifying the genetic variations associated with interindividual variation in disease initiation, progression, or clinical severity that may provide more accurate prediction of disease than is provided by these intermediate traits. Genotypes at loci that influence interindividual variation in intermediate traits and are hypothesized to contribute to susceptibility of CAD, such as the one coding for apolipoprotein E (ApoE), may be better indicators of risk than intermediate traits because (1) an individual's genotype is more stable over time than intermediate traits, (2) an individual's genotype is not influenced by the disease process, and (3) each gene may have effects on multiple intermediate traits, some of which may be difficult to measure or inaccessible in vivo. On the other hand, it is also important to note that a genotype's influence on intermediate traits and on disease susceptibility may change over an individual's life cycle. For example, a genotype's influence may be modified by the disease process itself or the genotype may affect different sets of traits or disease outcomes in different environments at different ages.
For diseases with a single gene origin in which mapping between
genetic variation and variation in presence of disease is expected to
be 
1:1, such genetic information is expected to provide both a basis
for prediction of onset and severity of disease and for selection of
interventions to prevent or treat disease. However, for common complex
diseases like CAD, it is far from clear whether genetic variation will
have utility for predicting or treating disease in the population at
large. In general, it is expected that allelic variation in some CAD
susceptibility genes will influence variation in risk of CAD that is
not reflected by variation in the intermediate traits presently
considered and variation in some intermediate traits will capture
information about environmental effects and allelic variation in
susceptibility genes that have not yet been measured. Consequently, an
important research question in the study of the genetics of CAD is
whether information about genetic variation will improve our ability to
predict CAD beyond established risk factor traits. This question is
especially relevant to the goal of identifying young,
asymptomatic adults with increased risk of coronary
atherosclerosis who would benefit most from
interventions to reduce risk.
One of the best studied CAD susceptibility genes is ApoE.
The common alleles of the ApoE gene (the
2, 3, and 4) give rise to 6
genotypes; 3 of these genotypes (3/2,
3/3, and 4/3) are common in most
populations studied. ApoE genotypic information is
hypothesized to confer CAD risk information because of its association
with variation in lipid and apolipoprotein levels that are risk factors
for CAD.11 12 13 14 15 Evidence to support this hypothesis
comes from findings that the relative frequency of the 4
allele is elevated in high risk populations,16 17
increased in those individuals with CAD in many case-control and
cross-sectional studies,18 19 20 21 and elevated in elderly men
who died of CAD during a 5-year follow-up.22 Although
there is increasing evidence of ApoE as a CAD susceptibility
gene, there have also been several studies that have reported no
significant association between ApoE genotypes and
risk of clinically defined symptomatic
CAD.23 24 25 26 27
Most ApoE genotype association studies have measured only the marginal, independent association between variation in ApoE genotype and variation in risk of CAD either before or after statistical adjustment for variation in other risk factors. The marginal effect of a genotype is a statistical measure of the impact of a particular genotype, averaged over all of the phenotypes in a particular sample that are associated with that particular genotype. However, we point out that the marginal effect of a genotype is only 1 way of looking at the influence of variation in a gene on variation in risk of disease or its associated risk factors. For example, it is becoming better understood that the influence of ApoE genotype on biochemical and physiological CAD risk factors is dependent on the levels of other risk factors such as age, body size, gender, and smoking status14 15 28 29 30 31 and that the relationship between intermediate biochemical and physiological traits is also dependent on ApoE genotype.12 32 33 These studies suggest that variation in ApoE genotype may provide additional information about CAD risk beyond its marginal association when it is combined with knowledge about other risk factors. To investigate these issues, we considered 3 research questions in our study: (1) Does ApoE genotype predict the probability of having CAC when considered alone? (2) Does ApoE genotype predict the probability of having CAC after statistical adjustment for other CAD risk factors? and (3) Does ApoE genotype influence the statistical relationship between the CAD risk factors and the probability of having CAC?
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Sample
As a part of the Rochester Family Heart Study, previously sampled individuals between the ages of 20 and 5934 35 were recruited to participate in a study of the epidemiology of CAC. Anyone who had undergone coronary surgery was not eligible to be in the CAC study. Further details of the sampling for this study are presented in Maher et al.36 In the present study, we restricted the sample to those individuals who were measured for the ApoE genotype (as of January 1995) and had 1 of the 3 most common ApoE genotypes (ie,
3/2,
3/3, or 4/3 genotypes). Six
individuals with definite hypertension, 2 individuals with diabetes, 1
individual with a metal rod in his spine, 1 individual with
symptomatic CAD, 1 individual with Klinefelter's syndrome,
and 1 woman who reported that she was lactating were excluded from the
present study because their conditions either substantially
influence the measurement of CAC or influenced the risk factor levels
that were measured in each individual. Women who reported they had
taken exogenous hormones in the past month were not removed because
they represent only 1 of a large number of environmental
factors that influence normal variation in lipid and
apolipoprotein levels.
Laboratory Methods
Age at examination (in years) was calculated using an
individual's birth date. Height and weight were measured with a wall
stadiometer and a beam balance. Body mass index (BMI)
(kg/m2) was used as a measure of body size.
Systolic and diastolic blood pressure was measured
in the right arm with a random-zero sphygmomanometer (Hawkslet & Sons,
Ltd) at the Korotkoff phase I sound and the Korotkoff phase V
sound, respectively. Three measures taken at least 2 minutes apart were
averaged. Total cholesterol and triglyceride
levels were measured by standard enzymatic methods (Beckman kits) as
described by Kottke et al.37 HDL cholesterol
(HDL-C) was measured according to the method of Izzo et
al.38 Plasma levels of ApoA1, ApoB, and ApoE were measured
by the use of radioimmunoassays.37 The isoforms of the
ApoE molecule were determined from frozen plasma samples by isoelectric
focusing according to methods described in Kamboh et al.39
The correspondence between ApoE genotypes assayed at
the DNA level and plasma ApoE isoforms has been shown in a sample of
unrelated individuals from the Rochester Family Heart
Study.40 Therefore, the ApoE isoforms
phenotype were used to infer the ApoE
genotypes referred to in the present study. Use of the term
"ApoE genotype" also avoids confusion when
referring to the plasma level of ApoE, which is also an ApoE
phenotype.
Electron Beam CT Protocol
Details of the electron beam CT protocol are given in Maher et
al.36 Briefly, individuals were placed in a supine
position on the scanning bench of the electron beam CT scanner (Imatron
C-100, Imatron Inc). Forty contiguous 3-mm thick computed tomograms
(transverse 2-dimensional images) were obtained from the root of the
aorta to the apex of the heart. Scans were obtained with a
30-cm2 field of view and transferred into a 512x512
reconstruction matrix in which 1 pixel=0.343 mm.2
Electrocardiographic triggering was used so that all images were
obtained at late diastole when cardiac motion was minimal.
Image acquisition time was 100 ms.
CT scan results were reviewed for technical quality and then scored by
the use of information software by a single radiologic technologist who
identified each focus of CAC, defined as an area of at least 2 adjacent
pixels (0.69 mm2 under field of view of 30) with a CT
number >130 Hounsfield Units (HU) within the borders of a
coronary artery, and marked it as a region of interest. Only 1
radiologic technologist was needed to score each examination because a
study of interobserver and intraobserver reliability indicated it is
not necessary for multiple observers to independently score a single
examination.41 In the present study, we defined
presence of CAC as any calcific area 
1.0
mm2 in size with a CT number >130 HU in any of
the 4 epicardial arteries. We used the 1.0
mm2 definition of calcific area rather than a
0.69 mm2 (2 pixel) definition to
discriminate more fully between individuals who do and do not have
reproducibly detectable CAC. This definition corresponded to a
calcification score, as defined by Agaston,42 of 1.0. In
our study, there were no individuals with a calcific area <1 that had
calcification scores >1.
Statistical Methods
All analyses were done in females and males separately.
Unless otherwise indicated, the results from the statistical
analyses discussed below were reported with 4 significance
intervals (0.10>P 0.05, 0.05>P>0.01,
0.01>P0.001, and P<0.001) to allow a
continuum of liberal to conservative assessment of results by the
reader. We chose to discuss the traditionally marginally significant
results (0.10>P0.05) in the text because in this
relatively small preliminary study we are more concerned with type II
errors than type I errors.
To address the first question in our study, logistic regression methods
were used to assess whether there were significant difference among the
3 most common ApoE genotypes in predicting the
probability of having CAC when considered alone. The effects of the
3/2 and 4/3 genotypes were modeled
by coding design variables in such a way that the 3/3
genotype represented the reference class (ie, the
design variable for 3/2 equaled 1 when
ApoE genotype was 3/2 and 0
otherwise; the same coding design was also used for the
4/3 genotype). We used the likelihood ratio test
to assess whether variation among these 3 ApoE
genotypes was a significant predictor of variation in the
probability of having CAC. Specifically, twice the natural logarithm of
the ratio of the maximum of the likelihood of a logistic model with the
ApoE genotype variables versus the maximum of
the likelihood of a logistic model with an intercept only was taken to
be approximately distributed as a 
2
distribution with 2 degrees of freedom.
To address the second question of whether ApoE
genotype was a significant predictor after considering
established CAD risk factor traits, we compared the maximum of the
likelihood of a logistic model that contained each CAD risk factor
considered separately [(ie, Model 1: logit of P (CAC is
present|risk factor)=
+ß (risk factor)] with the maximum of
the likelihood of a logistic model containing both the particular CAD
risk factor and ApoE genotype variation [(ie, Model
2: logit of P (CAC is present|risk factor, ApoE
genotype)=+ß1 (risk
factor)+ß2
(3/2)+ß3 4/3)] by
use of the likelihood ratio test. Twice the natural logarithm of the
ratio of Model 2 to Model 1 is approximately distributed as a
2 distribution with 2 degrees of freedom.
To address the third question of whether variation in the
ApoE genotype influenced the statistical
relationship between any of the CAD risk factors and the probability of
having CAC, we compared the maximum of the likelihood of a logistic
model containing a single CAD risk factor and ApoE
genotype (Model 2) versus the maximum of the likelihood of a
logistic model containing a single CAD risk factor, ApoE
genotype variables, and ApoE genotype
variables by CAD risk factor interactions [(ie, Model 3: logit of
P (CAC is present|risk factor, ApoE genotype,
ApoE genotype · risk factor
interactions)=+ß1 (risk
factor)+ß2
(3/2)+ß3
(3/4)+ß4
(3/2 · risk factor)+ß5
(4/3 · risk factor)] by use of the likelihood ratio
test. Twice the natural logarithm of the ratio of Model 3 to Model 2 is
approximately distributed as a 2 distribution
with 2 degrees of freedom. Interpretation of evidence for the
ApoE genotype-dependent relationship between
variation in the probability of having CAC and variation in risk
factors assumes homogeneity of the range of risk factor values among
genotypes. To test this assumption, we used a 2-sample
t test and F test to test the null hypothesis of no mean and
variance differences, respectively, in quantitative risk factors
between 2 common genotypes with P<0.05 as the level
of statistical significance. We considered 2 contrasts for each
statistic, 3/3=3/2 and
3/3=4/3. Satterthwaite's correction for
the t test was used when there was evidence of significant
heterogeneity of variances between genotype
classes.43
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Descriptive Statistics
In our sample of 169 women and 161 men between the ages of 20 and 60, we observed that 17% of women and 41% of men had CAC; the actual quantities of calcium observed in our sample ranged between 1.03 mm2 and 281.8 mm.2 In individuals under the age of 40, 14% of women (n=73) and 21% of men (n=82) had CAC. In individuals between the age of 40 and 60, 20% of women (n=96) and 62% of men (n=79) had CAC.
In Table 1
, we present the
descriptive statistics for the distribution of CAD risk factors in each
of the ApoE genotype and gender strata. In women,
those with the 3/2 genotype had a significantly
higher, and those with the 4/3 genotype had a
significantly lower mean ApoE level, than those with the
3/3 genotype. In men, both those with the
3/2 genotype and the 4/3
genotype had a significantly higher mean
triglyceride level and larger intragenotypic variance than
those with the 3/3 genotype. Also in men, those
with the 4/3 genotype had significantly lower
mean levels of HDL-C and ApoA1 than those with the 3/3
genotype. Men with the 3/2 genotype had a
significantly higher average level and larger intragenotypic variance
in ApoE than men with the 3/3 genotype.
|
ApoE Gene Variation as a Predictor of
CAC
To address our first research question, we used logistic
regression to model the association between the variation in
probability of having CAC and variation in ApoE
genotype. In Table 2, we
present the frequency distribution of ApoE
genotypes and CAC. In women, 19.0% of 3/2
individuals, 17.3% of 3/3 individuals, and 16.0% of
4/3 individuals had CAC. In men, 42.1% of
3/2 individuals, 43.1% of 3/3 individuals,
and 35.9% of 4/3 individuals had CAC. As expected given
this frequency distribution, the logistic regression results
presented in Table 3 indicate
that there is no statistically significant association between
variation in the probability of having CAC and variation in
ApoE genotype.
|
|
As a first step to address our second question of whether variation in
ApoE genotype was a predictor after considering
established CAD risk factors, we performed univariate
logistic regressions on each of the CAD risk factors (denoted Model 1)
and then compared Model 1 with a model containing only an intercept
term. The results from these comparisons are presented in
Column A of Tables 4 and 5 for women and men, respectively. When
each of the CAD risk factors were considered alone, we found that age,
BMI, systolic blood pressure (BP), diastolic BP,
HDL-C, and ApoB were significant predictors of the presence of CAC in
women. In men, all the CAD risk factors, except HDL-C and ApoA1, were
significant predictors. The logistic regression coefficients indicate
that increased levels of all the CAD risk factor traits, except ApoA1
and HDL-C, are associated with increased risk of CAC. The odds ratios
associated with a 1 SD increase or decrease from the mean value of each
risk factor are included in Tables 4 and 5 as an illustration of
the magnitude and range of the observed risk factor association with
presence of CAC in this sample. The second step in addressing our
second question was to add information about variation in
ApoE genotype to each logistic model containing a
CAD risk factor (ie, Model 2). The results from comparing Model 1 and
Model 2 are presented in Column B of Tables 4 and 5 in
women and men, respectively. In both genders, in no case was there a
significant improvement in the prediction of the probability of having
CAC when information about variation in ApoE
genotype was added.
|
|
To address our third question of whether ApoE
genotype influences the relationship between variation in the
probability of having CAC and variation in established CAD risk
factors, we constructed logistic models containing each CAD risk
factor, information about ApoE genotype, and
interaction terms between variation in ApoE genotype
and the particular CAD risk factor in the model (denoted Model 3). This
model was compared with Model 2, which contained no interaction terms.
The 2 from the likelihood ratio test between
Model 3 and Model 2 are presented in Column C of Tables 4
and 5 for women and men, respectively. In women,
ApoE genotype had a significant influence on the
relationship between variation in the probability of having CAC and
variation in BMI, triglycerides, ApoA1, and ApoE levels. In
the Figure, we plot the influence of each
ApoE genotype on the relationship between variation
in the estimated probability of having CAC and variation in these
coronary heart disease (CHD) risk factors in women. For BMI in
women, the 3/2 genotype was associated with a
sharp increase in the estimated probability of having CAC for BMI's
>25, whereas the 4/3 and 3/3
genotypes were similar in having more gradual increases for a
given BMI level. For both plasma triglycerides and plasma
ApoE levels in women, the 4/3 genotype had the
steepest increase in the estimated probability of having CAC for a
given increase in risk factor level, whereas the 3/3
genotype was associated with a more moderate increase and the
3/2 was associated with a decrease in the probability of
having CAC for a given increase in risk factor level. For ApoA1 in
women, we found that the 3/2 genotype had the
sharpest decrease in the estimated probability of having CAC for a
given increase in ApoA1 levels compared with the 4/3
genotype that had a more gradual decrease and the
3/3 genotype that had a slight increase. Overall,
women with the 3/2 genotype had a mixed
combination of risk factor relationships with the probability of having
CAC compared with women with the 4/3 genotype.
Specifically, women with the 3/2 had an expected increase
in probability of having CAC with increasing BMI and decreasing ApoA1
levels, but (counterintuitively) had a negative relationship between
the probability of having CAC and triglyceride or ApoE
levels. In contrast, women with the 4/3 genotype
had an expected increase in the probability of having CAC with
increasing BMI, triglycerides, and ApoE levels and expected
decreases in the probability of having CAC for increasing ApoA1
levels.
|
In men, ApoE genotype had a significant influence on
the relationship between variation in the probability of having CAC and
variation in BMI, total cholesterol, and ApoB levels (Table 5, Column C). In the Figure, we plot the influence of
each ApoE genotype on the relationship between
variation in the estimated probability of having CAC and variation in
these risk factors in men. For BMI in men, the 3/2 and
4/3 genotypes had greater increases in the
estimated probability of having CAC then the 3/3
genotype. For both total cholesterol and ApoB in
men, the 3/3 and 3/2 genotypes were
associated with a significant positive relationship between the
probability of having CAC and these risk factors, whereas the
probability of having CAC did not increase significantly with total
cholesterol or ApoB levels in those with the
4/3 genotype. Overall, men with the
3/2 genotype and 4/3
genotypes both had positive relationships between the
probability of having CAC and BMI, total cholesterol, and
ApoB levels, although in each of these instances men with the
3/2 genotype had larger logistic regression
coefficient estimates that indicate a stronger relationship between CAC
and these intermediate traits in 3/2 men versus
4/3 men.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Association of CAD With ApoE Genotype
Hundreds of studies have been carried out to evaluate the association between variation in risk of disease, variation in intermediate biochemical and physiological risk factors for disease,16 and variation in the gene coding for the ApoE molecule. Strong associations between the frequency of CAD, ischemic heart disease (IHD), or CHD and the frequency of the ApoE genotypes carrying the
4
allele have been found in case control
studies,18 19 20 21 44 45 46 in studies of special
population subgroups,16 17 47 and recently in a
prospective study.22 These results have raised
expectations that the ApoE polymorphism may be useful in
identifying asymptomatic individuals who are at higher risk
of CAD. However, we also point out that there have also been numerous
studies that have found no associations,23 24 25 26 27 28 or the
relative frequency of the 4 allele was actually lower
in cases than in control subjects.48 Other studies
have found associations between the frequency of disease (CAD, IHD, and
CHD) and the frequency of the 2 allele in addition to
an association with the frequency of the 4
allele,49 50 whereas 1 study only found an
association between the presence of carotid artery disease and the
presence of the 2 allele.51 These
heterogeneous results among studies suggest that the
influence of variation in the ApoE genotype on
variation in disease susceptibility is likely to be
heterogeneous within and between populations; that is, it
may change over an individual's life time, it may be influenced by the
disease process itself, and it may be different in different
environments or genetic backgrounds. Our finding that the relationship
between variation in established risk factor traits and variation in
presence of CAC is influenced by variation in the ApoE
genotype, even though we found no marginal statistical
association, also supports the general conclusion that the influences
of variation in the ApoE genotype on the variations
in biochemical and physiological traits that
determine risk of CAD may be biologically significant in some subgroups
or populations but not in others. To our knowledge, there have only been 2 studies that have investigated the association between variation in atherosclerotic lesions histologically measured in vitro in the coronary arteries or aorta52 53 and variation in the ApoE genotype. In both studies, arteries were removed at autopsy and evaluated to determine (1) the total percent of intimal surface involved in all atherosclerotic lesions (fatty streaks, fibrous plaques, complicated lesions, and calcified lesions) and (2) the total percent of intimal surface area involved in raised lesions (fibrous plaques, complicated lesions, and calcified lesions). Hixson et al 199152 observed no association between variation in either total percent surface area with atherosclerotic lesions or total percent surface area associated with raised lesions in the right coronary artery and variation in the ApoE genotype in white males 15 to 34 years of age who died of accidental causes. In both the thoracic and abdominal aorta, only variation in the total surface area associated with all atherosclerotic lesions, not raised lesions, was associated with variation in the ApoE genotype. In a sample of Alaskan natives 9 to 85 years in age who died of external causes (eg, accidents, homicide, suicide), again only variation in the total percent surface area defined by all atherosclerotic lesions, not raised lesions, was significantly associated with variation in the ApoE genotype.53 In as much as the results of our study can be compared with the results of these studies (our measure of atherosclerotic lesions was restricted to calcified lesions), our findings are consistent with these studies in that we found no marginal association of calcified lesions with variation in the ApoE genotype. Taken together, studies of lesion phenotypes suggest that variation in initiation of CAD may be more strongly associated with variation in the ApoE genotype than the progression of CAD to the raised lesion status.52 53
ApoE Genotype Influences the
Relationship Between Established Risk Factors and CAC
Conceptually, our finding that variation in ApoE
genotype influences the relationship between variation in the
probability of having CAC and variation in established risk factors
provides new insight into the complexity of the relationships between
genome variability and variability in phenotypes that are
measures of health. In basic terms, our findings further demonstrate
that the mapping function between risk factor variation and
interindividual variation in lifetime risk of progressive
atherosclerosis is dependent of the context defined by
an individual's genotype. This conclusion is both congruent
with our understanding of the complexity of the etiology of
CAD54 that leads to the expectation that not everyone
with high levels of atherogenic risk factors will necessarily develop
disease, and it is incongruent with the commonly held belief that high
levels of an atherogenic risk factor are invariably associated with
increased risk. How could high levels of atherogenic risk factors not
be associated with increased risk of CAD? Or conversely, how could low
levels of atherogenic risk factors be associated with increased risk?
Although we do not have knowledge of the biological mechanisms that
might provide the answers to these particular questions, we have a
basic understanding of how such events happen because we know that
variation in disease susceptibility among individuals is a consequence
of the interdependency55 of the effects of susceptibility
genotypes at many loci and the effects of different
environmental exposures.54 So, for instance, our findings
that the 4/3 genotype in women was associated
with increased probability of having CAC at increased BMI, increased
triglyceride levels, decreased Apo AI levels, and increased
ApoE levels compared with the 3/3 genotype
(Figure) are congruent with our understanding of CAD risk
factors relationship with the risk of developing
atherosclerosis. In contrast, the 4/3
genotype in men was associated with increased probability of
having CAC for increased BMI but not total cholesterol or
ApoB levels compared with the 3/3 genotype. Does
this mean that the level of total cholesterol or ApoB in
4/3 men is not a significant determinant of their
probability of having CAC as much as other factors such as BMI, or
other risk factors that were not genotype-gender dependent? We
can only speculate that there are 2 underlying reasons for what we are
seeing: (1) there are multiple other factors varying (age,
smoking, etc) such that what we see with these 2-dimensional figures is
a distortion of a multivariate reality where everything
would be congruent if we could measure all the relevant variables
for each individual in our sample, and (2) the variation in the ApoE
molecules, represented by the genotype, has
systemic, wide ranging effects on many biochemical and
physiological interrelationships that are different
in the preclinical and clinical manifestations of
atherosclerosis. We know so little about the
progression of atherosclerosis in free living
individuals because up to now we have not had the ability to measure
this stage of the atherosclerotic process. We note that our study only
provides preliminary evidence for these ApoE
genotype dependent effects and its inferences are limited
because of our relatively small sample size and the relatively young
age of our sample. Consequently, we can only speculate about the
etiological underpinnings of the particular statistical associations
between probability of having CAC and the particular ApoE
genotype, gender, and risk factor level combinations that we
observed in our study.
Uncertainty in the predictive ability of each particular risk factor
(intermediate trait, susceptibility genotype, environmental
factor) arises because each risk factor has a range of influences that
changes according to the context defined by the states of the other
intermediate traits, genotypes at other loci, and environmental
factors. In other words, many individuals with a particular atherogenic
risk factor level or a particular CAD susceptibility genotype
may remain healthy because of compensatory effects determined by the
states of their genetic or environmental backgrounds. In general, the
distribution of the etiologies of common diseases such as CAD among
individuals in the population at large is expected to involve only a
small number of genetic or environmental factors with large marginal
effects on risk of disease54 56 that are not dependent on
other genetic and environmental factors. Variations in most causal
agents, genetic and environmental, will have very small marginal
effects in humankind at large; although, in a particular subsets of
individuals, families, or populations, representing a
particular subset of etiologies defined by the levels of other genetic
or environmental agents, an agent may have a relatively large effect.
Our findings demonstrate both of these points. First, we did not
observe a statistically significant marginal association between
variation in the probability of having CAC and variation in the
ApoE genotype. Second, we illustrated
(Figure) that only at particular levels of intermediate risk
factors is there an association between variation in the probability of
having CAC and variation in the ApoE genotype. Thus,
our study suggests that variation in a gene that has a small marginal
association can still contribute to prediction of disease through
interaction effects between interdependent agents and that we may miss
a major portion of the genetic architecture of common diseases by
assuming that only genotypes with large marginal effects acting
in an independent and additive fashion on disease risk are important
for prediction. In other words, whether the 2 or
4 allele is a risk factor for having CAC depends on
its context, such as gender and biochemical risk factor level. This
concept is not new in traditional animal and plant genetics in which
the influence of a genotype is often viewed from 2 vantage
points: (1) how it impacts the initial conditions underlying the
development of the adult organism and (2) how it sets boundaries on the
response to each shift in environment.
Our results also raise the question of how many biological models of disease will be necessary to explain the prevalence of CAD in the population at large—1 for each ApoE genotype and gender? At one end of the spectrum, we often assume that everyone has the same etiological processes operating to determine their disease risk and we seek to develop a single, "best" explanation in the statistical sense, model that relates agents involved in the causes of disease to prediction of risk of disease. We have gained much in terms of our general understanding of the causes of the risk of CAD from this approach. However, at the other end of the spectrum, we know that each individual is genetically and environmentally unique and therefore has their own specific, complex multifactorial etiology defined by the states of only a subset of all agents that are involved in determining the prevalence of disease in humankind in general. We suggest that the next generation of studies should focus on understanding the distribution of etiologies within and between populations and on identifying that subset of genotype-environment combinations that are associated with high risk of CAD in particular subsets of humankind.
|
Acknowledgments |
|---|
We would like to thank Dr Stephen T. Turner for his supervision of the collection of the pedigrees in the Rochester Family Heart Study. We would like to thank Dr Patrick Sheedy for his supervision of the collection of the CAC measurements. Dr Kardia would especially like to thank Dr Patricia Peyser for helpful discussions about this work. Measurements of plasma lipids and apolipoproteins were carried out in the Mayo Atherosclerosis Research laboratory under the direction of Dr B. A. Kottke. This research was supported by National Institutes of Health grants HL39107, HL46292, and HL58240.
Received December 23, 1997; accepted July 10, 1998.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
1. American Heart Association. Heart and Stroke Facts Statistics: 1996 Statistical Supplement. Dallas, Tex: American Heart Association National Center; 1996.
2. Hopkirk JA, Uhl GS, Hickman JR Jr, Fischer J, Medina A. Discriminant value of clinical and exercise variables in detecting significant coronary artery disease in asymptomatic men. J Am Coll Cardiol. 1984;3:887–894.[Abstract]
3. Enos WF, Holmes RH, Beyer J. Coronary disease among United States soldiers killed in action in Korea. JAMA. 1953;256:2859–2862.
4. Blankenhorn D. Coronary arterial calcification: a review. Am J Med Sci. 1961;242:41–49.
5.
McNamara JJ, Molot MA, Stremple JF, Cutting RT.
Coronary artery disease in combat casualties in Vietnam.
JAMA. 1971;216:1185–1187.
6.
Bartel AG, Chen JT, Peter RH, Behar VS, Kong Y, Lester
RG. The significance of coronary calcification detected by
fluoroscopy. Circulation. 1974;49:1247–1253.
7. Simons DB, Schwartz RS, Edwards WD, Sheedy PF, Breen JF, Rumberger JA. Noninvasive definition of anatomic coronary artery disease by ultra fast computed tomographic scanning: a quantitative pathologic comparison study. J Am Coll Cardiol. 1992;20:1118–1126.[Abstract]
8. Detrano RC, Wong ND, Tang W, French WJ, Georgiou D, Young E, Brezden OS, Doherty TM, Narahara KA, Brundage BH. Prognostic significance of cardiac cinefluoroscopy for coronary calcific deposits in asymptomatic high risk subjects. J Am Coll Cardiol. 1994;24:354–358.[Abstract]
9. Puentes G, Detrano R, Tang W, Wong N, French W, Narahara K, Brundage B, Baksheshi H. Estimation of coronary calcium mass using electron beam computed tomography: a promising approach for predicting coronary events? Am J Card Imaging. 1995;9(suppl 1):5. Abstract.
10. Detrano R, Hsiai T, Wang S, Puentes G, Fallavollita J, Shields P, Stanford W, Wolfkiel C, Georgiou D, Budoff M, Reed J. Prognostic value of coronary calcification and angiographic stenoses in patients undergoing coronary angiography. J Am Coll Cardiol. 1996;27:285–290.[Abstract]
11. Enholm C, Lukka M, Kuusi C, Nikkila E, Utermann G. Apolipoprotein E polymorphism in the Finnish population: gene frequencies and relation to lipoprotein concentrations. J Lipid Res. 1986;27:227–235.[Abstract]
12. Boerwinkle E, Visvikis S, Welsh D, Steinmetz J, Hanash SM, Sing CF. The use of measured genotype information in the analysis of quantitative phenotypes in man, II: the role of the apolipoprotein E polymorphism in determining levels, variability, and covariability of cholesterol, betalipoprotein, and triglycerides in a sample of unrelated individuals. Am J Hum Genet. 1987;27:567–582.
13. Havekes LM, de Knijff P, Smit M, Frants R. The effect of apolipoprotein E allele substitutions on plasma lipid and apolipoprotein levels. Adv Exp Med Biol. 1988;243:87–93.[Medline] [Order article via Infotrieve]
14.
Kaprio J, Ferrell RE, Kottke BA, Kamboh MI, Sing CF.
Effects of polymorphisms in apolipoprotein E, A-IV, and H on
quantitative traits related to risk of cardiovascular
disease. Arterioscler Thromb. 1991;11:1330–1348.
15.
Xhignesse M, Lussier-Cacan S, Sing CF, Kessling AM,
Davignon J. Influences of common variants of apolipoprotein E on
measures of lipid metabolism in a sample selected for
health. Arterioscler Thromb. 1991;11:1100–1110.
16.
Davignon J, Gregg RR, Sing CF. ApoE polymorphism
and atherosclerosis.
Arteriosclerosis. 1988;8:1–21.
17. Gerdes LU, Klausen IC, Sihm I, Faergeman O. Apolipoprotein E polymorphism in a Danish population compared to findings in 45 other study populations around the world. Genet Epidemiol. 1992;9:155–167.[Medline] [Order article via Infotrieve]
18.
Kuusi T, Nieminen MS, Ehnholm C, Yki-Jarvinen H, Valle
M, Nikkila EA, Taskinen MR. Apolipoprotein E polymorphism and
coronary artery disease: increased prevalence of apolipoprotein
E4 in angiographically verified coronary patients.
Arteriosclerosis. 1989;9:237–241.
19. Davignon J. Apolipoprotein E polymorphism and atherosclerosis. In: Born GVR, Schwartz CJ, eds. New Horizons in Coronary Heart Disease. London: Current Science; 1993:5.1–5.2.
20.
Luc G, Bard J, Arveiler D, Evans A, Cambou J, Bingham
A, Amouyel P, Schaffer P, Ruidavets J, Cambien F, Fruchart J,
Ducimetiere P. Impact of apolipoprotein E polymorphism on
lipoproteins and risk of myocardial infarction: the ECTIM Study.
Arterioscler Thromb. 1994;14:1412–1419.
21.
Wilson PW, Myers RH, Larson MG, Ordovas JM, Wolf PA,
Schaefer EJ. Apolipoprotein E alleles, dyslipidemia,
and coronary heart disease: the Framingham Offspring Study.
JAMA. 1994;272:1666–1671.
22.
Stengård JH, Zerba KE, Pekkanen J, Ehnholm C, Nissinen
A, Sing CF. Apolipoprotein E polymorphism predicts death from
coronary heart disease in a longitudinal study of elderly
Finnish men. Circulation. 1995;91:265–269.
23.
Menzel HJ, Kladetzky RG, Assmann G. Apolipoprotein E
polymorphism and coronary heart disease.
Arteriosclerosis. 1983;3:310–315.
24. Stuyt PMJ, Brenninkmeijer BJ, Demacker PNM, Hendriks JCM, Van Elteren P, Stalenhoef AFH, Van't Laar A. Apolipoprotein E phenotypes, serum lipoproteins and apolipoprotein in angiographically assessed coronary heart disease. Scand J Clin Lab Invest. 1991;51:425–435.[Medline] [Order article via Infotrieve]
25. Payne MN, Green E, Walker MR, Beattie JM, Murray RG, Jones AF. Apolipoprotein E4 and coronary artery disease. Lancet. 1992;340:1350. Letter.
26. Lehtinen S, Lehtimäki T, Sisto T, Salenius J-P, Nikkilä M, Jokela H, Koivula T, Ebeling F, Ehnholm C. Apolipoprotein E polymorphism, serum lipids, myocardial infarction and severity of angiographically verified coronary artery disease in men and women. Atherosclerosis. 1995;114:83–91.[Medline] [Order article via Infotrieve]
27.
Lenzen HJ, Assmann G, Buchwalsky R, Schulte H.
Association of apolipoprotein E polymorphism, low-density
lipoprotein cholesterol, and coronary disease.
Clin Chem. 1986;32:778–781.
28.
Hanis CL, Hewett-Emmett D, Douglas TC, Bertin TK,
Schull WJ. Effects of the apolipoprotein E polymorphism on levels
of lipids, lipoproteins, and apolipoproteins among Mexican-Americans in
Starr County, Texas. Arterioscler Thromb. 1991;11:362–370.
29. Reilly SL, Ferrell RE, Kottke BA, Kamboh MI, Sing CF. The gender-specific apolipoprotein E genotype influence on the distribution of lipids and apolipoproteins in the populations of Rochester, Minnesota, II: regression relationships with concomitants. Am J Hum Genet. 1992;51:1311–1324.[Medline] [Order article via Infotrieve]
30. Srinivasan SR, Ehnholm C, Wattigney WA, Berenson GS. Relationship between obesity and serum lipoproteins in children with different apolipoprotein E phenotypes: the Bogalusa Heart Study. Metabolism. 1994;43:470–475.[Medline] [Order article via Infotrieve]
31. Zerba KE, Ferrell RE, Sing CF. Genotype-environment interaction: apolipoprotein E (ApoE) gene effects and age as an index of time and spatial context in the human. Genetics. 1996;143:463–478.[Abstract]
32. Reilly SL, Ferrell RE, Sing CF. The gender-specific apolipoprotein E genotype influence on the distribution of plasma lipids and apolipoproteins in the population of Rochester, MN, III: correlations and covariances. Am J Hum Genet. 1994;55:1001–1018.[Medline] [Order article via Infotrieve]
33. Haviland MB, Lussier-Cacan S, Davignon J, Sing CF. Impact of apolipoprotein E genotype variation on means, variances, and correlations of plasma lipid, lipoprotein, and apolipoprotein traits in octogenarians. Am J Med Genet. 1995;58:315–331.[Medline] [Order article via Infotrieve]
34.
Turner ST, Weidman WH, Michels VV, Reed TJ, Ormson CL,
Fuller T, Sing CF. Distribution of sodium-lithium countertransport and
blood pressure in Caucasian five to eighty-nine years of age.
Hypertension. 1989;13:378–391.
35. Moll PP, Michels VV, Wiedman WH, Kottke BA. Genetic determination of plasma apolipoprotein AI in a population-based sample. Am J Hum Genet. 1989;44:124–139.[Medline] [Order article via Infotrieve]
36.
Maher JA, Raz JA, Bielak LF, Sheedy PF, Schwartz
RS, Peyser PA. Potential of quantity of coronary artery
calcification to identify new risk factors for asymptomatic
atherosclerosis. Am J Epidemiol. 1996;144:943–953.
37. Kottke BA, Moll PP, Michels VV, Weidman WH. Levels of lipids, lipoproteins, and apolipoproteins in a defined population. Mayo Clin Proc. 1991;66:1198–1208.[Medline] [Order article via Infotrieve]
38.
Izzo C, Grillo F, Murador E. Improved method for
determination of high-density-lipoprotein cholesterol, I:
isolation of high density lipoproteins by use of polyethylene
glycol-6000. Clin Chem. 1981;27:371–374.
39. Kamboh MI, Ferrell RE, Sepehrnia B. Genetic studies of human apolipoproteins, IV: structural heterogeneity of apolipoprotein h (beta-2 glycoprotein 1). Am J Hum Genet. 1988;42:452–457.[Medline] [Order article via Infotrieve]
40. Mailly F, Moll PP, Kottke BA, Kamboh MI, Humphries SE, Ferrell RE. Estimation of the frequency of isoform-genotype discrepancies at the apolipoprotein E locus in heterozygotes for the isoforms. Genet Epidemiol. 1992;9:239–248.[Medline] [Order article via Infotrieve]
41.
Kauffman RB, Sheedy PF, Breen JF, Kelzenberg JR, Kruger
BL, Schwartz RS, Moll PP. Detection of heart calcification with
electron beam CT: interobserver and intraobserver reliability for
scoring quantification. Radiology. 1994;190:347–352.
42. Agaston AS, Janowitz WR, Hildner FJ, Zusmer NR, Viamonte M, Detrano R. Quantification of coronary artery calcium using ultra fast computed tomography. J Am Coll Cardiol. 1990;15:827–832.[Abstract]
43. Satterthwaite FE. An approximate distribution of estimates of variance components. Biometrics Bull. 1946;2:110–114.
44. Cumming AM, Robertson F. Polymorphism at the apo E locus in relation to risk of coronary disease. Clin Genet. 1984;25:310–313.[Medline] [Order article via Infotrieve]
45. Kuusi T, Nieminen MS, Ehnholm C, Yki-Karvinen H, Valle M, Nikkila EA, Taskinen M. Apolipoprotein E polymorphism and coronary artery disease: increased prevalence of apolipoprotein E-4 in angiographically verified coronary patients. Arteriosclerosis. 1989;9:237–241.
46. van Bockxmeer FM, Mamotte CD. Apolipoprotein epsilon 4 homozygosity in young men with coronary heart disease. Lancet. 1992;340:879–880.[Medline] [Order article via Infotrieve]
47.
Nieminen MS, Mattila KJ, Aalto-Setala K, Kuusi T,
Kontula K, Kauppinen-Makelin R, Ehnholm C, Jauhiainen M, Valle M,
Taskinen M-R. Lipoproteins and their genetic variation in subjects with
and without angiographically verified coronary artery disease.
Arterioscler Thromb. 1992;12:58–69.
48. Utermann G, Hardewig A, Zimmer F. ApoE phenotypes in patients with myocardial infarction. Hum Genet. 1984;65:237–241.[Medline] [Order article via Infotrieve]
49. Eto M, Watanabe K, Makino I. Increased frequencies of apolipoprotein 2 and 4 alleles in patients with ischemic heart disease. Clin Genet. 1989;36:183–188.[Medline] [Order article via Infotrieve]
50. Eichner JE, Kuller LH, Orchard TJ, Grandits GA, McCallum LM, Ferrell RE, Neaton JD. The relationship of apolipoprotein E phenotype to myocardial infarction and mortality from coronary artery disease. Am J Cardiol. 1993;71:160–165.[Medline] [Order article via Infotrieve]
51. de Andrade M, Thandi I, Brown S, Gotto A Jr, Patsch W, Boerwinkle E. Relationship of the apolipoprotein E polymorphism with carotid artery atherosclerosis. Am J Hum Genet. 1995;56:1379–1390.[Medline] [Order article via Infotrieve]
52.
Hixson JE. Apolipoprotein E polymorphisms affect
atherosclerosis in young males: Pathobiological
Determinants of Atherosclerosis in Youth (PDAY)
Research Group. Arterioscler Thromb. 1991;11:1237–1244.
53. Scheer WD, Boudreau DA, Malcom GT, Middaugh JP. Apolipoprotein E and atherosclerosis in Alaska natives. Atherosclerosis. 1995;114:197–202.[Medline] [Order article via Infotrieve]
54. Sing CF, Haviland MB, Reilly SL. Genetic architecture of common multifactorial diseases. In: Chadwick D, Cardew G, eds. Variation in the Human Genome. Ciba Foundation Symposia 197. Chichester, UK: John Wiley and Sons. 1996:211–232.
55. Bar-Yam Y. Dynamics of complex systems. Reading, Mass: Addison-Wesley; 1997:12.
56. Weiss KM. Genetic Variation and Human Disease: Principles and Evolutionary Approaches. Cambridge, UK: Cambridge University Press; 1993.
This article has been cited by other articles:
 |
|
 |
|
  G. Kolovou, D. Damaskos, K. Anagnostopoulou, and D. V. Cokkinos Apolipoprotein E Gene Polymorphism and Gender Ann. Clin. Lab. Sci., January 1, 2009; 39(2): 120 - 133. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P. Reilly, M. L. Wolfe, T. Rhodes, C. Girman, N. Mehta, and D. J. Rader Measures of Insulin Resistance Add Incremental Value to the Clinical Diagnosis of Metabolic Syndrome in Association With Coronary Atherosclerosis Circulation, August 17, 2004; 110(7): 803 - 809. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. M. Doherty, L. A. Fitzpatrick, D. Inoue, J.-H. Qiao, M. C. Fishbein, R. C. Detrano, P. K. Shah, and T. B. Rajavashisth Molecular, Endocrine, and Genetic Mechanisms of Arterial Calcification Endocr. Rev., August 1, 2004; 25(4): 629 - 672. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Song, M. J. Stampfer, and S. Liu Meta-Analysis: Apolipoprotein E Genotypes and Risk for Coronary Heart Disease Ann Intern Med, July 20, 2004; 141(2): 137 - 147. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. M. Doherty, L. A. Fitzpatrick, A. Shaheen, T. B. Rajavashisth, and R. C. Detrano Genetic Determinants of Arterial Calcification Associated With Atherosclerosis Mayo Clin. Proc., February 1, 2004; 79(2): 197 - 210. [Abstract] [PDF]
|
|
|
|
 |
|
 C. F. Sing, J. H. Stengard, and S. L.R. Kardia Genes, Environment, and Cardiovascular Disease Arterioscler. Thromb. Vasc. Biol., July 1, 2003; 23(7): 1190 - 1196. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. A. Peyser, L. F. Bielak, J. S. Chu, S. T. Turner, D. L. Ellsworth, E. Boerwinkle, and P. F. Sheedy II Heritability of Coronary Artery Calcium Quantity Measured by Electron Beam Computed Tomography in Asymptomatic Adults Circulation, July 16, 2002; 106(3): 304 - 308. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. J. O'Donnell, I. Chazaro, P. W.F. Wilson, C. Fox, M. T. Hannan, D. P. Kiel, and L. A. Cupples Evidence for Heritability of Abdominal Aortic Calcific Deposits in the Framingham Heart Study Circulation, July 16, 2002; 106(3): 337 - 341. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. A. Manolio and D. E. Bild Coronary Calcium, Race, and Genes Arterioscler. Thromb. Vasc. Biol., March 1, 2002; 22(3): 359 - 360. [Full Text] [PDF]
|
|
|
|
|
|
L. A. Lange, E. M. Lange, L. F. Bielak, C. D. Langefeld, S. L. Kardia, P. Royston, S. T. Turner, P. F. Sheedy II, E. Boerwinkle, and P. A. Peyser Autosomal Genome-Wide Scan for Coronary Artery Calcification Loci in Sibships at High Risk for Hypertension Arterioscler. Thromb. Vasc. Biol., March 1, 2002; 22(3): 418 - 423. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. E. Wagenknecht, D. W. Bowden, J. J. Carr, C. D. Langefeld, B. I. Freedman, and S. S. Rich Familial Aggregation of Coronary Artery Calcium in Families With Type 2 Diabetes Diabetes, April 1, 2001; 50(4): 861 - 866. [Abstract] [Full Text]
|
|
|
|
 |
|
 J. M. HAGBERG, K. R. WILUND, and R. E. FERRELL APO E gene and gene-environment effects on plasma lipoprotein-lipid levels Physiol Genomics, December 18, 2000; 4(2): 101 - 108. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||