Clinical Science (2000) 99, 247251 (Printed in Great Britain) 247

Angiotensin-converting enzyme gene

polymorphism and premature
coronary heart disease

Frank M. VAN BOCKXMEER*, Cyril D. S. MAMOTTE, Valerie BURKE

and Roger R. TAYLORR
*Department of Pathology, The University of Western Australia, Perth, Western Australia, Australia, Department of
Biochemistry, Royal Perth Hospital and the West Australian Heart Research Institute, Perth, Western Australia, Australia,
Department of Clinical Immunology and Biochemical Genetics, Royal Perth Hospital and the West Australian Heart Research
Institute, Perth, Western Australia, Australia, Department of Medicine, The University of Western Australia, Perth, Western
Australia, Australia, and RDepartment of Cardiology, Royal Perth Hospital and the West Australian Heart Research Institute,
Perth, Western Australia, Australia

A B S T R A C T

Since the initial report of the association of the deletion/insertion (D/I) polymorphism in the
gene for angiotensin-converting enzyme (ACE) with myocardial infarction (MI), there has been
considerable controversy. Some have found the D allele to be associated with MI, coronary heart
disease (CHD) or other cardiac pathology, while others have not. In the present study 713
consecutive patients, 50 years of age, documented prospectively with angiographic CHD
(
50 % diameter stenosis of at least one coronary artery), with or without MI, were studied,
along with 688 community control subjects, also 50 years of age, selected randomly from the
electoral rolls and without a history of CHD or MI. Genotyping was done by standard methods.
Most of the subjects in both groups were AngloCeltic Caucasians (547 in the CHD group and
642 in the community group), and the report concerns primarily these subjects. ACE genotype
distributions were not different between the Caucasian community control group and the CHD
or the MI subgroups ; the odds ratios and 95 % confidence limits for the CHD group were 0.96
(0.731.27) for the D allele and 1.02 (0.801.31) for D homozygotes ; for the MI group these values
were 1.00 (0.831.20) and 0.99 (0.741.32) respectively. This negative result was supported in
multivariate analysis accounting for conventional risk factors. There was a significant racial
difference in ACE genotypes between Caucasians, Asians and Australian Aborigines in the CHD
group (P 0.001) ; for example, in this group, 158 of 540 (29 %) Caucasians had the DD genotype
compared with eight of 84 (10 %) Aboriginals (P 0.001) and six of 59 (10 %) Asians (P l 0.002).
Failure to account for such racial differences would have led to erroneous conclusions. In
conclusion, we found no evidence that the D/I ACE gene polymorphism plays a role in the
development of CHD or MI at an early age in a Western Australian Caucasian population. While
this result refers uniquely to premature CHD and MI, and could be population specific, it is in
general agreement with recent meta-analysis of the larger previous studies.

Key words : ACE gene, coronary heart disease, myocardial infarction.

Abbreviations : ACE, angiotensin-converting enzyme ; CHD, coronary heart disease ; CI, confidence interval ; D\I, deletion\
insertion ; MI, myocardial infarction.
Correspondence : Professor Roger R. Taylor, Department of Cardiology, Royal Perth Hospital, GPO Box X2213, Perth 6847,
Western Australia, Australia (e-mail heletoey!rph.health.wa.gov.au).

# 2000 The Biochemical Society and the Medical Research Society

248 F. M. van Bockxmeer and others
INTRODUCTION
Controversy surrounds the role of the angiotensin-
converting enzyme (ACE) deletion\insertion (D\I) poly-
morphism in coronary heart disease (CHD) and myo-
cardial infarction (MI). Since the D allele was related to
MI in the ECTIM study [1], there have been con-
firmatory, negative and even contrary (implicating the I
allele) reports, as reviewed by Samani et al. [2] and more
recently by Keavney et al. [3]. In the present study we
have documented prospectively over 6 years patients
aged less than 50 years, with angiographic CHD with or
without MI. A large, randomly recruited community
group of similar age was also studied.
METHODS
Study subjects
A total of 713 subjects of age 50 years (meanp
S.E.M. 43.6p0.2 years) with CHD and 688 com-
munity control subjects, of similar age, were studied.
However, this report concentrates on the 547 Cau-
casian patients in the former group and the 642
Caucasian subjects in the latter. The study was approved
by the Ethics Committee of the hospital, and all sub-
jects gave informed consent.
The CHD subjects presenting to the hospital, over
approximately 6 years, were documented prospectively
for inclusion in the study and for risk factors for CHD
and genetic analysis. Patients were required to have at
least one obstruction of a major coronary artery ( 50 %
diameter) at angiography, either with (451 of 713 ; 63 %)
or without (37 %) current or previous MI (based on
historical, electrocardiographic and cardiac enzyme
documentation).
The community group were subjects between 30 and
50 years of age (40p0.2 years) who were selected at
random from the electoral roll and took part in a survey
of CHD risk factors in 1994. Each participant completed
a questionnaire concerning lifestyle and medical history ;
their blood pressure, height and weight were recorded,
and blood was taken for DNA extraction and genetic
analysis. Only those without a history suggestive of
CHD were included in the study ; three subjects with
such a history were excluded from this group.
The patient group, with premature CHD, was com-
prised predominantly of males (86 %), while equal
numbers of males and females were recruited in the
community group. Racial background was documented
accurately in the CHD group : the great majority of
patients [547 of 713 (77 %)] were Caucasian, 12 % were
of Australian Aboriginal descent and 8 % were Asian.
Australian Aboriginals (n l 84) and Asians (n l 59) had
a significantly higher II genotype frequency than
Caucasians (n l 540) (51 % and 37 % compared with
# 2000 The Biochemical Society and the Medical Research Society
22 % respectively ; P 0.01 for each) and a significantly
lower DD frequency (10 % and 10 % compared with
29 % ; P 0.001 and P l 0.002 respectively). Likewise,
in the community group, Caucasians and Asian-born
subjects had significantly different genotypes [Caucasian :
II, 21 % ; ID, 49 % ; DD, 29 % (n l 634) ; Asian : II, 45 % ;
ID, 38 %, DD, 17 % (n l 29) ; P l 0.011], while there
were no Aboriginals in this group.
Because of these racial differences, and the lack of
adequate numbers of controls for other racial groups, the
results for Caucasians only with regard to the relationship
between ACE genotype and CHD are presented. ACE
genotype did not differ between the sexes, and results
concerning the influence of ACE genotype on CHD and
MI were analysed regardless of gender.
Genotyping
Genomic DNA was extracted from EDTA-anti-
coagulated blood by a standard Triton X-100 procedure.
Genotyping for the 287 bp D\I polymorphism in intron
16 of the ACE gene was carried out, with precautions
against mistyping including the use of 63 mC as the
annealing temperature and the use of PAGE (12 % total ;
3.3 % cross-linker), rather than agarose-gel electro-
phoresis, for better resolution of the I and D products, as
described previously [4].
Statistical methods
Variables are quoted as meanspS.E.M. Analysis of data
was carried out with SPSS for Windows (SPSS Inc.,
Chicago, IL, U.S.A.). # tests were used to examine
categorical variables in cross-tabulations. Generalized
linear models were used for comparison of means, with
adjustment for co-variates as required. Relationships
with continuous dependent variables were examined in
multiple linear regression with logistic models for di-
chotomous dependent variables. A P value of 0.05 was
considered significant.
RESULTS
Table 1 presents conventional risk factors for the
Caucasian subjects in the CHD and community groups.
History of hypertension and of diabetes, current or pre-
vious smoking and body mass index all differed signifi-
cantly between the groups (P 0.001). The lipid values
quoted in Table 1, like the other variables, are for
males and females together ; however, since the value
for each lipid variable in males was more adverse than in
females, generalized linear models with adjustment for
sex were used to compare groups. Each of the lipid
variables was significantly different between groups (P
0.001), except for low-density lipoprotein cholesterol.
The latter was also not significantly different between
groups when males and females were analysed separately.
 Angiotensin-converting enzyme and coronary disease 249

Table 1 Characteristics of Caucasian CHD and community Table 3 Odds ratios for ACE genotypes and conventional risk
(control) groups factors in Caucasians with CHD and with MI compared with
Abbreviations : BMI, body mass index ; HDL, high-density lipoprotein ; LDL, low- the community group
density lipoprotein. Continuous variables are expressed as meanspS.E.M., except Abbreviations : BMI, body mass index ; HDL, high-density lipoprotein ; LDL, low-
for triacylglycerols, which are expressed as the geometric mean and 95 % density lipoprotein ; NS, not significant. Values in parentheses are 95 % CI. Odds
confidence limits. For binary variables, percentages are given in parentheses. Each ratios for continuous variables refer to a change of 1 kg/m2 for BMI, 1 year for
of the risk factor variables was significantly different between groups (P age and 1 mol/l for each of the lipid variables.
0.001), except for LDL cholesterol. The lipid values quoted are for males and
Factor CHD group MI subgroup
females together, but the values were compared between groups with adjustment
for sex. (a) Univariate models
ACE
CHD group Control group
D allele 0.96 (0.73, 1.27) 1.00 (0.83, 1.20)
Parameter (n l 547) (n l 642)
DD versus II 0.98 (0.71, 1.36) 0.99 (0.68, 1.43)
Age (years) 43.9p0.2 39.9p0.2 DD versus IDjII 1.02 (0.80, 1.31) 0.99 (0.74, 1.32)
Males 478 (87 %) 329 (51 %) Hypertension (history) 2.47 (1.88, 3.24) 2.20 (1.62, 3.00)
MI Diabetes (history) 12.56 (6.02, 26.19) 10.15 (4.70, 21.91)
Previous 167 (31 %) Smoking (current or ex) 3.64 (2.80, 4.72) 4.29 (3.12, 5.90)
Current 164 (30 %) BMI 1.81 (1.14, 1.22) 1.17 (1.13, 1.21)
Previous and current 18 (3 %) Total cholesterol 1.30 (1.17, 1.44) 1.24 (1.10, 1.40)
Diabetic 76 (14 %) 8 (1 %) HDL cholesterol 0.02 (0.01, 0.04) 0.02 (0.01, 0.03)
Smoking LDL cholesterol 1.26 (1.11, 1.42) 1.19 (1.03, 1.37)
Current 192 (35 %) 160 (25 %) Triacylglycerols 2.57 (2.21, 2.99) 2.40 (2.05, 2.82)
Ex 252 (46 %) 186 (29 %) (b) Multivariate model : forward stepwise regression
History of hypertension 181 (33 %) 107 (17 %) Sex 3.2 (2.53, 5.43) 3.82 (2.41, 6.06)
BMI (kg/m2) 28.4p0.2 25.3p0.2 Age 1.15 (1.12, 1.19) 1.14 (1.10, 1.18)
Total cholesterol (mmol/l) 5.46p0.04 5.29p0.05 Hypertension 1.45 (1.03, 2.09) NS
HDL cholesterol (mmol/l) 1.03p0.01 1.31p0.01 Diabetes (history) 9.67 (3.80, 24.61) 10.33 (3.60, 29.62)
LDL cholesterol (mmol/l) 3.42p0.04 3.39p0.04 Smoking (current or ex) 2.98 (2.13, 4.18) 3.53 (2.57, 5.26)
Triacylglycerols (mmol/l) 1.88 (1.80, 1.97) 1.06 (1.02, 1.11) BMI 1.07 (1.03, 1.11) 1.07 (1.03, 1.12)
HDL cholesterol 0.08 (0.04, 0.14) 0.05 (0.03, 0.11)

Table 2 presents the ACE genotypes of the Caucasian

subjects in the total CHD group, the MI subgroup
and the community group ; there were no significant
differences in the distribution of genotypes.
Table 3 presents the odds ratios and 95 % confidence
interval (CI) when comparing the total CHD group and
the MI subgroup with the community group for ACE
genotype and conventional risk factors. By univariate
logistic regression analysis, the odds ratio for the D allele
was 0.96 (95 % CI 0.731.27) for the total CHD group
and 1.00 (0.831.20) for the MI subgroup, and for the DD
genotype these values were 1.02 (0.801.31) for the total
CHD group and 0.99 (0.741.32) for the MI subgroup.
All variables significant in univariate logistic models were
used in forward (likelihood ratio) stepwise regression.
The final model is shown in the lower section of Table 3.
In this analysis high-density lipoprotein cholesterol was
the only lipid variable contributing independently to the
prediction of CHD or MI ; the other lipid variables and
ACE genotype functions failed to improve the model
with either CHD or MI as the dependent variable. Two
subgroups in whom a genetic influence might be expected
to play a particular role were analysed separately by
logistic regression, namely those who had never smoked
and those with a family history of CHD or MI in a first-
degree relative occurring at age 60 years. In neither
subgroup was there a significant relationship between

Table 2 ACE genotypes in Caucasian CHD, MI and control groups

Percentages are given in parentheses.

CHD patients (Caucasians) Community Caucasian group

ACE genotype Total group (n l 540) MI subgroup (n l 349) Total group (n l 634)
II 120 (22 %) 75 (21 %) 135 (21 %)
ID 262 (49 %) 173 (50 %) 313 (49 %)
DD 158 (29 %) 101 (29 %) 186 (29 %)

# 2000 The Biochemical Society and the Medical Research Society

250 F. M. van Bockxmeer and others
ACE genotype and CHD or MI. Notably, the results
were quite different if the analysis was not restricted to
Caucasians. Inclusion of Australian Aboriginals and
Asians, because the former were only represented in the
patient group and because of the higher I allele and lower
D allele frequencies in these ethnic groups, resulted in a
significant over-representation of the II compared with
the DD genotype in both the total CHD group (odds
ratio 1.40 ; 95 % CI 1.101.88) and the MI subgroup
(odds ratio 1.40 ; 95 % CI 1.101.95).
DISCUSSION
The main findings of the present study are that the ACE
D\I allele frequencies in Caucasians were almost identical
in young subjects with CHD or MI and in a randomly
recruited community group. Our study is among the
largest of the individual studies on the effect of ACE
genotype on CHD and MI. It is also unique in con-
centrating strictly on those less than 50 years of age, since
we have been particularly interested in seeking deter-
minants of premature CHD and MI.
The frequency of the D allele of the ACE gene in our
community Caucasian subjects, who are of U.K. and
continental European background, was 54 %, which is
the same as the overall frequency in the Caucasian-based
studies reviewed by Samani et al. [2] and in the CHD
cases and controls in the recently reported results from
the large ISIS study [3]. We also found very similar D\I
genotype frequencies in Caucasians with and without
CHD or MI. In the meta-analysis of 15 studies to 1995 by
Samani et al. [2], the DD genotype was found to confer an
odds ratio for MI of 1.26 (95 % CI 1.151.39). The largest
individual study was published more recently, concern-
ing MI in the ISIS study [3]. The odds ratio for the DD
compared with other genotypes was 1.10 (95 % CI
1.001.21), and that for D homozygotes and hetero-
zygotes compared with I homozygotes was 0.97 (95 %
CI 0.871.08). Keavney et al. [3] also analysed the
combined results from 35 smaller studies with less than
200 cases of MI and from the 14 larger studies with more
cases. While the relationship between the DD genotype
and MI was significant (odds ratio 1.57 ; 99 % CI
1.381.78) in the smaller studies, it was not in the larger
studies (odds ratio 0.99 ; 99 % CI 0.901.08). This
evidence of publication bias, suggesting that small nega-
tive studies tend not to be published, was similarly
documented in the earlier meta-analysis [2]. Between
these two meta-analyses [2,3], mixed results continued to
be published, with negative results concerning MI from
studies in North Karelian Finns [5], Austrians [6], for all
patients in a study of Germans [7], and for CHD patients
in Welsh subjects [8]. In a subsequent study from the
German group [9], based on a large number of patients
coming to diagnostic coronary angiography, the D allele
# 2000 The Biochemical Society and the Medical Research Society
was related to the extent of coronary artery disease,
although not to MI, in young subjects only ( 61.7
years). Two of the above studies [7,8] reported positive
results in selected low-risk subgroups, as noted in the
original ECTIM report [1], but the variable and rather
arbitrary nature of subgrouping could be questioned.
Low-risk subgroups were examined in the substantial
number of MI cases and controls in the ISIS study [3],
with negative results, leading to the conclusion that most
of the positive results with low-risk subgroups were
probably spurious.
Because current or past cigarette smoking was such a
dominant risk factor in our study, being present in 80 %
of CHD patients, we examined only non-smokers as a
low-risk group, but found no effect of the ACE genotype.
However, there may be a real effect of the ACE geno-
type in some particular groups. For example, a recent
study examined the effect in a high-risk group (213 sub-
jects with heterozygous familial hypercholesterolaemia
or familial defective apolipoprotein B100) of average age
57 years, and found an odds ratio for the DD genotype of
2.5 for MI and 2.2 for CHD in males [10].
The role of the D allele in contributing to MI or CHD
may also depend upon the frequency of the D allele in the
population and upon other racial aspects ; the odds ratio
for the DD genotype and MI was higher in the Japanese
than in the Caucasian studies reviewed by Samani et al.
[2]. Race is an important consideration in studies of the
ACE genotype, as emphasized by others [2,11] and
illustrated in the present study. There was a lower D
allele frequency in our Asian compared with our
Caucasian subjects (36 % and 54 % respectively), in line
with the values of 39 % found in the three Japanese
studies reviewed by Samani et al. [2], 36 % reported for
Han Chinese [12] and 30 % for Singaporean Chinese [13]
and Thais [14]. In our study, the D allele frequency in
Australian Aboriginals was also low, at 29 %. If our data
had been analysed regardless of race, a fallacious con-
clusion would have been reached.
Another important consideration is the nature of the
recruitment of both cases and controls. Our documen-
tation of cases was carried out prospectively, at the time
of hospital presentation, specifically for the purpose of
risk factor and genetic studies. Recruitment of controls
varies between studies and is particularly contentious ; we
consider that our use of a randomly selected community
group is the optimum. Admittedly, some controls , even
aged under 50 years, might have covert CHD. A control
group with no or minimal angiographic coronary
obstructive disease at angiography has been used in
numerous studies. Obviously such patients have usually
had coronary angiography because of symptoms, and a
substantial number are likely to have a pathological or
functionally abnormal coronary circulation.
Despite our negative result and the recent conclusion,
from the large ISIS study and meta-analysis of the larger

earlier studies, that ACE genotype plays little, if any, role
in MI [3], the greater frequency of the D allele or the DD
genotype reported in children with parents [15] or
grandparents [16] with a history of MI or CHD
strengthens the concept that there is a real relationship
between these factors in some populations. If so, the
commonly studied D\I polymorphism may not itself be
responsible, but instead one or more of the many
polymorphisms in the ACE gene that have been de-
scribed recently [17].
In conclusion, we did not find a relationship between
the ACE genotype and CHD or MI in our Caucasian
population. There probably is a relationship in some
populations, but the nature of the CHD or MI group
studied, the method of recruitment (especially of the
control group) and racial differences are important
considerations.
ACKNOWLEDGMENTS
This study was supported by the National Health and
Medical Research Council (Australia), the Medical Re-
search Fund of Western Australia and the Royal Perth
Hospital Medical Research Foundation.
REFERENCES
1 Cambien, F., Poirier, O., Lecerf, L. et al. (1992) Deletion
polymorphism in the gene for angiotensin-converting
enzyme is a potent risk factor for myocardial infarction.
Nature (London) 359, 641644
2 Samani, N. H., Thompson, J. R., OToole, L., Channer, K.
and Woods, K. L. (1996) A meta-analysis of the
association of the deletion allele of the angiotensin-
converting enzyme gene with myocardial infarction.
Circulation 94, 708712
3 Keavney, B., McKenzie, C., Parish, S. et al., for the
International Studies of Infarct Survival (ISIS)
Collaborators. (2000) Large-scale test of hypothesised
associations between the angiotensin-converting-enzyme
insertion\deletion polymorphism and myocardial
infarction in about 5000 cases and 6000 controls. Lancet
355, 434442
4 van Bockxmeer, F. M., Mamotte, C. D. S., Gibbons, F. A.,
Burke, V. and Taylor, R. R. (1995) Angiotensin-converting
enzyme and apolipoprotein E genotypes and restenosis
after coronary angioplasty. Circulation 92, 20662071
Received 17 January 2000/10 April 2000; accepted 30 May 2000
Angiotensin-converting enzyme and coronary disease 251
5 Miettinen, H. E., Korpela, K., Hamalainen, L. and
Kontula, K. (1994) Polymorphisms of the apolipoprotein
and angiotensin converting enzyme genes in young North
Karelian patients with coronary heart disease. Hum.
Genet. 94, 189192
6 Friedl, W., Krempler, F., Paulweber, B., Pichler, M. and
Sandhofer, F. (1995) A deletion polymorphism in the
angiotensin converting enzyme gene is not associated with
coronary heart disease in an Austrian population.
Atherosclerosis 112, 137143
7 Gardemann, A., Weiss, T., Schwartz, O. et al. (1995) Gene
polymorphism but not catalytic activity of angiotensin
I-converting enzyme is associated with coronary artery
disease and myocardial infarction in low-risk patients.
Circulation 92, 27962799
8 Mattu, R. J., Needham, E. W. A., Galton, D. J., Frangos,
E., Clark, A. J. L. and Caulfield, M. (1995) A DNA variant
at the angiotensin-converting enzyme gene locus associates
with coronary artery disease in the Caerphilly Heart
Study. Circulation 91, 270274
9 Gardemann, A., Fink, M., Stricker, J. et al. (1998) ACE
I\D gene polymorphism : presence of the ACE D allele
increases the risk of coronary artery disease in younger
individuals. Atherosclerosis 139, 153159
10 OMalley, J. P., Maslen, C. L. and Illingworth, R. (1998)
Angiotensin-converting enzyme DD genotype and
cardiovascular disease in heterozygous familial
hypercholesterolemia. Circulation 97, 17801783
11 Barley, J., Blackwood, A., Carter, N. D. et al. (1994)
Angiotensin converting enzyme insertion\deletion
polymorphism : association with ethnic origin.
J. Hypertens. 12, 955957
12 Chuang, L. M., Chiu, K. C., Chiang, F. T. et al. (1997)
Insertion\deletion polymorphism of the angiotensin
I-converting enzyme gene in patients with hypertension,
non-insulin-dependent diabetes mellitus, and coronary
heart disease in Taiwan. Metab. Clin. Exp. 46, 12111214
13 Lee, E. J. (1994) Population genetics of the angiotensin-
converting enzyme in Chinese. Br. J. Clin. Pharmacol. 37,
212214
14 Nitiyanant, W., Sriussadaporn, S., Ploybutr, S.,
Watanakejorn, P., Tunlakit, M. and Bejrachandra, S. (1997)
Angiotensin converting enzyme gene polymorphism in
healthy Thais and patients with non-insulin dependent
diabetes mellitus. J. Med. Assoc. Thailand 80, 747752
15 Tiret, L., Kee, F., Poirier, O. et al. (1993) Deletion
polymorphism in angiotensin-converting enzyme gene
associated with parental history of myocardial infarction.
Lancet 341, 991992
16 Wang, X. L., McCredie, R. M. and Wilcken, D. E. L.
(1996) Genotype distribution of angiotensin-converting
enzyme polymorphism in Australian healthy and coronary
populations and relevance to myocardial infarction and
coronary artery disease. Arterioscler. Thromb. Vasc. Biol.
16, 115119
17 Rieder, M. J., Taylor, S. L., Clark, A. G. and Nickerson,
D. A. (1999) Sequence variation in the human angiotensin
converting enzyme. Nat. Genet. 22, 5962
# 2000 The Biochemical Society and the Medical Research Society