Larsen CT, Dahlin J, Blackburn H, et al. Prevalence and prognosis of electrocardiographic left ventricular hypertrophy, ST segment depression and negative T wave; the Copenhagen City Heart Study. Eur Heart J 2002;23:315-24

ArticleinEuropean Heart Journal 23(4):315-24 · March 2002with112 Reads
Impact Factor: 15.20 · DOI: 10.1053/euhj.2001.2774 · Source: PubMed
Abstract

To evaluate the prevalence and the independent prognosis of electrocardiographic left ventricular hypertrophy by voltage only, ST depression and negative T wave, isolated negative T wave and left ventricular hypertrophy plus ST depression and negative T wave for cardiac morbidity and mortality, without known ischaemic heart disease at baseline. METHODS and Follow-up data from the Copenhagen City Heart Study were used. Subjects were 5243 men and 6391 women, age range 25-74 years. End-points were (1) myocardial infarction, (2) ischaemic heart disease and (3) cardiovascular disease mortality. Relative risk was age- and sex-adjusted, and multivariately adjusted for known cardiovascular risk factors. During 7 years follow-up, left ventricular hypertrophy plus ST depression and negative T wave had an age-adjusted relative risk of 3.78 (95% confidence interval 2.29-6.25) for myocardial infarction, 4.27 (2.95-6.16) for ischaemic heart disease and 3.75 (2.41-5.85) for cardiovascular disease. A negative T wave, ST depression and negative T wave changes, and left ventricular hypertrophy with negative T wave also carry independent prognostic information for myocardial infarction, ischaemic heart disease and cardiovascular disease. Electrocardiographic left ventricular hypertrophy with ST depression and negative T wave changes are the electrocardiographic abnormalities with the greatest prognostic information for future cardiac events. Electrocardiographic negative T waves, ST depression and negative T wave abnormalities and left ventricular hypertrophy with negative T waves, also have prognostic information.

Full-text

Available from: Peter Schnohr
European Heart Journal (2002) 23, 315–324
doi:10.1053/euhj.2001.2774, available online at http://www.idealibrary.com on
Prevalence and prognosis of electrocardiographic left
ventricular hypertrophy, ST segment depression and
negative T-wave
The Copenhagen City Heart Study
C. T. Larsen
1,2
, J. Dahlin
1,3
, H. Blackburn
4
, H. Scharling
1
, M. Appleyard
1
,
B. Sigurd
1,3
and P. Schnohr
1
1
Copenhagen City Heart Study, Epidemiological Research Unit, Bispebjerg University Hospital, Copenhagen,
Bispebjerg, 2400 NV, Denmark;
2
Department of Cardiology, Gentofte University Hospital, 2900 Hellerup,
Denmark;
3
Department of Cardiology, Bispebjerg University Hospital, Copenhagen, 2400 NV, Denmark;
4
Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, MN, U.S.A.
Aims To evaluate the prevalence and the independent
prognosis of electrocardiographic left ventricular hyper-
trophy by voltage only, ST depression and negative T wave,
isolated negative T wave and left ventricular hypertrophy
plus ST depression and negative T wave for cardiac mor-
bidity and mortality, without known ischaemic heart
disease at baseline.
Methods and Results Follow-up data from the
Copenhagen City Heart Study were used. Subjects were
5243 men and 6391 women, age range 25–74 years. End-
points were (1) myocardial infarction, (2) ischaemic heart
disease and (3) cardiovascular disease mortality. Relative
risk was age- and sex-adjusted, and multivariately adjusted
for known cardiovascular risk factors. During 7 years
follow-up, left ventricular hypertrophy plus ST depression
and negative T wave had an age-adjusted relative risk of
3·78 (95% confidence interval 2·29–6·25) for myocardial
infarction, 4·27 (2·95–6·16) for ischaemic heart disease and
3·75 (2·41–5·85) for cardiovascular disease. A negative T
wave, ST depression and negative T wave changes, and left
ventricular hypertrophy with negative T wave also carry
independent prognostic information for myocardial infarc-
tion, ischaemic heart disease and cardiovascular disease.
Conclusions Electrocardiographic left ventricular hyper-
trophy with ST depression and negative T wave changes are
the electrocardiographic abnormalities with the greatest
prognostic information for future cardiac events. Electro-
cardiographic negative T waves, ST depression and nega-
tive T wave abnormalities and left ventricular hypertrophy
with negative T waves, also have prognostic information.
(Eur Heart J 2002; 23: 315–324, doi:10.1053/euhj.2001.2774)
2001 The European Society of Cardiology
Key Words: Electrocardiographic, left ventricular
hypertrophy, prognosis, epidemiology, risk factors.
See page 268, doi: 10.1053/euhj.2001.2849 for the Editorial
comment on this article
Introduction
The resting electrocardiogram (ECG) is a simple, non-
invasive method, useful both for diagnosing heart dis-
ease and as a screening tool. ECG manifestations of left
ventricular hypertrophy include an increase in QRS
voltage. The exact mechanism of this increase is not
clear, but in addition to heart muscle mass other factors
such as intracavitary blood volume, proximity to the
chest wall, conduction properties, location of the heart
within the thorax, and intraventricular and transmural
pressures may play a role
[1]
. ECG manifestations of left
ventricular hypertrophy are often seen in association
with repolarization abnormalities with ST segment
depression and flat or negative T waves; this is called
an ECG left ventricular hypertrophy strain pattern.
ECG left ventricular hypertrophy is an ominous
harbinger of impending cardiovascular catastrophes in
Revision submitted 24 April 2001, accepted 25 April 2001, and
published online 3 August 2001.
Correspondence: Carsten Toftager Larsen, MD, Department of
Cardiology, Copenhagen University Hospital, Amtssygehuset i
Gentofte, 2900 Hellerup, Denmark.
0195-668X/02/040315+10 $35.00/0 2001 The European Society of Cardiology
Page 1
the hypertensive patient, and the risk associated with
ECG left ventricular hypertrophy is particularly great
when repolarization abnormalities are present
[26]
. There
is some information about the prevalence, incidence,
determinants and prognostic value of left ventricular
hypertrophy from epidemiological studies
[26]
. The pres-
ence of a resting ECG negative T wave is associated
with common clinical conditions including myocar-
dial ischaemia and infarction, cardiomyopathy, acute
pulmonary embolism, electrolyte abnormalities and
drugs
[1]
, and the presence of ECG ST segment depres-
sion and negative T waves are documented to be associ-
ated with sudden death
[7,8]
and a predictor of coronary
heart disease
[9,13]
. Some have shown that isoelectric or
small negative T waves are important risk indicators for
coronary heart disease
[11,14]
.
The purpose of the present study was to evaluate the
prevalence and the independent long-term prognosis of
voltage-only ECG left ventricular hypertrophy, isolated
ECG negative T wave, ECG ST depression and negative
T wave, ECG left ventricular hypertrophy with negative
T wave and ECG left ventricular hypertrophy with
ST depression and negative T wave in men and women
free of ischaemic heart disease and not under antihyper-
tensive treatment at baseline.
Methods
Subjects
The Copenhagen City Heart Study is a prospective
cardiovascular population study comprising a random
sample of 19 329 men and women, 20 years or older,
selected from a population of approximately 90 000
residents, living in a defined area of Copenhagen in
1976. At the first examination in 1976–1978, 14 223
subjects attended (response rate 73·6%); 6511 were men
and 7712 were women. Less than 2% of the population
was of non-Caucasian origin, and socio-economically
the majority was middle class. Details of the selection
procedure, a description of the eligible non-participants
and the complete examination programme and informa-
tion on the subjects have been presented elsewhere
[15,16]
.
To evaluate the independent prognostic influence of
ECG changes, we excluded subjects with a previous
history of ischaemic heart disease, self-reported and/or
discharged from a hospital with a diagnosis of ischaemic
heart disease (ICD8: 410-414), as ischaemic heart disease
itself is independently prognostic. The Rose question-
naire was used to evaluate previous history of ischaemic
heart disease
[17]
. For the evaluation of prevalences of
ECG abnormalities in the population, 548 subjects
younger than 25 and older than 74 years of age, and 621
subjects with incomplete data were excluded. Data from
13 054 subjects, 5942 men and 7112 women were used in
the evaluation of prevalences. In the evaluation of the
prognostic information of ECG findings we restricted
the age range to 35–74 years of age, due to a very low
number of events in the youngest ages. Subjects with
ECG Minnesota codes I
1-2
(Q waves), VI
1
(third degree
atrioventricular block), VI
2
(second degree atrioven-
tricular block), VI
4
(Wol–Parkinson–White syndrome,
VI
5
(PQ<0·12 s), VII
1
(left bundle branch block), VII
2
(right bundle branch block), VII
4
(QRS duration
>0·12 s), VIII
2
(ventricular tachycardia), VIII
3
(atrial
fibrillation), VIII
4
(supraventricular tachycardia), VIII
5
(Idioventricular rhythm), and VIII
6
(atrioventricular
nodal rhythm), at baseline were excluded from
the present analysis. Subjects taking antihypertensive
medication were excluded. The Minnesota codes IV
1-3
describing ST depression has an optional code V
describing the T wave. Subjects with isolated ST depres-
sion, code IV
4
were excluded, as were code V
4
describing
non-specific T wave changes. A total of 3241 subjects
were excluded. Thus the study group evaluating the
prognostic information of ECG findings comprises
10 982 subjects; 4940 men and 6042 women.
Follow-up
The National Patient Register provides information
on all hospital admissions in Denmark, and the dis-
charge diagnoses are registered after the World Health
Organization’s International Classification of Diseases
(ICD), ICD8 to the end of 1992, and ICD10 from 1993.
Information on fatal and non-fatal myocardial infarc-
tion (ICD8: 410 and ICD10: I 21), fatal and non-fatal
ischaemic heart disease events (ICD8: 410-414 and
ICD10: I20-I25), cardiovascular disease mortality (ICD8
390-458 and ICD10 I00-I99 and G45) during follow-up
were obtained from the Register of Causes of Death and
the National Patient Register.
Electrocardiogram
A resting supine 12-lead ECG was recorded in
each subject. ECGs were evaluated according to the
Minnesota code
[17]
by two independent technicians, and
in case of disagreement a third person settled the dis-
agreement. The ECG findings were divided into six
mutually exclusive groups: Voltage-only left ventricular
hypertrophy as code III
1,3
, and no other ECG abnor-
mality; negative T wave as code V
1-3
, and no other ECG
abnormality; ST depression and negative T wave
changes as code IV
1-3
, and no other ECG abnormality,
except that according to the Minnesota code there is an
optional code V (T waves) in subjects with ST depres-
sion codes IV
1-3
; left ventricular hypertrophy with nega-
tive T wave as codes III
1,3
and V
1-3
, and no other
ECG abnormality; left ventricular hypertrophy with ST
depression and negative T wave changes as codes III
1,3
,
IV
1-3
, and V
1-3
, and no other ECG abnormality. ECGs
were classified as normal in the absence of Minnesota
codes I
1-2
(Q waves), III
1,3
(ECG left ventricular
hypertrophy), IV
1-4
(ST segment depression), V
1-4
316 C. T. Larsen et al.
Eur Heart J, Vol. 23, issue 4, February 2002
Page 2
(negative/flat T-wave), VI
1
(third degree atrioventricular
block), VI
2
(second degree atrioventricular block), VI
4
(Wol–Parkinson–White syndrome), VI
5
(PQ<0·12 s),
VII
1
(left bundle branch block), VII
2
(right bundle
branch block), VII
4
(QRS duration >0·12 s), VIII
2
(ven-
tricular tachycardia), VIII
3
(atrial fibrillation), VIII
4
(supraventricular tachycardia), VIII
5
(Idioventricular
rhythm), and VIII
6
(AV nodal rhythm).
Covariates
Clinical and biochemical measurements were made in all
participating subjects using standardized and validated
methods
[17]
. Blood pressure was measured in a sitting
position using the left arm, and a London School of
Hygiene sphygmomanometer, after 5 min of rest. Cu
size was adjusted to arm circumference. The body mass
index was calculated as weight (kg) divided by height
squared (m
2
). A non-fasting blood sample included
analysis for plasma total cholesterol and glucose.
Information on smoking behaviour, family history
of ischaemic heart disease, antihypertensive medi-
cation, history of diabetes, alcohol consumption and
physical exercise during leisure time was obtained by
questionnaire.
Statistical methods
Age-standardized prevalence and age-standardized inci-
dence rates were calculated by the method of direct
standardization using the Danish population in year
1977 as reference. Associations between ECG findings
and end-points were analysed by Cox’s proportional
hazards regression model with time since study entry as
the underlying timescale and adjusted by age at study
entry and treated as categorical dummy-variables: (35–
44, 45–54, 55–64, 65–74). In the models, additional
covariates were also treated as categorical dummy-
variables: body mass index (<20, [20–25], 25–
30, d30 kg . m
2
), systolic blood pressure ([<140],
140–160, d160 mmHg), heart rate ([<60], 60–90,
d90 beats . min
1
), total serum cholesterol ([<5], 5–7,
d7mmol.l
1
), smoking status ([no], former, current),
diabetic status ([no], NIDDM/non-fasting glucose
>11mmol.l
1
IDDM), family history of ischaemic
heart disease ([no], yes), alcohol consumption ([no], 1–2,
3–4, d5 beverages a day), and physical exercise during
leisure time ([no], moderate exercise 2–4 h a week,
moderate exercise >4 h a week), where [] indicates the
reference levels.
The reference for sex was female and the reference for
ECG findings was subjects with a normal ECG. Relative
risks (RR) were calculated as the proportional hazard
ratio. The proportional assumptions of the hazards
were checked by log–log survivor plots for each
variable controlling for sex and age. When we tested for
interactions between ECG findings and sex, we had all
the five interactions terms in the Cox models and
considered both the tests for each interaction term df=1
and the test for all the interaction terms df=5. When we
tested for interactions between ECG findings and age
groups and ECG findings and systolic blood pressure
groups we did it the same way but for each sex at a time.
All analyses were performed using SAS (SAS System
for windows, release 6.12, SAS institute Inc., Cary,
NC, U.S.A.) and Stata (StataCorp. 1999. Stata
Statistical Software: Release 6.0. College Station, Stata
Corporation, TX, U.S.A.) software.
Results
General findings
Table 1 displays the prevalence of ECG findings in men
and women in a 10-year age strata, and the age-
standardized prevalences of ECG findings in the popu-
lation. The percentage of subjects with a normal ECG
decreased with age in women, and from age 35 years in
men. The prevalence of voltage-only left ventricular
hypertrophy is high in younger aged men, with a gradual
decrease of prevalence with increasing age. In contrast,
the prevalence of voltage-only left ventricular hyper-
trophy in women increased with age from age 35. The
prevalence of negative T wave, ST depression and
negative T wave changes and left ventricular hyper-
trophy with ST depression and negative T wave
increased with age, and the prevalence of these ECG
findings was low before age 55. The prevalence of
negative T wave and ST depression and negative T wave
changes were similar in both sexes, but left ventricular
hypertrophy with negative T wave, and left ventricular
hypertrophy with ST depression and negative T wave
was about twice as frequent in men as in women.
Baseline characteristics of the study sample, at the start
of the follow-up period are given in Table 2. The
cardiovascular risk profile is typical for Danish men and
women in the 1970s, where there was a high percentage
of smokers. Subjects with negative T wave, ST depres-
sion, left ventricular hypertrophy with negative T wave
and subjects with left ventricular hypertrophy with ST
depression and negative T wave changes were older, had
higher systolic blood pressure, and a higher prevalence
of diabetes compared to those with normal ECG (Table
2). Information on myocardial infarction (n=1127),
ischaemic heart disease events (n=2104) and cardio-
vascular disease mortality (n=1924) were complete until
the end of 1997.
Fatal and non-fatal myocardial infarction
We found that the age standardized incidence per 1000
patients years of fatal and non-fatal myocardial infarc-
tion during 7 and 21 years follow-up were 7·5 and 10·0
in men, and 2·5 and 3·9 in women. To evaluate the
The Copenhagen City Heart Study 317
Eur Heart J, Vol. 23, issue 4, February 2002
Page 3
influence of abnormal ECG findings, we performed a
Cox proportional analysis. There was no interaction
between sex and ECG changes, so we continued the
analyses for both sexes combined. The age standardized
incidence of fatal and non-fatal myocardial infarction in
relation to initial ECG findings is shown in Table 3.
Table 3 also shows the age- and sex-adjusted and the
multivariately adjusted RR for fatal and non-fatal myo-
cardial infarction. In the multivariate Cox proportional
model we adjusted for age, sex, systolic blood pressure,
heart rate, body mass index, plasma cholesterol, smok-
ing, diabetes, alcohol, physical exercise and family
history of ischaemic heart disease. ECG findings with
a significantly independent association with fatal and
Table 1 Prevalence of ECG findings in the age range 25–74 years stratified by age and sex
Normal
Voltage-only
LVH
T ST/T
LVH with
neg. T
LVH with
ST/T
Excluded subjects*
n % n % n%n%n%n% n %
Men
25–34 (n=303) 193 62·1 95 30·5 7 2·3 3 1·0 3 1·0 2 0·6 8 2·6
35–44 (n=1167) 866 71·9 238 19·8 45 3·7 4 0·3 9 0·7 5 0·4 37 3·1
45–54 (n=1477) 1095 68·1 250 15·6 93 5·8 13 0·8 14 0·9 12 0·7 129 8·0
55–64 (n=1709) 1086 59·2 235 12·8 122 6·7 36 2·0 42 2·3 25 1·4 288 15·7
65–74 (n=750) 460 46·6 86 8·7 116 11·8 26 2·6 27 2·7 35 3·5 237 24·0
Age-standardized 3700 62·6 904 19·3 383 5·3 82 1·2 95 1·4 79 1·1 699 9·0
Women
25–34 (n=349) 320 90·1 18 5·1 9 2·5 2 0·6 0 0 0 0 6 1·7
35–44 (n=1302) 1171 87·2 62 4·6 42 3·1 23 1·7 3 0·2 1 0·1 41 3·1
45–54 (n=1963) 1671 78·5 121 5·7 102 4·8 49 2·3 6 0·3 14 0·7 165 7·8
55–64 (n=2029) 1605 69·5 134 5·8 176 7·6 72 3·1 18 0·8 24 1·0 281 12·2
65–74 (n=748) 506 51·8 71 7·3 103 10·5 34 3·5 18 1·8 16 1·6 228 23·4
Age-standardized 5273 77·3 406 5·6 432 5·3 180 2·2 45 0·5 55 0·6 721 8·6
ECG=electrocardiographic; LVH=left ventricular hypertrophy; ST/T=ST depression and negative T wave; LVH with neg. T=left
ventricular hypertrophy with negative T wave; LVH with ST/T=left ventricular hypertrophy with ST depression and negative T wave.
*Excluded subjects include those with other ECG abnormalities, those with a previous history of ischaemic heart disease, and those taking
antihypertensive medication.
Table 2 Characteristics of subjects in the age range 35–74 years in the dierent ECG groups without ischaemic heart
disease at baseline
Normal
Voltage-only
LVH
T ST/T
LVH with
neg. T
LVH with
ST/T
Men n=3507 n=809 n=376 n=79 n=92 n=77
Age (years) 53·19·6 51·89·8 58·19·6 60·38·3 59·39·2 61·59·5
Systolic blood pressure (mmHg) 136·918·0 141·520·9 145·419·8 146·922·3 151·221·7 161·827·3
Diastolic blood pressure (mmHg) 84·011·5 86·112·5 87·113·0 84·914·4 91·112·7 91·315·8
Heart rate (beats . min
1
) 74·013·2 71·812·2 77·514·0 80·515·0 76·412·4 77·514·2
BMI (kg . m
2
) 25·83·6 25·33·1 27·24·1 25·93·5 25·63·1 25·63·0
Cholesterol (mmol . l
1
) 6·01·1 5·91·1 6·1 1·2 6·31·0 6·11·2 6·11·1
Smoking (%) 73·4 67·4 68·9 60·8 69·6 64·9
Diabetes (%) 3·4 2·7 6·7 10·1 5·4 9·1
Women n=4953 n=388 n=423 n=178 n=45 n=55
Age (years) 52·79·2 55·09·5 58·18·6 56·48·8 61·79·0 59·87·2
Systolic blood pressure (mmHg) 131·619·4 143·924·5 141·522·5 142·921·6 155·330·3 166·329·3
Diastolic blood pressure (mmHg) 80·410·8 84·912·7 84·211·4 83·511·9 90·415·0 94·115·0
Heart rate (beats . min
1
) 74·012·2 73·612·3 77·414·4 79·014·7 81·816·8 79·614·2
BMI (kg . m
2
) 24·54·2 23·94·2 26·35·0 24·93·9 24·94·1 24·14·2
Cholesterol (mmol . l
1
) 6·31·4 6·21·2 6·7 1·3 6·91·4 6·51·2 6·41·2
Smoking (%) 60·0 58·8 55·6 51·7 53·3 58·2
Diabetes (%) 1·2 2·1 2·6 3·4 2·2 1·8
IHD=ischaemic heart disease; BMI=body mass index; ECG=electrocardiographic; LVH=left ventricular hypertrophy; ST/T=ECG ST
depression and negative T wave; T-ECG=negative T wave; LVH with neg. T=ECG left ventricular hypertrophy with negative T wave;
LVH with ST/T=ECG left ventricular hypertrophy with ST depression and negative T wave.
Values given as meanSD for continuous variables.
318 C. T. Larsen et al.
Eur Heart J, Vol. 23, issue 4, February 2002
Page 4
non-fatal myocardial infarction were isolated negative
T wave, left ventricular hypertrophy with negative T
wave, and left ventricular hypertrophy with ST depres-
sion and negative T wave during 7 years follow-up, and
isolated negative T wave, ST depression and negative T
wave changes, left ventricular hypertrophy with negative
T wave, and left ventricular hypertrophy with ST de-
pression and negative T wave during 21 years follow-up.
During 7 years follow-up, there was interaction between
age and voltage-only left ventricular hypertrophy for
men (ratio 4·48 (95% confidence interval: 1·76–11·43,
P=0·002); this ECG finding was free of excess risk in
men below 55 years of age (RR 0·43 (0·19–1·00,
P=0·05), but predictive in men above 55 years (RR 1·94
(1·28–2·93, P=0·002).
Ischaemic heart disease events
We found that the age standardized incidence per 1000
patient years of fatal and non-fatal ischaemic heart
disease events during 7 and 21 years follow-up were 12·6
and 18·2 in men, and 5·1 and 9·2 in women, but there
was no significant interaction between sex and ECG
findings. Therefore we performed the analyses for both
sexes combined (Table 4). Voltage-only left ventricular
hypertrophy was not associated with excess future
ischaemic heart disease events. A negative T wave, ST
depression and negative T wave changes, left ventricular
hypertrophy with negative T wave and left ventricular
hypertrophy with ST depression and negative T wave
had significantly independent associations with fatal and
Table 3 Age-standardized incidence of fatal and non-fatal myocardial infarction, age- and sex-adjusted relative risk
and multivariately adjusted relative risk for fatal and non-fatal myocardial infarction during 7 years and 21 years
follow-up in relation to ECG findings among those without ischaemic heart disease at baseline, for the age range
35–74 years
ECG
7 years follow-up (number of MI=322) 21 years follow-up (number of MI=1127)
n (Inc.)
Age-adjusted RR
(95% CI)
RR* (95% CI) n (Inc.)
Age-adjusted RR
(95% CI)
RR* (95%CI)
Normal ECG, n=8460 203 (3·8) 1 1 760 (5·6) 1 1
Voltage-only LVH, n=1197 41 (6·3) 1·18 (0·84–1·66) 1·16 (0·82–1·63) 156 (8·9) 1·21 (1·02–1·45)† 1·20 (1·00–1·43)†
T, n=799 40 (6·6) 1·60 (1·13–2·25)‡ 1·42 (1·00–2·01)† 116 (8·3) 1·51 (1·24–1·84)§ 1·30 (1·07–1·59)‡
ST/T, n=257 9 (4·3) 1·28 (0·66–2·51) 1·15 (0·59–2·26) 39 (8·9) 1·82 (1·32–2·52)§ 1·68 (1·21–2·33)‡
LVH with neg. T, n=137 12 (9·0) 2·30 (1·27–4·13)‡ 2·12 (1·17–3·82)‡ 26 (11·6) 1·66 (1·12–2·45)† 1·56 (1·05–2·31)†
LVH with ST/T, n=132 17 (15·1) 3·78 (2·29–6·25)§ 2·83 (1·68–4·76)‡ 30 (12·9) 2·43 (1·68–3·51)§ 1·86 (1·27–2·71)‡
CI=confidence interval; RR=relative risk; MI=myocardial infarction; IHD=ischaemic heart disease; ECG=electrocardiographic;
LVH=left ventricular hypertrophy; ST/T=ST depression and negative T wave; T=negative T wave; LVH with neg. T=left ventricular
hypertrophy with neg. T wave; LVH with ST/T=left ventricular hypertrophy with ST depression and negative T wave; Inc.=age-
standardized number of myocardial infarction per 1000 years of follow-up.
*Multivariate Cox regression with adjustments for all risk factors shown in Table 2 and alcohol, physical exercise, and family history of
ischaemic heart disease.
P<0·05; P<0·01; §P<0·001.
Table 4 Age-standardized incidence of ischaemic heart disease events, age- and sex-adjusted relative risk and
multivariately adjusted relative risk for ischaemic heart disease events during 7 years and 21 years follow-up in relation
to ECG findings among those without ischaemic heart disease at baseline, for the age range 35–74 years
ECG
7 years follow-up (number of IHD=560) 21 years follow-up (number of IHD=2104)
n (Inc.)
Age-adjusted RR
(95% CI)
RR* (95% CI) n (Inc.)
Age-adjusted RR
(95% CI)
RR* (95% CI)
Normal ECG, n=8460 345 (6·7) 1 1 1·431 (11·3) 1 1
Voltage-only LVH, n=1197 66 (10·2) 1·16 (0·89–1·52) 1·15 (0·88–1·51) 253 (15·0) 1·10 (0·96–1·26) 1·10 (0·96–1·26)
T, n=799 73 (11·6) 1·69 (1·31–2·18)§ 1·56 (1·21–2·03)‡ 238 (18·2) 1·62 (1·41–1·86)§ 1·45 (1·26–1·67)§
ST/T, n=257 27 (14·1) 2·25 (1·52–3·34)§ 2·07 (1·39–3·08)§ 81 (19·3) 1·96 (1·57–2·46)§ 1·81 (1·44–2·27)§
LVH with neg. T, n=137 17 (16·2) 1·95 (1·20–3·19)‡ 1·89 (1·15–3·09)† 42 (20·9) 1·46 (1·07–1·99)† 1·37 (1·00–1·87)†
LVH with ST/T, n=132 32 (30·5) 4·27 (2·95–6·16)§ 3·62 (2·47–5·30)§ 59 (28·0) 2·62 (2·01–3·40)§ 2·17 (1·66–2·83)§
CI=confidence interval; RR=relative risk; IHD=ischaemic heart disease; ECG=electrocardiographic; LVH=left ventricular hyper-
trophy; ST/T=ST depression and negative T wave; T=negative T wave; LVH with neg. T=left ventricular hypertrophy with negative T
wave; LVH with ST/T=left ventricular hypertrophy with ST depression and negative T wave; Inc.=age-standardized number of ischaemic
heart disease per 1000 years of follow-up.
*Multivariate Cox regression with adjustments for all risk factors shown in Table 2 and alcohol, physical exercise, and family history of
ischaemic heart disease.
P<0·05; P<0·01; §P<0·001.
The Copenhagen City Heart Study 319
Eur Heart J, Vol. 23, issue 4, February 2002
Page 5
non-fatal ischaemic heart disease during 7 and 21 years
follow-up. There was a higher predictive value for
ischaemic heart disease in the ECG findings after 7 years
of follow-up than after 21 years. During 7 years
follow-up there was interaction between age and
voltage-only left ventricular hypertrophy for men (ratio
2·69 (1·35–5·39, P=0·005); this ECG finding was free of
excess risk in men below 55 years of age (RR 0·59
(0·32–1·08, P=0·09), but predictive in men above 55
years (RR 1·58 (1·13–2·23, P=0·008). There was inter-
action between systolic blood pressure and the ECG in
male subjects with respect to ischaemic heart disease
events during 21 years follow-up (P=0·02, df=10). This
interaction was due to a disappearance of the eect of
systolic blood pressure in subjects with left ventricular
hypertrophy and negative T wave and subjects with left
ventricular hypertrophy and ST depression and negative
T wave changes.
Cardiovascular disease mortality
We found that the age-standardized incidence per 1000
patients years of cardiovascular disease mortality during
7 and 21 years follow-up were 8·7 and 16·8 in men, and
3·2 and 9·2 in women. There was significant interaction
between sex and ECG findings with respect to 21 years
follow-up (P=0·02, df 5). This interaction was due to
voltage-only left ventricular hypertrophy (ratio 0·74,
P=0·06), ST depression and negative T wave changes
(ratio 1·53, P=0·07) and left ventricular hypertrophy
with negative T wave (ratio 0·49, P=0·02). Although
these interactions were by sex and ECG with respect to
21 years follow-up of cardiovascular disease, we per-
formed the Cox analysis with both sexes together. The
age- and sex-adjusted and the multivariately adjusted
RR for cardiovascular disease mortality is shown in
Table 5. Voltage-only left ventricular hypertrophy was
not associated with future cardiovascular disease mor-
tality. A negative T wave, ST depression and negative T
wave changes, left ventricular hypertrophy with negative
T wave and with left ventricular hypertrophy with
ST depression and negative T wave had significantly
independent associations with cardiovascular disease
mortality during 7 and 21 years follow-up.
ECG and other cardiovascular risk factors
In order to estimate quantitatively the relative impact of
ECG findings on ischaemic heart disease, the results of
the multivariately adjusted Cox proportional hazard
model for 7 years and 21 years follow-up in relation to
fatal and non-fatal ischaemic heart disease events is
shown in Table 6. The reference levels of the confound-
ing risk factors is described in statistical methods. Table
6 shows all risk factors in this model. We did this
analysis for both 7 and 21 years follow-up to evaluate
short- and long-term associations with ECG findings
and ischaemic heart disease events. During the 7 year
follow-up, ECG findings of isolated negative T waves,
ST depression and negative T wave changes, left ven-
tricular hypertrophy with negative T wave and left
ventricular hypertrophy with ST depression and nega-
tive T wave changes were independently associated with
excess future ischaemic heart disease events. The risk
ratio of left ventricular hypertrophy with ST depression
and negative T wave was the most important risk
factor, and left ventricular hypertrophy with negative T
wave and ST depression and negative T wave changes
had a higher risk ratio than systolic blood pressure
>160 mmHg. Evaluation after 21 years follow-up
showed a diminution in the predictive power of
ECG findings in relation to other risk factors. Both
short- and long-term diabetes history and male sex were
accompanied by high risk ratios for future ischaemic
Table 5 Age-standardized incidence of cardiovascular disease mortality, age- and sex-adjusted relative risk and
multivariately adjusted relative risk for cardiovascular disease mortality during 7 years and 21 years follow-up in
relation to ECG findings among those without ischaemic heart disease at baseline, for the age range 35–74 years
ECG
7 years follow-up (number of CVD=376) 21 years follow-up (number of CVD=1924)
n (Inc.)
Age-adjusted RR
(95% CI)
RR* (95% CI) n (Inc.)
Age-adjusted RR
(95% CI)
RR* (95% CI)
Normal ECG, n=8460 223 (4·5) 1 1 1259 (10·9) 1 1
Voltage-only LVH, n=1197 45 (6·6) 1·23 (0·89–1·69) 1·28 (0·92–1·77) 221 (13·2) 1·10 (0·95–1·27) 1·07 (0·92–1·24)
T, n=799 55 (8·5) 1·82 (1·35–2·45)§ 1·61 (1·19–2·18)‡ 253 (18·1) 1·76 (1·53–2·01)§ 1·54 (1·34–1·76)§
ST/T, n=257 17 (9·1) 1·98 (1·21–3·26)‡ 1·68 (1·02–2·77)† 79 (18·6) 1·91 (1·52–2·40)§ 1·71 (1·36–2·15)§
LVH with neg. T, n=137 14 (13·4) 2·28 (1·32–3·92)‡ 2·16 (1·25–3·74)‡ 46 (19·5) 1·59 (1·18–2·14)‡ 1·37 (1·01–1·85)†
LVH with ST/T, n=132 22 (17·8) 3·75 (2·41–5·85)§ 2·96 (1·87–4·68)§ 66 (28·0) 2·55 (1·98–3·27)§ 1·90 (1·47–2·46)§
CI=confidence interval; RR=relative risk; IHD=coronary heart disease; CVD=cardiovascular disease; ECG=electrocardiographic;
LVH=left ventricular hypertrophy; ST/T=ST depression and negative T wave; T=negative T wave; LVH with neg. T=left ventricular
hypertrophy with negative T wave; LVH with ST/T=left ventricular hypertrophy with ST depression and negative T wave;
Inc.=age-standardized number of cardiovascular disease per 1000 years of follow-up.
*Multivariate Cox regression with adjustments for all risk factors shown in Table 2 and alcohol, physical exercise, and family history of
ischaemic heart disease.
P<0·05; P<0·01; §P<0·001.
320 C. T. Larsen et al.
Eur Heart J, Vol. 23, issue 4, February 2002
Page 6
heart disease events. Because of the significant interac-
tion between age and voltage-only left ventricular hyper-
trophy in men we performed an analysis of the subjects
in the age range 55–74 years (Table 6). In this subgroup
of subjects voltage-only left ventricular hypertrophy was
associated with ischaemic heart disease.
Discussion
In the Copenhagen City Heart Study, a prospective
cardiovascular population study, findings show that
men below 55 years of age with voltage-only ECG left
ventricular hypertrophy have a good prognosis. ECG
left ventricular hypertrophy with ST depression and
negative T wave changes are associated with a two- to
fourfold risk of future myocardial infarction events,
ischaemic heart disease events and cardiovascular dis-
ease mortality, compared to subjects with a normal
ECG. Our results show that the ECG carries important
predictive information for future cardiac events. The
Minnesota Code and classification system
[17]
has been
used to formalize the comparison of ECG findings
within and between studies. Many epidemiological
studies demonstrated that ECG findings are associ-
ated with subsequent cardiovascular morbidity and
mortality
[214]
. In the present study, interaction analysis
showed no significant interaction between sex and ECG
findings with regard to myocardial infarction, ischaemic
heart disease, cardiovascular disease, except for cardio-
vascular disease during 21 years follow-up.
Prevalence of ECG findings
The Copenhagen City Heart Study is a prospective
cardiovascular population study comprising a random
sample of 19 329 men and women in the age range 20 to
80+ years, with a participation rate of 73·6%. The
present study is restricted to subjects in the age range
25–74 years of age, without a previous history of ischae-
mic heart disease, and without antihypertensive medi-
cation. When comparing the results of our study with
those obtained from other studies it is important to
Table 6 Multivariately adjusted risk ratios for ischaemic heart disease events related to ECG findings and
cardiovascular risk factors 7 and 21 years follow-up
35–74 years
7 years follow-up
Risk ratio (95% CI)
35–74 years
21 years follow-up
Risk ratio (95% CI)
55–74 years
7 years follow-up
Risk ratio (95% CI)
55–74 years
21 years follow-up
Risk ratio (95% CI)
Age 45–54 2·23 (1·52–3·26)‡ 1·94 (1·65–2·28)‡
Age 55–64 4·27 (2·97–6·14)‡ 3·18 (2·72–3·71)‡
Age 65–74 6·87 (4·69–10·06)‡ 5·79 (4·87–6·89)‡ 1·60 (1·30–1·96)‡ 1·81 (1·62–2·04)‡
Male sex 2·62 (2·14–3·21)‡ 2·12 (1·92–2·35)‡ 2·17 (1·71–2·75)‡ 1·90 (1·67–2·16)‡
ECG voltage only LVH 1·15 (0·88–1·51) 1·10 (0·96–1·26) 1·36 (0·99–1·86) 1·20 (1·01–1·43)*
ECG negative T 1·56 (1·21–2·03)‡ 1·45 (1·26–1·67)‡ 1·68 (1·26–2·24)‡ 1·48 (1·26–1·74)‡
ECG ST/T 2·07 (1·39–3·08)‡ 1·81 (1·44–2·27)‡ 2·17 (1·42–3·34)‡ 1·86 (1·44–2·41)‡
ECG LVH with neg. T 1·89 (1·15–3·09)† 1·37 (1·04–1·87)* 1·97 (1·16–3·35)* 1·33 (0·94–1·88)
ECG LVH with ST/T 3·62 (2·47–5·30)‡ 2·17 (1·66–2·83)‡ 3·73 (2·47–5·64)‡ 2·37 (1·77–3·18)‡
Former smoking 1·12 (0·79–1·58) 1·24 (1·06–1·47)† 1·18 (0·81–1·71) 1·27 (1·05–1·53)*
Current smoking 1·88 (1·41–2·49)‡ 1·85 (1·62–2·12)‡ 1·71 (1·25–2·33)‡ 1·67 (1·42–1·95)‡
Cholesterol 5–7 mmol . l
1
0·93 (0·72–1·20) 1·19 (1·03–1·37)* 0·86 (0·63–1·16) 1·04 (0·86–1·26)
Cholesterol d7·0 mmol . l
1
1·19 (0·89–1·59) 1·44 (1·23–1·69)‡ 0·99 (0·70–1·39) 1·14 (0·93–1·41)
SBP 140–160 mmHg 1·25 (1·03–1·53)* 1·26 (1·14–1·40)‡ 1·09 (0·87–1·38) 1·11 (0·98–1·27)
SBP d160 mmHg 1·49 (1·17–1·88)‡ 1·65 (1·46–1·87)‡ 1·48 (1·14–1·92)† 1·46 (1·26–1·70)‡
Exercise 2–4 hours/week 0·75 (0·61–0·92)† 0·77 (0·69–0·85)‡ 0·71 (0·56–0·91)† 0·81 (0·70–0·93)†
Exercise d4 hours/week 0·66 (0·52–0·84)‡ 0·76 (0·67–0·86)‡ 0·74 (0·56–0·96)* 0·82 (0·70–0·96)*
NIDDM 1·88 (1·33–2·67)‡ 2·01 (1·62–2·48)‡ 1·93 (1·33–2·79)‡ 2·01 (1·58–2·54)‡
IDDM 3·23 (1·52–6·86)‡ 2·86 (1·82–4·52)‡ 3·21 (1·18–8·69)* 2·42 (1·21–4·89)*
Alcohol 1–2 beverages/day 0·78 (0·62–0·99)* 0·85 (0·75–0·96)* 0·77 (0·58–1·01) 0·89 (0·77–1·04)
Alcohol 3–4 beverages/day 0·93 (0·70–1·24) 0·88 (0·75–1·03) 0·87 (0·61–1·24) 0·81 (0·66–1·01)
Alcohol d5 beverages/day 0·66 (0·47–0·94)* 0·89 (0·74–1·06) 0·70 (0·46–1·07) 1·04 (0·82–1·31)
BMI <20 0·97 (0·65–1·45) 1·22 (1·00–1·49)* 0·95 (0·59–1·54) 1·25 (0·97–1·61)
BMI 25–30 1·06 (0·88–1·28) 1·20 (1·09–1·32)‡ 1·04 (0·83–1·29) 1·20 (1·05–1·35)†
BMI d30 1·07 (0·81–1·40) 1·32 (1·15–1·51)‡ 1·10 (0·80–1·50) 1·32 (1·11–1·57)†
Family IHD 1·44 (1·16–1·78)‡ 1·27 (1·14–1·42)‡ 1·43 (1·10–1·85)† 1·17 (1·01–1·36)*
HR 60–90 1·15 (0·85–1·56) 1·10 (0·94–1·28) 1·16 (0·82–1·66) 1·15 (0·95–1·39)
HR d90 1·23 (0·86–1·76) 1·20 (0·99–1·44) 1·25 (0·82–1·89) 1·19 (0·94–1·50)
IHD=ischaemic heart disease; CI=confidence interval; ECG=electrocardiographic; LVH=left ventricular hypertrophy; ST/T=ST
depression and negative T wave; T=T wave; LVH with neg. T=left ventricular hypertrophy with negative T wave; LVH with ST/T=left
ventricular hypertrophy with ST depression and negative T wave; SBP=systolic blood pressure; IDDM=insulin dependent diabetes
mellitus; NIDDM=non-insulin dependent diabetes mellitus; BMI=body mass index.
*P<0·05; P<0·01; P<0·001.
The Copenhagen City Heart Study 321
Eur Heart J, Vol. 23, issue 4, February 2002
Page 7
notify the dierences between the selection of subjects
(working subjects vs population based samples), defini-
tions of ECG abnormalities, participation rates, whether
the studies included subjects free of ischaemic heart
disease or subjects in antihypertensive treatment. Our
prevalence data are similar to others
[2,1821]
which also
described a high prevalence of voltage-only ECG left
ventricular hypertrophy in the younger ages, and an
increased prevalence with increasing age of ST depres-
sion and negative T wave changes, negative T wave, left
ventricular hypertrophy with negative T wave and left
ventricular hypertrophy with ST depression and nega-
tive T wave changes. Our study showed a higher age-
adjusted prevalence of voltage-only left ventricular
hypertrophy (19·3% in men, 5·6% in women) compared
to other studies
[2,20,21]
, which might be explained by
dierences in ECG definitions of left ventricular hyper-
trophy between the studies. In the present study, subjects
with ST depression and negative T wave changes, nega-
tive T wave, left ventricular hypertrophy with negative T
wave and left ventricular hypertrophy with ST depres-
sion and negative T wave changes were older, had higher
systolic blood pressure, higher diastolic blood pressure
and higher heart rate than subjects with normal ECG
(Table 2). Sigurdsson et al.
[13]
studied men, where 20%
were on antihypertensive medication and found ST
depression and negative T wave prevalence of 2% at
age 40 and 30% at age 80, which is higher than our
prevalence of ST depression and negative T wave. This
dierence might be explained by the fact that we
excluded subjects taking antihypertensive medication. In
a population of 47 358 subjects, including subjects with
symptoms of ischaemic heart disease or on antihyperten-
sive medication, from four large studies, De Bacquer
et al.
[21]
reported the prevalences of ECG findings. They
found age-adjusted prevalences of ST depression and
negative T wave changes and negative T waves, respect-
ively, of 2·3% and 6·5% in men, and of 2·6% and 7·6% in
women. The Minnesota codes for T waves and ST
depression included in the study by De Bacquer et al.
were the same as ours, but the ECG groups diered. Our
ECG groups were mutually exclusive, whereas the ECG
abnormalities, such as left ventricular hypertrophy, ST
depression and negative T wave changes and negative T
wave in the study by De Bacquer et al. were coexistent
with each other and with other ECG abnormalities. This
dierence in subjects and definitions of ECG groups
might explain why the results of our study showed a
lower prevalence of ST depression and negative T wave
changes (1·2% in men, and 2·2% in women) and negative
T waves (5·3% in men, and 5·3% in women) than De
Bacquer et al. found. The Framingham data
[2]
showed
that the estimated prevalence of left ventricular hyper-
trophy with ST depression and negative T wave changes
increased from 0·6% in 40-year-old men to 5·5% in
70-year-old men. In women, the prevalence was about
half that in men at all ages. In our results, we divided
the subjects with left ventricular hypertrophy with ST
depression and negative T wave abnormalities from the
subjects with left ventricular hypertrophy with T wave
abnormalities. Our results are in line with the data from
Framingham, both concerning the increase in prevalence
with age and that women have half the prevalence of
men (Table 1).
Voltage-only ECG left ventricular
hypertrophy
In the present study, we found a high prevalence of
voltage-only left ventricular hypertrophy in men.
Subjects having voltage-only left ventricular hyper-
trophy had a higher systolic blood pressure than those
with a normal ECG. There was a significant interaction
between voltage-only left ventricular hypertrophy and
age in men with respect to myocardial infarction and
ischaemic heart disease during 7 years follow-up. Men
younger than 55 years of age with voltage-only left
ventricular hypertrophy had the same risk as those with
a normal ECG with respect to myocardial infarction and
ischaemic heart disease, whereas those older than 55
years with this ECG finding had an RR of 1·94 (1·28–
2·93) for myocardial infarction and an RR of 1·58
(1·13–2·23) for ischaemic heart disease. Others have
also described the same good prognosis in subjects
with voltage-only left ventricular hypertrophy
[19,22,23]
.
Voltage-only left ventricular hypertrophy was initially
described in the Framingham study as carrying half the
prognostic information of ECG left ventricular hyper-
trophy with ST depression and negative T wave with
respect to cardiovascular disease, but later information
from the Framingham study indicates that voltage-only
left ventricular hypertrophy appears to reflect chiefly the
severity and duration of the associated hypertension,
and when adjustment was made for coexistent hyperten-
sion, the excess cardiovascular risk associated with this
type of ECG left ventricular hypertrophy was virtually
eliminated
[2,24]
. From our results we suggest that
voltage-only left ventricular hypertrophy may not be
considered an abnormal finding in men below 55 years
of age.
ECG ST depression and negative T wave
and ECG negative T wave
Many studies dealt with the prognostic value for ischae-
mic heart disease and cardiovascular disease of ST
depression and negative T wave findings and negative T
wave in men
[911,13,14,19,20,23,2528]
, and the results are
generally consistent. Most studies reported an excess
risk of 2 in those having ST depression and negative T
wave abnormalities, and a little lower risk in those
having negative T waves. Liao et al.
[28]
reported that ST
depression and negative T wave abnormalities indicate
an increased risk of subsequent death from ischaemic
heart disease, independent of major coronary risk
factors in men, but this was not clearly so for women.
De Bacquer et al.
[20]
reported that ST depression and
322 C. T. Larsen et al.
Eur Heart J, Vol. 23, issue 4, February 2002
Page 8
negative T wave in the baseline ECG is strongly associ-
ated with subsequent all-cause, cardiovascular disease,
and ischaemic heart disease mortality, with a similar
predictive value for men and women. They found that
ST depression and negative T wave changes in men and
women were associated with adjusted RRs of 4·28 and
6·00 for cardiovascular disease mortality, and these are
higher RRs than we found. In our study, ST depression
and negative T wave findings were associated with an
age- and sex-adjusted RR of 2·25 (1·52–3·34) for fatal
and non-fatal ischaemic heart disease and 1·98 (1·21–
3·26) for cardiovascular disease mortality, and a nega-
tive T wave with an age- and sex-adjusted RR of 1·60
(1·13–2·25) for myocardial infarction, of 1·69 (1·31–2·18)
for ischaemic heart disease events and 1·82 (1·35–2·45)
for cardiovascular disease mortality during 7 years
follow-up. During 21 years follow-up, ST depression
and negative T wave abnormalities, and an isolated
negative T wave were associated with myocardial infarc-
tion, ischaemic heart disease and cardiovascular disease
mortality. We think that the RR for all ECG abnormali-
ties is under-estimated, especially during 21 years follow-
up, since the prevalence of abnormal ECG findings
increases with increasing age, and some of the subjects
classified in the normal ECG group might have devel-
oped an ECG abnormality prior to an event during
follow-up.
ECG left ventricular hypertrophy with
negative T wave and ECG left ventricular
hypertrophy with ST depression and negative
T wave findings
Our data support other reports
[2,3,5,29]
, that left ventricu-
lar hypertrophy with ST depression and negative T wave
findings is an ominous harbinger of cardiovascular dis-
ease in the general population, markedly increasing the
risk of myocardial infarction, ischaemic heart disease
and cardiovascular disease. During 7 years follow-up
our results showed that left ventricular hypertrophy with
ST depression and negative T wave is a marker of future
myocardial infarction, ischaemic heart disease and
cardiovascular disease, with significant age and sex-
adjusted RRs of 3·78 (2·29–6·25) for myocardial infarc-
tion, 4·27 (2·95–6·16) for ischaemic heart disease and
3·75 (2·41–5·85) for cardiovascular disease. Even after
21 years follow-up, left ventricular hypertrophy with ST
depression and negative T wave was associated with
myocardial infarction, ischaemic heart disease and
cardiovascular disease.
ECG and other cardiovascular risk factors
De Bacquer et al.
[20]
have previously reported that major
ECG findings (Code IV
1-2
, or Code V
1-2
, or Code VI
1-2
,
or Code VII
1-2
, or Code VIII
1-3
) have a high relative
importance for later cardiovascular disease and ischae-
mic heart disease mortality compared to established risk
factors. Our findings that ECG changes have great
importance for later ischaemic heart disease events is in
agreement with De Bacquer et al. in relation to blood
pressure, diabetes, smoking, hyperlipidaemia, body mass
index and sex. Left ventricular hypertrophy with ST
depression and negative T wave is the most powerful
ECG predictor of ischaemic heart disease with an
adjusted RR of 3·62 (2·47–5·30) (Table 6), and even after
21 years follow-up left ventricular hypertrophy with ST
depression and negative T wave rated number two after
IDDM. Left ventricular hypertrophy with negative T
wave had a multivariately adjusted RR of 1·89 (1·15–
3·09), ST depression and negative T wave abnormalities
of 2·07 (1·39–3·08) and isolated negative T wave of 1·56
(1·21–2·03) during 7 years follow-up. ECG with negative
T wave, ST depression and negative T wave abnormali-
ties, and left ventricular hypertrophy with negative T
wave were also independently associated with ischaemic
heart disease after 21 years. We found an interaction
between age and voltage-only left ventricular hyper-
trophy in males and therefore we analysed the subjects
older than 55 years separately. In this model, voltage-
only left ventricular hypertrophy had borderline signifi-
cance for future ischaemic heart disease events, with an
RR of 1·36 (0·99–1·86) during 7 years follow-up and an
RR of 1·20 (1·01–1·43) during 21 years follow-up. One
explanation for the high prevalence of voltage-only left
ventricular hypertrophy in the younger ages and the
good prognosis of this ECG finding could be that this
ECG finding is associated with the good physical status
of these subjects. In men older than 55 years voltage-
only left ventricular hypertrophy might indicate a car-
diac disorder due to high blood pressure. As previously
reported in the Copenhagen City Heart Study
[30,31]
,
moderate alcohol intake was associated with a reduced
risk of future ischaemic heart disease. Moreover, exer-
cise of more than 2–4 h a week was associated with a
reduced adjusted RR of 0·75 for ischaemic heart disease
during 7 years of follow-up.
Conclusion
Left ventricular hypertrophy with ST depression and
negative T wave changes is the ECG finding with the
highest prognostic information for future ischaemic
heart disease and cardiovascular disease, and the risk is
similar for both men and women. Left ventricular hyper-
trophy with negative T, ST depression and negative T
wave abnormalities and negative T waves are signifi-
cantly associated with myocardial infarction, ischaemic
heart disease and cardiovascular disease. Men below 55
years of age with voltage-only ECG left ventricular
hypertrophy have the same prognosis as those with a
normal ECG, but this finding needs further evaluation
in future studies.
Financial support, is gratefully acknowledged from the Danish
Heart Foundation.
The Copenhagen City Heart Study 323
Eur Heart J, Vol. 23, issue 4, February 2002
Page 9
References
[1] Fisch C. Electrocardiography. In: Braunwald E, ed.
Heart Disease. A Textbook of Cardiovascular Medicine.
Philadelphia: W.B Saunders Company, 1997: 108–52.
[2] Kannel WB, Gordon T, Outt D. Left ventricular hyper-
trophy by electrocardiogram. Prevalence, incidence, and mor-
tality in the Framingham study. Ann Intern Med 1969; 71:
89–105.
[3] Kannel WB, Gordon T, Castelli WP, Margolis JR. Electro-
cardiographic left ventricular hypertrophy and risk of cor-
onary heart disease. The Framingham study. Ann Intern Med
1970; 72: 813–22.
[4] Kannel WB. Prevalence and natural history of electrocardio-
graphic left ventricular hypertrophy. Am J Med 1983; 75:
4–11.
[5] Kannel WB, Dannenberg AL, Levy D. Population impli-
cations of electrocardiographic left ventricular hypertrophy.
Am J Cardiol 1987; 60: 85I–93I.
[6] Kannel WB. Left ventricular hypertrophy as a risk factor in
arterial hypertension. Eur Heart J 1992; 13 (Suppl D): 82–8.
[7] Kreger BE, Cupples LA, Kannel WB. The electrocardiogram
in prediction of sudden death: Framingham Study experience.
Am Heart J 1987; 113: 377–82.
[8] Rabkin SW, Mathewson FL, Tate RB. The electrocardiogram
in apparently healthy men and the risk of sudden death. Br
Heart J 1982; 47: 546–52.
[9] Cullen K, Stenhouse NS, Wearne KL, Cumpston GN.
Electrocardiograms and 13 year cardiovascular mortality in
Busselton study. Br Heart J 1982; 47: 209–12.
[10] Macfarlane PW. British Regional Heart Study: the electro-
cardiogram and risk of myocardial infarction on follow-up.
J Electrocardiol 1987; 20 (Suppl): 53–6.
[11] O} sto¨r E, Schnohr P, Jensen G, Nyboe J, Hansen AT. Electro-
cardiographic findings and their association with mortality in
the Copenhagen City Heart Study. Eur Heart J 1981; 2:
317–28.
[12] Schouten EG, Dekker JM, Pool J, Kok FJ, Simoons ML.
Well shaped ST segment and risk of cardiovascular mortality.
BMJ 1992; 304: 356–9.
[13] Sigurdsson E, Sigfusson N, Sigvaldason H, Thorgeirsson G.
Silent ST-T changes in an epidemiologic cohort study — a
marker of hypertension or coronary heart disease, or both: the
Reykjavik study. J Am Coll Cardiol 1996; 27: 1140–7.
[14] Reunanen A, Pyo¨ra¨la¨ K, Punsar S, Aromaa A. Predictive
value of ECG findings with respect to coronary heart disease
mortality. Adv Cardiol 1978; 21: 310–2.
[15] Jensen G. Epidemiology of chest pain and angina pectoris.
Acta Med Scand Suppl 1984; 602: 1–120.
[16] Appleyard M, Hansen AT, Schnohr P, Jensen G, Nyboe J.
The Copenhagen City Heart Study. A book of tables with
data from the first examination (1976–78) and a five-year
follow-up (1981–1985). Scand J Soc Med 1989; (Suppl 41):
1–160.
[17] Rose GA, Blackburn H, Gillum RF, Prineas RJ. Cardio-
vascular Survey Methods, 2nd edn. Geneva: World Health
Organization, 1982.
[18] Higgins ITT, Kannel WB. The electrocardiogram in epide-
miological studies. reproducebility, validity and international
comparision. Br Prev Soc Med 1965; 19: 53–68.
[19] Rose G, Baxter PJ, Reid DD, McCartney P. Prevalence and
prognosis of electrocardiographic findings in middle-aged
men. Br Heart J 1978; 40: 636–43.
[20] De Bacquer D, De Backer G, Kornitzer M, Blackburn H.
Prognostic value of ECG findings for total, cardiovascular
disease, and coronary heart disease death in men and women.
Heart 1998; 80: 570–7.
[21] De Bacquer D, De Backer G, Kornitzer M. Prevalences of
ECG findings in large population based samples of men and
women. Heart 2000; 84: 625–33.
[22] Blackburn H, Taylor HL, Keys A. Coronary heart disease in
seven countries. XVI. The electrocardiogram in prediction of
five-year coronary heart disease incidence among men aged
forty through fifty-nine. Circulation 1970; 41: I154–I161.
[23] Tervahauta M, Pekkanen J, Punsar S, Nissinen A. Resting
electrocardiographic abnormalities as predictors of coronary
events and total mortality among elderly men. Am J Med
1996; 100: 641–5.
[24] Kannel WB, Sorlie P. Left ventricular hypertrophy in
hypertension, prognostic and pathogenetic implications: The
Framingham Study. In: Strauer BE, ed. The Heart in
Hypertension. Berlin: Springer-Verlag, 1981: 223–42.
[25] Knutsen R, Knutsen SF, Curb JD, Reed DM, Kautz JA,
Yano K. The predictive value of resting electrocardiograms
for 12-year incidence of coronary heart disease in the
Honolulu Heart Program. J Clin Epidemiol 1988; 41: 293–302.
[26] De Bacquer D, De Backer G, Kornitzer M, Myny K, Doyen
Z, Blackburn H. Prognostic value of ischemic electrocardio-
graphic findings for cardiovascular mortality in men and
women. J Am Coll Cardiol 1998; 32: 680–5.
[27] Kannel WB, Anderson K, McGee DL, Degatano LS,
Stampfer MJ. Nonspecific electrocardiographic abnormality
as a predictor of coronary heart disease: the Framingham
Study. Am Heart J 1987; 113: 370–6.
[28] Liao Y, Liu K, Dyer A et al.Sexdierential in the relationship
of electrocardiographic ST-T abnormalities to risk of cor-
onary death: 11.5 year follow-up findings of the Chicago
Heart Association Detection Project in Industry. Circulation
1987; 75: 347–52.
[29] Levy D, Salomon M, D’Agostino RB, Belanger AJ, Kannel
WB. Prognostic implications of baseline electrocardiographic
features and their serial changes in subjects with left
ventricular hypertrophy. Circulation 1994; 90: 1786–93.
[30] Grønbæk M, Deis A, Sørensen TI et al. Influence of sex, age,
body mass index, and smoking on alcohol intake and
mortality. BMJ 1994; 308: 302–6.
[31] Grønbæk M, Mortensen EL, Mygind K et al. Beer, wine,
spirits and subjective health. J Epidemiol Community Health
1999; 53: 721–4.
324 C. T. Larsen et al.
Eur Heart J, Vol. 23, issue 4, February 2002
Page 10