Journal of Human Hypertension (2006) 20, 451459

& 2006 Nature Publishing Group All rights reserved 0950-9240/06 $30.00
www.nature.com/jhh

ORIGINAL ARTICLE

Left ventricular hypertrophy determined

by SokolowLyon criteria: a different
predictor in women than in men?
RL Antikainen1,2, T Grodzicki1,3, AJ Palmer1, DG Beevers4, J Webster5 and CJ Bulpitt1 for
The Department of Health and Social Security Hypertension Care Computer Project
(DHCCP)
1

Imperial College Faculty of Medicine, London, UK; 2Department of Internal Medicine, University of Oulu
and Oulu City Hospital, Oulu, Finland; 3Jagiellonian University, Cracow, Poland; 4City Hospital,
Birmingham, UK and 5Aberdeen Royal Infirmary, Aberdeen, UK

The clinical usefulness of the SokolowLyon voltage

criteria in the assessment of electrocardiographic left
ventricular hypertrophy (ECG LVH) is addressed. We
prospectively studied 3338 women and 3330 men
referred with hypertension, with an average follow-up
of 11.2 years. The voltage amplitude sum SV1 max
(RV5 or RV6) was calculated and ECG LVH was defined
as a sum X3.5 mV. We adjusted survival for age,
treatment status before presentation and a previous
myocardial infarction or cerebrovascular accident. The
risk of stroke, coronary heart disease (CHD) and
cardiovascular disease (CVD) mortality increased significantly for each quantitative 0.1 mV increase in baseline electrocardiogram (ECG) voltage, in women within
the range of 1.63.9% and in men 1.43.0%. After further
adjustments for race, body mass index, smoking and
systolic blood pressure, increasing voltage indepen-

dently predicted CVD mortality in both men and women.

In women, both increasing voltage and the presence of
left ventricular hypertrophy (LVH) were predictors of
stroke mortality, whereas in men this risk was attenuated. In men, the adjusted association between increasing voltage and CHD mortality tended to be stronger
than in women. The use of different thresholds for the
two genders made little difference. For stroke and CHD
mortality, the population attributable fractions associated with LVH were 15.2 and 5.4% in women and 12.8
and 8.5% in men, respectively. In conclusion, the greater
the baseline ECG voltage sum, the greater the associated CVD mortality risk. Women tended to have a high
risk of stroke mortality owing to LVH despite adjustments.
Journal of Human Hypertension (2006) 20, 451459.
doi:10.1038/sj.jhh.1002006

Keywords: left ventricular hypertrophy; mortality; prospective study

Introduction
The presence of left ventricular hypertrophy (LVH)
in a hypertensive patient is evidence of hypertensive target organ damage,1,2 although a number of
factors other than blood pressure are known to be
involved in the development of LVH.3 Left ventricular hypertrophy has been shown to be associated
with a significant increase in risk of cardiovascular
mortality among the general population and hypertensive patients, independently of whether it is
determined by electrocardiogram (ECG LVH), echocardiography (ECHO LVH) or radiology.2,413
Echocardiography is the gold standard for assessing anatomical LVH.2,14 However, it is less readily
Correspondence: Dr RL Antikainen, Oulu City Hospital, Pl 8,
90015 Oulun Kaupunki, Oulu, Finland.
E-mail: Riitta.Antikainen@ouka.fi
Received 19 August 2005; revised 19 January 2006; accepted 23
January 2006

available and sometimes difficult to perform in

obese and elderly subjects.12,15 Because of availability and low cost, the ECG has traditionally been
the principal method for recognising LVH1 and it
provides additional information about cardiac
rhythm, conduction and ischaemia. The commonly
used SokolowLyon voltage criteria16 are very easy
to employ in the assessment of LVH. Despite the
high specificity, the ability of SokolowLyon ECG
voltage criteria to detect anatomical ECHO LVH or
the increased left ventricular mass measured by
autopsy is relatively low compared with several
different ECG criteria.14,17 The question has therefore been asked, whether or not the low sensitivity
limits its clinical utility.17 Although there are
surprisingly few studies concerning the prognostic
effect of employing only the ECG method, in recent
guidelines, SokolowLyon ECG voltage criteria are
still on the list of routine tests to determine target
organ damage, together with the Cornell index.1

ECG LVH and risk of mortality

RL Antikainen et al
452

In this study, we assessed the long-term mortality

from stroke, coronary heart disease (CHD) and
cardiovascular disease (CVD) with an increase in
baseline SokolowLyon ECG voltage among hypertensive women and men. We also compared the risk
of mortality among subjects with and without LVH
at baseline. The independent effect of LVH on
mortality was studied after adjustments for confounders, especially for specified blood pressure
levels.

Materials and methods

The Department of Health and Social Security
Hypertension Care Computer Project (DHCCP) is a
multicentre computer-based observational study,
started in 1971, of patients being treated for
hypertension in the United Kingdom.18,19 The
DHCCP contains records of 11 663 patients referred
with hypertension and recruited up to December 31,
1987. Ninety-five per cent of patients entered the
system on presentation to one of the hospital clinics
(Hammersmith Hospital, London; Kings College
Hospital, London; John Radcliff Hospital, Oxford;
City Hospital, Birmingham; Aberdeen Royal Infirmary). The remaining 5% were recruited from the
general practices involved in the study (Kentish
Town, London; Harlow; Oxford and Norwich).
The present analytical database was created on
December 31, 1986 and 1477 patients presenting
after this date were excluded. Of the remaining
10 186 subjects, 6698 subjects had an ECG recorded
at entry. At presentation, 30 subjects were excluded
from the analyses: 25 patients who were aged below
18 years and five in whom gender was not recorded.
Thus, the results presented in this report refer to
a database of 6668 patients from the DHCCP.
The mean age of all subjects was 50.9 years (range
18.089.8 years).

The assessors were trained by clinical doctors

familiar with ECGs and they evaluated the standard
12-lead resting ECGs according to the Sokolow
Lyon criterion at presentation. All ECGs had a 1 mV
calibration signal. The voltage sum SV1 max (RV5
or RV6) was calculated for the analyses using
ECG voltage as a continuous variable. Electrocardiographic left ventricular hypertrophy was diagnosed
when the voltage amplitude sum of either
SV1 RV5 or SV1 RV6 at baseline was equal to
or above 3.5 mV. Body mass index (BMI) was
calculated as weight (kg)/height squared (m2). Blood
was not necessarily taken in the fasting state. Blood
glucose measurements at baseline were not formally
standardised although they were performed in
laboratories linked to the United Kingdom quality
assurance scheme. Smokers were those who currently smoked or who had smoked in the past. Blood
pressure measurements were not formally standardised but were performed by experienced observers.
Most observers used phase V Korotkoff as diastolic
blood pressure (DBP). Baseline systolic blood pressure (SBP) and DBP measurements were those taken
closest to the presentation or as the most recent
measurement before presentation while not on
antihypertensive drug treatment, and were available
in 6237 subjects (Table 1). Follow-up SBP and DBP
were assessed during the first year of follow-up
(N 4442). For 2768 patients, antihypertensive drug
treatment had been started before presentation and
for 5205 by 1 year of follow-up.
All patients were registered with the offices of
National Statistics for England and Wales or the
Registrar General for Scotland, who provided copies
of the death certificates. Codes 43004389 were
classified as stroke deaths, codes 41004149 as CHD
deaths and codes 39004589 and 79507959 as CVD
deaths (International Classification of Diseases,
eighth revision). These analyses are based on an
average of 11.2 years of follow-up.

Table 1 Baseline characterics in hypertensive women and men with the absence (LVH) or presence (LVH+) of left ventricular
hypertrophy determined using SokolowLyon criterion
Women

Age (years) 7s.e.

Previous cv disease (%)
Treated before presentation (%)
Black (%)
Smoking (%)
BMI7s.e. (kg/m2)
SBP0 (mm Hg) 7s.e.
DBP0 (mm Hg) 7s.e.
SBP1 (mm Hg) 7s.e.
DBP1 (mm Hg) 7s.e.

Men

LVH
(%)

LVH

LVH+

LVH
(%)

LVH

LVH+

3338
3338
3338
3179
2916
2895
3107
3107
2222
2222

14
14
14
14
15
14
14
14
16
16

51.170.26
5.3
41.3
10.0
23.9
27.570.11
16770.57
10070.29
15770.52
9370.23

53.470.66
8.4
46.7
24.1
26.9
26.070.25
18371.62
10670.85
16471.39
9470.60

0.0014
0.0078
0.0258
o0.0001
0.1934
o0.0001
o0.0001
o0.0001
o0.0001
0.0086

3330
3330
3330
3169
2915
2878
3130
3132
2220
2220

25
25
25
25
26
25
26
26
29
29

49.970.63
9.1
38.8
6.0
36.2
27.470.08
16270.57
10070.31
15170.52
9270.25

51.470.43
11.9
47.3
15.4
42.3
25.970.13
17471.00
10570.63
15770.95
9470.46

0.0034
0.0214
o0.0001
o0.0001
0.0032
o0.0001
o0.0001
o0.0001
o0.0001
0.0100

Abbreviations: BMI, body mass index; DBP0, baseline diastolic blood pressure (DBP); DBP1, follow-up DBP; LVH, left ventricular hypertrophy;
LVH (%); proportion of subjects with LVH, N, number of subjects in whom information of each variable is available; P, P-value for statistical
significance; previous cv disease, previous history of myocardial infarction or cerebrovascular accident; SBP0, baseline systolic blood pressure
(SBP); SBP1, follow-up SBP; s.e., standard error.
Journal of Human Hypertension

ECG LVH and risk of mortality

RL Antikainen et al
453

Statistical methods

Table 2 Number of deaths (%) from specified causes in women

In descriptive analyses, the differences in mean

values and standard errors among subjects with or
without ECG LVH were compared using Students
t-test. Comparisons between proportions were performed using the w2 test for independent groups.
Multivariate analyses were performed by using the
Cox proportional hazards model.20 The estimates of
relative risks and their 95% confidence intervals
were calculated from this model. The relative risk of
mortality for each 0.1 mV increment in baseline ECG
voltage was calculated. In the categorical analysis,
subjects with LVH were compared with the subjects
without LVH. The analyses to estimate the relative
risk of mortality were first adjusted for age, previous
history of myocardial infarction or cerebrovascular
accident and treatment status before presentation.
To also control for the influence of BMI, smoking
and race on mortality, we added these confounders
simultaneously to a model in two subgroups of subjects, where the information on all of these variables
was available (Groups 0 and 1). The additional variables were age, previous history of myocardial infarction or cerebrovascular accident, treatment status
before presentation and baseline SBP (SBP0) in
Model 0 (N 4530) or first-year follow-up SBP
(SBP1) in Model 1 (N 3532). The variables in the
analyses were selected from variables present in the
vast majority of patients. Each of these variables was
a significant predictor in a stepwise model for at least
one end point either in women or in men.
The population attributable fraction (expressed as
a percentage) for the excess deaths from stroke, CHD
and CVD is the proportion of mortality associated
with the presence of LVH. It was calculated as
100(IIc)/I, where I is the overall population risk of
death and Ic is the expected risk of death for subjects
who do not have LVH.
Statistical analyses were performed using the SAS
statistical programs (The SAS Institute Inc., Cary,
NC, USA).

and men with and without left ventricular hypertrophy (LVH+

and LVH)

Results
In this hypertensive population, the prevalence of
ECG LVH defined using the above SokolowLyon
criterion was 14% among women and 25% among
men (Table 1). The blood pressure of subjects with
ECG LVH was higher, BMI was lower and they were
more often black or a smoker than subjects without
LVH. The subjects with LVH more often presented
with a history of previous myocardial infarction or
cerebrovascular accident and they were more often
on antihypertensive drug treatment before presentation. The numbers and percentages dying from
stroke, CHD and CVD stratified by the presence or
absence of LVH in women and men are presented in
Table 2. During a mean follow-up period of 11.2
years, there were a total of 305 deaths from strokes,
611 from CHD and a total of 1087 CVD deaths.

Women

Stroke
CHD
CVD

Men

LVH+
(N 475)

LVH
(N 2863)

LVH+
(N 846)

LVH
(N 2484)

42 (9)
44 (10)
105 (21)

112 (4)
189 (7)
368 (13)

53 (6)
120 (14)
198 (23)

98 (4)
258 (10)
416 (17)

Abbreviations: CHD, coronary heart disease; CVD, cardiovascular

disease; LVH, left ventricular hypertrophy.

After adjustments for age, treatment status before

presentation and previous myocardial infarction
and cerebrovascular accident, there was a significant increase in the relative risk of fatal stroke, CHD
and CVD with each 0.1 mV increase in baseline ECG
voltage (SV1 max (RV5 or RV6)) in both genders
over an average of 11.2 years of follow-up (Table 3).
In women, the increase in the risk ranged from 1.6
(CHD) to 3.9% (stroke). In men, these risk estimates
tended to be lower than risk estimates in women
varying from 1.4 (CHD) to 3.0% (stroke). When LVH
was expressed as a categorical variable, the risk
associated with the presence of ECG LVH after these
adjustments ranged from 1.3 (CHD) to 2.1 (stroke)
in women and from 1.3 (CHD) to 1.5 (stroke) among
men (Table 3). Only the 1.3-fold risk of CHD
mortality in women did not achieve statistical
significance (P 0.17).
In women, controlling for all seven variables
including baseline (Group 0, Model 0) or follow-up
SBP (Group 1, Model 1), each 0.1 mV increase in
ECG voltage was associated with an independent
2.73.6% increase in the risk of stroke mortality
and 1.71.8% increase in the CVD mortality. The
increase in the risk of CHD mortality was not
statistically significant and adjustments for both
baseline and follow-up SBP further attenuated these
risk estimates (Table 4).
In men, after full adjustments including baseline
(Group 0, Model 0) or follow-up SBP (Group1,
Model 1), each 0.1 mV increase in ECG voltage was
associated with a significant 1.31.4% increase in
the risk of CVD mortality. Adjustments for age and
follow-up SBP partly attenuated the risk of stroke
mortality. The increase in the risk of CHD mortality
was partly dependent on baseline SBP (Table 4).
In women, the presence of LVH was associated
with a 1.6-fold (95% Cl 1.02.6, P 0.0679) adjusted
risk of stroke mortality in Model 0 and 1.9-fold (95%
Cl 1.13.3, P 0.0239) risk in Model 1 (data not
shown). The risk of CVD mortality was 1.3-fold in
Model 0 (95% Cl 1.0-1.8, P 0.0518) and 1.4-fold in
Model 1 (95% Cl 1.01.9, P 0.0520). The adjusted
nonsignificant risk of CHD mortality for women was
1.2 in Model 0 and 1.1 in Model 1. In men, the
adjusted risks of stroke, CHD and CVD mortality
Journal of Human Hypertension

ECG LVH and risk of mortality

RL Antikainen et al
454

Table 3 Adjusted RR of mortality from specified causes among hypertensive women and men with each 0.1 mV increase in baseline
electrocardiographic voltage (a) and with the presence of left ventricular hypertrophy (LVH), compared with women and men without
LVH (b)
a

RR

(95% CI)

RR

(95% CI)

Women (N 3338)
Stroke
CHD
CVD

1.039
1.016
1.027

1.0241.054
1.0031.030
1.0181.036

o0.0001
0.0195
o0.0001

2.1
1.3
1.6

1.53.0
0.91.7
1.32.0

o0.0001
0.1737
o0.0001

Men (N 3330)
Stroke
CHD
CVD

1.030
1.014
1.018

1.0161.044
1.0051.024
1.0101.025

o0.0001
0.0038
o0.0001

1.5
1.3
1.4

1.12.1
1.11.6
1.11.6

0.0131
0.0133
0.0004

Abbreviations: CHD, coronary heart disease; CI, confidence interval; CVD, cardiovascular disease; P, P-value for statistical significance; RR,
relative risks.
Adjusted for age, past history of myocardial infarction and cerebrovascular accident and for antihypertensive drug treatment status before
baseline.

Table 4 Adjusteda RR of mortality from specified causes among hypertensive women and men with each 0.1 mV increase in baseline
electrocardiographic voltage
Adjustment for

RR

Women (95% CI)

RR

N 2265

Men (95% CI)

N 2265

Group 0
Stroke
Age (years)
and SBP0 (mm Hg)
Model 0

1.034
1.032
1.027

1.0141.054
1.0121.053
1.0061.049

0.0006
0.0014
0.0105

1.026
1.022
1.019

1.0071.045
1.0031.041
0.9991.040

0.0067
0.0249
0.0622

CHD
Age (years)
and SBP0 (mm Hg)
Model 0

1.010
1.005
1.008

0.9931.028
0.9871.023
0.9891.027

0.2557
0.6114
0.4064

1.014
1.011
1.012

1.0021.027
0.9981.023
0.9991.026

0.0270
0.1064
0.0788

CVD
Age (years)
and SBP0 (mm Hg)
Model 0

1.019
1.016
1.018

1.0081.031
1.0041.028
1.0051.031

0.0013
0.0095
0.0061

1.016
1.013
1.013

1.0061.026
1.0031.023
1.0021.023

0.0013
0.0116
0.0177

N 1627

N 1625

Group 1
Stroke
Age (years)
and SBP1 (mm Hg)
Model 1

1.046
1.041
1.036

1.0201.071
1.0151.067
1.0101.064

0.0003
0.0016
0.0070

1.025
1.015
1.014

1.0051.046
0.9951.036
0.9921.036

0.0129
0.1474
0.2064

CHD
Age (years)
and SBP1 (mm Hg)
Model 1

1.012
1.006
1.007

0.9931.032
0.9871.027
0.9861.028

0.2249
0.5296
0.5388

1.016
1.013
1.017

1.0031.030
0.9991.027
1.0031.032

0.0189
0.0719
0.0212

CVD
Age (years)
and SBP1 (mm Hg)
Model 1

1.023
1.018
1.017

1.0091.037
1.0041.033
1.0031.032

0.0010
0.0096
0.0192

1.016
1.012
1.014

1.0051.027
1.0011.023
1.0031.026

0.0033
0.0311
0.0144

Abbreviations: CHD, coronary heart disease; CI, confidence interval; CVD, cardiovascular disease; Group 0, information on baseline systolic blood
pressure (SBP0) available; Group 1, information on systolic blood pressure during the first year of follow-up (SBP1) available; P, P-value for
statistical significance; RR, relative risks; SBP0, baseline systolic blood pressure (SBP); SBP1, follow-up SBP.
a
All RR adjusted also for past history of myocardial infarction and cerebrovascular accident and for antihypertensive drug treatment before
baseline. Also adjusted for race, body mass index, smoking and SBP0 in Model 0 (or SBP1 in Model 1).

associated with the presence of LVH ranged from 1.1

to 1.3 without statistical significance (data not
shown).
Journal of Human Hypertension

After adjustments for baseline or follow-up

DBP instead of SBP, the relative risks associated
with increasing ECG voltage or with the presence

ECG LVH and risk of mortality

RL Antikainen et al
455

of LVH remained virtually unchanged (data not

shown).
The mean blood glucose concentration at baseline
in women without LVH was 5.38 mmol/l and in
women with LVH 5.25 mmol/l. The corresponding
values in men were 5.38 and 5.26 mmol/l. When
differences in age, treatment status and previous
myocardial infarction and cerebrovascular accident
were taken into account, an additional adjustment
for blood glucose level or heart rate did not affect the
magnitude of the risk of stroke and CVD-associated
LVH in Table 3. Adjustment for blood glucose level
attenuated the risk of CHD associated with increasing ECG voltage in both women (RR 1.015,
P 0.09) and men (RR 1.010, P 0.1520) as did
adjustment for heart rate for the risk of CHD in
women (RR 1.013, P 0.07) (data not shown).
The population attributable fraction for women
associated with LVH was 15.2% for stroke, 5.4% for
CHD and 9.3% for CVD mortality. For men, the
corresponding percentages were 12.8, 8.5 and 9.2%,
respectively.

Discussion
In women, a statistically significant and independent increase in the risk of stroke and CVD mortality
was associated with increasing ECG voltage at
baseline. The presence of LVH was also a significant
risk factor for stroke in women in Model 1 and
marginally significant in Model 0. Similarly, men
had an independent increase in the risk of overall
CVD mortality associated with increasing baseline
ECG voltage. In men, adjusting for baseline SBP
partly attenuated the effect of LVH on the risk of
CHD mortality, whereas follow-up SBP attenuated
the effect on the risk of stroke mortality. The
adjustments for DBP only had a minor effect on
the risk of mortality. In our study, the continuous
voltage variable was found to be, as expected, a
better predictor because, unlike a dichotomous
variable, it retains all the information. However,
for a clinician, it is easier to work with a cutoff
value, even if this is slightly less good as a predictor.
Initially, in the Framingham study, the presence of
ECG LVH predicted an increase in the risk of stroke,
cardiac failure, coronary, peripheral vascular, total
cardiovascular diseases and sudden death.2,21 Subsequently, several studies have found an independent increase in mortality from CHD,69 stroke4,6,9,22
and CVD5,10,11,23 associated with ECG LVH among
both the general population and hypertensive
patients. A limitation in making comparisons between these studies is the fact that the ECG criteria
used to determine LVH varied substantially between
studies.
There are few published reports on the risk of
cardiovascular mortality associated with ECG-LVH
using SokolowLyon ECG voltage sum SV1 max
(RV5 or RV6) as the sole criterion.10,11,24,25 The

European Working Party on High Blood Pressure

Study (EWPHE) did not find any significant correlation between LVH and total or CVD mortality during
a 4-year follow-up.24 In the Italian Section of the
Seven Countries Study, the association between
non-sudden CHD mortality and LVH was even
inverse.25 The Progetto Ipertensione Umbria Monitoraggio Ambulatoriale (PIUMA) Study compared
the prognostic value of different ECG LVH criteria
among patients with essential hypertension.10 During an average of 3.3 years of follow-up, a 1.8-fold
increase in CVD mortality associated with the
presence of ECG LVH by SokolowLyon criteria
was not statistically significant, whereas some newer criteria identified a significant risk. In a study of
Swedish men,11 the presence of LVH defined using
the Cornell product independently predicted CVD
and all cause mortality, whereas using Sokolow
Lyon criteria, LVH only significantly predicted all
cause mortality. In the study from the Glasgow
Blood Pressure Clinic,9 the presence of ECG LVH
was diagnosed when the sum of SVI RV5 was
above 35 mm or additionally the deepest precordial
S and tallest R were above 40 mm. After adjustments, a significant increase in the risk of CHD and
stroke mortality associated with the presence of ECG
LVH was demonstrated among both hypertensive
women and men. In the Framingham Study,21
subjects with ECG LVH (one criterion or a combination of several voltage criteria together with repolarisation abnormalities) had an increased risk of CHD
compared with all subjects, whereas the risk among
subjects with only voltage criteria for ECG LVH
did not remain significant. Of these tracings,
85% fulfilled the SokolowLyon voltage criterion
SV1 RV5X3.5 mV. The protocols of these studies911,21,24 differed from our study in several ways:
among others, the number of subjects with ECG LVH
defined using SokolowLyon criteria was smaller
and the follow-up shorter than in our study. The
participants in the study of Swedish men11 and in
the PIUMA Study10 had lower mean blood pressure
levels than subjects in the Glasgow Study9 and in
the present study. Hsieh et al.26 compared the effect
of 17 different LVH ECG criteria among 19 434 male
veterans. No single criterion was significantly superior for predicting CVD mortality. Using selected
SokolowLyon criteria ((SV1 RV5)X3.5, RV5/
6X2.6, RaVLX1.1 or RaVFX2 mV), the adjusted
hazard ratio for CVD mortality was 1.9 (95% Cl 1.6
2.2) and higher than the corresponding relative risk
estimates among men in our study.26
The present study demonstrates the value of LVH
defined solely by the SokolowLyon voltage criterion SV1 max (RV5 or RV6) X3.5 mV as a predictor
of stroke mortality especially among hypertensive
women. Moreover, the population attributable risk
was highest for stroke mortality among women
(15%). It is not surprising that LVH predicts death
from stroke, as both LVH and stroke have a close
relationship with blood pressure level.1,2 Both
Journal of Human Hypertension

ECG LVH and risk of mortality

RL Antikainen et al
456

hypertension and ECG LVH have been shown to

predict the development of atrial fibrillation, which
is itself a major independent risk for stroke.27 In our
study, information on the presence of atrial fibrillation was not available at baseline or during follow-up.
The gender-specific risk of dying from cardiac
causes has been reported to be higher among women
than among men with ECHO LVH.28 The Framingham study found higher risk estimates for CHD,
stroke and any CVD morbidity for women versus
men with ECG LVH,2 whereas in a Belgian study the
risk of CVD mortality was statistically significant
among men only.5 In the Glasgow Blood Pressure
Clinic, ECG LVH among men was associated with a
significant 2.7-, 1.8- and 1.6-fold risk of dying from
ischaemic heart disease, stroke and all causes. The
corresponding relative risks among women were 2.0,
4.0 and 2.4, respectively.9 Similarly, the present
study suggests that women with ECG LVH may have
a relatively greater increase in the risk of stroke
death than do men, whereas the adjusted risk of
CHD mortality seems to be higher in men.
According to the guidelines in the 1980s, the
decision to begin drug treatment was largely based
on DBP and male gender has been one of the
associated risk factors that influenced the decision
to begin drug treatment at lower DBP levels.29 In our
study, the mean SBP levels as well as the mean age
at presentation were higher among women than
among men. It is possible that less active antihypertensive treatment strategies among women led to
these differences in blood pressure levels between
men and women and to the tendency for higher risk
estimates of stroke mortality associated with LVH
among women than among men in our study. An
earlier and more effective antihypertensive treatment may be required to reduce the risk of stroke
mortality especially in women. However, adjustments for SBP did not remove the excess risk of
stroke or CVD associated with increasing voltage
among women.
Echocardiographic left ventricular hypertrophy
may be more common among middle-aged and older
women than among men.30 However, the sensitivity
of the 3.5 mV SokolowLyon criterion to detect
anatomical LVH may be lower in women than in
men, the former having lower voltages, particularly
in the younger age ranges. This is complicated by
the fact that the sensitivity of ECG LVH criteria will
vary depending on other factors such as age, BMI or
severity of LVH.30,31 The question remains, whether
or not the presence of ECG LVH in our study
identifies a more serious anatomical LVH with a
greater risk in women than in men. Therefore, we
additionally assessed the risk associated with the
presence of LVH using the new gender-specific
partition values for SokolowLyon criteria for LVH
(X3.4 mV for women and X3.8 mV for men) proposed by Alfakih et al.32 The risk estimates using the
new partition values remained similar to the
estimates in Table 3. In both genders, only the 1.3Journal of Human Hypertension

fold risk for CHD mortality did not achieve statistical significance (data not shown). After full
adjustments, only the 1.8-fold risk of stroke in
women in Model 1 remained statistically significant
(P 0.0343). Thus, the gender difference may simply reflect a superiority of the SokolowLyon
criteria to predict stroke in women compared with
men, with the possibility that a lower voltage sum in
women indicates a higher degree of hypertrophy
than in men. Although the confidence limits for
the relative risks in women and men overlap, the
suggestion of differential prediction is supported by
the tendency for the criteria to be a better predictor
of CHD in men. This concept is supported to a large
extent by the three studies discussed above.2,5,9
Non-insulin-dependent diabetes mellitus has
been independently associated with an increase in
left ventricular mass.33,34 In our previous study
among DHCCP patients, blood glucose concentration tended to be inversely related to the presence of
ECG LVH by SokolowLyon criterion.35 In the
Framingham Heart Study after adjusting for BMI
and other risk factors, severity of hyperglycaemia
was observed to be related to left ventricular mass
more strongly in women than in men.36 In the
present study, the blood glucose level was measured
in 3970 subjects and heart rate in 5862 subjects.
After adjustments for variables in the Table 3,
additional adjustment for blood glucose level did
not affect the magnitude of the LVH-associated risk
of stroke and CVD mortality, but further attenuated
the association between CVD mortality in both
sexes.
Physical activity may be an important factor for
physiological LVH especially in younger people.2 In
our study, the same voltage threshold to detect LVH
was used for both young and old subjects. To
exclude the effect of a possible physiological LVH
in younger people, an additional analysis was
restricted to women (N 2621) and men (N 2603)
aged 40 years or more. In these analyses, the
adjusted relative risks of mortality associated with
both increasing voltage and the presence of LVH
remained similar to the results in women and men
aged 18 years or more. However, a 1.4-fold risk of
CVD mortality in women reached statistical significance in both Model 0 and Model 1 (data not
shown).
Electrocardiographic left ventricular hypertrophy
has also been shown to diminish in parallel with a
reduction in blood pressure level.8,37,38 The results
of the Losartan Intervention for Endpoint Reduction
in Hypertension (LIFE) Study39 in patients with ECG
LVH demonstrated that the risk of cardiovascular
morbidity and death, especially the risk of stroke,
was lower and reversing LVH more effective in those
assigned losartan than assigned atenolol treatment.
Regression of LVH by SokolowLyon voltage criteria
has been shown to result in a reduction of comorbidity in some40,41 but not in all studies.42 In
elderly patients with systolic hypertension in the

ECG LVH and risk of mortality

RL Antikainen et al
457

Systolic Hypertension In Europe (Syst-Eur) Trial, a

decrease in ECG voltage (RaVL SV1 RV5) independently predicted a lower incidence of cardiac
events.23
In our study, it was not possible to unequivocally
exclude the effect of antihypertensive drug treatment
on the gender difference in the risks of mortality.
Diuretic treatment was significantly more prevalent in
women than in men, 48.1 and 44.8%, respectively
(P 0.007). On the other hand, angiotensin-converting
enzyme (ACE) inhibitor use was less prevalent in
women than in men, 6.0 and 7.3%, respectively
(P 0.036). We analysed the data after we had
excluded subjects on ACE inhibitors (N 442) and
we also determined the relative risk of mortality
adjusted for diuretic treatment. In both analyses, as
well as in analyses where subjects with a history of
myocardial infarction were excluded, the relative risks
remained virtually the same as the risks in Table 3.
The absence of ECG LVH by the SokolowLyon
criterion does not exclude the presence of anatomical LVH. On the other hand, ECHO LVH and ECG
LVH have been shown to predict mortality independent of each other.11 These findings suggest that
anatomical ECHO and electrical ECG diagnoses of
LVH might reflect certain differences in pathogenesis.2,11 The SokolowLyon voltage index SV1 max
(RV5 or RV6) in hypertensive subjects is very
commonly used, inexpensive to measure and very
easy to employ. Despite its limited sensitivity, the
data in our study support the clinical utility of the
SokolowLyon criterion. However, in patients in
whom LVH is not discovered by this investigation,
other recommended tests1 should be employed to
determine the overall cardiovascular risk for antihypertensive treatment strategies.
The sensitivity of ECG LVH by SokolowLyon
criteria to recognise ECHO LVH among obese
subjects, particularly obese women, has been reported to be especially low.16 We have discussed in
detail the association between BMI and LVH in our
previous paper.32 We plan to estimate survival in
both obese and non-obese hypertensive subjects,
with the SokolowLyon voltage criterion for LVH.
Strengths and limitations

The main strength of our study is the high number

of hypertensive subjects with information on the
presence or absence of ECG LVH, as defined by the
commonly employed SokolowLyon criterion. Untreated blood pressure was often measured only
once. However, patients were referred with hypertension to the study and so the single blood pressure
measurement might not overestimate the patients
usual blood pressure level, as these subjects will
have been accustomed to blood pressure recording.
In the 1980s, although controlled prospective studies on the effectiveness and safety of treatment of
SBP were lacking, the recommendations on antihypertensive treatment were based on DBP.33 As

most patients referred with hypertension between

1971 and 1986 presented with elevated DBP, this
study clearly was biased towards diastolic hypertension. Long-term mortality data were obtained from
death certificates, known for their potential inaccuracies. Information on baseline SBP was available
in 94% of subjects, follow-up SBP in 67%, variables
in Model 0 in 65% and in Model 1 in 49% of
subjects. Thus, the adjusted results do not reflect all
the patients. Also, ECG LVH at baseline was not
always measured at the same time as baseline blood
pressure and the latter may have been determined
some time previously. However, these discrepancies
are likely to lead to falsely negative results and not
to the positive findings of the study. Also, in this
study, it was not possible to compare the effect of
LVH measured using SokolowLyon criteria with
other ECG measurements for LVH on mortality.
In conclusion, in both men and women, the
greater the baseline voltage sum, the greater the
associated CVD mortality risk. Women remained
with a high risk of stroke mortality due to LVH
despite adjustments.

What is known on this topic?

K ECG LVH is associated with increased cardiovascular risk
K Previous studies have often used complex criteria for ECG
LVH
K The commonly used and easily employed SokolowLyon
voltage criteria max (SV1+RV5 or RV6) significantly predicted
all cause mortality
What this study adds?
LVH determined using the SokolowLyon voltage criteria
predicted stroke CHD and CVD mortality in hypertensive
subjects
K The prediction of CHD tended to be better in men, possibly
owing to larger number of events in men
K The prediction of stroke tended to be slightly better in
women, agreeing with the fact that the SokolowLyon
criterion predicts MRI LVH better in women than in women
K

Acknowledgements
This study was supported by grants from the
Finnish Medical Foundation and the UK Department of Health and Social Security. The following
also participated in the DHSS: Beilin LJ, Cohen A,
Coles EC, Dollery CT, Fletcher AE, Jeffers TA,
Ledingham JGG, Munro-Faure AD, Petrie JC, Robb
OJ, Rossiter C, Rylance PB, Struthers A. From
general practice: Davies JB, Douglas-Jones AP, Grant
D, Horder E, Kenworthy-Browne JM.

References
1 European Society of Hypertension-European Society
of Cardiology Guidelines Committee. 2003. European
Society of Hypertension-European Society of Cardiology. Guidelines for the management of arterial hypertension. J Hypertens 2003; 21: 10111053.
Journal of Human Hypertension

ECG LVH and risk of mortality

RL Antikainen et al
458

2 Kannel WB. Left ventricular hypertrophy as a risk

factor: the Framingham experience. J Hypertens 1991;
9(Suppl 2): S3S9.
3 Kahan T. The importance of left ventricular hypertrophy in human hypertension. J Hypertens 1998;
16(Suppl 8): S23S29.
4 Bots ML, Nikitin Y, Salonen JT, Elwood PC, Malyutina
S, Freire de Concalves A et al. Left ventricular
hypertrophy and the risk of fatal and non-fatal stroke.
EUROSTROKE: a collaborative study among research
centres in Europe. J Epidemiol Commun Health 2002;
56(Suppl I): i8i13.
5 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: 570577.
6 Kahn S, Frishman WH, Weissmann S, Ooi WL,
Aronson M. Left ventricular hypertrophy on electrocardiogram: prognostic implications from a 10-year
cohort study of older subjects. A report from the Bronx
Longitudinal Ageing Study. J Am Geriatr Soc 1996; 44:
524529.
7 Brown DW, Giles WH, Croft JB. Left ventricular
hypertrophy as a predictor of coronary heart disease
mortality and the effect of hypertension. Am Heart J
2000; 140: 848856.
8 MacMahon S, Collins G, Rautaharju P, Cutler J, Neaton
J, Prineas R et al. for the Multiple Risk Factor
Intervention Trial Research Group. Electrocardiographic left ventricular hypertrophy and effects of
antihypertensive drug therapy in hypertensive participants in the Multiple Risk Factor Intervention Trial.
Am J Cardiol 1989; 63: 202210.
9 Dunn FG, McLenachan J, Isles CG, Brown I, Dargie HJ,
Lever AF et al. Left ventricular hypertrophy and
mortality in hypertension: an analysis of data from
the Glasgow Blood Pressure Clinic. J Hypertens 1990;
8: 775782.
10 Verdeccia P, Schillaci G, Borgioni C, Ciucci A,
Gattobigio R, Zampi I et al. Prognostic value of
new electrocardiographic method for diagnosis of
left ventricular hypertrophy in essential hypertension.
J Am Coll Cardiol 1998; 31: 383390.
rnlov J, Zethelius B, Andren B,
11 Sundstrom J, Lind L, A
Lithell HO. Echocardiographic and electrocardiographic diagnoses of left ventricular hypertrophy
predict mortality independently of each other in a
population of elderly men. Circulation 2001; 103:
23462351.
12 Levy D, Garrison RJ, Savage DD, Kannel WB, Castelli
WP. Prognostic implications of echocardiographically
determined left ventricular mass in Framingham Heart
Study. N Engl J Med 1990; 322: 15611566.
13 Koren MJ, Devereux RB, Casale PN, Savage DD, Laragh
JH. Relation of left ventricular mass and geometry to
morbidity and mortality in uncomplicated essential
hypertension. Ann Int Med 1991; 114: 345352.
14 Devereux RB, Koren MJ, de Simone G, Okin PM,
Kligfield P. Methods for detection of left ventricular
hypertrophy: application to hypertensive heart disease. Eur Heart J 1993; 14(Suppl D): 815.
15 Gardin JM, Siscovick D, Anton-Culver H, Lynch JC,
Smith VE, Klopfenstein HS et al. Sex, age, and disease
affect echocardiographic left ventricular mass
and systolic function in the free-living elderly. The
Cardiovascular Health Study. Circulation 1995; 91:
17391748.
Journal of Human Hypertension

16 Sokolow M, Lyon TP. The ventricular complex in

left ventricular hypertrophy as obtained by unipolar
precordial and limbs leads. Am Heart J 1949; 37:
161186.
17 Abergel E, Tase M, Menard J, Chatellier G. Influence of
obesity on the diagnostic value of electrocardiographic
criteria for detecting left ventricular hypertrophy. Am J
Cardiol 1996; 77: 739744.
18 Bulpitt CJ, Beillin LJ, Coles EC, Dollery CT, Johnson BF,
Munro-Faure AD et al. Randomised controlled trial
of computer-held medical records in hypertensive
patients. BMJ 1976; 1: 677679.
19 Beilin LJ, Bulpitt CJ, Coles EC, Dollery CT, Johnson BF,
Mearns C et al. Computer-based hypertension clinic
records: a co-operative study. BMJ 1974; 2: 212216.
20 Cox DR. Regression models and life tables. J R Stat Soc
(B) 1972; 34: 187220.
21 Kannel WB, Gordon T, Castelli W, Margolis JR.
Electrocardiographic left ventricular hypertrophy
and risk of coronary heart disease. Framingham Study.
Ann Int Med 1970; 72: 813822.
22 Tanne D, Yaari S, Goldbourt U. Risk profile and
prediction of long-term ischemic stroke mortality. A
21-year follow-up in the Israeli Ischemic Heart Disease
(IIHD) Project. Circulation 1998; 98: 13651371.
23 Fagard RH, Staessen JA, Thijs L, Celis H, Birkenhager
WH, Bulpitt CJ et al. for the Systolic Hypertension
in Europe (Syst-Eur) Trial Investigators. Prognostic
significance of electrocardiographic voltages and their
serial changes in elderly with systolic hypertension.
Hypertension 2004; 44: 459464.
24 Van Hoof R for the members of European Working
Party on High Blood Pressure in the Elderly. Left
ventricular hypertrophy in elderly hypertensive patients: a report from the European Working Party on
High Blood Pressure in the Elderly Trial. Am J Med
1991; 90(Suppl 3A): 55S59S.
25 Lanti M, Puddu PE, Menotti A. Voltage criteria of left
ventricular hypertrophy in sudden and nonsudden
coronary artery disease mortality: the Italian Section of
the Seven Countries Study. Am J Cardiol 1990; 66:
11811185.
26 Hsieh BP, Pham MX, Froelicher VF. Prognostic value
of Electrocardiographic criteria for left ventricular
hypertrophy. Am Heart J 2005; 150: 161167.
27 Kannel WB, Wolf PA, Benjamin EJ, Levy D. Prevalence,
incidence, prognosis, and predisposing conditions
for atrial fibrillation: population based estimates.
Am J Cardiol 1998; 82: 2N9N.
28 Liao Y Cooper RS, Mensah GA, McGee DL. Left
ventricular hypertrophy has a greater impact on
survival in women than in men. Circulation 1995; 92:
805810.
29 Report of the Joint National Committee on Detection,
Evaluation, and Treatment of High Blood Pressure.
A cooperative study. JAMA 1977; 237: 255261.
30 Levy D, Anderson KM, Savage DD, Kannel WB,
Christiansen JC, Castelli WP. Echocardiographically
detected left ventricular hypertrophy: prevalence and
risk factors. The Framingham Heart Study. Ann Int
Med 1988; 108: 713.
31 Levy D, Labib SB, Anderson KM, Christiansen JC,
Kannel WB, Castelli WP. Determinants of sensitivity
and specificity of electrocardiographic criteria for left
ventricular hypertrophy. Circulation 1990; 81: 815820.
32 Alfakih K, Walters K, Jones T, Ridgway J, Hall AS,
Sivananthan M. New gender-specific partition values

ECG LVH and risk of mortality

RL Antikainen et al
459

33

34

35

36

37

38

for ECG criteria of left ventricular hypertrophy.

Recalibration against cardiac MRI. Hypertension
2004; 44: 175179.
Palmieri V, Bella JN, Arnett DK, Liu JE, Oberman A,
Schuck MY et al. Effect of type 2 diabetes mellitus on
left ventricular hypertrophy geometry and systolic
function in hypertensive subjects. Hypertension
Genetic Epidemiology Network (HyperGEN) Study.
Circulation 2001; 103: 102107.
Devereux RB, Roman MJ, Paranicas M, OGrady MJ,
Lee ET, Welty TK et al. Impact of diabetes on cardiac
structure and function. The Strong Heart Study.
Circulation 2000; 101: 22712276.
Antikainen R, Grodzicki T, Palmer AJ, Beevers GD,
Coles EC, Webster J et al. Determinants of left
ventricular hypertrophy defined by SokolowLyon
criteria. J Hum Hypertens 2003; 17: 159164.
Rutter MK, Parise H, Benjamin EJ, Levy D, Larson MG,
Meigs JB et al. Impact of glucose intolerance and
insulin resistance on cardiac structure and function.
Sex-related differences in the Framingham Heart
Study. Circulation 2003; 107: 448454.
Levy D, Salomon M, DAgostino RB, Belanger AJ,
Kannel WB. Prognostic implications of baseline electrocardiographic features and their serial changes in
subjects with left ventricular hypertrophy. Circulation
1994; 90: 17861793.
Hypertension Detection and Follow-up Program cooperative group. Five years findings of Hypertension

39

40

41

42

detection and follow-up program. Prevention and

reversal of left ventricular hypertrophy with antihypertensive treatment. Hypertension 1985; 7: 105112.
Dahlof B, Devereux RB, Kjeldsen SE, Julius S, Beevers
G, de Faire U et al. Cardiovascular morbidity and
mortality in the Losartan Intervention for Endpoint
reduction in hypertension study (LIFE): a randomised
trial against atenolol. Lancet 2002; 359: 9951003.
Mathew J, Sleight P, Lonn E, Johnstone D, Pogue J, Yi Q
et al. for the Heart Outcomes Prevention Evaluation
(HOPE) Investigators. Reduction of cardiovascular risk
by regression of electrocardiographic markers of left
ventricular hypertrophy by angiotensin converting
enzyme inhibitor ramipril. Circulation 2001; 104:
16151621.
Okin PM, Devereux RB, Jern S, Kjeldsen SE, Julius S,
Nieminen MS et al. for the LIFE Study investigators.
Regression of electrocardiographic left ventricular
hypertrophy during antihypertensive treatment and
the prediction of major cardiovascular events. JAMA
2004; 292: 23432349.
Prineas RJ, Rautaharju P, Grandits G, Crow R for the
MRFIT Research Group. Independent risk for cardiovascular disease predicted by modified continuous
score electrocardiographic criteria for 6-year incidence
and regression of left ventricular hypertrophy among
clinically disease free men: 16-year follow-up for the
Multiple Risk Factor Intervention Trial. J Electrocardiol 2001; 34: 91101.

Journal of Human Hypertension