[Environmental Health and Preventive Medicine 7, 7981, May 2002]

Short Communication

Preventive Strategy for Hypertension Based on Attributable Risk Measures

Machi SUKA, Hiroki SUGIMORI and Katsumi YOSHIDA
Department of Preventive Medicine, St. Marianna University School of Medicine, Kanagawa

Abstract
Objectives: To examine the effective preventive strategy for hypertension in a Japanese male population, based on attributable risk measures.
Methods: A 7-year follow-up study of hypertension among 6,306 middle-aged male office workers in a
Japanese telecommunication company.
Results: In terms of population attributable risk percentage (PAR%), regular alcohol intake and physical inactivity showed great contributions to the development of hypertension in the population no less than
obesity. The PAR% of each risk factor varied by age group, and the total PAR% of the three modifiable risk
factors was considerably higher in the 3039 year old group (71%) than in the older groups.
Conclusions: Reduced alcohol intake and increased physical activity, as well as weight control, may
have a larger impact on prevention of hypertension in younger groups than in older groups.
Key words: primary prevention, attributable risk, follow-up study, hypertension, risk factors

140 mmHg and/or a diastolic blood pressure 90 mmHg (3)),

using annual health examination data of a Japanese telecommunication company (4).
The participants in the annual health examination were asked
to complete a structured and comprehensive questionnaire about
their health conditions and lifestyles. Questions about alcohol
drinking included weekly frequency of alcohol intake and that
about physical activity included with or without regular exercise.
Height and weight were measured while wearing an examining
gown without shoes. Systolic and diastolic blood pressures were
measured after five minutes of rest in a sitting position by welltrained nurses. For those with a systolic blood pressure
140 mmHg and/or a diastolic blood pressure 90 mmHg, blood
pressure was measured again after an additional five minutes of
rest. Blood samples (fasting) were taken for measurements of
biochemical variables. The above processes were conducted
according to the guidelines established by the employee health
management center of the company (4). Table 1 shows the baseline characteristics of the study subjects. Blood pressure, body
mass index and fasting blood glucose were classified according to
the Japanese expert committees guidelines: high-normal blood
pressure was defined as a systolic blood pressure of 130
139 mmHg and/or a diastolic blood pressure of 8589 mmHg (3);
obesity was defined as a body mass index 25 kg/m2 (5); glucose
intolerance was defined as taking antihyperglycemic therapy or a
fasting blood glucose 110 mg/dl (6). Based on information
collected through the questionnaire, those with alcohol drinking
5 days/week were classified as regular alcohol intake and those
without regular exercise were classified as physical inactivity.
Multivariate analyses were performed using the Coxs
proportional hazard models (7). The AR of each risk factor was

Introduction
Epidemiological studies have provided valuable information
on risk factors of chronic diseases. Relative risk (RR) and attributable risk (AR) are often used to estimate the magnitude of the
disease risk (1). A high RR indicates a major role in the causation
of disease, and RR measures can offer appropriate criteria for a
high-risk individual who requires intensive interventions to minimize the disease risk. On the other hand, a high AR indicates a
great potential to prevent the disease in populations, and AR
measures can offer appropriate guidelines for public health recommendations. Neither risk measure is a substitute for the other (1).
Although major studies have so far focused on RR rather than
AR (2), epidemiological findings with RR alone provide insufficient information for evidence-based public health. The present
study aimed to examine the effective preventive strategy for
hypertension in a Japanese male population, based on attributable risk measures.

Subjects and Methods

A cohort of normotensive male office workers aged 30 to
59 years (n=6,306) was followed up between 1991 and 1998
inclusive to observe the development of hypertension (either initiation of antihypertensive therapy or a systolic blood pressure
Received Oct. 11 2001/Accepted Jan. 4 2002
Reprint requests to: Machi SUKA
Department of Preventive Medicine, St. Marianna University School of
Medicine, 2-16-1 Sugao, Miyamae-ku, Kawasaki, Kanagawa 216-8511, Japan
TEL: +81(44)977-8111, FAX: +81(44)977-8356
E-mail: [email protected]
79

Preventive Strategy Based on AR Measures

Table 1

Baseline characteristics of study subjects

age (y)
all subjects
3039
factor

(n=6306)

4049

(n=561)

5059

(n=5283)

(n=462)

blood pressure, mmHg

<120/80 (optimal)
120129/8084 (normal)
130139/8589 (high-normal)

2,530
1,869
1,907

40.1
29.7
30.3

213
184
164

38.0
32.8
29.2

2,156
1,553
1,574

40.8
29.4
29.8

161
132
169

34.8
28.6
36.6

body mass index, kg/m2

<18.5 (thin)
18.524.9 (normal)
25(obese)

242
4,752
1,312

3.8
75.4
20.8

20
393
148

3.6
70.1
26.4

203
4,005
1,075

3.8
75.8
20.3

19
354
89

4.1
76.6
19.3

fasting blood glucose, mg/dl

<110 (normal)
110(glucose intolerance)

5,804
502

92.1
8.0

529
32

94.3
5.7

4,871
412

92.2
7.8

404
58

87.4
12.6

alcohol drinking
<5 days/week (no regular intake)
5 days/week (regular intake)

1,804
4,502

28.6
71.4

147
414

26.2
73.8

1,494
3,789

28.3
71.7

163
299

35.3
64.7

physical activity
regular exercise (active)
no regular exercise (inactive)

852
5,454

13.5
86.5

69
492

12.3
87.7

718
4,565

13.6
86.4

65
397

14.1
85.9

classified according to the expert committee guidelines from the Japanese Society of Hypertension.
classified according to the expert committee guidelines from the Japanese Society of Obesity.

classified according to the expert committee guidelines from the Japanese Society of Diabetes Mellitus.

and PAR% of each risk factor. In terms of RR, obesity (1.52,

95%CI: 1.381.66) was ranked first among the three modifiable
risk factors; regular alcohol intake (1.21, 95%CI: 1.101.34) and
physical inactivity (1.17, 95%CI: 1.031.33) followed after
obesity. On the other hand, in terms of the PAR%, regular alcohol
intake (13%, 95%CI: 720%) and physical inactivity (13%,
95%CI: 322%) were ranked beside obesity (10%, 95%CI: 7
12%).
Although lifestyles and obesity often play minor roles in the

estimated by population attributable risk percentage (PAR%) from

Levins formula (8).

Results and Discussions

After adjusting for age at the baseline, obesity, regular alcohol intake and physical inactivity, in addition to high-normal blood
pressure at the baseline and glucose intolerance, were recognized
as independent risk factors of hypertension. Table 2 shows the RR
Table 2

Relative risks and population attributable risk percentages for development of hypertension
age (y)
all subjects
3039
factor

(n=6306)

4049

(n=561)

(n=5283)
RR

PAR%

(95%CI)

(95%CI)

(95%CI)

(95%CI)

41
(2753)

3.73
(3.404.09)

45
(4248)

3.13
(2.304.25)

44
(3255)

1.72
(1.182.50)

16
(428)

1.57
(1.421.74)

10
(813)

1.08
(0.761.53)

1
(09)

1
(02)

1.79
(1.053.06)

4
(011)

1.07
(0.921.25)

1
(02)

1.28
(0.861.92)

3
(010)

1.21
(1.101.34)

13
(720)

1.30
(0.822.06)

18
(044)

1.16
(1.041.30)

11
(318)

1.47
(1.052.06)

23
(341)

1.17
(1.031.33)

13
(322)

1.68
(0.893.22)

37
(066)

1.20
(1.041.38)

15
(425)

0.82
(0.561.22)

(016)

PAR%

RR

(95%CI)

(95%CI)

(95%CI)

(95%CI)

high-normal blood pressure

3.62
(3.323.94)

44
(4147)

3.35
(2.294.90)

obesity

1.52
(1.381.66)

10
(712)

glucose intolerance

1.13
(0.981.29)

regular alcohol intake

physical inactivity

RR

(n=462)

PAR%

PAR%

RR

5059

RR: relative risk, PAR%: population attributable risk percent, CI: confidence interval.

calculated using the Cox's proportional hazards model, included simultaneously age(y) and all factors listed in the table.

calculated using Levin's formula.

refer to Table 1.
80

Preventive Strategy Based on AR Measures

Due to using the annual health examination data of a

company, the present study has the following potential limitations.
First, the definitions of the population at risk and the cases of developing hypertension depended on casual blood pressure readings.
As stated in previous studies, variability in measurements of
casual blood pressure makes it difficult to establish both normotension at the baseline and subsequent development of hypertension (3, 10). However, the diagnosis of hypertension in this study
was based on two measurements of blood pressure. Second,
those who had retired or had been transferred to another locality
(censored cases) were more likely to be seen in the 5059 years
old group. However, the mean (SD)s of the follow-up duration in
the 3039, 4049 and 5059 years old groups were 6.7 (0.8), 6.4
(1.3) and 6.3 (1.3) years, respectively, and there was no significant
difference among them. Thus, the sampling bias might have
negligible effects on the findings of this study. Finally, only qualitative information on alcohol drinking and physical activity was
used in the risk estimation. Previous studies suggested doseresponse relation of alcohol intake and physical activity in the
development of hypertension (1113). Moreover, some other lifestyles have been identified as risk factors of hypertension. In addition to weight control, increased physical activity, and reduced
alcohol intake, reduced salt intake is recommended as a nonpharmacological intervention on blood pressure (3, 14, 15). Lifestyle modifications may have a greater impact on the prevention of
hypertension than that expected in the present study. Further
studies may be required to improve the risk estimation.
Risk assessment based on population characteristics is important for evidence-based public health. Health professionals need
useful risk measures in the population for judgment of target risk
factors. Despite of the potential limitations of the present study,
the AR measures by age group can provide valuable information
for a population strategy.

causation of disease (low RR), their prevalence rates are high

(high AR): in the present study, regular alcohol intake, physical
inactivity and obesity accounted for 71% and 87% and 21%,
respectively. Rose has suggested two approaches for disease
prevention, that is, a high-risk strategy and a population strategy
(9). If means of improving a risk factor are safe enough for public
health recommendations, the latter approach is reasonable in those
with low RR and high AR rather than the former approach (9).
Magnitude ranking based on AR measures (obesity=regular alcohol
intake=physical inactivity) was different from that based on RR
measures (obesity>regular alcohol intake=physical inactivity).
Accordingly, priorities in preventive strategies for hypertension
should be different between the high-risk approach (i.e. weight
control is important) and the population approach (i.e. reduced
alcohol intake and increased physical activity, as well as weight
control, are important).
Moreover, the PAR% varied by age group with variations in
both the prevalence rate and the RR of each risk factor (Table 2).
The contribution of physical inactivity was greatest in the 30
39 years old group (37%), whereas that of regular alcohol intake
was greatest in the 5059 years old group (23%). The total PAR%
of the modifiable risk factors was considerably higher in the 30
39 years old group (71%) than in the older groups.
To promote a more effective preventive strategy for hypertension in the population, health professionals should change their
target risk factors by age group: increased physical activity should
be focused in the younger group; reduced alcohol intake should be
focused in the older group. Due to the high total PAR% of the
three modifiable risk factors in the 3039 years old group, reduced
alcohol intake and increased physical activity, as well as weight
control, may have a larger impact on prevention of hypertension in
the younger group than in the older groups. This reconfirms the
importance of early health education regarding good lifestyles and
ideal weight for the primary prevention of the disease.

References
disease. BMJ. 1981; 282: 18471851.
(10) Parati G, Ulian L, Santucciu C, Tortorici E, Villiani A, Di Rienzo
M, Mancia G. Clinical value of blood pressure variability. J.
Hypertens. Suppl. 1998; 16: S25S33.
(11) Nakanishi N, Nakamura K, Ichikawa S, Suzuki K, Kawashimo
H, Tatara K. Risk factors for the development of hypertension: a
6-year longitudinal study of middle-aged Japanese men. J.
Hypertens. 1998; 16: 753759.
(12) Tsuruta M, Adachi H, Hirai Y, Fujiura Y, Imazumi T. Association between alcohol intake and development of hypertension in
Japanese normotensive men: 12-year follow-up study. Am. J.
Hypertens. 2000; 13: 482487.
(13) Hayashi T, Tsumura K, Suematsu C, Okada K, Fujii S, Endo G.
Walking to work and the risk for hypertension in men: the Osaka
Survey. Ann. Intern. Med. 1999; 131: 2126.
(14) Ebrahim S, Smith GD. Lowering blood pressure: a systematic
review of sustained effects of non-pharmacological interventions. J. Public Health Med. 1998; 20: 441448.
(15) Cutler JA, Follmann D, Allender PS. Randomized trials of
sodium reduction: an overview. Am. J. Clin. Nutr. 1997; 65
(suppl): 643S651S.

( 1 ) Editorial. Relative or attributable risk? Lancet 1981; 2: 1211

1212.
( 2 ) Nakayama T, Zaman MM, Tanaka H. Reporting of attributable
and relative risks, 19661997. Lancet 1998; 351: 1179.
( 3 ) Expert committee in the Japanese Society of Hypertension.
Therapeutic guideline of hypertension 2000. Kyourinsya, Tokyo.
Japanese.
( 4 ) Suka M, Sugimori H, Iida Y, Yoshida K. Risk factors for hypertension: a longitudinal study of middle-aged Japanese male
workers. Jpn. J. Public Health 2001; 48: 543550. Japanese.
( 5 ) Expert committee in the Japanese Society of Obesity. Diagnostic
guideline of obesity. Obesity research 2000; 26: 1828.
Japanese.
( 6 ) Expert committee in the Japanese Society of Diabetes Mellitus.
Therapeutic guideline of diabetes mellitus 1999. Bunkoudou,
Tokyo. Japanese.
( 7 ) Cox DR. Regression models and life-tables. J. Royal Stat. Soc.
Series B 1972; 34: 187220.
( 8 ) Daly LE. Confidence limits made easy: interval estimation using
a substitution method. Am. J. Epidemiol. 1998; 147: 783790.
( 9 ) Rose G. Strategy of prevention: lessons from cardiovascular

81

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.