Content uploaded by Chih-Hsing Wu
Author content
All content in this area was uploaded by Chih-Hsing Wu on Aug 10, 2018
Content may be subject to copyright.
ORIGINAL ARTICLES
A Population-based Study of the
Prevalence and Associated Factors of
Diabetes Mellitus in Southern Taiwan
Feng-Hwa Lu, Yi-Ching Yang, Jin-Shang Wu, Chih-Hsing Wu, Chih-Jen Chang*
Department of Family Medicine, Medical College, National Cheng
Kung University, Tainan, Taiwan
Diabetes mellitus is one of the major health care problems in Taiwan, since the mortality
rate has increased from 7.91 per 100 000 in 1980 to 35.1 per 100 000 in 1996. To
determine the prevalence of diabetes in southern Taiwan and to investigate possible
associated factors, a stratified systematic cluster sampling of 1638 subjects (780 men and
858 women) aged ⱖ20 years living in Tainan city was investigated with a standard 75-g
oral glucose tolerance test. The crude prevalence of diabetes in Tainan was 9.0 % (10.3 %
men and 7.9 % women) and the age-adjusted prevalence was 9.2 % (10.4 % men and
8.1 % women). The crude prevalence of IGT was 14.0 % (13.8 % men and 14.1 %
women), and the age-adjusted prevalence was 15.5 % (15.0 % men and 15.9 % women).
The prevalence of diabetes by using the revised new diagnostic criteria was 7.5%. The
prevalence of diabetes and IGT increased significantly with age for both genders, although
the rises in prevalence of IGT in women was less consistent. Diabetic and IGT subjects
were older and had higher levels of BMI, triglyceride, systolic and diastolic blood pressure,
and higher prevalence of obesity, hypertension, and dyslipidemia but indulged in less
physical activity than non-diabetic subjects. The significant factors associated with the
newly diagnosed diabetes were age, family history of DM, BMI, systolic blood pressure,
physical activity, and serum triglyceride levels. 1998 John Wiley & Sons, Ltd.
Diabet. Med. 15: 564–572 (1998)
KEY WORDS
epidemiology; Type 2 diabetes mellitus, IGT; prevalence; Chinese
Received 4 September 1997; revised 6 January 1998; accepted 12 February 1998
Introduction
Diabetes mellitus (DM) in adults is a global health
problem. Although its prevalence varies widely between
different populations from low (⬍3 %), moderate (3–
10 %), high (11–20 %) to extremely high (⬎20 %), the
rate has generally increased worldwide.
1–4
To ascertain
the prevalence of diabetes through population-based
studies becomes an important issue for planning of
health services, analysing associated factors, evaluating
medical trends, and estimating community impact. In
Taiwan, the mortality rate of diabetes mellitus has almost
doubled over the past 10 years and its ranking in the
leading causes of death has increased: twelfth in 1982,
eighth in 1983, seventh in 1984 and fifth since 1987.
5
It has become one of the major health problems of the
country. The prevalence of diabetes in northern Taiwan
was previously established between 1985 to 1991 and
Abbreviations: BMI body mass index, CHO cholesterol, BP diastolic
blood pressure, FPG fasting plasma glucose, IFG impaired fasting
glucose, IGT impaired glucose tolerance, SBP systolic blood pressure,
TG triglyceride, WHR waist-hip ratio
Contract grant nos: DOH-85-TD-043 and DOH-86-TD-071
Sponsors: Department of Health, Taiwan; Department of Family
Medicine, National Cheng Kung University Hospital, Tainan, Taiwan
*Correspondence to: Dr Chih-Jen Chang, Department of Family
Medicine, National Cheng Kung University Hospital, 138 Sheng-Li
Road, Tainan, Taiwan 704, China
564
CCC 0742–3071/98/070564–09$17.50
1998 John Wiley & Sons, Ltd. DIABETIC MEDICINE, 1998; 15: 564–572
the rates were between 2.6 % and 5.8 %.
6–8
However,
none of these studies subjected the total study population
to the standard oral glucose tolerance test (OGTT). Most
of them used overnight fasting plasma glucose or capillary
whole blood glucose as a first step screening, then the
subjects whose fasting glucose level was between 5.6
and 7.8 mmol l
−1
received a 75 g OGTT. Therefore, the
prevalence of diabetes may be underestimated, since
some subjects with overnight fasting plasma glucose
⬍5.6 mmol l
−1
could have had OGTT 2-h glucose level
ⱖ11.1 mmol l
−1
.
8
The prevalence of impaired glucose
tolerance (IGT) also cannot be estimated by the two-
step method. For better comparison of prevalence data
in different populations, the standardized criteria of
World Health Organization (WHO)
9
should be applied
to determine the exact prevalence of diabetes and IGT
in Taiwan. The comparison of the prevalence of diabetes
between the current WHO criteria and the revised new
criteria of the Expert Committee of the American Diabetes
Association
10
is also potentially important. Moreover,
prevention of DM by identification and avoidance of its
risk factors, such as obesity and sedentary lifestyle are
becoming important issues in public health.
11,12
A
community-based epidemiological study was conducted
by using OGTT from January to December in 1996 to
determine the prevalence of diabetes and IGT in the
people who were aged ⱖ20 years living in Tainan
ORIGINAL ARTICLES
city, and to investigate possible risk factors associated
with DM.
Patients and Methods
The study was performed in Tainan city, the oldest city
in southern Taiwan, with a population of 700 000. The
sampling scheme was a three-stage process that generated
a stratified systemic cluster sample of households through-
out the city. First, areas of the city were grouped into
seven strata according to the administrative districts. One
area (Li) was randomly selected from each stratum. At
the second stage, every fifth household within each of
the seven selected areas were identified systematically.
In the third stage, the selected households were informed
about the survey by letter and telephone from the
medical centre to ask the consent for participation and
to arrange the schedule of examination. All of the
subjects whose age was 20 years or more according to
the government population register in 1995 were included
in the study.
With the help of specially trained assistants, all subjects
were interviewed according to a structured questionnaire.
The questionnaire included demographic information,
past medical history of diabetes and hypertension,
medication history, current smoking habit, current drink-
ing habit, physical activity over the past year, and family
history of DM and hypertension. Height and weight were
measured in light clothing without shoes. Body mass
index (BMI) was calculated as weight ×height
−2
(kg m
−2
).
Obesity was defined as a BMI ⱖ27 for men and ⱖ25
for women, according to the National Diabetes Data
Group criteria.
13
Waist (midway between the lower rib
margin and the iliac crest) and hip (widest circumference
over the great trochanters) circumferences were measured
with subjects standing relaxed without clothing and with
a standard dressmaker’s tape after normal expiration
14
and the waist-to-hip circumference ratio (WHR) was
calculated as an index of central obesity.
All subjects who did not have a medical history of
diabetes received a 75-g OGTT after a 10-h overnight
fast, a normal diet for 3 days before the test, and
abstention from smoking for ⬎24 h. None of the women
were pregnant when tested. Fasting venous blood was
sampled for assessment of serum lipids, lipoproteins and
plasma glucose levels and kept frozen (−80°C) until
analysis. Plasma glucose was analysed by the standard
glucose-oxidase method (Synchron CX3, Beckman). The
coefficients of variation (CV) for intra-assay was 1.2 %
and for inter-assay was 1.5 % at the upper end of the
reference range. The presence of DM was assessed
using questions about the person’s previously diagnosed
diabetes and using a 2-h OGTT. The participants were
designated to have previously diagnosed diabetes if in
either of the above questionnaires he/she reported having
diabetes treated with diet, oral hypoglycaemic agents or
insulin. They were tested only for fasting plasma glucose
levels. All the subjects who underwent a 2-h OGTT
565
POPULATION-BASED STUDY OF PREVALENCE OF DIABETES
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
were classified according to the current WHO criteria
9
for DM and IGT, and revised new criteria
10
for DM
and IFG (impaired fasting glucose). The subjects were
determined to have newly diagnosed diabetes if they
were not previously diagnosed diabetes, but the 2-h
OGTT demonstrated a diabetic level.
Fasting serum total cholesterol and triglyceride were
measured enzymatically using automated methods.
15,16
Intra-assay coefficients of variation (CVs) for total choles-
terol and triglyceride were 1.8 % and 1.2 %, respectively;
and the inter-assay CVs were 3.5 % and 2.5 %, respect-
ively. Hypercholesterolaemia was defined if serum chol-
esterol level was ⱖ5.18 mmol l
−1
,
17
and hypertriglyceri-
daemia was defined if serum triglyceride level was
ⱖ2.26 mmol l
−1
.
Two readings of systolic and diastolic blood pressure
were measured in the sitting position with a DINAMAP
vital sign monitor (Model 1846SX, Critikon Inc., Irvine,
CA)
18
after a 15-min rest period while subjects awaited
the second venepuncture for OGTT. Hypertension was
defined according to the WHO criteria:
19
a mean
blood pressure ⱖ160 mmHg or diastolic blood pressure
ⱖ95 mmHg, or a history of hypertension and currently
receiving treatment.
Total physical activity assessment was calculated in
MET-hours per week over all activities for the past year.
One MET (metabolic equivalent) represents the energy
expended at rest for an individual, expressed in kcal ×kg
−1
(body wt) ×h
−1
. First, the average number of hours per
week spent for each activity was calculated, including
leisure activity, occupational activity, and walking for
exercise. Then, hours per week of each activity was
multiplied by an estimate of the metabolic cost of that
activity, expressed as MET, and the MET-hours per week
for each activity was summed over all activities.
20–22
Statistical Methods
Data storage and retrieval were performed with DBASE
III PLUS software. Data analysis was conducted using
the Statistical Package for the Social Sciences for
Windows (SPSSWIN) statistical software. Truncated age-
adjusted prevalence rates were performed by a direct
method
23
based on the standard world population of
Segi.
24
Since triglyceride levels and physical activity
were clearly highly skewed, they were presented and
analysed by using log transformation. Due to a large
difference in mean age among different diabetic groups,
age was adjusted in the comparisons. The ANOVA test
and
2 test were used to test the differences of variables
among IGT, newly diagnosed, previously diagnosed, and
non-diabetic subjects. For the possible risk factor analysis,
multiple logistic regression was used to test for differences
between newly diagnosed diabetes and non-diabetes.
The level of significance was set at 0.05.
ORIGINAL ARTICLES
Results
A total of 2416 subjects were eligible for inclusion, 1170
(48.4 %) men and 1246 (51.6 %) women. Of the 2416
eligible people, 1638 (67.8 %) participated in the study,
780 (47.6 %) men and 858 (52.4 %) women. The non-
responders were slightly younger in age and consisted
of more males compared with the responders, but
the differences were not statistically significant (data
not shown).
Table 1 shows the prevalence of diabetes and IGT by
age and sex. The overall crude prevalence of diabetes
in subjects ⱖ20 years of age was 9.0 % (4.8 % previously
diagnosed and 4.2 % newly diagnosed). It was higher in
men (10.3 %) than in women (7.9 %), but there was
no significant difference. Both genders had a higher
prevalence of previously diagnosed diabetes (5.4 % men
and 4.4 % women) compared to newly diagnosed
diabetes (4.9 % men and 3.5 % women). Based on the
standard world population of Segi, the age-adjusted
prevalences of diabetes were 9.2 % for total (men 10.4 %,
women 8.1 %), 4.8 % for previously diagnosed, and
4.4 % for newly diagnosed diabetes. The prevalence of
diabetes increased with age for both genders. In the age
groups of 20–29 years, and 30–39 years, only newly
diagnosed diabetes was found for both genders. In the
age groups of 40–49 years, the number of newly
diagnosed diabetic subjects was more than the previously
diagnosed diabetic subjects, but in the age groups of 50
years and over, the prevalence of previously diagnosed
subjects was greater than the newly diagnosed subjects.
The overall crude prevalence of IGT was 14.0 %. It was
higher in women (14.1 %) than in men (13.8 %), but
there was no significant difference. The prevalence of
IGT increased with age for men, but was less consistent
for women. The age-adjusted prevalences of IGT were
15.5 % for total (men 15.0 %, women 15.9 %).
Table 2 shows the comparison of prevalence of diabetes
and IGT between the current WHO criteria and the
revised new criteria. Based on a fasting plasma glucose
(FPG) ⱖ7 mmol l
−1
(126 mg dl
−1
), prevalence of newly
diagnosed diabetes (2.6 %) was lower than that was
diagnosed from the use of the OGTT (4.2 %). The
prevalence of IFG was 3.9 %, which was much lower
than the prevalence of IGT (14 %).
Table 3 shows the clinical characteristics of the non-
diabetic, IGT, newly diagnosed, and previously diagnosed
diabetic subjects. The total diabetic and IGT subjects had
a higher mean age, body mass index (BMI), total
triglyceride (TG), systolic blood pressure (SBP), diastolic
blood pressure (DBP), fasting plasma glucose (FPG), 2-h
post-load plasma glucose, but lower total physical activity
levels than non-diabetic subjects. Only IGT subjects had
higher serum cholesterol (CHO) levels than non-diabetic
subjects. There was no significant difference in waist-to-
hip ratio (WHR) between non-diabetic and diabetic or
IGT groups. The diabetic and IGT subjects also had a
higher prevalence of obesity, hypertension, hypercholester-
566
FENG-HWA LU ET AL.
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
olaemia, and hypertriglyceridaemia than non-diabetic
subjects. The prevalence of family history of DM in the
diabetic groups was higher than non-diabetic subjects.
There was no significant difference in the prevalence of
smoking and drinking history between any two groups.
The IGT subjects had a lower mean WHR, FPG, 2-h post-
load plasma glucose, but had higher total physical activity
levels than total diabetic subjects. The IGT subjects had
a lower prevalence of obesity, and hypertension, but did
not have a different prevalence of hypercholesterolaemia
and hypertriglyceridaemia. The newly diagnosed and
previously diagnosed diabetic subjects had a higher mean
BMI, TG, SBP, FPG, and 2-h post-load plasma glucose
levels, and had a lower mean physical activity level than
non-diabetic subjects, but showed no significant difference
in WHR and CHO. Subjects with newly diagnosed and
previously diagnosed diabetes had a higher prevalence of
obesity, hypertension, hypercholesterolaemia, and hyper-
triglyceridaemia than non-diabetic subjects. Only subjects
with previously diagnosed diabetes had a higher preva-
lence of family history of DM than non-diabetic subjects.
The previously diagnosed diabetic subjects had signifi-
cantly higher FPG, but had lower SBP and DBP than
the newly diagnosed diabetic subjects. There were no
significant differences in age, BMI, WHR, CHO, TG, and
total physical activity between previously and newly
diabetic subjects.
Table 4 shows odds ratios and 95 % confidence
intervals for independent variables with newly diagnosed
diabetes in the multiple logistic regression model. The
best logistic regression model was fitted with a stepwise-
forward strategy. The independent variables were age,
sex, family history of DM, BMI, SBP, DBP, physical
activity, TG, CHO, smoking status, drinking status, and
WHR. According to the final model, the significant
factors associated with newly diagnosed diabetes were
age, family history of DM, BMI, SBP, physical activity,
and TG, but sex, DBP, CHO, smoking status, drinking
status, and WHR were not significantly correlated with
newly diagnosed diabetes.
Discussion
The present study is the first community-based epidemiol-
ogical investigation of the prevalence of diabetes by
using a standard OGTT for all subjects not being treated
for diabetes in Taiwan. It avoids the weakness of
underestimation of prevalence of diabetes by using OGTT
only in subjects with fasting plasma glucose larger than
5.6 mmol l
−1
, since subjects with fasting plasma glucose
less than 5.6 mmol l
−1
could have had 2-h glucose levels
equal to or higher than 11.1 mmol l
−1
if they had been
tested.
25
The overall crude prevalence of diabetes in this
study was 9.0 % and the age-specific standardized
prevalence was 9.2 %. In a comparison of the diabetes
prevalence studies in northern Taiwan (see Table 5), the
prevalence of our study is higher than that in the Taipei
city study,
6
the Kin-Hu study,
8
and the Ann-Lo study.
7
ORIGINAL ARTICLES
Table 1. Age and sex-specific prevalence of impaired glucose tolerance, newly diagnosed and previously diagnosed diabetes mellitus in Tainan, Taiwan
Age groups Number Total Men Women
IGT DM Previous New IGT DM Previous New IGT DM Previous New
PR PR cases cases PR PR cases cases PR PR cases cases
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
n
(%)
20–29 341 14 (4.1) 4 (1.2) 0 (0.0) 4 (1.2) 7 (4.4) 2 (1.3) 0 (0.0) 2 (1.3) 7 (3.8) 2 (1.1) 0 (0.0) 2 (1.1)
30–39 358 27 (7.5) 7 (2.0) 0 (0.0) 7 (2.0) 12 (7.1) 3 (1.8) 0 (0.0) 3 (1.8) 15 (7.9) 4 (2.1) 0 (0.0) 4 (2.1)
40–49 397 79 (19.9) 26 (6.5) 7 (1.8) 19 (4.8) 28 (15.8) 14 (7.9) 4 (2.3) 10 (5.6) 51 (23.2) 12 (5.5) 3 (1.4) 9 (4.1)
50–59 244 43 (17.6) 33 (13.5) 23 (9.4) 10 (4.1) 22 (20.0) 17 (15.5) 11 (10.0) 6 (5.5) 21 (15.7) 16 (11.9) 12 (9.0) 4 (3.0)
60–69 194 44 (22.7) 50 (25.8) 34 (17.5) 16 (8.2) 24 (23.3) 29 (28.2) 17 (16.5) 12 (11.7) 20 (22.0) 21 (23.1) 17 (18.7) 4 (4.4)
ⱖ70 104 22 (21.2) 28 (26.9) 16 (15.4) 12 (11.5) 15 (23.8) 15 (23.8) 10 (15.9) 5 (7.9) 7 (17.1) 13 (31.7) 6 (14.6) 7 (17.1)
Total 1638 229 (14.0) 148 (9.0) 80 (4.9) 68 (4.2) 108 (13.8) 80 (10.3) 42 (5.4) 38 (4.9) 121 (14.1) 68 (7.9) 38 (4.4) 30 (3.5)
Adjusted
a
15.5 9.2 4.8 4.4 15.0 10.4 5.2 5.2 15.9 8.1 4.5 3.6
PR, prevalence rate.
a
Age-adjusted rate by using of standard world population of Segi.
567
POPULATION-BASED STUDY OF PREVALENCE OF DIABETES
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
ORIGINAL ARTICLES
Table 2. Comparison of prevalences of diabetes mellitus and impaired glucose tolerance between the
current WHO criteria and the proposed new diagnostic criteria in Tainan in individuals ⱖ20 years old
Prevalence (%) of diabetes and IGT Total diabetes
by glucose criteria without a prevalence (%)
medical history of diabetes
Medical history of diabetes – 4.9
WHO criteria for IGT 14 –
FPG ⱖ6.1 and FPG ⬍7.0 mmol l
−1
(IFG) 3.9 –
WHO criteria for diabetes 4.2 9.0
FPG ⱖ7.0 mmol l
−1
2.6 7.5
IGT, impaired glucose tolerance; FPG, fasting plasma glucose; IFG, impaired fasting glucose.
The prevalence of diabetes in the Tainan study is the
highest rate that has been reported. Since the age of the
present study population was younger than the above
three Taiwanese studies, and the urbanization of the
present study was lower than in Taipei study, similar to
the Ann-Lo study and slightly higher than the Kin-Hu
study, we can be sure that an increase in the prevalence
of diabetes has been identified. The increase may be
related to an increasingly westernized life pattern during
the past decades in Tainan. In a comparison of the
diabetes prevalence studies in adult Chinese out of
Taiwan (see Table 5), the rate was higher than that in
the Da Qing, China study,
27
the Singapore study,
4
and
the Hong Kong study,
26
but lower than that of the
Mauritius study.
28
In a comparison of the prevalence
rate among Asian populations (see Table 5), the rate was
higher than that in the Vietnamese study,
29
the Korean
study,
3
close to that of the Japanese study,
2
but lower
than that in the Fiji Asian Indians study.
30
This moderate
prevalence of diabetes (3–10 %)
1
may be attributed to
the westernized life pattern of population in the process
of industrialization (environmental factor), genetic factor,
and interaction of both factors.
In the present study, the prevalence of diabetes was
higher among men than among women, which was the
same as the results in the Hong Kong study (men 5.1 %,
women 3.6 %), in the Bangladesh study (men 3.52 %,
women 1.13 %), and in the Korean study (men 10.6 %,
women 7.9 %). The prevalence rate of newly diagnosed
diabetes was close to that of previously diagnosed
diabetes, which has been shown in many studies.
31–33
This indicates that diabetes may be twice as prevalent
in Tainan city as rates found in medical history surveys.
Our study found that women had a lower prevalence of
newly diagnosed diabetes than men. It may be attributed
to a higher utilization of medical service in women,
so diabetes was easily diagnosed during screening
programmes. This can be seen from the higher partici-
pation rate in women in this study and other studies.
2,8
In the age groups from 20–29 years to 30–39 years, only
newly diagnosed diabetes was found. Whether to do
diabetes screening in younger age groups especially in
the high risk subjects who have a family history of DM,
obesity, hypertension, physical inactivity, and hypertrigly-
568
FENG-HWA LU ET AL.
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
ceridaemia, should be further investigated since the
newly diagnosed diabetic subjects were commonly
asymptomatic. The higher prevalence of diabetes in older
age groups identified in this study was similar to the
findings of other studies.
31,34
The crude prevalence of IGT was 14.0 %, and the
age-specific standardized prevalence was 15.5 %. Since
the studies in Taiwan previously did not use standardized
OGTT, prevalences of IGT in those studies could not be
determined. In a comparison of the IGT prevalence
studies in adult Chinese outside of Taiwan (see Table 5),
the rate was higher than that in the Da Qing, China
study,
27
the Singapore study,
4
and the Hong Kong study,
26
but close to that of the Mauritius study.
28
Comparison
of the prevalence rate among Asian populations (see
Table 5), showed the rate was higher than that in the
Vietnamese study,
29
the Korean study,
3
and the Fiji Asian
Indians study,
30
but close to that of the Japanese study.
2
Studies in different Chinese populations who have similar
genetic characteristics have shown substantial variation
in the prevalence of IGT. This result indicates that
environmental factors may play an important role, since
the levels of westernization were very different among
the Chinese in these areas.
The prevalence of IGT was higher among women than
men, which was the same as the Mauritius study,
28
and
the Fiji Asian Indians study.
30
However, the difference
in prevalence IGT between men and women varies
widely across populations.
1
The prevalence of IGT
increased with age, which was the same as the US study.
31
The prevalence of diabetes without a medical history
of diabetes (newly diagnosed diabetes) by using the
revised new criteria was 2.6 %, which was lower than
that using WHO criteria (4.2 %). The same result was
also found in the US NHANES III study.
10
From univariate
analyses, after adjusting for age, the study showed that
the newly and previously diagnosed diabetic subjects
had higher mean BMI, TG, SBP, DBP, FPG, 2-h post-
load plasma glucose than non-diabetic subjects, which
was also demonstrated before.
35
This means that the
newly and previously diagnosed diabetic subjects had
the same unfavourable risk factors for coronary heart
disease, so it is worthwhile for us to consider the
problems found in the newly diagnosed diabetic subjects.
ORIGINAL ARTICLES
Table 3. Comparisons of clinical characteristics of non-diabetic, IGT, newly diagnosed diabetic and previously diagnosed diabetic subjects
Non-diabetic IGT Diabetes mellitus
New cases Previous cases Total cases
Number 1261 229 68 80 148
Age (yr) 40.7⫾14.4 50.9 ⫾13.8
c
54.3 ⫾15.4
c
61.3 ⫾8.5
c
58.1 ⫾12.6
c
BMI (kg m
−2
) 23.2⫾3.4 25.1 ⫾3.9
c
25.6 ⫾3.5
b
25.8 ⫾3.3
b
25.7 ⫾3.4
c
WHR 0.845⫾0.270 0.890 ⫾0.085 0.920 ⫾0.078 0.946 ⫾0.077
e
0.934 ⫾0.078
e
Cholesterol (mmol l
−1
) 4.9⫾1.1 5.3 ⫾1.0
b
5.4 ⫾1.4 5.3 ⫾1.4 5.4 ⫾1.4
Triglyceride (mmol l
−1
)
j
0.040⫾0.235 0.177 ⫾0.224
c
0.249 ⫾0.330
c
0.232 ⫾0.255
c
0.246 ⫾0.291
c
Systolic blood pressure (mmHg) 115.0 ⫾18.5 127.8 ⫾22.1
c
137.0 ⫾28.5
c,d
133.9 ⫾21.6
c,h
135.3 ⫾25.0
c
Diastolic blood pressure (mmHg) 69.5⫾10.2 74.7 ⫾11.7
c
79.5 ⫾11.7
c,d
77.1 ⫾10.1
g
78.2 ⫾10.9
c
Fasting plasma glucose (mmol l
−1
) 5.0⫾0.5 5.4 ⫾0.6
c
7.9 ⫾3.2
c,f
9.5 ⫾3.7
c,f,h
8.8 ⫾3.6
c,f
2-h plasma glucose (mmol l
−1
) 5.5⫾1.2 8.9 ⫾0.9
c
14.7 ⫾4.4
c,f
15.4 ⫾7.7
c,f
15.0 ⫾6.4
c,f
Physical activity (MET-hr week
−1
)
j
1.65⫾0.47 1.59 ⫾0.48
a
1.48 ⫾0.42
b
1.37 ⫾0.55
c,d
1.42 ⫾0.49
c,d
Sex
Male 592 (46.9) 108 (47.2) 38 (55.9) 42 (52.5) 80 (54.1)
Female 669 (53.1) 121 (52.8) 30 (44.1) 38 (47.5) 68 (45.9)
Smoking
No 995 (78.9) 185 (80.8) 51 (75.0) 63 (78.8) 114 (77.0)
Yes 266 (21.1) 44 (19.2) 17 (25.0) 17 (21.3) 34 (23.0)
Drinking
No 1106 (87.7) 193 (84.3) 56 (82.4) 71 (88.8) 127 (85.8)
Yes 155 (12.3) 36 (15.7) 12 (17.6) 9 (11.3) 21 (14.2)
Family history of DM
No 1036 (82.2) 182 (79.5) 52 (76.5) 58 (72.5)
a
110 (74.3)
a
Yes 225 (17.8) 47 (20.5) 16 (23.5) 22 (27.5) 38 (25.7)
Obesity
BMI ⬍27 for M; ⬍25 for F 1015 (80.6) 154 (67.2)
c
40 (59.7)
c
43 (53.8)
c,d
83 (56.5)
c,d
BMI ⱖ27 for M; ⱖ25 for F 244 (19.4) 75 (32.8) 27 (40.3) 37 (46.3) 64 (43.5)
Hypertension
⬍160/95 mmHg 1154 (91.5) 176 (76.9)
c
43 (63.2)
c,d
55 (68.8)
c
98 (66.2)
c,d
ⱖ160/95 mmHg 107 (8.5) 53 (23.1) 25 (36.8) 25 (31.3) 50 (33.8)
Hypercholesterolaemia
⬍5.18 mmol l
−1
803 (63.7) 111 (48.5)
c
32 (47.1)
a
41 (51.3)
a
73 (49.3)
c
ⱖ5.18 mmol l
−1
458 (36.3) 118 (51.5) 36 (52.9) 39 (48.8) 75 (50.7)
Hypertriglyceridaemia
⬍2.26 mmol l
−1
1139 (90.3) 179 (78.2)
c
45 (66.2)
c,d
59 (73.8)
c
104 (70.3)
c
ⱖ2.26 mmol l
−1
122 (9.7) 50 (21.8) 23 (33.8) 21 (26.3) 44 (29.7)
Non-diabetics vs IGT, newly diagnosed, previously diagnosed, and total diabetics:
a
p
⬍0.05,
b
p
⬍0.01,
c
p
⬍0.001.
IGT vs newly diagnosed, previously diagnosed, and total diabetics:
d
p
⬍0.05,
e
p
⬍0.01,
f
p
⬍0.001.
Newly diagnosed vs previously diagnosed:
g
p
⬍0.05,
h
p
⬍0.01,
i
p
⬍0.001.
j
With log transformation; ANOVA test (age adjusted) or
2
test.
Data are presented either with % in brackets or as mean ⫾SD.
569
POPULATION-BASED STUDY OF PREVALENCE OF DIABETES
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
ORIGINAL ARTICLES
In several studies,
11,35
previously diagnosed diabetic
subjects had significantly higher FPG than newly diag-
nosed diabetic subjects, as was also demonstrated by
our study. This indicates that previously diagnosed
diabetic subjects did not have their blood glucose under
good control.
11,35
The result that newly diagnosed diabetic
subjects had significantly higher SBP than previously
diagnosed diabetic subjects also was found in the Pu-Li
study.
11
This result may be explained by having good
blood pressure control in previously diagnosed diabetic
subjects during treatment of their diabetes.
Subjects with IGT had the intermediate levels of risk
factors for coronary heart disease, such as age, BMI, TG,
SBP, DBP, FPG, 2-h plasma glucose, and physical
Table 4. Multiple logistic regression analysis on the newly
diagnosed diabetes mellitus in Tainan, Taiwan
Variables Odds 95% C.I.
ratio
Age (yr) 1.04 1.02–1.07
Sex (men =0; women =1) 0.91 0.50–1.66
Family history of DM (no =0; 2.07 1.04–4.08
yes =1)
BMI (kg m
−2
) 1.11 1.02–1.21
SBP (mmHg) 1.02 1.01–1.04
Physical activity (MET-hr wk
−1
)
a
0.33 0.18–0.62
Triglyceride (mmol l
−1
)
a
7.44 2.33–23.76
a
Log transformation; dependent variable is newly diagnosed diabetes
vs non-diabetes. Independent variables are age, sex, family history of
DM, body mass index (BMI), systolic blood pressure (SBP), diastolic
blood pressure (DBP), physical activity, triglyceride, cholesterol,
smoking, drinking, and waist-to-hip ratio (WHR).
Table 5. Prevalence of diabetes and IGT in Chinese and Asian populations
DM IGT Reference
Age range Sample Study Crude Adjusted Crude Adjusted
Country/Population (yr) size year rate rate rate rate
Chinese (in Taiwan)
Tainan ⱖ20 1 638 1996 9.0 9.2 14.0 15.5
d
Taipei
a
ⱖ40 11 478 1985 6.2 5.8 – – 6
Ann-Lo, Keelung
b
ⱖ40 2 448 1988 8.0 8.0 – – 7
Kin-Hu, Kinmen
c
ⱖ30 3 236 1991 6.5 4.9 – – 8
Chinese (outside of
Taiwan)
Da Qing, China 25–74 110 660 1986 0.77 1.6 0.6 0.9 27
Singapore ⱖ18 1 417 1985 4.0 M4.6,F4.9 0.6 M0.8,F0.3 4
Hong Kong 18–64 1 513 1991 4.5 7.7 7.3 M11.2,F8.8 26
Mauritius 25–74 412 1987 11.7 11.5 16.5 M16.5,F21.7 28
Asian
Vietnam ⱖ15 4912 1990 1.2 1.4 1.6 1.7 29
South Korea ⱖ30 2 497 1991 9.1 7.2 11.8 8.9 3
Japan ⱖ45 916 1990 11.9 10.4 16.6 15.3 2
Fiji, Indian (urban) ⱖ20 846 1980 13.1 M12.9,F11.0 10.3 M8.3,F11.8 30
a
Capillary blood glucose either after overnight fasting or 2-h postprandial by use of glucometer.
b
2-h postprandial blood glucose (Glucocheck SC B-2) or urinary glucose and adjusted with the population in Taiwan.
c
OGTT for subject with fasting plasma glucose levels between 5.6 and 7.8 mmol l
−1
.
d
Present study: M: male, F: female.
570
FENG-HWA LU ET AL.
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
activity, which has also been reported in a US study.
36
This indicates that IGT has the similar risk factors for
coronary heart disease and may be seen to be an
intermediate stage in diabetic development.
36
From the results of multiple logistic regression, the
significant risk factors associated with newly diagnosed
diabetes were age, family history of DM, BMI, SBP,
physical activity and TG. Age as an independent predictor
of diabetes has been reported in many studies.
37,38
Family
history of DM as a significant predictor for diabetes has
also been demonstrated in several studies.
3,11,39–41
The
relationship between physical activity and diabetes is
controversial depending on the questionnaire used.
21
In
our study, the structured questionnaire
20
accurately
assessed the activity patterns in work, walking and leisure
time; the results showed that higher physical activity can
reduce the risk of diabetes. Many prospective studies
have also shown that increased physical activity is
associated with a reduced risk of diabetes.
42–44
The
mechanism may be related to increasing insulin sensitivity
in skeletal muscle caused by physical activity.
45
In addition, BMI,
39,40,46,47
SBP,
7,8,47–50
and TG
47,51–53
have also been associated with prevalence of diabetes
from many studies. It could be explained by insulin
resistance and hyperinsulinemia suggested by Reaven in
his Banting lecture in 1988
54
and from many other
studies.
55,56
The mechanism is that obese subjects are
resistant to insulin-stimulated glucose uptake, which in
turn leads to an increase in insulin concentration,
enhanced hepatic very-low-density lipoprotein triglycer-
ide secretion and hypertriglyceridaemia, and leads to a
reduction of sodium excretion, enhanced sympathetic
nervous system activity, and hypertension.
57
ORIGINAL ARTICLES
In conclusion, the crude prevalence of diabetes was
9.0 % and 7.5 % by using WHO criteria and revised new
criteria, respectively, and the age-adjusted prevalence of
diabetes and IGT were 9.2 % and 15.5 %. The possible
risk factors of newly diagnosed diabetes were age, family
history of DM, BMI, SBP, triglyceride, and physical
activity. In Taiwan, the mortality rates of diabetes were
greatly increased over the past 10 years. Reduction of
the modifiable risk factors such as BMI, hypertension,
and hypertriglyceridaemia, and increase of physical
activity through public health programmmes may help
to reduce the risk of diabetes.
Acknowledgements
This study was supported by grants from the Taiwan
Department of Health (DOH-85-TD-043 and DOH-86-
TD-071). We thank the personnel of the Department
of Family Medicine, National Cheng Kung University
Hospital, Tainan, Taiwan for their full support and help.
References
1. King H, Rewers M, WHO Ad Hoc Diabetes Reporting
Group. Global estimates for prevalence of diabetes mellitus
and impaired glucose tolerance in adults. Diabetes Care
1993; 16: 157–177.
2. Sekikawa A, Tominaga M, Takahashi K, Eguchi H, Igarashi
M, Ohnuma H, et al. Prevalence of diabetes and impaired
glucose tolerance in Funagata area, Japan. Diabetes Care
1993; 16: 570–574.
3. Park Y, Lee H, Koh CS, Min H, Yoo K, Kim Y, Shin Y.
Prevalence of diabetes and IGT in Yonchon County,
South Korea. Diabetes Care 1995; 18: 545–548.
4. Thai AC, Yeo PPB, Lun KC, Hughes K, Ng WY, Lui KF,
Cheah JS. Diabetes mellitus and its chronic complications
in Singapore: an increasing healthcare problem. Annals
Academy of Med, Singapore 1990; 19: 517–523.
5. Department of Health. Health and Vital Statistics. I.
General Health Statistics, Taipei, Republic of China:
Department of Health, Executive Yuan, 1996.
6. Tai TY, Yang CL, Chang CJ, Chang SM, Chen YH, Lin
BJ, et al. Epidemiology of diabetes mellitus among adults
in Taiwan, R.O.C. J of the Medical Association of Thailand
1987; 70 (suppl 2): 42–48.
7. Lin JD, Shieh WB, Huang MJ, Huang HS. Diabetes
mellitus and hypertension based on the family history
and 2-h postprandial blood sugar in the Ann-Lo district
(Northern Taiwan). Diabetes Res Clin Pract 1993; 20:
75–85.
8. Chou P, Liao MJ, Kuo HS, Hsiao KJ, Tsai ST. A population
survey on the prevalence of diabetes in Kin-Hu, Kinmen.
Diabetes Care 1994; 17: 1055–1058.
9. WHO Study Group on Diabetes Mellitus. Diabetes
Mellitus. WHO Technical Report Series 727. Geneva:
WHO, 1985; 9–12.
10. The Expert Committee on the Diagnosis and Classification
of Diabetes Mellitus. Report of the expert committee on
the diagnosis and classification of diabetes mellitus.
Diabetes Care 1997; 20: 1183–1197.
11. Chou P, Chen HH, Hsiao KJ. Community-based epidemiol-
ogical study on diabetes in Pu-Li, Taiwan. Diabetes Care
1992; 15: 81–89.
571
POPULATION-BASED STUDY OF PREVALENCE OF DIABETES
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
12. WHO Study Group on Diabetes Mellitus. Diabetes
Mellitus. WHO Technical Report Series 727. Geneva:
WHO, 1985: 92.
13. National Diabetes Data Group. Classification and diag-
nosis of diabetes mellitus and other categories of glucose
intolerance. Diabetes 1979; 28: 1039–1057.
14. Seidell JC, Cigolini M, Charzewska J, Ellisinger BM, Biase
GD, Bjorntorp P, et al. Androgenicity in relation to body
fat distribution and metabolism in 38-year-old women.
The European fat distribution study. J Clin Epidemiol
1990; 43: 21–34.
15. Nobbs BT, Smith JM, Wallar AW. Enzymatic determination
of plasma cholesterol on discrete automatic analyzers.
Clin Chem Acta 1977; 79: 391–398.
16. Bucolo G, David H. Quantitative determination of serum
triglycerides by the use of enzymes. Clin Chem 1973;
19: 476–482.
17. The Expert Panel. Report of the National Cholesterol
Education Program: expert panel on detection, evaluation,
and treatment of high blood cholesterol in adults. Arch
Intern Med 1988; 148: 36–69.
18. Frohlich ED, Grim C, Labarthe DR. Recommendations
for human blood pressure determination by sphygmoman-
ometers: report of a task force appointed by the Steering
Committee, American Heart Association. Hypertension
1988; 11: 209A–222A.
19. World Health Organization. Hypertension and Coronary
Heart Disease: Classification and Criteria for Epidemiolog-
ical Study. WHO Technical Report Series 168. Geneva:
WHO, 1959.
20. Kriska AM, Bennett PH. An epidemiological prospective
of the relationship between physical activity and NIDDM:
from activity assessment to intervention. Diabetes Metab
Rev 1992; 4: 355–372.
21. Kriska AM, Knowler WC, LaPorte RE, Drash AL, Wing
RR, Blair SN, et al. Development of questionnaire to
examine relationship of physical activity and diabetes in
Pima Indians. Diabetes Care 1990; 13: 401–411.
22. Paffenbarger RS, Blair SN, Lee IM, Hyde RT. Measurement
of physical activity to assess health effects in free-living
populations. Med Sci Sports Exerc 1993; 25: 60–70.
23. Lilienfield DE, Stolley PD. Foundations of Epidemiology,
3rd edn. Oxford: Oxford University Press, 1994; 68–72.
24. Segi M. Cancer Mortality for Selected Sites in 24
Countries (1950–57). Sendai: Tohuku University School
of Medicine, 1960.
25. McLarty DG, Swai AB, Kitange HM, Masuki G, Mtinangi
BL, Kilima PM, Makene WJ, Chuwa IM, Alberti KG.
Prevalence of diabetes and impaired glucose tolerance
in rural Tanzania. Lancet 1989; i: 871–875.
26. Cockram CS, Woo J, Lau E, Chan JC, Chan AY, Lau J, et
al. The prevalence of diabetes mellitus and impaired
glucose tolerance among Hong Kong Chinese adults of
working age. Diabetes Res Clin Pract 1993; 21: 67–73.
27. Pan SR, Hu YH, Li GW, Liu PA, Bennett PH, Howard
BV. Impaired glucose tolerance and its relationship to
ECG-indicated coronary heart disease and risk factors
among Chinese-Da Qing IGT and diabetes study. Diabetes
Care 1993; 16: 150–156.
28. Dowse GK, Zimmet PZ, Gareeboo H, George K, Alberti
MM, Tuomilehto J, et al. Abdominal obesity and physical
inactivity as risk factors for NIDDM and impaired glucose
tolerance in Indian, Creole, and Chinese Mauritians.
Diabetes Care 1991; 14: 271–282.
29. Quoc PS, Charles MA, Cuong NH, Lieu LH, Tuan
NA, Thomas M, et al. Blood glucose distribution and
prevalence of diabetes in Hanoi. Am J Epidemiol 1994;
139: 713–722.
30. Zimmet P, Taylor R, Ram P, King H, Sloman G, Raper
ORIGINAL ARTICLES
LR, Hunt D. Prevalence of diabetes and impaired glucose
tolerance in the biracial (Melanesian and Indian) popu-
lation of Fiji: a rural-urban comparison. Am J Epidemiol
1983; 118: 673–688.
31. Harris MI, Hadden WC, Knowler WC, Bennett PH.
Prevalence of diabetes and impaired glucose tolerance
and plasma glucose levels in U.S. population aged 20–
74 yr. Diabetes 1987; 36: 523–534.
32. Fujumoto WY, Leonetti DL, Kinyoun JL, Neweel-Morris
L, Shuman WP, Stolov WC, Wahl PW. Prevalence of
diabetes mellitus and impaired glucose tolerance among
second-generation Japanese-American men. Diabetes
1987; 36: 721–729.
33. French LR, Boen JR, Martinez AM, Bushhouse SA, Sprafka
JM, Goetz FC. Population-based study of impaired glucose
tolerance and type II diabetes in Wadena, Minnesota.
Diabetes 1990; 39: 1131–1137.
34. Zimmet P. Type 2 (non-insulin-dependent) diabetes:
an epidemiological overview. Diabetologia 1982; 22:
399–411.
35. Chang CJ, Fu CC, Chen MS, Yang CL, Chen YJ, Chuang
LM, et al. A comparison of newly and previously
diagnosed diabetics in Taiwan. J Formosan Med Assoc
1990; 89: 264–269.
36. Harris MI. Impaired glucose tolerance—prevalence and
conversion to NIDDM. Diabetic Med 1996; 13: S9–S11.
37. Lipton RB, Liao Y, Cooper RS, McGee D. Determinants of
incident non-insulin-dependent diabetes mellitus among
blacks and whites in a national sample: the NHANES I
Epidemiologic Follow-up Study. Am J Epidemiol 1993;
138: 826–839.
38. Wilson PW, Anderson KM, Kannel WB. Epidemiology of
diabetes mellitus in the elderly: The Framinghan Study.
Am J Med 1986; 80 (supply 5A): 3–9.
39. Butler WJ, Ostrander LD, Carman WJ, Lamphiear DE.
Diabetes mellitus in Tecumseh, Michigan: prevalence,
incidence, and associated conditions. Am J Epidemiol
1982; 116: 971–980.
40. Knowler WC, Bennett PH, Hamman RF, Miller M.
Diabetes incidence and prevalence in Pima Indians: a
19-fold greater incidence than in Rochester, Minnesota.
Am J Epidemiol 1978; 108: 497–505.
41. Zimmet P, Taft P, Guinea A, Guthrie W, Thoma K. The
high prevalence of diabetes mellitus on a central Pacific
Island. Diabetologia 1977; 13: 113–115.
42. Helmrich SP, Ragland DR, Leung RW, Paffenbarger RS.
Physical activity and reduced occurrence of non-insulin-
dependent diabetes mellitus. N Engl J Med 1991; 325:
147–152.
572
FENG-HWA LU ET AL.
1998 John Wiley & Sons, Ltd. Diabet. Med. 15: 564–572 (1998)
43. Manson JE, Rimm EB, Stampfer MJ, Colditz GA, Willett
WC, Krolewski AS, et al. Physical activity and incidence
of non-insulin-dependent diabetes mellitus in women.
Lancet 1991; 338: 774–778.
44. Manson JE, Nathan DM, Krolewski AS, Stampfer MJ,
Willett WC, Hennekens CH. A prospective study of
exercise and incidence of diabetes in US male physicians.
J Am Med Assoc 1992; 268: 63–67.
45. Devlin JT. Effects of exercise on insulin sensitivity in
humans. Diabetes Care 1992; 15: 1690–1693.
46. Keen H, Thomas BJ, Jarrett RJ, Fuller JH. Nutrient intake,
adiposity and diabetes. Br Med J 1979; 1: 655–658.
47. Medalie JH, Papier CM, Glodbourt MA, Herman JB.
Major factors in the development of diabetes mellitus in
10,000 men. Arch Intern Med 1975; 135: 811–817.
48. Harlan LC, Harlan WR, Land, Goldstein NG. Factors
associated with glucose tolerance in adults in United
States. Am J Epidemiol 1987; 126: 80–84.
49. Feskens EJM, Kromhout D. Cardiovascular risk factors
and the 25-year incidence of diabetes mellitus in middle-
aged men: The Zuphen Study. Am J Epidemiol 1989;
130: 1101–1108.
50. Morales PA, Mitchell BD, Valdez RA, Hazuda HP, Stern
MP, Haffner SM. Incidence of NIDDM and impaired
glucose tolerance in hypertensive subjects: The San
Antonio Heart Study. Diabetes 1993; 42: 154–161.
51. Wilson PW, McGee DL, Kannel WB. Obesity, very low
density lipoproteins, and glucose intolerance over fourteen
years. Am J Epidemiol 1981; 114: 697–704.
52. Barrett-Connor E, Grundy SM, Holdbrook MJ. Plasma
lipids and diabetes mellitus in an adult community. Am
J Epidemiol 1982; 115: 657–663.
53. Cruz-Vidal M, Costas R Jr, Garcia-Palmieri MR, Sorlie
PH, Hertzmark E. Factors related to diabetes mellitus in
Puerto Rican men. Diabetes 1979; 28: 300–307.
54. Reaven GM. Banting lecture 1988: Role of insulin
resistance in human disease. Diabetes 1988; 37: 1595–
1607.
55. DeFronzo RA, Ferrannini E. Insulin resistance: a multifa-
ceted syndrome responsible for NIDDM, obesity, hyper-
tension, dyslipidemia, and atherosclerotic cardiovascular
disease. Diabetes Care 1991; 14: 173–194.
56. Reaven GM, Lithell H, Landsberg L. Hypertension and
associated metabolic abnormalities: the role of insulin
resistance and the sympathoadrenal system. N Engl J Med
1996; 334: 374–381.
57. Flack JM, Sowers JR. Epidemiologic and clinical aspects
of insulin resistance and hyperinsulinemia. Am J Med
1991; 91(suppl 1A): 11S–21S.
