Content uploaded by Jianzhong Xiao
Author content
All content in this area was uploaded by Jianzhong Xiao on Feb 06, 2015
Content may be subject to copyright.
E
p i d e m i o I o g y / H e a
11
h
Se
r v i
c
e
s
/ P
s
y
c
h o
s oc
i a I
Research
ORIGINAL ARTICLE
Effects of Diet and Exercise in Preventing
NIDDM in People With Impaired Glucose
Tolerance
The Da Qing IGT and Diabetes Study
XIAO-REN
PAN, MD
GUANG-WEI
Li, MD
YlNG-HUA HU, MD
JI-XING WANG,
MD
WEN-YING YANG,
MD
ZUO-XIN
AN, MD
ZE-XI
HU, MD
JUAN-LIN,
MD
JIAN-ZHONG XIAO,
MD
HUI-BI
CAO, MD
PING-AN
LIU, MD
XI-GUI JIANG,
MD
YA-YAN JIANG,
MD
JIN-PING WANG,
MD
HUI ZHENG,
MD
HUI ZHANG,
MD
PETER
H.
BENNETT,
MB,
FRCP
BARBARA
V.
HOWARD,
PHD
OBJECTIVE— Individuals with impaired glucose tolerance (IGT) have a high risk of
developing NIDDM. The purpose of this study was to determine whether diet and exercise
interventions in those with IGT may delay the development of NIDDM, i.e., reduce the inci-
dence of
NIDDM,
and thereby reduce the overall incidence of diabetic complications, such as
cardiovascular, renal, and retinal disease, and the excess mortality attributable to these com-
plications.
RESEARCH DESIGN AND METHODS— In 1986,110,660 men and women from 33
health care clinics in the city of
Da
Qing, China, were screened for IGT and NIDDM. Of these
individuals, 577 were classified (using World Health Organization criteria) as having
IGT.
Sub-
jects were randomized by clinic into a clinical trial, either to a control group or to one of three
active treatment groups: diet only, exercise only, or diet plus exercise. Follow-up evaluation
examinations were conducted at 2-year intervals over a 6-year period to identify subjects who
developed NIDDM. Cox's proportional hazard analysis was used to determine if the incidence
of NIDDM varied by treatment assignment.
RESULTS — The cumulative incidence of diabetes at 6 years was 67.7% (95% CI, 59.8-75.2)
in the control group compared with
43.8%
(95% CI, 35.5-52.3) in the diet
group,
41.1%
(95%
CI,
33.4-49.4) in the exercise group, and
46.0%
(95% CI, 37.3-54.7) in the diet-plus-exercise
group (P < 0.05). When analyzed by clinic, each of the active intervention groups differed
significantly from the control clinics (P < 0.05). The relative decrease in rate of development
of diabetes in the active treatment groups was similar when subjects were stratified as lean or
overweight (BMI < or >25 kg/m
2
). In a proportional hazards analysis adjusted for differences
in baseline
BMI
and fasting glucose, the diet, exercise, and diet-plus-exercise interventions were
associated with
31%
(P < 0.03), 46% (P < 0.0005), and 42% (P < 0.005) reductions in risk
of developing diabetes, respectively.
CONCLUSIONS — Diet and/or exercise interventions led to a significant decrease in the
incidence of diabetes over a 6-year period among those with IGT.
From
the
Departments
of
Endocrinology
(X.R, G.L, Z.H., J.L,
J.-Z.X.,
H.C.,
H.Zhe.,
H.Zha.)
and
Nutrition
(Z.A.),
China-Japan
Friendship
Hospital,
Beijing;
the
Department
of
Cardiovascular
Disease
(Y.-H.H.,Ji-x.W,
WY.,
P-A.L,
X.J.,
Y.J.,
Jin.W), Da
Qing
First
Hospital,
Da
Qing,
China;
Phoenix
Epidemiology
and
Clinical
Research Branch
(EH.B.),
National
Institutes
of
Health,
National
Institute
of
Diabetes
and
Digestive
and Kid-
ney
Diseases,
Phoenix,
Arizona;
and
Medlantic
Research Institute
(B.VH.),
Washington,
DC.
Address correspondence
and
reprint
requests
to
Barbara
V
Howard,
PhD,
Medlantic
Research
Institute,
108
Irving
St., NW,
Washington,
DC 20010-2933.
E-mail:
bvhl@mhg.edu.
Received
for
publication
19
April
1996 and
accepted
in
revised
form
14
November
1996.
ANOVA,
analysis
of
variance;
IGT,
impaired
glucose
tolerance;
WHO,
World
Health
Organization.
D
iabetes and its complications are
major and increasing health prob-
lems in many parts of
the
world. The
most frequent form, NIDDM, leads to vas-
cular complications that give rise to con-
siderable morbidity and premature
mortality. Impaired glucose tolerance
(IGT),
a lesser degree of hyperglycemia,
represents an intermediate stage in the
development of NIDDM that is associated
with a high risk of developing NIDDM
(1-3).
One- to three-quarters of
those
with
IGT develop diabetes within a decade of
discovery of IGT (4), and annual progres-
sion rates from IGT to diabetes range from
1 to 10% (5-11). Thus, if progression
could be slowed, the incidence of diabetes
would be reduced and the onset of its com-
plications prevented or delayed. Risk fac-
tors known to influence the rate of
progression from IGT to diabetes include
age,
obesity, hyperinsulinemia, and insulin
resistance (4,12).
The effect of interventions on the pro-
gression from IGT to diabetes has been
examined in a few studies. In two small
English studies (5,7), no measurable effect
of either diet or oral antidiabetic agents was
found on the incidence of subsequent dia-
betes,
whereas in the Malmohus Study in
Sweden
(6),
subjects with
IGT
who received
oral tolbutamide over a 10-year period had
a lower incidence of diabetes. In another
Swedish study, the Malmo Study, in which
treatment was not randomized, adherence
to a diet/exercise program for 5 years
reduced the incidence of diabetes (14).
In 1986,577 people with
IGT,
identified
during a population-based survey of dia-
betes and IGT in Da Qing, China, agreed to
participate in a randomized controlled trial
to evaluate the effects of diet and/or exercise
interventions on the incidence of diabetes
(15).
This report presents the results of this
trial over a 6-year follow-up period.
RESEARCH DESIGN AND
METHODS — The trial
was
designed as
a controlled clinical trial in which subjects
were randomized by
clinic
to investigate the
DIABETES CARE, VOLUME 20, NUMBER 4, APRIL 1997
537
Preventing NIDDM in people with IGT
Table
1—Activities
required for one unit of exercise
Intensity Time (min)
Exercise
Mild
Moderate
Strenuous
Very strenuous
30
20
10
5
Slow walking, traveling by bus, shopping,
housecleaning
Faster walking or walking down stairs, cycling, doing
heavy laundry, ballroom dancing (slow)
Slow running, climbing
stairs,
disco dancing for the elderly,
playing volleyball or table tennis
Jumping rope, Playing basketball, swimming
effects of dietary and exercise intervention
separately and in combination, on the inci-
dence of diabetes in people with IGT.
Eligibility and exclusion criteria
Da Qing
is
an industrial
city,
primarily con-
cerned with oil exploration and produc-
tion, in the Hei Long Jiang province in the
northern part of
China.
In 1986, the popu-
lation of Da Qing included 281,589 people
over the age of 25, all of whom received
health care in designated clinics located
throughout the city. Half of these clinics,
which served 126,715 people over the age
of 25, were selected to participate in a
screening
study.
Between June and Decem-
ber 1986, most (87.3%) of
the
target popu-
lation (110,660 total: 55,391 men and
55,269 women) underwent screening at
nearby
hospitals.
The screening consisted
of
measurement of plasma glucose concentra-
tion
2
h
(± 5
min) after
a
standard breakfast,
followed by a 75-g oral glucose tolerance
test in those who screened positive (15).
Details of the study population and valida-
tion of the screening procedures have been
described previously (15). From the initial
screening
study,
577 people who met World
Health Organization (WHO) criteria for
IGT
agreed to participate in the intervention
study described below. Of these, 530 sub-
jects were followed systematically until end-
points had been reached or for a
6-year
period. Most of the 47 lost to follow-up
were lost because of migration from the
region (see below). Enrollment and treat-
ment of subjects were conducted in accor-
dance with the Helsinki Declaration.
Randomization and baseline
measures
Intervention was provided by 33 local
health clinics associated with the oil factory
communities that are dispersed throughout
the city. The number of subjects attending
each of these clinics ranged from 5 to 33.
Each clinic, rather than each subject, was
randomized to carry out the intervention
on each of the eligible subjects attending
that clinic according to one of the four
specified intervention
protocols.
Study par-
ticipants in each clinic were categorized
according to BMI, with 208 individuals
categorized as lean (BMI <25 kg/m
2
) and
322 as overweight (BMI ^25 kg/m
2
).
A
baseline examination
was
conducted
on each participant after a 10- to 12-h
overnight fast as described previously (15).
Briefly, blood pressure, height, and weight
were measured in light clothing without
shoes following methods used in the WHO
multinational study of vascular disease in
diabetes (17). After a fasting blood sample
was taken, each subject ingested 75 g of
glucose monohydrate dissolved in 300 ml
water within a
2-min
period. Plasma glu-
cose and lipids were measured in the fast-
ing sample, and glucose was measured in
the samples obtained at 60 and 120 min
after the glucose load.
A
urine sample was
collected over the 2-h time period of the
glucose tolerance test to quantify urinary
glucose and albumin excretion. Past med-
ical history and family history of diabetes
were assessed by questionnaire. The oral
glucose tolerance test was repeated in each
subject during systematic evaluation exam-
inations conducted at ~2-year intervals.
Food intake and physical activity were
quantified
at baseline
and at each evaluation
examination using standardized forms and
interviews. For dietary intake, quantity per
day for the past 3 days was ascertained for
major food/beverage items, such as pork,
beef,
shrimp, fowl, eggs, milk, bean curd,
bean and pork oils, peanuts, sunflower
seeds,
fruits, vegetables, wine, and beer.
These were converted to major food con-
stituents using a food nutrition database
(Database of Nutrition for the Peoples
Republic of China version
1.0,1993).
Phys-
ical activity was assessed in a standardized
way For occupational activity, the kind of
activity and its frequency, as well as the
mode and duration of transportation to and
from work were assessed. Leisure physical
activity was ascertained in minutes per day
for major activities, such as walking, run-
ning, cycling, ball playing, aerobics, danc-
ing, gardening, and
swimming.
Activity
was
ascertained for the previous week and con-
verted to units per day
as
shown in
Table
1.
Interventions
Diet group. In clinics assigned to the diet-
only intervention, participants with BMI
<25 kg/m
2
were prescribed a diet contain-
ing 25-30 kcal/kg body wt (105-126
kj/kg),
55-65% carbohydrate, 10-15%
protein, and 25-30% fat. These partici-
pants were encouraged to consume more
vegetables, control their intake of alcohol,
and reduce their intake of simple sugars.
Subjects with BMI >25 kg/m
2
were
encouraged to reduce their calorie intake so
as to gradually lose weight at a rate of
0.5-1.0
kg per month until they achieved
a BMI of 23 kg/m
2
. Individual goals were
set for total calorie consumption and for
daily quantities of
cereals,
vegetables, meat,
milk, and oils. This was accomplished by
providing a list to each individual of the
recommended daily intake of commonly
used foods and a substitution list to allow
exchange within food groups. Patients
received individual counseling by physi-
cians concerning daily food
intake.
In addi-
tion, counseling sessions (in small groups)
were conducted weekly for 1 month,
monthly for 3 months, and then once every
3 months for the remainder of the study.
Exercise group. Participants in clinics
assigned to the exercise group were taught
and encouraged to increase the amount of
their leisure physical exercise by at least 1
U/day (as defined in Table 1) and by 2
U/day if possible for those <50 years of
age
with no evidence of cardiovascular disease
or
arthritis.
As
in the diet group, counseling
sessions were conducted weekly for 1
month, monthly for 3 months, and then
once every 3 months for the remainder of
the study. The rate of increase and type of
exercise recommended depended on age,
past exercise patterns, and the existence of
health problems other than
IGT.
Appropri-
ate indoor activities were suggested for
winter. Exercise units were computed as
described in Table 1.
Diet-plus-exercise group. Participants
from clinics assigned to this group received
instructions and counseling for both diet
538
DIABETES CARE, VOLUME 20, NUMBER 4, APRIL 1997
Pan and Associates
and exercise interventions that were similar
to those for the diet-only and the exercise-
only intervention groups.
Control group. Subjects from clinics
assigned to the control group were exposed
to general information about diabetes and
IGT.
Clinic physicians also dispensed infor-
mational brochures with general instruc-
tions for diet and/or increased leisure
physical activities to control group sub-
jects,
but no individual instruction or for-
mal group counseling sessions were
conducted.
Training. All local physicians, nurses, and
technicians involved in the study attended
a 2-day training session each year in which
they received standardized instruction on
the diet and exercise interventions and pro-
cedures for the examination. The Da Qing
Study Steering Committee provided edu-
cational materials on diabetes and IGT via
videotapes and brochures. Members of the
Steering Committee also talked to the
groups to supplement the education classes
on diet and/or exercise for the appropriate
groups in 1986 and again in 1988.
Follow-up procedures
Systematic evaluation examinations were
carried out in 1988, 1990, and 1992. In
these examinations, variables, such as
blood pressure, weight, skinfold measure-
ments, and diet and physical activity (as
used at baseline), were remeasured as
described
below.
All participants were seen
at 3-month intervals by local physicians.
The general health of each participant was
assessed by the physician, and compliance
with the intervention regimen was dis-
cussed with the nurses and clinic
staff.
Physicians repeated their counseling and
instructions concerning diet and exercise.
At each 3-month follow-up visit, weight
and blood pressure were measured and
urine glucose was assessed using a dip-
stick. Plasma glucose was measured 2 h
after a standard breakfast (100 g steamed
bread) if
the
urinary
glucose
was
positive.
If
the postmeal plasma glucose concentration
was >200 mg/dl (11.1 mmol/1), or if the
local physician suspected that the subject
had developed diabetes, the subject
received
a
75-g oral glucose tolerance test at
the city hospital
or,
occasionally, at
a
district
hospital. If, at any time during the course of
the study, a participant exhibited symp-
toms of diabetes and repeated fasting
plasma glucose measurements were >:140
mg/dl (7.8 mmol/1) or a casual glucose
measurement was ^200 mg/dl (11.1
mmol/1), a clinical diagnosis of diabetes
was made. A standard oral glucose toler-
ance test was performed on these individ-
uals.
If the subject met WHO criteria for
diabetes on the basis of these tests, his or
her formal participation in the study ended.
All decisions concerning whether or not
participants had reached endpoints based
on the 3-month follow-up examinations
were made by the vice chairman of the
Study Steering Committee.
Outcome assessment
At 2-year intervals (1988,1990, and 1992),
a systematic evaluation examination of each
participant, including those diagnosed at
the 3-month follow-up examinations, was
performed using methods similar to those
of the baseline examination. Physicians
from the China-Japan Friendship Hospital
in Beijing recorded diet and exercise
changes and provided individual advice on
intervention adherence. Height, weight,
and blood pressure were measured and
fasting 2-h plasma glucose
was
determined
after a 75-g oral glucose load. If fasting
plasma glucose was ^140 mg/dl (7.8
mmol/1) and 2-h glucose was ^200 mg/dl
(7.8-11.0
mmol/1),
then the assigned treat-
ment regimens were continued. If the
results were indicative of diabetes (fasting
glucose >140 mg/dl [7.8 mmol/1] or 2-h
glucose >200 mg/dl [11.1 mmol/1]), then
the oral glucose tolerance test was repeated
after 7-14 days. If the repeat results were
normal or in the range of IGT, then the
assigned treatment regimen was resumed.
If the diagnosis of diabetes was confirmed,
the subjects were considered to have
reached an endpoint and were referred to
receive standard diabetes treatment.
Of the 263 diabetes diagnoses made
during the 6 years, 55 (21%) were made
initially by the local physicians and con-
firmed at the city hospital by glucose toler-
ance test; 208 (79%) were made as a result
of the systematic oral glucose tolerance tests
performed in 1988,1990, and 1992. Those
who left in
1988
very early in the study and
before the first follow-up for reasons unre-
lated to their randomization group were
not included in the analysis. The 11 who
died were retained, although none had
developed diabetes before death.
Statistical analysis
The cumulative number of subjects who
had developed diabetes in each treatment
group was determined after conducting the
6-year
evaluation examination. Because the
randomization was performed at the clinic,
rather than at the individual, subject level,
data were analyzed in each treatment group
by comparing the incidence of diabetes in
the clinics assigned to each of the treat-
ments. The Ryan-Einot-Gabriel-Welsch
multiple F test was used to compare the
clinic groups. We also analyzed the data as
if individuals had been assigned to specific
treatment groups, including clinic as a
covariate in the
analyses.
Multivariate analy-
sis was performed using
Cox's
proportional
hazards analysis taking into account the
time to diagnosis. The proportional hazard
model was used because a number of indi-
viduals (21%) were diagnosed at interme-
diate points and because the characteristics
of the outcome evaluation conform more to
the assumptions of the Cox's model than to
multiple logistic regression, which might
have been more appropriate if there were
only
a
6-year
fixed
follow-up.
A
backwards
stepwise procedure was used to identify
possible
covariates.
The level of significance
was taken
as
P < 0.05. Interaction terms for
glucose * obesity and diet * exercise were
included in the
model.
A
subgroup analysis
was performed stratifying individuals by
BMI. For analysis of data on exercise and
diet, groups were compared using analysis
of variance (ANOVA).
RESULTS
Incidence of diabetes
Baseline and
6-year
follow-up characteristics
for the four study groups are summarized in
Table 2. Of the 577 subjects with IGT who
were randomized, 530 completed the study.
Of the remainder, 7 people refused follow-
up,
29 left
Da
Qing in 1988 (mostly because
of the establishment of
a
new oil
field
else-
where),
and
11
died during
the
course of the
study. No deaths occurred in the exercise-
only group. Three deaths occurred in the
control group (one pneumonia, two cirrho-
sis),
three in the diet group (two cancer, one
septicemia), and five in the diet-plus-exer-
cise
group (one stroke, two cancer, one acci-
dental, one Crohn's disease). None of these
11 people were known to have developed
diabetes before death. There
were
no signifi-
cant differences in
baseline values
among the
four groups.
The mean for
6-year
diabetes incidence
in each of
the
clinics was calculated accord-
ing to the treatment group assigned to that
clinic (Table 3, Fig. 1). When the means of
diabetes incidence in each clinic by treat-
ment group were compared, there was a
DIABETES CARE, VOLUME 20, NUMBER 4, APRIL 1997
539
Prexenting NIDDM in people with IGT
Table 2—Characteristics of participants at the baseline and
6-year
evaluation examinations by intervention group
n
Baseline characteristics
Age (years)
Sex (M/F)
BMI (kg/m
2
)
Fasting plasma glucose (mmol/1)
2-h fasting glucose (mmol/1)
Results at or before
6-year
evaluation
Diabetes by WHO criteria
No.
with 2-h plasma glucose
> 11.1 mmol/1 (%)
Incidence/100 person-years (95% Cl)
Fasting hyperglycemia
No.
with fasting plasma glucose
>7.8 mmol/1 (%)
Incidence/100 person-years (95% CI)
Weight change (kg)
No diabetes
Diabetes
Glucose (mmol/1)
Fasting plasma glucose
2-h glucose
Control
133
46.5
±
9.3
73/60
26.2
±
3.9
5.52
±
0.82
9.03
±
0.89
90 (67.7)
15.7 (12.7-18.7)
55 (41.4)
9.6 (7.2-12.0)
0.27
-1.55
7.59
±
2.59
12.99 ±4.19
Diet
130
44.7
±
9.4
59/71
25.3 ±3.8
5.56 ±0.81
9.03
±
0.94
57 (43.8)
10.0 (7.5-12.5)
21 (16.2)
3.7(2.1-5.3)
0.93
-2.43
6.94
±
4.49
10.51 ±4.89
Exercise
141
44.2
±
8.7
81/60
25.4 ±3.7
5.56 ±0.83
8.83
±
0.79
58(41.1)
8.3 (6.4-10.3)
37 (26.2)
5.3 (3.6-7.0)
0.71
-1.93
6.83
±
2.24
10.51 ±3.93
Diet + exercise
126
44.4
±
9.2
70/56
26.3 ±3.9
5.67
±
0.80
9.11 ±0.93
58 (46.0)
9.6 (7.2-12.0)
33 (26.2)
5.5 (3.7-7.3)
-1.77
-3.33
7.15 ±2.72
10.76 ±4.37
Total
530
45.0 ±9.1
283/247
25.8 ±3.8
5.59 ±0.81
9.0
±
0.89
263 (49.6)
10.8 (9.6-12.0)
14.6 (27.5)
6.0 (5.06-6.94)
-0.31
-0.87
7.13*3.11
11.05 ±4.41
Data are means ± SD.
highly significant difference between the
groups (P < 0.0035), and each group of
clinics providing active treatments differed
significantly from the clinics in the control
group (P < 0.05). There
were,
however, no
statistically significant differences in the inci-
dences of diabetes between the groups of
clinics providing
active
treatments.
Among individual subjects in the con-
trol group, the incidence of diabetes
(defined using WHO criteria) was 15.7/100
person-years (95% CI, 12.7-18.7%). In
each of the three intervention groups, the
incidence of diabetes was significantly
lower than in the control group (10.0 [95%
CI,
7.5-12.5], 8.3 [6.4-10.3], and 9.6
[7.2-12.0] per
100
person-years in
the
diet,
exercise, and diet-plus-exercise groups,
respectively) (P < 0.05 for all) (Table 2).
Incidence rates did not differ significantly
among the three intervention groups (P >
0.05). If an alternative endpoint is defined
as fasting glucose ^140 mg/dl (7.8
mmol/1), incidence rates were 9.6 (95% CI
7.2-12.0)
in the control group and 3.7
(2.1-5.3),
5.3 (3.6-7.0), and 5.5 (3.7-7.3)
per 100 person-years in the diet, exercise,
and diet-plus-exercise groups, respectively
(P < 0.05 for each).
Comparison of lean and overweight
subgroups
Because the dietary advice differed according
Table
3—6-year
cumulative incidence of diabetes by clinic and treatment assignment
Control clinics
No.
of subjects
% developing diabetes
Diet clinics
No.
of subjects
% developing diabetes
Exercise clinics
No.
of subjects
% developing diabetes
Diet-plus-exercise clinics
No.
of subjects
% developing diabetes
11
73
12
58
18
56
16
38
15
60
5
60
16
38
13
31
24
58
17
67
16
44
28
57
16
64
6
67
20
30
15
53
15
53
27
48
20
55
16
50
6
67
8
38
7
29
10
40
33
85
28
32
28
21
28
43
13
62
15
40
8
75
—
—
—
—
12
50
8
50
—
—
Clinic
mean (%)
65.9
47.1
44.2
44.6
SD(%)
10.0
10.4
16.8
9.2
95%
CI (%)
57.5-76.3
38.7-55.5
31.3-57.1
36.1-53.1
The overall test for a difference among group means is statistically significant (P
=
0.0035,
ANOVA).
Pair-wise test indicates statistically significant differences between
the control group of clinics and the diet, exercise, and diet-plus-exercise groups (P < 0.05, Ryan-Einot-Gabriel-Welsch multiple F test). There were no statistical
differences among the diet, exercise, and diet-plus-exercise groups.
540
DIABETES CARE, VOLUME 20, NUMBER 4, APRIL 1997
Pan and Associates
90
70
50
30
10
Control Diet Exercise Diet
+ Exercise
Figure
1—Mean rate
of diabetes for
each
clinic
at
6-year
follow-up,
by
intervention
group.
Means
(±
SD)
were control,
66 ±
10; diet,
47 ±11;
exercise,
45 ±
9;
and
diet plus exercise,
44 ±17.
to BMI, leading to the possibility of different
effects
of
the interventions
in
lean
and
over-
weight individuals, the incidence of diabetes
was evaluated separately
in
those
who had
BMI
at
baseline
<25 or S:25
kg/m
2
(Tables
4
and 5,
Fig.
2).
Incidence rates
of
diabetes
in
the
control group
of
overweight partici-
pants were higher than those
in the
control
group
of
lean subjects
(17.2 vs.
13.3/100
person-years
[P <
0.05]).
On the
other
hand,
in
both
BMI
categories, compared
with
the
corresponding control groups,
there were significantly lower incidence rates
of diabetes
in the
intervention groups
(except the lean diet
group):
8.3
(diet) (NS),
5.1 (exercise)
(P <
0.01),
and 6.8
(diet plus
exercise) (P
<
0.05)
for
the lean groups;
and
11.5,10.8,
and
11.4, respectively (P
< 0.05
for
all),
for
the overweight groups. The rela-
tive decrease
in
rate
of
development
of
dia-
betes
in the
active treatment groups
(compared with the control group) was sim-
ilar
in
the overweight
and
lean strata. Those
who developed diabetes lost weight,
and
the change
was
significant
in the
lean
and
obese groups. When weight change
was
added
to the
model,
the
conclusions were
not changed (data
not
shown).
Influence of type of intervention and
baseline characteristics on the
development of diabetes
When the three intervention strategies were
compared with
the
control group
in a pro-
portional hazards model, there was
an
over-
all reduction
in the
incidence
of
diabetes
of
33%
in the
diet-only group
(P <
0.03),
47%
in the
exercise-only group
(P <
0.0005),
and
38%
in the
diet-plus-exercise
group
(P <
0.005) (Table
6).
Inclusion
of
clinic as
a
variable did not change the results
and
had a
nonsignificant effect
(P =
0.65).
Because plasma glucose level
and
BMI both
influenced the incidence of NIDDM
in
sub-
jects with
IGT, the
effects
of
intervention
after adjustment
for
these baseline factors
was examined. Only modest changes
in the
effects
of
intervention were seen
in the
risk
reductions
of 31, 46, and 42% for
diet,
exercise,
and
diet plus exercise, respectively
Baseline physical activity was
not
predictive
of
the
development
of
NIDDM
in any
model. The interaction terms, BMI * Fasting
plasma glucose
and
Diet
*
Exercise, were
not significant (data
not
shown).
Changes in diet and exercise
Baseline caloric intake and diet composition
were similar in all four intervention groups.
After 6 years of follow-up, estimated caloric
intake appeared lower in the diet and diet-
plus-exercise groups, but these differences
did not reach statistical significance. Analy-
sis of calorie composition showed a slightly
lower proportion of carbohydrates and pro-
teins and a slightly higher proportion of fat
at follow-up, but the differences were not
statistically significant. Physical exercise,
expressed in units per
day,
was significantly
higher at baseline in the diet-plus exercise
group than in the control group. At the 6-
year follow-up, average units per day of
exercise were significantly higher than at
baseline in the exercise and in the diet-
plus-exercise groups (Table 7).
CONCLUSIONS— This study has
demonstrated in a large group of men and
women with IGT, identified by screening,
that institution of a lifestyle intervention
over a
6-year
period led to a significant
decrease in the incidence of diabetes.
Groups randomized by clinic to receive
diet, exercise,