Article

Finnish Diabetes Prevention Study Group: Sustained reduction in the incidence of type 2 diabetes by lifestyle intervention: follow-up of the Finnish Diabetes Prevention Study

Department of Dental Public Health, University of Helsinki, Helsinki, Uusimaa, Finland
The Lancet (Impact Factor: 45.22). 12/2006; 368(9548):1673-9. DOI: 10.1016/S0140-6736(06)69701-8
Source: PubMed

ABSTRACT

Lifestyle interventions can prevent the deterioration of impaired glucose tolerance to manifest type 2 diabetes, at least as long as the intervention continues. In the extended follow-up of the Finnish Diabetes Prevention Study, we assessed the extent to which the originally-achieved lifestyle changes and risk reduction remain after discontinuation of active counselling.
Overweight, middle-aged men (n=172) and women (n=350) with impaired glucose tolerance were randomly assigned to intensive lifestyle intervention or control group. After a median of 4 years of active intervention period, participants who were still free of diabetes were further followed up for a median of 3 years, with median total follow-up of 7 years. Diabetes incidence, bodyweight, physical activity, and dietary intakes of fat, saturated fat, and fibre were measured.
During the total follow-up, the incidence of type 2 diabetes was 4.3 and 7.4 per 100 person-years in the intervention and control group, respectively (log-rank test p=0.0001), indicating 43% reduction in relative risk. The risk reduction was related to the success in achieving the intervention goals of weight loss, reduced intake of total and saturated fat and increased intake of dietary fibre, and increased physical activity. Beneficial lifestyle changes achieved by participants in the intervention group were maintained after the discontinuation of the intervention, and the corresponding incidence rates during the post-intervention follow-up were 4.6 and 7.2 (p=0.0401), indicating 36% reduction in relative risk.
Lifestyle intervention in people at high risk for type 2 diabetes resulted in sustained lifestyle changes and a reduction in diabetes incidence, which remained after the individual lifestyle counselling was stopped.

Full-text

Available from: Sirkka Aunola
Articles
www.thelancet.com Vol 368 November 11, 2006
1673
Sustained reduction in the incidence of type 2 diabetes by
lifestyle intervention: follow-up of the Finnish Diabetes
Prevention Study
Jaana Lindström, Pirjo Ilanne-Parikka, Markku Peltonen, Sirkka Aunola, Johan G Eriksson, Katri Hemiö, Helena Hämäläinen, Pirjo Härkönen,
Sirkka Keinänen-Kiukaanniemi, Mauri Laakso, Anne Louheranta, Marjo Mannelin, Merja Paturi, Jouko Sundvall, Timo T Valle, Matti Uusitupa,
Jaakko Tuomilehto, on behalf of the Finnish Diabetes Prevention Study Group
Summary
Background Lifestyle interventions can prevent the deterioration of impaired glucose tolerance to manifest type 2
diabetes, at least as long as the intervention continues. In the extended follow-up of the Finnish Diabetes Prevention
Study, we assessed the extent to which the originally-achieved lifestyle changes and risk reduction remain after
discontinuation of active counselling.
Methods Overweight, middle-aged men (n=172) and women (n=350) with impaired glucose tolerance were randomly
assigned to intensive lifestyle intervention or control group. After a median of 4 years of active intervention period,
participants who were still free of diabetes were further followed up for a median of 3 years, with median total
follow-up of 7 years. Diabetes incidence, bodyweight, physical activity, and dietary intakes of fat, saturated fat, and
bre were measured.
Findings During the total follow-up, the incidence of type 2 diabetes was 4·3 and 7·4 per 100 person-years in the
intervention and control group, respectively (log-rank test p=0·0001), indicating 43% reduction in relative risk. The
risk reduction was related to the success in achieving the intervention goals of weight loss, reduced intake of total and
saturated fat and increased intake of dietary fi bre, and increased physical activity. Benefi cial lifestyle changes achieved
by participants in the intervention group were maintained after the discontinuation of the intervention, and the
corresponding incidence rates during the post-intervention follow-up were 4·6 and 7·2 (p=0·0401), indicating 36%
reduction in relative risk.
Interpretation Lifestyle intervention in people at high risk for type 2 diabetes resulted in sustained lifestyle changes
and a reduction in diabetes incidence, which remained after the individual lifestyle counselling was stopped.
Introduction
The pandemic of type 2 diabetes is an enormous public
health problem.
1,2
Studies using lifestyle intervention in
people with impaired glucose tolerance have shown
that the progress to manifest type 2 diabetes can be
prevented or postponed.
3–8
Lifestyle intervention in
these studies lasting for 3–6 years emphasised
bodyweight control, physical activity, and dietary
modifi cation. Reduction in relative risk achieved in the
intervention group compared with the control group
ranged from 30% to 67%, as shown in a recent
meta-analysis.
9
The Finnish Diabetes Prevention Study
5
and the US Diabetes Prevention Program
6
both revealed
a 58% relative risk reduction in the progression from
impaired glucose tolerance to type 2 diabetes, during a
mean intervention period of about 3 years.
However, whether the risk reduction achieved during
active counselling for lifestyle changes will last
after discontinuation of the intervention is not known.
The extended follow-up of the Diabetes Prevention
Study was designed to assess the long-term results of
the lifestyle intervention originally aimed at reducing
the risk for developing type 2 diabetes in high-risk
individuals.
Methods
The Diabetes Prevention Study was a randomised
controlled trial aimed at prevention of type 2 diabetes by
lifestyle intervention. The study design has been
described in detail previously.
10
The study protocol was
approved by the ethics committee of the National Public
Health Institute in Helsinki, Finland, and all study
participants gave written informed consent. Random-
isation started in 1993 and was completed in 1998
(fi gure 1). The fi rst interim analysis was done in March,
2000.
5
According to the recommendation of the endpoint
committee, the intervention period was discontinued at
each participant’s next yearly clinic visit, after a median
follow-up of 4 years. Subsequently, we decided to continue
to monitor the participants who had remained free of
diabetes. This report consists of the data obtained until
Dec 31, 2004, ie, post-intervention follow-up for a median
of 3 years, with median total follow-up of 7 years.
Participants
Originally, 522 men and women in fi ve study centres
were randomised at the baseline visit to one of the two
treatment modalities, the intervention group with
intensive diet-exercise counselling (n=265, the proportion
Lancet 2006; 368: 1673–79
See Comment page 1634
Diabetes Unit, Department of
Health Promotion and Chronic
Disease Prevention, National
Public Health Institute,
Mannerheimintie 166,
00300 Helsinki, Finland
(J Lindström MSc,
M Peltonen PhD,
Prof J G Eriksson MD,
K Hemiö MSc, T T Valle MD,
Prof J Tuomilehto MD);
The Diabetes Centre, Finnish
Diabetes Association, Tampere,
Finland (P Ilanne-Parikka MD);
Research Unit of Tampere
University Hospital, Tampere,
Finland (P Ilanne-Parikka);
Laboratory for Population
Research, National Public Health
Institute, Turku, Finland
(S Aunola PhD); Research
Department, Social Insurance
Institution, Turku, Finland
(H Hämäläinen PhD);
Department of Sports Medicine,
Oulu Deaconess Institute, Oulu,
Finland (P Härkönen MSc,
M Mannelin MSc); Department
of Public Health Science and
General Practice, University of
Oulu, Oulu, Finland
(Prof S Keinänen-Kiukaanniemi
MD, M Laakso MD); Unit of
General Practice, Oulu University
Hospital, Oulu, Finland
(S Keinänen-Kiukaanniemi,
M Laakso); Department of Public
Health and Clinical Nutrition,
University of Kuopio, Kuopio,
Finland (A Louheranta PhD,
Prof M Uusitupa MD); Nutrition
Unit, Department of
Epidemiology and Health
Promotion, National Public
Health Institute, Helsinki,
Finland (M Paturi MSc);
Laboratory of Analytical
Biochemistry, Department of
Health and Functional Capacity,
National Public Health Institute,
Helsinki, Finland
(J Sundvall MSc); Health Centre
of Oulu, Oulu, Finland
(S Keinänen-Kiukaanniemi);
Department of Public Health,
University of Helsinki, Helsinki,
Finland (J Lindström,
J Tuomilehto, J G Eriksson);
South Ostrobothnia Central
Hospital, Seinäjoki, Finland
(J Tuomilehto)
Correspondence to:
Jaana Lindström
jaana.lindstrom@ktl.fi
Page 1
Articles
1674
www.thelancet.com Vol 368 November 11, 2006
of women 66%) or the control group (n=257, the
proportion of women 69%). Overweight (mean body-mass
index 31·1 kg/m
2
), middle-aged (mean age 55 years)
participants with impaired glucose tolerance based on
two 75 g oral glucose tolerance tests by the WHO 1985
criteria
11
were eligible for the study. Mean fasting plasma
glucose at baseline was 6·1 (SD 0·8) mmol/L and mean
plasma glucose value 2 h after the 75 g oral glucose load
was 8·9 (1·5) mmol/L without signifi cant diff erences
between the two groups. The overall proportion of
participants who were lost to follow-up was 10% in the
intervention group and 8% in the control group (p=0·3619
Fisher’s exact test; fi gure 1).
Intervention
The main goals of the intervention were: weight reduction
of 5% or more; less than 30% of the daily energy intake
from fat; less than 10% of the daily energy intake from
saturated fat; fi bre intake 15 g per 1000 kcal or more; and
moderately intense physical activity 30 min per day or
more. The duration of intervention ranged from less
than 1 year (indicating withdrawal before the fi rst yearly
visit) up to 6 years, with median length of 4 years. The
implementation of the intervention programme has been
previously reported.
12
Briefl y, the participants in the
intervention group were given detailed and individualised
counselling to achieve the lifestyle goals. They had seven
personal counselling sessions with the study nutritionist
during the fi rst year and every 3 months thereafter. The
median number of dietary counselling sessions per
participant was 20 thus indicating excellent compliance
with the study protocol. The participants were also
advised to increase their level of physical activity, and
were off ered free of charge, supervised, individually
tailored circuit-type moderate-intensity resistance
training sessions to improve the functional capacity and
strength of the large muscle groups of the upper and
lower body.
The participants in the control group were given
general verbal and written health behaviour information
at baseline without specifi c individualised advice. At the
last intervention period visit all the participants were
given a summary of their laboratory test results during
the intervention period, including the glucose values,
and they were also told about the fi ndings of the
randomised trial.
Post-intervention follow-up
All individuals who participated in the Diabetes
Prevention Study were invited to take part in the
post-intervention follow-up. During this follow-up, all
study participants had a yearly visit with the study nurse.
The visits included the same procedures as during the
intervention period, and were similar for all participants
irrespective of their former randomisation group. No
specifi c diet or exercise counselling was provided.
Procedures and measurements
The parameters measured every year included fasting
and post load (75 g oral glucose tolerance test) plasma
glucose after a 12-h fast. During the intervention, plasma
glucose was measured locally according to standard
guidelines. During the post-intervention follow-up,
centralised glucose assays were established enzymatically
with the hexokinase method (Thermo Electron Oy,
Vantaa, Finland).
A clinical examination was done and questionnaires
including questions about physical activity were obtained
at baseline and at every yearly visit. Individuals who
reported that they “mostly read, watch TV, and spend
time in other ways that are not physically demanding
during their spare time were categorised as physically
inactive, and those who reported “walking, bicycling, or
other exercise for at least 4 hours per week” were
categorised as achieving the physical activity goal. All
study participants completed a 3-day food record with a
picture booklet of portion sizes of typical foods.
13
The
average intakes of total fat (proportion of the total daily
energy intake), saturated fat (proportion of the total daily
energy intake), and dietary fi bre (g per 1000 kcal) from
the baseline and 1-year, 2-year, and 3-year visits of the
intervention period were calculated using a dietary
analysis programme and the Finnish Food Composition
27 diabetes cases
265 intervention group
First interim analysis
March 28, 2000
257 control group
522 randomisedRandomisation
November, 1993–June, 1998
23 withdrawals
17 new diabetes cases
1 new withdrawal
17 new diabetes casesEnd of intervention
1 new withdrawal
17 new diabetes cases
1 new withdrawal
31 new diabetes casesPost intervention follow-up
December 31, 2004
8 rejoined the trial
11 new withdrawals
17 new diabetes cases
1 new withdrawal
75 diabetes cases
163 without diabetes
27 lost to follow up†
17 new diabetes cases
1 new withdrawal
110 diabetes cases
127 without diabetes
20 lost to follow up†
17 new diabetes cases
1 new withdrawal
38 new diabetes cases
5 rejoined the trial
7 new withdrawals
13 new diabetes cases
1 new withdrawal
59 diabetes cases
17 withdrawals
Figure 1: Trial profi le
*After the decision to end the intervention period, the intervention was continued until each participant’s next
scheduled yearly clinic visit. End date thus varied from March, 2000, to Dec, 2001. †Participants who were lost to
follow-up were treated as censored observations in the analyses.
Page 2
Articles
www.thelancet.com Vol 368 November 11, 2006
1675
Database (Fineli) developed at the National Public Health
Institute, Helsinki, Finland.
14
The dietary analyses were
repeated at the fi rst post-intervention follow-up visit to
clarify the maintenance of the dietary changes after the
intervention had been discontinued.
The study participants were categorised according to
their success in achieving the fi ve predefi ned lifestyle
goals (0=not achieved, 1=achieved) by the 3-year visit,
with mean physical activity and nutrient intakes during
the years 1, 2, and 3. For those who either dropped out or
were diagnosed with diabetes before the 3-year visit, the
last observation for bodyweight was used for calculating
weight reduction. A success score from 0 to 5 was
calculated as the sum of the achieved goals. The analysis
was repeated at the fi rst post-intervention follow-up-
visit.
The development of type 2 diabetes was the primary
endpoint. Since the study was started before the cur-
rent criteria for diabetes were introduced,
15
diabetes was
defi ned according to WHO 1985 criteria,
11
ie, either
fasting plasma glucose of 7·8 mmol/L or more, or
2-hour post-challenge plasma glucose of 11·1 mmol/L or
more. The diagnosis of diabetes was confi rmed by a
second oral glucose tolerance test.
Statistical analysis
Kaplan-Meier survival curves were calculated to estimate
the probability of remaining free of diabetes in the two
groups. Participants who were lost during follow-up
were treated as censored observations. The diff erence
between the survival curves was tested with the log-rank
test. The Cox proportional hazards model was used to
estimate the hazard ratio for development of diabetes.
The proportionality assumption of the model was
assessed with graphical methods (ie, the log-log plot). All
comparisons of the endpoints were based on the
intention-to-treat principle.
Mean levels of bodyweight, nutrient intakes, and
physical activity during the study were compared between
the groups with analysis of covariance, adjusting for the
level of respective variable at baseline. Further, analysis of
covariance was used to examine changes in these variables
from the last intervention period visit until the fi rst
post-intervention examination. In this analysis, adjustment
was made for the level of respective variable at the last visit
during intervention. In further analyses, the Cox model
was used to analyse the relation between the success score
and the incidence of diabetes. First, the success score
variable was included in the model as categorical variable,
with those who did not achieve any of the lifestyle goals as
reference category. Additionally, test of linear trend was
done including the success score as continuous variable
in the model. In these analyses the groups were pooled.
The analyses were adjusted for treatment group, study
centre, sex, age, and the baseline 2-h post-challenge
plasma glucose concentration. Analyses were done with
the statistics package Stata version 8.0.
Role of the funding source
The sponsors of the study had no role in study design,
the collection, analysis, or interpretation of the data, or in
the writing of the report. The corresponding author had
full access to all data in the study and had the fi nal
responsibility to submit for publication.
Results
The total number of cases of diabetes diagnosed during
the overall follow-up of 7 years was 75 in the intervention
group and 110 in the control group (fi gure 1). The
incidence rates were 4·3 (95% CI 3·4–5·4) and
7·4 (6·1–8·9) per 100 person-years in the intervention
and control group, respectively (p=0·0001 log-rank test).
The corresponding hazard ratio was 0·57 (0·43–0·76;
gure 2). The cumulative incidence of diabetes at year 6
was 23% in the intervention group and 38% in the control
group, with an absolute risk reduction of 15% (7·2–23·2).
The number of people needed to be treated to prevent
one case of type 2 diabetes by lifestyle intervention was
22 for 1 year. The mean bodyweight and the intake of
total and saturated fat were lower in the intervention
group compared with that in the control group during
the intervention (table 1). Further, intake of dietary fi bre
and physical activity were higher in the intervention
group.
In the intervention and the control group, respectively,
10% and 27% of the participants did not achieve any of
the predefi ned goals by the 3-year examination, whereas
14% and 6% achieved four or fi ve goals (p<0·0001 for
Fisher’s exact test). There was a strong inverse correlation
between the success score and the incidence of diabetes
during the total follow-up. Incidence rate per
0·00
0·75
0·50
0·25
1·00
012345678
261
251
265
257
250
231
238
209
228
192
214
176
191
157
174
140
118
91
Log-rank test: p=0·0001
Hazard ratio=0·57 (95% CI 0·43–0·76)
Number at risk, intervention/control:
Kaplan-Meier estimate of probability of remaining free of diabetes
Follow up time (years)
Intervention
Control
Figure 2: Diabetes by treatment group
Follow-up time is truncated at 8 years, since number of participants at risk beyond this point was low, but they are
included in the calculation of hazard ratios.
Page 3
Articles
1676
www.thelancet.com Vol 368 November 11, 2006
100 person-years ranged from 8·4 (95% CI 6·2–11·3) in
the participants who did not achieve any of the goals at
the 3-year visit, to 2·0 (1·0–4·3) in those who achieved
four or fi ve of the goals. The hazard ratios were 1·00,
0·85 (0·57–1·28), 0.66 (0·40–1·09), 0·69 (0·38–1·26),
and 0·23 (0·10–0·52) for success score from 0, 1, 2, 3, to
4–5, respectively (test for trend p=0·0004).
To assess the independent eff ects of achieving the
success score components at the 3-year examination on
diabetes incidence during the total follow-up, all fi ve
variables for lifestyle goal were fi rst individually included
in a Cox model. Univariate hazard ratios (95% CI) were
0·45 (0·31–0·64) for weight reduction from baseline,
0·65 (0·45–0·95) for intake of fat, 0·59 (0·31–1·13) for
intake of saturated fat, 0·69 (0·49–0·96) for intake of
bre, and 0·62 (0·46–0·84) for physical activity,
comparing those who did or did not achieve the respective
goal. When all fi ve success score components were
simultaneously included in the Cox model, the
multivariate-adjusted hazard ratios for diabetes (95% CI)
were 0·43 (0·30–0·61) for weight reduction, 0·80
(0·48–1·34) for intake of fat, 0·55 (0·26–1·16) for intake
of saturated fat, 0·97 (0·63–1·51) for intake of fi bre, and
0·80 (0·57–1·12) for physical activity. Furthermore,
weight change from baseline was signifi cantly associated
with the achievement of each of the other four lifestyle
goals, and consequently, success score was strongly and
inversely correlated with weight reduction. The 3-year
weight reduction was 0·5%, 2·1%, 4·3%, 4·7%, and
8·7% for success score from 0, 1, 2, 3, to 4–5, respectively
(test for trend p<0·0001). Additionally, all the dietary
goals (total fat, saturated fat, and fi bre) were signifi cantly
associated with each other (p for all <0·0001).
Achievement of the fat intake goal or the fi bre intake goal
was associated also with the physical activity goal
(p=0·0019 and p<0·0001, respectively).
To explore whether the reduced long-term risk of type 2
diabetes in the intervention group could be attributed
solely to a reduced risk during the actual intervention of
the study, we excluded all participants who were
diagnosed with diabetes during the intervention (n=116)
and calculated the incidence rates exclusively for the
post-intervention follow-up. The median post-intervention
follow-up time was 3 years, and the number of incident
new cases of type 2 diabetes was 31 in the intervention
group of 221 people at risk, and 38 in the control group of
185 people at risk. The corresponding incidence rates
were 4·6 and 7·2 per 100 person-years, respectively
(log-rank test p=0·0401), ie, 36% relative risk reduction
(fi gure 3).
Bodyweight, physical activity, and nutrient intakes in
those without diabetes at the end of the intervention are
shown in table 2. The diff erences in these variables
between the groups remained favourable for the
intervention group during the post-intervention
follow-up. The proportion of physically active individuals
decreased in the control group. Conversely, the
participants in the control group reduced their intake of
saturated fat more but, since they had a higher intake to
start, still maintained a higher intake of saturated fats
than the intervention group.
Intervention Control p*
n Mean n Mean
Bodyweight (kg)
Baseline 265 86·7 257 85·5 0·3267
Year 1 256 82·2 250 84·8 <0·0001
Year 3† 256 83·4 251 85·2 <0·0001
Last intervention period visit† 257 84·3 251 85·6 <0·0001
Proportion of physically active (%)‡
Baseline 261 64 257 67 0·5192
Year 1 252 86 245 69 <0·0001
Year 3† 256 82 251 71 0·0003
Last intervention period visit† 256 81 251 71 0·0013
Energy proportion of fat (%)
Baseline 264 36 255 37 0·0670
Year 1 254 33 245 35 0·0001
Year 3† 254 32 246 34 <0·0001
Energy proportion of saturated fat (%)
Baseline 264 16 255 17 0·0188
Year 1 254 14 245 16 <0·0001
Year 3† 254 13 246 15 <0·0001
Dietary fi bre (g per 1000 kcal)
Baseline 264 11·7 255 11·7 0·9431
Year 1 254 14·2 245 12·5 <0·0001
Year 3† 254 14·1 246 12·7 <0·0001
*p for test of equality between groups, adjusting for baseline level.Last observation brought forward for individuals
who dropped out or became diabetic during the study. ‡Individuals who reported walking, cycling, or other moderate
intensity activity for at least 4 h per week categorised as physically active.
Table 1: Bodyweight, physical activity, and dietary intake during the intervention period of the study
0·00
0·75
0·50
0·25
1·00
01
194
165
2
180
148
3
167
134
4
102
73
221
185
Log-rank test: p=0·0401
Hazard ratio=0·61 (95% CI 0·38–0·98)
Number at risk, intervention/control:
Kaplan-Meier estimate of probability of remaining free of diabetes
Follow up time (years)
Intervention
Control
Figure 3: Diabetes by treatment group during the post-intervention follow-up period
Follow-up time is truncated at 4 years, since number of participants at risk beyond this point was low, but they are
included in calculation of hazard ratios.
Page 4
Articles
www.thelancet.com Vol 368 November 11, 2006
1677
The success score analysis was repeated to analyse the
eff ect of maintained lifestyle changes on the diabetes
incidence during the post-intervention follow-up. In the
intervention and control groups, respectively, 7% and
14% of the participants did not achieve any of the lifestyle
goals at the fi rst follow-up visit, 32% and 40% achieved
one, while 18% and 7% achieved at least four out of the
ve goals (p=0·0042 for Fisher’s exact test). The incidence
rate of diabetes per 100 person-years was 8·0 (95% CI
4·2–15·4) in the group that did not achieve any of the
goals, compared with 3·8 (1·7–8·5) in the group with 4
or 5 goals achieved. The hazard ratios were 1·00, 0·96
(0·45–2·04), 0·37 (0·15–0·93), 0·78 (0·32–1·91) and
0·54 (0·20–1·49) for the success score from 0, 1, 2, 3, to 4
or 5, respectively (p=0·1089).
Univariate hazard ratios (95% CI) for diabetes
incidence during the post-intervention follow-up were
0·55 (0·30–1·02) for achieving the weight reduction
goal, 0·74 (0·44–1·27) for achieving the fat intake goal,
1·01 (0·54–1·89) for achieving the saturated fat intake
goal, 0·72 (0·40–1·30) for achieving the fi bre intake goal,
and 0·62 (0·36–1·06) for achieving the physical activity
goal, compared with those who did not achieve the
respective goal at the fi rst post-intervention follow-up
examination. When all fi ve variables for lifestyle goals
were simultaneously analysed, the adjusted hazard ratios
were 0·52 (0·28–0·96) for weight reduction from
baseline, 0·67 (0·35–1·31) for the intake of fat, 1·62
(0·68–3·85) for the intake of saturated fat, 0·77
(0·38–1·57) for the intake of fi bre, and 0·82 (0·46–1·48)
for physical activity.
Discussion
Individually randomised controlled lifestyle intervention
studies have shown the benefi t of healthy lifestyle on
delaying the deterioration of glucose tolerance to
manifest type 2 diabetes, at least as long as the inter–
vention continues.
5–8
Our study with a median of 7 years
total follow-up shows that a marked diff erence in the
cumulative incidence of diabetes can be sustained after
the discontinuation of active counselling. The absolute
diff erence in diabetes risk between the intervention and
control groups was about 15% during the initial trial
period and also remained the same during the
post-intervention follow-up. The relative risk reduction
of 43% was, however, less than the 58% seen during the
original study,
5
as expected from the increasing
cumulative diabetes incidence in both groups.
The earlier Da Qing IGT and Diabetes Study
4
with
clinics randomly assigned either to diet, exercise, or diet
plus exercise intervention showed a 31%, 46%, and 42%
risk reduction, respectively, after a 6-year intervention.
The relative risk reduction achieved in our study was
about the same after a similar period even though the
duration of active intervention was shorter. Thus, from a
public health point of view there is an important message:
an intensive lifestyle intervention lasting for a limited
time can yield long-term benefi ts in reducing the risk of
type 2 diabetes in high-risk individuals.
The achieved changes in physical activity and dietary
habits seemed to be maintained at least 1 year after the
discontinuation of the intervention. The diff erences
between the groups persisted despite a possible dilution
eff ect, since the control group participants can be
considered to have received a reinforced mini-
intervention when they were provided with their own
glucose results and told about the main fi ndings of the
Diabetes Prevention Study at the end of the intervention
period. Still, a modest diff erence in bodyweight change
from baseline between the intervention and control
groups was preserved. Our results confi rm the fi ndings
from earlier studies showing that interventions can
have long-term eff ect on lifestyle,
16,17
and off er
encouraging evidence for the effi cacy of comprehensive
lifestyle intervention even without large reduction in
weight.
Analysis of the success score showed that most people
who maintained the lifestyle goals at 3-year visit remained
free of diabetes during the extended follow-up. This
nding indicates that the true eff ect of healthy lifestyle
results in a dramatically better outcome than that seen by
the intention-to-treat analysis of the treatment eff ect.
Each of the success score components at the 3-year visit
Intervention Control p* p†
n Mean n Mean
Bodyweight (kg)
Baseline 190 84·9 165 84·0 0·5174
Last intervention visit 190 81·8 165 83·3 <0·0001
First post-intervention follow-up visit 190 83·1 165 84·0 0·0032 0·1482
Proportion of physically active (%)‡
Baseline 184 70 164 70 0·9102
Last intervention visit 187 88 164 76 0·0035
First post-intervention follow-up visit 187 86 164 71 0·0005 0·0273
Energy proportion of fat (%)
Baseline 187 36 159 37 0·1879
Year 3‡ 187 31 159 34 0·0002
First post-intervention follow-up visit 187 31 159 33 0·0174 0·1189
Energy proportion of saturated fat (%)
Baseline 187 16 159 17 0·0676
Year 3‡ 187 13 159 15 <0·0001
First post-intervention follow-up visit 187 12 159 14 0·0001 0·0128
Dietary fi bre (g per 1000 kcal)
Baseline 187 11·9 159 11·9 0·9750
Year 3‡ 187 14·5 159 12·9 0·0003
First post-intervention follow-up visit 187 13·6 159 12·6 0·0071 0·4577
*p for test of equality between the groups, adjusting for the baseline level. †p for test of equal change between the
groups from the last intervention period visit to the fi rst post-intervention follow-up visit, adjusting for the level at the
last intervention visit. ‡Individuals who reported walking, cycling, or other moderate intensity activity for at least 4 h
a week were categorised as physically active.
Table 2: Bodyweight, physical activity, and dietary intakes of participants of the post-intervention
follow-up period who were without diabetes at the end of the intervention
Page 5
Articles
1678
www.thelancet.com Vol 368 November 11, 2006
(except that for saturated fat intake) was signifi cantly
associated with the reduction in diabetes risk in
univariate analyses, but when all the components were
included into the model simultaneously, only the eff ect
of weight reduction remained signifi cant. Analyses from
the fi rst post-intervention follow-up visit revealed
similar tendency; however, the only signifi cant
association was between weight reduction and diabetes
risk in the multivariate model. The fi ndings suggest
that dietary composition and physical activity are
important in diabetes prevention but their eff ect on
diabetes risk is in large part, although not entirely,
mediated through resulting weight reduction.
Nevertheless, because of the multicollinearity shown by
the fact that weight change correlated with all the other
intervention goals, the interpretation of the results
should be made cautiously.
Our fi ndings do not allow us to distinguish between
the carry-over eff ect from the intervention, and the
ongoing eff ect of lifestyle during the post-intervention
follow-up, on diabetes incidence. In a subgroup
analysis of the Diabetes Prevention Study population,
we showed a marked improvement in insulin
sensitivity concomitantly with weight loss, whereas
insulin secretion did not change signifi cantly.
18
This
nding suggests that the prolonged benefi t of the
lifestyle intervention on the diabetes risk could partly
be attributed to a correction of insulin resistance,
which, on the other hand, might result in a preservation
of the beta cell function. Even so, we cannot rule out
the eff ect of maintaining lifestyle changes after the
original intervention period. The question concerning
the risk reduction in those in whom the success score
or its components changed during the post-intervention
follow-up period would be of interest, but unfortunately
at the present our data have restricted statistical power
for this kind of subgroup analyses.
About a third of participants in the intervention group
met none or only one of the predefi ned goals 1 year after
the intervention. Adherence to the intervention is a
specifi c challenge for future diabetes prevention
programmes. Oral antidiabetic medications have been
shown to prevent diabetes, and could be an option for
those who have not responded satisfactorily to lifestyle
intervention. However, medications seem to lower blood
glucose as long as they are taken, but much of their
eff ect dissipates as soon as the drug is discontinued.
6,19–21
Unfortunately, there has thus far been no pharmacological
trial specifi cally targeted to people at high risk of diabetes
and who were unable to change their lifestyle.
Some limitations of the present study have to be
addressed. The analyses related to the post-intervention
follow-up period of the Diabetes Prevention Study were
not planned in the original study protocol, and post hoc
analyses have to be interpreted with caution. The
post-intervention follow-up was not foreseen while
calculating the original sample size,
10
and because of
low numbers of people at risk and cases of diabetes the
statistical power remains restricted. Furthermore, the
study participants were volunteers and willing to take
part in a long-lasting trial and thus were probably more
health-conscious than the general population. A low
number of withdrawals is a marker of high commitment,
but since there was no diff erence between the groups it
is also an advantage in the analyses. Future studies will
reveal if the results from this clinical trial can be
transposed into usual health-care settings. Also the
generalisability of our fi ndings in other populations
must be studied.
Based on the Kaplan-Meier analysis, around 50% of
people with impaired glucose tolerance will develop
diabetes during 10 years when no active intervention is
applied. Although a lifestyle intervention alone, even if
successful, does not necessarily prevent type 2 diabetes
in all individuals, it will still postpone the onset of the
disease. Even delaying the onset of diabetes can have a
substantial eff ect on subsequent morbidity, and therefore
on the cost-eff ectiveness of diabetes prevention.
22
Whether the lifestyle intervention used in the Diabetes
Prevention Study reduces diabetes-related microvascular
and macrovascular complications in the long run is still
to be proven, and such an assessment is planned in the
future after an adequate number of cases and
person-years have been accumulated.
The high diabetes incidence even in the intervention
group of our study suggests that preventive actions
should probably be targeted to all high-risk individuals,
even before impaired glucose tolerance is present. The
lifestyle intervention used in the Diabetes Prevention
Study has formed the basis for the implementation
programme for the prevention of type 2 diabetes in
Finland.
23
This programme identifi es high risk
individuals with a simple, validated risk score
questionnaire
24,25
and thus is likely to reach people at an
earlier stage in the process leading to diabetes. Although
a population-based strategy to fi ght the pandemic of
type 2 diabetes is urgently needed, an individualised
approach to guide people at high risk is also warranted.
A simple lifestyle intervention seems to work well.
However, further research is needed to reveal the
optimum and most cost-effi cient strategy, intensity, and
duration of such an intervention. The results from the
extended follow-up of the Finnish Diabetes Prevention
Study nevertheless show that the eff ect of lifestyle
intervention on diabetes risk does not disappear after
active lifestyle counselling is stopped.
Contributors
J Lindström and P Ilanne-Parikka had joint responsibility for writing this
manuscript and share the primary authorship of this paper. M Peltonen
did the statistical analyses and participated in writing the manuscript.
S Aunola, J G Eriksson, K Hemiö, H Hämäläinen, P Härkönen,
S Keinänen-Kiukaanniemi, M Laakso, A Louheranta, M Paturi, J Sundvall,
and T T Valle contributed to data extraction and revised the manuscript.
J Tuomilehto and M Uusitupa are the principal investigators of the study
and participated in writing the manuscript.
Page 6
Articles
www.thelancet.com Vol 368 November 11, 2006
1679
Acknowledgments
We thank Zygimantas Cepaitis, Martti Hakumäki, Anja Ilmanen,
Hilkka Kaisto, Kaija Kettunen, Hannele Kivelä, Arja Kylliäinen,
Ritva Läärä, Paula Nyholm, and Arja Putila for their skillful assistance in
the study; and Marja-Riitta Taskinen and Antti Aro for their participation
in the endpoint committee. This study was supported by the Academy of
Finland (grants 8473/2298, 40758/5767, 38387/54175, 46558), Juho Vainio
Foundation, Ministry of Education, Novo Nordisk Foundation, Yrjö
Jahnsson Foundation and Finnish Diabetes Research Foundation, and
EVO funds from Tampere, Kuopio, and Oulu University Hospitals.
Confl ict of interest statement
We declare that we have no confl ict of interest.
References
1 King H, Aubert RE, Herman WH. Global burden of diabetes,
1995–2025: prevalence, numerical estimates, and projections.
Diabetes Care 1998; 21: 1414–31.
2 Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence of
diabetes: estimates for the year 2000 and projections for 2030.
Diabetes Care 2004; 27: 1047–53.
3 Eriksson KF, Lindgärde F. Prevention of type 2 (non-insulin-
dependent) diabetes mellitus by diet and physical exercise. The
6-year Malmo feasibility study. Diabetologia 1991; 34: 891–98.
4 Pan XR, Li GW, Hu YH, et al. Eff ects of diet and exercise in
preventing NIDDM in people with impaired glucose tolerance. The
Da Qing IGT and Diabetes Study. Diabetes Care 1997; 20: 537–44.
5 Tuomilehto J, Lindström J, Eriksson JG, et al. Prevention of type 2
diabetes mellitus by changes in lifestyle among subjects with
impaired glucose tolerance. N Engl J Med 2001; 344: 1343–50.
6 The Diabetes Prevention Program Research Group. Reduction in
the incidence of type 2 diabetes with lifestyle intervention or
metformin. N Engl J Med 2002; 346: 393–403.
7 Ramachandran A, Snehalatha C, Mary S, Mukesh B, Bhaskar AD,
Vijay V. The Indian Diabetes Prevention Programme shows that
lifestyle modifi cation and metformin prevent type 2 diabetes in
Asian Indian subjects with impaired glucose tolerance (IDPP-1).
Diabetologia 2006; 49: 289–97.
8 Kosaka K, Noda M, Kuzuya T. Prevention of type 2 diabetes by
lifestyle intervention: a Japanese trial in IGT males.
Diabetes Res Clin Pract 2005; 67: 152–62.
9 Yamaoka K, Tango T. Effi cacy of lifestyle education to prevent type 2
diabetes: a meta-analysis of randomized controlled trials.
Diabetes Care 2005; 28: 2780–86.
10 Eriksson J, Lindström J, Valle T, et al. Prevention of Type II diabetes
in subjects with impaired glucose tolerance: the Diabetes
Prevention Study (DPS) in Finland. Study design and 1-year interim
report on the feasibility of the lifestyle intervention programme.
Diabetologia 1999; 42: 793–801.
11 WHO. Diabetes Mellitus: report of a WHO Study Group. Technical
Report Series No 727. Geneva: World Health Organization, 1985.
12 Lindström J, Louheranta A, Mannelin M, et al. The Finnish
Diabetes Prevention Study (DPS): lifestyle intervention and 3-year
results on diet and physical activity. Diabetes Care 2003; 26: 3230–36.
13 Haapa E, Toponen T, Pietinen P, Räsänen L. Annoskuvakirja.
Helsinki: Painokaari Oy, 1985.
14 Ovaskainen M-L, Valsta L, Lauronen J. The compilation of food
analysis values as a database for dietary studies–the Finnish
experience. Food Chemistry 1996; 57: 133–36.
15 WHO. Defi nition, diagnosis and classifi cation of diabetes mellitus
and its complications: report of a WHO Consultation: Part 1:
Diagnosis and classifi cation of diabetes mellitus. Geneva: World
Health Organization, Department of Noncommunicable Disease
Surveillance, 1999.
16 Pereira MA, Kriska AM, Day RD, Cauley JA, LaPorte RE, Kuller LH.
A randomized walking trial in postmenopausal women: eff ects on
physical activity and health 10 years later. Arch Intern Med 1998;
158: 1695–701.
17 Simkin-Silverman LR, Wing RR, Boraz MA, Kuller LH. Lifestyle
intervention can prevent weight gain during menopause: results
from a 5-year randomized clinical trial. Ann Behav Med 2003;
26: 212–20.
18 Uusitupa M, Lindi V, Louheranta A, Salopuro T, Lindström J,
Tuomilehto J. Long-term improvement in insulin sensitivity by
changing lifestyles of people with impaired glucose tolerance: 4-year
results from the Finnish Diabetes Prevention Study. Diabetes 2003;
52: 2532–38.
19 The Diabetes Prevention Program Research Group. Eff ects of
withdrawal from metformin on the development of diabetes in the
Diabetes Prevention Program. Diabetes Care 2003; 26: 977–80.
20 The Diabetes Prevention Program Research Group. Prevention of
type 2 diabetes with troglitazone in the Diabetes Prevention
Program. Diabetes 2005; 54: 1150–56.
21 Chiasson J-L, Josse RG, Gomis R, et al. Acarbose for prevention of
type 2 diabetes mellitus: the STOP-NIDDM randomised trial. Lancet
2002; 359: 2072–77.
22 Herman WH, Hoerger TJ, Brandle M, et al. The cost-eff ectiveness
of lifestyle modifi cation or metformin in preventing type 2 diabetes
in adults with impaired glucose tolerance. Ann Intern Med 2005;
142: 323–32.
23 The Finnish Diabetes Association. Programme for the prevention
of type 2 diabetes in Finland 2003–2010. www.diabetes.fi /english.
Tampere, 2003.
24 Lindström J, Tuomilehto J. The Diabetes Risk Score: A practical tool
to predict type 2 diabetes risk. Diabetes Care 2003; 26: 725–31.
25 Saaristo T, Peltonen M, Lindström J, et al. Cross-sectional
evaluation of the Finnish Diabetes Risk Score: a tool to identify
undetected type 2 diabetes, abnormal glucose tolerance and
metabolic syndrome. Diab Vasc Dis Res 2005; 2: 67–72.
Page 7
  • Source
    • "Even the accelerated bodily deterioration caused by well-defined genetic substrates as in Huntington's and Duchenne's diseases is influenced by the environment (Wexler et al., 2004; Jansen et al., 2013 ). As a consequence, environmental interventions can prevent or postpone cardiovascular disease, diabetes mellitus, and cancer (Lindström et al., 2006; Estruch et al., 2013; Rasmussen-Torvik et al., 2013). Infectious diseases, accidents, and natural disasters require environmental risk factors, but cannot be uncoupled from the body's vulnerability that increases over age. "
    [Show abstract] [Hide abstract] ABSTRACT: Intrinsic and extrinsic mortality are often separated in order to understand and measure aging. Intrinsic mortality is assumed to be a result of aging and to increase over age, whereas extrinsic mortality is assumed to be a result of environmental hazards and be constant over age. However, allegedly intrinsic and extrinsic mortality have an exponentially increasing age pattern in common. Theories of aging assert that a combination of intrinsic and extrinsic stressors underlies the increasing risk of death. Epidemiological and biological data support that the control of intrinsic as well as extrinsic stressors can alleviate the aging process. We argue that aging and death can be better explained by the interaction of intrinsic and extrinsic stressors than by classifying mortality itself as being either intrinsic or extrinsic. Recognition of the tight interaction between intrinsic and extrinsic stressors in the causation of aging leads to the recognition that aging is not inevitable, but malleable through the environment. Copyright © 2015. Published by Elsevier Inc.
    Full-text · Article · Apr 2015 · Experimental gerontology
  • Source
    • "The direct cost of obesity in Canada was recently estimated at 4.3 billion dollars, in addition to 5.3 billion dollars related to sedentary lifestyles: these costs represent 4.8% of the global Canadian health care system's budget [5]. Nonetheless, moderate weight loss in obese subjects has been proven to markedly ameliorate DM2, blood pressure, lipid profile and may decrease mortality678910. In response to the alarming evolution of obesity in Canada, the 2005 federal budget provided $300 million over 5 years for a strategy focusing on healthy living and prevention of chronic disease [11]. "
    [Show abstract] [Hide abstract] ABSTRACT: The majority of obese subjects are treated by primary care physicians (PCPs) who often feel uncomfortable with the management of obesity. In a previous study, we successfully developed, implemented and evaluated an obesity management system based on training and coaching of health professionals of family medicine groups (FMGs) by a team of experts in obesity management. Using a pre/post design, this study suggested a positive impact on health professionals' perceptions and reported obesity care. The current research project is aimed at evaluating the impact on obesity screening and care of this integrated obesity management system. We hypothesize that our program combining preceptorships with a virtual community and on-site coaching will improve: (1) management and weight loss of obese/overweight subjects treated by PCPs for hypertension, type 2 diabetes or impaired glucose tolerance; and (2) screening and initial management of obesity among a regular follow-up group of patients of PCPs who practice in FMGs. Ten FMGs will be approached for a practice monitoring project and will be randomised to receive the intervention developed in our previous project or will only be provided clinical practice guidelines. In the participating FMGs, we will enrol 22 patients per FMG with weight related targeted disease and 24 patients with regular follow-up. These patients will be evaluated for the care they received regarding screening and/or management of obesity using medical chart reviews, and will fill out a questionnaire on their lifestyle and satisfaction. They will also be examined for anthropometric measures, vital signs, blood markers for chronic diseases and physical fitness. The same patients will be assessed again after 18 months. The impact of the program on health professionals will be evaluated at baseline, and at 1 year. Qualitative data will also be collected from both professional and patient participants. Direct and indirect costs and QALYs will be evaluated as indicators of cost-effectiveness. In the context of the dramatic increase in obesity prevalence and the low perception of PCPs' self-efficacy, providing efficient strategies to PCPs and interdisciplinary health care teams for management of obesity is crucial. Clinicaltrials.gov Identifier: NCT00991640.
    Full-text · Article · Dec 2014
  • Source
    • "It has been shown that T2D can be efficiently prevented by lifestyle intervention in high-risk individuals in research setting . Furthermore, these lifestyle interventions seem to have a long-lasting effect on T2D risk factors [9] [10]. Finland was one of the first countries in the world to establish a national program for prevention of T2D, accompanied by a prevention implementation project within the primary health care [11]. "
    [Show abstract] [Hide abstract] ABSTRACT: Aims: To evaluate feasibility and effectiveness of lifestyle counseling in occupational setting on decreasing risk for diabetes and cardiovascular disease. Methods: A health check-up including physical examination, blood tests, questionnaires and health advice was completed on 2312 employees of an airline company. Participants with elevated risk for type 2 diabetes based on FINDRISC score and/or blood glucose measurement (n=657) were offered 1-3 additional lifestyle counseling sessions and 53% of them agreed to participate. After 2.5 years, 1347 employees of 2199 invited participated in a follow-up study. Results: Among women and men with low baseline diabetes risk, cardiovascular risk factors increased slightly during follow-up. Larger proportion of the men who attended interventions lost weight at least 5% compared with the non-attendees (18.4% vs. 8.4%, p=0.031) and their FINDRISC score increased less (0.6 vs. 1.5, p=0.037). Older age associated with participation in follow-up and higher baseline FINDRISC score and presence of clinical and lifestyle risk factors and problems in sleep and mood increased attendance in interventions. Conclusions: Identification of employees with cardiovascular and diabetes risk, and the low intensity lifestyle intervention were feasible in occupational health-care setting. However, the health benefits were modest and observed only for men with increased risk.
    Full-text · Article · Aug 2014 · Primary Care Diabetes
Show more