The
new england journal
of
medicine
n engl j med 359;15 www.nejm.org october 9, 2008
1577
original article
10-Year Follow-up of Intensive Glucose
Control in Type 2 Diabetes
Rury R. Holman, F.R.C.P., Sanjoy K. Paul, Ph.D., M. Angelyn Bethel, M.D.,
David R. Matthews, F.R.C.P., and H. Andrew W. Neil, F.R.C.P.
From the Diabetes Trials Unit (R.R.H.,
S.K.P., M.A.B.), the Division of Public
H e a l t h a n d P r i m a r y H e a l t h C a r e ( H . A . W . N . ) ,
and the National Institute of Health Re-
search (NIHR) School for Primary Care
Research (H.A.W.N.), Oxford Centre for
Diabetes, Endocrinology, and Metabo-
lism (R.R.H., S.K.P., M.A.B., D.R.M.,
H.A.W.N.); and the NIHR Oxford Bio-
medical Research Centre (R.R.H., D.R.M.,
H.A.W.N.) — both in Oxford, United
Kingdom. Address reprint requests to Dr.
Holman at the Diabetes Trials Unit, Ox-
ford Centre for Diabetes, Endocrinology,
and Metabolism, Churchill Hospital, Head-
ington, Oxford OX3 7LJ, United Kingdom,
or at rury.holman@dtu.ox.ac.uk.
This article (10.1056/NEJMoa0806470) was
published at www.nejm.org on September
10, 2008.
N Engl J Med 2008;359:1577-89.
Copyright © 2008 Massachusetts Medical Society.
Abstract
Background
During the United Kingdom Prospective Diabetes Study (UKPDS), patients with
type 2 diabetes mellitus who received intensive glucose therapy had a lower risk of
microvascular complications than did those receiving conventional dietary therapy.
We conducted post-trial monitoring to determine whether this improved glucose con-
trol persisted and whether such therapy had a long-term effect on macrovascular
outcomes.
Methods
Of 5102 patients with newly diagnosed type 2 diabetes, 4209 were randomly assigned
to receive either conventional therapy (dietary restriction) or intensive therapy (either
sulfonylurea or insulin or, in overweight patients, metformin) for glucose control.
In post-trial monitoring, 3277 patients were asked to attend annual UKPDS clinics
for 5 years, but no attempts were made to maintain their previously assigned thera-
pies. Annual questionnaires were used to follow patients who were unable to attend
the clinics, and all patients in years 6 to 10 were assessed through questionnaires.
We examined seven prespecified aggregate clinical outcomes from the UKPDS on
an intention-to-treat basis, according to previous randomization categories.
Result s
Between-group differences in glycated hemoglobin levels were lost after the first
year. In the sulfonylurea–insulin group, relative reductions in risk persisted at 10
years for any diabetes-related end point (9%, P = 0.04) and microvascular disease
(24%, P = 0.001), and risk reductions for myocardial infarction (15%, P = 0.01) and
death from any cause (13%, P = 0.007) emerged over time, as more events occurred.
In the metformin group, significant risk reductions persisted for any diabetes-relat-
ed end point (21%, P = 0.01), myocardial infarction (33%, P = 0.005), and death from
any cause (27%, P = 0.002).
Conclusions
Despite an early loss of glycemic differences, a continued reduction in microvascu-
lar risk and emergent risk reductions for myocardial infarction and death from any
cause were observed during 10 years of post-trial follow-up. A continued benefit after
metformin therapy was evident among overweight patients. (UKPDS 80; Current
Controlled Trials number, ISRCTN75451837.)
The New England Journal of Medicine
Downloaded from nejm.org at UQ Library on March 28, 2017. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
The
new england journal
of
medicine
n engl j med 359;15 www.nejm.org october 9, 2008
1578
T
he United Kingdom Prospective Dia-
betes Study (UKPDS), a randomized, pro-
spective, multicenter trial, showed that in-
tensive glucose therapy in patients with newly
diagnosed type 2 diabetes mellitus was associ-
ated with a reduced risk of clinically evident mi-
crovascular complications and a nonsignificant
reduction of 16% in the relative risk of myocar-
dial infarction (P = 0.052).
1
In patients whose body
weig ht w a s more t h a n 1 20% of t he i r idea l weig ht
2
and who pr ima rily received metformin, reduct ions
in the risk of myocardial infarction of 39%
(P = 0.01) and of death from any cause of 36%
(P = 0.01) were observed. The results of the UKPDS,
which were published in 1998, have appeared to be
influential in subsequent diabetes management.
3,4
In patients with type 1 diabetes, the Diabetes
Control and Complications Trial/Epidemiology of
Diabetes Interventions and Complications (DCCT/
EDIC) study reported a postinterventional micro-
vascular benefit and the emergence of macrovas-
cular risk reduction from earlier improved glyce-
mic control during an 8-year period.
5
The Steno-2
Study repor ted a similar outcome duri ng a 5.5-year
period after earlier multifactorial risk reduction
among patients with type 2 diabetes.
6
In both
trials, enhanced risk reductions occurred despite
the loss of within-trial differences in glucose
levels and, in the Steno-2 Study, diminished dif-
ferences in blood pressure and lipid levels, sug-
gesting the persistence of effects of earlier im-
proved metabolic management.
We report here the results of a 10-year, post-
interventional follow-up of the UKPDS survivor
cohort that examined whether a continued micro-
vascular benefit from earlier improved glucose
control was evident and whether such therapy had
a long-term effect on macrovascular outcomes.
Methods
Patients
The recruitment of patients, study protocol, and
methods for the UKPDS have been reported pre-
viously.
7,8
Approval was obtained from the ethics
committees at all 23 clinical centers, and the
study conformed to the Declarations of Helsinki
guidelines. Brief ly, 5102 of 7616 pat ients who un-
derwent screening were enrolled from 1977 to
1991. All patients provided written informed con-
sent. Patients were between the ages of 25 and 65
years and had a fasting plasma glucose level of
more than 108 mg per deciliter (6.0 mmol per li-
ter) on two occasions after their general practitio-
ners had diagnosed type 2 diabetes. By self-report,
81% of the patients were white, 10% Asian Indi-
an, a nd 9% A f ro-Caribbea n. A tot al of 2514 pat ients
were excluded because of the following condi-
tions: ketonuria, a serum creat inine level of more
than 175 μmol per liter (2.0 mg per deciliter), myo-
cardial infarction in the previous year, current
angina or heart failure, more than one major vas-
cular event, retinopathy requiring laser treatment,
malignant hypertension, uncorrected endocrine
disorder, occupations precluding insulin therapy,
severe concu rrent i llness l imit ing li fe expect anc y,
inadequate understanding of the study protocol,
or unwillingness to enter the study.
After a 3-month dietary run-in period, patients
with a fasting plasma glucose level of more than
108 mg per deciliter but less than 270 mg per
deciliter (15.0 mmol per liter) were randomly as-
signed to receive conventional glucose control
(diet) or intensive glucose control (sulfonylurea or
insulin or, if more than 120% of ideal body
weight, metformin
2
). All patients were seen quar-
terly in UKPDS clinics.
7
The median follow-up for
the sulfonylurea–insulin and metformin groups
was 10.0 years
1
and 10.7 years,
9
respectively.
Post-Trial Monitoring
All surviving patients entered the post-trial mon-
it or i ng pro g ra m a f t er t he int er vent ion t r i a l clo sed
on September 30, 1997. A 10-year follow-up was
planned to coincide wit h a projected deat h rate of
50%. In September 1998, when the UKPDS results
were published,
10,11
patients and clinicians were
advised that it was necessary to lower levels of
blood glucose and blood pressure as much as pos-
sible. Patients returned to community or hospital-
based diabetes care according to their clinical
needs, with no attempt to maintain previously
randomized therapies. They were seen annually
for 5 years in UKPDS clinics, with continued
standardized collection of outcome data; mea-
surements of blood pressure, fasting plasma glu-
cose, glycated hemoglobin, plasma creatinine, and
the ratio of albumin to creatinine; and results on
two questionnaires, the European Quality of Life–5
Dimensions (EQ-5D)
12
and a questionnaire on the
use of health resources.
Clinical examinations every 3 years were con-
tinued. Patients who were unable to attend clinics
were sent EQ-5D and health-resource question-
The New England Journal of Medicine
Downloaded from nejm.org at UQ Library on March 28, 2017. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
10-Year Follow-up of Intensive Glucose Control
n engl j med 359;15 www.nejm.org october 9, 2008
1579
naires, and additional questionnaires were sent
to their general practitioners to capture possible
clinical outcomes. In years 6 to 10, these ques-
tionnaires were used to follow patients remotely,
since funding for clinic visits was not available.
Final questionnaires were sent to all remaining
patients after the cutoff for the censoring of post-
trial data on September 30, 2007.
Clinical Outcomes
The study administrator obtained full documen-
tation for all putative outcomes from hospitals
and general practitioners, whether reported at
clinic visits or by means of questionnaires. The
vit al status for a ll patients who were st ill living in
the United Kingdom was obtained from the Of-
f ice of National Statistics. Members of the UKPDS
end-point committee, who were unaware of as-
signments to study groups, adjudicated outcomes
exactly as they had during the original trial. The
seven prespecified UKPDS aggregate clinical out-
comes
7
were any diabetes-related end point (sud-
den deat h, death from hyperglycemia or hypogly-
cemia, fatal or nonfatal myocardial infarction,
angina, heart failure, fatal or nonfatal stroke, re-
nal failure, amputation, vitreous hemorrhage,
retinal photocoagulation, blindness in one eye, or
cataract extraction), diabetes-related death (sud-
den death or death from myocardial infarction,
stroke, peripheral vascular disease, renal disease,
hyperglycemia, or hypoglycemia), death from any
cause, myocardial infarction (sudden death or fa-
39p6
4209 Were randomly assigned to receive
either intensive or conventional therapy
5102 Patients with newly diagnosed
type 2 diabetes were enrolled
2729 Were assigned to receive
intensive therapy with
sulfonylurea or insulin
342 Were assigned to
receive intensive
therapy with metformin
50 Died
7 Emigrated
6 Had no final-
year data
2118 Were available for
post-trial monitoring
279 Were available for
post-trial monitoring
1138 Were assigned to receive
conventional therapy
primarily with diet
880 Were available for
post-trial monitoring
489 Died
57 Emigrated
65 Had no final-
year data
213 Died
19 Emigrated
26 Had no final-
year data
102 Died
6 Emigrated
35 Had no final-
year data
1010 Completed post-trial
monitoring
136 Completed post-trial
monitoring
379 Completed post-trial
monitoring
674 Died
44 Emigrated
390 Had no final-
year data
324 Died
13 Emigrated
164 Had no final-
year data
AUTHOR:
FIGURE:
JOB: ISSUE:
4-C
H/T
RETAKE
SIZE
ICM
CASE
EMail
Line
H/T
Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
REG F
Enon
1st
2nd
3rd
Holman
1 of 4
10-09-08
ARTIST: ts
35915
Figure 1. Enrollment and Outcomes.
Of the 4209 patients who underwent randomization in the original United Kingdom Prospective Diabetes Study, 78% entered post-trial
monitoring. Of the 1138 patients who were assigned to receive conventional therapy, 411 were overweight. These were compared with
the 342 overweight patients who were assigned to receive intensive therapy with metformin. Vital status was not available for any of the
patients who left the United Kingdom during follow-up. For patients for whom no final-year data were available, the analysis may not
have captured all nonfatal events. Overall, 3.5% of patients were lost to follow-up.
The New England Journal of Medicine
Downloaded from nejm.org at UQ Library on March 28, 2017. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
The
new england journal
of
medicine
n engl j med 359;15 www.nejm.org october 9, 2008
1580
Table 1. Baseline Characteristics of the Patients.*
Variable
Data Available in Final Year of Post-Trial
Monitoring Sulfonylurea–Insulin Group Metformin Group
Yes
(N = 1525)
No
(N = 705) P Value†
Conventional
Therapy
(N = 880)
Intensive
Therapy
(N = 2118) P Value†
Conventional
Therapy
(N = 309)
Intensive
Therapy
(N = 279) P Value†
Age — yr 62±8 60±9 0.002 63±9 63±9 0.78 63±9 64±9 0.56
Male sex — no. (%) 892 (58.5) 391 (55.5) 0.22 532 (60.5) 1248 (58.9) 0.44 142 (46.0) 127 (45.5) 0.92
Race or ethnic group — no. (%)‡ <0.001 0.59 0.42
White 1161 (76.1) 564 (80.0) 710 (80.7) 1717 (81.1) 262 (84.8) 235 (84.2)
Afro-Caribbean 143 (9.4) 51 (7.2) 58 (6.6) 159 (7.5) 23 (7.4) 28 (10.0)
Asian Indian 209 (13.7) 84 (11.9) 105 (11.9) 230 (10.9) 21 (6.8) 12 (4.3)
Other 12 (0.8) 6 (0.8) 7 (0.8) 12 (0.6) 3 (1.0) 4 (1.4)
Weight — kg 0.97 0.01 0.42
Median 81.0 81.00 79.0 80.0 87.0 86.0
Interquartile range 71.00–92.00 70.00–91.00 69.0–90.0 71.0–92.0 76.0–97.0 75.0–95.8
Body-mass index 29.4±5.5 29.4±5.4 0.86 28.7±5.6 29.3±5.5 0.005 32.2±5.7 31.7±5.4 0.34
Blood pressure — mm Hg
Systolic 137±19 137±19 0.98 138±21 139±20 0.52 139±22 141±18 0.43
Diastolic 77±10 78±10 0.22 77±10 77±10 0.06 77±10 78±10 0.22
Fasting plasma glucose — mg/dl 164±59 168±61 0.34 178±58 161±61 <0.001 182±55 177±64 0.12
Glycated hemoglobin — % 0.25 <0.001 0.12
Median 8.0 8.1 8.5 7.9 8.9 8.4
Interquartile range 6.9–9.4 7.0–9.6 7.3–9.7 6.8–9.2 7.5–10.0 7.2–9.7
Cholesterol — mg/dl
Total 198±39 198±37 0.66 197±37 197±39 0.63 200±37 204±41 0.37
Low-density lipoprotein 127±34 127±32 0.81 126±32 126±34 0.92 129±32 130±36 0.98
High-density lipoprotein 42±12 43±13 0.87 43±12 42±13 0.23 40±12 42±13 0.08
The New England Journal of Medicine
Downloaded from nejm.org at UQ Library on March 28, 2017. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
10-Year Follow-up of Intensive Glucose Control
n engl j med 359;15 www.nejm.org october 9, 2008
1581
tal or nonfatal myocardial infarction), stroke (fa-
tal or nonfatal stroke), peripheral vascular disease
(amputation of at least one digit or death from
peripheral vascular disease), and microvascular
disease (vitreous hemorrhage, retinal photocoag-
ulation, or renal failure).
Statistical Analysis
We performed the analyses according to the inten-
tion-to-treat principle, with descriptive statistics
presented as numbers and percentages or appro-
priate measures of central tendency and disper-
sion. Continuous and categorical study variables
were compared between the conventional-thera-
py group and the intensive-therapy groups with
the use of nonparametric tests. Kaplan–Meier
time-to-event analyses were used for aggregate
clinical outcomes, with log-rank tests used for
differences between previous study-group assign-
ments. Since recruitment was performed during
a 14-year period, patients could have been in the
interventional trial for 6 to 20 years. Since there
was no common time from randomization to the
st art of post-trial monitoring, we used serial haz-
ard-ratio plots with confidence intervals (after
checking that proportional-hazards assumptions
were not violated) to illustrate possible changes
in post-trial relative risks, with P values calcu-
lated only at the end of follow-up. Absolute risk
rates are expressed as the number of events per
1000 person-years.
Post-trial monitoring was initiated by the in-
vestigators and was sponsored for 5 years by the
Medical Research Council and then by the Uni-
versit y of Oxford. The investigators designed and
conducted the study, analyzed the data, and pre-
pared the manuscript, independently of any fund-
ing bodies. The investigators vouch for the com-
pleteness and accuracy of the data.
Results
Patients
A total of 4209 patients were randomly assigned
to receive either conventional therapy or intensive
therapy (Fig. 1). Of these patients, baseline char-
acteristics for the 3277 patients who entered post-
trial monitoring are shown in
Table 1
. In the sul-
fonylurea–insulin group, patients who had been
assigned to receive intensive therapy had lower
levels of mean glycated hemoglobin and fasting
plasma glucose than those who received conven-
tional therapy (P<0.001 for both comparisons)
Triglycerides — mg/dl 0.60 0.97 0.10
Median 127 132 128 127 143 157
Interquartile range 84–184 88–190 88–180 85–182 103–203 107–231
Plasma creatinine — mg/dl 0.87 0.61 0.03
Median 1.00 0.98 1.02 1.02 0.96 1.03
Interquartile range 0.87–1.15 0.89–1.14 0.89–1.17 0.90–1.17 0.83–1.11 0.85–1.23
Ratio of albumin to creatinine§ 0.99 0.42 0.48
Median 12.3 12.7 14.9 14.3 19.9 19.8
Interquartile range 6.2–33.9 5.8–36.7 6.5–49.7 6.8–43.8 8.1–82.8 8.0–61.2
* Plus–minus values are means ±SD. The body-mass index is the weight in kilograms divided by the square of the height in meters. To convert the values for glucose to millimoles per liter,
multiply by 0.05551. To convert the values for cholesterol to millimoles per liter, multiply by 0.02586. To convert the values for triglycerides to millimoles per liter, multiply by 0.01129. To
convert the values for creatinine to micromoles per liter, multiply by 88.4.
† P values were calculated with the use of the chi-square test for categorical variables and the Wilcoxon signed-rank test for continuous variables.
‡ Race or ethnic group was self-reported.
§ Albumin was measured in milligrams, and creatinine was measured in grams.
The New England Journal of Medicine
Downloaded from nejm.org at UQ Library on March 28, 2017. For personal use only. No other uses without permission.
Copyright © 2008 Massachusetts Medical Society. All rights reserved.
