Professional Documents
Culture Documents
Prospective
Diabetes
Study
ukpds
UK Prospective Diabetes Study
multi-centre
randomised controlled trial
of different therapies
of Type 2 diabetes
ukpds
UKPDS : need for a long-term study
ukpds
Co-ordinating Staff
Chief Investigators : Robert Turner, Rury Holman
Statisticians : Irene Stratton, Carole Cull
Ziyah Mehta, Heather McElroy
Modeller : Richard Stevens
Epidemiologists : Andrew Neil, Amanda Adler
Diabetologists : David Matthews, Valeria Frighi
Biochemists : Susan Manley, Iain Ross
Administrators : Philip Bassett, Suzy Oakes
Retinopathy Grading Centre : Eva Kohner, Steve Aldington
Health Economics : Alastair Gray, Maria Raikou
Grant Applications : Ivy Samuel, Caroline Wood
Computing Support : Ian Kennedy, John Veness
And many others
ukpds
Acknowledgements
• patients
• physicians
• nurses
• dietitians
• retinal photographers
ukpds
Major Funding Bodies
ukpds
UK Prospective Diabetes Study
ukpds
Blood Glucose Control Study : Aims
to determine whether
• improved glucose control of Type 2 diabetes
will prevent clinical complications
• therapy with
sulphonylurea - first or second generation
insulin
metformin
has any specific advantage or disadvantage
ukpds
Patient Characteristics
5102 newly diagnosed Type 2 diabetic patients
ukpds
Randomisation
5102 newly diagnosed Type 2 diabetic patients
Diet therapy
14%
ukpds
Randomisation of Treatment Policies
Main Randomisation
n=4209 (82%)
342 allocated to
metformin
3867
Sulphonylurea Insulin
n=1573 n=1156
ukpds
Treatment Policies in 3867 patients
Conventional Policy
n = 1138
• initially with diet alone
• aim for
near normal weight
best fasting plasma glucose < 15 mmol/L
asymptomatic
ukpds
Treatment Policies in 3867 patients
80 pharmacological therapy
60
intensive
40 pharmacological
therapy
20 diet alone
0
1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 7 8 9 10 11 12
ukpds
HbA1c
cross-sectional, median values
9
Conventional
8
HbA1c (%)
Intensive
7
Intensive
5.0
kg
2.5
Conventional
0.0
-2.5
0 3 6 9 12 15
Years from randomisation ukpds
Hypoglycaemic Episodes
ukpds
Hypoglycaemic episodes per annum
Actual Therapy analysis
50 5
40 4
30 3
20 2
10 1
0 0
0 3 6 9 12 15 0 3 6 9 12 15
Years from randomisation
ukpds
Any Diabetes Related Endpoint
1401 of 3867 patients (36%)
First occurrence of any one of:
• diabetes related death
• non fatal myocardial infarction, heart failure or angina
• non fatal stroke
• amputation
• renal failure
• retinal photocoagulation or vitreous haemorrhage
• cataract extraction or blind in one eye
ukpds
Any Diabetes Related Endpoint (cumulative )
1401 of 3867 patients (36%)
60%
Conventional (1138)
% of patients with an event
Intensive (2729)
p=0.029
40%
20%
ukpds
Diabetes Related Deaths (cumulative)
414 of 3867 patients (11%)
30%
Conventional (1138)
% of patients with an event
Intensive (2729)
p=0.34
20%
10%
0%
0 3 6 9 12 15
Years from randomisation ukpds
Microvascular Endpoints (cumulative)
renal failure or death, vitreous haemorrhage or photocoagulation
346 of 3867 patients (9%)
30%
Conventional
% of patients with an event
Intensive
p=0.0099
20%
10%
Intensive
p=0.052
20%
10%
ukpds
Progression of Retinopathy
Two step change in Early Treatment Diabetic Retinopathy Study (ETDRS) scale
Relative Risk
& 99% CI
RR p 0.5 1 2
0 - 3 years 1.03 0.78
0 - 6 years 0.83 0.017
0 - 9 years 0.83 0.012
0 - 12 years 0.79 0.015
Favours Favours
intensive conventional
ukpds
Microalbuminuria
Urine albumin >50 mg/L
Relative Risk
& 99% CI
RR p 0.5 1 2
Baseline 0.89 0.24
Three years 0.83 0.043
Six years 0.88 0.13
Nine years 0.76 0.00062
Twelve years 0.67 0.000054
Fifteen years 0.70 0.033 <
Favours Favours
intensive conventional
ukpds
Glucose Control Study Summary
The intensive glucose control policy maintained a lower
HbA1c by mean 0.9 % over a median follow up of 10 years
from diagnosis of type 2 diabetes with reduction in risk of:
12% for any diabetes related endpoint p=0.029
25% for microvascular endpoints p=0.0099
16% for myocardial infarction p=0.052
24% for cataract extraction p=0.046
21% for retinopathy at twelve years p=0.015
33% for albuminuria at twelve years p=0.000054
ukpds
Conclusion
ukpds
UK Prospective Diabetes Study
Does insulin or
sulphonylurea therapy have
specific advantages or
disadvantages?
ukpds
Sulphonylurea Therapy
advantages
• known to improve glycaemic control
• stimulates endogenous insulin production
disadvantages
• in the heart sulphonylurea mimics ATP
and may prevent vasodilation in ischaemia
• 1st generation agents may increase arrhythmia
ukpds
Insulin Therapy
advantages
• well-used therapy to improve glycaemic control
• may be essential for many patients
disadvantages
• need for injections
• risk of weight gain and hypoglycaemia
• raised insulin levels may promote atherosclerosis
ukpds
Randomisation
3041 patients
in 15 centres
ukpds
HbA1c
cohort, median data
9
Conventional InsulinChlorpropamide Glibenclamide
8
%
6
0
0 2 4 6 8 10
Years from randomisation ukpds
change in weight
cohort, mean data
10.0
Conventional Insulin Chlorpropamide Glibenclamide
7.5
5.0
kg
2.5
0.0
-2.5
0 2 4 6 8 10
Years from randomisation ukpds
Hypoglycaemic episodes per annum
Actual Therapy analysis
50 10
40 8
30 6
20 4
10 2
0 0
0 3 6 9 12 15 0 3 6 9 12 15
Years from randomisation
ukpds
Blood Pressure
cohort, mean data
150
Conventional Insulin Chlorpropamide Glibenclamide
145
140
mmHg
135
85
80
75
70
0 2 4 6 8 10
Years from randomisation ukpds
Any diabetes-related endpoints
0.6
0.3
0.2
0.1
CvGvI
0.0 p = 0.36
0 3 6 9 12 15
Years from randomisation ukpds
Myocardial Infarction
0.4
Conventional (896)
0.2
0.1
CvGvI
0.0 p = 0.66
0 3 6 9 12 15
Years from randomisation ukpds
Progression of Retinopathy : 2 step change
favours favours
RR p 0.2 intensive 1 conventional 5
0-3 rs
y e a p=0 . 9 9
Chlorpro pmida e 1 . 0 02 . 9 2
Glib e lamid
n c e 1 . 0 04 . 8 1
I n lins u 1 . 0 01 . 9 2
0-6 rs
y e a p=0 . 5 8
Chlorpro pmida e 0 . 9 02 . 4 7
Glib e lamid
n c e 0 . 9 04 . 5 9
I n lins u 0 . 8 04 . 0 9 6
0-9 rs
y e a p=0 . 6 5
Chlorpro pmida e 0 . 9 01 . 3 3
Glib e lamid
n c e 0 . 8 03 . 0 6 8
I n lins u 0 . 8 03 . 0 4 8
0-1 2 rs y e a p=0 . 0 0 5 9
Chlorpro pmida e 1 . 0 00 . 9 9
Glib e lamid
n c e 0 . 6 08 . 0 0 4 4
I n lins u 0 . 7 02 . 0 0 4 1
ukpds
Sulphonylurea or Insulin : Summary 1
• all three therapies were similarly effective in
reducing HbA1c
ukpds
Sulphonylurea or insulin : Summary 2
Sulphonylurea therapy
• no evidence of deleterious effect on myocardial
infarction, sudden death or diabetes related deaths
Insulin therapy
• no evidence for more atheroma-related disease
ukpds
UK Prospective Diabetes Study
Does metformin in
overweight diabetic patients
have any advantages or
disadvantages?
ukpds
Introduction
• the UKPDS has shown that an intensive glucose
control policy using sulphonylurea or insulin
therapy is effective in reducing the risk of
complications in both overweight and normal
weight patients
• overweight (>120% Ideal Body Weight) UKPDS
patients could be randomised to an intensive
glucose control policy with metformin instead of
diet, sulphonylurea or insulin
ukpds
Randomisation
Main Randomisation
4209
Non overweight
2505
Overweight
1704
8
HbA 1c (%)
6
0 0
2 4 6 8 10
Years from randomisation ukpds
Change in Weight
overweight patients cohort, mean values
10 Conventional Insulin Chlorpropamide Glibenclamide Metformin
weight change (kg)
-5 0
2 4 6 8 10
Years from randomisation
ukpds
Hypoglycaemic episodes per annum
overweight patients Actual Therapy analysis
any episode major episodes
50 8
Proportion of patients (%)
40
6
30
4
20
2
10
0 0
0 2 4 6 8 10 0 2 4 6 8 10
Years from randomisation
ukpds
Any diabetes related endpoint
overweight
patients 0.6 Conventional (411)
Proportion of patients with events Intensive (951)
Metformin (342)
0.4 M v C
p=0.0023
0.2
MvI
p=0.0034
0.0
0 3 6 9 12 15 ukpds
Years from randomisation
Diabetes related deaths
overweight
0.4
patients Conventional (411)
Proportion of patients with events Intensive (951)
0.3 Metformin (342)
Mv C
p=0.017
0.2
0.1
MvI
0.0 p=0.11
0 3 6 9 12 15
ukpds
Years from randomisation
Myocardial Infarction
overweight 0.4 Conventional (411)
Proportion of patients with events
patients
Intensive (951)
0.3 Metformin (342)
MvC
p=0.010
0.2
0.1
MvI
p=0.12
0.0
0 3 6 9 12 15
Years from randomisation ukpds
Microvascular endpoints
overweight 0.3 Conventional (411)
Proportion of patients with events
patients
Intensive (951)
Metformin (342)
0.2 M v C
p=0.19
0.1
MvI
p=0.39
0.0
0 3 6 9 12 15
Years from randomisation ukpds
Metformin Comparisons
overweight patients RR (95% CI)
RR p 0.2 1 5
Any dia b e tse re l ate d e n dpo i n t
M e tor f min 0 . 68 0 . 002 3
favours favours
metformin conventional
ukpds
Metformin Comparisons
overweight patients RR (95% CI)
M v Int RR p 0.2 1 5
Any dia b e tse re l ate d e n dpo i n tp=0 . 0 0 3 4
M e tor f min 0 . 68 0 . 002 3
I n tnesiv e 0 . 93 0 . 46
Diab e t se re l ate d d e ahst p=0 . 1 1
M e tor f min 0 . 58 0 . 017
I n tnesiv e 0 . 80 0 . 19
All ca use m ort ali ty p=0 . 0 2 1
M e tor f min 0 . 64 0 . 011
I n tnesiv e 0 . 92 0 . 49
Myo c ardia l in f arc tio n p=0 . 1 2
M e tor f min 0 . 61 0 . 01
I n tnesiv e 0 . 79 0 . 11
favours favours
metformin or conventional
intensive
ukpds
Sulphonylurea plus Metformin
• patients primarily randomised to intensive therapy with
sulphonylurea were not given additional metformin until
their fpg was >15 mmol/L or they developed
hyperglycaemic symptoms
ukpds
Aim
• the aim of this secondary randomisation was to assess
the degree to which glycaemic control might be
improved by early combination therapy with metformin
ukpds
Aggregate Endpoints
Relative Risk
& 95% CI
Median follow up 6.6 years RR p 0.1 1 10
Any diabetes related endpoint 1.04 0.78
Diabetes related deaths * 1.96 0.039
All cause mortality 1.60 0.041
ukpds
Metformin : Summary
ukpds
UK Prospective Diabetes Study
Blood Pressure
Control Study
ukpds
Blood Pressure Control Study : Aims
to determine whether
ukpds
Inclusion criteria
patients NOT on anti-hypertensive therapy
systolic >160 and/or diastolic > 90 mmHg
excluded if:
required strict blood pressure control; severe illness;
contraindication to study medication or declined
informed consent
ukpds
Patient Characteristics
1148 Type 2 diabetic patients
age 56 years
gender male / female 55% / 45%
ethnic groups Caucasian 87%
Asian 6%
Afro-caribbean 7%
Body Mass Index 29 kg/m2
HbA1c 6.8 %
systolic / diastolic blood pressure 160 / 94 mmHg
urine albumin > 50 mg/l 18%
ukpds
Randomisation
randomisation
ukpds
Blood Pressure : Tight vs Less Tight Control
cohort, median values
180 Less tight control Tight control
160
140
mmHg
100
80
60
0 2 4 6 8
Years from randomisation
ukpds
Mean Blood Pressure
ukpds
Therapy requirement
number of antihypertensive agents
None one two > two
80
60
40
20
0
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8
Years from randomisation
ukpds
Any diabetes-related endpoints
50% Less tight blood pressure control (390)
30%
20%
10%
risk reduction
0%
24% p=0.0046
0 3 6 9
Years from randomisation ukpds
Diabetes-related deaths
20% Less tight blood pressure control (390)
Tight blood pressure control (758)
% of patients with events
15%
10%
5%
risk reduction
0%
32% p=0.019
0 3 6 9
Years from randomisation ukpds
Myocardial Infarction
25% Less Tight Blood Pressure Control (390)
Tight Blood Pressure Control (758)
% of patients with event 20%
15%
10%
5%
risk reduction
0%
21% p=0.13
0 3 6 9
Years from randomisation ukpds
Stroke
25%
Less Tight Blood Pressure Control (390)
Tight Blood Pressure Control (758)
20%
% patients with event
15%
10%
5%
risk reduction
0% 44% p=0.013
0 3 6 9
Years from randomisation ukpds
Microvascular endpoints
25%
Less Tight Blood Pressure Control (390)
15%
10%
5%
risk reduction
0%
37% p=0.0092
0 3 6 9
Years from randomisation ukpds
Heart Failure
25%
Less Tight Blood Pressure Control (390)
15%
risk reduction
56% p=0.0043
10%
5%
0%
0 3 6 9
Years from randomisation ukpds
Progression of Retinopathy : 2 step change
p=0.38 p=0.019 p=0.004
60
51
40 37
% patients
34
28
23
20
20
0
243 461 207 411 152 300
3 years 6 years 9 years
Years from randomisation
numbers above bars are % affected
ukpds
Deterioration of Vision : 3 lines on ETDRS chart
p=0.40 p=0.47 p=0.004
30
19
% patients
20
10
10 9
7 8
5
0
293 575 257 523 180 332
3 years 6 years 9 years
10
0
317 618 274 543 166 299
3 years 6 years 9 years
Do ACE inhibitors or
Beta Blockers
have any specific advantages
or disadvantages?
ukpds
Blood Pressure : ACE inhibitor vs Beta blocker
cohort, median values
180
Less tight control ACE inhibitor Beta blocker
160
140
mm Hg
100
80
60
0 2 4 6 8
Years from randomisation
ukpds
Reasons for non-compliance
Captopri l Atenolo l
(n=400 ) (n=358 ) p
non- compl iant 88 (22%) 125 (35%) <0.0001
ukpds
Any Diabetes Related Endpoint (cumulative)
429 of 1148 patients (37%)
20%
10%
5%
Beta blocker
15% ACE inhibitor
Less tight vs Tight
p=0.0092
10%
5%
ukpds
Surrogate endpoints
Relative Risk & 99% CI
RR p 0.1 1 10
Reti nopathy 2 step progress ion
median 1.5 years 0.99 0.75
median 4.5 years 0.99 0.82
median 7.5 years 0.91 0.28
Ur ine albumi n > 50 mg/L
3 year s 1.11 0.55
6 year s 0.93 0.65
9 year s 1.20 0.31
Ur ine albumi n > 300 mg/L
3 year s 1.41 0.44
6 year s 0.75 0.43
9 year s 0.48 0.090
favours ACE favours Beta
inhibitor blocker
ukpds
Conclusion
ACE inhibitors and Beta blockers were equally
effective in lowering mean blood pressure in
hypertensive patients with type 2 diabetes and in
reducing the risk of:
ukpds
UK Prospective Diabetes Study
Potential implications
for clinical care of
diabetic patients
ukpds
UK Prospective Diabetes Study
An intensive glucose control policy HbA1c 7.0 % vs 7.9 %
reduces risk of
any diabetes-related endpoints 12% p=0.030
microvascular endpoints 25% p=0.010
myocardial infarction 16% p=0.052
ukpds
Choice of Therapies
diabetes :
• each of the available therapies studied can be used
• in overweight, diet-treated patients, metformin may
be advantageous
hypertension :
• Beta blockers and ACE inhibitors each provide
protection
ukpds
Which goals of therapy?
• current guidelines suggest HbA1c <7%
ukpds
Which goals of therapy?
• current guidelines suggest blood pressure
<140 / 85 mmHg or <130 / 85 mmHg
ukpds
Polypharmacy
• glycaemia
combinations of agents with different actions
will be needed
more patients will require insulin
• blood pressure
many patients will need 3 or more different
types of agents
ukpds
Differences between Therapies
• sulphonylurea, insulin and metformin are each
effective in reducing the risk of any diabetes related
endpoints and microvascular endpoints
ukpds