Download No Slide Title

A randomised controlled trial of the prevention
of CHD and other vascular events by BP and
cholesterol lowering in a factorial study design
B.Dahlof (Co-chair), P.Sever (Co-chair), N. Poulter (Secretary)
H. Wedel (Statistician), G. Beevers, M. Caulfield, R. Collins
S. Kjeldsen, A. Kristinsson, J. Mehlsen, G. McInnes, M. Nieminen
E. O’Brien, J. Östergren, on behalf of the ASCOT Investigators
ASCOT- BPLA: Rationale
• Insufficient outcome data on newer
types of BP-lowering agents, especially
in specific combination treatment
regimens
• Less-than-expected CHD prevention
using standard therapy
ASCOT- BPLA
Primary Objective
To compare the effect on non-fatal
myocardial infarction (MI) and fatal
CHD of the standard antihypertensive
regimen (-blocker ± diuretic) with a more
contemporary regimen
(CCB ± ACE inhibitor)
Additional objectives include:
Secondary end points
• Total stroke
• All coronary events
• Primary end point minus
silent MI
• Total cardiovascular
(CV) events and
procedures
• CV mortality
• All-cause mortality
• Heart failure
Tertiary end points
• Development of diabetes
• Impairment of renal function
• Pre-specified end points in
pre-specified subgroups
• Life-threatening arrhythmias
Other objectives
• Interaction between statins
and antihypertensive
treatment
• Health economic analyses
Study design
19,257
hypertensive
patients
atenolol ±
bendroflumethiazide
ASCOT-BPLA
PROBE
design
amlodipine ±
perindopril
10,305 patients
TC ≤ 6.5 mmol/L (250 mg/dL)
atorvastatin 10 mg
Double-blind
ASCOT-LLA
placebo
Investigator-led, multinational
randomised controlled trial
Treatment algorithm to BP targets < 140/90 mm Hg
or < 130/80 mm Hg in patients with diabetes
amlodipine 5-10 mg
atenolol 50-100 mg
add
add
bendroflumethiazide-K
1.25-2.5 mg
perindopril 4-8 mg
add
doxazosin GITS 4-8 mg
add
additional drugs, eg, moxonidine/spironolactone
Patient inclusion criteria
• Screening and baseline BP
–  160/100 mm Hg untreated
–  140/90 mm Hg following treatment with 1 or more drugs
• Age 40-79 years
• No previous MI or current clinical CHD
• 3 or more CV risk factors
Study power and recruitment
(Feb 1998-May 2000)
1567
3984
2226
2382
POWER: PRIMARY END POINT
Relative additional benefit (ITT)
Significance level
9098
Power
Estimated sample size
Persons with events
TOTAL RECRUITED 19,257
16%
5%
80%
18,000
1150
Baseline characteristics
amlodipine  perindopril atenolol  thiazide
Demographics and clinical characteristics
Woman
White
Current smoker
Age (years)
SBP (mm Hg)
DBP (mm Hg)
Heart rate (bpm)
BMI (kg/m2)
n = 9639
2258 (23.4%)
9187 (95.3%)
3168 (32.9%)
63.0 (8.5)
164.1 (18.1)
94.8 (10.4)
71.9 (12.7)
28.7 (4.6)
n = 9618
2257 (23.5%)
9170 (95.3%)
3110 (32.3%)
63.0 (8.5)
163.9 (18.0)
94.5 (10.4)
71.8 (12.6)
28.7 (4.5)
Drug therapy
Previous antihypertensive treatments
0
1
≥2
Lipid-lowering therapy
Aspirin
1841 (19.1%)
4280 (44.4%)
3518 (36.5%)
1046 (10.9%)
1851 (19.2%)
1825 (19.0%)
4283 (44.5%)
3510 (36.5%)
1004 (10.4%)
1837 (19.1%)
Values are number of patients, (%) or mean (SD)
Data safety monitoring board (DSMB)
In October 2004 the DSMB recommended that the BP
arm of ASCOT should be stopped on account of
concerns that those patients receiving atenolol  thiazide
would continue to be disadvantaged compared with the
comparator group
The Steering Committee endorsed the recommendation
of the DSMB, and trial closure began Dec, 2004 and
ended June 2005.
Statistical methods
• Based on an intention-to-treat analysis
• Time to first primary event
• Log-rank procedure and Cox’s Proportional
Hazards were used to calculate confidence
intervals
• Cumulative incidence curves were generated
using the Kaplan-Meier method
Sever PS, Dahlöf B, Poulter N, Wedel H, et al, for the ASCOT Investigators. Lancet. 2003;361:1149-58
ASCOT patient population
risk factor profile
All patients in ASCOT have hypertension plus ≥ 3 risk factors for CHD
100
Hypertension
Age ≥ 55 years
84
Male
77
Microalbuminuria/proteinuria
61
Smoker
30
Family history of CHD
27
Plasma TC:HDL-C ≥ 6
24
Type 2 diabetes
24
Certain ECG abnormalities
14
LVH
13
Previous cerebrovascular events
11
Peripheral vascular disease
6
0
10
20
30
40
50
60
70
Patients with risk factor (%)
80
90
100
trial design
19,342 randomised to
antihypertensive therapy
85 excluded before end of
study due to irregularities
19,257 randomised
9,639 assigned and
received amlodipine 
perindopril
121 with incomplete
information
• 81 alive at last visit
• 24 withdrawn consent
• 16 lost to follow-up
• 9639 assessed for
primary endpoint on
intention-to-treat basis
• 9,518 with complete
information (8,780 alive
738 dead)
9,618 assigned and
received atenolol  thiazide
171 with incomplete
information
• 102 alive at last visit
• 36 withdrawn consent
• 33 lost to follow-up
• 9618 assessed for
primary endpoint on an
intention-to-treat basis
• 9,447 with complete
information ( 8,627 alive
820 dead)
Systolic and diastolic
blood pressure
atenolol  thiazide
amlodipine  perindopril
180
160
164.1 SBP
163.9
Mean difference 2.7
137.7
mm Hg
140
136.1
120
100
94.8
DBP
Mean difference 1.9
94.5
79.2
80
77.4
60
Baseline
0.5
1
1.5
2
2.5
3
Time (years)
3.5
4
4.5
5
5.5
Last
visit
Mean proportion of time on antihypertensive
medication by treatment group
Year 1
All Study
Amlodipine
88.2
82.5
Perindopril
46.2
58.5
Amlodipine + perindopril
39.1
49.5
Atenolol
87.4
79.4
Bendroflumethiazide
56.6
65.7
Atenolol + bendroflumethiazide
49.1
54.9
Randomised to Amlodipine
Randomised to atenolol
Primary end point:
Non-fatal MI, fatal CHD
%
5.0
Atenolol  thiazide
(No. of events =474)
4.0
Amlodipine  perindopril
(No. of events = 429)
3.0
2.0
HR = 0.90 (0.791.02)
p = 0.1052
1.0
Years
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
9639
9618
1.0
2.0
3.0
9475
9470
9337
9290
9168
9083
4.0
8966
8858
5.0
7863
7743
Endpoints: Non-fatal MI
(excl silent) + fatal CHD
%
5.0
4.0
Atenolol  thiazide
(No. of events 444)
3.0
Amlodipine  perindopril
(No. of events 390)
2.0
HR = 0.87 (0.761.00)
p = 0.0458
1.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
9639
9618
1.0
9485
9475
2.0
3.0
9354
9302
9193
9099
4.0
8998
8881
5.0
7895
7768
Years
Primary end point + coronary
revascularisation procedures
%
8.0
Atenolol  thiazide
(No. of events 688)
7.0
6.0
5.0
Amlodipine  perindopril
(No. of events 596)
4.0
3.0
2.0
HR = 0.86 (0.770.96)
p = 0.0058
1.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9458
9447
9288
9236
9086
8986
8857
8719
7732
7590
Years
Total coronary end point
%
10.0
Atenolol  thiazide
(No. of events 852)
8.0
6.0
Amlodipine  perindopril
(No. of events 753)
4.0
2.0
HR = 0.87 (0.790.96)
p = 0.0070
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
9639
9618
9400
9373
9204
9136
8984
8864
8744
8591
5.0
7614
7470
Years
Fatal and non-fatal stroke
%
5.0
Atenolol  thiazide
(No. of events 422)
4.0
3.0
Amlodipine  perindopril
(No. of events 327)
2.0
HR = 0.77 (0.660.89)
p = 0.0003
1.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
9639
9618
9483
9461
9331
9274
9156
9059
8972
8843
5.0
7863
7720
Years
Total CV events and procedures
%
18.0
Atenolol  thiazide
(No. of events 1602)
16.0
14.0
12.0
Amlodipine  perindopril
(No. of events 1362)
10.0
8.0
6.0
4.0
HR = 0.84 (0.780.90)
p < 0.0001
2.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9277
9210
8957
8848
8646
8465
8353
8121
7207
6977
Years
CV mortality
%
3.5
Atenolol  thiazide
(No. of events 342)
3.0
2.5
2.0
Amlodipine  perindopril
(No. of events 263)
1.5
1.0
HR = 0.76 (0.650.90)
p = 0.0010
0.5
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9544
9532
9441
9415
9322
9261
9167
9085
8078
7975
Years
All-cause mortality
%
10.0
Atenolol  thiazide
(No. of events 820)
8.0
6.0
Amlodipine  perindopril
(No. of events 738)
4.0
2.0
HR = 0.89 (0.810.99)
p = 0.0247
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9544
9532
9441
9415
9332
9261
9167
9085
8078
7975
Years
Fatal and non-fatal heart failure
%
1.8
Atenolol  thiazide
(No. of events 159)
1.6
1.4
1.2
Amlodipine perindopril
(No. of events 134)
1.0
0.8
0.6
0.4
HR = 0.84 (0.661.05)
p = 0.1257
0.2
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
9639
9618
9524
9501
9409
9369
9275
9195
9101
9011
5.0
8004
7901
Years
Unstable angina
%
1.2
Atenolol  thiazide
(No. of events 106)
1.0
0.8
Amlodipine  perindopril
(No. of events 73)
0.6
0.4
HR = 0.68 (0.510.92)
p = 0.0115
0.2
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9536
9510
9416
9374
9285
9198
9123
9007
8021
7888
Years
Chronic stable angina
%
2.5
Atenolol  thiazide
(No. of events = 208)
2.0
Amlodipine  perindopril
(No. of events = 205)
1.5
1.0
0.5
HR = 0.98 (0.811.19)
p = 0.8323
0.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
9639
9618
1.0
2.0
3.0
9493
9482
9350
9327
9198
9135
4.0
5.0
9017
8924
7917
7817
Years
Peripheral arterial disease
%
2.5
2.0
Atenolol  thiazide
(No. of events = 202)
1.5
Amlodipine  perindopril
(No. of events = 133)
1.0
0.5
HR = 0.65 (0.520.81)
p = 0.0001
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
9639
9618
1.0
9523
9495
2.0
3.0
9382
9348
9237
9163
4.0
9070
8958
5.0
7958
7828
Years
Life-threatening arrhythmias
%
0.3
Atenolol  thiazide
(No. of events = 25)
0.2
Amlodipine  perindopril
(No. of events = 27)
0.1
HR = 1.07 (0.62 1.85)
p = 0.8009
0.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
9639
9618
1.0
2.0
9539
9527
9433
9405
3.0
9310
9243
4.0
5.0
9153
9069
8060
7961
Years
New-onset renal impairment
%
5.0
Atenolol  thiazide
(No. of events = 469)
4.0
Amlodipine  perindopril
(No. of events = 403)
3.0
2.0
HR = 0.85 (0.750.97)
p = 0.0187
1.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9426
9431
9277
9247
9093
9021
8877
8782
7775
7640
Years
New-onset diabetes mellitus
%
10.0
Atenolol  thiazide
(No. of events = 799)
8.0
6.0
Amlodipine  perindopril
(No. of events = 567)
4.0
HR = 0.70 (0.630.78)
p < 0.0001
2.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9383
9295
9165
9014
8966
8735
8726
8455
7618
7319
Years
CV death + MI + stroke
%
10.0
Atenolol  thiazide
(No. of events = 937)
8.0
6.0
Amlodipine  perindopril
(No. of events = 796)
4.0
HR = 0.840 (0.760.92)
p < 0.0003
2.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9415
9400
9228
9152
9007
8891
8778
8629
7655
7500
Years
Summary of all end points
Primary
Non-fatal MI (incl silent) + fatal CHD
Unadjusted Hazard
ratio (95% CI)
0.90 (0.79-1.02)
Secondary
Non-fatal MI (exc. Silent) +fatal CHD
Total coronary end point
Total CV event and procedures
All-cause mortality
Cardiovascular mortality
Fatal and non-fatal stroke
Fatal and non-fatal heart failure
0.87 (0.76-1.00)
0.87 (0.79-0.96)
0.84 (0.78-0.90)
0.89 (0.81-0.99)
0.76 (0.65-0.90)
0.77 (0.66-0.89)
0.84 (0.66-1.05)
Tertiary
Silent MI
Unstable angina
Chronic stable angina
Peripheral arterial disease
Life-threatening arrhythmias
New-onset diabetes mellitus
New-onset renal impairment
1.27 (0.80-2.00)
0.68 (0.51-0.92)
0.98 (0.81-1.19)
0.65 (0.52-0.81)
1.07 (0.62-1.85)
0.70 (0.63-.078)
0.85 (0.75-0.97)
Post hoc
Primary end point + coronary revasc procs
CV death + MI + stroke
0.86 (0.77-0.96)
0.84 (0.76-0.92)
0.50
0.70
1.00
Amlodipine  perindopril better
1.45
2.00
Atenolol  thiazide better
The area of the blue square is proportional to the amount of statistical information
Total CV events and procedures among subgroups
Diabetes
No diabetes
Current smoker
Non-current smoker
Obese
Non-obese
Older (>60 years)
Younger (≤60 years)
Female
Male
LVH according to ECG or ECHO
No LVH according to ECG or ECHO
Previous vascular disease
No previous vascular disease
Renal dysfunction
No renal dysfunction
With metabolic syndrome
Without metabolic syndrome
All patients
0.60 0.70 0.80 0.90 1.00
1.50
Amlodipine  perindopril better Atenolol  thiazide better
The area of the black square is proportional to the amount of statistical information
Total CV events and procedures among subgroups
p value
Diabetes
No diabetes
Heterogeneity p
0.0283
<0.0001
0.5205
0.0001
0.0030
0.1138
Obese
Non-obese
0.0162
<0.0001
0.6753
Older (>60 years)
Younger (≤60 years)
<0.0001
0.0227
0.7816
0.0015
0.0001
0.2889
0.0056
<0.0001
0.6364
0.0019
0.0001
0.4863
<0.0001
0.0055
0.7130
0.0015
0.0002
0.9417
Current smoker
Non-current smoker
Female
Male
LVH according to ECG or ECHO
No LVH according to ECG or ECHO
Previous vascular disease
No previous vascular disease
Renal dysfunction
No renal dysfunction
With metabolic syndrome
Without metabolic syndrome
All patients
<0.0001
0.60 0.70 0.80 0.90 1.00
1.50
Amlodipine  perindopril better Atenolol  thiazide better
The area of the black square is proportional to the amount of statistical information
Adverse events leading to treatment
discontinuation
Adverse event
Total
Serious
* p<0.0001
Amlodipine  perindopril
(%)
Atenolol  thiazide
(%)
2358
(24.5)
2402
(25.0)
162
(1.7)
254
(2.6)*
Adverse events
Adverse event*
Bradycardia
Chest pain
Amlodipine 
perindopril
n
(%)
34
(0.4)
Atenolol  thiazide
n
(%)
p-value
536
(6)
<0.0001
740
(8)
849
(9)
0.0040
1859
(19)
782
(8)
<0.0001
Diarrhoea
377
(4)
548
(6)
<0.0001
Dizziness
1183
(12)
1555
(16)
<0.0001
Dyspnoea
599
(6)
987
(10)
<0.0001
Eczema
493
(5)
383
(4)
0.0002
Erectile dysfunction
556
(6)
707
(7)
<0.0001
Fatigue
782
(8)
1556
(16)
<0.0001
1371
(14)
308
(3)
<0.0001
202
(2)
525
(6)
<0.0001
Oedema peripheral
2188
(23)
588
(6)
<0.0001
Peripheral coldness
81
(1)
579
(6)
<0.0001
642
(7)
745
(8)
0.0039
Cough
Joint swelling
Lethargy
Vertigo
* Adverse events with incidence >5% and difference of more than 1%
ASCOT Committees
Executive and Writing Committee
B Dahlöf, Co-Chairman, Gothenburg, P Sever, Co-Chairman, London, N Poulter, Secretary, London,
H Wedel, Statistician, Gothenburg.
Steering Committee
A Adderkin, London, DG Beevers, Birmingham, J Buch, New York (non-voting), M Caulfield, London, R
Collins, Oxford, B Dahlöf, Gothenburg, A Jarl, Stockholm (non-voting), SE Kjeldsen, Oslo, A Kristinsson,
Reykjavik, J Mehlsen, Copenhagen, G McInnes, Glasgow, M Nieminen, Helsinki, N Poulter, London, E
O'Brien, Dublin, P Sever, London, H Wedel, Gothenburg, J Östergren, Stockholm, Servier representative,
Paris (non-voting).
Working Group
A Adderkin, London, J Buch, New York, S Cavanaugh (up to 2003), New York, R Chamberlain, New York, B
Dahlöf, Gothenburg, S Gee, London, A Holmner, Gothenburg, A Jarl, Stockholm, N Poulter, London, P Sever,
London, H Wedel, Gothenburg.
Data Safety Monitoring Board
J Cohn, Minneapolis, L Erhardt, Malmö, K Fox, London, A Oden, Gothenburg, S Pocock, London,
J Tuomilehto, Helsinki.
Endpoint Committee
U Dahlström, Linköping, F Fyhrquist, Helsinki, H Hemingway, London, K Midtbo, Oslo.
Substudy Committee
M Caulfield, London, B Dahlöf, Gothenburg, T Kahan, Stockholm, J Mehlsen, Copenhagen, M Nieminen,
Helsinki, E O'Brien, Dublin, I Os, Oslo, N Poulter, London, P Sever, London, S Thom, London.
Possible explanations for the observed
differences in outcomes
• Better BP lowering with amlodipine  perindopril
• Non-BP-lowering benefits of amlodipine  perindopril
• Non-BP-related disadvantages of atenolol  thiazide
• Adverse interaction between atenolol  thiazide and
statin
• Beneficial interaction between amlodipine  perindopril
and statin
Variables which differed significantly (baseline final visit) between treatment regimens
Mean differences Changes baseline
to final visit
(Amlodipine  perindopril
- Atenolol  thiazide)
p-value
Systolic BP (mm Hg)
-1.78
<0.0001
Diastolic BP (mm Hg)
-2.05
<0.0001
Heart rate (bpm)
11.12
<0.0001
Weight (kg)
-0.79
<0.0001
HDL-cholesterol (mmol/L)
0.11
<0.0001
Triglycerides (mmol/L)
-0.23
<0.0001
Glucose (mmol/L)
-0.20
<0.0001
Creatinine (µmol/L)
-5.06
<0.0001
Potassium (mmol/L)
0.05
<0.0001
Endpoints evaluated
1. Primary endpoint + coronary revascularisation
(coronary events)
2. Non-fatal and fatal stroke
Rationale
•
Significantly different rates
•
Potentially different mechanisms
•
Sufficient power
The role of BP differences?
Methods
•
Temporal association
•
Serial mean matching
•
Updated Cox regression adjustment
Which BP measure?
•
SBP
•
DBP
•
MBP - [(SBP + DBP) / 2]
•
Pulse pressure
When?
•
Latest
•
One year recurrent average
•
Accumulated mean
Differences in coronary event rates and
in BP over time
Time interval
Accumulated differences (mmHg)
Amlodipine 
perindopril better
Atenolol  thiazide - Amlodipine  perindopril
0-6 months
0.5 - 1 year
SBP
4.95
4.03
DBP
1.72
1.99
1 - 2 years
2.62
1.89
2 - 3 years
3 - 4 years
4 - 5 years
> 5 years
All study
2.02
1.85
1.62
1.55
2.76
1.76
1.88
1.88
1.78
1.91
0.25
Time interval
0.50
Accumulated differences (mmHg)
Atenolol  thiazide - Amlodipine  perindopril
SBP
DBP
0-6 months
4.95
1.72
0 - 1 year
0 - 2 years
4.51
3.67
1.86
1.89
0 - 3 years
3.24
1.87
0 - 4 years
0 - 5 years
All study
3.01
2.85
2.76
1.89
1.91
1.91
1.45
1.00
0.70
Hazard ratio
Amlodipine 
perindopril better
0.25
0.50
Atenolol  thiazide
better
2.00
2.85
Atenolol  thiazide
better
1.45
0.70
1.00
Hazard ratio
2.00
2.85
Differences in stroke event rates
and in BP over time
Time interval
0-6 months
0.5 - 1 year
1 - 2 years
2 - 3 years
3 - 4 years
4 - 5 years
> 5 years
All study
Accumulated differences (mmHg)
Amlodipine 
perindopril better
Atenolol  thiazide - Amlodipine  perindopril
SBP
4.95
4.03
2.62
2.02
1.85
1.62
1.55
2.76
DBP
1.72
1.99
1.89
1.76
1.88
1.88
1.77
1.91
0.25
Time interval
0.50
Atenolol  thiazide
better
1.45
1.00
0.70
Hazard ratio
2.00
2.85
Accumulated differences (mmHg)
Atenolol  thiazide - Amlodipine  perindopril
SBP
DBP
0-6 months
4.95
1.72
0 - 1 year
0 - 2 years
0 - 3 years
0 - 4 years
0 - 5 years
All study
4.51
3.67
3.24
3.01
2.85
2.76
1.86
1.89
1.87
1.89
1.91
1.91
Amlodipine 
perindopril better
0.25
0.50
Atenolol  thiazide
better
1.45
0.70
1.00
Hazard ratio
2.00
2.85
Serial mean matching: Methods
• Take Amlodipine ± perindopril cohort at each of five time points plus baseline
%
Mean
Amlodipine ± perindopril
(all patients)
Atenolol ± thiazide matched subset
NB: n> 7500 (86%)
SBP
• Select “atenolol ± thiazide” person closest to mean after randomisation
• Add others sequentially to maintain group SBP matching
• Maximum mean SBP group difference = 0.02 mm Hg
• 9% of coronary events and 14% of stroke events excluded
• Multiple Cox regression - adjusted HR in 6 periods
• Further adjustment for age and number of risk factors
Serial mean matching: Results
Hazard ratio
unadjusted
p-value
Pooled hazard ratio
for adjusted
SMM adjusted
HR
Primary endpoint +
Coronary
revascularisation
0.86 (0.77 - 0.96)
0.87 (0.78 - 0.98)
0.0177
Fatal and non-fatal
stroke
0.77 (0.66 - 0.89)
0.83 (0.71 - 0.96)
0.0147
Hazard ratios for treatment effect on coronary
events adjusted for accumulated mean levels of
variables that differed
Unadjusted
Systolic BP
Diastolic BP
Mean BP*
Pulse pressure
Heart rate
Glucose
HDL cholesterol
Triglycerides
Creatinine
Potassium
Weight
* Mean BP = (SBP/DBP)/2
Coronary events
HR
p-value
0.86
0.0058
0.88
0.0258
0.86
0.0065
0.88
0.0205
0.87
0.0170
0.85
0.0201
0.85
0.0041
0.90
0.0610
0.85
0.0043
0.86
0.0091
0.85
0.0045
0.86
0.0053
Hazard ratios for treatment effect on stroke
events adjusted for accumulated mean levels of
variables that differed
Unadjusted
Systolic BP
Diastolic BP
Mean BP*
Pulse pressure
Heart rate
Glucose
HDL cholesterol
Triglycerides
Creatinine
Potassium
Weight
* Mean BP = (SBP/DBP)/2
Fatal and non-fatal stroke
HR
p-value
0.77
0.0003
0.83
0.0144
0.80
0.0033
0.84
0.0170
0.80
0.0026
0.74
0.0007
0.78
0.0007
0.76
0.0002
0.78
0.0008
0.79
0.0014
0.76
0.0002
0.76
0.0002
Impact on the treatment effect on
coronary events after adjustment for BP
and all variables that differed
Hazard
Hazard
ratio
ratio
95%
95% CI
CI
Unadjusted
0.86
0.77 - 0.96
0.0058
SBP
SBP
SBP
0.88
0.88
0.79
0.79 -- 0.98
0.98
0.0258
SBP + covariates
0.93
0.81 - 1.07
0.3276
SBP + DBP + covariates
0.92
0.80 - 1.06
0.2744
MBP** ++ covariates
covariates
MBP**
PP
PP +
+ covariates
covariates
0.94
0.91
0.91
0.81 - 1.08
0.79 -- 1.04
1.04
0.79
0.3519
Amlodipine 
perindopril better
p-value
0.1791
0.50
** MBP = (SBP+DBP)/2
Atenolol 
thiazide better
0.70
1.00
Hazard ratio
1.45
Impact on the treatment effect on stroke
events after adjustment for BP and all
variables that differed
Hazard
Hazard
ratio
ratio
95% CI
95% CI
Unadjusted
Unadjusted
0.77
0.66 - 0.89
0.0003
Mean
Mean BP
BP
0.84
0.72 - 0.97
0.0170
SBP
SBP ++ covariates
covariates
0.85
0.71 - 1.02
0.0836
SBP
SBP ++ DBP
DBP ++ covariates
covariates
0.87
0.73 - 1.05
0.1386
MBP**
MBP**++ covariates
covariates
0.87
0.73 - 1.05
0.1380
PP ++ covariates
covariates
PP
0.80
0.67 - 0.96
0.0164
Amlodipine 
perindopril better
0.50
** MBP = (SBP+DBP)/2
Atenolol 
thiazide better
0.70
1.00
Hazard ratio
1.45
p-value
Summary
•
Several potentially important variables including BP
differed post-randomisation between 2 treatment groups
•
No temporal association between BP differences and
event rate differences
•
SMM for SBP and updated Cox regression analyses
suggested BP accounted for about 15% of coronary
differences and 30% of stroke differences
•
Full multivariate adjustment accounted for about 50% of
coronary differences (mainly HDL-C differences) and 40%
of strokes
•
Residual differences are large for stroke but nonsignificant for coronary and stroke events
Conclusions
•
These analyses are compatible with the possibility
that CV event differences were explained by the
variables considered
•
BP differences unlikely single explanation
•
Residual differences, albeit non-significant, are large
especially for stroke
•
ASCOT provides implications for optimal CV
prevention independent of these analyses
ASCOT: BPLA and LLA combined: Insight
into optimal CV prevention (2)
Rates / 1000 patient years
Amlodipine 
perindopril +
statin
Atenolol 
thiazide +
placebo
Relative risk
reduction
Fatal MI and non-fatal CHD
4.8
9.2
48%
Fatal and non-fatal stroke
4.6
8.2
44%
Endpoint
Final conclusions
• Amlodipine  perindopril based therapy confers
an advantage over atenolol  thiazide based
therapy on all major CV end points, all-cause
mortality and new-onset diabetes
• Irrespective of the reasons for benefit, the standard
regimen of beta-blocker  thiazide should not be
preferred to the amlodipine  perindopril regimen for
most patients
• Compared with standard antihypertensive therapy
without statin therapy, the amlodipine  perindopril
regimen plus atorvastatin reduced coronary and
stroke events by almost 50%
AHA 2006
Results CAFE study, a substudy of ASCOT
Central vs brachial bloodpressure
•Are antihypertensive drugs equal in lowering central
bloodpressure?
• What determines central bloodpressure ?
• What are the consequences of lower central bloodpressure?
CAFE study: Rationale
• Brachial blood pressure is a strong predictor of clinical
outcomes in people with hypertension
• It is assumed that brachial blood pressure accurately
reflects pressures in the central aorta and thus left
ventricular load
• This assumption may not be valid in all circumstances
because different classes of blood pressure-lowering drugs
may differentially influence central aortic blood pressures
• In clinical trials comparing different blood pressure loweringdrugs, clinical outcomes could be influenced by drug effects
on central aortic pressures, despite similar effects on
brachial blood pressure
CAFE : Background (1)
• In clinical trials comparing different blood pressure
lowering-drugs, clinical outcomes could be influenced by
drug effects on central aortic pressures, despite similar
effects on brachial blood pressure
• This hypothesis required testing in a prospective clinical
outcomes trial evaluating 2 different blood pressurelowering treatment regimens
CAFE : Background (2)
• The Conduit Artery Functional Evaluation
(CAFE) study was designed to test this
hypothesis as a major substudy of the AngloScandinavian Cardiac Outcomes Trial
(ASCOT)
CAFE Hypothesis
Primary Objective
• The different blood pressure-lowering regimens in
ASCOT (atenolol ± thiazide versus amlodipine ±
perindopril) would produce different effects on central
aortic pressures and hemodynamics despite similar
effects on brachial blood pressure
Secondary Objective
• Central aortic pressures would be an important
determinant of clinical outcomes in ASCOT
Williams B. et al. J Hum Hypertens. 2001; 15 (suppl 1): S69-S73
Methods
• Seated brachial blood pressure measured
using Omron 705CP
• Pulse wave analysis 10 second sampling of
radial artery pulse waves were recorded by
tonometry
• The radial pulse wave was calibrated to
brachial blood pressure and transformed to
derive a central aortic pressure wave using
the Sphygmocor© apparatus (v7)
CAFE study Profile
2199 subjects recruited from 5
UK ASCOT centers
126 excluded due to heart rate
irregularity/poor waveforms
2073 evaluable
for tonometry
1042 received
Amlodipine-based regimen
4 subjects incomplete
information, 1 alive at last visit,
2 withdrawn consent, 1 lost to
follow-up
1042 assessed on an intentionto-treat basis 1038 complete
information
(997 alive, 41 dead)
1031 received atenolol-based
regime
1 subject incomplete
information, withdrawn
consent
1031 assessed on an intentionto-treat basis 1030 complete
information
(989 alive, 41 dead)
Baseline Demographics (1)
CAFE
ASCOT
Atenololbased
n=1031
Amlodipinebased
n=1042
Atenolol-based
n=9639
Amlodipinebased
n=9618
Women
189 (18.3%)
208 (20%)
2257 (23.5%)
2258 (23.4%)
Age (years)
62.6 (8.3)
62.9 (8.2)
63.0 (8.5)
63.0 (8.5)
White
886 (85.9%)
892 (85.6%)
9170 (95.3%)
9187 (95.3%)
Current smoker
251 (24.3%)
267 (25.6%)
3109 (32.3%)
3168 (32.9%)
SBP mm Hg
159.9 (16.6)
161 (18.4)
163.9 (18)
164.1 (18.1)
DBP (mm Hg)
92.4 (9.6)
92.6 (9.8)
94.5 (10.4)
94.8 (10.4)
Heart rate (bpm)
71.8 (12.3)
71.2 (12.4)
71.8 (12.6)
71.9 (12.7)
BMI (kg/m2)
29 (4.5)
29.1 (4.7)
28.7 (4.5)
28.7 (4.6)
Height (cm)
170.7 (8.7)
170.2 (9.4)
NA
NA
Total cholesterol
(mg/dL)
224.3 (38.7)
224.3 (42.5)
228.2 (42.5)
228.2 (42.5)
LDL-cholesterol (mg/dL)
143.1 (34.8)
143.1 (34.8)
146.9 (38.7)
146.9 (38.7)
HDL-cholesterol
(mg/dL)
50.3 (15.5)
50.3 (15.5)
50.3 (15.5)
50.3 (15.5)
Triglycerides (mg/dL)
159.4 (88.6)
159.4 (8.6)
168.3 (88.6)
159.4 (88.6)
Glucose (mg/dL)
110 (38)
110 (38)
112 (38)
112 (38)
Creatinine (mg/dL)
1.08 (0.18)
1.09 (0.19)
1.09 (0.19)
1.09 (0.18)
Baseline Demographics (2)
CAFE
ASCOT
Atenolol-based
n=1031
Amlodipinebased
n=1042
Atenolol-based
n=9639
Amlodipinebased
n=9618
Prior stroke/TIA
76 (7.4%)
62.9 (8.2)
63.0 (8.5)
63.0 (8.5)
Diabetes
252 (24.4%)
892 (85.6%)
9170 (95.3%)
9187 (95.3%)
LVH (echo or ECG)
237 (23%)
267 (25.6%)
3109 (32.3%)
3168 (32.9%)
Peripheral vascular
disease
61 (5.9%)
59 (5.7%)
613 (6.4%)
586 (6.1%)
Other relevant CV
disease
22 (2.1%)
27 (2.6%)
486 (5.1%)
533 (5.5)
Mean (SD) # risk factors
3.7 (0.9)
3.7 (0.9)
3.7 (0.9)
3.7 (0.9)
BP treatment naive
109 (10.6%)
100 (9.6%)
1825 (19%)
1841 (19.1%)
Lipid-lowering therapy
120 (11.6%)
120 (11.5%)
1004 (10.4%)
1046 (10.9%)
Aspirin use
244 (23.7%)
274 (26.3%)
1837 (19.1%)
1851 (19.2%)
Medical history
Drug therapy
Mean Proportion of Time (%) on BP
Lowering Medication by Treatment Group*
Year 1
All study
Amlodipine
90.0
80.7
Perindopril
56.0
66.7
Amlodipine + perindopril
47.8
55.5
Atenolol
88.1
73.5
Bendroflumethiazide
69.8
74.2
Atenolol + bendroflumethiazide
60.2
59.6
Randomized to Amlodipine
Randomized to Atenolol
* From time of randomization into ASCOT
CAFE Study: Different waveforms demonstrate
lower central BP in CCB/ACEI arm
amlodipine ± perindopril
Peripheral waveform
atenolol ± bendroflumethiazide
Central aortic waveform
Williams B. Circulation 2006.
BP lowering drugs have different effects on
central bloodpressure
Central aortic SBP difference:
4.3 mm Hg (P < 0.0001)
Central aortic PP difference:
3.0 mm Hg (P < 0.0001)
Williams B. Circulation 2006.
Results summary (1)
• Atenolol±thiazide was associated with higher central
aortic systolic pressure and higher central aortic
pulse pressure, despite similar brachial pressures,
when compared with amlodipine±perindopril
• Central aortic outgoing pressure wave (P1 height)
was lower with atenolol±thiazide vs
amlodipine±perindopril
• Pulse wave augmentation and the percentage of the
central aortic pressure wave attributable to wave
reflection was increased by atenolol±thiazide
compared with amlodipine±perindopril
Central vs brachial bloodpressure
•Are antihypertensive drugs equal in lowering central
bloodpressure?
• What determines central bloodpressure ?
• What are the consequences of lower central bloodpressure?
Pressure Wave Reflection at the Heart
What is it?
•
If there was no wave reflection
(ie. the aorta was an open-ended
tube providing a simple
resistance to flow), then:
•
the pressure wave in the aortic
root would be the same as the
flow wave (see graph).
Figure 1
Figure 1
Menu
Pressure Wave Reflection at the Heart
What is it?
•
Now if we connect up the
network of arteries with all its
bifurcations and vascular beds,
then:
•
as this primary wave travels
along the arteries it will generate
reflected waves from each
bifurcation and from the
peripheral vascular beds.
•
all these small reflected waves
return to the heart, summing to
create a reflected wave as
shown, starting even before the
end of systole.
Figure 2
Menu
Pressure aorta is sum of wave
reflections at the Heart
•
So the pressure in the aortic root
is the
sum of the outgoing and reflected
wave
(the green wave).
•
Note importantly how the
reflected wave boosts the
coronary artery perfusion
pressure – the aortic root
pressure – during diastole when
over 95% of perfusion of the subendocardium takes place.
Figure 3
Figure 3
Menu
What determines wave reflection at the
heart?
•
The speed at which the outgoing
and reflected waves travel is
dependent on the stiffness of the
arteries along which they are
travelling.
•
So if a person has stiffer arteries,
the waves will travel out and
back quicker, arriving earlier back
at the heart (see graph).
Figure 4
Figure 4
Menu
Pressure Wave Reflection at the Heart
•
Now when the outgoing and
reflected waves are added there
is a very different aortic root
pressure waveform.
•
There are three important
clinical implications.
Figure 5
Figure 5
Menu
Central vs brachial bloodpressure
•Are antihypertensive drugs equal in lowering central
bloodpressure?
• What determines central bloodpressure ?
•Clinical consequences of lower central bloodpressure and
possible explanations?
Pressure Wave Reflection at the Heart
Consequences
•
First, the central systolic
pressure and central pulse
pressure is increased.
•
An increase in the central pulse
pressure
that drives cerebral blood flow
increases stroke risk.
•
NOTE: this change in central
systolic pressure can occur
without any changes occurring in
peripheral cuff systolic pressure.
Increased Central
Pulse Pressure
PP
Figure 6
Menu
Fatal and non-fatal stroke
%
5.0
Atenolol  thiazide
(No. of events 422)
4.0
3.0
Amlodipine  perindopril
(No. of events 327)
2.0
HR = 0.77 (0.660.89)
p = 0.0003
1.0
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
9639
9618
9483
9461
9331
9274
9156
9059
8972
8843
5.0
7863
7720
Years
Pressure Wave Reflection at the Heart
Increased LV Load
 LVL
•
Second, there is an increase in
left ventricular load (LV load).
•
Increase in LV load accelerates
increase in LV mass and
increases risk of LV
hypertrophy.
•
The area under the pressuretime curve during systole is by
definition LV load.
•
This increase in LV Load (late
systolic “afterload”) is shown by
the black
arrowed region.
Figure 7
Menu
Pressure Wave Reflection at the Heart
•
Third, the pressure that is
perfusing the coronary arteries
during the critical diastole period
is reduced, increasing the risk
of myocardial ischemias.
•
Conclusion: Increasing arterial
stiffness independently
increases the risk of all three
major cardiovascular
outcomes.
Decreased Coronary Artery
Perfusion Pressure in
Diastole
Figure 8
Menu
Total coronary end point
%
10.0
Atenolol  thiazide
(No. of events 852)
8.0
6.0
Amlodipine  perindopril
(No. of events 753)
4.0
2.0
HR = 0.87 (0.790.96)
p = 0.0070
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
9639
9618
9400
9373
9204
9136
8984
8864
8744
8591
5.0
7614
7470
Years
CV mortality
%
3.5
Atenolol  thiazide
(No. of events 342)
3.0
2.5
2.0
Amlodipine  perindopril
(No. of events 263)
1.5
1.0
HR = 0.76 (0.650.90)
p = 0.0010
0.5
0.0
Number at risk
Amlodipine  perindopril
Atenolol  thiazide
0.0
1.0
2.0
3.0
4.0
5.0
9639
9618
9544
9532
9441
9415
9322
9261
9167
9085
8078
7975
Years
BP lowering drugs have different effects on
central bloodpressure
Brachial BP
Decrease in BP (mmHg)
2,3
SBP
DBP
Central BP (aorta)
SBP
PP
-5
-10
-15
-20
Perindopril / indapamide (2 / 0,625 mg)
Atenolol (50 mg)
-25
Adjusted difference
between group
-6,02
-0,32
-12,52
-10,34
p<0,001
P= 0,715
p< 0,001
p< 0,01
(mmHg)
Asmar R et al. Hypertension. 2001;38:922-926.
REASON study / AI decrease
P<0.001
Ao-AI (%)
2
1
+1.8
0
-1
-2
-3.1
-3
-4
Perindoprilindapamide
Atenolol
P=0.112
P=0.002
Asmar R et al. Hypertens. 2001;38:922-926.
BP lowering drugs have different effects on
structure small arteries
P < 0.01
P < 0.05
NS
8
Arterial
media-to-lumen 6
ratio (%)
4
7.94
5.96
5.82
Normotensive Before
Normotensive
n=25
Mean BP (mmHg)
Lumen ø (µm)
7.14
After
Perindopril
n=13
6.79
Before
After
Atenolol
n=12
90
122
101**
126
98**
237
208
247*ii
222
208
vs before *P <0.05. **P <0.01 - vs atenolol iiP < 0.01
Thybo NK, et al. Hypertension 1995; 25: 474-481
BP lowering drugs have different effects
on structure large arteries
carotid artery elasticity
(% improvement)
20

Double-blind randomized study

41 hypertensive patients (DBP : 95 - 110 mmHg)

Perindopril : n = 20 - Diuretic : n = 21)

6-week placebo run-in

6-month study
p < 0.05
15
16*
10
5
HCTZ 25 mg + amiloride 2,5 mg
Perindopril 4 mg
0
1
*p<0.05 versus baseline
Kool M.J. Van Bortel L.M. et al. J Hypertens 1995 ; 13 :839 - 848
CAFE Study Conclusions
• Brachial blood pressure overestimated the
hemodynamic benefit of atenolol ± thiazide-based
treatment and underestimated the benefit of
amlodipine ± perindopril-based treatment on central
aortic pressures and hemodynamics
• Central aortic pressure may be an important
independent determinant of clinical outcomes
• Results of the CAFE study suggest that the “central
aortic blood pressure hypothesis” is a plausible
mechanism to explain the better clinical outcomes for
hypertensive patients treated with amlodipine ±
perindopril-based therapy in ASCOT