Professional Documents
Culture Documents
Carlene M.M. Lawes, Derrick A. Bennett, Valery L. Feigin and Anthony Rodgers
Stroke. 2004;35:776-785; originally published online February 19, 2004;
doi: 10.1161/01.STR.0000116869.64771.5A
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright 2004 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/35/3/776
An erratum has been published regarding this article. Please see the attached page for:
http://stroke.ahajournals.org/content/35/4/1024.full.pdf
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/
Methods
Prospective Cohort Studies
Published overviews of cohort studies were identified by a search of
MEDLINE and EMBASE databases with key words overview,
meta-analysis, blood pressure, stroke or cerebrovascular or cardiovascular disease, cohort study or prospective study, and comparative
study. Overviews were included if they were published or data were
accessible by May 1, 2003, and if they provided data on inclusion
criteria, number and regions of the cohort studies included, number
of participants and demographics, duration of follow-up, disease end
points, and methods of analyses. Results were independently extracted and summarized by 2 reviewers, but no additional analyses
were conducted.
Received August 14, 2003; final revision received October 7, 2003; accepted November 19, 2003.
From the Clinical Trials Research Unit, Department of Medicine, University of Auckland, Auckland, New Zealand.
Correspondence to Carlene Lawes, Clinical Trials Research Unit, Department of Medicine, University of Auckland, Private Bag 92019, Auckland, New
Zealand. E-mail c.lawes@ctru.auckland.ac.nz
2004 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
DOI: 10.1161/01.STR.0000116869.64771.5A
776
Downloaded from http://stroke.ahajournals.org/
by guest on November 3, 2014
Lawes et al
meta-analysis, blood pressure or hypertension, stroke or cerebrovascular or cardiovascular disease, and randomized controlled trial.
Data from individual trials included in these meta-analyses were
included if they were published before May 1, 2003, and if they
provided data on inclusion criteria, total number of participants and
demographics, randomization method, trial interventions, trial end
points, duration of follow-up, and net reduction in BP (ie, treatment
effect). Data from individual studies, from both the trial and
meta-analytic publications, were independently extracted by 2 reviewers. In some cases, existing meta-analyses of trials311 included
the most recent data on individual trials, and therefore these were
included in the final meta-analyses.
Core baseline and outcome data were extracted, and trials included
in the meta-analysis were stratified into the following 3 subgroups to
reduce clinical heterogeneity: (1) drug versus placebo or no treatment, (2) more intensive dose of drug versus less intensive dose of
drug, and (3) drug versus drug. The net difference in BP between, for
example, the treatment and placebo groups was calculated; this
reflects the true difference in BP achieved as the influence of
regression to the mean (which would be evident in both groups) is
removed. The updated meta-analysis was based on summary data
rather than individual participant data. The relative risk was computed as the measure of effect size for each trial and subgroup of
TABLE 1.
777
Results
Prospective Cohort Studies
Five overviews of cohort studies assessing the association
between BP and cardiovascular disease were identified. The
No. of cohorts
Countries included
Inclusion criteria
61
37
958 074
425 251
% male
No. of participants
NR
NR
47 (20107) years
Follow-up duration
Mean 12 years
Mean 7 years
11 960 strokes
5556 strokes
100% fatal
50% fatal
NR
NR
Association by sex
No significant difference
Association by age
778
Stroke
March 2004
Lawes et al
TABLE 2.
779
Trial
Placebo
Net Difference*
Follow-Up,
y
Mean Age,
y
%
Male
SBP
DBP
RR and 95% CI
13
1721
22
1706
50
63
NR
0.59 (0.301.16)
58
58
56
43
15
10
0.99 (0.049.56)
Carter et al52
10
49
21
48
NA
57
NR
NR
0.47 (0.250.88)
53
Barraclough51
Coope
20
419
39
465
4.4
69
31
18
11
0.57 (0.340.96)
Dutch TIA54
52
732
62
741
2.6
65
64
0.85 (0.601.21)
EWPHE55
32
416
48
424
4.7
72
30
21
0.68 (0.441.04)
HDFP5660
102
5485
158
5455
51
55
NR
0.64 (0.500.82)
0.78 (0.541.11)
HSCSG I II III61
43
233
52
219
2.3
59
41
25
12
MRC older62
101
2183
134
2213
5.8
70
42
14
0.76 (0.590.98)
MRC young63
60
8700
109
8654
52
52
11
0.55 (0.400.75)
406
379
5.5
45
100
17
10
0.08 (0.001.53)
Oslo64,65
66
SHEP
105
2365
162
2371
4.5
72
43
12
0.65 (0.510.83)
STOP H67
30
812
55
815
2.1
76
37
20
0.55 (0.350.85)
USPHS68
193
196
44
80
15
10
0.17(0.021.39)
VA6970
254
23
257
3.8
51
100
17
10
0.26 (0.110.64)
VA-NHLBI71,72
508
504
1.5
38
81
NR
0.99 (0.0249.9)
Wolff et al73
45
42
1.4
49
32
NR
20
1.87 (0.1819.8)
577
24 579
897
24 547
4.6
57
53
13
0.65 (0.590.73)
156
4654
226
4652
66
73
0.69 (0.570.84)
308
309
61
82
1.76 (0.525.94)
307
3051
420
3054
64
70
0.73 (0.640.84)
QUIET
878
872
58
82
NR
NR
0.99 (0.0615.8)
SCAT77
229
231
61
89
0.22 (0.051.03)
Summary
ACE inhibitors vs placebo
HOPE74
PART 275
PROGRESS27
76
TEST78
Summary
372
69
348
2.6
70
60
1.00 (0.751.35)
547
9492
729
9466
4.3
64
73
0.72 (0.680.83)
417
408
57
80
0.98 (0.293.35)
Syst-Eur47
47
2398
77
2297
70
33
11
0.58 (0.410.84)
Summary
52
2815
82
2705
68
40
10
0.61 (0.440.85)
nnumber of events; Nnumber of participants. SBP indicates systolic blood pressure; DBP, diastolic blood pressure; RR, relative risk; 95% CI, 95% confidence
interval; NR, not recorded.
*Net BP differenceblood pressure difference between the randomized groups in mm Hg (ie, BP change in treatment groupBP change in placebo group).
Summary follow-up, mean age, % male, and net SBP and DBP change are weighted by study size for those studies that provided data on net changes in both
SBP and DBP. Weighting by number of events made negligible difference. The pooled RR and 95% CI (using the fixed effects approach) for each of the 3 trial subgroups
are presented, and the test for heterogeneity (Q statistic and P value) are Q14.23, P0.36; Q8.90, P0.11; and Q0.61, P0.43, respectively.
These trials also included older classes of drugs such as methyldopa and alkaloids.
There were no stroke events in the Barraclough or VA-NHLBI trial; however, for the purposes of meta-analyses, 0.5 was added to all cells in the 22 table.
780
Stroke
March 2004
results were of borderline statistical significance. The latter
subgroup was the only one in which there was evidence of
statistical heterogeneity (P0.09), and a random effects
model gave a nonsignificant relative risk reduction of 5%.
Figure 2. Randomized controlled trials comparing antihypertensive drugs with a placebo (or no treatment) by subgroup. The
meta-analyses of BP-lowering trials are presented stratified into
subgroups on the basis of mean age of trial participants at
entry, baseline SBP level, and whether trial participants predominantly had a history of stroke/transient ischemic attack (TIA) or
vascular disease. The diamonds are centered on the pooled
estimate of effect and represent 95% CI. The solid diamond
represents the pooled relative risk and 95% CI for all contributing trials.
More
ABCD22
21
UKPDS-HDS
Summary
Discussion
Cohort studies have demonstrated that the strong association
between BP and stroke appears to be continuous down to
levels of at least 115/75 mm Hg. The strength of the associations was similar for men and women and for fatal and
nonfatal events but attenuated with age.2,1518 Despite a
potentially weaker relative association for older age groups,
the higher overall stroke rate in this group means that in
absolute terms, the benefits of a given reduction in BP are
likely to be greater. The analyses have not consistently
demonstrated a difference in the association between intracerebral hemorrhage and ischemic stroke; however, only approximately one half of stroke subtypes were confirmed (CT,
MRI, or autopsy). Regional comparisons are very limited
Randomized Controlled Trials Comparing More Intensive vs Less Intensive Antihypertensive Drug Regimens
More Intensive vs
Less Intensive Trials
HOT
23,24
Less
N
Net Difference*
N
Follow-Up,
y
Mean Age,
y
%
Male
SBP
DBP
RR and 95% CI
58
67
0.98 (0.402.43)
237
233
89
6262
205
12528
62
53
0.87 (0.681.11)
38
758
34
390
56
55
10
0.58 (0.370.90)
136
7257
248
13 151
4.2
62
53
0.80 (0.650.99)
nno. of events, Nno. of participants. SBP indicates systolic blood pressure; DBP, diastolic blood pressure; RR, relative risk; 95% CI, 95% confidence interval.
* Net BP differenceblood pressure difference between the randomized groups (mm Hg) (ie, BP change in more intensive drug treatment groupBP change in
less intensive drug treatment group).
Summary follow-up, mean age, % male, and net SBP and DBP change are weighted by study size for those studies that provided data on net changes in both
SBP and DBP. Weighting by number of events made negligible difference. The pooled RR and 95% confidence interval (using the fixed effects approach) are presented,
and the test for heterogeneity (Q statistic and P value) are Q2.70, P0.26.
Lawes et al
TABLE 4.
781
Drug vs Drug
Trials
Second Drug
N
Net Difference*
Follow-Up,
y
Mean Age,
y
%
Male
SBP
DBP
RR and 95% CI
675
15255
457
9054
4.9
67
53
0.98(0.780.98)
ANBPS81
107
3039
112
3044
4.1
72
49
0.95(0.741.23)
CAPPP82
148
5493
189
5492
53
53
0.78(0.630.97)
237
2213
215
2205
76
33
1.10(0.931.32)
17
358
21
400
56
54
0.90(0.491.69)
1184
26 358
994
20 195
5.1
65
51
0.91(0.830.99)
1.06(0.931.20)
46
STOP-2
UKPDS-HDS
Summary
23,24
675
15255
377
9048
67
53
CASTEL83
205
146
73
51
0.71(2.440.21)
118
8361
133
8241
66
44
0.88(0.681.12)
ELSA85
14
1157
1177
56
55
1.59(0.693.70)
INSIGHT86
84
3164
79
3157
65
46
1.06(0.781.43)
MIDAS87
441
442
59
78
0.50(0.132.00)
NICS-EH88
214
215
70
33
1.33(0.473.85)
89
NORDIL
196
5471
159
5410
60
49
1.22(0.991.49)
STOP-246
237
2213
207
2196
76
33
1.14(0.951.35)
VHAS90,91
707
707
54
49
0.80(0.222.94)
1344
37 188
986
30 739
4.2
65
48
1.08(0.991.16)
CONVINCE84
Summary
FACET
STOP-246
Summary
11
235
235
58
67
1.57(0.623.98)
377
9048
457
9054
67
53
0.83(0.720.94)
10
191
189
207
2196
215
2205
11 683
11 683
11 683
11 683
3.5
63
41
2.47(0.797.75)
76
33
0.97(0.811.16)
4.9
68
49
0.89(0.800.99)
nno. of events, Nno. of participants; SBP indicates systolic blood pressure; DBP, diastolic blood pressure; RR, relative risk; 95% CI, 95% confidence interval.
*Net BP differenceblood pressure difference between the randomized groups (mm Hg). Presented as the BP change in (first listed drug)(second listed drug),
ie, a negative value means that the second listed drug achieved a greater blood pressure reduction than the first listed drug.
Summary follow-up, mean age, % male, and net SBP and DBP change are weighted by study size for those studies that provided data on net changes in both
SBP and DBP. Weighting by number of events made negligible difference. The pooled RR and 95% confidence interval (using the fixed effects approach) for each
of the 3 trial subgroups are presented, and the test for heterogeneity (Q statistic and P value) are Q6.91, P0.14; Q7.36, P0.59; and Q6.53, P0.09,
respectively. The pooled RR and 95% CI using a random effects approach for the last trial subgroup was 0.95 (0.76 1.19).
782
Stroke
March 2004
Heart Attack Trial (ALLHAT)30 and the Losartan Intervention
For Endpoint reduction in hypertension study (LIFE)31,32 were
not included in the meta-analysis because no other trials included the same classes of drugs. They indicated that there was
a higher risk of stroke in the -adrenergic blocker group
compared with diuretics30 and that angiotensin II blockers may
have greater impact on stroke than -blockers.31,32
TABLE 5. Relative Risk Reduction in Stroke With a Reduction in Systolic Blood Pressure Predicted From
Cohort Studies and Observed From Randomized Controlled Trials
Observed Effects on Stroke With a Reduction
in SBP of 10 mm Hg
6069 Years
7079 Years
Approximately 73 Years
34%
29%
31%
36%
25%
Lawes et al
data improve the understanding of the associations between
BP and stroke and the potential benefits of BP lowering in
different population subgroups. Additional analyses are able
to examine the combined impact of multiple risk factors on
stroke risk. BP intervention strategies would be greatly
enhanced if these data were also used in the improvement of
risk prediction frameworks to better identify those at higher
absolute risk of cardiovascular disease. Development of
simple versions of these risk prediction tools would be of
particular value in developing regions where much of the
future cardiovascular disease burden is likely to occur.1,49
Future trial analyses should focus on the presence or
absence of age attenuation in the effect of BP lowering. This
would require comparison of age-specific, follow-up time
specific, and end pointspecific analyses of individual participant data from both observational studies (such as APCSC)
and clinical trials (such as the Blood Pressure Lowering
Treatment Trialists Collaboration11). Given that most individuals require combination therapy, it is also important to
shift the focus of trials from comparisons of the short-term
benefits of individual drugs. Instead, future trials should aim
to determine the optimal combination therapy and the potential long-term benefits of different antihypertensive regimens
in different population subgroups. While there may be some
differences between BP-lowering agents, the relative differences between these agents are minor compared with the
relative benefits of any BP-lowering agent versus no
treatment.
Finally, it is important to acknowledge that treatment of
high-risk groups will only partially address the growing
cardiovascular disease epidemic. Data presented in this article
indicated that the relative benefits of BP lowering are not
only limited to those classified as hypertensive but would also
be evident in those with normal or below normal BP. These
relative benefits appear broadly consistent across many population subgroups. Other important areas of research therefore include exploration of cost-effective strategies to bring
about population changes in cardiovascular risk factors such
as BP. Such initiatives would complement targeted treatment
approaches and would not only benefit the developed world
but could potentially arrest the projected increase in cardiovascular disease in developing countries.
Acknowledgments
This study was supported by a Health Research Council (New
Zealand) fellowship to Dr Lawes. Dr Rodgers is a National Health
Foundation (New Zealand) Senior Research Fellow. We thank
Varsha Parag and Ruey-Bin Lin for statistical assistance and Clarissa
Could-Thorpe for secretarial support.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
References
1. World Health Organisation. The World Health Report 2002: Reducing
Risks, Promoting Healthy Life. Geneva, Switzerland: World Health
Organisation; 2002.
2. MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R,
Godwin J, Dyer A, Stamler J, et al. Blood pressure, stroke, and coronary
heart disease, part I: prolonged differences in blood pressure: prospective
observational studies corrected for the regression dilution bias. Lancet.
1990;335:765774.
3. Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA,
Godwin J, Qizilbash N, Taylor JO, Hennekens CH, et al. Blood pressure,
stroke, and coronary heart disease, part 2: short-term reductions in blood
23.
24.
25.
26.
783
784
Stroke
March 2004
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
Lawes et al
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
785
Correction
The article entitled Blood Pressure and Stroke: An Overview of Published Reviews, by Lawes
et al1 contained incorrect data resulting from a production error. The corrected article appears
below. The publisher regrets the error.
Received August 14, 2003; final revision received October 7, 2003; accepted November 19, 2003.
From the Clinical Trials Research Unit, Department of Medicine, University of Auckland, Auckland, New Zealand.
Correspondence to Carlene Lawes, Clinical Trials Research Unit, Department of Medicine, University of Auckland, Private Bag 92019, Auckland, New
Zealand. E-mail c.lawes@ctru.auckland.ac.nz
1
[Correction for Vol 35, Number 3, March 2004. Pages 776 785.]
2004 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
DOI: 10.1161/01.STR.0000126208.14181.DD
1024
Lawes et al
Methods
Prospective Cohort Studies
Published overviews of cohort studies were identified by a search of
MEDLINE and EMBASE databases with key words overview,
meta-analysis, blood pressure, stroke or cerebrovascular or cardiovascular disease, cohort study or prospective study, and comparative
study. Overviews were included if they were published or data were
accessible by May 1, 2003, and if they provided data on inclusion
criteria, number and regions of the cohort studies included, number
of participants and demographics, duration of follow-up, disease end
points, and methods of analyses. Results were independently extracted and summarized by 2 reviewers, but no additional analyses
were conducted.
TABLE 1.
1025
effect). Data from individual studies, from both the trial and
meta-analytic publications, were independently extracted by 2 reviewers. In some cases, existing meta-analyses of trials311 included
the most recent data on individual trials, and therefore these were
included in the final meta-analyses.
Core baseline and outcome data were extracted, and trials included
in the meta-analysis were stratified into the following 3 subgroups to
reduce clinical heterogeneity: (1) drug versus placebo or no treatment, (2) more intensive dose of drug versus less intensive dose of
drug, and (3) drug versus drug. The net difference in BP between, for
example, the treatment and placebo groups was calculated; this
reflects the true difference in BP achieved as the influence of
regression to the mean (which would be evident in both groups) is
removed. The updated meta-analysis was based on summary data
rather than individual participant data. The relative risk was computed as the measure of effect size for each trial and subgroup of
trials. Statistical heterogeneity was assessed with the use of the
standard Q statistic for heterogeneity, with a probability value of
0.1 used to indicate evidence of heterogeneity. A less stringent
probability value of 0.1 was used because it is well known that tests
of heterogeneity have low statistical power. If there was no evidence
of statistical heterogeneity when this criterion was used, then the
results were combined with the use of the standard inverse-variance
weighted fixed effect approach.12 If there was evidence of statistical
heterogeneity between the trials, attempts were made to explore
No. of cohorts
Countries included
Inclusion criteria
61
37
No. of participants
958 074
425 325
% male
NR
NR
47 (20107) years
Mean 12 years
Mean 7 years
11 960 strokes
5178 strokes
Follow-up duration
100% fatal
50% fatal
NR
NR
Association by sex
No significant difference
Association by age
1026
Stroke
April 2004
Results
Prospective Cohort Studies
Five overviews of cohort studies assessing the association
between BP and cardiovascular disease were identified. The
MacMahon2 and Prospective Studies Collaboration (PSC)15,16
overviews included cohorts predominantly from the United
Kingdom, United States, and Europe. Cohorts in the MacMahon2 and first PSC15 review were subsequently included in
the later PSC review.16 The Eastern Stroke and Coronary
Heart Disease Collaborative Research17 overview is a subset
of the Asia Pacific Cohort Studies Collaboration (APCSC),18
which comprises cohorts from Mainland China, Hong Kong,
Taiwan, Japan, Singapore, South Korea, Australia, and New
Zealand. The PSC and APCSC overviews included almost 1
million and 420 000 individuals, respectively; both analyzed individual participant data and corrected for regression
dilution bias to reflect true or usual BP levels.16,18 In total,
7 cohorts were included in both the PSC and APCSC
overviews, representing 5% and 10% of participants,
respectively. The main characteristics and results of the
overviews are summarized in Table 1.
An important finding in all overviews was that the association between BP and risk of stroke was continuous and log
linear. Therefore, a given absolute difference in usual BP was
associated with a similar relative risk reduction of stroke at all
levels of BP. There was no evidence of a threshold below
which levels of BP were no longer associated with lower
relative risk of stroke, down to approximately 115 mm Hg
systolic blood pressure (SBP) or 75 mm Hg DBP.16,18 All
overviews commented on the statistical heterogeneity between studies, and age appeared to be an important factor,16,18
explaining approximately 80% of the between-study hetero-
Lawes et al
TABLE 2.
1027
Trial
Placebo
Net Difference*
N
Follow-Up,
y
Mean Age,
y
%
Male
SBP
DBP
RR and 95% CI
13
1721
22
1706
4.0
50
63
NR
0.59 (0.301.16)
58
58
2.0
56
43
15
10
0.99 (0.049.56)
Carter et al52
10
49
21
48
4.0
NA
57
NR
NR
0.47 (0.250.88)
53
Coope
20
419
39
465
4.4
69
31
18
11
0.57 (0.340.96)
Dutch TIA54
52
732
62
741
2.6
65
64
0.85 (0.601.21)
EWPHE55
32
416
48
424
4.7
72
30
21
0.68 (0.441.04)
HDFP5660
102
5485
158
5455
5.0
51
55
NR
0.64 (0.500.82)
0.78 (0.541.11)
Barraclough51
HSCSG I II III61
43
233
52
219
2.3
59
41
25
12
MRC older62
101
2183
134
2213
5.8
70
42
14
0.76 (0.590.98)
MRC young63
60
8700
109
8654
5.0
52
52
11
0.55 (0.400.75)
406
379
5.5
45
100
17
10
0.08 (0.001.53)
0.65 (0.510.83)
Oslo64,65
66
SHEP
105
2365
162
2371
4.5
72
43
12
STOP H67
30
812
55
815
2.1
76
37
20
0.55 (0.350.85)
USPHS68
193
196
7.0
44
80
15
10
0.17 (0.021.39)
VA6970
254
23
257
3.8
51
100
17
10
0.26 (0.110.64)
508
504
1.5
38
81
NR
0.99 (0.0249.9)
45
42
1.4
49
32
NR
20
1.87 (0.1819.8)
577
24 579
897
24 547
4.6
57
53
13
0.65 (0.590.73)
156
4654
226
4652
4.5
66
73
0.69 (0.570.84)
308
309
4.7
61
82
1.76 (0.525.94)
307
3051
420
3054
3.9
64
70
0.73 (0.640.84)
QUIET
878
872
2.3
58
82
NR
NR
0.99 (0.0615.8)
SCAT77
229
231
4.0
61
89
0.22 (0.051.03)
71,72
VA-NHLBI
Wolff et al73
Summary
ACE inhibitors vs placebo
HOPE74
PART 275
PROGRESS27
76
TEST78
Summary
74
372
69
348
2.6
70
60
1.00 (0.751.35)
547
9492
729
9466
4.3
64
73
0.72 (0.680.83)
417
408
3.0
57
80
0.98 (0.293.35)
Syst-Eur47
47
2398
77
2297
2.6
70
33
11
0.58 (0.410.84)
Summary
52
2815
82
2705
2.8
68
40
10
0.61 (0.440.85)
nnumber of events; Nnumber of participants. SBP indicates systolic blood pressure; DBP, diastolic blood pressure; RR, relative risk; 95% CI, 95% confidence
interval; NR, not recorded.
*Net BP differenceblood pressure difference between the randomized groups in mm Hg (ie, BP change in treatment groupBP change in placebo group).
Summary follow-up, mean age, % male, and net SBP and DBP change are weighted by study size for those studies that provided data on net changes in both
SBP and DBP. Weighting by number of events made negligible difference. The pooled RR and 95% CI (using the fixed effects approach) for each of the 3 trial subgroups
are presented, and the test for heterogeneity (Q statistic and P value) are Q14.23, P0.36; Q8.90, P0.11; and Q0.61, P0.43, respectively.
These trials also included older classes of drugs such as methyldopa and alkaloids.
There were no stroke events in the Barraclough or VA-NHLBI trial; however, for the purposes of meta-analyses, 0.5 was added to all cells in the 22 table.
1028
Stroke
April 2004
Figure 2. Randomized controlled trials comparing antihypertensive drugs with a placebo (or no treatment) by subgroup. The
meta-analyses of BP-lowering trials are presented stratified into
subgroups on the basis of mean age of trial participants at
entry, baseline SBP level, and whether trial participants predominantly had a history of stroke/transient ischemic attack (TIA) or
vascular disease. The diamonds are centered on the pooled
estimate of effect and represent 95% CI. The solid diamond
represents the pooled relative risk and 95% CI for all contributing trials.
More
ABCD22
21
UKPDS-HDS
Summary
Randomized Controlled Trials Comparing More Intensive vs Less Intensive Antihypertensive Drug Regimens
More Intensive vs
Less Intensive Trials
HOT
23,24
Less
N
Net Difference*
N
Follow-Up,
y
Mean Age,
y
%
Male
SBP
DBP
RR and 95% CI
5.3
58
67
0.98 (0.402.43)
237
233
89
6262
205
12528
3.8
62
53
0.87 (0.681.11)
38
758
34
390
8.4
56
55
10
0.58 (0.370.90)
136
7257
248
13 151
4.2
62
53
0.80 (0.650.99)
nno. of events, Nno. of participants. SBP indicates systolic blood pressure; DBP, diastolic blood pressure; RR, relative risk; 95% CI, 95% confidence interval.
* Net BP differenceblood pressure difference between the randomized groups (mm Hg) (ie, BP change in more intensive drug treatment groupBP change in
less intensive drug treatment group).
Summary follow-up, mean age, % male, and net SBP and DBP change are weighted by study size for those studies that provided data on net changes in both
SBP and DBP. Weighting by number of events made negligible difference. The pooled RR and 95% confidence interval (using the fixed effects approach) are presented,
and the test for heterogeneity (Q statistic and P value) are Q2.70, P0.26.
Lawes et al
TABLE 4.
1029
Drug vs Drug
Trials
Second Drug
N
Net Difference*
Follow-Up,
y
Mean Age,
y
%
Male
SBP
DBP
RR and 95% CI
675
15255
457
9054
4.9
67
53
0.88 (0.780.98)
ANBPS81
107
3039
112
3044
4.1
72
49
0.95 (0.741.23)
CAPPP82
148
5493
189
5492
6.1
53
53
0.78 (0.630.97)
237
2213
215
2205
5.0
76
33
1.10 (0.931.32)
17
358
21
400
8.4
56
54
0.90 (0.491.69)
1184
26 358
994
20 195
5.1
65
51
0.91 (0.830.99)
46
STOP-2
UKPDS-HDS
Summary
23,24
675
15255
377
9048
4.9
67
53
1.06 (0.931.20)
CASTEL83
205
146
7.0
73
51
0.71 (2.440.21)
118
8297
133
8179
3.0
66
44
0.88 (0.681.12)
ELSA
14
1157
1177
4.0
56
55
1.59 (0.693.70)
INSIGHT86
74
3164
67
3157
4.0
65
46
1.06 (0.781.43)
MIDAS87
441
442
3.0
59
78
0.50 (0.132.00)
NICS-EH88
1.33 (0.473.85)
CONVINCE84
85
214
215
5.0
70
33
89
NORDIL
196
5471
159
5410
5.0
60
49
1.22 (0.991.49)
STOP-246
237
2213
207
2196
5.0
76
33
1.14 (0.951.35)
VHAS90,91
707
707
2.0
54
49
0.80 (0.222.94)
1334
37 124
974
30 677
4.2
65
48
1.08 (0.991.16)
Summary
FACET
11
235
235
5.3
58
67
1.57 (0.623.98)
377
9048
457
9054
4.9
67
53
0.83 (0.720.94)
2.47 (0.797.75)
10
191
189
3.5
63
41
STOP-246
207
2196
215
2205
5.0
76
33
0.97 (0.811.16)
Summary
605
11 670
683
11 683
4.9
68
49
0.89 (0.800.99)
nno. of events, Nno. of participants; SBP indicates systolic blood pressure; DBP, diastolic blood pressure; RR, relative risk; 95% CI, 95% confidence interval.
*Net BP differenceblood pressure difference between the randomized groups (mm Hg). Presented as the BP change in (first listed drug)(second listed drug),
ie, a negative value means that the second listed drug achieved a greater blood pressure reduction than the first listed drug.
Summary follow-up, mean age, % male, and net SBP and DBP change are weighted by study size for those studies that provided data on net changes in both
SBP and DBP. Weighting by number of events made negligible difference. The pooled RR and 95% confidence interval (using the fixed effects approach) for each
of the 3 trial subgroups are presented, and the test for heterogeneity (Q statistic and P value) are Q6.91, P0.14; Q7.36, P0.59; and Q6.53, P0.09,
respectively. The pooled RR and 95% CI using a random effects approach for the last trial subgroup was 0.95 (0.76 1.19).
Discussion
Cohort studies have demonstrated that the strong association
between BP and stroke appears to be continuous down to
levels of at least 115/75 mm Hg. The strength of the associations was similar for men and women and for fatal and
nonfatal events but attenuated with age.2,1518 Despite a
potentially weaker relative association for older age groups,
the higher overall stroke rate in this group means that in
absolute terms, the benefits of a given reduction in BP are
likely to be greater. The analyses have not consistently
demonstrated a difference in the association between intracerebral hemorrhage and ischemic stroke; however, only approximately one half of stroke subtypes were confirmed (CT,
MRI, or autopsy). Regional comparisons are very limited
because most of the cohort studies have been conducted
primarily in the developed parts of the world; however, the
1030
Stroke
April 2004
size for men and women or for fatal and nonfatal events.3 8
Limited data were available on treatment effects by age, and the
treatment effects by age within individual trials27 and between
trials4,5,9,10,28 have not been consistent. However, the trials were
often not powered for age subgroup analyses and often included
only a narrow age band. Few data are available on stroke
subtypes, limiting the scope for meta-analyses. The limited data
from the Perindopril Protection Against Recurrent Stroke Study
(PROGRESS)27 and the Heart Outcomes Prevention Evaluation
(HOPE) study29 are insufficient to draw definitive conclusions
regarding the impact of BP lowering on stroke subtypes. Anal-
TABLE 5. Relative Risk Reduction in Stroke With a Reduction in Systolic Blood Pressure Predicted From
Cohort Studies and Observed From Randomized Controlled Trials
Observed Effects on Stroke With a Reduction
in SBP of 10 mm Hg
6069 Years
7079 Years
Approximately 73 Years
34%
29%
31%
36%
25%
Lawes et al
suggest a dose-response relationship. At mean age at baseline
of approximately 63 years, a 10 mm Hg reduction in SBP was
associated with a risk reduction in stroke of 31%. Given that
most of these individuals would not have experienced their
stroke event until approximately a decade later, the results are
consistent with the predicted reduction in stroke risk from the
cohort study data. This indicates that most of the epidemiologically expected benefit of BP lowering may be achieved
within a few years for stroke.
Acknowledgments
This study was supported by a Health Research Council (New
Zealand) fellowship to Dr Lawes. Dr Rodgers is a National Health
1031
References
1. World Health Organisation. The World Health Report 2002: Reducing
Risks, Promoting Healthy Life. Geneva, Switzerland: World Health
Organisation; 2002.
2. MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R,
Godwin J, Dyer A, Stamler J. Blood pressure, stroke, and coronary heart
disease, part I: prolonged differences in blood pressure: prospective
observational studies corrected for the regression dilution bias. Lancet.
1990;335:765774.
3. Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA,
Godwin J, Qizilbash N, Taylor JO, Hennekens CH. Blood pressure,
stroke, and coronary heart disease, part 2: short-term reductions in blood
pressure: overview of randomised drug trials in their epidemiological
context. Lancet. 1990;335:827 838.
4. MacMahon S, Rodgers A. The effects of anti-hypertensive treatment on
vascular disease: reappraisal of the evidence in 1994. J Vasc Med Biol.
1993;4:265271.
5. MacMahon S, Rodgers A. The effects of blood pressure reduction in older
patients: an overview of five randomized controlled trials in elderly
hypertensives. Clin Exp Hypertens. 1993;15:967978.
6. Insua JT, Sacks HS, Lau TS, Lau J, Reitman D, Pagano D, Chalmers TC.
Drug treatment of hypertension in the elderly: a meta-analysis. Ann Intern
Med. 1994;121:355362.
7. Gueyffier F, Boutitie F, Boissel JP, Pocock S, Coope J, Cutler J, Ekbom
T, Fagard R, Friedman L, Kerlikowske K, Perry M, Prineas R, Schron E,
for the INDANA Investigators. Effect of antihypertensive drug treatment
on cardiovascular outcomes in women and men: a meta-analysis of
individual patient data from randomized, controlled trials. Ann Intern
Med. 1997;126:761767.
8. He J, Whelton PK. Elevated systolic blood pressure as a risk factor for
cardiovascular and renal disease. J Hypertens Suppl. 1999;17:S7S13.
9. Gueyffier F, Boutitie F, Boissel JP, Coope J, Cutler J, Ekbom T, Fagard
R, Friedman L, Perry HM, Pocock S. INDANA: a meta-analysis on
individual patient data in hypertension: protocol and preliminary results.
Therapie. 1995;50:353362.
10. Gueyffier F, Froment A, Gouton M. New meta-analysis of treatment trials
of hypertension: improving the estimate of therapeutic benefit. J Hum
Hypertens. 1996;10:1 8.
11. Neal B, MacMahon S, Chapman N, for the Blood Pressure Lowering
Treatment Trialists Collaboration. Effects of ACE inhibitors, calcium
antagonists, and other blood-pressure-lowering drugs: results of prospectively designed overviews of randomised trials. Lancet. 2000;356:
19551964.
12. Hedges LV, Olkin I. Statistical Methods for Meta-analysis. London, UK:
Academic Press; 1985.
13. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin
Trials. 1986;7:177188.
14. Light RJ, Pillemer DB. Summing Up: The Science of Reviewing Research.
Cambridge, Mass: Harvard University Press; 1984.
15. Prospective Studies Collaboration. Cholesterol, diastolic blood pressure,
and stroke: 13,000 strokes in 45,000 people in 45 prospective cohorts.
Lancet. 1995;346:16471653.
16. Prospective Studies Collaboration. Age-specific relevance of usual blood
pressure to vascular mortality: a meta-analysis of individual data for one
million adults in 61 prospective studies. Lancet. 2002;360:19031913.
17. Eastern Stroke and Coronary Heart Disease Collaborative Research
Group. Blood pressure, cholesterol, and stroke in eastern Asia. Lancet.
1998;352:18011807.
18. Asia Pacific Cohort Studies Collaboration. Blood pressure and cardiovascular disease in the Asia Pacific region. J Hypertens. 2003;21:
707716.
19. Svensson P, de Faire U, Sleight P, Yusuf S, Ostergren J. Comparative
effects of ramipril on ambulatory and office blood pressures: a HOPE
substudy. Hypertension. 2001;38:E28 E32.
20. Sterne JA, Egger M, Smith GD. Systematic reviews in health care:
investigating and dealing with publication and other biases in meta-analysis. BMJ. 2001;323:101105.
21. Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfeldt D, Julius S,
Menard J, Rahn KH, Wedel H, Westerling S, for the HOT Study Group.
Effects of intensive blood-pressure lowering and low-dose aspirin in
1032
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
Stroke
April 2004
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
Lawes et al
69. Effects of treatment on morbidity in hypertension: results in patients with
diastolic blood pressures averaging 115 through 129 mm Hg. JAMA.
1967;202:1028 1034.
70. Effects of treatment on morbidity in hypertension, II: results in patients
with diastolic blood pressure averaging 90 through 114 mm Hg. JAMA.
1970;213:11431152.
71. Perry HM Jr. Treatment of mild hypertension: preliminary results of a
two-year feasibility trial. Circ Res. 1977;40(suppl 1):I180 I187.
72. Veterans AdministrationNational Heart, Lung, and Blood Institute Study
Group for Evaluating Treatment in Mild Hypertension. Evaluation of
drug treatment in mild hypertension: VA-NHLBI feasibility trial: plan
and preliminary results of a two-year feasibility trial for a multicenter
intervention study to evaluate the benefits versus the disadvantages of
treating mild hypertension. Ann N Y Acad Sci. 1978;304:267292.
73. Wolff FW, Lindeman RD. Effects of treatment in hypertension: results of
a controlled study. J Chron Dis. 1966;19:227240.
74. Yusuf S, Sleight P, Pogue J, Bosch J, Davies R, Dagenais G, for the Heart
Outcomes Prevention Evaluation Study Investigators. Effects of an angiotensin-converting-enzyme inhibitor, ramipril, on cardiovascular events
in high-risk patients. N Engl J Med. 2000;342:145153.
75. MacMahon S, Sharpe N, Gamble G, Clague A, Mhurchu CN, Clark T,
Hart H, Scott J, White H, for the PART-2 Collaborative Research Group.
Randomized, placebo-controlled trial of the angiotensin-converting
enzyme inhibitor, ramipril, in patients with coronary or other occlusive
arterial disease: Prevention of Atherosclerosis with Ramipril. J Am Coll
Cardiol. 2000;36:438 443.
76. Cashin-Hemphill L, Holmvang G, Chan RC, Pitt B, Dinsmore RE, Lees
RS, for the QUIET Investigators. Angiotensin-converting enzyme inhibition as antiatherosclerotic therapy: no answer yet: QUinapril Ischemic
Event Trial. Am J Cardiol. 1999;83:43 47.
77. Teo KK, Burton JR, Buller CE, Plante S, Catellier D, Tymchak W, et al.
Long-term effects of cholesterol lowering and angiotensin-converting
enzyme inhibition on coronary atherosclerosis: the Simvastatin/Enalapril
Coronary Atherosclerosis Trial (SCAT). Circulation. 2000;102:
1748 1754.
78. Eriksson S, Olofsson BO, Wester PO. Atenolol in secondary prevention
after stroke. Cerebrovasc Dis. 1995;5:2125.
79. Pitt B, Byington RP, Furberg CD, Hunninghake DB, Mancini GB, Miller
ME, Riley W. Effect of amlodipine on the progression of atherosclerosis
and the occurrence of clinical events. Circulation. 2000;102:15031510.
80. The ALLHAT Officers and Coordinators for the ALLHAT Collaborative
Research Group. Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel
blocker vs diuretic: the Antihypertensive and Lipid-Lowering Treatment
to Prevent Heart Attack Trial (ALLHAT). JAMA. 2002;288:29812997.
81. Wing LM, Reid CM, Ryan P, Beilin LJ, Brown MA, Jennings GL,
Johnston CI, McNeil JJ, Macdonald GJ, Marley JE, Morgan TO, West
MJ; Second Australian National Blood Pressure Study. A comparison of
outcomes with angiotensin-converting enzyme inhibitors and diuretics
for hypertension in the elderly. N Engl J Med. 2003;348:583592.
82. Hansson L, Lindholm LH, Niskanen L, Lanke J, Hedner T, Niklason A,
Luomanmaki K, Dahlof B, de Faire U, Morlin C, Karlberg BE, Wester
PO, Bjork JE. Effect of angiotensin-converting-enzyme inhibition
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
1033