Articles

Long-term mortality after blood pressure-lowering and

lipid-lowering treatment in patients with hypertension in
the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT)
Legacy study: 16-year follow-up results of a randomised
factorial trial
Ajay Gupta, Judith Mackay, Andrew Whitehouse, Thomas Godec, Tim Collier, Stuart Pocock, Neil Poulter, Peter Sever

Summary
Background In patients with hypertension, the long-term cardiovascular and all-cause mortality effects of different Published Online
blood pressure-lowering regimens and lipid-lowering treatment are not well documented, particularly in clinical trial August 26, 2018
http://dx.doi.org/10.1016/
settings. The Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT) Legacy Study reports mortality outcomes after 16 S0140-6736(18)31776-8
years of follow-up of the UK participants in the original ASCOT trial.
See Online/Comment
http://dx.doi.org/10.1016/
Methods ASCOT was a multicentre randomised trial with a 2 × 2 factorial design. UK-based patients with hypertension S0140-6736(18)31943-3
were followed up for all-cause and cardiovascular mortality for a median of 15·7 years (IQR 9·7–16·4 years). At William Harvey Research
baseline, all patients enrolled into the blood pressure-lowering arm (BPLA) of ASCOT were randomly assigned to Institute, Queen Mary
University of London, London,
receive either amlodipine-based or atenolol-based blood pressure-lowering treatment. Of these patients, those who had
UK (A Gupta MRCP); National
total cholesterol of 6·5 mmol/L or lower and no previous lipid-lowering treatment underwent further randomisation Heart and Lung Institute,
to receive either atorvastatin or placebo as part of the lipid-lowering arm (LLA) of ASCOT. The remaining patients Imperial College London,
formed the non-LLA group. A team of two physicians independently adjudicated all causes of death. London, UK (J Mackay MRCP,
A Whitehouse MBBS,
Prof P Sever FRCP); Department
Findings Of 8580 UK-based patients in ASCOT, 3282 (38·3%) died, including 1640 (38·4%) of 4275 assigned to of Medical Statistics, London
atenolol-based treatment and 1642 (38·1%) of 4305 assigned to amlodipine-based treatment. 1768 of the 4605 patients School of Hygiene & Tropical
in the LLA died, including 903 (39·5%) of 2288 assigned placebo and 865 (37·3%) of 2317 assigned atorvastatin. Of Medicine, London, UK
(T Godec MSc, T Collier MSc,
all deaths, 1210 (36·9%) were from cardiovascular-related causes. Among patients in the BPLA, there was no overall Prof S Pocock PhD); and
difference in all-cause mortality between treatments (adjusted hazard ratio [HR] 0·90, 95% CI 0·81–1·01, p=0·0776]), Imperial Clinical Trials Unit,
although significantly fewer deaths from stroke (adjusted HR 0·71, 0·53–0·97, p=0·0305) occurred in the amlodipine- Imperial College London,
based treatment group than in the atenolol-based treatment group. There was no interaction between treatment London, UK
(Prof N Poulter FMedSci)
allocation in the BPLA and in the LLA. However, in the 3975 patients in the non-LLA group, there were fewer
Correspondence to:
cardiovascular deaths (adjusted HR 0·79, 0·67–0·93, p=0·0046) among those assigned to amlodipine-based treatment
Prof Peter Sever, National Heart
compared with atenolol-based treatment (p=0·022 for the test for interaction between the two blood pressure and Lung Institute, Imperial
treatments and allocation to LLA or not). In the LLA, significantly fewer cardiovascular deaths (HR 0·85, 0·72–0·99, College London, London
p=0·0395) occurred among patients assigned to statin than among those assigned placebo. W12 0NN, UK
p.sever@imperial.ac.uk

Interpretation Our findings show the long-term beneficial effects on mortality of antihypertensive treatment with a
calcium channel blocker-based treatment regimen and lipid-lowering with a statin: patients on amlodipine-based
treatment had fewer stroke deaths and patients on atorvastatin had fewer cardiovascular deaths more than 10 years after
trial closure. Overall, the ASCOT Legacy study supports the notion that interventions for blood pressure and cholesterol
are associated with long-term benefits on cardiovascular outcomes.

Funding Pfizer.

Copyright © 2018 Elsevier Ltd. All rights reserved.

Introduction Outcomes Trial (ASCOT)3 and Avoiding Cardiovascular

Guidelines for the management of patients with hyper­
tension highlight the importance of blood pressure
control, although the target blood pressures, particularly
for systolic pressure, remain controversial. Some guide­
lines advocate preferred drug treatment regimens1,2 on
the basis of the results of cardiovascular disease
outcome trials such as the Anglo-Scandinavian Cardiac
Events through Combination Therapy in Patients
Living with Systolic Hypertension (ACCOMPLISH)
trials,4,5 whereas others simply focus on blood
pressure control irrespective of particular drug
classes.6–10 The trials on which the guidelines are based
have typically involved follow-up durations of about
5 years.

www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8 1

 Articles
Research in context
Evidence before this study
We searched PubMed from Jan 1, 1990, to May 1, 2018, using
the search terms “long term outcomes” or “long term benefits”
for clinical trials involving blood pressure-lowering treatment or
lipid-lowering with statins. Randomised clinical trials were only
included if they reported morbidity or mortality outcomes. For
blood pressure, most trials were placebo-controlled and, judging
from the single meta-analysis reported so far, the majority were
not trials in patients with hypertension (more commonly trials in
patients with heart failure and ischaemic heart disease). These
trials showed that during a post-trial follow-up of, on average,
41 months, there was 15% reduction in all-cause mortality
among patients originally assigned to active drug treatment.
To date, the only other trial to have studied long-term outcomes
following blood pressure treatment with two or more active
treatment groups is ALLHAT. In the case of ALLHAT, the in-trial
benefits apparent with the use of chlorthalidone compared with
amlodipine and lisinopril, respectively, were no longer evident in
the long-term follow-up. A meta-analysis of eight trials of lipid-
lowering drugs, including five with statins, reported long-term
follow up outcomes, with a significant overall benefit for
patients assigned active treatment. Over an average follow-up of
71 months, all-cause mortality was lowered by 10% in patients
formerly assigned a statin compared with those assigned
placebo. Only two trials reported follow-up longer than 15 years.
Added value of this study
Our findings in patients with hypertension with no previous
history of a coronary event show the long-term beneficial
effects of antihypertensive treatment with a calcium channel
blocker-based treatment regimen and lipid-lowering with a
statin. In particular, assignment to amlodipine-based treatment
(with perindopril added as required) was associated with a
Previously, hypertension trials comparing active drug
treatment with placebo, which have reported substantial
post-randomisation (in-trial) blood pressure differences
between the two groups, have linked active treatment
with long-term benefits.11 In comparison, long-term post-
trial follow-up data from trials comparing active
treatments are sparse.12 It is uncertain, therefore, whether
more recent trials that compared active treatment
regimens and showed the benefits of a regimen based on
a calcium channel blocker and an angiotensin-converting
enzyme inhibitor3,4 would also show a long-lasting
beneficial effect. Several long-term follow-up studies of
placebo-controlled statin trials have been reported,13–15
which have found persistent benefits in patients
For the ASCOT website see previously assigned statin treatment, but none have
www.ascotstudy.org involved patients with hypertension who were also
assigned interventions with different anti­ hyper­tensive
strategies.
ASCOT was designed to compare two antihypertensive
treatment strategies and, using a factorial design, to also
compare atorvastatin with placebo.16 In this report, we
2 www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8
reduction in the number of stroke deaths throughout 16 years
of follow-up. Extrapolating our findings from data accrued
during the in-trial period, it seems that this benefit is not
related to blood pressure differences between treatment
regimens. Our findings also add further evidence for the
long-term benefits of statin therapy in reducing risk of death
from cardiovascular causes. This study is, to our knowledge,
the first to report that both blood pressure-lowering and
lipid-lowering treatments confer such long-term benefits.
Furthermore, our findings from the non-lipid-lowering arm of
the trial, a subgroup at higher cardiovascular risk at baseline,
showed that these patients also receive significant long-term
benefits from blood pressure-lowering therapies.
Implications of all the available evidence
The long-term or legacy benefits of assignment to statins during
clinical trials are now well established, with benefits persisting
for up to 20 years after the original trials. The mechanisms
underlying these observations remain unproven, but it is
possible that statin-induced plaque stabilisation occurs during
the initial trial, which confers the long-term benefit.
The long-term benefit of the calcium channel blocker treatment
regimen, which seems to be independent of mean blood
pressure levels reported in the original trial, suggests the
existence of additional mechanisms for long-term protection.
We can only speculate on the nature of these mechanisms, with
the most likely being the substantial effects that the calcium
channel blocker-based regimen had on blood pressure
variability, although other possibilities include the negative
effects the atenolol-based regimen had on patients’ metabolic
profiles. Overall, our findings support the view that
interventions on blood pressure and cholesterol are associated
with long-term benefits in terms of cardiovascular outcomes.
evaluated the mortality data from the cohort of patients
originally recruited into ASCOT from the UK (the
ASCOT Legacy study), roughly 16 years after entry into
the trial and 10 years after trial closure, to establish
whether assignment to either of the original blood
pressure- lowering regimens, or to atorvastatin compared
with placebo, conferred long-term benefits in terms of
all-cause and cause-specific mortality outcomes.
Methods
Study design and participants
The detailed ASCOT protocol, including study design,
conduct, and baseline characteristics has been published16
and further detailed information is available on the
ASCOT website). ASCOT was designed to compare two
antihypertensive treatment strategies in a blood pressure-
lowering arm (BPLA): amlodipine, to which perindopril
was added as necessary (amlodipine-based treatment),
and atenolol, to which bendroflumethiazide was added
as necessary (atenolol-based treatment). Furthermore,
using a 2 × 2 factorial design, the trial also compared
 Articles

atorvastatin and placebo in a lipid-lowering arm (LLA)

consisting of patients with total cholesterol less than
6·5 mmol/L who were not currently taking a statin or Blood pressure-lowering arm Lipid-lowering arm (n=4605)
(n=8580)
a fibrate.
The study population consisted of men and women Amlodipine Atenolol Atorvastatin Placebo
(n=4305) (n=4275) (n=2317) (n=2288)
with hypertension who were aged 40–79 years at
randomisation and had at least three additional risk Age (years) 64 (8) 64 (8) 64 (8) 64 (8)
factors for cardio­ vascular disease, but no history of Male sex 3492 (81·1%) 3468 (81·1%) 2016 (87·0%) 2004 (87·6%)
coronary heart disease events, currently treated angina, Ethnicity
or a cerebro­vascular event within 3 months of random­ White/European 3861 (89·7%) 3840 (89·8%) 2045 (88·3%) 2019 (88·2%)
isation. The primary outcome was non-fatal myocardial South Asian 130 (3·0%) 109 (2·5%) 72 (3·1%) 80 (3·5%)
infarction and fatal coronary heart disease. Patients East Asian 7 (0·2%) 3 (0·1%) 2 (0·1%) 2 (0·1%)
were originally recruited between Feb 18, 1998, and Mixed/other 85 (2·0%) 86 (2·0%) 36 (1·6%) 33 (1·4%)
May 26, 2000, mostly from family practices. In the Nordic African 222 (5·2%) 237 (5·5%) 162 (7·0%) 154 (6·7%)
countries, individual patients from 686 family practices Socioeconomic status (age at leaving full-time education)*
were enrolled and entered into randomisation, whereas 12–14 years 1282 (30·0%) 1272 (29·6%) 682 (29·8%) 658 (28·4%)
in the UK and Ireland, most patients who underwent 15–16 years 2091 (48·9%) 2165 (50·3%) 1121 (49·0%) 1119 (48·3%)
randomisation were referred from family practices to 17–18 years 484 (11·3%) 465 (10·8%) 245 (10·7%) 287 (12·4%)
32 regional study centres. In all, 19  257 patients >18 years 416 (9·7%) 400 (9·3%) 239 (10·5%) 252 (10·9%)
were randomised to the BPLA, of whom 10 240 were Data missing 2 3 1 1
randomised to the LLA. There were 9017 patients in the Body-mass index (kg/m²) 28·9 (4·7) 28·9 (4·6) 28·8 (4·9) 28·8 (4·6)
non-LLA group, of whom, about a third were on previous Current smoker 1035 (24·0%) 1006 (23·5%) 547 (23·6%) 541 (23·6%)
lipid-lowering or aspirin therapy. Alcohol status
In late 2002 and after a median follow-up of 3·3 years, Non-drinker 1088 (25·3%) 1089 (25·5%) 574 (24·8%) 571 (25·0%)
the LLA was stopped prematurely17 at the recommendation 1–13 units per week 1816 (42·2%) 1831 (42·8%) 1010 (43·6%) 983 (43·0%)
of the data safety monitoring board because of substantial ≥14 units per week 1401 (32·5%) 1355 (31·7%) 733 (31·6%) 734 (32·1%)
benefits of atorvastatin on the primary endpoint. Patients Systolic blood pressure 162 (18) 162 (17) 162 (17) 162 (18)
in the LLA continued to be followed up until the end of (mm Hg)
the BPLA. During that period, these patients were offered Diastolic blood pressure 92 (10) 92 (10) 92 (10) 93 (10)
open-label atorvastatin in addition to their assigned blood (mm Hg)
pressure-lowering treatment. A similar number of Heart rate (beats per minute) 71 (13) 71 (12) 70 (12) 71 (13)
patients (approximately two-thirds), formerly assigned Total cholesterol (mmol/L) 5·9 (1·1) 5·9 (1·1) 5·5 (0·8) 5·5 (0·8)
either atorvastatin or placebo, went on to receive a statin HDL-cholesterol (mmol/L) 1·3 (0·4) 1·3 (0·4) 1·3 (0·3) 1·3 (0·3)
for the 2·2 years of the LLA extension. LDL-cholesterol (mmol/L) 3·8 (1.0) 3·8 (1.0) 3·5 (0.7) 3·5 (0.8)
The BPLA was also prematurely stopped at the Serum triglycerides (mmol/L) 1·6 (1·2–2·3) 1·6 (1·2–2·3) 1·4 (1·0–2·0) 1·4 (1·1–2·0)
recommendation of the data safety monitoring board, Fasting plasma glucose 5·6 (5·1–6·6) 5·6 (5·1–6·6) 5·6 (5·1–6·5) 5·6 (5·1–6·6)
mainly because of the significantly higher mortality (mmol/L)
among patients allocated atenolol-based treatment than Serum creatinine (µmol/L) 99 (89–109) 98 (89–109) 99 (90–109) 99 (90–109)
among those on amlodipine-based treatment. Database Presence of diabetes mellitus 1139 (26·5%) 1145 (26·8%) 621 (26·8%) 630 (27·5%)
lock was in June, 2005, with the last follow-up of patients Number of cardiovascular risk factors
ranging from December, 2004, to April, 2005.3 Median 3† 2055 (47·7%) 2044 (47·8%) 1201 (51·8%) 1141 (49·9%)
follow-up in the BPLA was 5·5 years. 4 1416 (32·9%) 1417 (33·1%) 716 (30·9%) 746 (32·6%)
The study conformed to Good Clinical Practice ≥5 834 (19·4%) 814 (19·0%) 400 (17·3%) 401 (17·5%)
guidelines and the Declaration of Helsinki. The protocol History of stroke or transient 507 (11·8%) 492 (11·5%) 233 (10·1%) 239 (10·4%)
and all subsequent amendments were reviewed and ischaemic attack (>3 months
ago)
ratified by central and regional ethics review boards in the
History of peripheral vascular 359 (8·3%) 383 (9·0%) 160 (6·9%) 150 (6·6%)
UK and by national ethics and statutory bodies in Ireland, disease
Sweden, Denmark, Iceland, Norway, and Finland. Presence of atrial fibrillation 60 (1·4%) 60 (1·4%) 36 (1·6%) 32 (1·4%)
Previous antihypertensive 3961 (92·0%) 3924 (91·8%) 2118 (91·4%) 2106 (92·0%)
ASCOT Legacy cohort treatment
The ASCOT Legacy cohort consisted of all 8580 ASCOT Previous lipid-lowering 490 (11·4%) 478 (11·2%) 29 (1·3%) 22 (1·0%)
trial participants from the UK. These patients treatment
were followed up until the end of the BPLA, during Previous aspirin use 1083 (25·2%) 1040 (24·3%) 533 (23·0%) 519 (22·7%)
which period 717 died. Of the remaining patients, Data are n (%), mean (SD), or median (IQR). *Socioeconomic status was missing for five patients. †Including 37 who
7302 (ie, patients from all but two trial sites in the UK had two risk factors only.
from which patient consent for follow-up was not
Table 1: Baseline characteristics of patients in the ASCOT Legacy cohort
obtained) were flagged with the Office for National

www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8 3

 Articles
8580 patients from the UK enrolled in ASCOT-Legacy cohort
8580 randomly assigned treatment in the BPLA
4275 assigned to atenolol-based
treatment
1640 died overall
1969 in non-LLA group 2306 enrolled in LLA
4605 randomly assigned treatment
in the LLA
2288 assigned to
placebo
769 died 1200 alive 903 died 1385 alive
Figure 1: ASCOT-Legacy patient population stratified by treatment allocation
BPLA=blood pressure-lowering arm. LLA=lipid-lowering arm.
Statistics and the General Register Office for Scotland
for post-trial follow-up. In this report, we have included
all reported deaths up to and including Dec 31, 2015.
However, no data on morbidity and treat­ ment were
available after the end of the BPLA.
A team of two physicians (JM and AW) independently
adjudicated the cause of death, using prespecified criteria
that were consistent with the definitions used during the
trial period. In these analyses, we report on all-cause
mortality, and deaths from cardiovascular and non-
cardiovascular causes. All cardiovascular deaths were
further adjudicated to report on deaths due to coronary
heart disease or stroke. Similarly, non-cardiovascular
deaths were sub-categorised to report on cancer-related
deaths.
Statistical methods
All analyses were done in accordance with the intention-to-
treat principle, and thus in-trial follow-up is included for
the 561 patients from two UK sites who were not flagged
after the closure of the BPLA, and those who dropped out
of the BPLA early. For patients who were still alive,
censoring was defined as the end of follow-up (Dec 31, 2015)
or the end of the BPLA for those who did not consent to
long-term follow-up. The end of the LLA period was
defined as Oct 1, 2002, and the end of the BPLA was the
last follow-up before the database lock in June, 2005.
4 www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8
4305 assigned to amlodipine-based
treatment
1642 died overall
2299 enrolled in LLA 2006 in non-LLA group
2317 assigned to
atorvastatin
865 died 1452 alive 745 died 1261 alive
For both the BPLA and LLA, and for each death
outcome, separate Cox proportional hazards models
were developed to estimate hazard ratios (HRs) and
95% CIs comparing treatment groups. Both unadjusted
and adjusted analyses were done. For the analyses of
each cause-specific death, all deaths from other causes
were handled as censoring events. We adjusted for the
following prespecified covariates at baseline: age, sex,
ethnicity, age at leaving full-time education (reflecting
socio-economic status), body-mass index (BMI), systolic
blood pressure, total cholesterol, presence of diabetes,
smoking history, and the other treatment comparison
(for example, for the comparisons of the treatment
groups in the BPLA, we adjusted for the allocation to
statin or a placebo, with a dummy variable for those in
the non-LLA group).
For each Cox model, we tested the assumption of
proportionality using Schoenfield’s residuals,18 and we
found no evidence of any deviation. We prespecified
tests for interactions between the two treatment
comparisons: ie, between blood pressure lowering
treatment regimens and between statin and placebo.
Tests for interactions were also done to establish
whether the effect of the two blood pressure-lowering
treatments differed between subgroups such as those
based on the presence of diabetes, age, or allocation to
the LLA or not.
 Articles

In-trial period (median follow-up 5·5 years) Post-BPLA period (median follow-up 10·7 years) Total follow-up (median follow-up 15·7 years)
Atenolol-based Amlodipine-based Atenolol-based Amlodipine-based Atenolol-based Amlodipine-based
treatment (n=4275) treatment (n=4305) treatment (n=3613) treatment (n=3688) treatment (n=4275) treatment (n=4305)
Deaths Incidence Deaths Incidence Deaths Incidence Deaths Incidence Deaths Incidence Deaths Incidence
per 100 per 100 per 100 per 100 per 100 per 100
person-years person-years person-years person-years person-years person-years
All-cause mortality 370 1·62 347 1·50 1270 3·97 1295 3·98 1640 2·99 1642 2·95
Cardiovascular causes 149 0·65 115 0·50 474 1·48 472 1·45 623 1·13 587 1·05
Coronary heart disease 86 0·38 66 0·29 127 0·40 132 0·41 213 0·39 198 0·36
Stroke 30 0·13 21 0·09 69 0·22 51 0·16 99 0·18 72 0·13
Non-cardiovascular causes 221 0·97 232 1.00 796 2·49 823 2·53 1017 1·85 1055 1·90
Cancer 135 0·59 146 0·63 440 1·37 451 1·39 575 1·05 597 1·07
The atenolol-based regimen consisted of atenolol, with thiazide diuretic added as required, and the amlodipine-based regimen consisted of amlodipine, with perindopril added as required. BPLA=blood
pressure-lowering arm.

Table 2: All-cause and cause-specific mortality in the BPLA of the ASCOT Legacy cohort

Statistical tests were two-sided, with p values less than
0·05 considered statistically significant. All statistical
analysis was done using Stata 15.
Role of the funding source
The funder of the study had no role in study design, data
collection, data analysis, data interpretation, or writing of the
report. The corresponding author had full access to all the
data in the study and had final responsibility for the decision
to submit for publication.
Results
The ASCOT Legacy cohort consisted of 8580 patients
from the UK, with a mean age at baseline of 64·1 years
(SD 8). Baseline characteristics of these patients
were similar to those from Sweden, Denmark, Iceland,
Norway, and Finland,16 except that patients in the UK
ASCOT Legacy cohort were more ethnically diverse (10%
non-white participants vs 1%), more were men (81·1% vs
72·9%), and fewer were current smokers (23·8% vs
36·1%).
Table 1 shows the baseline characteristics of patients
in the ASCOT Legacy cohort. In the BPLA, 4305 pat­
ients were assigned to amlodipine-based treatment
and 4275 were assigned to atenolol-based treatment.
4605 patients were also included in the LLA, with
2317 ran­domly assigned to atorvastatin and 2288 to
placebo. The remaining 3975 patients comprised the
non-LLA group. The baseline characteristics of patients
in the LLA and non-LLA groups are shown in the
appendix. Compared with the LLA, the non-LLA group
contained more women, more patients with more than
four cardiovascular risk factors at baseline, and more
patients with a history of previous use of lipid-lowering
therapy or previous vascular disease, as well as having
higher mean baseline total-cholesterol and LDL-
cholesterol.
Figure 1 shows the numbers of patients and deaths in
the ASCOT Legacy cohort by treatment allocation.
During a median follow-up of 15·7 years (IQR 9·7–16·4),
3282 (38·3%) of all patients died, including 1640 (38·4%)
of 4275 allocated atenolol-based treatment and 1642
(38·1%) of 4305 allocated amlodipine-based treatment.
1768 (53·9%) of all deaths occurred in patients in the
LLA, including 903 (39·5%) of 2288 allocated placebo
and 865 (37·3%) of 2317 allocated atorvastatin. In total,
1210 (36·9%) deaths were from cardiovascular-related
causes.
Table 2 shows the number of events and incidence
rates (per 100 person-years) for total and cause-specific
mortality with the two blood pressure-lowering treatment
regimens for the in-trial period, post-trial period, and
throughout all follow-up. During the in-trial period
(median 5·5 years follow-up), there were more deaths
among patients on atenolol-based treatment than among
those on amlodipine-based treatment for all-causes
and cardiovascular causes, including the coronary heart
disease and stroke com­ ponents of the cardiovascular
mortality. However, during the post-trial period following
the closure of the BPLA (an extra 10·7 years of median
follow-up), the only treatment differences in mortality
were for stroke mortality, for which the differential in the
event rates between patients on the two treatment
regimens persisted.
Table 3 and figure 2 show that overall (during a median
follow-up of 15·7 years), in the BPLA, there was no
significant difference in all-cause mortality between the
treatments (HR 0·97, 95% CI 0·90–1·04, p=0·3411).
The effect of amlodipine-based versus atenolol-based
treatment on cardiovascular death decreased between the See Online for appendix
in-trial period (0·74, 0·58–0·95, p=0·0177) and total
follow-up (0·90, 0·81–1·01, p=0·0776). However, for
stroke mortality, the effect of amlodipine-based versus
atenolol-based treatment during the in-trial period (HR
0·69, 0·40–1·21, p=0·2013) was similar to that at the end
of the follow-up (HR 0·71, 0·53–0·97, p=0·0305),
becoming statistically significant with the accrual of
more events.

www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8 5

 Articles

In-trial period Total follow-up

Unadjusted Adjusted* Unadjusted Adjusted*
HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value
All-cause mortality 0·93 (0·80–1·07) 0·3130 0·91 (0·78–1·05) 0·2012 0·99 (0·92–1·06) 0·6722 0·97 (0·90–1·04) 0·3411
Cardiovascular causes 0·76 (0·60–0·97) 0·0302 0·74 (0·58–0·95) 0·0177 0·93 (0·83–1·04) 0·1909 0·90 (0·81–1·01) 0·0776
Coronary heart disease 0·76 (0·55–1·05) 0·0930 0·74 (0·53–1·02) 0·0625 0·92 (0·76–1·11) 0·3786 0·88 (0·73–1·07) 0·2148
Stroke 0·69 (0·40–1·21) 0·1969 0·69 (0·40–1·21) 0·2013 0·72 (0·53–0·97) 0·0316 0·71 (0·53–0·97) 0·0305
Non-cardiovascular causes 1·04 (0·86–1·25) 0·6965 1·02 (0·85–1·23) 0·8292 1·02 (0·94–1·11) 0·6403 1·01 (0·92–1·10) 0·8880
Cancer 1·07 (0·85–1·35) 0·5720 1·06 (0·84–1·34) 0·6101 1·02 (0·91–1·15) 0·7090 1·01 (0·90–1·14) 0·8304
The atenolol-based regimen consisted of atenolol, with thiazide diuretic added as required, and the amlodipine-based regimen consisted of amlodipine, with perindopril
added as required. HR=hazard ratio. *Adjusted for baseline age, sex, ethnicity, systolic blood pressure, total cholesterol, body-mass index, diabetes, smoking status, years of
education (socioeconomic status), randomisation in the lipid-lowering arm (placebo or atorvastatin) or not.

Table 3: All-cause and cause-specific mortality with amlodipine-based treatment versus atenolol-based treatment in the blood pressure-lowering arm of
the ASCOT Legacy cohort

A All-cause death B Cardiovascular death

0·6 Atenolol 0·2 HR 0·90 (95% CI 0·81–1·01), p=0·08
Amlodipine
0·5 HR 0·97 (95% CI 0·90–1·04), p=0·34
Cumulative incidence

0·4

0·3 0·1

0·2

0·1

0 0
0 3 6 9 12 15 0 3 6 9 12 15
Number at risk
Atenolol 4275 4107 3617 3292 2784 2374 4275 4107 3617 3292 2784 2374
Amlodipine 4305 4149 3692 3324 2843 2423 4305 4149 3692 3324 2843 2423

C Coronary heart disease death D Stroke death

0·06 HR 0·88 (95% CI 0·73–1·07), p=0·22 0·03 HR 0·71 (95% CI 0·53–0·97), p=0·0305
Cumulative incidence

0·04 0·02

0·02 0·01

0 0
0 3 6 9 12 15 0 3 6 9 12 15
Time since randomisation (years) Time since randomisation (years)
Number at risk
Atenolol 4275 4107 3617 3292 2784 2374 4275 4107 3617 3292 2784 2374
Amlodipine 4305 4149 3692 3324 2843 2423 4305 4149 3692 3324 2843 2423

Figure 2: Cumulative incidence of all-cause and cause-specific death in the BPLA of the ASCOT Legacy cohort
Risk of death from all-causes (A), cardiovascular causes (B), coronary heart disease (C), and stroke (D) is shown among patients in the BPLA allocated amlodipine-based
treatment versus those allocated atenolol-based treatment during the 16 year follow-up of the ASCOT-Legacy cohort. BPLA=blood pressure-lowering arm.

There was no interaction between treatment allocation
in the BPLA and the LLA (appendix). Other subgroup
analyses provided no evidence for treatment interactions
with either age or diabetes status at baseline (data not
shown). However, there were differences in the effect of
the two blood pressure treatment regimens, depending on
whether a patient was allocated to the LLA or not (p=0·0220
for interaction). Table 4 shows the numbers of events and
incidence rates for cause-specific deaths in the two treatment
groups, stratified by allocation to the LLA. In the non-LLA

6 www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8

 Articles

Atenolol-based treatment Amlodipine-based treatment Adjusted hazard p value pinteraction†

ratio (95% CI)*
Deaths Incidence per 100 Deaths Incidence per 100
person-years person-years
All-cause mortality
Non-LLA group 769 3·05 745 2·84 0·91 (0·83–1·01) 0·0784 ··
LLA 871 2·93 897 3·05 1·02 (0·93–1·12) 0·7093 0·1222
Cardiovascular causes
Non-LLA group 325 1·29 275 1·05 0·79 (0·67–0·93) 0·0046 ··
LLA 298 1·00 312 1·06 1·03 (0·88–1·21) 0·6999 0·0220
Coronary heart disease
Non-LLA group 125 0·50 102 0·39 0·76 (0·59–0·99) 0·0439 ··
LLA 88 0·30 96 0·33 1·06 (0·80–1·42) 0·6706 0·0952
Stroke
Non-LLA group 49 0·19 34 0·13 0·67 (0·43–1·04) 0·0751 ··
LLA 50 0·17 38 0·13 0·76 (0·50–1·16) 0·1977 0·6982
Non-cardiovascular causes
Non-LLA group 444 1·76 470 1·79 1·00 (0·88–1·14) 0·9803 ··
LLA 573 1·93 585 1·99 1·01 (0·90–1·13) 0·8569 0·9193
Cancer
Non-LLA group 247 0·98 261 1·00 1·01 (0·85–1·20) 0·9436 ··
LLA 328 1·11 336 1·14 1·02 (0·87–1·19) 0·8163 0·9207
The atenolol-based regimen consisted of atenolol, with thiazide diuretic added as required, and the amlodipine-based regimen consisted of amlodipine, with perindopril
added as required. The non-LLA group consisted of all patients not included in the LLA because of high baseline total cholesterol. LLA=lipid-lowering arm. *Adjusted for age,
sex, ethnicity, socioeconomic status, body-mass index, systolic blood pressure, total cholesterol, presence of diabetes, and smoking history. †p value from test for interaction
in adjusted models.

Table 4: All-cause and cause-specific mortality in patients assigned to amlodipine-based or atenolol-based treatment, by the inclusion in the LLA

In-trial period (median follow-up 3·3 years) Post-LLA period (median follow-up 13·2 years) Total follow-up (median follow-up, 15·7 years)
Placebo (n=2288) Atorvastatin (n=2317) Placebo (n=2198) Atorvastatin (n=2234) Placebo (n=2288) Atorvastatin (n=2317)
Deaths Incidence Deaths Incidence Deaths Incidence Deaths Incidence Deaths Incidence Deaths Incidence
per 100 per 100 per 100 per 100 per 100 per 100
person-years person-years person-years person-years person-years person-years
All-cause mortality 90 1·28 83 1·18 813 3·67 782 3·42 903 3·09 865 2·89
Cardiovascular causes 36 0·51 30 0·43 289 1·30 255 1·11 325 1·11 285 0·95
Coronary heart disease 19 0·27 19 0·27 84 0·38 62 0·27 103 0·35 81 0·27
Stroke 8 0·11 6 0.09 35 0·16 39 0·17 43 0·15 45 0·15
Non-cardiovascular causes 54 0·77 53 0·75 524 2·36 527 2·30 578 1·98 580 1·94
Cancer 37 0·53 38 0·54 297 1·34 292 1·28 334 1·14 330 1·10
LLA=lipid-lowering arm.

Table 5: All-cause and cause-specific mortality in the LLA of the ASCOT Legacy cohort

group, compared with patients receiving atenolol-based
treatment, those allocated to amlodipine-based treatment
had significantly fewer cardiovascular deaths (adjusted
HR 0·79, 95% CI 0·67–0·93, p=0·0046) and coronary
heart disease deaths (adjusted HR 0·76, 0·59–0·93,
p=0·0439), and numer­ically, but not significantly, fewer
stroke deaths (adjusted HR 0·67, 0·43–1·04, p=0·0751;
Kaplan-Meier plots in the LLA and non-LLA groups are
available in the appendix).
Table 5 shows the numbers of events and incidence
rates for cause-specific mortality among patients assigned
to atorvastatin or a placebo in the LLA of the ASCOT
Legacy cohort, during the in-trial period, post-trial period,
and throughout all follow-up.
Table 6 shows estimates for mortality differences
between atorvastatin and placebo in the LLA of the
ASCOT Legacy cohort during the in-trial period and
overall follow-up. Significantly fewer patients assigned to
atorvastatin died from cardiovascular causes (HR 0·85,
95% CI 0·72–0·99, p=0·0395) than did those assigned to
placebo. Although fewer all cause and coronary heart
disease deaths occurred with atorvastatin compared with

www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8 7

 Articles

In-trial period Total follow-up

Unadjusted Adjusted* Unadjusted Adjusted*
HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value HR (95% CI) p value
All-cause mortality 0·92 (0·68–1·24) 0·5970 0·93 (0·69–1·25) 0·6379 0·93 (0·85–1·02) 0·1211 0·92 (0·84–1·01) 0·0913
Cardiovascular causes 0·83 (0·51–1·35) 0·4483 0·85 (0·52–1·38) 0·5128 0·85 (0·73–1·00) 0·0459 0·85 (0·72–0·99) 0·0395
Coronary heart disease 0·99 (0·53–1·87) 0·9805 1·02 (0·54–1·92) 0·9594 0·77 (0·57–1·03) 0·0735 0·78 (0·58–1·04) 0·0884
Stroke 0·75 (0·26–2·17) 0·5995 0·80 (0·27–2·32) 0·6774 1·02 (0·67–1·55) 0·9353 1·02 (0·67–1·55) 0·9238
Non-cardiovascular causes 0·99 (0·67–1·44) 0·9393 0·98 (0·67–1·43) 0·9227 0·97 (0·87–1·09) 0·6420 0·96 (0·86–1·08) 0·5440
Cancer 1·03 (0·65–1·62) 0·9011 1·01 (0·64–1·59) 0·9598 0·96 (0·82–1·12) 0·5932 0·95 (0·82–1·11) 0·5023
*Adjusted for age, sex, ethnicity, systolic blood pressure, total cholesterol, body-mass index, diabetes at baseline, smoking status, years of education (socioeconomic status),
and randomisation to the blood pressure-lowering treatment.

Table 6: All-cause and cause-specific mortality with atorvastatin versus placebo in the lipid-lowering arm of the ASCOT Legacy cohort

A All-cause death B Cardiovascular death

0·5 Placebo 0·2 HR 0·85 (95% CI 0·72–0·99), p=0·0395
Atorvastatin
0·4 HR 0·92 (95% CI 0·84–1·01), p=0·09
Cumulative incidence

0·3
0·1
0·2

0·1

0 0
0 3 6 9 12 15 0 3 6 9 12 15
Number at risk
Placebo 2288 2206 1949 1746 1472 1251 2288 2206 1949 1746 1472 1251
Atorvastatin 2317 2236 1975 1793 1529 1306 2317 2236 1975 1793 1529 1306

C Coronary heart disease death D Stroke death

0·06 HR 0·78 (95% CI 0·58–1·04), p=0·09 0·03 HR 1·02 (95% CI 0·67–1·55), p=0·92
Cumulative incidence

0·04 0·02

0·02 0·01

0 0
0 3 6 9 12 15 0 3 6 9 12 15
Time since randomisation (years) Time since randomisation (years)
Number at risk
Placebo 2288 2206 1949 1746 1472 1251 2288 2206 1949 1746 1472 1251
Atorvastatin 2317 2236 1975 1793 1529 1306 2317 2236 1975 1793 1529 1306

Figure 3: Cumulative incidence of all-cause and cause-specific death in the LLA of the ASCOT Legacy
Risk of death from all-causes (A), cardiovascular causes (B), coronary heart disease (C), and stroke (D) is shown among patients in the LLA allocated atorvastatin
versus those allocated placebo during the 16 year follow-up of the ASCOT Legacy cohort. LLA=lipid-lowering arm.

placebo, differences were not significant (all cause Discussion

mortality 0·92, 0·84–1·01, p=0·0913 and coronary heart Our findings from patients with hypertension and no
disease mortality 0·78, 0·58–1·04, p=0.0884; figure 3). previous coronary events show the long-term benefits
However, there was no evidence of a treatment difference of antihypertensive treatment with a calcium channel
in terms of stroke deaths (figure 4; appendix). Subgroup blocker-based regimen and lipid lowering with a statin.
analyses by age or presence of diabetes at baseline In particular, assignment to amlodipine-based treatment
showed no effect modification (data not shown). (with perindopril added as required) was associated

8 www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8

 Articles

with fewer stroke deaths throughout 16 years of follow-

A
up (figure 2). Our results also support the long-term Adjusted HRs (95% CI)
benefits of statin therapy in reducing risk of death from
cardiovascular causes.13–15,19,20 This study is the first to Cardiovascular causes 0·90 (0·81–1·01), p=0·08
Coronary heart disease 0·88 (0·73–1·07), p=0·21
report that both blood pressure-lowering and lipid-
Stroke 0·71 (0·53–0·97), p=0·0305
lowering treatments confer such long-term benefits.
Non-cardiovascular causes 1·01 (0·92–1·10), p=0·89
Furthermore, findings for the higher risk subgroup, the
Cancer 1·01 (0·90–1·14), p=0·83
non-LLA group, supported the long-term benefits of All causes 0·97 (0·90–1·04), p=0·34
blood pressure lowering therapies in such patients.
Additionally, the benefits conferred differed between the 0·50 0·75 1·0 1·25 1·50

blood pressure lowering regimens, despite the in-trial
gain in blood pressure control being similar (appendix)
The only other large trial to have studied cardiovascular
outcomes with both antihypertensive therapy and lipid-
lowering with statins, the Antihypertensive and Lipid
Lowering to prevent Heart Attacks Trial (ALLHAT),21,22
compared different monotherapies with similar add-on
antihypertensive treatments. Both ASCOT and ALLHAT
assessed the potential benefits of statins in patients with
hypertension. In ASCOT, atorvastatin was compared
with placebo, whereas in ALLHAT, pravastatin was
compared with usual care. Unfortunately, in ALLHAT,
many patients in the usual care group received statins
and only a small difference in cholesterol was detected
between the treatment groups, which resulted in the
trial being underpowered to compare effects on major
cardiovascular endpoints.22,23 The findings from the
extended follow-up of patients assigned to three arms
of the antihypertensive study, the chlorthalidone,
amlodipine, and lisinopril arms, were mixed.24,25 In-trial
benefits that were apparent with the use of chlorthalidone
versus amlodipine and lisinopril, respectively, were no
longer evident in the long-term follow-up.
In long-term follow-up of trials in patients with hyper­
tension11 that compared active treatment with placebo, and
in which blood pressure differences were associated with
substantial reductions in cardiovascular events, a legacy or
carryover effect has been seen in the post-trial period,26–29
with long-term reductions—on average 9%—in mortality
in the groups previously receiving active treatment.11
However, in ASCOT, which compared different active
(alternative) treatment strategies,16 the long-term benefits
associated with the amlodipine-based regimen could not
be attributed to earlier differences in blood pressures
during the trial.30 First, as previously reported with the
main trial outcome3 and in a post-hoc analysis,30 there was
only a small difference in blood pressure of 2·9/1·8 mm Hg
between the two treatment groups. Furthermore, in the
ASCOT UK Legacy popu­lation, overall mean differences
between blood pressure-lowering regimens in blood
pressure recorded during the trial were only around
1·2/1·6 mm Hg (appendix). This small difference in
blood pressure cannot account for the sustained and
significant differences in stroke mortality apparent in this
long-term follow-up.
Initially, we attributed the benefits of the amlodipine-
based treatment to differences in the metabolic profile,
Favours amlodipine-based Favours atenolol-based
treatment treatment
B
Adjusted HRs (95% CI)
Cardiovascular causes 0·85 (0·72–0·99), p=0·0395
Coronary heart disease 0·78 (0·58–1·04), p=0·09
Stroke 1·02 (0·67–1·55), p=0·92
Non-cardiovascular causes 0·96 (0·86–1·08), p=0·54
Cancer 0·95 (0·82–1·11), p=0·50
All causes 0·92 (0·84–1·01), p=0·09
0·50 0·75 1·0 1·25 1·50
Favours atorvastatin Favours placebo
Figure 4: Risk of all-cause and cause-specific mortality among patients in the BPLA (A) and the LLA (B) of the
ASCOT-Legacy cohort
BPLA=blood pressure-lowering arm. LLA=lipid-lowering arm. HR=hazard ratio.
including adverse effects of glycaemia associated with the
atenolol-based therapy, and other differences including
small changes in lipids and electrolytes.30 However, these
changes alone were insufficient to explain the observed in-
trial effects on mortality and cardiovascular events.30 We
have subsequently described other important differences
between the two treatment arms, with atenolol-based
treatment lowering central aortic pressure sub­ stantially
less than amlodipine-based treatment.31 Additionally, blood
pressure variability was a major determinant of cardio­
vascular outcome in ASCOT,32 and was reduced to a much
greater extent with the amlodipine-based treatment
compared with the atenolol-based regimen.33
We believe that these mechanisms, and potentially
others that are currently unknown, are possible
explanations for the findings in ASCOT-BPLA, and that
they also contribute to the long-term benefits we identified
in the ASCOT Legacy population. We are in the process
of investigating these mechanisms further, together
with studies on potential genetic and other biomarker
predictors of cardiovascular events, which might explain
the differences we saw during this 16 year follow-up.
For the blood pressure trial, the most striking treatment
difference between regimens was in stroke death, which,
according to our earlier reports,33 was closely associated
with blood pressure variability. Our observations also
show that, although there was a significant reduction in
cardiovascular mortality with amlodipine versus atenolol
during the trial, this effect was attenuated in subsequent
follow-up, except for stroke mortality, where the beneficial

www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8 9

 Articles
effect remained essentially the same even 10 years after
trial closure. These dilutional effects might reflect that,
after the trial, both groups of patients would have received
similar overall treatment strategies. Nevertheless, the
persistence of the effect size for stroke deaths might
reflect the close relationship that this outcome has with
blood pressure treatment.34
We also observed a differential legacy effect between the
two treatment groups, according to the baseline risk of the
patients. In our evaluation of event rates between patients
included in the LLA and those in the non-LLA group
(table 4), among those assigned amlodipine-based
treatment, there was little difference in stroke mortality
and cardiovascular mortality and very small differences in
coronary heart disease mortality. These findings suggest
that amlodipine-based treatment confers a similar effect
irrespective of baseline risk. However, for patients
assigned atenolol-based treatment there was, compared
with those in the LLA, an increased event rate among
those in the non-LLA group.
Following the early stopping of the LLA, the beneficial
effects of atorvastatin on mortality persisted, with little
dilution of the effect during the long-term follow-up. In
the ASCOT-LLA Legacy cohort, the effect sizes of
cardiovascular and all-cause mortality at 13 years
remained similar to those at the end of the trial, but, with
an increasing number of events, the mortality benefits
became significant for cardiovascular deaths, with non-
significant suggestions of long-term benefit for all cause
and coronary heart disease mortality. Similar findings,
particularly the relationship between the use of a statin
and reduction in cardiovascular mortality in long-term
follow-up have been reported in other trials.13–15,19,20 The
mechanism for these durable and legacy effects remains
unclear, but it is possible that statin-induced plaque
stabilisation occurs during the initial trial period, which
confers a long term benefit on cardiovascular outcomes.
The present analyses have several limitations. After the
closure of the trial, we have no data on anti­hypertensive
and lipid-lowering medications, or indeed on any other
treatments. Thus, we cannot reliably ascertain what
differences, if any, in post-trial blood pressures and
treatments existed in the longer term. However, after the
closure of the LLA, a similar number of patients
previously assigned to atorvastatin or placebo received a
statin during the 2·2 year extension during final years of
the BPLA of the trial.35 Another limitation is that we have
no morbidity data after trial closure. The availability
of such data would allow for a more comprehensive
evaluation, and we are in the process of developing
appropriate registry linkages to acquire them. Further­
more, the generalisability of the study population could
be questioned. The patients in this cohort had a median
age of 64 years, they were hypertensive, and they had
several common risk factors, which would make them
representative of individuals of a similar age in the
community. Indeed, in previous unpublished work, we
10 www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8
have compared our population with the patients with
hypertension from the community, as reported in the
Health Survey of England, and we found them to be
similar. Lastly, our findings need to be replicated in other
studies. The greatest strength of this study is that it is the
first to report on a large cohort of patients with
hypertension and investigate both blood pressure-
lowering and lipid-lowering treatments and their effects
on long-term mortality, with substantial power to
evaluate differences in mortality between treatments.
The legacy outcomes from the ASCOT trial show the
long-term benefits of antihypertensive treatment with a
calcium channel blocker-based treatment regimen and
lipid-lowering with a statin, 16 years after entry into the
trial and more than 10 years after its closure. To our
knowledge, we report for the first time that the legacy
benefit from the amlodipine-based regimen in terms of
reducing risk of stroke mortality seems to be independent
of blood pressure levels achieved. Our data contribute to
evidence to support the long-term benefits of statins for
reducing cardiovascular mortality. Overall, our findings
support the notion that interventions for blood pressure
and cholesterol are associated with long-term benefits for
cardiovascular outcomes.
Contributors
AG and PS had the original idea for the Legacy study, designed the
protocol, supervised the data collection and statistical analyses, and
wrote the initial drafts of the manuscript. JM and AW did the
independent and blinded classification of all deaths and commented
on the manuscript. Statistical analyses were done by AG, TG, and TC.
SP provided statistical oversight and reviewed the manuscript.
NP contributed to protocol design and reviewed the manuscript.
Declaration of interests
AG reports support from the Foundation for Circulatory Health, during
the conduct of the study, and non-financial support from Servier, outside
the submitted work. NP reports grants from Pfizer and Servier, during
the conduct of the study; grants and other support from Servier for
serving on advisory boards, speaking at educational meetings, and
conducting the PREMIUM trial; and other support from AstraZeneca
and Napi for speaking at educational meetings, outside the submitted
work. Additionally, NP is President of the International Society of
Hypertension. PS reports grants and Personal fees from Pfizer and
Servier, during the conduct of the study, and grants and personal fees
from Pfizer, Servier, and Amgen, outside the submitted work. All other
authors declare no competing interests.
References
1 Krause T, Lovibond K, Caulfield M, McCormack T, Williams B.
Management of hypertension: summary of NICE guidance.
BMJ 2011; 343: d4891.
2 McManus RJ, Caulfield M, Williams B. NICE hypertension
guideline 2011: evidence based evolution. BMJ 2012; 344: e181.
3 Dahlof B, Sever PS, Poulter NR, et al. Prevention of cardiovascular
events with an antihypertensive regimen of amlodipine adding
perindopril as required versus atenolol adding bendroflumethiazide
as required, in the Anglo-Scandinavian Cardiac Outcomes
Trial-Blood Pressure Lowering Arm (ASCOT-BPLA): a multicentre
randomised controlled trial. Lancet 2005; 366: 895–906.
4 Jamerson K, Weber MA, Bakris GL, et al. Benazepril plus
amlodipine or hydrochlorothiazide for hypertension in high-risk
patients. N Engl J Med 2008; 359: 2417–28.
5 Bakris GL, Sarafidis PA, Weir MR, et al. Renal outcomes with
different fixed-dose combination therapies in patients with
hypertension at high risk for cardiovascular events (ACCOMPLISH):
a prespecified secondary analysis of a randomised controlled trial.
Lancet 2010; 375: 1173–81.

6 James PA, Oparil S, Carter BL, et al. 2014 evidence-based guideline
for the management of high blood pressure in adults: report from
the panel members appointed to the Eighth Joint National
Committee (JNC 8). JAMA 2014; 311: 507–20.
7 Mancia G, Fagard R, Narkiewicz K, et al. 2013 ESH/ESC guidelines
for the management of arterial hypertension. Eur Heart J 2013;
34: 2159–219.
8 Leung AA, Nerenberg K, Daskalopoulou SS, et al. Hypertension
Canada’s 2016 Canadian hypertension education program
guidelines for blood pressure measurement, diagnosis, assessment
of risk, prevention, and treatment of hypertension. Can J Cardiol
2016; 32: 569–88.
9 Stephan D, Gaertner S, Cordeanu EM. A critical appraisal of the
guidelines from France, the UK, Europe and the USA for the
management of hypertension in adults. Arch Cardiovasc Dis 2015;
108: 453–59.
10 Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/
ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the
prevention, detection, evaluation, and management of high blood
pressure in adults: a report of the American College of Cardiology/
American Heart Association Task Force on Clinical Practice
Guidelines. Hypertension 2018; 71: e13–115.
11 Kostis WJ, Thijs L, Richart T, Kostis JB, Staessen JA. Persistence of
mortality reduction after the end of randomized therapy in clinical
trials of blood pressure-lowering medications. Hypertension 2010;
56: 1060–68.
12 Chowdhury EK, Owen A, Ademi Z, et al. Short- and long-term
survival in treated elderly hypertensive patients with or without
diabetes: findings from the Second Australian National Blood
Pressure study. Am J Hypertens 2014; 27: 199–206.
13 Kostis WJ, Moreyra AE, Cheng JQ, Dobrzynski JM, Kostis JB.
Continuation of mortality reduction after the end of randomized
therapy in clinical trials of lipid-lowering therapy. J Clin Lipidol 2011;
5: 97–104.
14 Ford I, Murray H, McCowan C, Packard CJ. Long term safety and
efficacy of lowering LDL-cholesterol with statin therapy: 20-year
follow up of West of Scotland Coronary Prevention Study.
Circulation. 2016; 133: 1073–80
15 Hague WE, Simes J, Kirby A et al. Long-term effectiveness and
safety of pravastatin in patients with coronary heart disease: 16 years
of follow-up of the LIPID Study. Circulation 2016: 133: 1851–60.
16 Sever PS, Dahlöf B, Poulter NR, et al. Rationale, design, methods
and baseline demography of participants of the Anglo-Scandinavian
Cardiac Outcomes Trial. ASCOT investigators. J Hypertens 2001;
19: 1139–47.
17 Sever PS, Dahlöf B, Poulter NR, et al. Prevention of coronary and
stroke events with atorvastatin in hypertensive patients who have
average or lower-than-average cholesterol concentrations, in the
Anglo-Scandinavian Cardiac Outcomes Trial—Lipid Lowering Arm
(ASCOT-LLA): a multicentre randomised controlled trial.
Lancet 2003; 361: 1149–58.
18 Schoenfeld D. Partial residuals for the proportional hazards
regression-model. Biometrika 1982; 69: 239–41.
19 Strandberg TE, Pyorala K, Cook TJ, et al. Mortality and incidence of
cancer during 10-year follow-up of the Scandinavian Simvastatin
Survival Study (4S). Lancet 2004; 364: 771–77.
20 Heart Protection Study Collaborative Group. Effects on 11-year
mortality and morbidity of lowering LDL cholesterol with
simvastatin for about 5 years in 20 536 high-risk individuals:
a randomised controlled trial. Lancet 2011; 378: 2013–20.
www.thelancet.com Published online August 26, 2018 http://dx.doi.org/10.1016/S0140-6736(18)31776-8 11
Articles
21 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: 2981–97.
22 The ALLHAT Officers and Coordinators for the ALLHAT
Collaborative Research Group. Major outcomes in moderately
hypercholesterolemic, hypertensive patients randomized to
pravastatin vs usual care: the antihypertensive and lipid-lowering
treatment to prevent heart attack trial (ALLHAT-LLT). JAMA 2002;
288: 2998–3007.
23 Margolis KL, Davis BR, Baimbridge C, et al. Long-term follow-up of
moderately hypercholesterolemic hypertensive patients following
randomization to pravastatin vs usual care: the Antihypertensive
and Lipid-Lowering Treatment to Prevent Heart Attack Trial
(ALLHAT-LLT). J Clin Hypertens 2013; 15: 542–54.
24 Cushman WC, Davis BR, Pressel SL, et al. Mortality and
morbidity during and after the antihypertensive and
lipid-lowering treatment to prevent heart attack trial.
J Clin Hypertens 2012; 14: 20–31.
25 Yamal JM, Oparil S, Davis BR, et al. Stroke outcomes among
participants randomized to chlorthalidone, amlodipine or lisinopril
in ALLHAT. J Am Soc Hypertens 2014; 8: 808–19.
26 Shulman N, Tuttle E, Entwisle G, et al. Persistence of reduction in
blood pressure and mortality of participants in the hypertension
detection and follow-up program. JAMA 1988; 259: 2113–22.
27 Kostis JB, Wilson AC, Freudenberger RS, Cosgrove NM, Pressel SL,
Davis BR. Long-term effect of diuretic-based therapy on fatal
outcomes in subjects with isolated systolic hypertension with and
without diabetes. Am J Cardiol 2005; 95: 29–35.
28 Staessen JA, Thijisq L, Fagard R, et al. Effects of immediate versus
delayed antihypertensive therapy on outcome in the Systolic
Hypertension in Europe Trial. J Hypertens 2004; 22: 847–57.
29 Nelson MR, Chowdhury EK, Doust J, Reid CM, Wing LM. Ten-year
legacy effects of baseline blood pressure ‘treatment naivety’ in the
Second Australian National Blood Pressure study. J Hypertens 2015;
33: 2331–37.
30 Poulter NR, Wedel H, Dahlöf B, et al. Role of blood pressure
and other variables in the differential cardiovascular event rates noted
in the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure
Lowering Arm (ASCOT-BPLA). Lancet 2005; 366: 907–13.
31 Williams B, Lacy PS, Thom SM, et al. Differential impact of blood
pressure-lowering drugs on central aortic pressure and clinical
outcomes: principal results of the Conduit Artery Function
Evaluation (CAFE) study. Circulation 2006; 113: 1213–25.
32 Rothwell PM, Howard SC, Dolan E, et al. Prognostic significance of
visit-to-visit variability, maximum systolic blood pressure,
and episodic hypertension. Lancet 2010; 375: 895–905.
33 Rothwell PM, Howard SC, Dolan E, et al. Effects of beta blockers
and calcium-channel blockers on within-individual variability in
blood pressure and risk of stroke. Lancet Neurol 2010; 9: 469–80.
34 Ettehad D, Emdin CA, Kiran A, et al. Blood pressure lowering for
prevention of cardiovascular disease and death: a systematic review
and meta-analysis. Lancet 2016; 387: 957–67.
35 Sever PS, Poulter NR, Dahlöf B, et al. The Anglo-Scandinavian
Cardiac Outcomes Trial lipid lowering arm: extended observations
2 years after trial closure. Eur Heart J 2008; 29: 499–508.