Research
Effects of candesartan in acute stroke on vascular events during long-term
follow-up: results from the Scandinavian Candesartan Acute Stroke Trial (SCAST)
Astrid G. Hornslien1,2, Else C. Sandset3, Jannicke Igland4, Andreas Ternt5, Gudrun Boysen6,
Philip M. W. Bath7, Gordon D. Murray8, and Eivind Berge1*
Background Randomized-controlled trials have shown no
beneficial short-term effects of blood pressure lowering treat-
ment in the acute phase of stroke.
Aim We aimed to see whether blood pressure lowering treat-
ment with candesartan in the acute phase can lead to benefits
that become apparent over a longer period of follow-up.
Methods The Scandinavian Candesartan Acute Stoke Trial was
a randomized- and placebo-controlled trial of candesartan in
2,029 patients with acute stroke and systolic blood pres-
sure 140 mmHg. Trial treatment was given for seven-days,
and the primary follow-up period was six-months. We have
used the national patient registries and the cause of death
registries in the Scandinavian countries to collect data on vas-
cular events and deaths up to three-years from randomization.
The primary end-point was the composite of stroke, myocar-
dial infarction, or vascular death, and we used Cox propor-
tional hazards regression model for analysis.
Results Long-term data were available for 1,256 of the 1,286
patients (98%) from Scandinavia. The risk of the primary com-
posite end-point did not differ significantly between the
groups (candesartan 178/632 events, placebo 203/624 events,
hazard ratio = 087, 95% confidence interval 071107). very early treatment of hemorrhagic stroke (3), and some indica-
There were also no statistically significant differences for the
Correspondence: Eivind Berge*, Department of Internal Medicine, Oslo
University Hospital, Ullevl, P.O. Box 4956 Nydalen, NO-0424 Oslo,
Norway.
E-mail: eivind.berge@medisin.uio.no
1
Department of Internal Medicine, Oslo University Hospital, Oslo,
Norway
2 Cilexetil Therapy in Stroke Survivors (ACCESS) study showed
University of Oslo, Oslo, Norway
3
Department of Neurology, Oslo University Hospital, Oslo, Norway
4
Department of Global Public Health and Primary Care, University of
Bergen, Bergen, Norway
5
Department of Medical Sciences, Uppsala University, Uppsala, Sweden
6
Department of Neurology, Bispebjerg Hospital, and University of
Copenhagen, Copenhagen, Denmark
7
Stroke Trials Unit, Division of Clinical Neuroscience, University of Not-
tingham, Nottingham, UK
8
Centre for Population Health Sciences, University of Edinburgh, Edin-
burgh, UK
Received: 14 February 2015; Accepted: 6 January 2015; Published online
22 March 2015
Conflict of interest: AT has accepted support to his institution for aca-
demic trials from AstraZeneca. EB has received payment for lectures at
meetings arranged by AstraZeneca and accepted support for the trial to
his institution from AstraZeneca and Takeda. All other authors declare
that they have no conflicts of interest.
Trial Registration: URL: http://www.clinicaltrials.gov. Unique identifier:
NCT00120003.
Funding: The trial was funded by grants from the South-Eastern Norway
Regional Health Authority and Oslo University Hospital. AstraZeneca
supplied the study drugs, and AstraZeneca and Takeda supported the trial
with limited, unrestricted grants.
DOI: 10.1111/ijs.12477
830 Vol 10, August 2015, 830835
secondary end-points stroke and all-cause death, or in any of
the pre-specified subgroups.
Conclusions Treatment with candesartan in the acute phase of
stroke was not associated with clear long-term clinical ben-
efits. This result supports the conclusion from trials with short-
term follow-up, that blood pressure lowering treatment with
candesartan should not be given routinely to patients with
acute stroke and raised blood pressure.
Key words: acute stroke therapy, hypertension, blood pressure,
candesartan, long-term results
Introduction
High blood pressure is common in the acute phase of stroke and
is associated with poor outcome, but several randomized-
controlled trials have shown no beneficial effect of routine blood
pressure lowering treatment in this setting (1,2). There are,
however, strong indications that benefits can be obtained with
tions that the same is true for ischemic stroke (2,4,5), and new
trials are ongoing to see if benefits can be obtained in other
subgroups, or from other ways of blood pressure lowering in the
acute phase of stroke (6,7).
It is also possible that benefits from treatment will only become
apparent after long-term follow-up. The Acute Candesartan
that treatment with the angiotensin receptor blocker candesartan
in the acute phase was associated with a significantly reduced risk
of vascular events and deaths during one-years follow-up, and
that this risk reduction became more apparent with time (8). In
all other trials of blood pressure lowering in the acute phase of
stroke, the majority of patients have only been followed for three-
to six-months (1,2) and it is possible that this is too short. For
example, the Third International Stroke Trial (IST-3) showed that
an early increase in deaths from thrombolytic treatment was
offset by a similar-sized reduction in deaths at longer-term
follow-up (9). Also, in studies of secondary prevention,
angiotensin-receptor blockers have been shown to have long-term
neurovascular-protective effects beyond the effect of blood pres-
sure lowering, related to beneficial effects on endothelial, vascular
smooth muscle and neuronal cells (10), which may partly explain
the beneficial long-term effects of these agents (11,12).
The Scandinavian Candesartan Acute Stroke Trial (SCAST)
was designed to test whether the long-term benefits of candesar-
tan seen in the ACCESS study could be reproduced in a larger
trial, but found no such benefits after six-months follow-up (5).
Here we present the results of a pre-specified analysis of the effect
of treatment in patients who have been followed for an extended
period of three-years.
2015 World Stroke Organization
 A. G. Hornslien et al. Research
Methods other), systolic blood pressure, and Scandinavian Stroke Scale
score at baseline. A pre-specified subgroup analysis was done for
Study design and participants
the primary end-point. As a sensitivity analysis, we used ordinal
SCAST was a North European, multicenter, randomized- and
logistic regression for the total number of recurrent strokes at
placebo-controlled trial of candesartan in 2,029 patients present-
three-years (categorized as no stroke, non-fatal stroke, and fatal
ing within 30 h of acute ischemic or hemorrhagic stroke and with
stroke), to see if treatment affected stroke severity (21). We used
systolic blood pressure 140 mmHg. The study design and par-
SPSS Statistics version 20 for all analyses.
ticipants have been described elsewhere (5,13). In brief, patients
were randomly allocated to treatment with candesartan or
placebo for seven-days, with doses increasing from 4 to 16 mg Results
once daily during the first three-days. After the seven-days treat-
ment period, treatment was left to the clinicians discretion, but Long-term data were available from 1,256 patients of all 1,286
candesartan was the advised antihypertensive treatment and was Scandinavian patients in the trial (98%) (Fig. 1). For the last 78
provided free of charge for six-months. The primary follow-up Swedish patients and 18 Danish patients included in the trial, we
period was six-months, and the two co-primary effect variables did not have follow-up data for the full period of three-years,
were the composite end-point of stroke, myocardial infarction, or because of delays in the updating of the Swedish National Patient
vascular death during follow-up, and functional outcome at six- Registry and the Danish Cause of Death Registry. The median
months. Written consent was sought from all patients, and the follow-up time for these 96 patients was 2 years and 45 months,
protocol was approved by the ethics committees in all countries. and they were evenly distributed between the candesartan group
and the placebo group (data not shown).
Long-term follow-up Table 1 shows the baseline characteristics of the included
We performed a prolonged follow-up for up to three-years of the patients. Demographic and clinical characteristics were well
1,286 patients from the Scandinavian countries (Norway, Sweden,
and Denmark), as pre-specified in the protocol. The patients
Table 1 Baseline characteristics
personal identification numbers were linked with the national
patient registries and cause of death registries (1419), and we Candesartan Placebo
extracted data for up till three-years after randomization, or until (n = 632) (n = 624)
death, whichever came first. The national patient registries Women 250 (40%) 275 (44%)
contain information about patients contact with all public and Age (years) 719 (112) 720 (112)
private hospitals, and for each hospitalization the registries Systolic blood pressure (mmHg) 1712 (199) 1721 (194)
include information about the date of admission and discharge, Diastolic blood pressure (mmHg) 895 (137) 892 (141)
Qualifying event
the primary diagnosis and a set of secondary diagnoses. In
Ischemic stroke 537 (85%) 531 (85%)
Norway, data linkage with the patient registry was only possible Hemorrhagic stroke 87 (14%) 85 (14%)
from 2008 onwards, so for Norwegian patients included before Other 7 (1%) 8 (1%)
2008 we used data from the Cardiovascular Disease in Norway Unknown 1 (<1%) 0
(CVDNOR) project (20). The cause of death registries include SSS score 414 (124) 411 (130)
OCSP syndrome
information about the date and cause of death. Data extraction
Total anterior 55 (9%) 57 (9%)
from all registries was performed blinded to information about Partial anterior 293 (46%) 288 (46%)
treatment allocation. Posterior 111 (18%) 91 (15%)
Lacunar 169 (27%) 182 (29%)
Effect variables Other 4 (<1%) 6 (1%)
The primary end-point was the composite end-point of stroke, Duration of symptoms (h) 181 (78) 184 (78)
myocardial infarction, or vascular death. Secondary end-points Premorbid mRS score 0 (0-0) 0 (0-0)
were recurrent stroke (non-fatal or fatal) and death of all causes. Medical history
Hypertension 362 (57%) 360 (58%)
We used the 10th version of the International Classification of
Diabetes mellitus 90 (14%) 92 (15%)
Diseases (ICD-10) to define end-points. A patient was said to have Current or previous atrial 111 (18%) 108 (17%)
had a stroke if he was hospitalized after the first 28 days of ran- fibrillation
domization with a diagnosis of I61, I63 (except I636), I64, or Previous stroke or TIA 166 (26%) 123 (20%)
G45. Myocardial infarction was defined as I21 or I22, occurring at Current use of an ACE 118 (19%) 109 (18%)
inhibitor
any time after randomization. The diagnosis of vascular death was
Thrombolytic treatment before 30 (5%) 29 (5%)
made if death occurred within 28 days of onset of stroke or randomization
myocardial infarction, or if the cause of death was G45 or R96, or
Data represent numbers (%), mean (SD), or median (IQR). Percentages
any diagnosis from chapter I.
are proportions of valid data entries, which might be lower than the
Statistical analysis number of patients in each group. SSS, Scandinavian Stroke Scale;
OCSP, Oxfordshire Community Stroke Project (both ischemic and hem-
We used the Cox proportional hazards regression model for
orrhagic strokes); mRS, modified Rankin Scale; TIA, transient ischemic
analysis of all end-points, after adjustment for the following pre- attack; ACE, angiotensin-converting enzyme.
defined prognostic variables; age, type of stroke (ischemic vs. all

2015 World Stroke Organization Vol 10, August 2015, 830835 831
 Research A. G. Hornslien et al.

2029 patients included in the trial

743 patients included in

non-Scandinavian countries (37%)

1286 patients included in

Scandinavia (63%)

642 patients allocated to 640 patients allocated to

candesartan (50%) placebo (50%)

10 patients with no long-term 16 patients with no long-term

data (2%) data (2%)

632 patients with long-term 624 patients with long-term

data (98%) data (98%)

Fig. 1 Trial profile.

Table 2 Composite vascular end-point during three-years follow-up

Candesartan (n = 632) Placebo (n = 624) Hazard ratio (95% CI) P-value

Composite vascular end-point 178 (28%) 203 (33%) 087 (071107)* 019*
Components of composite end-point
Stroke 107 (17%) 124 (20%)
Non-fatal 103 113
Fatal 4 10
Myocardial infarction 15 (2%) 20 (3%)
Non-fatal 11 19
Fatal 4 1
Vascular death of other causes 56 (9%) 60 (10%)

Data represent numbers (%). *Adjusted for age, cause of stroke, systolic blood pressure, and Scandinavian Stroke Scale score at baseline.

balanced between treatment groups, except that there were more
patients with a history of previous stroke or TIA in the candesar-
tan group. As for all patients in the trial (5), blood pressure was
around 6/2 mmHg lower in patients allocated candesartan from
day 2 onwards (p < 0001). Blood pressure was similar in the two
groups at one-, three- and six-months, and at six-months, mean
blood pressure was 143/80 mmHg and 142/79 mmHg in the can-
desartan and placebo group, respectively (p = 025).
Figure 2 shows the cumulative risk of the composite end-point
of stroke, myocardial infarction or vascular death during three-
years follow-up. There were 178/632 (282%) events in the can-
desartan group and 203/624 (325%) events in the placebo group,
but the difference was not statistically significant [adjusted hazard
ratio (HR) 087, 95% confidence interval (CI) 071107, p = 019;
Table 2]. The figure shows that the two curves cross, which raises
concern over the proportional hazards assumption. We therefore
performed a sensitivity analysis using logistic regression, and
found virtually identical results (adjusted odds ratio [OR] 081,
95% CI 063104, p = 011).
Also presented in Fig. 2 is the cumulative risk of recurrent
stroke. There were 107/632 (169%) recurrent strokes in the can-
desartan group and 128/624 (205%) in the placebo group, which
was not significantly different (adjusted HR 083, 95% CI 064
107, p = 015, Table 2). Additionally, we performed an ordinal
regression analysis to assess the effect of treatment on the severity
of recurrent stroke (no stroke, non-fatal stroke, and fatal stroke),

832 Vol 10, August 2015, 830835 2015 World Stroke Organization
 A. G. Hornslien et al. Research
a b

04 Candesartan 04
Placebo
Composite vascular endpoint

03 03

Recurrent stroke
02 02

01 01

00 00

0 365 730 1095 0 365 730 1095

Follow-up (days) Follow-up (days)

Number at risk
Candesartan 632 510 461 402 632 513 467 411

Placebo 624 507 432 370 624 511 444 379

Number of events
Candesartan 0 114 154 178 0 68 96 107

Placebo 0 109 167 203 0 73 105 128

Fig. 2 Cumulative risk of stroke, myocardial infarction, or vascular death (a) and stroke (b) during three-years follow-up.

but there was no statistically significant difference (adjusted

04 Candesartan
common OR 078, 95% CI 059104, p = 009).
Placebo
Figure 3 shows the cumulative risk of death of any cause during
follow-up. During three-years, 113/632 (179%) patients died in
03
the candesartan group and 117/624 (188%) died in the placebo
group (adjusted HR 100, 95% CI 077130, p = 100).
All-cause death

For the primary end-point, there was no evidence of a differ-

02
ential effect in any of the pre-specified subgroups (Fig 4). For the
subgroup of patients treated very late (>24 h), there was a differ-
ence in favor of candesartan (p = 002), and for patients treated
01
very early (<6 h) there was a small, statistically non-significant
difference in favor of candesartan (p = 013), but the interaction
was not statistically significant (p = 070).
00

Discussion 0 365 730 1095

Follow-up (days)

This analysis of the Scandinavian Candesartan Acute Stroke Trial

Number at risk
showed no clear beneficial long-term effects of blood pressure
Candesartan 632 575 542 487
lowering treatment with candesartan in the acute phase. This
Placebo 624 576 529 473
result adds to the neutral results seen during short-term
follow-up in previous trials (1,2), and supports the conclusions
from these trials that there is no indication for routine blood Number of events
Candesartan 0 57 84 113
pressure lowering treatment with candesartan in the acute phase
Placebo 0 48 87 117
of stroke.
For recurrent stroke we observed a small difference in favor of Fig. 3 Cumulative risk of death from any cause during three-years
candesartan, similar to what was observed in the ACCESS study follow-up.

2015 World Stroke Organization Vol 10, August 2015, 830835 833
 Research A. G. Hornslien et al.

P-value
Stroke pathology

Ischemic
0.93
Hemorrhagic

Systolic blood pressure (mm Hg)

140 - 160

160 - 180 0.20

180 - 200

>200

Duration of symptoms (h)

<6

6 - 11
0.70
12 - 23

>24

History of hypertension

No
0.80
Yes

Thrombolytic treatment

No
0.71
Yes

Cardiac rhythm

Sinus rhythm
0.59
Atrial fibrillation

Use of ACE inhibitor

No
0.69
Yes

0.3 0.5 1.0 2.0 3.0

Favors treatment HR (95% CI) Favors control

Fig. 4 Effects of treatment on the composite vascular end-point during three-years follow-up in different subgroups of patients. P-values are for the
interaction between subgroup and allocated treatment. ACE, angiotensin-converting enzyme; HR, hazard ratio.

(8) and in trials of angiotensin receptor blockers for secondary
prevention (11,12). However, the difference was statistically non-
significant, the ordinal analysis showed no effect of candesartan
on the severity of recurrent stroke, and there was no effect on
all-cause death. We therefore believe that the difference in recur-
rent stroke seen during long-term follow-up is most likely due to
the play of chance.
There was also no evidence of benefit in any of the pre-
specified subgroups. The second Intensive blood pressure Reduc-
tion in Acute Cerebral haemorrhage Trial (INTERACT2) has
indicated that benefits can be obtained with early treatment in
patients with hemorrhagic stroke (3) and there are some sugges-
tions from other trials that the same is true for ischemic stroke
(2,5). We also observed small differences in favor of treatment in
patients treated very early, but the difference was far from statis-
tically significant, and must not be taken as evidence of effect.
The present analysis represents the first report from a
randomized-controlled trial of the long-term effects of blood
pressure lowering treatment in the acute phase of stroke. The
main strengths of this analysis is the randomization of a large
number of patients, the low risk of observation bias (due to
masking of study treatment and blinded collection of long-term
data), long follow-up, and near complete data. The complete
follow-up was possible due to the linkage with nation-wide,
automatically updated patient registries in the Scandinavian
countries.
The analysis includes only two-thirds of the patients included
in the trial, but this was according to plan, since data linkage with
national patient registries was only possible for the Scandinavian
patients. The patients included were broadly similar to those from
other countries, and is a representative sample of stroke patients
admitted to hospitals in the western world. However, with only
two-thirds of the original study population, the analysis was not
adequately powered to detect a treatment effect, and this limits
our possibilities to draw firm conclusions. The use of national
patient registries may also raise some questions. First, we have
only identified patients who were hospitalized with a diagnosis of
stroke or myocardial infarction, and the true incidence of these
end-points may have been underestimated, since some patients
may not have been hospitalized. However, the randomization of a

834 Vol 10, August 2015, 830835 2015 World Stroke Organization
 A. G. Hornslien et al.
large number of patients will probably have ensured similar dis-
tribution of different severities of recurrent stroke and myocardial
infarction in the two treatment groups, and so the estimate of the
treatment effect will not be biased. Second, the diagnostic accu-
racy of the ICD-10 diagnostic codes relies on the quality of coding
of a large number of physicians. But, again, any misclassification
of end-points would have been at random, and this would have
diluted, rather than changed the direction of any treatment effect.
Moreover, we performed an analysis of end-points up to six-
months using data from the national patient registries, and com-
pared this with the analysis of data from the six-months visit in
the trial, and found virtually identical results (data not shown).
In summary, blood pressure lowering treatment with candesar-
tan in the acute phase of stroke had no clear beneficial effect on
the risk of vascular events or death during long-term follow-up.
This result compares to the result seen during short-term
follow-up in previous trials, and supports the conclusions from
these trials that there is no indication for routine blood pressure
lowering treatment with candesartan in the acute phase of stroke.
This should not, however, deter clinicians from starting blood
pressure lowering treatment for secondary prevention before dis-
charge from hospital.
Acknowledgements
We thank Tomislav Dimoski at the Norwegian Knowledge Centre
for Health Services, Oslo, Norway, for his contribution by devel-
oping software necessary for obtaining data from Norwegian hos-
pitals, conducting the data collection and quality assurance of
data in the CVDNOR project.
References
1 He J, Zhang Y, Xu T et al. Effects of immediate blood pressure reduc-
tion on death and major disability in patients with acute ischemic
stroke: the CATIS randomized clinical trial. JAMA 2014; 311:47989.
2 Bath PM, Woodhouse L, Scutt P et al. Efficacy of nitric oxide, with or
without continuing antihypertensive treatment, for management of
high blood pressure in acute stroke (ENOS): a partial-factorial ran-
domised controlled trial. Lancet 2015; 385:61728.
3 Anderson CS, Heeley E, Huang Y et al. Rapid blood-pressure lowering
in patients with acute intracerebral hemorrhage. N Engl J Med 2013;
368:235565.
4 Ankolekar S, Fuller M, Cross I et al. Feasibility of an ambulance-based
stroke trial, and safety of glyceryl trinitrate in ultra-acute stroke: the
rapid intervention with glyceryl trinitrate in Hypertensive Stroke Trial
(RIGHT, ISRCTN66434824). Stroke 2013; 44:31208.
5 Sandset EC, Bath PM, Boysen G et al. The angiotensin-receptor
blocker candesartan for treatment of acute stroke (SCAST): a ran-
domised, placebo-controlled, double-blind trial. Lancet 2011;
377:74150.
2015 World Stroke Organization
Research
6 Qureshi AI, Palesch YY. Antihypertensive Treatment of Acute Cerebral
Hemorrhage (ATACH) II: design, methods, and rationale. Neurocrit
Care 2011; 15:55976.
7 Anderson C. Enhanced Control of Hypertension and Thrombolysis
Stroke Study (ENCHANTED). 2014. Available at https://
clinicaltrials.gov/show/NCT01422616. (accessed 10 March 2015).
8 Schrader J, Luders S, Kulschewski A et al. The ACCESS Study: evalu-
ation of Acute Candesartan Cilexetil Therapy in Stroke Survivors.
Stroke 2003; 34:1699703.
9 Sandercock P, Wardlaw JM, Lindley RI et al. The benefits and harms of
intravenous thrombolysis with recombinant tissue plasminogen acti-
vator within 6 h of acute ischaemic stroke (the third international
stroke trial [IST-3]): a randomised controlled trial. Lancet 2012;
379:235263.
10 Sigurdsson ST, Strandgaard S. Blood pressure lowering in acute
ischaemic stroke: an update on the role of angiotensin receptor block-
ers. J Hypertens 2007; 25:7435.
11 Dahlof B, Devereux RB, Kjeldsen SE et al. Cardiovascular morbidity
and mortality in the Losartan Intervention for endpoint reduction in
hypertension study (LIFE): a randomised trial against atenolol. Lancet
2002; 359:9951003.
12 Lithell H, Hansson L, Skoog I et al. The Study on Cognition and
Prognosis in the Elderly (SCOPE): principal results of a randomized
double-blind intervention trial. J Hypertens 2003; 21:87586.
13 Sandset EC, Murray G, Boysen G et al. Angiotensin receptor blockade
in acute stroke. The Scandinavian Candesartan Acute Stroke Trial:
rationale, methods and design of a multicentre, randomised- and
placebo-controlled clinical trial (NCT00120003). Int J Stroke 2010;
5:4237.
14 Guidelines for the distribution of data from the Norwegian
Patient Register. The Norwegian Directorate of Health, 2009. Avail-
able at https://helsedirektoratet.no/Documents/English/guidelines
.pdf. (accessed 10 March 2015).
15 The National Patient Register Sweden. The National Board of Health
and Welfare, 2015. Available at http://www.socialstyrelsen.se/register/
halsodataregister/patientregistret/inenglish. (accessed 10 March
2015).
16 Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Reg-
ister. Scand J Public Health 2011; 39:303.
17 Cause of Death Registry Norway. Norwegian Institute of Public
Health, 2015. Available at http://www.fhi.no/eway/default.aspx?pid
=240&trg=MainContent_6898&Main_6664=6898:0:25,7847:1:0:0:::0
:0&MainContent_6898=6706:0:25,9012:1:0:0:::0:0 (accessed 10 March
2015).
18 Cause of Death Register Sweden. The National Board of Health and
Welfare, 2015. Available at http://www.socialstyrelsen.se/register/
dodsorsaksregistret. (accessed 10 March 2015).
19 Helweg-Larsen K. The Danish Register of Causes of Death. Scand J
Public Health 2011; 39:269.
20 Igland J, Tell GS, Ebbing M, Nygrd O, Vollset SE, Dimoski T. The
CVDNOR project: Cardiovascular Disease in Norway 19942009.
Description of data and data quality 2013. CVDNOR, 2013. Available
at http://cvdnor.b.uib.no/files/2013/08/CVDNOR-Data-and-Quality-
Report1.pdf (accessed 10 March 2015).
21 Bath PM, Geeganage C, Gray LJ, Collier T, Pocock S. Use of ordinal
outcomes in vascular prevention trials: comparison with binary out-
comes in published trials. Stroke 2008; 39:281723.
Vol 10, August 2015, 830835 835