THE LANCET
Articles
CAST: randomised placebo-controlled trial of early aspirin use in
20 000 patients with acute ischaemic stroke
CAST (Chinese Acute Stroke Trial) Collaborative Group*
Summary
Background Aspirin is effective in the treatment of acute
myocardial infarction and in the long-term prevention of
serious vascular events in sur vivors of stroke and
myocardial infarction. There is, however, no reliable
evidence on the effectiveness of early aspirin use in acute
ischaemic stroke.
Methods The Chinese Acute Stroke Trial (CAST) was a large
randomised, placebo-controlled trial of the effects in
hospital of aspirin treatment (160 mg/day) star ted within
48 h of the onset of suspected acute ischaemic stroke and
continued in hospital for up to 4 weeks. The primar y
endpoints were death from any cause during the 4-week
treatment period and death or dependence at discharge, and
the analyses were by intention to treat. 21 106 patients
with acute ischaemic stroke were enrolled in 413 Chinese
hospitals at a mean of 25 h after the onset of symptoms
(10 554 aspirin, 10 552 placebo). 87% had a CT scan before
randomisation. It was prospectively planned that the results
would be analysed in parallel with those of the concurrent
International Stroke Trial (IST) of 20 000 patients with acute
stroke from other countries.
Findings There was a significant 14% (SD 7) propor tional
reduction in mor tality during the scheduled treatment period
(343 [3·3%] deaths among aspirin-allocated patients vs 398
[3·9%] deaths among placebo-allocated patients; 2p=0·04).
There were significantly fewer recurrent ischaemic strokes
in the aspirin-allocated than in the placebo-allocated group
(167 [1·6%] vs 215 [2·1%]; 2p=0·01) but slightly more
haemorrhagic strokes (115 [1·1%] vs 93 [0·9%]; 2p>0·1).
For the combined in-hospital endpoint of death or non-fatal
stroke at 4 weeks, there was a 12% (6) propor tional risk
reduction with aspirin (545 [5·3%] vs 614 [5·9%]; 2p=0·03),
an absolute difference of 6·8 (3·2) fewer cases per 1000. At
discharge, 3153 (30·5%) aspirin-allocated patients and
3266 (31·6%) placebo-allocated patients were dead or
dependent, corresponding to 11·4 (6·4) fewer per 1000 in
favour of aspirin (2p=0·08).
Interpretation There are two major trials of aspirin in acute
ischaemic stroke. Taken together, CAST and the similarly
large IST show reliably that aspirin star ted early in hospital
produces a small but definite net benefit, with about 9 (SD
3) fewer deaths or non-fatal strokes per 1000 in the first few
weeks (2p=0·001), and with 13 (5) fewer dead or
dependent per 1000 after some weeks or months of follow-
up (2p<0·01).
Lancet 1997; 349: 1641–49
*Members of group and writing committee listed at end of paper
Correspondence to: Dr Z heng-Ming Chen, Clinical Trial Ser vice
Unit, Nuffield Depar tment of Clinical Medicine, Radcliffe Infirmar y,
Oxford OX2 6HE, UK
Vol 349 • June 7, 1997
Introduction
During the current decade in China there will be about 15
million deaths from stroke, plus much disability.1 Although
the proportion of haemorrhagic strokes is somewhat higher
than in western populations, ischaemic stroke still accounts
for the majority of new cases and deaths in China.2,3 If a
simple and widely practicable treatment for acute
ischaemic stroke could be shown reliably to produce even a
moderate improvement in outcome, the population benefit
could be substantial. Aspirin is effective in the treatment of
acute myocardial infarction,4 and a systemic overview in
1994 of all previous trials of long-term antiplatelet therapy
among patients with a history of previous myocardial
infarction, stroke, or transient ischaemic attack showed that
about 40 serious vascular events (myocardial infarction,
stroke, or vascular death) are avoided per 1000 patients
treated for a few years with aspirin.5 As a result, many
patients admitted to hospital with strokes are now being
discharged on long-term low-dose aspirin (or other
antiplatelet agents), not only in western countries,6 but also
in China7 and elsewhere.
There is, however, little evidence on the balance of
benefits and risks of antiplatelet therapy started during the
initial acute phase of ischaemic stroke.8,9 Consequently,
there is much variation in routine clinical practice.6,7,10,11
The large, randomised, placebo-controlled Chinese Acute
Stroke Trial (CAST), and the parallel International Stroke
Trial (IST)12 conducted in other countries, were designed
to provide reliable evidence about the effects on early
mortality and major morbidity of early aspirin treatment in
a wide range of patients presenting with definite or
suspected ischaemic stroke. Both trials planned to enrol
20 000 patients, thereby yielding a total of 40 000.
Methods
Eligibility
Patients admitted to the 413 participating hospitals in China were
eligible for CAST if they were judged to be within 48 h of the
onset of symptoms of suspected acute ischaemic stroke, and had
no clear indications for, or contraindications to, aspirin.
Contraindications were specified not by the protocol but by the
responsible physician, and were generally based on an expectation
of either an increased risk of adverse effects (eg, because there was
a recent history of serious gastric bleeding or a known allergy to
aspirin), or little likelihood of any worthwhile benefit in hospital
(eg, because the patient had only a minor stroke, other major life-
threatening disease, or severe pre-existing disability). The
fundamental criterion for entry was that the responsible physician
was uncertain whether or not aspirin treatment was indicated for a
particular patient. The protocol stated that the physician should
discuss the study in an appropriate way with the patient or
relatives. No written consent was requested, since such a request
may not be appropriate in an emergency setting.13
Recruitment began first in ten hospitals, as a pilot phase in
which 600 patients were randomised. The aims of this phase were
to refine the trial practical procedures and to estimate the likely
recruitment rates. Under the protocol, these patients would be
included in the total study sample.
1641
THE LANCET

death or non-fatal stroke during the scheduled treatment period.

Other major clinical events during the scheduled treatment period,
such as transfused or fatal extracranial haemorrhage and
pulmonary embolus (clinical diagnosis), were also considered.
Non-fatal events were defined as those occurring among patients
who survived to the end of the scheduled treatment period. The
few patients who had more than one type of non-fatal event can
appear more than once in the tabulations. The main prespecified
subgroup analyses were of the effect of aspirin on death or non-
fatal stroke during the scheduled treatment period, subdivided by
hours of delay from onset of symptoms, previous CT scan, and
consciousness level at entry. Subgroup analyses according to
various other baseline charactersitics were also done.

Planned study size

Figure 1: Trial profile The planned sample size was 20 000 patients,17 which was based
on the assumption of 9% mortality in the placebo group during
the scheduled trial treatment period, with 15% of these deaths
CT scanning avoidable by aspirin. In a trial of 10 000 patients, even if the true
A CT scan before randomisation was mandatory only for patients effect of aspirin on early mortality were about 15% (which would
who were comatose. Other patients could be entered for lead to the prediction of about 380 aspirin-allocated vs 450
randomisation without a CT scan provided that the responsible placebo-allocated deaths), the effects of chance could well reduce
physican was reasonably certain that haemorrhagic stroke could be the observed difference to a not conventionally significant value
excluded. (eg, 400 vs 440 deaths). The aim of CAST, therefore, was to enter
at least 20 000 patients to reduce the risk of such a false-negative
Randomisation result. In addition, it was planned prospectively that this study
Randomisation was by prepacked, sequentially numbered trial would be analysed in parallel with the concurrent IST.
envelopes, each containing calendar-packed aspirin or placebo,
produced centrally. Within each hospital, random allocation to
aspirin or placebo was, unknown to the participating physicians, Number in allocated treatment group
balanced for every ten consecutive patients. Initially, 20 sealed Aspirin (n=10 554) Placebo (n=10 552)
envelopes were sent to each participating hospital, strung together
through punched holes in the order they were to be used. Sex
Male 6633 (63%) 6734 (64%)
Subsequent resupply depended on the speed of recruitment, Female 3921 (37%) 3818 (36%)
which was assessed by monitoring of the sequential return of entry
Age (years)
forms. To enter a patient into the trial, the responsible physician
<60 3361 (32%) 3369 (32%)
had simply to complete the one-page entry form attached to the 60–69 4214 (40%) 4202 (40%)
outside of the next trial envelope in the bundle. 肁70 2938 (28%) 2948 (28%)
Unknown 41 33

Trial treatment Time from onset (h)

0–3 453 (4%) 434 (4%)
After the entry form had been completed, the envelope was
4–6 823 (8%) 801 (8%)
opened, and the 4-week calendar pack of either 160 mg film- 7–12 1465 (14%) 1498 (14%)
coated aspirin tablets or matching placebo tablets was removed. 13–24 2888 (27%) 2859 (27%)
The doctor recorded the patient’s name on the drug pack, and 25–48 4906 (47%) 4932 (47%)
then gave the first trial tablet immediately, either crushed or Unknown 19 28
chewed, followed by one tablet each day for 4 weeks (or until Consciousness
earlier discharge or death). If the patient could not swallow the Drowsy/coma 1339 (13%) 1341 (13%)
tablets, nasogastric feeding tubes were used. All other decisions Fully alert 9201 (87%) 9198 (87%)
Unknown 14 13
about management of patients were at the discretion of the
responsible doctor, except that non-trial aspirin or other Atrial fibrillation
antiplatelet drugs were not allowed during the trial period unless Present 696 (7%) 715 (7%)
Absent 9811 (93%) 9790 (93%)
the doctor believed that the patient had developed some strong Unknown 47 47
indication for aspirin, such as acute myocardial infarction. Chinese
herbal products could be prescribed if considered indicated, even CT scan at entry
CT done, lesion visible 7843 (74%) 7859 (74%)
though some may affect platelets.14 CT done, lesion not visible 1379 (13%) 1385 (13%)
No CT 1264 (12%) 1257 (12%)
Discharge Unknown 68 51

At discharge (or death before discharge), a single-sided form was Systolic blood pressure (mm Hg)
completed and returned to the trial office in Beijing. This form <140 2599 (25%) 2596 (25%)
140–159 2861 (27%) 2838 (27%)
provided brief details of compliance with the study treatment, 160–179 2425 (23%) 2438 (23%)
possible side-effects, major clinical events, disability by the 180–199 1814 (17%) 1833 (17%)
modified Rankin disability scale,15,16 and, if the patient had died, 肁200 844 (8%) 835 (8%)
the likely primary cause of death. The protocol encouraged the Unknown 11 12
prescription of low-dose aspirin at discharge for the long-term Type of stroke*
prevention of recurrent stroke and other major vascular events. Total anterior 938 (9%) 954 (9%)
Partial anterior 5772 (55%) 5673 (54%)
Posterior circulation 698 (7%) 746 (7%)
Outcome measures and comparisons Lacunar infarct 3117 (29%) 3146 (30%)
The primary prespecified comparisons were of death from any Other unknown 29 33
cause during the scheduled treatment period (ie, in hospital within Discharge form available 10 335 (98%) 10 320 (98%)
4 weeks), and of death or dependence at discharge. Subsidiary *Based on Oxford Community Stroke Project (OCSP) classification.21
comparisons were of fatal or non-fatal recurrent stroke, and of
Table 1: Baseline characteristics by treatment group

1642 Vol 349 • June 7, 1997

 THE LANCET

Number (%) of those with Absolute 2p were male, 72% were younger than 70 years, 26% were
discharge forms in benefit with within 12 h of symptom onset, 13% were drowsy or
allocated group aspirin per comatose, and 7% had atrial fibrillation. 87% had had a
1000 (SD)*
Aspirin Placebo CT scan before randomisation. Discharge forms were
(n=10 335) (n=10 320)
available by April, 1997, for 20 655 (97·9%) of the
Deaths randomised patients; forms were unavailable for similar
All deaths 343 (3·3%) 398 (3·9%) 5·4 (2·6) 0·04 numbers in the two treatment groups (figure 1). Of
Due to initial stroke 144 (1·4%) 175 (1·7%) 3·0 (1·7) 0·08 patients with discharge forms available, 94% (with 87%
Due to recurrent stroke (any type) 99 (1·0%) 108 (1·2%) 0·9 (1·4) >0·1
Due to other (or unknown) causes 100 (1·0%) 115 (1·1%) 1·5 (1·4) >0·1 scanned before randomisation) had at least one CT scan in
Recurrent stroke (fatal or not)
hospital, and the diagnosis of ischaemic stroke was
All 335 (3·2%) 351 (3·4%) 1·6 (2·5) >0·1 eventually confirmed in about 98% of the randomised
Ischaemic 167 (1·6%) 215 (2·1%) 4·7 (1·9) 0·01 patients. 302 patients (1·5%; 152 aspirin vs 150 placebo)
Haemorrhagic† 115 (1·1%) 93 (0·9%) ⫺2·1 (1·4) >0·1
Unknown 53 (0·5%) 43 (0·4%) ⫺1·0 (0·9) >0·1
were found to have been misdiagnosed at entry, including
174 (0·8%) with haemorrhagic strokes and 128 (0·6%)
Death or non-fatal stroke 545 (5·3%) 614 (5·9%) 6·8 (3·2) 0·03
with brain tumours or other diseases. All were included in
Pulmonary embolism the analyses for the intention-to-treat comparisons. The
All 12 (0·1%) 20 (0·2%) 0·8 (0·6) >0·1
Fatal 5 (0·1%) 10 (0·1%) 0·5 (0·4) >0·1 final diagnosis was not reported on the discharge form for
Non-fatal 7 (0·1%) 10 (0·1%) 0·3 (0·4) >0·1 152 patients (0·7% of those with forms; 73 vs 79). The
Transfused (or fatal) extracranial bleed average stay in hospital was about 4 weeks. 97% in each
All 86 (0·8%) 58 (0·6%) ⫺2·7 (1·2) 0·02 treatment group had completed their 4 weeks of trial
Fatal 39 (0·4%) 31 (0·3%) ⫺0·8 (0·8) >0·1
Non-fatal 47 (0·5%) 27 (0·3%) ⫺1·9 (0·8) 0·02
tablets or had continued to take the tablets throughout the
hospital stay if they died or were discharged during the first
*Negative numbers indicate more events occurred in aspirin than in placebo group.
†Includes cerebral haemorrhage or haemorrhagic transformation of original infarct. 4 weeks.
Table 2: Main clinical events in hospital during scheduled
(4-week) treatment period by allocated treatment Effects on mortality and recurrent stroke
There were 343 (3·3%) in-hospital deaths within 4 weeks
Statistical analyses among patients allocated aspirin compared with 398
The analyses compare all patients allocated aspirin with all those (3·9%) in the placebo group. The proportional reduction in
allocated placebo (intention to treat). Most involve simple the odds of death in the aspirin group (14% [SD 7]) is
comparisons18,19 of total numbers affected during the scheduled statistically significant (2p=0·04) and corresponds to an
treatment period. Absolute differences are given as benefits per absolute difference of 5·4 (SD 2·6) fewer deaths per 1000
1000 patients treated with aspirin, and proportional differences patients allocated aspirin for about 4 weeks
are given as odds ratios (or percentage reductions in the odds
(table 2). Of the deaths in hospital, about one third each
ratio), with SDs or CIs. Two-sided p values (2p) are given
throughout.
were attributed directly to the initial stroke, recurrent
Interim analyses of the primary endpoints were planned after stroke (which, by definition, includes ischaemic stroke,
5000, 10 000, and 15 000 patients had been recruited. In the light haemorrhagic stroke, or haemorrhagic transformation of
of these interim analyses and of any other evidence or advice they the original infarct), and other causes, most of which are
wished to seek, the data-monitoring committee would advise the likely to have been due indirectly to the initial stroke.There
chairman of the steering committee if there was at any time proof were slightly fewer deaths in each category of attributable
beyond reasonable doubt that treatment was either clearly causes of death among patients allocated aspirin than
indicated or clearly contraindicated for all, or for some, types of among those allocated placebo, but none of these
patient. Otherwise, the steering committee, collaborators, and differences was statistically significant (table 2).
administrative staff would not be informed of the interim results.
For the endpoint of fatal and non-fatal recurrent strokes
In September, 1996, the preliminary results from IST20 and
interim results on the first 14 000 patients in CAST were during the scheduled treatment period, there were 335
presented, and the collaborating doctors in CAST were informed (3·2%) among aspirin-allocated patients compared with
of the main results from both studies. Although the preliminary 351 (3·4%) in the placebo group. This difference, though
results were promising, the steering committee decided to not significant, combined a significant reduction of 4·7
continue randomisation in CAST to its scheduled end, with the (1·9) recurrent ischaemic strokes per 1000 (2p=0·01) with
principal entry criterion (as before) being the uncertainty of the
responsible physicians as to whether or not to use aspirin early in
the treatment of any particular patient presenting with suspected Number (%) of those with Absolute 2p
ischaemic stroke. discharge forms in benefit with
allocated group aspirin per
1000 (SD)*
Aspirin Placebo
Results (n=10 335) (n=10 320)
Characteristics of participants
Functional status of survivors†
21 106 patients with entry forms available by April, 1997, Total number of sur vivors 9915 ( 9853 (
were enrolled from 413 Chinese hospitals between Fully recovered 3840 (38·7%) 3716 (37·7%) ⫺10·1 (6·9) >0·1
November, 1993, and March, 1997 (figure 1). This large Independent, not fully recovered 3299 (33·3%) 3307 (33·6%) 2·9 (6·7) >0·1
Dependent 2776 (28·0%) 2830 (28·7%) 7·2 (6·4) >0·1
size ensured good balance between the aspirin and placebo
groups for the main prerandomisation prognostic features Dead 377 (3·6%) 436 (4·2%) 5·8 (2·7) 0·03

that were measured (table 1). The mean ages of the aspirin Dead or dependent 3153 (30·5%) 3266 (31·6%) 11·4 (6·4) 0·08
and placebo groups were 63·2 years (SD 10·4) and 63·1 *Negative numbers indicate more events occurred in aspirin than in placebo group.
years (10·3) years; the mean times from onset of symptoms †Information on disability at discharge not available for 74 sur vivors (43 aspirin, 31
placebo) who were excluded from this analysis.
were 24·9 h (14·1) and 25·1 h (14·2); and the mean Fully recovered=Rankin scale 0; independent, not fully recovered=Rankin scale 1–2;
systolic blood pressures were 156·8 mm Hg (27·9) and dependent=Rankin scale 肁3.
156·9 mm Hg (28·0). Of all patients randomised, 63% Table 3: Functional status at discharge

Vol 349 • June 7, 1997 1643

THE LANCET

Figure 2: Effects of early aspirin treatment in acute ischaemic stroke on death or non-fatal stroke during scheduled treatment period,
subdivided by prerandomisation features (CAST)
Odds ratio (black square, with area propor tional to amount of information contributed5) and 99% CI (horizontal line) plotted for various subgroups of
study population. A black square to the left of solid vertical line suggests a benefit, significant at 2p<0·01 only if whole 99% CI is to the left of solid
ver tical line. Overall result for all individuals entered (and 95% CI) is represented by a diamond. In each subdivision, any patients with missing
information were omitted.

1644 Vol 349 • June 7, 1997


Figure 3: CAST, IST, and MAST-I—overview of proportional
effects of early aspirin treatment in acute ischaemic stroke
Recurrent ischaemic stroke, haemorrhagic stroke, and death or non-fatal
stroke=within scheduled treatment periods—ie, in hospital up to 4
weeks for CAST, 2 weeks for IST, and 10 days for MAST-I.
Dead or dependent=at end of follow-up—ie, at discharge for CAST, at 6
months for IST and MAST-I.
Symbols and conventions as in figure 2. MAST-I data exclude patients
allocated streptokinase.
a non-significant excess of 2·1 (1·4) haemorrhagic strokes
per 1000. For the combined endpoint of death or non-fatal
stroke, there were 545 (5·3%) events in the aspirin group
and 614 (5·9%) in the placebo group (2p=0·03). This 12%
(6) proportional reduction corresponds to 6·8 (3·2) fewer
cases per 1000 patients allocated aspirin.
Effects on disability at discharge
Survivors were classified at discharge as dependent,
independent but not fully recovered, or fully recovered. For
the primary outcome of dead or dependent at discharge,
there was a non-significant trend favouring those allocated
aspirin (3153 [30·5%] vs 3266 [31·6%]; 2p=0·08),
corresponding to a reduction of 11·4 (6·4) per 1000
Vol 349 • June 7, 1997
THE LANCET
(table 3). Among patients who were independent at
discharge, 3840 (53·8% of those independent) aspirin-
allocated patients achieved full recovery compared with
3716 (52·9%) patients in the placebo group (2p>0·1).
Effects on other clinical events
Few cases of pulmonary embolism were recorded during
the scheduled treatment period, and there was no
significant difference between the treatment groups
(table 2). Aspirin was, however, associated with a small but
significant excess of 2·7 (SD 1·2) transfused or fatal
extracranial bleeds per 1000 during the scheduled
treatment period (86 [0·8%] vs 58 [0·6%]; 2p=0·02).
There is substantial overlap between deaths and major
events reported in hospital. To avoid double counting in
the assessment of any effect of aspirin, the analysis was
restricted to patients alive at the end of scheduled
treatment period. In this analysis, the numbers of non-fatal
pulmonary emboli reported in each group were similar
(seven vs ten), and the excess of non-fatal transfused
bleeds was 1·9 (SD 0·8) per 1000 (47 [0·5%] vs 27
[0·3%]; 2p=0·02). Information on non-fatal myocardial
infarction was not sought, but this event is relatively rare in
Chinese stroke patients.
Major subgroup analyses
Figure 2 shows the observed effects of aspirin on death or
non-fatal stroke during the scheduled treatment period
according to baseline features. There were no significant
differences in the effects of aspirin between the different
subgroups studied, including those prospectively identified
for study. For patients treated early or late after onset of
symptoms, the odds reductions for those randomised at
0–3 h, 4–6 h, 7–12 h, 13–24 h, and 25–48 h were: 36%
(SD 18), 15% (18), ⫺3% (16), 12% (11), and 13% (9),
respectively. For those randomised with or without a prior
CT scan the odds reductions were 11% (6) and 28% (17),
and for those with or without impaired consciousness at
entry the odds reductions were 15% (10) if drowsy and
11% (8) if alert. The analysis in figure 2 includes more
than 2600 patients who were drowsy or comatose when
enrolled; the prognosis in this large subgroup is poor (the
risk of death or non-fatal stroke during the treatment
period was 16·2% aspirin vs 18·5% placebo). The
prognosis for the 17 881 who were fully alert was much
better (3·7% deaths vs 4·1%). Thus, a similar proportional
benefit in the poor-prognosis and good-prognosis
subgroups would correspond to a greater absolute benefit
in those with a poor prognosis.
Although aspirin appeared to have little effect in some
subgroups (such as those randomised at 7–12 h, or
younger patients), this is not good evidence that the real
effects in these subgroups differ from the effect observed
overall, for, the effect of aspirin on death or non-fatal
strokes in CAST as a whole is not highly significant and so
results in particular subgroups will not be reliable. Hence,
the chief need is not to consider subgroups of CAST but to
consider the CAST results together with the randomised
evidence from the other trials of aspirin.
Discussion
Before CAST and IST12 (which also randomised about
20 000 acute stroke patients), no reliable evidence was
available on the value of early antiplatelet therapy in acute
ischaemic stroke, and there was particular concern about
the potential risk of intracranial haemorrhage.6,11 These two
1645
THE LANCET
Figure 4: CAST, IST, and MAST-I—overview of absolute effects
of early aspirin treatment in acute ischaemic stroke on clinical
events during scheduled treatment periods (as in figure 3)
Numbers on bars indicate actual number of cases. Error bars=1 SE for
percentage.
large trials in about 40 000 randomised patients, together
with the non-fibrinolytic component of the smaller
MAST-Italy trial,8 show, however, that immediate
treatment of acute ischaemic stroke with medium-dose
aspirin (160 or 300 mg daily) produces a modest, but
definite (2p=0·001), net reduction in early death or non-
fatal stroke (figures 3 and 4: a systematic search identified
no other aspirin trials9). In these trials, which included a
wide range of patients, aspirin caused about two (SD 1)
haemorrhagic strokes among every 1000 patients treated,
but prevented about 11 (SD 3) other strokes or deaths in
hospital. Almost all the difference in fatal or non-fatal
recurrent ischaemic or unknown stroke was confined to
ischaemic stroke (322 aspirin vs 457 control cases), a
difference of 6·7 (SD 1·4) per 1000. For stroke with
unknown aetiology, there were 175 aspirin and 181 control
cases.
Overall, therefore, about nine (SD 3) deaths or non-fatal
strokes were avoided for every 1000 patients treated with
aspirin for a few weeks from soon after the onset of acute
ischaemic stroke.
The background risk of haemorrhagic stroke was much
greater in CAST than in IST (nine per 1000 in CAST
placebo group vs three per 1000 in those allocated no
aspirin or heparin in IST), but the small absolute increase
in this risk associated with early use of aspirin was similar
in the two studies (excess two per 1000 in both trials).
In many countries, though perhaps not in China, 1–2%
of acute stroke patients would also suffer a non-fatal
myocardial infarction during the first few weeks in hospital.
This risk is also likely to be reduced by aspirin,5 increasing
the net clinical benefit of nine per 1000 slightly, perhaps to
about ten per 1000. This benefit from just a few weeks of
aspirin given early in acute ischaemic stroke seems
worthwhile, being about the same as the benefit from a
whole year of antiplatelet therapy in the long-term
secondary prevention of stroke.5 Moreover, there was no
evidence of any adverse effect among survivors in
functional recovery after stroke. Overall in CAST, there
1646
were 11 (SD 6) fewer patients dead or dependent at
discharge per 1000 allocated aspirin, and in IST (which
assessed disability at 6 months) there was a similar benefit
of 13 per 1000.12 For these two trials together with the
small MAST-Italy trial, the benefit is 13 (SD 5) fewer dead
or dependent per 1000 (2p=0·007). This finding provides
substantial reassurance that the early use of aspirin does
not materially increase the prevalence of serious disability
among survivors.
Overall, the patients enrolled in CAST were mainly at
low risk, with in-hospital mortality only half that in IST,
even though the two studies had virtually identical entry
criteria. This difference in mortality is due partly to the
difference in the age distribution of the patients: in CAST,
72% were younger than 70, whereas in IST only 38% were
under age 70, reflecting demographic differences between
China and the countries where IST took place. It is also
due partly to the exclusion from CAST of some (though
not all) patients with severe stroke, perhaps because of
particular concerns among Chinese doctors about possible
haemorrhagic side-effects of aspirin or because of their
reluctance to use nasogastric tubes in patients unable to
swallow because of severe stroke. Another contributing
factor may be differences in the clinical subtypes of stroke
prevalent in Chinese and western populations. For
example, previous studies have suggested that stroke in
western populations is characterised by predominantly
extracranial carotid-artery disease (with a substantial
prevalence of cardioembolic stroke), whereas in China
intracranial small-vessel disease, which results in lacunar
infarction and haemorrhage, occurs more frequently.22,23
Another difference is that in CAST there was widespread
use of other treatments (eg, 53% of patients received a
calcium antagonist and 51% received various Chinese
herbal products), whereas virtually no other treatments
were used commonly in IST. No substantial benefit for any
of these additional treatments in acute stroke has been
proven, however, and even if some do produce modest
benefits they could not account for the large difference in
prognosis between the two trials.
Despite the low overall risk of the CAST population, the
trial was large and included substantial numbers of patients
at high risk of early death (for example, more than 2600
patients were drowsy or comatose at entry, of whom about
16% died or suffered non-fatal stroke; figure 2). The results
of CAST and IST should therefore be applicable to stroke
patients at both low risk and high risk.
Although CAST and IST each involved about 20 000
patients, their results had to be combined, yielding
evidence from a total of 40 000, before the effects of
aspirin could be assessed reasonably reliably, since the
benefits are not as large in acute ischaemic stroke as in
myocardial infarction.4 If this highly significant evidence of
a small but real benefit does substantially alter medical
practice, one important research implication is the need for
more large-scale randomised trials in stroke medicine. As
in the treatment of acute myocardial infarction, the
expectation of large reductions in mortality and major
morbidity from any single therapeutic agent is probably
unrealistic.24 Moderate treatment effects are more
plausible, and may well be worthwhile if produced by
widely practicable treatments. Reliable assessment of such
moderate effects may, however, require strict
randomisation of several thousands, or even tens of
thousands, of patients.
Aspirin irreversibly inhibits platelet cyclo-oxygenase,
Vol 349 • June 7, 1997

thereby inhibiting platelet aggregability, and is the most
convenient and widely tested antiplatelet agent. There are
many different pathways involved in the aggregation of
platelets, and a combination of two effective antiplatelet
agents working through different mechanisms might be
more effective than a single agent. For example,
ticlopidine, which inhibits ADP-induced platelet
aggregation, may add to the proven benefit of aspirin in
patients at high risk of occlusive arterial disease,25 as may
dipyridamole26 (although earlier trials had not suggested
this5). However, particularly large trials will be needed to
test whether the addition to aspirin of ticlopidine,
dipyridamole, or some newer antiplatelet agent (such as
clopidogrel27) can produce significantly greater clinical
benefit than aspirin alone.28
CAST and IST are directly relevant to the practicalities
of stroke management. CAST took place in a wide variety
of specialist and non-specialist hospitals throughout China,
and IST in a similar range of hospitals in 36 different
countries. The identification of suitable patients did not
involve any material change in the normal patterns of
investigation or diagnosis in either trial, and the trial
treatments did not interfere with the use of most other
treatments. Aspirin treatment does not necessitate careful
monitoring, so it can be used widely not only in developed
countries but also in countries with limited medical
resources. For about 800 of the 40 000 patients in CAST
and IST, the initial diagnosis was wrong, and the patient
had had a cerebral haemorrhage before randomisaton.
There was no indication in either trial, however, that these
misdiagnosed patients were damaged by aspirin, so any
such hazard cannot be large.
Hence, aspirin treatment should now be considered for
almost all patients presenting with acute ischaemic stroke,
provided that there are no strong contraindications and
that haemorrhagic stroke can be excluded with reasonable
probability, whether based on CT scan or not. If early
aspirin treatment were to be given in hospital to one
million new patients a year—a fraction of the worldwide
total with acute ischaemic stroke1—about 10 000 early
deaths, non-fatal strokes, or myocardial infarctions would
be avoided without any apparent increase in the prevalence
of disability among survivors. Furthermore, early aspirin
treatment in acute ischaemic stroke is likely to lead to more
patients being sent home on long-term low-dose aspirin,
which is already of proven value,5 and this consequence will
reduce the mortality and major morbidity after stroke still
further.
CAST study organisation
Writing committee—Z M Chen, R Collins, L S Liu, H C Pan, R Peto,
J X Xie.
Steering committee—Z M Chen (clinical coordinator), J M Hui, L S Liu
(chair), Z M Liu, R Peto, P Sandercock (IST liaison), W Q Wang,
Y X Wang, Z B Wang, J X Xie (coordinator), G X You, F L Zhang,
H Q Zhang, Z Y Zhao.
Data monitoring committee—R Collins,Y P Guo, Z Qin.
Neurological advisory group—Y P Guo, Z Qin, X F Tang.
Coordinating centre—Beijing H Q Fang, X Y He, L S Liu, H L Ma,
G X Xiao, J X Xie, S Y Xu, A R Zhang, X E Zhang.
Clinical Trial Service Unit, Oxford—J Boreham, Z M Chen, R Collins,
J Godwin, H C Pan, R Peto.
Participating centres (by province)—Anhui: Chuzhou Third People’s
Hospital (Z P Wang); Tonglin People’s Hospital (Y C Kang); Anhui
Provincial Hospital (B X Zhou); Bengbu Medical College Hospital
(Y W Zhang); Anqing City Hospital (Z X Chen, J J Mo); Anhui Medical
College First Hospital (C Q Wang); Suzhou First People’s Hospital
(X M Wang); Hefei Second People’s Hospital (J X Liu); Huannan Dongshui
Miner Hospital (F F Wang); Tongcheng County People’s Hospital
(Z G Xiang); Fuyang Third People’s Hospital (T X Wang); Caohu City
People’s Hospital (S Q Guo); Wangjiang County People’s Hospital
Vol 349 • June 7, 1997
THE LANCET
(W H Ai); Feixi County People’s Hospital (S S Chen); Anqing Second
People’s Hospital (J C Yang); Shucheng County People’s Hospital
(Y Y Song); Anqing Petrochemical General Hospital (J Hua); Qianshan
Country Hospital (L Z Ge); Xuancheng Prefecture Second People’s
Hospital (Z L Tang); Taihu County People’s Hospital (Z J Zhang).
Beijing: Beijing CCP School Hospital (Y C Xu); Jiuxianqiao Worker’s
Hospital (S S Wang); Capital Iron and Steel Hospital
(J Q Wang); Beijing Fengshen Hospital (X J Yang); Fengtai Great Wall
Hospital (J Wang); Beijing 402 Hospital (L J Shi); Beijing Boai Hospital
(W J Zhou); China Aerospace Central Hospital (C Z Wang); Beijing
Aerospace 731 Hospital (X M Wang); Beijing Railway Hospital
(J F Zhou); Beijing Haidian Hospital (Y H Li); PLA 304 Hospital
(X M Meng); PLA 309 Hospital (P X Liu); Fanshan District First Hospital
(J Z Li); Beijing Military District General Hospital (W M Yin); China Civil
Aviation Beijing Hospital (S Q Han); PLA General Hospital (G P Wang);
Beijing Tiantan Hospital (J Long); Beijing Medical College Third Hospital
(S Q Zhao); PLA 292 Hospital (W L Wei, S P Kang); Chanping County
Beijiao Hospital (Z C Li); Tong County Luhe Hospital (W L He); Beijing
Chuiyangliu Hospital (M Y Wu); Shunyi County Hospital (T H Dong);
Beijing Shuili Hospital (J Fan); Changping County Hospital (K Y Song);
Guanganmen Hospital (G D Gao); Beijing Mentougou Hospital (X Liu);
Beijing Air Force General Hospital (C S Shi); PLA Naval General Hospital
(G F Liu). Fujian: Fujian Provincial Hospital (X Z Zhou); Lian Jian County
Hospital (J Q Chen). Gansu: Lanzhou Military District General Hospital
(J S Yang); Jinchun Nonferrous Metal Co. Hospital (P Q Wang).
Guangdong: Zhaoqin First People’s Hospital (K Z Lian). Guangxi: Guangxi
Mingzhu Hospital (Y A Zhong); Liuzhou People’s Hospital (J Su).
Guizhou: Zunyi Medical College Hospital (W L Du); Guiyang Huaxi
Prefecture People’s Hospital (L P Zhao); Dujun Qiannanzhou People’s
Hospital (X J Liu). Hebei: Xingtai Third People’s Hospital (H C Ning);
Hebei Provincial People’s Hospital (X Z Wang); Qinghuangdao First
People’s Hospital (J Li); Baoding Second Hospital (Q S Du); Chengde
Central Hospital (X L Zhao, R P Cheng); Shijiazhuang Baiqiuhen
International Peace Hospital (C Z Wang); Tangshan Third People’s Hospital
(X C Wu); Baoding Second Central Hospital (Y S Li); Chengde Medical
College Hospital (Z J Dou); Chengde County Hospital (J Y Ji); Chengde
Fengning County Hospital (G L Fu); Shiyan Dongfeng Automobile
Hospital (Q D Zhou); Wuhan Military Industry Hospital (J D Liu); Hubei
Finance and Trade Hospital (C Z Zhu); Jingsha Central Hospital
(MH Zhang); Zigui County First People’s Hospital (F R Zeng);
Qinhuangdao Port Hospital (J Y Liu); Fengrun County Hospital
(J F Hao). Heilongjiang: Harbin Fourth Hospital (W Z Wang); Harbin
Medical College First Hospital (Z B Wang); Harbin Medical College
Second Hospital (Y W Zheng); Mudanjiang First People’s Hospital
(Y J Zhao); Qiqihaer First Hospital (L H Zhao); Ningan People’s Hospital
(L Men); Heilongjiang RedCross Hospital (S Y Yu); Harbin Second
Hospital (Y Y Chi); Jiamushi Medical College First Teaching Hospital
(C H Jiang); Qiqihar Nenjiang Agriculture Administration Central Hospital
(S J Zhu); Archen People’s Hospital (K X Ju); Shanzhi People’s Hospital
(A X Fu); Hailuan First People’s Hospital (Y Z Xiang); Qiqihar PLA 203
Hospital (N G Li); Daqing Fourth Hospital (L Y Ma); Bin County People’s
Hospital (G C Zhang); Mudanjiang Railway Hospital (Z J Lu); Harbin First
Hospital (X H Zhang); Qiqihar Second Hospital (Y X Song);Yichun
Forestry Administration Central Hospital (G R Chen); Daqing Oil and
Petroleum Administration Worker’s Hospital (Y T Dou); Harbin Railway
Central Hospital (R H Han); Daxinganlin Forest Bureau Central Hospital
(S Y Li). Henan: Henan Medical College First Hospital (Z H Fan);
Zhengzhou Second People’s Hospital (L H Feng); Henan Hebi City
Hospital (W P Zhao); Xingxian First People’s Hospital (W P Zhao); Henan
Lushan County Hospital (Y X Li); Henan Xi County Gongliao Hospital
(S L Zhang); Henan Traditional Medicine College First Teaching Hospital
(J L Wang); Xuchang Central Hospital (Z Z Hu); Zhengzhou PLA 460
Hospital (Z H Cheng); Zhengzhou Guomian Third Hospital (J Z Lu);
Zhengzhou Guomian Fourth Hospital (Q Y Tie); Luohe First People’s
Hospital (Y F Wei); Qinyang People’s Hospital (M C Tian); Zhongyuan
Oilfield Worker’s Central Hospital (Q Q Ma); Luoyang Second People’s
Hospital (H Q Tian); Henan Construction Worker’s Hospital (J X Dong);
Kaifeng County Central Hospital (S D Shi); Xinxiang Medical College
First Hospital (G M Li); Runan Country Worker’s Hospital (D H Peng);
Mengjin County People’s Hospital (Y L Lu); Jiaozuo People’s Hospital
(Q F Pan); Xinxian Central Hospital (X E Chen); Changge People’s
Hospital ( B C Li); Zhumadian Prefecture Hospital (W J Bo); Nanyan
County People’s Hospital (X F Li); Zhumadian Prefecture People’s
Hospital (H B Ma, F T Zhang); Zhoukou Prefecture People’s Hospital
(L L Tian); Luoyang Cotton Worker’s Hospital (Y Y Zhang); Kaifen First
People’s Hospital (M D Yin); Zhumadian People’s Hospital (S Q Wang);
Xinxian Second People’s Hospital (P He); Sanmenxia People’s Hospital
(Z B Sun); Linbao People’s Hospital (T H Tu); Zhengzhou Seventh
People’s Hospital (Q Z Yang); Zhengzhou Fourth People’s Hospital
(Z H Quan); Henan Provincial People’s Hospital (Y Li); Zhengzhou Great
Wall Aluminum Worker’s Hospital (G B He); Zhengzhou Railway Central
Hospital (S Y Ma); Xinyan Prefecture People’s Hospital (J Liu); Henan
Traditional Chinese Medicine College (Z S Li); Anyan People’s Hospital
(S Yao); Henan Medical College Second Hospital (Q Q Liu). Hubei: Jingsha
First People’s Hospital (C S Zhao); Hubei Gezhouba Central Hospital
(J Zhao); Xiangfan First People’s Hospital (J J Yang);Wuhan First People’s
1647
THE LANCET
Hospital (J E Wang); );Wuhan Eleventh People’s Hospital (L Hu); Wuhan
First Metallurgy Construction Company Hospital (H Z Cheng); Dongxi
District People’s Hospital (S B Xu); Jingmen First People’s Hospital
(X D Liu); Hubei Li Yuan Hospital (S B Xu); Tianmen Second People’s
Hospital (Z R Tang); Guangshui First People’s Hospital (X C Peng);
Yingcheng People’s Hospital (X B Yin); Zigui County First People’s
Hospital (C M Zhao); Gongan County People’s Hospital (F D Chen);
Nanzhang County People’s Hospital (Q L Zhang); Enshi Autonomous
Region People’s Hospital (Z S Han); Erzhou Second People’s Hospital
(Y S Wang); Laohekou First People’s Hospital (D C Xu); Xiangfan Central
Hospital (H X Peng); Gezhouba Three Gorges Hospital (P S Lei); Deye
People’s Hospital (Z L Zheng); Suizhou First People’s Hospital (C Q Jiang);
Wuhan 161 Central Hospital (Y Z Chen). Hunan: Yiyan Prefecture Hospital
(J X Cao);Yiyan People’s Hospital (G F Cao); Chenzhou 331 Hospital
(L M Xie); Xiangtan Third People’s Hospital (L J Chen); Xianxian People’s
Hospital (C L Zhou); Hunan Medical College Second Hospital
(Y M Zhou); Cili County People Hospital (D Q Yan). Jiangsu: Xinghua
City Hospital (C W Gao); Nianjing Railway Central Hospital (X G He);
Xuzhou Fourth People’s Hospital (ZG Wei); Lianyungan First People’s
Hospital (S Y Li);Yancheng First People’s Hospital (J G Gu); Zhangjiagang
First People’s Hospital (L G Li); Suzhou Medical College Second Hospital
(G F Shao); Xuzhou Medical College Hospital (M L Lian); Nanjing
Military District General Hospital (Y Q Jin); Nanjing Medical College First
Hospital (Q Tong,Y L Cheng); Nanjing Bayi Hospital (S Z Shao); Nanjing
Gulou Hospital (X J Tang); Nanjing Navy 414 Hospital (H Wang); Hanmen
People’s Hospital (S R Zhu); Tongzhou People’s Hospital (C S Yao);
Xuzhou First People’s Hospital (L S Zhou); Rugao People’s Hospital
(X H Shen). Jiangxi: Nanchang First Hospital (M X Ou); Jiangxi Medical
College Second Hospital (Z L Fen); Gannan Medical College Hospital
(X K Liu). Jilin: Changchun Jilin University Hospital (Y R Shi); Changchun
Diesel Engine Factory Hospital (Y H Feng); Jilin Medical College Hospital
(X P Wang); Changchun Baiqiuhen Medical College Third Hospital
(Y Q Wu); Changchun Central Hospital (X Gao);Yanji People’s Hospital
(Y F Zhu); Dongbei Normal University Hospital (Y J Liu); Jihua Co.
Worker’s Hospital (C N Zhu, P Y Sun); Jilin Prison Administration Central
Hospital (L J Wang); Gongzhulin City First Veteran Hospital
(R Y Shi); Jilin University Hospital (Y X Gong); Changchun PLA 461
Hospital (R Q Gao); Jilin Medical College Cerebrovascular Institute
(G Li); Changchun Second Hospital (X Peng); Changchun Baiqiuhen
Medical College Second Hospital (Y Zhang); Jilin BaichengCentral
Hospital (Y Yu); Taonan Hospital (Y S Wang); Jilin Central Hospital
(W Q Li, SQ Wang); Dumen Hospital (Z J Li); Liaoyuan Central Hospital
(G Y Wang); Dunhua Hospital (S C Wen);Yanbian Medical College
Hospital (X J Han, G Wu);Yanji Hospital (C J Piao); Jilin Dianli Hospital
(M Fang); Jilin Huichun Hospital (R Z Luo); Jilin Baicheng Hospital
(H R He); Siping Central Hospital (Y F Xue); Changchun China First
Automobile Hospital (Y R Tong); Jilin Provincial Hospital (G L Zhang,
R L Cao); Jilin Paper Industry Worker’s Hospital (L Y Shao);Yanbian
Miaolin Cement Factory Hospital (S K LI); Helong People’s Hospital
(J Cui); Daan First Hospital (Z L Wang); Liaoyuan Mine Administration
Workers’ Hospital (X S Wang); Taonan Neurological and Psychiatry
Hospital (S S Li); Jilin Hospital (S Liu); Shenyang Railway Bureau Dumen
Hosptial (E H Li); Tumen Shiyan Paper Mill Hospital (C Y Li). Liaoning:
Shenyang Railway Bureau Hospital (S L Li); Dalian Thrombosis Institute
(H Q Zhang); Anshan Cerebrovascular Centre (Y X Wang); Shenyang Tiexi
District Central Hospital (F Zhang); PLA 324 Hospital (Z C Tang);
Shenyang Red Cross Hospital (B X Guo); Fushun Miner’s Hospital
(B H Chen); Benxi Central Hospital (B D Zhou); Liaoning Paramilitary
Police General Hospital (A P Xu); Liaomei Co. Danai Hospital (F S Zhou);
Dalian 523 Factory Worker’s Hospital (Y J Lin); Shenyang Fifth People’s
Hospital (G Q Yang); Angang Tiedong Hospital (J Y Li); Shenyang 242
Hospital (A H Wang); Angang Tiexi Hospital (J Sun,Y C Wang); Liaoning
Shihua Co.Worker’s Hospital (W H Yu); Liaoyang Third People’s Hospital
(M Chen); Railway Ministry 191 Hospital (CR Li); Chaoyang Second
People’s Hospital (X H Zhang); Dandong First People’s Hospital
(S Q Zhou); Liaoyan Central Hospital (G X Cao); Fushun Central Hospital
(T Xie); Shenyang Seventh People’s Hospital (Y Q Yang); Dalian Friendship
Hospital (G Y Zhao); Shenyang Medical College Second Hospital
(X L Zhang); Angang Lishan Hospital (L J Zhao); Shenyang PLA 463
Hospital (W F Ma); Shenyan Dadong District Hospital (X R Tang);
Shenyang Fourth People’s Hospital (B H Chen); Fushun Steel Hospital
(B J Han); Dalian Medical College First Hospital (X J Hong); Dalian PLA
210 Hospital (Z R Yu); Dalian PLA 469 Hospital (F Tao); Chaoyang
Central Hospital (G X Zhao); Shenyang 157 Hospital (L Li); Shenyang
Chinese and Western Medicine Thrombosis Centre (B R Zhao); Jingzhou
Central Hospital (C X Liu). Neimongol: Neimongol Women & Children
Hospital (G Z Men); Neimongol Autonomous Region’s Hospital
(JC Zhou); Baotou Medical College Second Hospital (G F Zhao). Ningxia:
Ningxia Medical College Hospital (S Q Yan) Shaanxi: Xian Fourth Military
Medical College Tangdu Hospital (Z Y Wu); Xian Fourth Hospital
(S Q Liu); Xian Central Hospital (H X Wu), Xian Medical College Second
Hospital (J M Li). Shandong: Laiwu Iron & Steel General Hospital
(X X Ma,Y J Yang); Qingdao Second People’s Hospital (P P Li); Huaifang
City Women and Children’s Hospital (K F Zhang); Huaifang Second
People’s Hospital (Z B Zhang); Binzhou Prefecture People’s Hospital
(X F Li, J Y Cui); Cao County People Hospital (T P Li); Qingdao Laoshan
1648
401 Hospital (G W Zhong); Dongying People’s Hospital (Q Y Xue);
Huaifang Huaichai Hospital (S H Liu); Shangdong Zhuanye Co.Worker’s
Hospital (Q D Wang); Shandong Qianfushan Hospital (R H Zhu);Yuchen
People’s Hospital (W M Wang); Huaifang Hospital (N L Guo); Linqu
County People’s Hospital (X R Ma); Jiaxiang County People’s Hospital
(D Q Zhuang); Qufu People’s Hospital (R F Du); Juye County People’s
Hospital (G X Hao); Longri Kuangwuju Central Hospital (Z Z Fu);
Zhangqiu People’s Hospital (Q J Men); Shandong Medical Academy Keyun
Hospital (M H Liu); Shandong Huanhe Hospital (C S Shi); Wudi County
People’s Hospital (J C Li); Dezhou People’s Hospital (Z J Song, X G Ren);
Jinan Third People’s Hospital (W R Ma); Wuaifan People’s Hospital
(F Z Meng); Zibo Central Hospital (K R Tang); Binzhou Prefecture Central
Hospital (X Z Yan); Taian First People’s Hospital (Y S Wang); Shandong
Prison Administration Hospital (X M Du); Feicheng People’s Hospital
(Y H Sun); Linyi PLA 146 Hospital (G F Zhou); Jinan Central Hospital
(X H Liu). Shanghai: Shanghai First People’s Hospital (F L Zhang);
Shanghai Hypertension Institute (X Y Wang); Shanghai Second Medical
College Ninth Hospital (J X Jin); Ruijin Hospital (D M Hu); Xinhua
Hospital (L H Yu); Shanghai Railway Central Hospital (J Y Sun); Huadong
Hospital (Y S Jin); Songjiang County Central Hospital (Y F Zhang).
Shanxi: Yunchen Prefecture Hospital (D Tong), Shanxi Medical College
First Hospital (L H Wei); Shanxi Medical College Second Hospital
(R Z Zhang); Xinzhou Prefecture People’s Hospital (H R Yang);Yangquan
Mine Administration Hospital (Y S Wang); Meng County People’s Hospital
(X M Cui). Sichuan: Chongqin Medical College First Hospital (P Man);
Chengdu Third People’s Hospital (K L Zhang); Jinyan Central Hospital
(X Y Wang); Chongqing Southwest Hospital (S Q Shao); Huaxi Medical
University First Hospital (M Liu, G G Yuan); Chengdu Traditional Chinese
Medicine Hospital (J Liu); Chengdu 95 Hospital (M Ye); Chunbei Medical
College Hospital (X Q Zheng); Shifang County Central Hospital
(J S Wang); Third Military Medical College Third Hospital (J Z Wang);
Chengdu Fifth Yejin Construction Workers’ Hospital (M Chen); Chengdu
Stroke Prevention Hospital (C Yang); Pengan County People’s Hospital
(Z X Pu); Deyang People’s Hospital (B Zhang); Deyang Second People’s
Hospital (B H Dang); Dazhou Prefecture People’s Hospital (P J Zhou);
Jintang County First People’s Hospital (C F Xu); Pangan General Hospital
(S L Liu); Shuanliu County First People’s Hospital (G X Liu); Shuanliu
County Second People’s Hospital (Y Wang); Mianyan Third People’s
Hospital (G J Wang); Chongqing Emergency Medical Centre (S J Mao);
Pujiang County People’s Hospital (J R Yang); Chongqing Second People’s
Hospital (Z F Ma); Chinese Engineering Academy Hospital (B X Hu);
Wanxian Fourth People’s Hospital (HY Liao); Pengzhou People’s Hospital
(Y Y Xiao); Guanghan Third People’s Hospital (Q R Yang); Lianshanzhou
First People’s Hospital (L Liu); Jianyan People’s Hospital (Z L Gong);
Chinese Nuclear Industry 416 Hospital (H T Fu); Chengdu First People’s
Hospital (M L Chen); Chengdu Sixth People’s Hospital (C J Song);
Chengdu 107 Hospital (Y Q Peng); Sichuan Provincial People’s Hospital
(X R Zeng); Shuidianbu Wuju Central Worker’s Hospital (Y Y Huang);
Mianyan People’s Hospital (Y Q Li); Xichong County People’s Hospital
(R Z Feng); Zigong First People’s Hospital (M L Liu); Sichuan Medical
Administration College Hospital (R X Xie); Sichuan Chemistry Industry
Worker’s Hospital (S Y Qin); Mianzhu County People’s Hospital
(T L Cao); Jiajiang County People’s Hospital (C Z Xiao); Pi County
People’s Hospital (S Liu); Neihui Third People’s Hospital (W Chen); Lezhi
County People’s Hospital (M L Liu). Tianjin: Tianjin Brain Disease
Hospital (S M Wang); Tianjin Hospital (Q Z Meng); Tianjin Red Cross
Hospital (Q L Jiang); Tianjin Xiqing Hospital (K Lin); Tianjin Nankai
Hospital (R J Zhang); Tianjin Third Hospital (P Wang); Tianjin Tanggu
Hospital (J Y Zheng); Tianjin Institute of Geratology (Q F Meng); Tianjin
Tianhe Hospital (J Y Li); Tianjin 272 Hospital (T Cai, Z F Liu); Tianjin
Hongqiao Hospital (D M Fang). Xinjiang: Xinjiang Medical College First
Hospital (D R Fang); Xinjiang Autonomous Region People’s Hospital
(Y H Tan);Wulumuqi Railway Central Hospital (D Y Liu);Wulumuqi
Friendship Hospital (T Q Zhou); Kalamayi Oil Administration General
Hospital (L Wang) Yunnan: Kunming First People’s Hospital (J M Hui);
Yunnan Provincial Hospital (Z Z Zhu);Kunming Medical College Second
Hospital (Y Li); Funming Medical College First Hospital (H W Jin);
Lincang Prefecture Hospital (X B Liu); Chengdu Military District
Kunming Hospital (J B Hou). Z hejiang: Shaoxin People’s Hospital
(Z Q Sa).
Acknowledgments
Thousands of patients and hundreds of doctors and nurses collaborated.
The study was funded through the Medical Research Council’s support for
the Clinical Trial Service Unit, and trial tablets were donated by Shandong
Xinhua Pharmaceuticals and tested by Dr J McVittie, Oxford Department
of Biochemistry. The study was conducted, analysed, and interpreted
independently of any companies.
References
1 World Health Organization.World Health Statistics Annual, 1994.
Geneva, Switzerland:World Health Organization; 1995.
2 Li SC, Schoenberg BS,Wang CC, Cheng XM, Bolis CL,Wang KJ.
Cerebrovascular disease in the People’s Republic of China:
epidemiologic and clinical features. Neurology 1985; 35: 1708–13.
Vol 349 • June 7, 1997

3 Chen DY,Wu GS,Wu CS, Hong SG, Zhang M, Zhao D. Incidence and
diagnostic features of stroke in the Beijing MONICA population.
Chinese Med J 1987; 7: 382–386.
4 ISIS-2 (Second International Study of Infarct Survival) Collaborative
Group. Randomised trial of intravenous streptokinase, oral aspirin, both,
or neither among 17187 cases of suspected acute myocardial infarction:
ISIS-2. Lancet 1988; ii: 349–60.
5 Antiplatelet Trialist’ Collaboration. Collaborative overview of
randomised trials of antiplatelet treatment. Part I: Prevention of death,
myocardial infarction and stroke by prolonged antiplatelet therapy in
various categories of patients. BMJ 1994; 308: 81–106.
6 Lindley RI, Amayo EO, Marshall J, Sandercock PAG, Dennis M,
Warlow CP. Acute stroke treatment in UK hospitals: The Stroke
Association Survey of consultant opinion. J R Coll Phys Lond 1995; 29:
4790–84.
7 Chen ZM, Sandercock P, Xie JX, Peto R, Collins R, Liu LS. Hospital
management of acute ischaemic stroke in China. J Stroke Cerebrovasc Dis
1997; 6: 361-67.
8 Multicentre Acute Stroke Trial - Italy (MAST-I) Group. Randomised
controlled trial of streptokinase, aspirin, and combination of both in
treatment of acute ischaemic stroke. Lancet 1995; 346: 1509–14.
9 Counsell C, Sandercock P. The efficacy and safety of antiplatelet
therapy in patients with acute presumed ischaemic stroke: a systematic
review of the randomised trials comparing immediate antiplatelet
therapy with control. In:Warlow C,Van Gijn J, Sandercock P (eds.)
Stroke Module of The Cochrane Database of Systematic Reviews,
[updated 02 December 1996]. Available in The Cochrane Library [disk
and CDROM]. The Cochrane Collaboration; Issue 1. Oxford: Update
Software; 1997.
10 Marsh EE, Adams HP, Biller J,Wasek P, Banwart K, Mitchell V,Woolson
R. Use of antithrombotic drugs in the treatment of acute ischaemic
stroke. A survey of neurologists in practice in the United States.
Neurology 1989; 39: 1631–34.
11 Kent J, Bamford J. Consultant views on the use of aspirin in acute
cerebrovascular disease: implications for clinical trials. Postgrad Med J
1994; 70: 185–87.
12 International Stroke Trial Collaborative Group. The International Stroke
Trial (IST): a randomised trial of aspirin, subcutaneous heparin, both,
or neither among 19 435 patients with acute ischaemic stroke. Lancet
1997; 349: 1569–81.
13 Collins R, Doll R, Peto R. Ethics of clinical trials. In:Williams CJ, ed.
Introducing new treatment for cancer: practical, ethical and legal
problems. London:Wiley, 1992: 49–65.
14 Chen KJ, Chen K. Ischaemic stroke treated with Ligusticum
chuanxiong. Chin Med J (English) 1992; 105: 870–873.
Vol 349 • June 7, 1997
THE LANCET
15 Chen ZM, Xie JX, Peto R, Collins R, Liu LS (for the CAST
Collaborative Group). Chinese Acute Stroke Trial (CAST): Rationale,
design and progress. Cerebrovasc Dis 1996; 6 (suppl 2): 23 (abstr).
16 Bamford JM, Sandercock PAG,Warlow CP, Slattery J. Inter-observer
agreement for the asseeement of handicap in stroke patients. Stroke
1989; 20: 8.
17 Dennis M,Wellwood I, O’Rourke S, Machale S,Warlow C. How reliable
are simple questions in assessing outcome after stroke? Cerebrovasc Dis
1997; 7: 19–21.
18 Peto R, Pike M, Armitage P, Breslow NE, Cox DR, Howard SV, et al.
Design and analysis of randomised clinical trials requiring prolonged
observation of each patient: I: introduction and design. Br J Cancer
1976; 34: 585–612.
19 Antiplatelet Trialist’ Collaboration. Secondary prevention of vascular
disease by prolonged antiplatelet treatment. BMJ 1988; 296: 320–331.
20 International Stroke Trial Collaborative Group. The International Stroke
Trial. Preliminary results, Part I: Effects of aspirin and heparin
separately and in combination. Cerebrovasc Dis 1996; 6 (suppl 2): 23
(abstr).
21 Bamford J, Dennis M, Sandercock P, Burn J,Warlow C. The frequency,
causes and timing of death within 30 days of acute stroke: the
Oxfordshire Community Stroke Project (OCSP). J Neurology Neurosurg
Psychiatry 1990; 53: 824–29.
22 Huang CY, Chan FL,Yu YL,Woo E, Chin D. Cerebrovascular disease in
Hong Kong Chinese. Stroke 1990; 21: 230–35.
23 Feldmann E, Daneault N, Kwan E, Ho KJ, Pessin MS, Langenberg P,
Caplan LR. Chinese-white differences in the distribution of occlusive
cerebrovascular disease. Neurology 1990; 40: 1541–45.
24 Collins R, Peto R, Gray R, Parish S. Large-scale randomized evidence:
trials and overviews. In:Weatherall D, Ledingham JGG,Warrell DA, eds.
Oxford Textbook of Medicine, Volume 1. Oxford: Oxford University
Press, 1996: 21–32.
25 Leon MB, Baim DS, Gordon P et al. Clinical and angiographic results
from the Stent Anticoagulation Regimen Study (STARS). Circulation
1996; 94: I-685 (abstr).
26 Diener H, Cunha L, Forbes C, Sivenius J, Smets P, Lowenthal A.
European Stroke Prevention Study 2. Dipyridamole and acetylsalicylic
acid in the secondary prevention of stroke. J Neurologi Sci 1996; 143:
1–13.
27 CAPRIE Steering Committee. A randomised, blinded, trial of
clopidogrel versus aspirin in patients at risk of ischaemic events
(CAPRIE). Lancet 1996; 348: 1329–1339.
28 Born GVR, Collins R. Aspirin versus clopidogrel: the wrong question?
Lancet 1997; 349: 806.
1649