Aspirin in the prevention of progressing stroke: a randomized controlled study

Authors


Å. Rödén-Jüllig, Karolinska Institutet, Division of Internal Medicine, Danderyd Hospital, Stockholm, Sweden (fax: 46 8 4274399; e-mail: asa.roden@previa.se).

Abstract.

Objectives.  In acute stroke, progression has a severe impact on patient outcome and no effective treatment is known. The main objective was to evaluate the efficacy of aspirin for prevention of stroke progression thereby improving outcome.

Design.  The trial was randomized, double-blind and placebo-controlled.

Setting.  The patients were treated in stroke units of four hospitals in Sweden.

Subjects.  Patients with ischaemic stroke but not complete paresis were included. No antiplatelet drugs were allowed within the last 72 h before onset. Delay until first trial dosage was maximized to 48 h. The trial was designed to detect a 20% reduction of the rate of stroke progression, which was estimated to take place in 20% of cases. Totally, 441 patients (220 aspirin, 221 placebo) completed the trial. Baseline comparisons between the groups showed no differences.

Interventions.  Aspirin (325 mg) or placebo was given once daily for five consecutive days.

Main outcome measures.  Neurological assessments were carried out three times daily during the treatment period to detect progression of at least two points in the Scandinavian Stroke Supervision Scale. Patient outcome was followed up at discharge and at 3 months.

Results.  Aspirin treatment did not significantly reduce the frequency of stroke progression. Amongst aspirin-treated patients, stroke progression occurred in 15.9% as compared with 16.7% in the placebo group, which is less frequent than expected. The relative risk was 0.95 (95% CI 0.62–1.45) in the treatment group. As regards patient outcome at discharge and after 3 months, aspirin treatment did not show any difference.

Conclusion.  No positive effect of aspirin, of the expected size, could be shown on the frequency of stroke progression or patient outcome.

Introduction

Approximately 20–25% of stroke patients deteriorate during the first days of hospital care according to several reports [1–3]. These patients stay longer in hospital, become significantly more disabled and need more institutional care than those without progression [4–5]. Progressing stroke thus constitutes a serious problem. Neither reliable clinical predictors of progression nor an effective treatment are known [6–11]. Preventive measures would be ideal but have not been addressed in any clinical trial.

The pathophysiological mechanism for progressing stroke symptoms is not known. Theories present the following aetiology: extension of the brain oedema, absence of recanalization, thrombus propagation and various systemic causes [12–14]. There is some evidence that metabolites from the infarct elevate the level of prostaglandins, resulting in propagation of neuronal damage [15]. Autopsy studies show that thrombus in a cerebral arterial vessel is platelet-rich [16]. Aspirin might block both thrombus propagation and prostaglandin formation, thereby preventing stroke progression [17–21].

This trial was set up to test the efficacy of aspirin for prevention of deterioration of acute stroke symptoms and thereby improvement of patient outcome.

Materials and methods

The trial was a randomized, double-blind, placebo-controlled clinical trial conducted in four centres. The regional ethics committee of each participating centre has approved the trial and the patient information. Approval from the Swedish Medical Products Agency was also obtained.

Inclusion criteria

Patients of all ages were included if they had an acute ischaemic stroke (CT confirmed) and focal neurological impairment of at least one point severity according to the Scandinavian Stroke Supervision Scale (SSSS) [22], but not a total paralysis. The patients should not have used antiplatelet drugs (including NSAID) within 72 h before the onset of stroke symptoms. The delay until first trial dosage was maximized to 48 h from onset. Informed consent had to be obtained from the patient or a legal representative. Baseline data from the patient history and examinations at admission were collected in a case report form where laboratory tests and final diagnosis were later registered (Table 1).

Table 1.  Baseline comparisons of the treatment groups. For continuous data 95% confidence intervals are presented. None of the differences between proportions were statistically significant
 Aspirin (n = 220)Placebo (n = 221)
  1. SSSS, Scandinavian Stroke Supervision Scale; SBP, systolic blood pressure; DBP, diastolic blood pressure; PACI, partial anterior circulation infarcts; TACI, total anterior circulation infarcts; LACI, lacunar infarcts; POCI, posterior circulation infarcts; TIA, transient ischaemic attack.

  2. aRef. 30.

Mean age (years)74.1 (72.7–75.4)74.2 (72.9–75.5)
Gender (men, %)52.749.8
Hypertension (%)34.637.3
Diabetes mellitus (%)14.616.8
Stroke (%)7.34.6
Atrial fibrillation (%)14.116.2
Condition at admission
 SSSS (7–27), mean points13.2 (12.6–13.7)12.3 (11.8–12.7)
 SBP (mean, mmHg)165 (161–168)166 (162–170)
 DBP (mean, mmHg)90 (88–92)89 (87–91)
 Blood glucose (mmol L−1)5.9 (5.6–6.3)6.4 (6.0–6.8)
 Haemoglobin (g L−1)142 (140–144)141 (139–143)
 Leucocytes (×109 L−1)8.6 (8.0–9.1)8.3 (8.0–8.6)
 Platelets (×109 L−1)229 (221–237)235 (225–245)
 Delay onset – admission (h)10.0 (8.8–11.2)9.3 (8.2–10.3)
 PACIa (%)76.075.1
 TACI (%)2.31.8
 LACI (%)18.419.9
 POCI (%)3.23.2
Final diagnosis (%)
 Cerebral thrombosis80.976.0
 Cerebral embolism16.221.3
 TIA2.32.7

Exclusion criteria

Patients with severe concomitant medical conditions or pre-existing neurological illness were excluded from the trial. So were patients with known bleeding diathesis, allergic reactions against aspirin or bleeding gastric ulcer during the last year. Ongoing anticoagulation treatment was another reason for exclusion as was a blood pressure level above 240/140 mmHg. The patient must not participate in any other clinical trial.

Randomization and blinding

The patient randomization, stratified for gender, was produced by Danderyd Hospital pharmacy in accordance with standard randomization tables. The randomization tables were kept in the pharmacy until the trial was finished and the safety board had approved the data for all patients. When a patient was accepted for inclusion, one package (in numerical order) was taken. Except for trial number the packages were identical in appearance. Patients, trial personnel and ward staff were thus totally blinded as to the trial medication. A disclosure envelope was kept in the patient medical record. The envelope should only be opened at an emergency. At the check after the termination of the trial, only two disclosure envelopes had been opened, one because of suspected allergy (placebo) and one because of a mistake by a nurse (placebo).

Treatment

Tablets of 325 mg aspirin or placebo, water solvable, were administered orally once a day for five consecutive days. The first dosage was given as soon as possible after inclusion. The trial treatment was discontinued if the main outcome event occurred, i.e. progression of stroke symptoms. When this happened, patients were treated according to the ordinary routines in the participating hospitals with intravenously administered and monitored heparin and thereafter with oral anticoagulants. The heparin dose was somewhat (30%) reduced and the APT-time was kept 30% lower than normal during the first 3 days to lessen the risk of bleeding for the patients whose platelets could be inhibited.

All patients were managed in the same way for coexisting diseases and with physical, occupational and speech therapy in accordance with standard practice in each Stroke Unit. No treatment with drugs affecting platelets was allowed during the period of trial treatment. If acute anticoagulant treatment was indicated before the trial medication was finished, the trial medication was discontinued. Treatment was also stopped on patient request and when side-effects were suspected. Secondary prevention (antiplatelet drugs or anticoagulation) was carried out in accordance with local practice but delayed until the trial medication had been completed.

Outcome

The main outcome event of the trial was a progression of stroke symptoms measurable as two points or more (not necessarily in the same item) worsening on the SSSS, compared with baseline, during 5 days of trial treatment. Therefore the neurological status was continuously tested with the SSSS at inclusion and three times a day during the 5 days and at discharge. Patient outcome was also evaluated at discharge from the SU and 3 months after the stroke onset. At discharge the following were assessed: SSSS, activities of daily living (ADL) according to Barthel [23], ability to walk without aid (by a person or mechanical aid) and if the patient could be discharged home. Items evaluated at 3 months were: living at home, ability to walk unaided, the need for more help for ADL than before the stroke and the total number of days of institutional care during the 3-month period.

Statistical methods and data management

The trial was designed to detect a 20% reduction of the rate of stroke progression (primary end-point), which was estimated to take place in 20% of cases. As a type I error cut-off level 0.05 (two-tailed) was chosen. The type II error cut-off level was set at 0.20. The required total number of patients was thereby calculated to be 500. The secondary end-points (functional status at 3-month follow-up) were measured to ensure that a reduction in progression rate would have resulted in factors of real importance for patients. Repeated measurement analysis was used to analyse time-dependent data. Statistical comparisons to test differences between the two groups were made by use of the Student's t-test for uncorrelated means, after validation for normal distribution by use of the Shapiro–Wilke's test. The within-group analysis was made by using parewise Student's t-test for correlated means. Multiple comparisons of continuous data were performed by analysis of variance. The procedure proposed by Fisher was used to control for multiplicity [24–25]. In order to evaluate hypotheses of variables in contingency tables, the chi-square test was used or, in the case of small expected frequencies, Fisher's Exact Test. All analyses were carried out by using the SAS system [26], and the 5, 1 and 0.1% levels of significance were considered.

Results

The trial was terminated when 444 patients had been included. Three patients were excluded from further evaluations because they did not meet the inclusion criterion of ischaemic stroke: one patient had a brain tumour, one had a peripheral facial palsy and the third had a cerebral haemorrhage, see flow chart in Fig. 1. Thus, the final analysis was performed on data from 441 patients. Nine of these patients did not strictly fulfil the inclusion criteria: three could have taken antiplatelet drugs within 72 h before onset of symptoms, two had a few hours longer delay than 48 h after onset, and in four the symptoms according to the SSSS had resolved at inclusion. Of these nine patients, five belonged to the treatment group and four to the controls. All 441 are included in the following evaluations according to the intention-to-treat method. Totally, of the 441 patients 220 received aspirin and 221 placebo. Baseline comparisons between the two treatment groups are shown in Table 1 and there are no significant differences between the groups.

Figure 1.

Flow diagram of 444 patients randomly allocated. The number of patients actively followed up for the primary outcome at 5 days, and for the secondary outcomes at discharge and after 3 months are presented.

The main results of the trial showed that aspirin treatment did not significantly reduce the rate of stroke progression. The progression rate was 15.9% amongst patients treated with aspirin and 16.7% for those on placebo. In the aspirin group, the relative risk was 0.95 (95% CI 0.62–1.45). Treatment conditions and the main trial outcomes are shown in Table 2.

Table 2.  Treatment conditions and trial outcomes. The 95% confidence intervals for continuous data are shown within brackets. None of the differences between proportions were statistically significant
 Aspirin-treated (n = 220)Placebo-treated (n = 221)
  1. SSSS, Scandinavian Stroke Supervision Scale; SU, Stroke Unit; ADL, activities of daily living.

Stroke progression n (%)35 (15.9%)37 (16.7%)
Treatment
 Delay arrival-treatment (mean, h)12.2 (11.1–13.4)12.1 (10.9–13.3)
 Tablets (mean number given before progression)1.97 (1.57–2.38)2.27 (1.92–2.62)
 Time (mean hours from initiation of treatment to progression)27.7 (17.7–37.6)28.9 (20.8–37.0)
Outcome at discharge
 Time in SU (days)11.0 (9.8–12.2)10.8 (9.7–12.0)
 Discharge home (%)60.664.1
 Walking without aid (%)51.958.4
 Normal ADL index (%)59.566.2
 SSSS, mean points10.9 (10.3–11.5)10.2 (9.7–10.6)
 Dead, n (%)6 (2.7%)7 (3.2%)
Outcome at 3 months follow-up
 Walking without aid (%)60.264.8
 Living at home (%)80.384.0
 Increased need for ADL help (%)46.144
 Living in an institution (%)10.59.0
 Hospital/institutional care (mean days/3 months)25.7 (21.7–29.6)24.7 (20.9–28.4)
 Dead, n (%)15 (6.8%)12 (5.4%)

As regards patient outcome, the other main end-point of the trial, aspirin treatment made no difference. Patient outcome at discharge and at 3 months was similar in the treatment and placebo groups, also when patients with and without progression were analysed separately. Thus, the ability to live at home, to walk unaided, or need for further institutionalized care after discharge was not significantly improved in the aspirin group (Table 2).

Of all included patients, 80.7% completed the trial medication during the five intended days. For the remaining patients, the reasons for discontinuation were: stroke progression in 14.8%, death 0.4%, on the patients request 0.7%, need for acute anticoagulant treatment 0.2%, and 2.3% miscellaneous. Study medication was interrupted due to suspected side-effects on four occasions (0.9%): two patients on aspirin had stomach bleedings. On placebo, one woman had vaginal bleedings and allergy was suspected in one man. One to three of the intended five tablets (never the first two) were missed in nine patients, three in the aspirin and six in the placebo group. As all patients received at least the two initial doses of aspirin, it was not considered meaningful to perform a separate efficacy analysis.

Two post hoc analyses were made. The importance of stroke severity at inclusion and time to treatment was divided into quartiles. The rate of progression increased with increasing severity. For the best two quartiles the rate was nonsignificantly lower in the aspirin than in the placebo group whereas in the worst two groups there was a reversed tendency, Table 3. The mean SSSS did not differ between the groups. Still, there was a slight imbalance as regards stroke severity. Of the patients given aspirin, 43% had a SSSS lower than 12 compared with 52% in the placebo group. This imbalance was accounted for in an analysis of variance, which showed no statistically significant difference in outcome between the groups.

Table 3.  Progress in relation to stroke severity at inclusion. The Scandinavian Stroke Supervision Scale (SSSS) was divided into quartiles with increasing degree of severity. None of the differences between the groups were statistically significant
SSSSAspirin (n = 221)Placebo (n = 221)
ProgressionNo progressionProgressionNo progression
7–9 (n = 97)4 (8.5%)437 (14.0%)43
10–11 (n = 112)5 (10.4%)439 (14.1%)55
12–14 (n = 116)11 (18.6%)4910 (17.9%)46
15–25 (n = 116)15 (22.5%)5011 (21.6%)40

There were no significant differences in time to treatment between the groups.

Unpredicted circumstances

The study started in 1988 at the SU of Danderyd Hospital, was interrupted for the years 1989–91 and restarted at Danderyd and St Göran Hospital in 1992. Recruitment gradually slowed down, whilst two more centres were included (Umeå and Uppsala University Hospitals) with very little effect on inclusion.

Reasons for exclusion of 300 consecutive patients in the first half of the study were noted: already on antiplatelet medication (25%), focal neurological deficit but normal SSSS (14%), too long a delay (13%), no symptoms at the time for possible inclusion (12%), complete deficit at arrival (10%), anticoagulant medication (6%) and miscellaneous (19%). The fraction of patients already being medicated with aspirin gradually increased over time so that recruitment completely ran out of steam. It was therefore considered the only solution finally to terminate the study in 1997 when 444 of the planned 500 patients were included.

To compare patients included early and late in the trial they were divided into four groups according to year of inclusion. Mean age at inclusion was nonsignificantly higher in the later period. The mean level of neurological deficit at inclusion, and the progression rate did not systematically change during the trial.

Discussion

This trial could not show any significant positive effects of aspirin on the acute phase of ischaemic stroke for the prevention of progression or for patient outcome. Neither could any significant negative effects of aspirin treatment be demonstrated.

The progression rate in this trial turned out to be lower than expected when the size of the trial was calculated. Exclusions of patients with the most severe neurological deficit and a gradual change over time of the natural history of stroke may be the explanations. With the present rate of progression the result would not have been altered if all 500 calculated patients had been included in the trial.

The huge International Stroke Trial (IST) and Chinese Acute Stroke Trial (CAST) were not designed to detect a preventive effect of aspirin on stroke progression, and progression in the acute phase was not specifically looked for [27–28]. It is not known whether the slight positive effect of aspirin shown in IST and CAST is related to the prevention of progress. The present trial did not have the power to detect a general effect of that size on patient outcome.

In the subgroup analyses of the present trial there was an uncertain trend that aspirin might be more effective in milder stroke cases for prevention of stroke progression and thereby for improving patient outcome. In the conjunct analysis of IST and CAST the aspirin effect was similar in three prognostic groups of patients [29]. Thus, today's collected data do not provide evidence for different aspirin strategies according to severity of the stroke.

It is a draw back of this study that it was so protracted. On the positive side and in spite of this, there have been a very limited number of persons involved and the main investigators have been able to take part in and supervise the work the whole time. Patients have been continuously investigated, data carefully collected and monitored with very few missing. The trial was randomized in a strictly concealed way and kept completely blinded. All data were delivered and checked by the safety board before the code was broken. No patients were lost to follow-up and an intention-to-treat analysis was used. Data can therefore be used as a background for, or together with data from future studies on the problem of stroke progression.

We conclude that the present trial could not prove effects of the calculated size nor disprove smaller positive effects of aspirin in the prevention of stroke progression. However, no negative effects of aspirin treatment were found whilst the recommendations from the IST and CAST studies to give aspirin in the acute phase of an ischaemic stroke could be supported by the present findings. However, the serious clinical problem with progressing stroke still remains and new treatments, which can prevent progression, must be looked for. We believe that before an effective treatment could be found, the pathophysiological mechanisms underlying stroke progression must be better clarified.

Conflict of interest statement

No conflict of interest was declared.

Acknowledgements

We thank Bo Norrving, MD, PhD and Claes Helmers, MD, PhD for their participation in the safety board. We also thank Senior Statistician Per Näsman, PhD, Center for Safety Research, Royal Institute of Technology, Stockholm for statistical advice and calculations. The placebo tablets were provided free of charge by Miles Inc. through Bayer AG and the aspirin tablets were purchased from Miles Inc. Otherwise the companies were not involved in the conduct of this trial.

The trial is supported by grants from the Serafimer Hospital Foundation, the County Council of Stockholm, Department of Research, Development & Education, the Claes Groschinsky Foundation, the Loo and Hans Osterman Foundation, the Eirs 50-year Foundation, the SALUS 50-year Foundation, the 1987 Foundation for Stroke Research, the Tore Nilsson Foundation for Medical Research and the Karolinska Institutet Foundations for Research.

Ancillary