| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2000;31:355.)
© 2000 American Heart Association, Inc.
Original Contributions |
Placebo Treatment in Acute Stroke Trials
Benefit or Harm to Patients?
From the Departments of Neurology (R.L., R.B., L.D., W.B.) and Emergency Medicine (T.K.), Long Island Jewish Medical Center, Long Island Campus for the Albert Einstein College of Medicine, New Hyde Park, NY.
Correspondence to Richard B. Libman, MD, Department of Neurology, Long Island Jewish Medical Center, The Long Island Campus for the Albert Einstein College of Medicine, New Hyde Park, NY 11040. E-mail libman{at}lij.edu
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Background and Purpose—Some stroke patients and their families express reservations about participating in trials of experimental therapies for acute stroke. Among many reasons given for this is the concern that by participating, patients may be deprived of some component of routine care. We sought to determine the effect on outcome of participating in a clinical stroke trial while being treated with placebo.
Methods—Prospective clinical information was collected for all
patients admitted with acute ischemic stroke between July 1995
and July 1996. A subgroup of these patients was enrolled in a clinical
trial of acute stroke therapy and had been randomly assigned to the
placebo group. The control group was selected from concurrent stroke
patients who were not enrolled in any clinical trial. The National
Institutes of Health Stroke Scale (NIHSS) was performed on admission
and on day 7 after admission. The Glasgow Outcome Scale (GOS) was also
performed at discharge. Stroke severity was classified as "severe"
if NIHSS was 
9 or GOS 3. Group comparisons were performed with
2 tests.
Results—One hundred twenty-six patients were evaluated. Forty-seven were placebo patients, and 79 were selected as control subjects. There were no significant differences between the groups with respect to age, sex, hematocrit, blood glucose level, history of hypertension, diabetes, smoking, or initial NIHSS. In addition, there was no difference between groups in terms of the frequency of baseline stroke subtype. Among our controls, 55 patients (70%) were on antithrombotic treatment during hospitalization, whereas none of our placebo patients were on any antithrombotic treatment. For the GOS at follow-up, a good outcome was attained by 76% of the control subjects and 72% of placebo patients (not significant). A severe NIHSS (>9) at follow-up, however, was documented in 15% of controls and 59% of placebo patients (P<0.001). There was a trend toward a higher ("worse") mean follow-up NIHSS among placebo patients (mean NIHSS, 11) versus controls (mean NIHSS, 6) (P=0.09).
Conclusions—Patients enrolled in the placebo arms of some acute clinical stroke trials have similar functional outcomes but more severe neurological deficits at 1 week than did a control group. These findings might be partially explained by the withholding of antithrombotic medication and the exclusion criteria inherent in most trials. Vigilance is required to ensure that all patients participating in stroke studies be guaranteed optimal known medical therapy.
Key Words: cerebrovascular disease • aspirin • outcome • placebo • clinical trials
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Some stroke patients and their families express reservations about participating in trials of experimental therapies for acute stroke. Among many reasons given for this is the fear that they will do worse in the trial than if treated conventionally. In part, this reflects concern about the risks associated with an experimental drug. In addition, however, is the concern that by participating, patients may be deprived of some component of routine care, even though most if not all protocols allow routine management to proceed, with the experimental treatment being an addition to but not supplanting standard therapy. We sought to determine the effect on outcome of participating in a clinical stroke trial while being treated with placebo. Our hope was that this would isolate the effect of participation per se, excluding the potential effect, either positive or negative, of an experimental treatment.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Prospective clinical information was collected for all patients admitted with acute ischemic stroke between July 1995 and July 1996. A subgroup of these patients was enrolled in a clinical trial of acute stroke therapy and had been randomly assigned to the placebo group. The control group was selected from concurrent stroke patients who were not enrolled in any clinical trial. The National Institutes of Health Stroke Scale (NIHSS) was performed on admission and on day 7 after admission. The Glasgow Outcome Scale (GOS) was also performed at discharge. Stroke severity was classified as severe if NIHSS was
9 or
GOS 3 (conversely, stroke outcome was classified as good if GOS was
<3). Discharge disposition was classified as good if the patient was
discharged to home or rehabilitation and poor if the patient was
discharged to a nursing home or died. For the purposes of this study,
the term "antithrombotic" was defined as including any
antiplatelet agent (primarily aspirin or ticlopidine) or
anticoagulant (warfarin or heparin) administered during the first week
of hospitalization. Group comparisons were performed by
2 tests. |
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
One hundred twenty-six patients were evaluated. Forty-seven were placebo patients, and 79 were selected as control subjects (controls). Thirty-seven of the placebo patients were drawn from a trial of acute anticoagulation, and the remaining 10 patients had been enrolled in 2 different trials of neuroprotective agents. Baseline characteristics of the 2 groups are shown in the Table
. There were no significant
differences between the groups with respect to age, sex, hematocrit,
blood glucose level, history of hypertension, diabetes, smoking, or
initial NIHSS. In addition, there was no difference between groups in
terms of the frequency of baseline stroke subtype. Among our controls,
55 patients (70%) were on antithrombotic treatment during
hospitalization, whereas none of our placebo patients were on any
antithrombotic treatment. The majority of our placebo patients were
drawn from a clinical trial of an anticoagulant, and treatment with
other antithrombotics was prohibited during the treatment phase. The
majority of the patients in our study who received antithrombotic
treatment received aspirin (40 patients), with 1 patient receiving
ticlopidine, 9 warfarin in combination with heparin, and 5 heparin
alone.
|
For the GOS at follow-up, a good outcome was attained by 76% of
the controls and 72% of placebo patients
(2=0.17, P=NS). For discharge
disposition, 90% of controls and 89% of placebo patients had a good
outcome, being discharged either to home or to rehabilitation. A severe
NIHSS (>9) at follow-up, however, was documented in 15% of controls
and 59% of placebo patients, which was statistically significant
(2=24.60, P<0.001). There was a
trend toward a higher (worse) mean follow-up NIHSS among placebo
patients (mean NIHSS, 11) versus controls (mean NIHSS, 6)
(P=0.09). Among all 19 patients whose NIHSS worsened during
the first week, 18 were not on antithrombotic treatment, and only 1 was
on an antithrombotic.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Effective therapy for acute stroke is the subject of numerous ongoing and planned clinical trials. Patients are frequently approached to participate in these studies with the understanding, either implicit or explicit, that general care will be equivalent if not superior to standard care. Additional nursing, medical, and neurological attention from the stroke team provides 1 potential benefit. Among our controls, 55 patients (70%) were on antithrombotic treatment during hospitalization, whereas none of our placebo patients were on any antithrombotic treatment. The majority of our placebo patients were drawn from a clinical trial of an anticoagulant, and for safety reasons, treatment with other antithrombotics was prohibited during the treatment phase. The remainder of the placebo patients, who were participating in trials of neuroprotective agents, did not receive antithrombotic medication during the first week, probably because investigations as to stroke mechanism were ongoing and data from subsequently published trials showing benefit to early antithrombotic treatment were not yet available.1 2 Therefore, at that time, there was no compelling reason to begin antithrombotic therapy as soon as possible after admission. Recent evidence suggests that the early initiation of aspirin after ischemic stroke has a statistically significant, although clinically modest, beneficial effect on outcome.1 2 The International Stroke Trial showed a beneficial effect of aspirin in terms of early recurrence of stroke, but functional outcome was examined only at 6 months.1 The Chinese Acute Stroke Trial, however, suggested that patients treated with aspirin may have a better early functional outcome, ie, death or dependency at discharge from the hospital, although this did not reach statistical significance (2P=0.08).2 At least part of this improvement may have been a result of less severe neurological deficits, although this was not specifically measured in the trial. It is possible that aspirin lessens neurological worsening during the early stroke period. Among all 19 patients whose NIHSS worsened during the first week, 18 were not on antithrombotic treatment, whereas only 1 was on an antithrombotic. Our results suggest that if there is no contraindication, placebo patients in clinical trials may benefit from antithrombotic treatment. It is possible that any kind of additional antithrombotic treatment might be deemed contraindicated in the treatment arms of trials using experimental medications with potential hemorrhagic complication, eg, thrombolytics, anticoagulants, or such potent antiplatelet agents as platelet glycoprotein receptor antagonists. Under such circumstances, perhaps a dummy antithrombotic could be administered to the treatment arm, whereas the placebo patients would receive active medication. This would help to prevent unblinding of the investigators but would clearly introduce imbalances among the treatment arms of the study. We are not certain as to how to resolve this dilemma.
The placebo effect has been noted in many areas of clinical research and has probably been studied most intensively in treatment of various pain disorders. Thirty percent to 40% of pain patients are considered placebo reactors, showing a satisfactory response to placebo treatment.3 Would such a placebo response be expected in trials of new stroke therapies? It is not unreasonable to think so, although this question has not been explored. In addition, many clinical stroke investigators believe (and hope) that patients participating in stroke trials may benefit, even if randomized to placebo, because of additional medical and nursing attention, which is integral to the design of many trials. Why, then, would our placebo patients tend to worsen slightly in terms of their neurological examinations (ie, NIHSS) (although, fortunately, the final functional outcome as measured by the GOS was not significantly different between groups)? One explanation lies in the exclusion criteria of virtually all stroke trials, ie, that patients who show rapid improvement in the emergency department, regardless of initial stroke severity, are not enrolled. Although we did not systematically collect data regarding rapid early improvement among our controls, it is reasonable to expect that at least some of them followed this clinical course, whereas, almost by definition, none of our placebo patients could have had such early spontaneous improvement. It is tempting to speculate that had the groups been balanced in terms of antithrombotic treatment and rapid early improvement, perhaps a better outcome would have been observed among placebo patients. This remains entirely unproven. The majority of our placebo patients, enrolled in an anticoagulant trial, were evaluated in <24 hours from stroke onset, as specified in the study inclusion criteria. The remainder of the placebo patients, enrolled in a neuroprotective trial, were evaluated within <6 hours from onset. We did not systematically collect data on time from stroke onset in our control patients, but late arrival to the hospital is a frequent reason for exclusion from stroke trials. It is plausible, therefore, that many of our control patients arrived later and could be considered to be in more stable condition than the placebo patients, whose stroke syndromes might have been more likely to evolve. Given the possible bias against our placebo patients having a more favorable outcome, it seems reasonable that we could at least reassure our patients and their families that it is highly unlikely that they would be harmed by participation in a stroke trial, especially if all placebo patients receive therapies known to be effective.
Limitations to this study include the relatively small sample size,
particularly of our placebo group, rendering it more difficult to
detect more subtle differences in outcome. The outcome measure at 
7
days addresses only the short term, ie, before discharge from hospital.
It is possible that significant differences might emerge in the longer
term. An example of this is the NINDS tPA stroke trial, in which
statistically significant differences in outcome were not detected at
24 hours (although tPA-treated patients did improve to a greater degree
than did placebo patients), but the 3-month follow-up significantly
favored tPA-treated patients.4 The interval to the second
assessment was 6.7 days for controls and 7.7 days for the placebo
group. This imbalance could theoretically have biased the results in
favor of one or the other group, depending on whether improvement or
worsening might have occurred during the extra day allowed the placebo
group. The time difference was not statistically significant, and it is
unlikely that this had a major effect on outcome measures.
In summary, we have found that patients enrolled in the placebo arms of some acute clinical stroke trials have similar functional outcomes but more severe neurological deficits at 1 week than did a control group. These findings might be partially explained by the withholding of antithrombotic medication and the exclusion criteria inherent in most trials, but both these explanations remain speculative. It is somewhat reassuring that the most relevant outcome measure, ie, functional outcome, was no different between groups. Nonetheless, it is worrisome that early neurological status was worse in the placebo group, because such deficits may predict greater long-term disability. It is the responsibility of ethics committees and institutional review boards to prohibit clinical trials in which placebo patients are denied a therapy known to be effective. Similarly, principal investigators should not participate in such trials. In studies of novel medications with antithrombotic properties for acute stroke (eg, anticoagulants, platelet receptor antagonists, etc), early administration of antithrombotic agents to all placebo patients should be ensured. We have focused on the aspirin question in a circumscribed way, only because it provided 1 potential explanation for the distinct patient sample we studied. The general principle we wish to stress is the need for ongoing vigilance to ensure that all patients participating in stroke studies be guaranteed optimal known medical therapy.
|
Acknowledgments |
|---|
We thank Dr Marc Lawrence Gordon for his helpful discussions.
Received September 22, 1999; revision received November 24, 1999; accepted November 24, 1999.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
1. International Stroke Trial Collaborative Group. The International Stroke Trial (IST): a randomised trial of aspirin, subcutaneous heparin, both, or neither among 19435 patients with acute ischaemic stroke. Lancet. 1997;349:1569–1581.[Medline] [Order article via Infotrieve]
2. CAST: randomised placebo-controlled trial of early aspirin use in 20,000 patients with acute ischaemic stroke. Lancet. 1997;349:1641–1649.[Medline] [Order article via Infotrieve]
3. Lasagna L, Mosteller F, von Felsinger JM, Beecher HK. A study of the placebo response. Am J Med. 1954;16:770–779.[Medline] [Order article via Infotrieve]
4.
The National Institute of Neurological Disorders and
Stroke rt-PA Stroke Study Group. Tissue plasminogen
activator for acute ischemic stroke. N
Engl J Med. 1995;333:1581–1587.
This article has been cited by other articles:
 |
|
 |
|
  K. Uchino, D. Billheimer, and S. C. Cramer Entry Criteria and Baseline Characteristics Predict Outcome in Acute Stroke Trials Stroke, April 1, 2001; 32(4): 909 - 916. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||