Donate Help Contact The AHA Sign In Home
American Heart Association
Stroke
Search: search_blue_button Advanced Search
« Previous Article | Table of Contents | Next Article »
Stroke. 1996;27:1507-1515

This Article
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Dorman, P. J.
Right arrow Articles by Sandercock, P. A.G.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Dorman, P. J.
Right arrow Articles by Sandercock, P. A.G.

(Stroke. 1996;27:1507-1515.)
© 1996 American Heart Association, Inc.

Articles

Considerations in the Design of Clinical Trials of Neuroprotective Therapy in Acute Stroke

Paul J. Dorman, MRCP Peter A.G. Sandercock, FRCP

the Neurosciences Trials Unit, Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital (Scotland). E-mail pd@skull.dcn.ed.ac.uk.

Correspondence to Dr Paul Dorman, Neurosciences Trials Unit, Department of Clinical Neurosciences, University of Edinburgh, Western General Hospital, Crewe Rd, Edinburgh, Scotland EH4 2XU.


Key Words: clinical trials • neuroprotection • stroke, acute


*    Introduction
up arrowTop
 *Introduction
down arrowWhich Agent to Choose?
 down arrowSelecting Patients for Trials...
 down arrowOptimum Duration of Treatment
 down arrowMeasuring Outcome
 down arrowSample Size Estimation
 down arrowCollaboration With Industry
 down arrowSummary
 down arrowReferences
 
Acute stroke is a substantial public health problem; over the next decade about 15 million patients will suffer acute stroke in Europe and the United States,1 and approximately 85% of these strokes will be ischemic in origin. Therefore, an effective and widely applicable treatment of acute cerebral ischemia would have an enormous public health impact, but as yet no such treatment has been found.2 In addition, there is no effective treatment for the 15% of patients with primary ICH.

The existence of an ischemic penumbra3 (a zone of ischemically threatened but potentially viable tissue in patients with acute ischemic stroke) in the tissue around a primary intracerebral hematoma has raised interest in whether neuronal damage in this zone may be limited by early neuroprotective intervention. Many agents that intervene at one or more points in the pathophysiological cascades initiated by ischemia have been identified4 5 6 7 and are currently entering clinical evaluation.

Well-designed RCTs are the most effective and efficient way to evaluate new treatments. Interest in the design, methodology, analysis, and interpretation of RCTs of treatment of acute stroke has increased substantially in the last decade.8 9 10 The design of the large simple trials in acute ischemic stroke, the IST11 and the Chinese Acute Stroke Trial, evolved from the large trials of antithrombotic and thrombolytic therapy in acute myocardial infarction.12 13 Because neuroprotective agents differ from antithrombotic and thrombolytic agents in their modes of action, pharmacological properties, and profiles of adverse effects,4 a number of methodological issues must be considered in the design of large controlled clinical trials of neuroprotective therapy in acute stroke. In this article, we review the choice of trial design: should it be a small study with very detailed data collection among a few hundred highly selected subjects or a large simple trial with collection of only key data items from several thousand (or perhaps tens of thousands of) patients?14


*    Which Agent to Choose?
up arrowTop
 up arrowIntroduction
 *Which Agent to Choose?
down arrowSelecting Patients for Trials...
 down arrowOptimum Duration of Treatment
 down arrowMeasuring Outcome
 down arrowSample Size Estimation
 down arrowCollaboration With Industry
 down arrowSummary
 down arrowReferences
 
The first step in the design of any study is to choose the "best" (or most promising) agent. However, a large number of neuroprotective agents are currently under development by the pharmaceutical industry. Several hundred are undergoing preclinical evaluation in experimental models. In a recent systematic review of phase 2, 3, and 4 RCTs of newer neuroprotective agents, at least 31 agents in various stages of clinical development were identified.7 Although there have been more than 80 trials (some ongoing) of these 31 different neuroprotective agents, involving approximately 12 000 patients with acute ischemic stroke, none has yet shown clear evidence of benefit.7 However, all the completed studies have been small (the largest included 1234 patients)15 and therefore lacked the power to show moderate but clinically worthwhile benefit.

How can the most promising of the many agents currently in development be identified? First, reliable evidence of preclinical efficacy must be sought; the key areas that need to be addressed are listed in Table 1Down (personal communication, Professor McCulloch). Second, the data from the early-phase clinical trials can be reviewed with respect to grounds, other than safety or efficacy, that would make a novel efficacious treatment more likely to be widely used (Table 2Down).


View this table:
[in this window]
[in a new window]
  		Table 1. Choice of Neuroprotective Agents for Evaluation in Large Phase 3 and 4 Studies: Key Markers of Efficacy From Experimental Studies


View this table:
[in this window]
[in a new window]
  		Table 2. Choice of Neuroprotective Agents for Evaluation in Large Phase 3 and 4 Studies: Important Clinical Qualities in Potential Neuroprotective Agents


*    Selecting Patients for Trials of Neuroprotection
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 *Selecting Patients for Trials...
down arrowOptimum Duration of Treatment
 down arrowMeasuring Outcome
 down arrowSample Size Estimation
 down arrowCollaboration With Industry
 down arrowSummary
 down arrowReferences
 
`Uncertainty Over Scope for Benefit' as Criterion of Eligibility
The "uncertainty principle" has been used as a fundamental design feature to define eligibility and facilitate recruitment in large international multicenter collaborative trials (eg, the ISIS-2,12 ECST,16 and IST11 ). It enabled clinicians who were uncertain about the benefits of treatment (eg, carotid endarterectomy) for individual patients to randomize such patients in controlled clinical trials. Because clinicians are likely to be uncertain about the benefits of unlicensed neuroprotective treatments in all patients, we propose that this principle is applied to whether there is uncertainty over scope for benefit in individual patients. Patients in whom there is substantial uncertainty over scope for benefit should be considered eligible for study entry. On the other hand, patients with minimal scope for benefit from treatment (eg, patients with trivially mild strokes who are likely to recover spontaneously or patients in a coma who are likely to die whatever treatment is given) should be excluded.

Should Patients Presenting More Than 6 Hours After Symptom Onset Be Included?
There is continuing controversy about the duration of the therapeutic window for neuroprotective therapy in acute stroke in humans.17 Experimental studies in a variety of animal models have failed to demonstrate significant neuroprotection beyond 6 hours. Consequently, there is a strong body of opinion suggesting that the "time window" for treatment (and hence the eligibility for clinical trials of neuroprotection) should be restricted to patients presenting within 6 hours of symptom onset.18 19 20 21 However, it was recently reported that administration of low-molecular-weight heparin to patients within 48 hours of onset of acute ischemic stroke improved clinical outcomes.22 An absolute therapeutic window is unlikely; it seems plausible that the effective time window for intervention might differ according to the type of stroke and the properties of the particular treatment; furthermore, treatment is not likely to be an all-or-nothing phenomenon.17 Thus, whereas neuroprotective therapy is likely to be most efficacious if administered early, there are many reasons to test a broader time window in controlled clinical trials, not the least that the benefit, though smaller, may still be worthwhile.

Several groups have attempted to demonstrate potentially viable brain tissue in the ischemic penumbra in humans by means of PET studies.23 24 Serial PET studies in acute ischemic stroke have demonstrated a dense core of ischemia with very low cerebral blood flow and cerebral metabolic rate of oxygen. Values similar to those seen in areas of fully developed infarction (indicating the nonviability of affected tissue) are found at an early stage in the infarct core. But, in 45% to 57% of cases of acute stroke studied within 4 days of symptom onset, areas of tissue adjacent to the ischemic core affected by "misery perfusion" (characterized by an increased oxygen extraction fraction) were identified.25 The fate of this penumbral tissue is unclear; it may either survive and surround the infarct as a viable rim or become necrotic.26 The observation of this potentially viable tissue at up to 4 days after stroke onset and the recent demonstration by serial diffusion-weighted MRI of progression of the ischemic lesion volume beyond 12 hours in patients with acute ischemic stroke27 raise the prospect that the therapeutic window for neuroprotection in humans may be significantly longer than that observed in experimental animals.

Experimental Data in Animals Relate Only Indirectly to Humans
The brain lesions in acute strokes are clinically heterogeneous, and they affect patients with a wide variety of comorbid conditions. In contrast, the infarcts induced in experimental models occur in a homogeneous population of inbred, young, healthy animals. The ischemic cerebrovascular lesion is generated in a consistent manner, and it affects a precise arterial territory with a clearly defined time of onset. Given these major differences between stroke in animals and humans, the relevance of the therapeutic window as measured in laboratory animals must be questioned.28 The potentially misleading nature of such experimental evidence was clearly illustrated in the development of thrombolytic therapy for the treatment of acute myocardial infarction. Experimental data suggested that thrombolysis could not possibly be effective more than 4 to 6 hours after the onset of coronary artery occlusion.29 30 31 In view of this, the design of several RCTs of thrombolytic therapy in acute myocardial infarction was constrained, and randomization of patients who presented more than 6 hours after onset of chest pain was not permitted.32 33 Subsequent trials with broader entry criteria established that thrombolytic therapy in selected patients presenting up to at least 12 hours after the onset of chest pain was still worthwhile.12 13

Broad Entry Criteria Facilitate Recruitment
Severely restrictive trial entry criteria substantially reduce the proportion of eligible patients, which may adversely affect the enthusiasm and morale of a collaborating center and ultimately its ability to recruit patients.34 Moreover, when the occasional patients who do fulfill the entry criteria arrive, they may be overlooked because the investigator never became familiar with the trial procedures, lost interest in the study for lack of eligible patients, or simply forgot about it. For these reasons, very restrictive trial eligibility criteria might so hamper the recruitment of suitable patients that the trial becomes unworkable. This phenomenon was clearly described in a trial of a calcium channel antagonist undertaken at Temple University Hospital (Philadelphia, Penn): very specific and detailed eligibility criteria resulted in just a single patient being recruited (after the screening of 192 patients) during a 2-year period.35 Paradoxically, a broadening of entry criteria may lead to increased recruitment of the target population. This phenomenon is illustrated by the IST, which aims to study the balance of risk and benefit for antithrombotic therapy in patients with acute ischemic stroke within 48 hours of symptom onset. At the end of trial recruitment (May 31, 1996), 19 436 patients were recruited, of whom 3165 (16%) were randomized within 6 hours of stroke onset (see Table 3Down). It is therefore already the largest study of early intervention undertaken in acute stroke (Table 4).Down Broad entry criteria can—in an appropriately designed study—facilitate the recruitment of patients within the first few hours of symptom onset.


View this table:
[in this window]
[in a new window]
  		Table 3. Time to Randomization in IST Based on Overall Trial Recruitment


View this table:
[in this window]
[in a new window]
  		Table 4. Patients Randomized in Major "Early Intervention" Studies

Broad Entry Criteria Can Lead to Increased Generalizability of Study Results
The final results of any trial ideally should be applicable to a wide variety of patients. Extrapolation of the results in a trial is not appropriate if the entry criteria for the trial were unduly restrictive. For instance, as the balance of risks and benefits of treatment may differ with increasing time from symptom onset, it might be inappropriate to generalize the results obtained in a population of patients treated within 6 hours of symptom onset to other patients treated much later. A variety of factors can delay the presentation of patients with acute stroke,8 36 37 38 some of which are potentially modifiable: education of potential patients, general practitioners, and the ambulance service personnel might facilitate the earlier admission of patients with acute stroke.39 However, those whose strokes develop during overnight sleep represent an irreducible fraction of patients who will almost always present more than 6 hours after symptom onset: in the Oxfordshire Community Stroke Project, approximately 24% of patients awoke with their symptoms.40 Although the exact time of symptom onset is unknown, in this group of patients the stroke presumably occurred a few hours before waking. In addition, a significant proportion of patients with acute stroke, particularly those living alone, might be unable to seek help as a direct consequence of the neurological deficits of the stroke; for example, hemiparesis or aphasia may prevent them from using the telephone, further increasing the delay before treatment can be started. Thus, even if attitudes toward stroke could be modified, it is still likely that a significant proportion (perhaps up to 40%) of all stroke patients would continue to present more than 6 hours after symptom onset. Given the size of this population, it is important that potentially effective stroke treatments are tested over a wide time window, to determine whether treatment is effective and safe in the many patients presenting "late."

If Uncertain, Randomize
At present, there is considerable scientific uncertainty regarding the duration of the therapeutic window for neuroprotection in humans.17 Ultimately, the human therapeutic window cannot be established by trials in experimental animals or by observational studies in human subjects. The only scientific way of assessing the width of the therapeutic window is to randomize patients over as broad a time window as possible (up to at least 12 hours) and subsequently analyze the treatment effect separately in prespecified time bands (eg, patients treated within 6 hours and patients treated after 6 hours, etc). Maintaining a broad time window in clinical trials of neuroprotection will not only allow the limits of the therapeutic window to be defined more clearly but will also serve to both enhance the final generalizability of the results and facilitate the rapid conclusion of the trial (by enhancing recruitment).

Should Only Patients Without Visible Infarction on Baseline CT Scan Be Included?
A post hoc subgroup analysis of the American Nimodipine Study suggested benefit for patients with normal rather than abnormal baseline CT scans who were treated with 120 mg nimodipine within 18 hours of symptom onset.41 A potential explanation for this observation is that the presence of infarction on the admission CT scan might indicate irreversible ischemic brain damage.41 On the other hand, given that this was a post hoc subgroup analysis in a trial with an overall null result, this result might have been due to chance.42 Subsequently, on the basis of this hypothesis, patients with clinically relevant infarction clearly visible on the prerandomization CT scan were excluded from entry to a recent trial of thrombolysis in acute stroke (ECASS),43 and it has been suggested that similar eligibility criteria should be used for future trials of neuroprotective therapy.41 Given the dangers of retrospective subgroup analysis,42 44 it would seem prudent not to exclude patients with "visible infarction" until the implications of this finding are better understood. Visible infarction may simply be a marker for time from symptom onset. In the National Institute of Neurological Disorders and Stroke study of thrombolysis within 180 minutes of symptom onset, visible infarction was not an exclusion criteria, but very few subjects had edema or mass effect visible on the prerandomization CT scan (18/624=2.9%),45 presumably because patients were scanned very early after onset. Determining the presence or absence of early infarct signs on CT is not easy; in the recent ECASS study, 10% of all the initially randomized patients were later deemed to have violated the CT scan eligibility criteria,43 which suggests that reliably excluding patients from entry to RCTs because of abnormalities on the prerandomization CT scan may prove impractical. Therefore, in trials of neuroprotective therapy it may be more appropriate to (1) randomize before CT has been performed to minimize the delay (and, indirectly, minimize the proportion of patients with visible infarction) and thereby maximize the scope for benefit, (2) prespecify that the effect of treatment may be different among those with and without infarction visible on the baseline CT and perform appropriate subgroup analyses, and (3) code scans "blindly" to all other data (and perform studies of interobserver and intraobserver reliability of CT scan interpretation).

Should Patients With Hemorrhagic Stroke Be Included?
Primary ICH accounts for about 10% to 15% of all acute strokes in Caucasians46 and perhaps a larger proportion among people of Oriental origin.47 There have been very few trials evaluating medical (or indeed any other) therapy for ICH, and optimal treatment remains controversial.48 Experimental models simulating spontaneous ICH, such as might occur in hypertensive hemorrhage, confirm the existence of an ischemic penumbral zone around the lesion.49 Neuroprotective therapies have not been widely tested in experimental models of ICH. However, available experimental data support the notion that neuroprotection might improve outcome in hemorrhagic stroke.50 Human studies on the ischemic penumbra in this setting are limited. Single-photon emission CT was performed in 8 patients with traumatic ICH51 ; perfusion deficits that appeared larger than the hematoma seen on the CT scan were demonstrated in 5 of the 8 patients, which suggested the existence of a zone of ischemic penumbral tissue around the hematoma. It is therefore plausible that the early institution of neuroprotective treatment in patients with primary ICH might protect these ischemically threatened neurons and so improve long-term outcome. There may be additional advantages to the inclusion of patients with hemorrhagic stroke in RCTs of neuroprotective therapy: for example, the need to perform CT scanning to exclude patients with hemorrhagic stroke before starting antihemostatic treatments such as thrombolysis or antithrombotic drugs delays the start of treatment. In the IST, the average delay attributable to performance of a CT scan before randomization was approximately 6 hours.52 A trial that includes patients with both ischemic and hemorrhagic strokes by not demanding CT before randomization will increase the number of patients randomized within the crucial first few hours after stroke onset. However, in view of the very limited preclinical and clinical experience with neuroprotection in primary ICH, further evaluation of this treatment strategy may be appropriate before it is tested in large RCTs.


*    Optimum Duration of Treatment
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 up arrowSelecting Patients for Trials...
 *Optimum Duration of Treatment
down arrowMeasuring Outcome
 down arrowSample Size Estimation
 down arrowCollaboration With Industry
 down arrowSummary
 down arrowReferences
 
The optimum duration of neuroprotective therapy in humans has not yet been fully defined by the experimental studies in animal models. It is likely to be influenced by a variety of factors that include patient-, agent-, and hospital-related factors (Table 5Down). The variability (from as little as a few hours up to 1 month of treatment) in the duration of therapy applied in the trials of agents modulating excitatory amino acid transmission (Table 6Down) is not explained by the variability in treatment-related factors and presumably reflects scientific uncertainty. A recent proton magnetic spectroscopic study demonstrated continuing loss of cerebral metabolites within an infarct region up to 10 days after infarction.53 This suggests that ischemic damage may continue for several days and highlights the potential need for the continued administration of neuroprotective therapy for up to 1 week or more after symptom onset.


View this table:
[in this window]
[in a new window]
  		Table 5. Variables Influencing Optimal Period of Administration of Neuroprotective Therapy


View this table:
[in this window]
[in a new window]
  		Table 6. Variability in Duration of Neuroprotective Therapy in Current Clinical Trials of Agents Modulating Excitatory Amino Acid Transmission


*    Measuring Outcome
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 up arrowSelecting Patients for Trials...
 up arrowOptimum Duration of Treatment
 *Measuring Outcome
down arrowSample Size Estimation
 down arrowCollaboration With Industry
 down arrowSummary
 down arrowReferences
 
Safety
The pharmacological profiles and potential hazards of many neuroprotective agents are not fully characterized. The safety and tolerability of these agents, if they really are effective, will be the key factors in determining their utility. Agents acting on excitatory amino acid transmission seem to produce a similar profile of adverse effects both in experimental models and humans. These may be principally classified into neuropsychiatric and cardiovascular effects.4 The typical neuropsychiatric adverse effects observed in the clinical trials included somnolence, visual hallucinations, and agitation. Other neurological adverse effects included nystagmus, unsteadiness, nausea, and vomiting. The potentially significant cardiovascular effects included variable effects on blood pressure (observed with the NMDA receptor and calcium channel antagonists) heart rate, as well as prolongation of the QT interval (with some of the sodium channel modulators). Although these adverse effects have been dose dependent and generally transient, resolving spontaneously on treatment withdrawal, there is a definite possibility that significant early hazard might result from the use of these agents in patients with acute stroke. Even if there is no early hazard, the nature of the adverse events may reduce the practicability and tolerability of some agents.

Some of the agents modulating excitatory amino acid transmission appear to disrupt memory and learning5 and cause morphological alterations, which are typified by the reversible cellular swelling and vacuolation seen with MK-801 and other noncompetitive NMDA antagonists in large cortical neurons in layers III to IV in rats.54 This raises concerns that long-term cognitive and neuropsychiatric adverse effects could be a problem, particularly in older stroke patients who are at high risk of developing cognitive impairment even without treatment.55

Future trials must be designed to detect these potential risks and provide precise estimates of the risk of any early hazard or of any long-term adverse neuropsychological consequences. An early hazard might manifest in a number of different ways: an increased risk of falls, injury, dehydration, impaired nutrition, or perhaps a consequent increased risk of life-threatening chest infection or thromboembolism. Clinical trials of neuroprotective therapy should monitor these complications prospectively. However, it is essential that protocols designed to detect these diverse complications do not become burdensome for investigators: they must be kept simple if thousands of patients are to be randomized (see below). Alternatively, more complex protocols to detect these adverse effects could be applied to a random sample of patients in the trial.

Do We Need Novel Outcome Measures?
The choice of the most appropriate measures of outcome is a key aspect of any clinical trial design. The simplicity, clinical relevance, reliability, and validity of the chosen measures will not only influence the ease and completeness of data collection but may even affect the eventual interpretation of the trial results.

Clinicians are reluctant to change their practice even after the publication of powerful studies that show significant advantages for new treatments.56 57 This might be partly explained by concerns regarding both the clinical relevance of the outcome measures used and the cost effectiveness of the intervention. Cost-benefit analyses and QOL evaluations have an important bearing in some countries on the licensing of novel therapies for routine clinical practice.56 58

In very early clinical trials of patients with acute stroke, outcome was measured solely in terms of case fatality or "neurological deterioration." However, the importance of describing functional outcome more precisely in survivors was soon appreciated because intervention might save lives at the cost of increasing the number of survivors with severe disability. Stroke scales, derived by arbitrarily summing scores related to various physical impairments detected during neurological examination, became a popular means for describing outcome. A number of serious shortcomings associated with the use of these scales have been recognized,59 perhaps the most serious of which is that stroke scales focus on impairments (the neurological signs), which are much less important to patients than disability (what the patient can actually do, ie, the end result of the impairments). Thus, outcomes in survivors of acute stroke are now frequently assessed by measurement of disability in performance of activities of daily living. However, even this approach has been criticized, since it takes no account of psychological and social health outcomes. Important psychological outcomes include cognitive functioning, emotional status, general perception of health, life satisfaction, and happiness. "Social health" can be assessed by the number and quality of social interactions. Interest in the measurement of these outcomes, which are potentially more relevant to the patient's perspective, has led to the development of a number of generic health-related QOL measures that try to measure outcomes not just in physical but also in psychological and social domains.

The information generated from some of these QOL instruments might be considered inappropriately detailed for use in large clinical trials, since it is widely held that relatively crude and simple measures of outcome can reliably distinguish differences between treatment and control groups.60 61 However, given the potential for long-term neuropsychiatric adverse effects after neuroprotective therapy, designers of trials should consider measuring psychological and social functioning, which could be done simply with a health-related QOL instrument. Furthermore, some of these QOL measures generate an overall estimate of health status62 for use in cost-benefit analyses, which will probably become a standard requirement of the regulatory agencies. Although a variety of generic QOL measures have been validated in a number of disease states and languages, the routine use of these instruments in stroke trials can only be justified once their feasibility, reliability, and validity have been demonstrated.

Surrogate Measures
It has been suggested that neuroimaging techniques might be a more efficient way to assess the efficacy of neuroprotective treatments.63 64 These include spin-echo MRI, perfusion and diffusion-weighted MRI, MR spectroscopy, and PET. Diffusion-weighted MRI is perhaps the most promising of these techniques; it has already been used to demonstrate neuroprotective effects in experimental models.65 66 MRI methods are attractive because they might allow a sensitive, very early assessment of treatment effect, which could be measured within a few days (instead of needing to wait 3 months for the measurement of stable clinical functional outcomes). This reduces the effect of the many potential sources of "noise" that affect measures of an individual's level of disability or QOL. However, the validity and reliability of these approaches are yet to be determined in patients with stroke. Even if such surrogate measures prove to be satisfactory as a means of screening agents for further development, large reliable trials with clinically relevant measures of outcome will still be required because clinicians are likely to change their practice only if there is proof beyond reasonable doubt that treatment not only saves lives but also reduces the probability of a poor clinical outcome (particularly if the outcome avoided is severe disability or handicap). Furthermore, regulatory agencies are likely to require convincing evidence of worthwhile clinical benefit before granting a product license.


*    Sample Size Estimation
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 up arrowSelecting Patients for Trials...
 up arrowOptimum Duration of Treatment
 up arrowMeasuring Outcome
 *Sample Size Estimation
down arrowCollaboration With Industry
 down arrowSummary
 down arrowReferences
 
What Size of Treatment Effect Can Reliably Be Detected With 1500 Patients?
Mortality
Most of the published or ongoing controlled placebo trials of neuroprotective therapy plan to randomize 1500 patients or fewer. How large a treatment effect can reliably be detected with this sample size? If approximately 10% of the control group are dead at 14 days, the statistical power of a study involving 1500 patients would be completely inadequate to detect a 20% relative reduction in the proportion of patients dead at 14 days (ie, 10% of control subjects reduced to 8% of treated patients) (Table 7Down). Yet this reduction corresponds to the avoidance of 20 deaths per 1000 patients treated, a moderate and humanly worthwhile absolute benefit if the treatment is both safe and not too costly.


View this table:
[in this window]
[in a new window]
  		Table 7. Effect of Trial Size on Reliability of a Trial Result*

Disability
A study with 1500 patients would have better power (approximately 90%) to detect a similar relative reduction in the proportion of patients dead or dependent at 6 months (ie, 60% of control patients reduced to 50% of treated patients); this effect would correspond to the avoidance of 100 patients being dead or dependent for every 1000 treated, an extremely large absolute benefit from treatment. But, more modest absolute effects, for example, 20 events avoided per 1000 patients treated, might still be considered worthwhile, especially if the treatment were simple, widely practicable, and relatively nontoxic (aspirin or thrombolytic therapy with streptokinase have this magnitude of benefit in patients with acute myocardial infarction and have been widely adopted).12 However, a trial with 1500 patients would have less than a 5% probability of reliably detecting this more modest but nonetheless worthwhile benefit (FigureDown).



View larger version (19K):
[in this window]
[in a new window]
  		Figure 1. Relationship between the size of treatment effect to be detected and the statistical power of a study to detect a given size of treatment effect using two different sample sizes (n=1500 and n=20 000). The measure of outcome is the number of patients dead or dependent at 6 months after randomization. Treatment effect is defined in absolute terms (ie, the number of events avoided per 1000 patients treated); for example, if 60% of control patients versus 57% of patients allocated active treatment are dead or dependent at 6 months, this is an absolute difference of 3% or 30 per 1000 patients treated. If {alpha}=.01, the power of a study with 1500 patients to detect such a difference is 8% and for a study with 20 000 it is 96%.

Extreme Benefits From Neuroprotection Are Unlikely
Treatments that have only a moderate effect on major outcomes such as death or "being dead or disabled" are more plausible than ones with extreme effects for a number of reasons.67 First, in ischemic neuronal damage many different pathophysiological mechanisms interact, so it is unlikely that modification of any one mechanism in the acute phase by a particular intervention will have a large effect on clinical outcome some months after the stroke (which may be influenced by many other factors). Furthermore, the indirect relationship between the volume of brain damaged by ischemia and the final degree of disability suggests that benefits from neuroprotection in acute ischemic stroke might plausibly be of the order of 20 to 50 events avoided per 1000 patients treated.

What Sample Size Would Be Required to Reliably Detect These Moderate Clinical Benefits?
The FigureUp illustrates the relationship between treatment effect and trial power. Reliable assessment of both moderate treatment effects and the effect of treatment on mortality requires a very large trial in which the intervention is tested in several tens of thousands of patients (FigureUp and Table 7Up). Importantly, such a trial would minimize the risk of falsely concluding that the treatment was ineffective when in truth it was beneficial. Furthermore, a trial of this magnitude would allow prespecified subgroup analyses of reasonable statistical power.

Limitations of Large Simple Trials
Large simple trials may not be the most appropriate first step in evaluating neuroprotective therapy because empirical evidence suggests they may suffer from null bias, which leads to an underestimation of the true effect size.68 Null bias arises when the contrast between the treatment and control groups is blunted by lack of constraint on nontrial therapies or by inaccuracy of data; both are more likely to occur in large studies when recruitment is distributed across many centers and countries. Thus, in the evaluation of neuroprotective treatments, a role exists for both large simple studies with high statistical power and exploratory studies that examine the efficacy of treatment in a smaller, homogenous, selected samples of patients.


*    Collaboration With Industry
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 up arrowSelecting Patients for Trials...
 up arrowOptimum Duration of Treatment
 up arrowMeasuring Outcome
 up arrowSample Size Estimation
 *Collaboration With Industry
down arrowSummary
 down arrowReferences
 
Physicians will need to collaborate with the pharmaceutical industry to evaluate most of the currently available neuroprotective treatments for stroke. This might take one of two different forms: at one extreme, physicians might enroll patients in clinical trials designed, coordinated, analyzed, and reported solely by the pharmaceutical company; at the other end of the spectrum the trial is conducted, analyzed, and—most importantly—reported independently of the company by a collaborative group. This latter model has advantages for both parties: the pharmaceutical company gains credibility from the independence of such a trial, and collaborating clinicians gain the opportunity and resources to conduct investigator-driven studies with correspondingly greater freedom for scientific innovation. Recent studies of neuroprotective agents conducted by the pharmaceutical industry all have followed very similar designs and have not taken advantage of the opportunity for innovation and simplification.14 Furthermore, the system of substantial reimbursements to investigators for every patient randomized may make very large trials simply too expensive.14

The selection of the neuroprotective agents for clinical evaluation has largely been driven by commercial factors. Thus, treatments with limited or no commercial potential, such as intravenous phenytoin, magnesium, and hypothermia, have not been adequately tested in patients with acute stroke. Factoring these treatments into sponsored trials of proprietary neuroprotective agents would allow the most efficient use of the collaborators' effort. Furthermore, if these treatments are not factored into trials testing other agents, it is unlikely that they ever will be tested on an appropriately large scale in a "single-agent" trial.


*    Summary
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 up arrowSelecting Patients for Trials...
 up arrowOptimum Duration of Treatment
 up arrowMeasuring Outcome
 up arrowSample Size Estimation
 up arrowCollaboration With Industry
 *Summary
down arrowReferences
 
Any therapeutic trial of a new treatment for stroke must provide sufficient reliable evidence to convince clinicians and healthcare purchasers of its merits. Clinicians are most likely to be convinced by large independent studies that provide clear evidence of benefit. If the trial is really to alter healthcare delivery, it should also confirm that the treatment is cost-effective enough to satisfy the increasingly critical demands of the healthcare purchasers. Although some of the current trials will be able to detect large benefits, reliable detection of the moderate benefits that seem more plausible with neuroprotection will need to wait until completion of trials that are perhaps an order of magnitude larger. If tens of thousands of patients are to be recruited into trials of neuroprotective therapy, it is essential that the trials have simple practicable designs that allow participation not only by interested university hospitals but also busy general hospitals with few research resources.


*    Selected Abbreviations and Acronyms
 
ECASS = European Cooperative Acute Stroke Study
ECST = European Carotid Surgery Trial
ICH = intracerebral hemorrhage
ISIS = International Study of Infarct Survival
IST = International Stroke Trial
NMDA = N-methyl-D-aspartate
PET = positron emission tomography
QOL = quality of life
RCT = randomized controlled trial


*    Acknowledgments
 
Dr Paul Dorman is supported by a UK Medical Research Council clinical training fellowship, and Dr Peter Sandercock is supported by a grant from the UK Medical Research Council. We would like to acknowledge the helpful comments of Professor Charles Warlow, Professor Peter Koudstaal, Dr Stephano Ricci, Dr Heinrich Mattle, and Dr Bobby Mulrooney in the preparation of this article.


*    Footnotes
 
Drs Sandercock and Dorman are both funded by the Medical Research Council of the United Kingdom. However, Dr Sandercock has received a substantial grant from GlaxoWellcome PLC for the planning of an investigator-driven, collaborative, multicenter trial of neuroprotection with 619 C89 (Pilot IST-2).

Received January 16, 1996; revision received April 22, 1996; accepted May 31, 1996.


*    References
up arrowTop
 up arrowIntroduction
 up arrowWhich Agent to Choose?
 up arrowSelecting Patients for Trials...
 up arrowOptimum Duration of Treatment
 up arrowMeasuring Outcome
 up arrowSample Size Estimation
 up arrowCollaboration With Industry
 up arrowSummary
 *References
 

  1. Sandercock PAG, Celani MG, Ricci S. The likely public health impact in Europe of simple treatments for acute ischaemic stroke. Cerebrovasc Dis. 1992;2:236. Abstract.
  2. Sandercock PAG, Willems H. Medical treatment of acute ischaemic stroke. Lancet. 1992;339:537-539.[Medline] [Order article via Infotrieve]
  3. Heiss W-D, Graf R. The ischaemic penumbra. Curr Opin Neurol. 1994;7:11-19.[Medline] [Order article via Infotrieve]
  4. Muir KW, Lees KR. Clinical experience with excitatory amino acid antagonist drugs. Stroke. 1995;26:503-513.[Abstract/Free Full Text]
  5. McCulloch J. Excitatory amino acid antagonists and their potential for the treatment of ischaemic brain damage in man. Br J Clin Pharmacol. 1992;34:106-114.[Medline] [Order article via Infotrieve]
  6. Counsell C, Sandercock PAG. The management of patients with acute ischaemic stroke. Curr Med Lit Ger. 1995;7:99-110.
  7. Dorman PJ, Counsell C, Sandercock PAG. Recently developed neuroprotective therapies for acute stroke: a qualitative systematic review of clinical trials. CNS Drugs. 1996;5:457-474.
  8. Davis SM, Donnan A. Patients, centres, admission procedures and non-randomized treatments around the world. Cerebrovasc Dis. 1995;5(suppl 1):16-21.
  9. Bogousslavsky J. Acute stroke trials: from morass to nirvana? Cerebrovasc Dis. 1995;5:3-6.
  10. van Gijn J. Extrapolation of trials data into practice: where is the limit? Cerebrovasc Dis. 1995;5:159-162.
  11. Sandercock P. Major ongoing stroke trials: the International Stroke Trial—Main. Stroke. 1995;26:1142. Abstract.
  12. ISIS-2 Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, both, or neither among 17 187 cases of suspected acute myocardial infarction. Lancet. 1988;2:349-360.[Medline] [Order article via Infotrieve]
  13. ISIS-3 Collaborative Group. ISIS-3: a randomised comparison of streptokinase vs tissue plasminogen activator vs anistreplase and of aspirin plus heparin vs aspirin alone among 41 299 cases of suspected acute myocardial infarction. Lancet. 1992;339:753-770.[Medline] [Order article via Infotrieve]
  14. Adams RJ, Fisher M, Furlan AJ, del Zoppo G. Acute stroke treatment trials in the United States: rethinking strategies for success. Stroke. 1995;26:2216-2218.[Free Full Text]
  15. Trust Study Group. Randomised, double-blind, placebo-controlled trial of nimodipine in acute stroke. Lancet. 1990;336:1205-1209.[Medline] [Order article via Infotrieve]
  16. Rothwell PM, Warlow CP. The European Carotid Surgery Trial (ECST). In: Greenhalgh RM, Hollier LH, eds. Surgery for Stroke. London, UK: WB Saunders Co; 1993.
  17. Baron JC, von Kummer R, del Zoppo GJ. Treatment of acute ischemic stroke: challenging the concept of a rigid and universal time window. Stroke. 1995;26:2219-2221.[Free Full Text]
  18. Fisher M, Bogousslavsky J. Evolving toward effective therapy for acute ischemic stroke. JAMA. 1993;270:360-364.[Abstract]
  19. Frey JL. The future of acute stroke treatment: lessons we have learned from laboratory and clinical research. Arch Intern Med. 1995;155:1005.[Medline] [Order article via Infotrieve]
  20. Barsan WG, Brott TG, Broderick JP, Haley EC, Levy DE, Marler JR. Urgent therapy for acute stroke: effects of a stroke trial on untreated patients. Stroke. 1994;25:2132-2137.[Abstract]
  21. Grotta J. The current status of neuronal protective therapy: why have all the neuronal protective drugs worked in animals but none so far in stroke patients? Cerebrovasc Dis. 1994;4:115-120.
  22. Kay R, Wong K, Yu Y, Chan Y, Tsoi T, Ahuja A, Chan F, Fong K, Law C, Wong A, Woo J. Low-molecular-weight heparin for the treatment of acute ischemic stroke. N Engl J Med. 1995;333:1588-1593.[Abstract/Free Full Text]
  23. Wise RJS, Bernardi S, Frackowiak RSJ, Legg NJ, Jones T. Serial observations on the pathophysiology of acute stroke: the transition from ischaemia to infarction as reflected in regional oxygen extraction. Brain. 1983;106:197-222.[Abstract/Free Full Text]
  24. Heiss WD, Huber M, Fink GR, Herholz K, Pietrzyk U, Wagner R, Wienhard K. Progressive derangement of periinfarct viable tissue in ischemic stroke. J Cereb Blood Flow Metab. 1992;12:193-203.[Medline] [Order article via Infotrieve]
  25. Baron JC. Ischemic stroke studied by 150-labelled compounds: misery perfusion and luxury perfusion. In: Heiss WD, Pawlik G, Herholz K, Wienhard K, eds. Clinical Efficacy of Positron Emission Tomography. Boston, Mass/Dortrecht, Netherlands: Martinus Nijhoff Publishing; 1987:15-23.
  26. Heiss WD, Fink GR, Huber M, Herholz K. Positron emission tomography imaging and the therapeutic window. Stroke. 1993;24(suppl):I-50-I-53.
  27. Warach S, Benfield A, Edelman RR. Prolonged evolution of lesion volume in human stroke determined by diffusion-weighted MRI. Neurology. 1995;45(suppl 4):287. Abstract.
  28. Wiebers DO, Adams HP, Whisnant JP. Animal models of stroke: are they relevant to human disease? Stroke. 1990;21:1-3.[Free Full Text]
  29. Mathur VS, Guinn GA, Burris WH. Maximal revascularization (reperfusion) in intact conscious dogs after 2 to 5 hours of coronary occlusion. Am J Cardiol. 1975;36:252-261.[Medline] [Order article via Infotrieve]
  30. Reimer KA, Lowe JE, Rasmussen MM, Jennings RB. The wavefront phenomenon of ischemic cell death, I: myocardial infarct size vs duration of coronary occlusion in dogs. Circulation. 1977;56:786-792.[Abstract/Free Full Text]
  31. O'Brien CM, Carroll M, O'Rourke PT, Rhodes EL, Gago O, Kirsh MM, Morris JD, Sloan HE. The reversibility of acute ischemic injury to the myocardium by restoration of coronary flow. J Thorac Cardiovasc Surg. 1972;64:840-846.[Medline] [Order article via Infotrieve]
  32. Wilcox RG, von der Lippe G, Olsson CG, Jensen G, Skene AM, Hampton JR. Trial of plasminogen activator for mortality reduction in acute myocardial infarction: Anglo- Scandinavian Study of Early Thrombolysis (ASSET). Lancet. 1988;2:525-530.[Medline] [Order article via Infotrieve]
  33. ISAM Study Group. A prospective trial of intravenous streptokinase in acute myocardial infarction (ISAM): mortality, morbidity and infarct size at 21 days. N Engl J Med. 1986;314:1465-1471.[Abstract]
  34. Morris AD, Grosset DG, Squire IB, Lees KR, Bone I, Reid JL. The experiences of an acute stroke unit: implications for multicentre acute stroke trials. J Neurol Neurosurg Psychiatry. 1993;56:352-355.[Abstract]
  35. LaRue LJ, Aler M, Traven ND, Sterman AB, Sobel E, Kleiner J. Acute stroke therapy trials: problems in patient accrual. Stroke. 1988;19:950-954.[Abstract/Free Full Text]
  36. Harper GD, Haigh RA, Potter JF, Castleden CM. Factors delaying hospital admission after stroke in Leicestershire. Stroke. 1992;23:835-838.[Abstract/Free Full Text]
  37. Feldmann E, Gordon N, Brooks JM, Brass LM, Fayad PB, Sawaya KL, Nazareno F, Levine SR. Factors associated with early presentation of acute stroke. Stroke. 1993;24:1805-1810.[Abstract/Free Full Text]
  38. Alberts MJ, Bertels C, Dawson DV. An analysis of time of presentation after stroke. JAMA. 1990;263:65-68.[Abstract]
  39. Alberts MJ, Perry A, Dawson DV, Bertels C. Effects of public and professional education on reducing the delay in presentation and referral of stroke patients. Stroke. 1992;23:352-356.[Abstract/Free Full Text]
  40. Wroe SJ, Sandercock P, Bamford J, Dennis M, Slattery J, Warlow C. Diurnal variation in incidence of stroke: Oxfordshire community stroke project. BMJ. 1992;304:155-157.
  41. American Nimodipine Study Group. Clinical trial of nimodipine in acute ischemic stroke. Stroke. 1992;23:3-8.[Abstract/Free Full Text]
  42. Counsell CE, Clarke MJ, Slattery J, Sandercock PAG. The miracle of DICE therapy for acute stroke: fact or fictional product of subgroup analysis? BMJ. 1994;309:1677-1681.[Abstract/Free Full Text]
  43. Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R, Boysen G, Bluhmki E, Hoxter G, Mahagne MH, Hennerici M, for the ECASS Study Group. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke: the European Cooperative Acute Stroke Study (ECASS). JAMA. 1995;274:1017-1025.[Abstract]
  44. Collins R, Gray R, Godwin J, Peto R. Avoidance of large biases and large random errors in the assessment of moderate treatment effects: the need for systematic overviews. Stat Med. 1987;6:245-250.[Medline] [Order article via Infotrieve]
  45. 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.[Abstract/Free Full Text]
  46. Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project 1981-1986, II: incidence, case fatality and overall outcome at one year of cerebral infarction, primary intracerebral haemorrhage and subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry. 1990;53:16-22.[Abstract]
  47. Tanaka H, Ueda Y, Date C, Baba T, Yamashita H, Hayashi M, Shoji H, Owada K, Baba K, Shibuya M, Kon T, Detels R. Incidence of stroke in Shibata, Japan: 1976-1978. Stroke. 1981;12:460-466.[Abstract/Free Full Text]
  48. Mendelow D. Spontaneous intracerebral haemorrhage. J Neurol Neurosurg Psychiatry. 1991;54:193-195.[Medline] [Order article via Infotrieve]
  49. Nehls DG, Mendelow D, Graham DI, Sinar EJ, Teasdale GM. Experimental intracerebral hemorrhage: progression of hemodynamic changes after production of a spontaneous mass lesion. Neurosurgery. 1995;23:439-444.
  50. Bullock R, McCulloch J, Graham DI, Lowe D, Chen MH, Teasdale GM. Focal ischemic damage is reduced by CPP-ene studies in two animal models. Stroke. 1990;21(suppl):III-32-III-36.
  51. Choksey MS, Costa DC, Iannotti F, Ell PJ, Crockard HA. 99TCm-HMPAO SPECT studies in traumatic intracerebral haematoma. J Neurol Neurosurg Psychiatry. 1991;54:6-11.[Abstract]
  52. Dorman PJ, Sandercock PAG. Should patients with visible infarction on the admission CT scan be excluded from acute stroke trials? An analysis of data from 12,450 patients randomised in the IST. In: Program of the Meeting of the Association of British Neurologists; April 11-12, 1996; Norwich, UK. Abstract.
  53. Saunders DE, Howe FA, van den Boogaart A, McClean MA, Griffiths JR, Brown MM. Continuing ischemic damage after acute middle cerebral artery infarction in humans demonstrated by short-echo proton spectroscopy. Stroke. 1995;26:1007-1013.[Abstract/Free Full Text]
  54. Olney JW, Labruyere J, Price MT. Pathological changes induced in cerebrocortical neurons by phenylcyclidine and related drugs. Science. 1989;244:1360-1362.[Abstract/Free Full Text]
  55. Tatemichi TK, Desmond DW, Stern Y, Paik M, Sano M, Bagiella E. Cognitive impairment after stroke: frequency, patterns and relationship to functional abilities. J Neurol Neurosurg Psychiatry. 1994;57:202-207.[Abstract]
  56. Anonymous. Clinical trials and clinical practice. Lancet. 1993;342:877-878.[Medline] [Order article via Infotrieve]
  57. Ketley D, Woods KL. Impact of clinical trials on clinical practice: example of thrombolysis for acute myocardial infarction. Lancet. 1993;342:891-894.[Medline] [Order article via Infotrieve]
  58. Freemantle N, Henry D, Maynard A, Torrance G. Promoting cost effective prescribing. BMJ. 1995;310:955-956.[Free Full Text]
  59. van Gijn J, Warlow C. Down with stroke scales. Cerebrovasc Dis. 1992;2:244-246.
  60. Peto R. Monitoring cancer patients in clinical trials need not be precise. In: Symington T, Williams AE, McVie JG, eds. Cancer: Assessment and Monitoring—Tenth Pfizer International Symposium. Edinburgh, Scotland: Churchill Livingstone Inc; 1980:377-381.
  61. Lindley RI, Waddell F, Livingstone M, Warlow C, Dennis M, Sandercock PAG, Smith B. Can simple questions assess outcome after stroke? Cerebrovasc Dis. 1994;4:314-324.
  62. The EuroQol Group. EuroQol: a new facility for the measurement of health related quality of life. Health Policy. 1990;16:199-208.[Medline] [Order article via Infotrieve]
  63. Yatsu FM, Yeakley JW, Fenstermacher MJ. The role of neuro-imaging in acute ischaemic stroke therapy. In: Amery WK, Bousser M-G, Clifford Rose F, eds. Clinical Trial Methodology in Stroke. London, UK: Bailliere Tindall; 1989:75-90.
  64. Warach S, Gaa J, Siewert B, Wielopolski P, Edelman RR. Acute human stroke studied by whole brain echo planar diffusion-weighted magnetic resonance imaging. Ann Neurol. 1995;37:231-241.[Medline] [Order article via Infotrieve]
  65. Muller TB, Haraldseth O, Jones RA, Sebastiani G, Lindboe CF, Unsgard G, Oksendal AN. Perfusion and diffusion-weighted MR imaging for in vivo evaluation of treatment with U74389G in a rat stroke model. Stroke. 1995;26:1453-1458.[Abstract/Free Full Text]
  66. Minematsu K, Fisher M, Li L, Davis MA, Knapp AG, Cotter RE, McBurney RN, Sotak CH. Effects of a novel NMDA antagonist on experimental stroke rapidly and quantitatively assessed by diffusion-weighted MRI. Neurology. 1993;43:397-403.[Abstract/Free Full Text]
  67. Yusuf S, Collins R, Peto R. Why do we need some large, simple randomized trials? Stat Med. 1984;3:409-420.[Medline] [Order article via Infotrieve]
  68. Woods KL. Mega-trials and management of acute myocardial infarction. Lancet. 1995;346:611-614.[Medline] [Order article via Infotrieve]
  69. Hommel M, Boissel JP, Cornu C, Boutitie F, Lees KR, Besson G, Leys D, Amarenco P, Bogaert M, for the Multicentre Acute Stroke Trial Study Group. Termination of trial of streptokinase in severe acute ischaemic stroke. Lancet. 1995;345:57.
  70. 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;346:1509-1514.[Medline] [Order article via Infotrieve]
  71. Donnan GA, Davis S, Chambers BR, Gates PC, Hankey GJ, McNeil JJ, Rosen D, Stewart-Wynne EG, Tuck RR. Trials of streptokinase in severe acute stroke: the Australian Streptokinase Study. Lancet. 1995;345:578-579.[Medline] [Order article via Infotrieve]
  72. Muir KW, Lees KR, Hamilton SJC, George CF, Hobbiger SF, Lunnon MW. A randomized, double-blind, placebo-controlled ascending dose tolerance study of 619C89 in acute stroke. In: Program of the Second International Conference on Neuroprotective Therapy; July 31-August 3, 1994; New York, NY. Abstract.
  73. Otomo E, Araki G, Mori A, Kurihara M. Clinical evaluation of GABA in the treatment of cerebrovascular disorders: multi-center double blind study in comparison with pyrithioxine and placebo. Arzneimittelforschung. 1981;9:1511-1523.
  74. Wahlgren NG, for the Class Study Group. The Chlormethiazole Acute Stroke Study (CLASS). Cerrebrovasc Dis. 1996:6:190. Abstract.
  75. Huber M, Kittner B, Hojer C, Fink GR, Neveling M, Heiss W-D. Effect of propentofylline on regional cerebral glucose metabolism in acute ischaemic stroke. J Cereb Blood Flow Metab. 1993;13:526-530.[Medline] [Order article via Infotrieve]
  76. Albers GW, Saenz RE, Moses JA, Choi DW. Safety and tolerance of oral dextromethorphan in patients at risk for brain ischemia. Stroke. 1991;22:1075-1077.[Abstract/Free Full Text]
  77. Albers GW, Atkinson R, Kelley R, Rosenbaum DM, for the Dextrorphan Study Group. Safety, tolerability, and pharmacokinetics of the N-methyl-D-aspartate antagonist dextrorphan in patients with acute stroke. Stroke. 1995;26:254-258. Abstract.[Abstract/Free Full Text]
  78. Fisher M, for the CNS 1102-003 Investigators. Cerestat (CNS 1102) a non-competitive NMDA antagonist, in ischemic stroke patients: dose-escalating safety study. Cerebrovasc Dis. 1994;4:245. Abstract.
  79. Strand T, Wester PO, for the CVD Group. A double blind randomized pilot trial of magnesium therapy in acute cerebral infarction. In: Program of the 7th Nordic Meeting on Cerebrovascular Disease; Jyraskytla Finland; August 14-17, p 37. Abstract.
  80. Muir KW, Lees KR. A randomized double-blind placebo-controlled pilot trial of intravenous magnesium sulfate in acute stroke. Stroke. 1995;26:1183-1188.[Abstract/Free Full Text]
  81. Grotta J, for the CGS19755 Study Group. Safety and tolerability of the glutamate antagonist CGS19755 in acute stroke patients. Stroke. 1994;25:255. Abstract.
  82. Clark WM, Coull BM. Randomised trial of CGS19755, a glutamate antagonist in acute ischaemic stroke treatment. Neurology. 1994;44(suppl 2):A270. Abstract.
  83. Diener HC, Hacke W, Hennerici M, de Keyser J, Radberg J, Hantson L, for the LUB-INT-4 Study Group. The effects of lubeluzole in the acute treatment of ischemic stroke: results of a phase-2 trial. Stroke. 1995;26:185. Abstract.
  84. Argentino C, Sacchetti ML, Toni D, Savoini G, D'Arcangelo E, Erminio F, Federico F, Milone FF, Gallai V, Gambi D, et al. GM1 ganglioside therapy in acute ischemic stroke. Stroke. 1989;20:1143-1149.[Abstract/Free Full Text]
  85. Hoffbrand BI, Bingley PJ, Oppenheimer SM, Sheldon CD. Trial of ganglioside GM1 in acute stroke. J Neurol Neurosurg Psychiatry. 1988;51:1213-1214.[Abstract]



This article has been cited by other articles:


Home page
 NeurologyHome page
D. S. Liebeskind, C. S. Kidwell, J. W. Sayre, and J. L. Saver
Evidence of publication bias in reporting acute stroke clinical trials.
Neurology, September 26, 2006; 67(6): 973 - 979.
[Abstract] [Full Text] [PDF]

Home page
 StrokeHome page
F. B. Young, K. R. Lees, C. J. Weir, and for the GAIN International Trial Steering Committe
Improving Trial Power Through Use of Prognosis-Adjusted End Points
Stroke, March 1, 2005; 36(3): 597 - 601.
[Abstract] [Full Text] [PDF]