Using Pathophysiology in Acute Stroke Trials
To the Editor:
Muir and Grosset1 recently gave their views on why neuroprotective therapies have not worked. They raised issues of trial methodology and some pathophysiological considerations. We would like to add further comments from the pathophysiological viewpoint. We believe that assessment of the state of cerebral perfusion—initial hypoperfusion as well as extent of early reperfusion—will permit optimal evaluation of the efficacy of these agents and the mechanisms by which they have their therapeutic effects.
Hypoperfusion is the final common pathway for all etiologic mechanisms of cerebral ischemia. Neuroprotective agents were developed as adjunctive or alternative approaches to the primary strategy of restoring blood flow. The efficacies of the various neuroprotective agents were demonstrated in experimental stroke models, where perfusion status was strictly controlled. In the case of permanent middle cerebral artery occlusion models, limitation of brain damage was found through salvage of the tissue with mild to moderate ischemia around the core of severe ischemia. Other agents were studied in temporary occlusion models of uniform duration.
Unlike experimental stroke models, human stroke is a very heterogeneous process, with marked variation in the mechanism, extent, severity, and duration of ischemia. Depending on the cause of the ischemia, fluctuations in perfusion may also occur. In addition, when early reperfusion does occur, it has been found to be a major determinant of clinical outcome at 3 months,2 and so far the only efficacious therapies have been those that restore perfusion.3 4 5 In MR studies, the major determinant of the enlargement of ischemic lesion volume is the presence of a larger surrounding region of hypoperfusion.6
Therefore, in the design and evaluation of clinical trials of neuroprotective agents, those pathophysiological factors that cannot be controlled for in the design (for example, the extent and severity of hypoperfusion and early reperfusion) could be controlled in the statistical analyses. In prior trials it is possible that beneficial effects were masked by uncontrolled differences in perfusion and reperfusion status between treatment and control groups, or that the agents themselves altered tissue perfusion.
The rational approach to drug development and evaluation is to make use of what knowledge we have about ischemic stroke pathophysiology in humans. The results of the PROACT II study, which tested intra-arterial recombinant prourokinase in acute stroke patients with angiographically demonstrated occlusion at the M1 or M2 segments of the middle cerebral artery, showed that the reperfusion treatment window in human stroke can be extended to 6 hours when patients are selected on the basis of their pathophysiology.4 A current trend is to advocate using MR patterns of diffusion and perfusion abnormalities to guide therapy and clinical trials of neuroprotective agents, but target patterns have yet to be standardized. Several clinical trials using MRI are in progress, and they may answer questions regarding the utility of pathophysiologically based measures in clinical trial selection and evaluation.6
- Copyright © 1999 by American Heart Association
Muir KW, Grosset DG. Neuroprotection for acute stroke: making clinical trials work. Stroke. 1999;30:180–182.
Baird AE, Austin MC, McKay WJ, Donnan GA. Changes in cerebral tissue perfusion during the first 48 hours of ischaemic stroke: relation to clinical outcome. J Neurol Neurosurg Psychiatry. 1996;61:26–29.
Furlan AJ, Higashida R, Wechsler L, Schulkz, PROACT II Investigators. PROACT II: recombinant prourokinase (r-ProUK) in acute cerebral thromboembolism: initial trial results. Stroke. 1999;30:234. Abstract.
Sherman DG, for the STAT Writers Group. Defibrinogenation with Viprinex TM (Ancrod) for the treatment of acute, ischemic stroke. Stroke. 1999;30:234. Abstract.
Warach S, Pettigrew LC, Dashe JF, Pullicino P, Sabounjian L. The effect of citicoline on lesion volume in acute stroke: a multicenter, double-blind placebo-controlled trial. Stroke. 1999;30:243.
We thank Drs Baird and Warach for their comments proposing an alternative approach to the problems that we attempted to highlight. We agree on the starting premise that pathophysiology must be taken into account in clinical trial design and that knowledge of cerebral perfusion is essential. Our suggested solution was to seek a homogeneous study population by restricting trial entry based on perfusion (or similar) imaging: their alternative is to enter all patients and to use post hoc stratification in statistical analyses. This solution has the attraction that it would ensure larger numbers of patients for trials and provide information on drug safety and tolerability in the eventual (broad) target population with clinically diagnosed stroke. Although this may be practical, reliance on post hoc analyses risks compromising the validity of trial results. Unless the baseline perfusion characteristics are factored in at the time of randomization, it is impossible to ensure balanced treatment allocation, and there is increased likelihood of false-positive or false-negative results. Clinical trials would have to be powered on the basis of this post hoc analysis and may therefore need to be much larger. We would suggest that prerandomization perfusion imaging and a clearly defined basis for stratification are essential for either proposal: provided that these prerequisites are addressed, the decision on whether to structure a trial as restrictive or all-inclusive may depend on the stage of drug development and the attitudes of the steering committee.