Steroids may have a role in stroke therapy.

Few therapeutic responses are more dramatic than the response to corticosteroids administered overnight to a drowsy and hemiplegic patient with a cerebral tumor, who next morning is alert with minimal neurological disability, even though this effect may be short lived. Unfortunately, no such dramatic response is seen in patients with ischemic or hemorrhagic stroke; but does this mean that such therapy is totally ineffective? The categorization of cerebral edema by Klatzo and Seitelberger1 into “cytotoxic” and “vasogenic” holds the key to this therapeutic response. Cytotoxic, or “intracellular,” edema represents the earliest response to cerebral damage, and whether ischemic or traumatic, is due to the breakdown of cellular ionic pumps causing ingress of water into the cell, and confined within the cellular membrane. The clinical effects and response to corticosteroids of this immediate type of edema is uncertain. Vasogenic edema, occurring hours later, is due to damage to the blood-brain barrier, which becomes “leaky,” allowing extravasation of water, electrolytes, and soon protein into the parenchyma. This produces clinically significant brain swelling, resulting in distortion and herniation of brain and causing neurological disability and death. In brain tumors, the normally tight vascular junctions are pathologically separated, and cellular physiology (such as pinocytosis) is otherwise disturbed,2 resulting in severe vasogenic edema, which is highly responsive to corticosteroids.3 This fundamental difference in pathophysiology between the 2 types of lesions may explain the apparent difference in their initial therapeutic response. There is overwhelming evidence experimentally in a variety of mammalian models that corticosteroids effectively reduce ischemic cerebral edema, both focal and generalized, although the credibility gap needed to extrapolate to the human brain, as always, remains uncertain. In a rather contrived but nevertheless convincing series of experiments, de Courten-Myers et al,4 using hyperglycemic cats, occluded the middle cerebral arteries for 4 hours, administering highdose corticosteroids 30 minutes after occlusion. Compared with untreated controls, there was a highly significant, 6-fold difference in the size of the resulting infarcts in favor of the treated group. Paradoxically, there was no significant difference in the acute death rate between the 2 groups of animals, but whether this represents the difference between the sensitivity of response to the drug in cytotoxic versus vasogenic edema must remain speculative. Also, there are other unexplored effects of corticosteroids on the brain, such as the unanticipated but significantly increased cerebral blood flow in the treated group of animals. Effective reduction of cerebral edema after corticosteroids has also been documented in animal models of global ischemia.5 Hemorrhagic stroke, in both clinical and experimental trials, has been relatively neglected, probably because of a sense of therapeutic nihilism, but there is a rationale for corticosteroid therapy in the reduction of perihematomal edema. Using a pig model, Wagner et al6 demonstrated that serum proteins accumulated in the white matter around hematoma (produced by injecting autologous blood into the frontal lobe), resulting in rapid and prolonged cerebral edema. Similarly, CT brain scanning of patients with spontaneous intracerebral hemorrhage showed that the volume of perihematomal edema increased by 75% in the first 24 hours after the event7 and so should be susceptible to edemareducing drugs such as corticosteroids. Early reports of the beneficial effects corticosteroid therapy in inflammatory and neoplastic brain lesions made it a victim of its own success in stroke patients, since a flurry of early clinical trials in anticipation of similar benefits were conducted before enough was known about either the natural history of stroke or correct methodology of clinical trials. Most trials that were conducted were too little or too late, resulting in the premature abandonment of this potential therapeutic avenue. In a recent Cochrane Review,8 only 7 of 22 published trials of corticosteroids in stroke were acceptable for further analysis, and these comprised woefully inadequate numbers of patients (only 453 in total) with no uniformity of evaluation or assessment, and totally disparate conclusions No trials have been conducted since 1986. These generally negative results did not deter physicians from the United States or China from administering corticosteroids in stroke patients, however, and 20% of physicians surveyed in both countries reported that they still used them routinely, at least in ischemic stroke.8 The Cochrane reviewers concluded that with the available data, they were reluctant to advocate a large-scale therapeutic trial, but that corticosteroid therapy could be factored in as an additional arm of a large trial of a more promising drug. Unfortunately, so far significantly more than 100 neuroprotective trials in acute stroke have proven negative, and it is The opinions expressed in this editorial are not necessarily those of the editors or of the American Stroke Association. From the Department of Neurology, University of Toronto, Ontario, Canada. Correspondence to John W. Norris, MD, Dept of Clinical Neurosciences, St Georges Hospital Medical School, London SW17 ORE, United Kingdom. E-mail [email protected] (Stroke. 2004;35:228-229.) © 2003 American Heart Association, Inc.