Stroke Treatment: Title and subTitle BreakPromising but Still Struggling

Just 2 scant decades ago, the stroke exhibit at the Museum of Science and Industry in Chicago, Ill, featured wheelchairs, braces, and various other devices that aided mobilization of patients with paralysis after strokes. While these aids help improve function for those with residual impairment, prevention and treatment were not addressed. Not surprisingly, stroke was then and remains today the single disease that older individuals most fear. No one wants to spend their crowning years dependent on others and with impaired mental and physical faculties. Even so, during the first three quarters of the 20th century, stroke was a disorder of little medical interest. Most physicians thought there was little that could be done to prevent stroke or treat patients with stroke. Few physicians were interested in stroke. Therapeutic nihilism prevailed.

The atmosphere has changed dramatically during the last decade of this century. Stroke, the third leading cause of death in the world¹ and an even more important cause of serious disability, is finally beginning to get its due. Thousands of physicians attended the last 3 international stroke meetings, and the mood of attendees and stroke specialists is distinctly optimistic. Journals now receive and publish more articles about stroke than ever before. Therapeutic stroke trials are flourishing.

Two catalysts are responsible for this dramatic turnabout: (1) advances in modern technology that now make possible imaging of the brain, heart, and cervicocranial vasculature, and (2) investment by the pharmaceutical industry in new substances to treat and prevent stroke and in therapeutic trials. Clinicians can now quickly and safely acquire all the information needed to determine the nature, cause, and likely severity of stroke in their patients. Brain imaging (computed tomographic [CT] and magnetic resonance [MR] scans) can differentiate ischemia from hemorrhage and often defines the location, size, and vascular territory of brain infarcts and hemorrhages. Vascular investigations (extracranial and transcranial ultrasound, spiral CT angiography, MR angiography) can demonstrate the location, nature, and severity of vascular occlusive lesions and detect most aneurysms and vascular malformations. Blood tests yield information about the cellular components of blood and its coagulability, and noninvasive cardiac imaging studies can reveal potential sources of brain embolism and the presence of coexisting cardiac disease. This information allows clinicians experienced in treating patients with stroke to choose the most appropriate treatment for their individual patients from a growing list of available therapeutic procedures and pharmacological agents. This theme issue of JAMA devoted to stroke contains 5 articles on stroke treatment^{2 - 6} and provides insights into the promise and problems of stroke therapeutic trials and stroke treatment today. In addition, the April 1998 issue of the Archives of Neurology is devoted to stroke.

The practice and outcomes of a commonly performed surgical procedure to prevent stroke, carotid endarterectomy, are the focus of 2 articles in this issue of JAMA.^{2 - 3} Recent publication of the results of 3 major carotid endarterectomy trials involving both symptomatic and asymptomatic patients^{7 - 9} has increased interest in this surgical procedure. All 3 trials showed that carotid surgery was beneficial in reducing stroke morbidity and mortality in selected patients with severe carotid artery stenosis who were operated on by selected surgeons and cared for by selected physicians at selected hospitals within therapeutic trials. Wennberg and colleagues² analyzed the perioperative mortality of all Medicare patients who had carotid endarterectomy during 1992 and 1993. On average, perioperative mortality was considerably higher than that attained in the trials. Even at trial hospitals, mortality was higher among patients not entered in a trial. Nontrial patients generally were older and had more comorbid illnesses. Mortality correlated with the volume of procedures performed in the hospital; hospitals in which fewer patients had carotid surgery had higher mortality rates.

Cebul et al³ studied the frequency of death and nonfatal stroke during the 30 days after carotid endarterectomy among Medicare patients not enrolled in health maintenance organizations in Ohio who had carotid surgery during 1 year (July 1993-June 1994). About half the patients operated on were asymptomatic or had nonspecific symptoms that were not classifiable as stroke or transient ischemic attack. As in the report by Wennberg et al,² mortality was considerably higher at hospitals that had a low volume of carotid surgeries, whereas surgery at higher-volume hospitals in Ohio conferred a 71% risk reduction for stroke or death.

These 2 studies of carotid surgery clearly show that the results of trials are not generalizable to the population at large. Outcomes depend highly on the risk factors, age, and comorbidities of the individual patient and the volume and track record of the hospital and surgeon. In the Ohio study,³ hospital volume was a more important factor than the number of procedures performed by individual surgeons. Preoperative and postoperative medical and neurologic care, neuroradiological expertise, and anesthesiology experience at high-volume hospitals probably have a greater effect on perioperative outcomes than has been appreciated. Physicians and their patients are entitled to know data regarding volume, mortality, and morbidity of hospitals and surgeons before patients undergo carotid surgery. Nonemergency carotid endarterectomies probably should not be performed at hospitals with low surgical volumes.

Although the results of the TOAST trial⁴ of a heparinoid, danaparoid sodium, in the treatment of patients with acute ischemic stroke are considered "negative" by the trialists, the results might be predicted from knowledge of the effects of heparin compounds. Heparins (heparin sodium, low molecular weight heparins, and heparinoids), when effective, prevent the formation, propagation, and embolization of erythrocyte-fibrin thrombi ("red clots"). The effect of treatment lasts about as long as the drug is given, in this trial 7 days. Heparins should be effective in patients with large-artery atherosclerosis because these patients have a high frequency of acute thromboembolism of red clots. If heparins were effective, they should maximally influence stroke outcome and recurrence in these patients during the 7 days of infusion. The subsequent clinical course depends on the causative vascular process and subsequent treatment. Patients with cardioembolism have a low rate of recurrence during the first week after brain embolization,¹⁰ and patients with penetrating artery disease do not have red clots. These subgroups of acute ischemic stroke would not be expected to respond to a week of heparinoid treatment.

In the TOAST trial, in patients classified as having large-artery atherosclerosis, heparinoid reduced the number of recurrences of stroke during the 7 days of infusion, and the rates of favorable and very favorable outcomes were significantly higher in patients given heparinoid compared with those given placebo.⁴ Among all patients, neurologic improvement and favorable and very favorable outcomes were more common after 7 days in patients treated with heparinoids compared with placebo-treated patients. The outcomes at 3 months were not significantly different but are heavily influenced by subsequent treatment that was chosen by the treating physicians. As expected, brain hemorrhage was more common in patients who received heparinoid.

The results of this trial and the larger International Stroke Trial¹¹ show that heparin compounds (heparin sodium, low molecular weight heparins, and heparinoids) should not be given indiscriminately to all patients with acute brain ischemia. These drugs may be effective in patients with large-artery thromboemboli and probably also are effective in patients with cardiac-source emboli who have a high risk of early recurrence (eg, those with acute myocardial infarction, mitral stenosis with atrial fibrillation, or atrial or ventricular thrombi shown by echocardiography). Selection of anticoagulants or antiplatelet aggregants should be based on the nature and severity of the cardiac and cerebrovascular lesions and the presence and severity of associated brain infarction in the individual patient.

The Stroke Prevention in Atrial Fibrillation (SPAF) Investigators⁵ report the results of their most recent study of the treatment of patients who have atrial fibrillation but do not have valvular heart disease. Warfarin prophylaxis has been shown to be effective in preventing stroke in patients with atrial fibrillation and is more effective than aspirin. However, warfarin treatment has significant risks, especially in elderly patients with comorbid illness. In this article⁵ and other communications,^{12 - 17} the SPAF Investigators have attempted to use historical data and the results of investigations (especially echocardiography) to identify groups of patients who have a relatively low risk of embolism and who could be satisfactorily managed with aspirin prophylaxis. In elderly patients who do not have left atrial enlargement, congestive heart failure, prior brain embolism, low cardiac ejection fraction, or left ventricular fractional shortening of less than 25%, the risk-benefit ratio of aspirin is probably better than that of warfarin, especially for patients with relative contraindications to warfarin.^{5 ,12 - 17} Patients with hypertension were at immediate risk.

Also in this issue of THE JOURNAL, Fisher and Bogousslavsky⁶ present an overview of 2 relatively new stroke therapies—thrombolysis and neuroprotective drug treatment. Unfortunately, randomized trials of these treatments have been performed almost entirely in groups of patients with acute brain ischemia in whom the vascular causes have not been defined. For nearly 3 decades physicians have designed therapeutic trials studying the effectiveness and safety of platelet antiaggregants and anticoagulants in patients grouped because they had brain ischemia (ie, transient ischemic attacks and minor strokes) but without mandated definition or stratification of the underlying cardiac, hematologic, and cerebrovascular causes of the ischemia. These studies have uniformly failed to find a panacea for brain ischemia. The most effective agent has improved outcomes about 20% to 25% vs placebo.¹⁸

Thrombolytic therapy has great promise. The 2 largest randomized trials of intravenous thrombolytic treatment showed that treatment of patients with acute stroke with recombinant tissue-type plasminogen activator was, in general, probably effective.^{19 - 20} Unfortunately, there was also a high rate of hemorrhage, sometimes fatal. In these 2 trials no vascular studies defining the vascular pathology or cause of stroke were reported, and vascular studies were not used as a basis for treatment. Thrombolytic drugs can be given intravenously or can be administered by interventionists through arterial catheters guided to the location of the clots. Intravenous treatment needs no special training or expertise, whereas intra-arterial treatment requires an available, experienced interventionist and necessitates cerebral angiography. Intra-arterial treatment allows delivery of the drug directly into the thrombus and also allows mechanical manipulation of the thrombus; angioplasty and vasodilating agents can be used along with thrombolysis. Drugs given intravenously may not reach a thrombus if the supply arteries are obstructed. If angiography or noninvasive vascular studies (CT angiography, MR angiography, or ultrasound) are not performed, treating physicians will not even know if a thrombus is present at the time of treatment.

Future trials of thrombolytic therapy must thoroughly explore which patients with acute brain ischemia will profit most from treatment and will have the least risk of brain hemorrhage. Some important information about the response of various vascular occlusive lesions is available from nonrandomized studies,²¹ but many questions remain: Which thrombolytic drug should be used at what dose given by which route to which patients with what vascular lesions with what time constraints? What extent, if any, of brain infarction determined by what technology before infusion represents a major risk for thrombolysis? Should heparin or another coagulation-modifying agent or angioplasty be used after thrombolysis and, if so, when?

I believe that present published recommendations for the use of thrombolytic treatments^{22 - 23} are premature. If thrombolytics are given indiscriminately by any physician to all patients suspected of having acute brain ischemia based on CT scans and a 3-hour time window, we will never obtain answers to the important, as yet unanswered questions, and we will needlessly cause many brain hemorrhages and deaths.^{24 - 25} Brain imaging alone is not sufficient to define etiology or guide treatment. In my opinion, thrombolytics should be given only under the direction of stroke specialists (usually neurologists or neurosurgeons) and only after vascular investigations show a thrombus within a large artery. As pointed out by Fisher and Bogousslavsky,⁶ the time window is variable; 3 hours is not sacrosanct but differs in individual patients. Many more therapeutic trial data are needed.

Similarly, although neuroprotective agents are promising, none has yet been shown to be effective in large groups of patients with brain ischemia.⁵ Outcome in patients with brain ischemia probably most depends on the etiology of the ischemia and whether reperfusion (by spontaneous or iatrogenic recanalization) occurs or adequate collateral circulation develops quickly. Neuroprotective drugs may not sufficiently reach the ischemic areas if the major supply arteries are blocked. Instead of spending large amounts of money on large groups of incompletely studied patients, I favor preliminary trials in selected centers using the most modern MR technology⁶ in small, well-studied groups of patients to attempt to define those patients likely and those not likely to respond and then designing trials with these preliminary data in mind. Diffusion and perfusion MR scans with MR angiography can define underlying vascular lesions and show regions that already are infarcted as well as those that are underperfused and ischemic and at risk for further infarction.

We have learned something about stroke treatment but have only scratched the surface. Further trials should be designed that compare various treatment strategies for specific conditions (such as carotid artery atherostenosis or atrial fibrillation). Trials that study large numbers of inadequately evaluated patients grouped only on the presence of ischemia may be a waste of effort, time, and money. Physicians must choose treatment for individual patients based on age, comorbidities, stroke etiology, social, psychological, and economic factors, desires of patients, treatment facilities available, and the risks of various treatments in the hands of the individuals who will administer the treatment. Trials should be designed so that physicians will be able to use the data to guide treatment of their individual patients and should include patients with various subtypes of stroke who have been evaluated fully for the cause of brain ischemia. Treatment trials should be designed in relation to the cause of the stroke or cerebrovascular disease and the expected effect of the treatments studied.