Reports

This report is based on medical evidence presented at sanctioned medical congress, from peer reviewed literature or opinion provided by a qualified healthcare practitioner. The consumption of the information contained within this report is intended for qualified Canadian healthcare practitioners only.

31st International Stroke Conference

Kissimmee, Florida / February 16-18, 2006

Acute stroke is a common but devastating clinical event which is a leading cause of death and disability in Canada and other industrialized countries. Neuroprotective agents have been vigorously pursued on the basis that much of the neuronal death contributing to disability occurs after the immediate infarction. With rapid administration of drugs that protect neurons from destructive elements, particularly oxygen free radicals, experimental studies have demonstrated that the area of the infarct can be reduced and brain function preserved.

Controlling Rather than Reversing Disability

One explanation for the failure of many of these agents in clinical trials is the use of inappropriate end points. “What is the goal in the treatment of stroke? We should not be thinking just about cures. A clinically meaningful reduction in disability is an important change in outcome and may be the most we can ask from a neuroprotective drug that cannot reverse damage that has already occurred,” stated Dr. Marc Fisher, Department of Neurology, University of Massachusetts School of Medicine, Worcester.

Several neuroprotectants that were shown to be highly effective in experimental models of stroke subsequently failed in clinical trials. Many of these agents were targeted at oxidative stress, an established mediator of neuronal death in the immediate period following an ischemic stroke. At least in animal models, oxidative stress enlarges the area of the infarct to increase the likelihood of death or major brain injury. Due to the compelling experimental evidence that neuroprotection is a viable approach to limiting brain damage after stroke, a variety of theories have been proposed to explain why past studies with promising agents were suboptimal. The potential explanations include the possibility that therapies were delivered too late or that therapies were not provided at sufficient doses; that too many patients with large infarcts unresponsive to neuroprotective agents were included in study populations, so that when outcomes were being measured, investigators were unable to demonstrate the benefits of neuroprotective agents.

“The scoreboard tells a grim story. There have been 49 neuroprotective agents studied in 114 controlled trials that randomized 21,445 patients. There have been zero positive results,” observed Dr. Jeffrey Saver, University of California, Los Angeles. Although many of these agents may have been ineffective, he suggested that failure to employ end points with the capability of demonstrating benefit may have defeated some proportion of these treatments strategies. According to Dr. Saver, “Most studies have taken a very dichotomous approach to evaluating neuroprotective agents. That is, patients were labelled as having benefited or not on a single outcome.” The advantage of this approach is that it is simple, but the disadvantage is that it can underestimate benefits in an individual patient. In the new approach, a shift analysis is used to measure modest but clinically significant improvements along a continuum. This is helpful for increasing the signal for benefit on an individual basis.

This method was employed and validated in a trial published just one week before the 2006 International Stroke Conference (Lees et al. N Engl J Med 2006; 354:588-600). SAINT I (Stroke-Acute Ischemic NXY Treatment) is the first major trial to associate a neuroprotective agent with significant benefit for the primary study end point. In the study, 1699 patients with an acute stroke were included in the efficacy analysis. The patients were randomized to NXY-059, an investigational neuroprotective agent that has been shown to trap free radicals to reduce infarct size in experimental studies, or to placebo. Treatment was administered within 6 hours of the onset of an acute ischemic stroke. The primary end point was disability at 90 days as measured according to scores on the modified Rankin scale for disability. In the trial, there were 118 participating centres in 24 countries.

“On the primary end point, [treatment] improved the odds ratio of a good outcome by 20% [95% CI 1.01- 1.42; P=0.038),” reported Dr. Kennedy R. Lees, Western Infirmary, Glasgow, UK. As senior author of this study, Dr. Lees reported that the improvement was observed across all categories of the modified Rankin scale on an intent-to-treat analysis. For example, while only 11% of the placebo patients had a Rankin score of 0, signifying no disability, this outcome climbed to 15.4% of the patients receiving NXY-059, an absolute increase of 4.4%. Conversely, the proportion of patients meeting criteria for a Rankin score of 4, which signifies inability to walk or attend to bodily needs without assistance, and Rankin score of 5, which signifies severe bedridden disability and a need for constant nursing care was only 41% on active treatment vs. 44.6% on placebo. The reduction in severe stroke on active teatment was explained by an overall shift toward less disability.

Table 1. Modified Rankin Scale Reduced Disability at 90 Days

Evaluating Neuroprotection

The per-protocol analysis produced very similar results. While the odds ratio for an improvement in modified Rankin score remained at 1.20, the proportion of patients with no disability was almost 5% greater among those receiving the neuroprotectant than placebo, and the overall shift toward better outcome in the active treatment group had a greater statistical significance (P=0.02). When groups were compared at earlier time points, its relative advantage was even greater, a result consistent with the activity of neuroprotection.

“At 90 days, you are already beginning to see regression to the mean, because these patients often have other medical problems that dilute the difference in death and disability with longer follow-up,” Dr. Lees explained. “The best time to evaluate neuroprotection is probably between seven and 30 days.”

The same phenomenon of early benefit was observed with the next end points evaluated in a hierarchal order after the primary outcome. For example, the first of these, the National Institutes of Health Stroke Scale (NIHSS) did not show a significant advantage for NXY-059 relative to placebo at 90 days, but the odds ratio of an improved outcome on NIHSS rose to 1.13 if evaluated at 30 days (P=0.09) and 1.46 (P=0.04) at seven days. The Barthel index did not associate the agent with a statistical advantage at 90 days either, but it did at 30 days (P=0.02) and seven days (P=0.0001).

The total number of adverse events was slightly lower on active therapy than it was in the placebo group. In a post hoc analysis of patients who received tissue plasminogen activator (tPA) therapy, there was a highly significant reduction in the risk of hemorrhagic transformation among patients receiving the neuroprotectant (P=0.001). Specifically, the rate of asymptomatic intracerebral hemorrhage after thrombolysis was 20.9% in the placebo group vs. 12.9% in the active treatment group. The rates of symptomatic intracerebral haemorrhage on placebo and active treatment were 6.4% and 2.5%, respectively.

Table
age after Thrombolysis

<img72|center>

“The point of the post hoc analysis was to evaluate whether tPA modified the benefit of NXY-059 by providing enough improvement in outcome that the neuroprotective agent could not show additional improvement. The protection against hemorrhagic transformation was unexpected, but the interaction was highly significant and does seem real. It may have something to do with the ability of NXY-059 to improve vascular endothelial function, but more studies are needed,” Dr. Lees told delegates.

When patients were stratified according to an extensive list of characteristics, including age above or below 71, gender, race, baseline plasma glucose, body temperature, and time from onset of stroke until treatment, the neuroprotectant was consistently more efficacous than placebo. The one exception was diastolic blood pressure in which there was a negative interaction. However, Dr. Lees noted that the agent has no known effect on blood pressure and suggested that this was likely due to a statistical anomaly.

Modest Benefits but Clinically Significant

The neuroprotective effect of NXY-059 is modest, but it was considered by Dr. Lees and several experts invited to discuss the trial results as clinically significant. According to Dr. Saver, who observed that “a modest shift towards less disability with a neuroprotective agent accords with our knowledge of stroke,” the benefit can be translated into practical terms. Based on the premise that a 1 category improvement in Rankin scale is clinically meaningful for patients, he calculated that the number needed to treat (NNT) for this degree of change is 9.8. In other words, approximately 10 individuals can expect to improve a clinically significant way for every 100 patients treated. Although he cautioned that more data are needed to confirm this finding, he characterized this NNT as supportive of the use of the neuroprotectant.

More data will be forthcoming with a second trial, called SAINT II. This study is randomizing 3200 patients and is similarly designed to SAINT I with the same primary end point. One of the major differences with SAINT I is that it includes a protocol to seek out for an interaction between NXY-059 and tPA. With or without tPA, the neuroprotectant is expected to be a major addition to current options for controlling the damage of acute stroke.

“When administered to appropriate candidates, tPA is effective, but only a small proportion of patients receive a thrombolytic. As a result, therapies with more modest effects could still represent an important step forward if they can be offered to more patients,” Dr. Lees observed.

Novel Trial Design Rivals Results for Importance

The innovative approach to capturing change in stroke outcome in SAINT I may be nearly as important as the results. The modified Rankin scale has been widely used to measure global disability, and it has been previously employed as a primary end point in at least 11 previously randomized studies of stroke treatments. However, the shift in the modified Rankin scale across the study population, rather than any specific threshold improvement in score, is expected to become a new standard for future neuroprotection trials.

“I believe the shift in the modified Rankin scale will replace other outcome measures because it looks at a result useful for determining if we are helping our patients,” Dr. Fisher observed. He noted that a 1 category change in Rankin score may not seem impressive in the context of efforts to prevent any disability, but this degree of shift may mean the difference between a long-term care facility and independent living. For example, a change from the moderate disability of a modified Rankin score of 3 to the slight disability of a modified Rankin score of 2 may not only be of critical importance to the patient but to patient’s spouse and family as well.

“One of the issues raised by this new approach is whether previous neuroprotective agents that were considered ineffective on the basis of less sensitive measures of benefit might also have done as well with this approach,” Dr. Fisher noted. He concurred with Dr. Saver that one of the major causes of the series of negative trials with neuroprotective agents may have been failure to measure benefits appropriately.

New Era in Clinical Studies

With completion of the first trial to demonstrate benefit from a neuroprotective agent on the primary end point, Dr. Fisher suggested that “this is the end of the beginning rather than the beginning of the end” in the effort to develop new treatments for acute stroke. He indicated that since valid tools have been developed for measuring outcome, it will be easier to consider studies in which modest but clinically significant results are anticipated. For example, he suggested that neuroprotective agents predicted to have additive or synergistic effects might now be evaluated in combination.

“Once we have confirmation that we have a drug that is safe and effective for the treatment of acute stroke, it will be unethical to conduct placebo-controlled trials. We are entering a new era,” Dr. Fisher told delegates.

Summary

After a long series of disappointing trial findings with highly promising neuroprotective agents, an encouraging outcome with an agent that traps free radicals has shown evidence that neuroprotection is a viable concept. It is considered significant that the study showing benefit, the SAINT I trial, employed a shift in modified Rankin score as its primary outcome. As opposed to evaluating the proportion of patients who improved according to a single threshold of benefit, this study design allowed improvement along the continuum of the Rankin scale to be captured. This approach is considered a more appropriate measure of treatment effect because neuroprotective agents limit the area of infarction but are not expected to reverse damage within the area that has already been infarcted. This suggests that benefit from these agents is most likely to be identified by modest improvements. A new era of neuroprotection studies is anticipated on the basis of both the promising results and the innovative study design.

Note: At the time of printing, NXY-059 is not available in Canada.