Neuroprotection Is Unlikely to Be Effective in Humans Using Current Trial Designs
What we need to carry out in the laboratory and at the bedside are experiments. If our experiments have been positive in the laboratory but negative at the bedside, it is logical to me that at the bedside we need to better emulate the conditions under which the laboratory experiment turns out positive; in other words, we need to do the “rat experiment” in man. Those clinical studies that have adhered to this dictum1–3 have been the only positive clinical trials to date.
Figure 1 depicts the general design of the rat transient middle cerebral artery occlusion model that is most often used to test neuroprotective drugs. It also depicts the general design of the sort of clinical trial that I postulate must be done to get positive results in stroke patients. There are 4 main areas where clinical trials have departed most from this model. In rats, we take pains to produce lesions of standardized severity in order to better detect a treatment effect; we start by giving the drug soon after the onset of stroke and then determine how much this time to treatment (TTT) can be lengthened; we use models of temporary rather than permanent arterial occlusion; and we increase doses of drug until we see a therapeutic effect.
These factors must all be addressed in the design of future clinical trials, and they are listed in Figure 2 along with one other important point. We need to find treatments that are substantially more potent than those that have failed in clinical trials to date.
Standardize Stroke Severity
There are two reasons why standardizing stroke severity is important in our experiments. Both relate to optimizing the ability to see a treatment effect between the drug and placebo. First, if stroke severity is too great, then animals die whether or not they receive treatment. If severity is too mild, the lesion is so small that any differences cannot be detected. Second, if, by chance, the distribution of initial stroke severity varies between the treatment groups, the effect of this imbalance could be much greater than any effect of the treatment.
Attempts to standardize stroke severity in patients randomized into clinical trials is important for the same reasons. The severity of stroke is reflected in the National Institute of Health Stroke Scale (NIHSS). In clinical studies, the baseline NIHSS is clearly the most important variable predicting outcome.4 Using low and high NIHSS cutoffs and ensuring that treatment groups are matched in distribution of NIHSS scores may be one way to achieve standardized stroke severity in our trials.
Stroke standardization might be better accomplished by assessing tissue viability using MRI. However, MRI criteria for predicting tissue outcome is still uncertain, and adhering to a very narrow time window often does not allow for ancillary tests to determine tissue viability. Keeping the time to treatment brief may itself help standardize stroke severity since all patients would have relatively brief, and therefore more reversible, ischemia.
Other variables affecting stroke severity in rats that have not been controlled in most human studies are the number of vessels occluded and location in cortex or white matter. Only 1 study3 was designed to limit patients enrolled to only one type of vascular lesion. The results were positive, again reflecting the wisdom of designing clinical studies to closely emulate what we do in animals. Noninvasive techniques such as transcranial ultrasound and MRI could help us standardize these variables.
Shorten Time to Treatment
In the laboratory, the investigator always starts with a brief TTT and then gradually prolongs it until an effect is no longer seen. This is just the opposite of what has been done in all clinical studies of neuroprotection.
Important lessons can be learned by comparing laboratory with clinical results using thrombolysis to achieve tissue reperfusion. In the laboratory, reperfusion must occur within 2 to 4 hours to see a reduction of infarct size.5 Clinical studies using intravenous rt-PA begun within 3 hours showed a positive effect,1 with more benefit associated with earlier treatment within that window. If begun after 3 hours, rt-PA had little or no benefit.6 Considering that it takes 30 to 90 minutes for a clot to dissolve after beginning intravenous rt-PA, the TTT for reperfusion in clinical studies correlates very nicely with what was found in laboratory models. Now we can carry out further studies in selected patients to see if we can find benefit with longer TTT.
Preclinical studies have shown that all neuroprotective drugs are less effective the later they are given, and most are ineffective if started more than 2 to 4 hours after the onset of ischemia. Yet no clinical trial has yet included enough patients within that 4-hour time window to reach any conclusions about efficacy. Pharmaceutical companies and their consultants, naturally interested in establishing the largest market for their drug, have abandoned the laboratory data and extended the TTT in most clinical studies to 6 hours or more.
Combine Neuroprotection with Reperfusion
Another lesson from laboratory stroke models is that neuroprotective therapy is generally more effective if given to animals with reversible rather than permanent arterial occlusion. One obvious reason is that for a neuroprotective drug to work, it must reach the injured tissue. Also, neuroprotective drugs are particularly effective by targeting cellular events triggered when injured brain tissue is reperfused.
Is it possible to design a study combining reperfusion and neuroprotection? In a recent safety study of lubeluzole combined with rt-PA,7 all patients received intravenous rt-PA within 3 hours of stroke onset, and were begun on study drug or placebo within 1 hour of starting rt-PA.
Give Sufficient Dose
Side effects have been the Achilles heel limiting the doses of neuroprotective drugs given to stroke patients. Recent stroke victims seem particularly vulnerable to cardiovascular or sedative effects of a drug, often not predicted by studies in young rats or in normal volunteers. These considerations make it difficult to predict the blood levels we can expect to reach safely in our acute stroke patients.
In designing a clinical neuroprotective trial, we should begin with a dose escalation study (with blood level correlation) in stroke patients similar to those intended for the pivotal trial. The highest tolerable dose should then be chosen for the pivotal trial. Ideally, the next step would be to demonstrate that the dose chosen has the desired effect on the target biological process.
Find a More Effective Drug
Most of the neuroprotective drugs subjected to clinical evaluation have been able to improve outcome by about 50% in well-standardized preclinical stroke models. It is possible that even if we adhere to the principles described in the preceding paragraphs, this effect is not enough to detect in our stroke patients.
Based on our inability to translate preclinical results to the bedside, we probably should no longer move forward with clinical evaluation of a monotherapy that reduces damage by only 50% in a rat. We need to find drugs that reduce damage by 80% in these models. While testing a drug that affects a single pathway in the process of brain injury has scientific and regulatory purity, this strategy has proven ineffective clinically. It is time to look for combinations of drugs that have a stronger effect. We need to work on the regulatory and financial impediments for conducting such trials.
The opinions expressed in this editorial are not necessarily those of the editors or of the American Stroke Association.
Furlan A, Higashida R, Wechsler L, Gent M, Rowley H, Kase C, Pessin M, Ahuja A, Callahan F, Clark WM, Silver F, Rivera F, for the PROACT Investigators. Intra-arterial prourokinase for acute ischemic stroke: the PROACT II study: a randomized controlled trial. JAMA. 1999; 282: 2003–2011.
The NINDS t-PA Stroke Study Group. Generalized efficacy of t-PA for acute stroke: subgroup analysis of the NINDS t-PA stroke trial. Stroke. 1997; 28: 2119–2125.
Kaplan B, Brint S, Tanabe J, Jacewicz M, Wang XJ, Pulsinelli W. Temporal thresholds for neocortical infarction in rats subjected to reversible focal cerebral ischemia. Stroke. 1991; 22: 1032–1039.
Hacke W, Kaste M, Fieschi C, Toni D, Lesaffre E, von Kummer R, Boysen G, Bluhmki E, Hoxter G, Mahagne MH, et al. Intravenous thrombolysis with recombinant tissue plasminogen activator for acute hemispheric stroke: the European Cooperative Acute Stroke Study (ECASS). JAMA. 1995; 274: 1017–1025.