tPA for Acute Ischemic Stroke and Its Controversies: A Review

Abstract

Background and Purpose:

Although neurologists consider intravenous tissue plasminogen activator (tPA) to be standard of care in the treatment of patients with acute ischemic stroke, its use remains contentious within the broader medical community, and particularly among emergency physicians. Why might this be? We provide a historical context to this ongoing controversy by reviewing how neurologists have conceptualized the acute stroke and its treatment, with the aim of bridging this gap.

Methods:

Based on historical sources in the Mackie Family History of Neuroscience Collection at the University of Calgary, as well as online resources, we trace the evolution of the concept of the “acute stroke,” which has come to mean a stroke that is potentially treatable with tPA. We frame this conceptualization in relation to historical “building blocks” in anatomy, pathology, and physiology. We then use these building blocks to explain why neurologists understand tPA to be effective and why emergency physicians often do not.

Results and Conclusions:

Arguments against the use of tPA reiterate 20-year-old concerns about its efficacy and safety. We believe these persistent concerns can be framed as a lack of understanding of the “building blocks” upon which neurologists’ conception of tPA is built. Our view suggests that the way forward to bridge the gap between neurology and other disciplines is not to conduct more trials but to offer a shared conceptualization of the trials already completed and of the intellectual tradition from which they emerged.

Introduction

On March 26, 2018, the New York Times published an in-depth investigative report on whether intravenous (IV) tissue plasminogen activator (tPA) is an effective treatment for stroke,¹ a controversy echoed in a recent debate on MedScape.² In the years since the National Institutes of Neurological Disease and Stroke (NINDS) trial reported the successful use of tPA for stroke,³ concerns and controversy have persisted.^4-7 Many of these concerns have been voiced by emergency physicians, while most of the enthusiasm for tPA has come from neurologists. For example, as many as 40% of emergency physicians surveyed in 2005 admitted that they would be unlikely to use tPA even “under ideal conditions.”⁸

In this narrative review, we provide a historical context to the ongoing controversy around tPA. By studying how neurologists have conceptualized the acute stroke and its treatment, we seek to bridge this gap in understanding and practice. Based mostly on historical sources in the Mackie Family History of Neuroscience Collection at the University of Calgary, we trace the evolution of the concept of the “acute stroke,” which has come to mean a stroke that is potentially treatable with tPA. We frame this conceptualization in relation to historical “building blocks” in anatomy, pathology, and physiology. We then use these building blocks to explain why neurologists understand tPA to be effective and why emergency physicians often do not. We believe that the persistent concerns surrounding tPA can be framed as a lack of understanding of these historical “building blocks.” Our view suggests that the way forward is not to conduct more trials but to offer a shared conceptualization of the trials also completed and of the intellectual tradition from which they emerged.

This historical analysis of the “acute stroke” exemplifies the importance of ideas to the way diseases are treated. In other words, the way doctors think about diseases impacts the way they treat those diseases. On some level, this may seem perfectly obvious: if fibromyalgia is conceived as a psychological problem, it is likely to be treated with psychotherapy, while if it is conceived as a problem of hyperexcitable neurons, it is likely to be treated with anticonvulsants. Similarly, if “acute stroke” is understood as a disease that is irreversibly debilitating and unresponsive to any treatment, then why would doctors put patients at risk trying? In contrast, if the “acute stroke” were thought to be treatable, then we would expect doctors to offer treatment. Differences in conceptualizations impact decision-making.

This fact underscores the reality that diseases should not be understood as things but as ideas: ways of understanding the suffering of others. And, importantly, ideas of disease change over time, for example, in response to changes in technological abilities, treatment availabilities, and value systems. Epilepsy is a classic example, which went from being a sign of the divine, to a mark of demonic possession, to a problem of neuronal excitability.⁹ In fact, we sometimes carve off new diseases from old ones because of changes in how we understand mechanisms, prognosticate, and treat.¹⁰

The Untreatability of Stroke

The term stroke was first used in the medical context in 1599, though it appeared rarely until the mid-20th century.¹¹ Through the early modern period, the term apoplexy was used to describe events of sudden onset, especially with neurological symptoms. While they should not be taken as synonymous, apoplexy and stroke were consistently understood as conditions that could not be treated. For example, in the 17th century, Thomas Willis wrote that apoplexy was associated with a prognosis that was “fatal and dubious.”¹² In 1820, British physician John Cooke declared “that it is impossible to cure the strong apoplexy” and urged doctors to avoid “unsuccessful attempts” for they “might bring disgrace on themselves and on the profession.”¹³ These attitudes persisted into the mid-20th century.

Against this tradition of nihilism, American neurosurgeons Bernard Sussman and Thomas Fitch were the first to elaborate and test a concept of treatability for stroke in their 1958 article “Thrombolysis with Fibrinolysin in Cerebral Arterial Occlusion.” They described how, in one patient, an occlusion of the middle cerebral artery seen on angiography disappeared after the IV administration of fibrinolysin. They considered this finding to be suggestive of fibrinolysin’s “specific usefulness in treatment” for stroke.¹⁴ Sussman and Fitch postulated that the success of any treatment for stroke would depend upon a “critical period” just after the onset of symptoms, during which time the stroke may be reversible.¹⁴ Sussman and Fitch’s work introduced the concept that would come to define the acute stroke: treatability in relation to time.

The question of which treatment, and administered over what time frame, remained to be explicated. In the NINDS trial, the concepts of treatability and time were strongly related, for it was the first trial to suggest that “acute stroke” was indeed treatable, but only within 3 hours of the onset of symptoms.³ Several other contemporaneous trials, which had attempted treatment up to 4.5, 6, and 8 hours after the onset of symptoms, failed to show a beneficial effect to patients.^15-18

In December 1995, the NINDS trial was published in the New England Journal of Medicine under the title “Tissue Plasminogen Activator for Acute Ischemic Stroke.” The article lays out the logic of its approach:

Cerebral angiography conducted soon after the onset of stroke demonstrates arterial occlusions in 80 percent of acute infarctions. Thrombolytic canalization of occluded arteries may reduce the degree of injury to the brain if it is done before the process of infarction has been completed…The safety of intravenous t-PA for the treatment of acute cerebral ischemia was previously tested in two open-label, dose-escalation studies, which emphasized very early treatment—within 90 and 180 minutes of the onset of the stroke—to reduce the risk of hemorrhage and to maximize the potential for recovery.^3(p1581)

In these few sentences, the authors enumerate key concepts in the idea of the acute stroke and frame their understanding at that point in time: that stroke is an ischemic vascular disease arising from arterial occlusion, that time is important in determining the success of potential reperfusion, and that tPA may have specific activity in this condition. As compared with patients given placebo, patients treated with IV tPA under 3 hours from symptom onset were at least 30% more likely to have minimal or no disability at 3 months.³ For the first time in the history of stroke, a treatment could be said to be effective in reversing its signs and symptoms, albeit if administered under particular conditions and to particular patients.

In light of the results of the NINDS trial, “acute stroke” was arguably carved off from stroke, becoming a disease of its own: it was managed by stroke teams rushing to the emergency department, treated with IV tPA, and had a different (better) prognosis. Starting in 1995, stroke went from an untreatable condition to one that was treatable with tPA. But that reconceptualization didn’t happen overnight. The idea of the “acute stroke” was made possible by a set of historical building blocks, which remain essential to understanding the ongoing debates surrounding the efficacy of tPA.

Building Blocks of the Acute Stroke

Vascular Anatomy

The first building block relevant to the acute stroke is recognition of the brain’s vascular anatomy, meaning the fact that and ways by which blood arrives in the brain. As late as 1616, the great anatomist William Harvey—who had proved that blood flowed from the heart to arteries to tissues to veins and back—had written that the brain “is cold, for it is bloodless.”¹⁹ In the 1680s, English physician Thomas Willis described the ways in which the 4 main arteries—2 carotids and 2 vertebrals—bring blood to the brain and share supply through a series of interconnections, which we now call the “circle of Willis.”¹² Willis himself suggested that these interconnections could protect against the clinical entity of apoplexy.

Localizability of Pathology

Diseases were previously thought to arise from a disruption of the body’s humors (blood, phlegm, yellow bile, and black bile), and treatments were targeted to specific imbalances in those humors. Starting in the 17th century, diseases began to be related to dysfunction of specific organs, rather than to the balance of humors in the body. Italian pathologist Giovanni Battista Morgagni (1682-1771) famously promoted this “anatomical concept” in his seminal work De Sedibus et causis morborum per anatomem indagatis of 1761. In De Sedibus, Morgagni paired case histories with autopsy reports, establishing a narrative and causal link between a patient’s symptoms in life and postmortem pathological findings.²⁰ Morgagni’s work began to establish that a patient’s symptoms could be related to processes within specific organs and that the location of these processes could be predicted by examining live patients, though would need to be confirmed at autopsy.

Ischemia and Hemorrhage

Building on the work of Willis and Morgagni, the clinicoanatomical concept allowed physicians to localize patients’ symptoms to regions of the brain, and even to particular arterial territories. While symptoms could be localized, it was still difficult (if not impossible) to differentiate between pathologies in live patients. Swiss physician Johannes Jakob Wepfer famously argued that the clinical entity of apoplexy—a sudden onset of neurological symptoms—could relate to any number of pathologies found at autopsy, including clots obstructing the carotid or vertebral arteries and intracerebral bleeding.²¹ Wepfer’s work established that clinical presentations and pathological causes could be understood as distinct processes, and that, at least in the case of apoplexy, it might not be possible to differentiate among pathological processes on clinical grounds alone. The ability to differentiate hemorrhage from ischemia in routine clinical assessment remained impossible until the dissemination of the computed tomography (CT) scanner in the late 20th century.

Computed Tomography Scans

Efforts to use thrombolytics to reverse the symptoms of stroke in the 1950s and 1960s were hampered by clinicians’ inability to differentiate between hemorrhagic and ischemic processes on clinical grounds. For example, early experiments with thrombolytics in acute stroke—such as those reported by Sussman and Fitch—likely included patients with intracerebral hemorrhage (ICH). With the dissemination of the CT scanner in the 1970s and 1980s, physicians gained a new means of quickly distinguishing ischemia from hemorrhage, which allowed for the selection of patients with stroke for enrollment into the clinical trials of the late 1980s and 1990s.

Thrombolysis

The idea of chemical thrombolysis—of dissolving blood clots through the administration of a drug—arose in the 1940s. In the case of stroke, it was a matter of selecting which agent, for which population, and in which time frame. Multiple thrombolytics were tested during the late 20th century, though tPA emerged as the prime candidate. Tissue plasminogen activator was first identified in 1947 as a naturally occurring compound that could activate the body’s clot-busting cascade.²² Its ability to dissolve arterial occlusions in animals was first demonstrated in 1980²² and was tested in patients with myocardial infarction starting in 1984. The stroke section of the NINDS began funding research into agents that might reverse the signs and symptoms of stroke in the late 1980s.²³ The NINDS rtPA Study Group performed 2 dose-escalation trials of tPA, the first including patients up to 8 hours after symptom onset, the second within only 3 hours. On the basis of these experiments, the NINDS trial was launched with a dose of 0.9 mg/kg of tPA to be administered to patients under 3 hours of symptom onset.³

Penumbra

The final building block essential to the contemporary meaning of the acute stroke is the physiological concept of the penumbra. The penumbra refers to the phenomenon that, even after arterial occlusion, some brain tissue that stops functioning may be saved with the restoration of blood flow. The physiologic reason for the existence of the penumbra is understood to be the presence of collateralization within the brain, a system of interconnection and redundancy first described by Willis in the 1680s. At the same time, the penumbra was also felt to be a dynamic process that would not last forever. In 1961, Sussman and Fitch had specifically proposed that the viability of ischemic brain be considered a function of time. The theory of the penumbra explains why the efficacy of treatment with tPA should be time dependent and why the acute stroke must therefore be defined in relation to time from the onset of symptoms.

These building blocks together justified the concept of the “therapeutic window,” a period during which the brain could be salvaged in whole or in part. Understanding the brain’s vascular and functional anatomy allowed neurologists to localize symptoms to specific arterial territories, and the CT scan allowed for the exclusion of patients with ICH. The physiological penumbra created an opportunity for thrombolysis to be effective, once the agent tPA was tested and proven. All of these foundational factors—anatomical, physiological, radiological, pharmacological—are essential to understanding why neurologists conceive of stroke as a condition that is treatable with tPA.

Controversies Around tPA

How does this historical review relate to the contemporary controversies surrounding tPA? Let us begin with 2 general points. First, that diseases are best understood as ideas that change along with developments in diagnosis and treatment. Indeed, a moral imperative exists to reframe our concepts of diseases in light of new knowledge. In fact, the definition of the acute stroke has further evolved over the last 5 years, with clinical evidence of the efficacy of mechanical thrombectomy for acute stroke suggesting that in some patients a stroke may be reversible up to 24 hours after the onset of symptoms.²⁴ Second, if diseases are understood as ideas, then we have to recognize that not all physicians working in the same context will always share the same conceptualization of a given disease. And this can lead to conflicts over treatments. For example, we would expect that physicians who do not conceive of acute stroke as a treatable condition may be less likely to use tPA as a treatment. This is indeed the situation that arose around the stroke field in the years after 1995. Why?

There have been 3 charges laid against tPA in the medical literature: that its endorsement in guidelines was untrustworthy because of collusion between the pharmaceutical industry and the American Heart Association (AHA); that the seminal trials failed to show tPA’s efficacy; and that the risks associated with tPA use are too high. Essential to understanding each of these claims are the historical building blocks outlined above.

The charge of collusion—advanced most famously by journalist Jeanne Lenzer in a 2002 article in the BMJ—relates to the fact that in its 2000 guidelines, the AHA upgraded tPA from a class IIb recommendation to a class I recommendation.²⁵ According to Ms Lenzer, this upgrade occurred because the distributor of tPA—a pharmaceutical firm known as Genentech—contributed $11 million to the AHA, thereby corrupting guideline writers. While we cannot deny the possibility of conflicts of interest, other guideline-writing organizations that did not receive funds from Genentech have similarly strongly endorsed tPA, speaking to the strength of the evidence.^26,27

We suspect that the concern about collusion overlaps with the second criticism of tPA, namely that the tPA trials failed to prove its efficacy. In our view, this is where the historical building blocks of the acute stroke come into play. Critics of tPA lump together the results of the NINDS trial with the results of other earlier and contemporaneous trials (ECASS I, ECASS II, ATLANTIS A, and ATLANTIS B) that had weaker or negative results, despite significant differences in protocol.

This practice is reflected on the website thennt.com, which pooled data from trials with different methodologies, time frames, and drugs to conclude that thrombolysis was ineffective and even dangerous.²⁸ At first glance, this approach may have appeal, but a significant difference exists between the NINDS trial and all other stroke trials of that era: Only the NINDS trial enrolled all of its patients under 3 hours from the onset of symptoms. The selection of the 3-hour window was based upon the concept of the penumbra, which is related to the principles of vascular anatomy and collateralization.

Interestingly, the concept of the penumbra is cited commonly in articles arguing for the treatability of stroke and is not mentioned in any articles we have found that challenged it. Without the concept of the penumbra, and its operational correlate of the 3-hour window, there would be no basis to differentiate between the NINDS trial and the other thrombolytic trials; it would be reasonable to conclude that tPA is not effective and that acute stroke is not treatable. The various trial designs reflect different ways of trying to translate the concept of the penumbra into clinical practice, and it is to be expected that these “competing attempts to identify the details of implementation” might lead to different results.⁴ They do not invalidate the results of any one trial and once again support the importance of different conceptualizations impacting treatment decisions and outcomes.

The third criticism focuses on the risk of ICH after tPA. In the NINDS trial, approximately 6% of patients who received tPA experienced an ICH, in contrast to 0.6% in those who did not.³ Critics argue that a 6% risk is “unacceptable” and renders moot any benefit of tPA. For a group of Emergency Room (ER) physicians who were surveyed in 2005, a 3.4% risk of ICH was deemed acceptable, while a 6% risk was not.⁸ Concerns about post-tPA ICH mostly originate in the emergency medicine community,^29-33 and we contend that this discomfort with treatment actually reflects a discomfort with diagnosis. In a 2014 listserv discussion about tPA, one New York-based neurologist suggested that “the dirty little secret” of the debate around tPA administration is that emergency physicians “feel inadequate to diagnose stroke” (AAN Listserv, February 9, 2014). In a recent debate about stroke treatment, emergency physician Dr Anand Swaminathan of New York University similarly expressed concerns about the fact that “we don’t know who to give the drug to.”³⁴ Emergency physicians appear to feel uncomfortable diagnosing stroke and so fret about causing unnecessary hemorrhage in patients who should never have received tPA in the first place. We would frame this concern in relation to the ability (or lack thereof) to positively diagnose stroke, an act that requires familiarity with the brain’s vascular and functional anatomies, derived from the acute stroke’s building blocks. This association between treatment likelihood and diagnostic certainty is reflected in the writings of tPA sceptic Dr Ryan Radecki, an emergency physician at the University of Texas. Radecki identifies “diagnostic certainty” as a central element that would render tPA use “more appropriate.”³⁵

So how should this divide between emergency physicians and neurologists be addressed? From our perspective, the question of tPA efficacy has been adequately answered with available evidence; more clinical trials are not needed.³⁶ What is needed now are efforts to build bridges between neurologists and colleagues in other medical communities. Neurologists should seek to share not only what they do but why they do it. The why—as we have argued—is a reflection of those building blocks that serve as the foundation of modern conceptualizations of both the acute stroke and its treatment.