The neurologist Camilo Gomez, MD, coined the phrase “time is brain” in 1993. (J Stroke Cerebrovasc Dis. 1993;3:1.) This was borrowed from the phrase “time is muscle” that had been promoted by the American Heart Association to support the use of IV tPA for acute myocardial infarction in a three-hour window and beyond.
Oddly, NINDS I and II were published together in 1995. (New Engl J Med. 1995;333:1581; http://bit.ly/2VYV1o9.) NINDS I (n=291) “proved” that time was not brain. No statistical difference was seen between IV tPA (0-3 hours) and placebo for stroke when measured in the first 24 hours of hospitalization. NINDS II (n=333) “proved” there was a 12 percent difference between IV tPA (0-3 hours) and placebo for stroke measured by the modified Rankin scale (mRS) three months later. Even in NINDS II, time was not brain beyond the first 90 minutes of stroke.
We have known from the beginning that “time is brain” is a false mantra. “Time is brain” has been used as a juggernaut marketing slogan by the American Heart Association, Genentech (Roche), and hospitals advertising themselves as stroke centers. A paper presenting a mathematical calculation initially based on rodent brain tissue was used, suggesting that “1.9 million neurons are lost per minute” in ischemic stroke. (Stroke. 2006:37:263; http://bit.ly/2LKGQiM.) This hypothesis was trumpeted as a call to action: Every second counts.
Dr. Gomez, the originator of “time is brain,” wrote last year: “It is no longer reasonable to believe that the effect of time on the ischemic process represents an absolute paradigm.... It is increasingly evident that the volume of injured tissue within a given interval of time of onset shows considerable variability in large part due to the beneficial effect of robust collateral circulation.” (J Stroke Cerebrovascular Dis. 2018;27:2214.)
It is disingenuous for anyone to claim that he knew all along that the time-is-brain mantra, the brain-attack banner, and the every-second-counts battle cry were misguided (or just blatantly false). They said nothing and looked the other way as the emperor of Genentech went by without clothes because they knew the cause of IV tPA for stroke would be advanced. Patrons who applauded as the stroke alert procession hurried past are now shouting: Tissue-based! Precision medicine! Theranostics! As if this were a new regime.
The public unveiling of CT perfusion with automated digital software (RAPID) forced us to admit what we already knew. Acute ischemic stroke is complex. It is magical thinking to give a $38,000 dose of IV tPA for “stroke” based upon a dubious “last seen normal” report, a brief history, a NIHSS score with poor interrater reliability, and a negative noncontrast CT and proclaim honorable results (though it certainly increases IV tPA usage). It is not about seconds or even minutes. It is about the precise automated measurement of a small ischemic core, a large reversible penumbra, and the use of thrombectomy in highly selected patients with detailed consent.
IV tPA has little if any place in treating acute ischemic stroke. This statement should shock no one because it is the only conclusion we ever had. The best we can hope for in the three-hour window is a 12 percent benefit with a 6.5 percent ICH harm, the vast majority of which happens in the first 90 minutes (New Engl J Med. 1995;333:1581; http://bit.ly/2VYV1o9), which is improbable in most EDs, to say nothing of the myriad other problems with the NINDS data.
ECASS III showed a seven percent benefit and a seven percent ICH harm (using the NINDS definition) in the 3-4.5-hour window among highly selected patients (which many now ignore selecting). (New Engl J Med. 2008;359:1317; http://bit.ly/2JBeVPz.) It is a wash, at best. The article by Josh Farkas, MD, on the fragility index lends further support to an already sizable argument that NINDS II and certainly ECASS III were actually negative studies. (EmCrit. July 5, 2016; http://bit.ly/2YAiUA3.)
Tissue, Not Time
Then there is the zero-to-six-hour window of thrombectomy, in which it appears convincing that the addition of IV tPA adds nothing to thrombectomy results for a large vessel occlusion (LVO) stroke. (AJNR Am J Neuroradiol. 2017;38:1464, http://bit.ly/2Ho2025; Acta Radiol. 2019;60:308, http://bit.ly/2Hrc3Ud.) What if we admit that the time-is-brain train was never on track? What if we accept that good stroke management is based on tissue instead of time? How would it look if we used automated CT perfusion software for thrombolysis like we do thrombectomy? It is a fascinating hypothesis, especially for the non-LVO stroke with an NIHSS score >9 in the 1.5-4.5 hour window.
Unfortunately, we don't have those data. The just-published EXTEND trial used CT perfusion for IV tPA in the 4.5-9-hour window, but the majority were LVOs (>65%). (New Engl J Med. (2019;380:1795; http://bit.ly/2VkCPBl.) The research parameter for core size was <70 cc, but the actual trial median size of the core was <5 cc with a median size penumbra of approximately 75 cc. These conditions will not come around with any frequency. And no one should treat LVOs with thrombolysis anymore, making the EXTEND study obsolete the day it was published. Furthermore, the study was equivocal with an absolute benefit of six percent and an absolute ICH harm of six percent. All benefit goes away when the definition of good outcome is redefined as mRS 0-2 instead of 0-1, the same way it did in ECASS III. So IV tPA doesn't work with CTP either, at least in the 4.5-9-hour window.
The editorial accompanying the EXTEND study admits that the time-is-brain mantra is wrong, but it still insists that the study “shatters” the time windows and supports using IV tPA beyond 4.5 hours everywhere “without the need for interventional radiologists.” (New Engl J Med. 2019;380:1865.) He could not be more wrong. If we had the ability to see the data instead of what we want to see, we should be looking at the radical minimization of IV tPA for stroke. We should be looking at doing primarily software perfusion-guided thrombectomy without the need for thrombolysis. (Is this the facilitated PCI argument all over again?)
The cost of the software, the specialized training of neuroradiologists and neurointerventionalists, the availability of these people and resources 24/7, and the tiny fraction of patients who will qualify to receive this therapy will make maintaining a stroke program a Machiavellian task that only a few select centers can pull off. We used IV tPA prematurely for stroke because we were a desperate and blinded people. We have allowed the naked emperor to parade around for 25 years. It may take us just as long to have a reliable and affordable automated technology in addition to finding an available army of the right personnel to find the right patient and perform the right therapy for stroke.
Dr. Mosleyis the medical director for student and resident education at Wesley Medical Center in Wichita, KS.