A number of somewhat conflicting studies have been published over the past year regarding triage and transport protocols for stroke. In light of the most recent of these publications, NeuroNews speaks to Ashutosh Jadhav (Barrow Neurological Institute, Phoenix, USA) in an attempt to make sense of the existing data.
Last year, new data from the TRIAGE-STROKE and RACECAT studies created a level of uncertainty regarding the advantages of directly triaging acute stroke patients to thrombectomy-capable centres. The former suggested a trend towards functional outcome benefits in patients transferred directly to comprehensive stroke centres (CSCs), but was ultimately unable to reach statistical significance on this front, while the latter indicated that bypassing a primary stroke centre (PSC) may even be harmful in patients with a final diagnosis of intracranial haemorrhage (ICH).
Speaking to NeuroNews following publication of these datasets, leading authors for both—Anne Behrndtz (Aarhus University Hospital, Aarhus, Denmark) and Anna Ramos-Pachón (Hospital Germans Trias i Pujol, Barcelona, Spain)—were in agreement that further research is needed to improve stroke diagnosis ‘in the field’, and to ascertain which patients are most likely to benefit from bypassing protocols.
These data were followed towards the end of last year by a statement from the Society of NeuroInterventional Surgery (SNIS), in which several leading US neurointerventionists asserted that the entire body of evidence to date—including findings from the aforementioned trials—points towards the fact that direct triage to a CSC is beneficial in stroke patients with emergent large vessel occlusions (LVOs). The statement urged against misinterpreting or inappropriately applying the data, noting that TRIAGE-STROKE produced results “overwhelmingly in favour” of direct triage, despite its lack of power, and attributing RACECAT’s somewhat negative outcomes to geographical and healthcare system-related factors in Catalonia, Spain—where the study was conducted— that are “vastly different” to those seen across much of the USA.
Recent updates
The beginning of 2024 has already seen two further publications that contribute additional data to this ever-growing body of evidence. In February, Amrou Sarraj (Case Western Reserve University, Cleveland, USA) delivered one-year outcomes from the SELECT2 trial at the International Stroke Conference (ISC; 7–9 February, Phoenix, USA), and his presentation was accompanied by a prespecified analysis of the study’s transfer protocols being published in JAMA Neurology.
Across a total of 352 enrolled patients, all of whom had severe ischaemic strokes—meaning an Alberta stroke programme early computed tomography score (ASPECTS) of 3–5, or a core volume of ≥50mL on imaging, or both—59.9% were transferred from a PSC to receive thrombectomy, while 40.1% presented directly to a CSC to undergo the procedure.
Sarraj and colleagues’ analysis showed that the positive treatment effect of thrombectomy versus medical management was observed across both groups. Treatment effect estimates were maintained and continued to favour thrombectomy in patients with low ASPECTS at referring hospitals, and in those who demonstrated a loss of ≥2 ASPECTS points during transfer.
The authors also note that, while improved functional outcomes were numerically better in patients presenting directly to a CSC for thrombectomy versus those triaged via a PSC, this did not constitute a significant effect modification. Another of their findings was that thrombectomy treatment effect estimates were lower in patients with transfer times of three or more hours, as compared to those with shorter transfer times.
Sarraj and colleagues conclude their paper by stating that “these findings emphasise the need for rapid identification of patients suitable for transfer and expedited transport”.
More recently, in March, the team behind RACECAT offered another contribution to the existing body of data in this space, publishing a substudy of the trial in Stroke: Vascular and Interventional Neurology (SVIN). This secondary, post-hoc analysis sought to evaluate factors that may influence functional outcomes among stroke patients initially assessed at a local stroke centre—either a telestroke centre or a PSC—compared to a thrombectomy-capable CSC.
Leading author Marta Olivé-Gadea (University Hospital Vall d’Hebron, Barcelona, Spain) and colleagues determined that, across a modified intention-to-treat population of 903 acute ischaemic stroke patients, 90-day functional outcomes were associated with door-to-needle times and local hospital level of care. More specifically, they observed a trend favouring direct transport to a CSC for patients whose assigned local stroke centre was a telestroke centre, rather than a PSC, and those whose door-to-needle time was greater than the global median of 31 minutes. The authors also found that these benefits of direct transfer to a thrombectomy-capable CSC were “more evident” in patients with a confirmed LVO.
“Direct transport to thrombectomy-capable centres may be preferable in areas primarily covered by telestroke or local stroke centres with poorer performance, especially in patients with LVO,” Olivé-Gadea and colleagues conclude. “These findings can contribute to refining prehospital triage strategies and optimising stroke systems of care.”
Deciphering the data
“It is confusing, because there are multiple studies that show slightly different things,” says Jadhav, reflecting on these varying recent publications. “For LVO stroke patients, the only thing that we’ve shown helps is to open up the blood vessel quickly. There’s no doubt that, at an individual level, the faster you take out the blockage, the better. But, apart from opening up the vessel quickly, nothing else has really been proven as something we can modify to improve outcomes.”
Here, Jadhav draws comparisons between these debates and similar discussions over percutaneous coronary intervention (PCI) triage for myocardial infarction (MI) in the cardiology space. However, while a rough “inflection point” of two hours has been settled on regarding direct PCI triage, there are additional challenges preventing the translation of similar findings into the neuro world.
Firstly, according to Jadhav, ‘how’ stroke triage is actually conducted is relevant, and door-in-door-out (DIDO) times constitute an “important metric and potential bottleneck”.
“Ideally, the patient receives lytics and then goes straight to a thrombectomy-capable centre,” he adds, “but, if DIDO times are long and the system is inefficient, that can be to the detriment of the patient, because you’re spending a lot of time giving them lytics and doing advanced imaging.”
“I think we’re trying to solve this problem the same way cardiology did, but cardiology had a different set of considerations and was able to get to a more definitive answer. The efficiency of stroke triage is so multivariable that it’s hard to translate results between regions.”
On this front, stroke triage is generally more complex and ultimately “much slower” compared to the equivalent process for MI, as it involves telemedicine consultations, a greater imaging burden, and more discussions around lytics and other medical options. In addition, the second key discrepancy Jadhav highlights between stroke and MI triage relates to a point right at the heart of debates surrounding stroke transport protocols: the uncertainty over whether a patient is suffering from an ischaemic or haemorrhagic stroke, which is estimated by emergency responders using stroke severity scales but cannot be diagnosed with certainty until they undergo in-hospital imaging.
According to Jadhav, this is one of the major factors behind RACECAT’s perceived failure to produce data supporting the direct triage of stroke patients to thrombectomy-capable hospitals. The study used Rapid arterial occlusion evaluation (RACE) scores to approximate an LVO diagnosis but, in doing so— because RACE scoring cannot discriminate between stroke patients with or without brain bleeds—saw many haemorrhagic strokes bypassing local hospitals and heading straight to CSCs.
“Most EMS [emergency medical service] triage protocols, if they assume it’s an ischaemic stroke caused by a blockage, tend to advise letting the patient’s blood pressure ride high, because we want that patient to perfuse their brain,” Jadhav notes. “But, then, you have patients with a high RACE score being driven around for an extra hour with uncontrolled blood pressure and, potentially, a brain haemorrhage. That’s why RACECAT saw that difference where harm was caused to the haemorrhage population.”
As such, the appropriate triage option for a given individual patient may be clear but, “at a population level”, the answer is far less simple. Jadhav believes that, “if you don’t have a diagnostic tool beyond just RACE or NIHSS [National Institutes of Health stroke scale], there are a lot of LVO mimics that you’re potentially going to be harming [via direct triage]”. According to Jadhav, the “ideal” comparative stroke triage study will initially exclude haemorrhagic patients and include, for example, a “very clean population” of hundreds of LVO patients who are then randomised to different strategies. However, such a study would be “hard to do” with current technologies and will likely require more sophisticated prehospital tools capable of discerning those with brain bleeds from more ‘standard’ ischaemic stroke patients.
Real-world factors
Touching on the potential future role of mobile stroke units (MSUs) in this space, Jadhav adds that—while “bringing the ED [emergency department] to the patient would be the ideal solution”—the reality is that many places cannot afford to deploy MSUs on a large enough scale for the approach to be used en masse. He feels this will be a “huge limiting factor”, stating that “while, conceptually, it may be what we want, practically, it’s unlikely to happen”.
Even if a cheaper, more practical alternative to MSUs—for example, thermal energy- or transcranial Doppler (TCD)-based methods for distinguishing between stroke types—can be introduced into ambulances, the prehospital phase remains “politically, very complicated”, particularly in the USA. Jadhav notes the “major challenge” posed by the fact the country’s Centers for Medicare & Medicaid Services (CMS) does not provide coverage for the prehospital setting as a place of care.
“So, who pays for it?” he asks. “That means the appetite to introduce expensive technologies in the prehospital setting is not going to be very high, without a financial pathway to do it. MSUs have struggled with this issue as well.”
Returning to the topic of RACECAT, Jadhav highlights another discrepancy brought about by regional variations.
“Before that study, I think a lot of people thought: ‘it’s definitely better to bypass the local ED’, because we know it’s going to take a while to get patients triaged,” he says. “One thing we learned from RACECAT is that the Catalonia region is phenomenally efficient at triaging and transporting patients. Their DIDO times were very good and, in a way, that actually hurt their study hypothesis because they did not have these long delays that we see in other regions. Their times are definitely faster than in the USA, for example.”
This was confirmed, in a sense, by the RACECAT group’s more recent March 2024 publication in the SVIN journal. Jadhav says this paper essentially showed that direct triage to a CSC is more beneficial when it comes to less efficient systems with longer DIDO times at local centres, because “you’re spending way too much time at the primary hospital”.
“Bypassing protocols might be favourable in one town but not another, and it really depends on some of these drivers of workflow efficiency. If you’re in an area like Catalonia where processes are super-efficient, bypass may not be necessary and could even be detrimental. The approach to triage is not going to be generally applicable to everybody—what works in Catalonia may not be what works across the rest of Europe, or even in other parts of Spain. We know there are differences between urban and rural environments. And, within countries and between countries, there are huge differences in how triage works.
“I think we’re trying to solve this problem the same way cardiology did, but cardiology had a different set of considerations and was able to get to a more definitive answer. The efficiency of stroke triage is so multivariable that it’s hard to translate results between regions.”
Another emerging stroke workflow paradigm is the direct-to-angio approach, bypassing the emergency room at the thrombectomy-capable centre and directly admitting a confirmed or suspected acute thrombectomy candidate to the angiography suite with the hope of achieving faster recanalisation. Jadhav looks forward to the results of the upcoming DIRECT trial, which will compare the efficacy of the direct-to-angio approach to traditional protocols triaging via the ED. Jadhav closes the conversation by emphasising the bearing of robust evidence on real-world stroke triage, and discussions around it.
“Distinct considerations arise when considering changes to the prehospital system, separate from those within the confines of a hospital,” he concludes. “While actions within a hospital may be more controlled, altering prehospital protocols requires a robust foundation of data, as there will be impact on additional stakeholders including paramedics and local hospitals. Many anticipated that RACECAT would unequivocally advocate for bypassing, based on their local experience, and the results were surprising to many.
“Getting the right patient to the right hospital at the right time remains a significant area of uncertainty in the rapidly evolving stroke landscape.”
