Management of acute ischemic strokeBMJ 2020; 368 doi: https://doi.org/10.1136/bmj.l6983 (Published 13 February 2020) Cite this as: BMJ 2020;368:l6983
- 1Department of Neurology, University of Maryland School of Medicine, Baltimore, MD, USA
- 2Department of Epidemiology and Public Health, University of Maryland School of Medicine, Baltimore, MD, USA
- Correspondence to: M S Phipps
Stroke is the leading cause of long term disability in developed countries and one of the top causes of mortality worldwide. The past decade has seen substantial advances in the diagnostic and treatment options available to minimize the impact of acute ischemic stroke. The key first step in stroke care is early identification of patients with stroke and triage to centers capable of delivering the appropriate treatment, as fast as possible. Here, we review the data supporting pre-hospital and emergency stroke care, including use of emergency medical services protocols for identification of patients with stroke, intravenous thrombolysis in acute ischemic stroke including updates to recommended patient eligibility criteria and treatment time windows, and advanced imaging techniques with automated interpretation to identify patients with large areas of brain at risk but without large completed infarcts who are likely to benefit from endovascular thrombectomy in extended time windows from symptom onset. We also review protocols for management of patient physiologic parameters to minimize infarct volumes and recent updates in secondary prevention recommendations including short term use of dual antiplatelet therapy to prevent recurrent stroke in the high risk period immediately after stroke. Finally, we discuss emerging therapies and questions for future research.
Worldwide, one in six people will have a stroke in their lifetime, more than 13.7 million have a stroke each year, and 5.8 million a year die as a consequence (http://world-stroke.org). Globally, more than 80 million people have survived a stroke. About 70% of incident strokes are ischemic (9.5 million), and the rest are intracerebral hemorrhage or subarachnoid hemorrhage—the proportion of ischemic strokes in the US is estimated to be higher, at about 85-87%.1 This review will focus on the treatment of ischemic stroke, specifically on treatment in the hyperacute and acute stages.
Acute ischemic stroke (AIS) is defined by the sudden loss of blood flow to an area of the brain with the resulting loss of neurologic function. It is caused by thrombosis or embolism that occludes a cerebral vessel supplying a specific area of the brain. During a vessel occlusion, there is a core area where damage to the brain is irreversible and an area of penumbra where the brain has lost function owing to decreased blood flow but is not irreversibly injured. Evidence based treatments such as intravenous thrombolysis and endovascular clot retrieval, which can remove the obstruction and restore blood flow to the affected areas of the brain, have been shown to improve outcomes in AIS when applied to appropriate patients, with substantial advances in these treatments occurring in the past few years.23456 Selecting the right patients involves critical clinical assessment and brain and vascular imaging, as well as systems that provide fast but safe care, because the speed at which AIS is treated is directly related to outcome.4789 Although other vascular causes of acute brain injury exist, such as intracerebral hemorrhage, and important management decisions must be made after the acute stroke phase, this review will focus only on the management of ischemic stroke in the hyperacute and acute phases of the disease.
Sources and selection criteria
Both authors independently searched PubMed and Embase for English language articles published between 1 January 2000 and 1 September 2019, using the keyword terms “prehospital scales ischemic stroke”, “prehospital diversion in AIS”, “imaging in AIS”, “thrombolysis or alteplase or tenecteplase in AIS”, “endovascular thrombectomy or mechanical thrombectomy in AIS”, “intubation versus conscious sedation or anesthesia in AIS”, “endovascular devices in AIS”, “blood pressure management in AIS”, “glucose management in AIS”, “oxygen therapy in AIS”, “patient position in AIS”, and “antiplatelet or anticoagulation or antithrombotic therapy for secondary stroke prevention”. We included articles of historical importance from the 1990s that include the pivotal trials for the use of intravenous alteplase. We also searched the reference lists of high quality articles and reviews, and we included selected randomized controlled trials (RCTs), observational studies, systematic reviews, and meta-analyses from these sources. We gave precedence to large key studies that have informed the guidelines, but we also reviewed a wider range of recent studies on topics for which conclusions and evidence are mixed, controversial, or both. We excluded case reports and small case series.
A stroke assessment system used by emergency medical services (EMS), initial management with a stroke protocol started in the field, and pre-notification of hospitals all have moderate evidence from non-randomized studies and are strongly recommended.10 Regional EMS systems should develop triage standards and protocols specific to stroke, using validated instruments, and an organization of hospitals with different levels of stroke care should be developed for rapid triage of the right patient to the right hospital for the right treatment, in the most efficient way.10
Levels of care
Hospitals have differing capabilities in terms of treatment of AIS, and an international consensus exists on levels of care 1 through 3.11 Level 1 stroke centers have the full spectrum of endovascular care, do a minimum number of mechanical thrombectomies, have dedicated neurointensive care and stroke units, and have full neurosurgical services. Level 2 requires at least 100 stroke patients a year, a stroke unit, and a minimum of 50 mechanical thrombectomies, but neurointensive and neurosurgical care are not required, whereas level 3 requires only a minimum of 50 patients a year and a stroke unit. Four designations of stroke centers exist in the US: comprehensive stroke center (CSC), thrombectomy ready stroke center (TSC), primary stroke center (PSC), and acute stroke ready hospital (ASRH).12 All have different capabilities regarding the care they can provide to patients with acute stroke.
As endovascular therapy for patients with large vessel occlusion (LVO) is primarily available at TSCs and CSCs, diversion of patients with suspected LVO to these centers has become common in an attempt to decrease time from when the patient was last known to be well (“last known well”—LKW) to clot retrieval. However, the benefit of diverting patients to different level of stroke centers, including bypassing the closest to go to a higher level of stroke care, is uncertain.1013 A large observational study with almost 1000 patients suggested that patients taken directly to endovascular centers did better (60% achieving functional independence—modified Rankin Scale (mRS) score 0-2) than patients who were transferred (52.2%; odds ratio 1.38, 95% confidence interval 1.06 to 1.79); the authors proposed that a bypass that adds less than 20 minutes would improve time to endovascular therapy and therefore outcomes.14 However, although decreasing time to endovascular therapy is likely to improve outcomes for some patients, the increased time to intravenous alteplase in a bypass may be detrimental to others. Guidelines from the American Heart Association and American Stroke Association (AHA/ASA) recommend direct transport to a CSC if the travel time is less than 15 minutes more than that to a PSC or ASRH; however, the evidence is insufficient to show that benefits at the higher level of stroke care outweigh the additional time added until evaluation.1012 A protocol to randomize patients in Denmark with likely large vessel ischemic stroke to the nearest PSC compared with bypass directly to a CSC has recently been published and may provide stronger evidence for this process.15
Several scales exist to assist EMS in identifying patients with LVO. Identification is the crucial first step in getting the right patient to the right treatment more quickly and, as the outcome depends on time to reperfusion, might improve outcomes.9 About 20 pre-hospital scales exist16; some of the most common scales used are the Los Angeles Motor Scale (LAMS),17 Cincinnati Prehospital Stroke Severity Scale (CPSS),18 and Rapid Arterial Occlusion Evaluation Scale (RACE).19 Many of the scales were designed initially to identify patients with stroke as opposed to conditions that mimic stroke, but some were specifically designed for identification of patients with stroke with LVO (for example, Vision, Aphasia, Neglect or VAN).20 However, a recent meta-analysis found substantial heterogeneity of sensitivity and specificity among studies.16 The conclusion of this meta-analysis suggested that the National Institutes of Health Stroke Scale (NIHSS), LAMS, and VAN had the best predictive value for LVO but that more testing in different populations is needed.
Populations in different settings (such as rural versus urban) may also benefit from different scales; for example, triage in rural areas may require a more specific scale given the time and distance a diversion might entail. Simple modification of the Face-Arm-Speech-Time (FAST) or the LAMS score might help to stratify the risk of an LVO, but neither has been prospectively validated in the pre-hospital setting.2122
Mobile stroke units
Mobile stroke units (MSUs), have been deployed since 2010 in a few settings as a means for decreasing time to treatment for patients with stroke, by bringing the diagnostic tools and treatments to the patient. These are essentially retrofitted ambulances that include a small bore computed tomography scanner and a laboratory unit that are sent to patients with a potential stroke for evaluation and treatment with thrombolytics onsite. However, more personnel are needed to provide thrombolysis onsite, including often a neurologist, a computed tomography technician, and a critical care nurse, in addition to paramedics.23
Most research on MSUs to date has examined the known metrics of time for treating with thrombolysis in AIS, such as alarm to treatment time and LKW to treatment time, and found better times with MSUs, but data on the outcomes of these patients compared with those not seen by the MSU are limited.2324 In addition, telemedicine is often used for remote assessment by a neurologist for these patients, and telestroke in general has been increasingly used to provide access to stroke expertise in rural, remote, and resource poor areas. Telestroke, which is a two way audiovisual communication between stroke specialists and physicians with limited neurologist coverage, has been shown to be safe and effective in both rural and urban situations.25 However, although telestroke can improve access and reduce times, whether clinical outcomes are improved is not clear,26 so guidelines give a IIa recommendation for the use of telestroke in decision making for thrombolytic treatment.10
Imaging in acute stroke
Acute ischemic stroke and intracerebral hemorrhage cannot be distinguished clinically, and treatment with thrombolytics is efficacious in the first and detrimental to the second. Therefore, all patients with suspected AIS must have emergent brain imaging, and in most situations a non-contrast head computed tomography scan is sufficient for initial management.10 As outcomes are time dependent, brain imaging should be done as quickly as possible, ideally within 20 minutes of the patient’s arrival. If it does not delay intravenous thrombolysis, non-invasive intracranial vascular imaging should be done in patients who otherwise meet criteria for endovascular clot retrieval. This can be done in combination with the initial imaging study but should not delay intravenous thrombolysis. One potential barrier to including computed tomography angiography (CTA) with the initial imaging is the concern about contrast induced nephropathy. However, evidence shows that the risk of doing CTA before obtaining a creatinine concentration in patients without known renal failure is low, and many radiology guidelines recommend that delays should not occur because of concerns about creatinine.2728293031
One measure of early ischemic changes on a non-contrast computed tomography head scan that has been used extensively in acute stroke trials is the Alberta Stroke Program Early CT Score (ASPECTS). ASPECTS is a prospectively validated score that gives points for each of 10 middle cerebral artery (MCA) territory regions that do not show early ischemic changes. Generally, an ASPECTS of 6-10 was used as an inclusion criterion in acute stroke endovascular trials, such as ESCAPE, SWIFT PRIME, and REVASCAT, to select patients with a relatively small core (that is, irreversible) infarct.
Perfusion imaging, using either computed tomography or magnetic resonance imaging (MRI), has been used to select patients for treatment who are outside typical time windows (4.5 hours for intravenous alteplase, 6 hours for endovascular therapy). Perfusion studies use contrast to measure the amount and timing of blood flow to certain areas of the brain, which can help to identify areas that have been irreversibly damaged or are at risk of damage if reperfusion is not achieved. Areas that have a very low blood flow have likely been irreversibly injured, whereas areas that have enough blood flow but high time to maximum of the residue function (Tmax) for the blood to reach that area are at risk but not yet irreversibly injured. The DAWN and DEFUSE 3 trials (see below) showed improved outcomes after thrombectomy for patients selected by specific parameters of the computed tomography or MRI perfusion study between six and 16-24 hours.56 The EXTEND alteplase study used similar perfusion studies to determine which patients might safely receive intravenous alteplase in the 4.5-9 hour window, with improved outcomes for those who received it.32
The mainstay of AIS management for the past two decades has been attempted reperfusion of ischemic tissue with intravenous thrombolysis. The recommended eligible patients and time frame for treatment have evolved over that time.
During the 1990s multiple AIS trials with intravenous alteplase treated patients up to six hours from LKW; ECASS I used a dose of 1.1 mg/kg, and ECASS II, ATLANTIS, and ATLANTIS A used 0.9 mg/kg.33343536 Each trial designated different primary and secondary endpoints ranging from acute resolution of symptoms (24 hour NIHSS score 0-1) to long term functional improvement (three month mRS, Glasgow Outcome Scale (GOS), or Barthel Index). Some trials showed a benefit of thrombolytic therapy in outcomes other than those that had been designated as the primary outcome for that study, but all failed to show a significant benefit of treatment in the primary outcome measure.
The NINDS trials (parts A and B combined for publication) treated patients up to three hours from LKW, with a requirement that half the patients had to be enrolled at less than 90 minutes from LKW.2 NINDS part A analyzed multiple endpoints, finding benefit of treatment in a global outcome score derived as a combination of the NIHSS, mRS, GOS, and Barthel Index at 90 days (odds ratio for good outcome 2.1; P=0.001). This global outcome measure was then used as the primary outcome measure for NINDS part B, which confirmed the increased rate of good outcome with intravenous alteplase (odds ratio 1.7; P=0.008). The benefit of treatment was present despite the increased risk of symptomatic intracerebral hemorrhage of 6.4% with intravenous alteplase compared with 0.6% with placebo, and no difference was seen in three month mortality (17% v 21%; P=0.30). A meta-analysis in 2004 of patient level data from all previous studies with intravenous alteplase for AIS showed that the odds ratio for favorable outcome crossed 1.0 at 270 minutes (4.5 h) from LKW.37 This analysis suggested that pushing the time window for alteplase out to 4.5 hours may be beneficial. A condition of intravenous alteplase being approved in Europe for use 0-3 hours from onset on the basis of the NINDS trials was the initiation of the ECASS III trial to evaluate the 3-4.5 hour treatment window further. ECASS III randomized 821 patients to intravenous alteplase or placebo and found that the chance of favorable outcome (three month mRS score 0-1) was 45.2% in the placebo arm and 52.4% in the treatment group.3
The Food and Drug Administration (FDA) approved alteplase for use in the US with the labeling instructions mirroring the inclusion/exclusion criteria from the NINDS trials. When alteplase was approved in Europe for AIS, the approval included more restrictive exclusion criteria based on retrospective analysis of data from the previous studies. The additional exclusion criteria mandated in Europe were age greater than 80 years, any oral anticoagulant treatment regardless of international normalized ratio, a history of both previous stroke and diabetes, and severe stroke (defined as NIHSS score >25 or with ischemic changes involving more than a third of the MCA territory on head computed tomography). After publication of ECASS III trial results, the AHA/ASA issued a scientific statement recommending use of intravenous alteplase 3-4.5 hours from LKW in patients who meet the inclusion/exclusion criteria used for the ECASS III trial.38
This difference between trials and the subsequent recommendations for treatment of patients sparked debate in the stroke community about which patients should be treated with intravenous alteplase and whether having different criteria depending on the treatment time window makes sense. A meta-analysis of individual patient data from 6756 patients specifically evaluated efficacy of treatment in different subgroups of patients and concluded that the proportional benefits of intravenous alteplase treatment were similar irrespective of age or severity of stroke.7
The AHA/ASA issued a Scientific Statement in 2016 in which the evidence base for each of the exclusion criteria was reviewed, concluding that the additional criteria from ECASS III should not be treated as strict exclusions to treatment with intravenous alteplase.39 The FDA prescribing information sheet for intravenous alteplase (Activase, alteplase) has also been updated in compliance with new labeling guidelines and factoring in the evidence base for exclusion criteria. The patient characteristics listed as exclusions have decreased significantly (see box 1), with many criteria downgraded to the “Warnings and precautions” section.39
Listed contraindications on FDA drug labeling for Activase (alteplase) use for AIS
Current intracranial hemorrhage
Active internal bleeding
Recent (within 3 months) intracranial or intraspinal surgery or serious head trauma
Presence of intracranial conditions that may increase the risk of bleeding (eg, some neoplasms, arteriovenous malformations, or aneurysms)
Current severe uncontrolled hypertension
AIS=acute ischemic stroke; FDA=Food and Drug Administration
Patient selection for intravenous thrombolytic therapy (imaging based criteria)
The DEFUSE trial was a prospective observational trial that used MRI diffusion weighted imaging (DWI) and perfusion weighted imaging (PWI) to determine the imaging characteristics associated with favorable and unfavorable responses to treatment with intravenous alteplase at three to six hours from LKW.40 This analysis defined a “target mismatch” profile that identified patients with a favorable response to reperfusion (<100 mL of DWI lesion with PWI lesion of 120% or more of the DWI lesion, but with <100 mL of PWI lesion with >8 s of Tmax delay). Another observational trial found that the presence of DWI positive lesions without corresponding lesions on fluid attenuated inversion recovery (FLAIR) correlated with patients known to be less than 4.5 hours from symptom onset.41 This DWI to FLAIR imaging mismatch was used to identify patients with unknown time of onset for randomization to intravenous alteplase or placebo in the WAKE-UP trial.42 A significant benefit of treatment with intravenous alteplase in these patients was seen, with a good outcome of mRS 0-1 in 53% of the alteplase group and 42% in the placebo group (odds ratio 1.61, 95% confidence interval 1.09 to 2.36). The EXTEND trial used automated perfusion imaging to identify patients with a target mismatch (<70 mL of core with perfusion lesion-ischemic core mismatch ratio >1.2) 4.5-9 hours from LKW or waking with symptoms for randomization to intravenous alteplase or placebo and showed a benefit in terms of a good outcome (mRS 0-1) in 35% of the alteplase group versus 29% of the placebo group (risk ratio 1.44, 95% confidence interval 1.01 to 2.06).32 This treatment benefit was confirmed in a meta-analysis combining patient data from EXTEND and two smaller trials (ECASS 4-EXTEND and EPITHET) using perfusion mismatch to select patients for randomization to alteplase or placebo in an extended time window.43 Although the exact imaging criteria used have varied, the principle of using imaging based selection criteria rather than time for patients in extended time windows has also been used for intra-arterial reperfusion strategies (see below) and seems to be safe and effective in identifying the subgroup of patients who have a slower progression than average from symptom onset to completed infarction. Using the WAKE-UP trial criteria of DWI to FLAIR mismatch to select patients who wake with symptoms for treatment with intravenous alteplase has received a class IIa (moderate) recommendation in the 2019 AHA/ASA guidelines for management of AIS.10
Tenecteplase, an alternate tissue plasminogen activator
Tenecteplase is another tissue plasminogen activator, which has been shown to have higher affinity for fibrin and a longer half life than alteplase. It is widely used for acute coronary events and has a lower rate of systemic hemorrhage than alteplase in that setting. Between 2012 and 2015 three phase II studies were published that compared standard dose alteplase (0.9 mg/kg) with variable doses of tenecteplase (0.1, 0.25, and 0.4 mg/kg) for AIS, with neutral to promising results.444546 Published in 2017, NOR-TEST was a phase III randomized, open label, blinded endpoint trial comparing tenecteplase 0.4 mg/kg with standard dose alteplase in AIS.47 This large trial randomized 1100 patients, with a predominance of mild strokes with a median NIHSS score of 4 (interquartile range 2-8), and did not show a difference between tenecteplase and alteplase in the primary outcome (mRS 0-1 at 90 days in 64% and 63%, respectively) or in safety (symptomatic intracerebral hemorrhage). Another study, EXTEND-IA TNK, randomized 202 patients with acute occlusion of the intracranial internal carotid artery (ICA), basilar artery, or MCA who were eligible for thrombolysis followed by mechanical thrombectomy to tenecteplase 0.25 mg/kg or alteplase 0.9 mg/kg.4849 Patients then underwent thrombectomy according to standard protocols. The primary outcome was based on assessment of the initial angiogram showing reperfusion of greater than 50% of the involved ischemic territory or absence of retrievable thrombus. This primary outcome occurred in 22% of the tenecteplase and 10% of the alteplase group (P=0.002 for non-inferiority and P=0.03 for superiority). Symptomatic ICH was not different, occurring in 1% of each group.
The 2019 AHA/ASA acute stroke management guidelines gave the following guidance for tenecteplase: “Tenecteplase administered as a 0.4-mg/kg single IV bolus has not been proven to be superior or noninferior to alteplase but might be considered as an alternative to alteplase in patients with minor neurological impairment and no major intracranial occlusion.”10 This class IIb recommendation was based primarily on the largest trial, NOR-TEST. Further support for the use of tenecteplase in AIS comes from a recent meta-analysis of all randomized trials comparing tenecteplase and alteplase for AIS.50 The analysis included 1585 patients and concluded that tenecteplase was non-inferior to alteplase in the treatment of AIS, with good outcome (mRS 0-1) achieved in 57.9% of patients treated with tenecteplase and 55.4% of those treated with alteplase. Symptomatic intracerebral hemorrhage occurred in 3% of both groups. As discussed, the optimal dose of tenecteplase remains unknown as the dosing was variable, but most patients received the highest dose (0.1 mg/kg in 6.8%, 0.25 mg/kg in 24.6%, and 0.4 mg/kg in 68.6%).
Complications of thrombolytic therapy
When treating patients with thrombolytic therapy, providers must be able to identify and manage the two main potential complications of treatment, intracerebral hemorrhage and angioedema. Hospitals should have protocols for reversal of coagulopathy, typically with cryoprecipitate or protein complex concentrate, although no studies have shown that these interventions are beneficial.51 Angioedema of the oropharynx can cause airway compromise and must be recognized and corrected quickly, typically with steroids, antihistamines, and intubation if needed. Further recommendations can be found in the AHA guidelines for management of AIS.10
Mechanical thrombectomy for acute ischemic stroke
Interventions in the less than six hours window
Endovascular treatment for acute stroke with LVO has revolutionized care for the most severe ischemic strokes. In 2013 initial trials of endovascular therapy versus standard care (including intravenous alteplase) for strokes with LVO (including IMS III,52 MR RESCUE,53 and SYNTHESIS54) were disappointing, as they did not show a benefit of thrombectomy over standard care. However, these previous studies were limited by lack of vessel imaging for patient selection, low use of stent retrievers, low recruitment, and slow recanalization times.55 Since then, nine studies, seven within the six hour period and two in a period up to 16-24 hours, have shown dramatically improved functional outcomes in patients undergoing endovascular treatment for LVO. The numbers needed to treat for these studies have tended to be about five for one patient to go from being dead or dependent to alive and independent.456
The clinical trial MR CLEAN was the first significantly positive trial to be presented that showed efficacy of mechanical thrombectomy in AIS. In this trial, 500 patients less than six hours from LKW were randomized across 16 medical centers in the Netherlands to either usual care (including receiving intravenous alteplase if eligible) or usual care plus endovascular clot retrieval.56 Eligible patients included those with a proximal arterial occlusion in the anterior circulation confirmed on vessel imaging, and the primary outcome was functional independence on the modified Rankin scale (mRS of 0-2) at 90 days. Most devices used in the treatment arm were stent retrievers, including Solitaire and Trevo (81.5%). An absolute difference of 13.5% (95% confidence interval 5.9% to 21.2%) in the rate of functional independence was seen in favor of endovascular therapy (32.6% v 19.1%). This was the only major trial of endovascular treatment within six hours using mostly stent retrievers that enrolled to completion.
The other four trials were ongoing at the time of this trial’s presentation, and the data safety monitoring boards of these trials were prompted to do an interim analysis. ESCAPE, EXTEND-IA, SWIFT PRIME, and REVASCAT were all stopped early owing to an overwhelming signal in the data pointing to significant improvement in functional outcomes with thrombectomy.57585960 A meta-analysis of these five trials in 2016 showed that endovascular thrombectomy in LVO significantly reduced disability at 90 days compared with control, with an adjusted common odds ratio of 2.49 (95% confidence interval 1.76 to 3.53), and estimated that the number needed to treat to reduce disability by at least one level on the mRS was 2.6.4 Importantly, pooled treatment with intravenous alteplase was 83% in the intervention group and 87% in the usual care group, and the average time from symptom onset to reperfusion in the intervention group was 285 minutes (4.75 hours). In addition, four of these trials did not go to completion, and only between 70 and 316 patients were enrolled when each study had planned for at least 500 participants.
THERAPY was another thrombectomy trial that was stopped early (108 of a planned 692 patients), comparing aspiration only thrombectomy plus intravenous alteplase with intravenous alteplase alone.61 The primary outcome of mRS of 0-2 at 90 days did not differ (38% v 30%; odds ratio 1.4, 0.6 to 3.3). The small numbers make these results difficult to interpret, but there was no suggestion of harm.
More recently, the HERMES collaboration included pooled individual data from the five previous trials, combined with data from THRACE, which randomized 414 AIS patients less than five hours from LKW to thrombectomy plus intravenous alteplase versus intravenous alteplase alone (53% with mRS of 0-2 at three months with thrombectomy versus 42% with intravenous alteplase alone),62 and a smaller study, PISTE.63 Patients were stratified by laterality of stroke and no significant differences were seen between right and left hemispheric stroke in the 90 day functional outcome (mRS ≤2: 40.7% v 37.6%; P=0.19).64 The fact that most of the endovascular treatment trials were stopped early could raise some questions about the validity of the findings, but with multiple trials and combined analyses showing similar results the evidence is strong that selected AIS patients with LVO benefit from thrombectomy less than six hours from LKW. It is estimated that about 10% of all hospital admissions for acute stroke would be eligible for thrombectomy on the basis of the eligibility criteria of these studies.65
Interventions in the greater than six hour window: DAWN and DEFUSE 3
After the multiple trials showing significantly better outcomes with endovascular therapy in LVO less than six hours from LKW, two studies (DAWN and DEFUSE 3) were published that showed large improvement in functional outcome in patients more than six hours from LKW at presentation. However, both trials had strict imaging criteria to select patients who would be most likely to benefit from delayed thrombectomy by identifying those without large areas of established infarct. The trials used software called RAPID to analyze computed tomography or magnetic resonance perfusion imaging to identify core infarction and areas of penumbra or potentially salvageable tissue (fig 1).
The DAWN trial enrolled patients who were identified six to 24 hours after LKW and had an occlusion of the ICA or the M1 segment of the proximal MCA.6 Criteria for inclusion included small core infarct volume and NIHSS score cut-off depending on age: for patients age 80 or above and NIHSS score greater than 10, infarct volume needed to be less than 21 mL; for younger patients aged under 80, infarct volume had to be less than 31 mL if the NIHSS score was above 10 and 31-50 mL if the NIHSS score was above 20. Mismatch of infarct core with clinical examinations was used for inclusion of patients for randomization. The trial was stopped early owing to an interim analysis showing pre-specified benefit. In total, 206 patients were enrolled, 107 in the intervention arm and 99 in the control arm. The rate of functional independence was dramatically better in the thrombectomy group than in the control group (49% v 13% with mRS 0-2 at 90 days). Even patients in the greater than 12 hour window had dramatic improvement in outcomes with thrombectomy; however, most of the included patients had an unwitnessed onset (so actual time from stroke onset could have been much less than time from LKW), with only 11.7% witnessed.6
DEFUSE 3 was similar to DAWN but had some differences in inclusion criteria. Patients were included if they were six to 16 hours from LKW with ICA or M1 occlusion, selected with RAPID software.5 In this trial, patients were randomized if their core infarct was less than 70 mL and the ratio of at risk ischemic tissue to core infarct volume was 1.8 or greater. Although the time window was shorter, about 60% more patients were eligible for DEFUSE 3 than for DAWN.5 This trial was also stopped early owing to pre-specified stopping rules that showed impressive benefit; 182 patients were enrolled, 92 in the intervention arm and 90 in the control arm. In this trial, 36% of cases had witnessed onset. The percentage of functionally independent patients was similar to DAWN, with 45% after thrombectomy and 17% of controls with mRS 0-2 at 90 days. These two high quality studies together led to the AHA guidelines designating mechanical thrombectomy up to 16 hours from onset for carefully selected patients as a class 1A recommendation.10 A single center study estimated that about 1.7-2.5% of all acute stroke admissions would be eligible for thrombectomy after six hours from LKW on the basis of DAWN and/or DEFUSE 3 criteria66 (fig 2).
Devices/techniques (stent retrievers, aspiration catheters)
Although stent retrievers such as Solitaire and Trevo were used in the main studies that showed efficacy for thrombectomy after intravenous alteplase versus alteplase alone,4566367 other devices and techniques are also used for thrombectomy that have varying levels of evidence. Previous devices, including the Merci device and early Penumbra aspiration catheters with separators, did not show efficacy in earlier trials,525354 and although problems such as patient selection likely also affected outcomes of these earlier trials, stent retrievers have consistently shown better recanalization rates (see table 1 for explanation of the Treatment in Cerebral Ischemia (TICI) rating system for recanalization68) and clinical outcomes.697071 More recently, another stent retriever, EMBO-TRAP, has been evaluated in the single arm prospective study ARISE II; although not compared directly with other devices, it showed similar rates of recanalization (TICI 2B-3 of 80.2% within three passes), functional independence (mRS 0-2 in 67%), and mortality (9%) at 90 days as studies involving Solitaire and Trevo.72 An important caveat is that this study had a single arm with no comparators, so biases may be present, and concluding that EMBO-TRAP is a safe and effective option for thrombectomy is difficult; future research will need to evaluate it against other devices.
The first line use of contact aspiration versus stent retriever thrombectomy in LVO is evolving as a more common technique, but evidence to support its use is mixed. This technique, specifically one called “a direct aspiration as first pass technique” (ADAPT), has been studied because observational studies suggested that it might lead to earlier reperfusion and lower cost of thrombectomy.7374 This led to two randomized clinical trials, ASTER (Contact Aspiration versus Stent Retriever for Successful Revascularization) in 2017 and COMPASS (Aspiration Thrombectomy Versus Stent Retriever Thrombectomy as First-Line Approach for Large Vessel Occlusion) in 2019,757677 which both examined the ADAPT technique versus stent retriever within six hours of onset. The ASTER trial was designed as a superiority trial and the COMPASS trial as non-inferiority. Both trials studied recanalization rates within three passes and mRS outcomes at 90 days. ASTER showed recanalization of TICI 2b or better of 85.4% in the ADAPT arm versus 83.1% in the stent retriever arm and a 90 day mRS score of 2 or less in 45.3% versus 50.0% (P=0.38).75 COMPASS showed TICI of at least 2b in 92% in the ADAPT arm and 89% in the stent retriever arm (P=0.54) and a 90 day mRS score of 2 or less in 52% versus 50% (P=0.001 for non-inferiority). Essentially, the current data support equivalency of first line aspiration with first line stent retriever, but not superiority of this technique.
Intubation versus conscious sedation for mechanical thrombectomy procedure
Controversy continues to exist around the optimal way to manage anesthesia during endovascular treatment. Multiple non-randomized and a few randomized studies have examined different options, including conscious sedation or general anesthesia. In general, non-randomized studies have found that general anesthesia leads to worse functional outcome and a higher mortality rate,787980818283 but some randomized, single center trials including AnStroke,84 SIESTA,85 and GOLIATH,86 showed no difference in mortality for general anesthesia versus conscious sedation, with better functional outcomes for general anesthesia in SIESTA but no difference in AnStroke or GOLIATH.
A recent meta-analysis of these three trials suggests better outcomes with general anesthesia.87 However, a pre-planned sub-study of DEFUSE 3 published in 2019 that examined the outcomes of patients who underwent general anesthesia versus conscious sedation, found significantly better functional independence (mRS ≤2) in patients who underwent conscious sedation, with no difference in mortality or symptomatic intracerebral hemorrhage.88 A pooled analysis of trials in the HERMES collaborative found that use of general anesthesia was associated with worse functional outcomes at 90 days compared with conscious sedation or non-general anesthesia, but with no difference in mortality or symptomatic intracerebral hemorrhage.89
The drawback of many of these studies, including the last two, is that the decision to use general anesthesia or conscious sedation was left up to the treating team, and some locations did all general anesthesia and others did all conscious sedation. Multiple comparisons have been done to adjust for different factors, but at this point definitively saying that general anesthesia or conscious sedation is better for outcomes is still difficult. It is also important to recognize that the trials were designed to avoid blood pressure extremes, that the rate of conversion from conscious sedation to general anesthesia was high (between 6.3% and 15.6%), that no significant difference was seen in time to groin puncture (although it was a little longer with general anesthesia), that no significant difference was seen in time to recanalization after groin puncture (although it was a little longer with conscious sedation), and that assessment of pneumonia outcomes was limited, although these are important when considering intubation as an intervention.
The 2019 AHA/ASA guidelines conclude that “It is reasonable to select an anesthetic technique during endovascular therapy for AIS on the basis of individualized assessment of patient risk factors, technical performance of the procedure, and other clinical characteristics.”10
Management of physiological factors
Blood pressure management in acute stroke
Management of blood pressure in AIS must balance multiple factors including elevated pressures improving tissue perfusion but also increasing the risk of hemorrhage or secondary damage to already infarcted areas of brain. With these competing factors, the optimal blood pressure for any given patient will vary depending on whether the patient has been treated with thrombolytic therapy or embolectomy, as well as the patency of major vessels and perfusion status of the brain, if these factors are known.
Some trials have evaluated blood pressure management in the acute phase. The recent RIGHT-2 trial had EMS personnel randomize patients in the ultra-acute period and start treatment using glyceryl trinitrate or placebo; it found no difference in functional outcome or death.90 A meta-analysis of 13 RCTs of antihypertensive agents started between 15 hours and three days after onset of AIS also had neutral results for both likelihood of death or dependency at three months and recurrent vascular events.91 The 2019 AHA/ASA acute stroke guidelines recommend: “In patients with BP ≥220/120 mm Hg who did not receive IV alteplase or EVT and have no comorbid conditions requiring acute antihypertensive treatment, the benefit of initiating or reinitiating treatment of hypertension within the first 48 to 72 hours is uncertain. It might be reasonable to lower BP by 15% during the first 24 hours after onset of stroke.”10
The protocols for the thrombolytic trials instructed that blood pressure should be controlled to below 180/105 mm Hg after thrombolytic treatment. This protocol had not been compared with alternate blood pressure goals until the recent ENCHANTED trial,92 in which patients eligible for thrombolytic therapy were randomized to standard goal systolic pressure (<180 mm Hg) or intensive management (goal 130-140 mm Hg). Fewer hemorrhages occurred in the intensive therapy group, but no difference was seen in three month mRS (however, the mean systolic blood pressure was not very different between the groups despite different goals (144.3 v 149.8 mm Hg)).
Most patients enrolled in mechanical thrombectomy trials were also eligible for and treated with intravenous alteplase, so blood pressure was controlled as per post-alteplase guidelines (<180/105 mm Hg). Debate remains regarding whether blood pressure should be lowered further post-thrombectomy if reperfusion is achieved. The DAWN trial protocol recommended systolic pressure below 140 mm Hg for 24 hours after reperfusion. A single site observational study of patients undergoing mechanical thrombectomy stratified by maximum blood pressure in the 24 hours post-thrombectomy: permissive hypertension (<180/105 mm Hg for patients treated with intravenous alteplase or <220/110 mm Hg for those not treated), moderate blood pressure goal (<160/90 mm Hg), or intensive blood pressure goal (<140/90 mm Hg).93 This study did not find a difference in symptomatic intracerebral hemorrhage but did find that high maximum blood pressure levels were independently associated with increased likelihood of three month mortality and functional dependence (mRS >2). As this was a non-randomized study, the maximum blood pressure achieved could be associated with poor outcomes but not causative. A survey of 58 StrokeNet institutions showed variability in current practice patterns for blood pressure management post-thrombectomy.94 Systolic blood pressure goals for patients with full recanalization varied widely: 120-139 mm Hg in 36%, 140-159 mm Hg in 21%, and below 180 mm Hg in 28%. For patients with unsuccessful reperfusion, blood pressure goals were generally higher, with 43% of respondents accepting any value below 180 mm Hg and 10% accepting below 220 mm Hg.
In addition to permissive blood pressure, positioning patients fully supine (that is, with the head of the bed flat) is another strategy that has been proposed to increase cerebral perfusion in the acute stroke setting. Small studies evaluating physiologic outcome parameters have shown that positioning patients fully supine improves cerebral blood flow velocities and cerebral blood volume in some patients with AIS, especially those with poor cerebral autoregulation.9596979899 This potential advantage to perfusion brings with it the potential increased risk of aspiration. The HeadPoST trial published in 2017 assigned 11 093 patients to either lying flat or lying with the head elevated at least 30 degrees, with positioning started soon after hospital admission and maintained for 24 hours.100 No difference was seen between the groups in disability outcomes at 90 days.
A critique of the HeadPoST protocol was published questioning the validity of the results.101 The main concerns were inclusion of all types of strokes, as the authors felt that minor strokes would be unlikely to benefit from the intervention, and the intervention being started too late to be beneficial (median time to intervention initiation was 14 hours from LKW and seven hours from hospital admission). Although no evidence based recommendation on head positioning exists that can be applied to all patients with AIS, certain patients with disrupted autoregulation and tenuous penumbral perfusion may benefit from positioning with head of the bed flat, especially in the hyperacute stages before definitive reperfusion strategies are started.
Blood sugar management
Persistent hyperglycemia in the acute stroke period has been associated with poor outcomes. However, whether tighter control of blood glucose would be beneficial was unclear. The SHINE trial randomized patients to receive either intensive glucose control for 72 hours post-stroke (target 80-130 mg/dL with intravenous insulin infusion) or standard therapy with sliding scale insulin targeted to 80-179 mg/dL.102 The study was stopped for futility at the fourth planned interim analysis after 82% (1151/1400) of the planned maximum number of patients had been enrolled. The primary endpoint of favorable three month outcome was reached in 20.5% of patients in the intensive treatment group and 21.6% of patients in the standard treatment group (relative risk 0.97; P=0.55). Severe hypoglycemia occurred in 2.6% of patients in the intensive group and none in the standard treatment group.
The AHA/ASA guidelines recommend supplemental oxygen to maintain O2 saturation above 94%. Some debate has taken place about whether all patients should be treated with supplemental oxygen in the acute stroke setting. The Stroke Oxygen Study (SO2S) investigated this in 8003 non-hypoxic patients with acute stroke within 24 hours of admission by randomizing them to continuous nasal cannula oxygen, nocturnal nasal cannula oxygen, or control (oxygen only if clinically indicated) for 72 hours.103 No difference was seen between the groups in likelihood of a good outcome, and no subgroups were identified that benefited from oxygen.
Acute antithrombotic management for secondary prevention
Management has two main objectives when patients present with acute ischemic stroke or transient ischemic attack (TIA): to minimize disability from the acute event and decrease the likelihood of another stroke. The risk of recurrent stroke is highest soon after presentation, when presumably the factors leading to the current event are still in play (for example, ruptured atherosclerotic plaque with thrombus). Therefore, secondary stroke prevention strategies will be most successful if implemented as soon as possible. Two RCTs published in 1997 formed the basis of the standard of care treatment of acute stroke patients with aspirin for secondary stroke prevention, the International Stroke Trial (IST) and the Chinese Acute Stroke Trial (CAST).104105 Both studies randomized patients as soon as possible after onset (within 48 hours for CAST) to aspirin (162 mg and 300 mg, respectively) or placebo. In total, 40 541 patients were randomized and the results analyzed together indicated that treatment with aspirin prevented about 10 deaths or recurrent strokes per 1000 patients treated during the first few weeks. The IST also randomized patients to heparin or placebo and concluded that “Patients allocated to heparin had significantly fewer recurrent ischaemic strokes within 14 days (2.9% vs 3.8%) but this was offset by a similar-sized increase in haemorrhagic strokes (1.2% vs 0.4%), so the difference in death or non-fatal recurrent stroke (11.7% vs 12.0%) was not significant.”
Two recent RCTs have re-examined this topic to determine whether dual antiplatelet therapy with aspirin and clopidogrel is superior to aspirin alone in preventing recurrent stroke soon after the presenting stroke or TIA. The CHANCE trial randomized 5170 patients in China presenting with TIA or minor stroke within 24 hours from onset to either aspirin 75 mg or aspirin 75 mg and clopidogrel (300 mg load and then 75 mg per day) for 21 days.106 Analysis at 90 days found that stroke occurred in 8.2% of the dual antiplatelet group compared with 11.7% of the aspirin group (hazard ratio 0.68; P<0.001). No difference was seen between the groups in the rate of moderate to severe hemorrhage or in intracerebral hemorrhage.
The POINT trial randomized 4881 patients at 269 international sites (with 82.8% enrolled in the US) to either aspirin (dose 50 mg to 325 mg) or aspirin and clopidogrel (600 mg load followed by 75 mg daily for 90 days) within 12 hours from onset.107 The Data Safety Monitoring Board stopped the trial early after 84% of the planned patients had been enrolled when the determination was made that the dual antiplatelet group had both a lower risk of major ischemic events (5.0% v 6.5%) and a higher risk of major hemorrhage (0.9% v 0.4%) at 90 days. Interestingly, most of the ischemic events occurred within the first week after the initial event, whereas the risk of hemorrhagic events remained relatively constant throughout the trial period. A recent BMJ Rapid Recommendations Panel did a meta-analysis of RCTs examining dual versus single antiplatelet agents started acutely after presentation with ischemic stroke or TIA.108 The panel recommended starting dual antiplatelet therapy within 24 hours of onset of TIA or minor stroke symptoms and continuing for 10-21 days on the basis of the finding that this practice reduces non-fatal recurrent stroke (ischemic or hemorrhagic) by 1.9% while increasing moderate to major extracranial bleeding by 0.2% and having no effect on all cause mortality, myocardial infarction, or recurrent TIA.
Emerging treatments in the area of AIS management include research into expanding indications for thrombectomy, stem cell therapy, and neuroprotection. As advanced imaging is increasingly being used to select patients who will most benefit from thrombectomy, a question remains as to whether even patients with less favorable imaging characteristics would also derive more benefit than risk from thrombectomy.109110 Recently initiated trials, including TELSA, SELECT 2, TENSION, and IN EXTREMIS (https://clinicaltrials.gov/) aim to establish the effectiveness of thrombectomy versus medical management in patients with moderate to large infarcts established on non-contrast computed tomography of the head (ASPECTS111 of 2-5 for TESLA and 3-5 for the others). These studies may help us to understand whether we can offer more for patients presenting with larger established infarcts than we currently do. In addition, whether patients with low severity (low NIHSS score) and LVO should get thrombectomy is unclear; some analyses of cohorts have reported either benefit of thrombectomy or no difference between those that received thrombectomy and best medical care.112113 A second part of IN EXTREMIS (called MOSTE or Minor Stroke Therapy Evaluation) and the Endovascular Therapy for Low NIHSS Ischemic Strokes (ENDOLOW) are planned randomized trials that will attempt to answer this question.
Time to reperfusion continues to predict outcome in endovascular treatment, so the possibility of direct triage to the angiography suite is potentially desirable. Small series of patients have examined “one stop imaging” with a direct-to-angiography suite for imaging with the flat panel detector computed tomography and decision making about intravenous alteplase and thrombectomy taking place there, with the goal of avoiding time in a separate computed tomography scan.114115 If the quality of the flat panel detector computed tomography to detect hemorrhage, ischemic changes, and LVO is confirmed, this might be an option to significantly reduce time at endovascular capable centers. Another potential time saver for patients eligible for thrombectomy might be to bypass intravenous alteplase to go directly to endovascular treatment. Ongoing studies, such as SWIFT DIRECT, are examining whether bridging with thrombolysis plus thrombectomy is superior to thrombectomy alone and, if not, whether time saved by going directly to thrombectomy could potentially improve outcomes.
The use of stem cells in recovery from stroke has been tested for decades in animal models, and more recently in humans, but now interest exists in testing stem cell treatments in the acute phase, with the hope of early recovery of neurologic function. A 2019 Cochrane Database review of randomized trials in stem cell transplantation in ischemic stroke identified seven trials involving 401 participants, and overall stem cell transplantation was associated with better clinical outcome when measured by NIHSS but not by mRS or other measures of disability.116 The two larger trials with 60 or more patients did not show any benefit, whereas the smaller trials did show a small benefit, and there were no safety concerns. The authors’ conclusions were that there is currently “insufficient evidence to support or refute the use of stem cell transplantation to treat ischemic stroke” and that more research is “urgently needed.”116 Even less evidence exists in the acute phase, but a search of clinicaltrials.gov using key words “acute stroke” and “stem cells” retrieves 17 either recently completed or ongoing trials worldwide. Intravenous or intra-arterial injection of stem cells might be a promising treatment for acute stroke in the future but is of uncertain value at this time.
Neuroprotection in ischemic stroke has been studied extensively, resulting from an interest in extending the time neurons can survive ischemia until recanalization or other forms of reperfusion can be achieved. An ideal neuroprotectant would have few adverse effects and would be easy to use in the earliest setting of stroke, either onsite or at least pre-hospital. Unfortunately, after more than 1000 published experiments and more than 100 clinical trials, although multiple animal studies have showed promising results, no data in humans have shown efficacy.117 The AHA/ASA 2019 guidelines recommend against the use of neuroprotective agents, as class III (no benefit), level of evidence A (highest quality).10 Nevertheless, research continues to find agents or strategies that could provide neuroprotection in AIS, and the most promising efforts at this point seem to be intra-arterial delivery of drugs, stem cells, and local hypothermia.118
Multiple stroke organizations worldwide periodically review the available literature and produce updated guidelines for the management of acute ischemic stroke. The Royal College of Physicians in the UK last updated the National Clinical Guidelines for Stroke in 2016, which can be found online at https://www.strokeaudit.org/Guideline/Guideline-Home.aspx. The Canadian Stroke Best Practice Recommendations for AIS treatment were updated in 2018 and are available as online modules at https://www.strokebestpractices.ca/recommendations/acute-stroke-management/acute-ischemic-stroke-treatment and also published in the International Journal of Stroke.119 The Australian Stroke Foundation states that its Clinical Guidelines for Stroke Management are “living guidelines,” which are updated as new evidence emerges and can be found at https://informme.org.au/en/Guidelines/Clinical-Guidelines-for-Stroke-Management. The most frequently cited stroke guidelines are published in Stroke by the American Heart Association and American Stroke Association. The most recent update to the Guidelines for the Early Management of Acute Ischemic Stroke was published in 2019.10 The format of the AHA/ASA guidelines has evolved in recent years to more clearly link the recommendation statements to the supporting scientific evidence and to clearly indicate the class (strength) of recommendation and the level (quality) of evidence.
Guidelines from these different organizations are mostly very similar to each other. Where differences exist, they are typically in the strength of the recommendation made about newly published data, with some guideline committees endorsing new protocols, whereas others require a higher level of data before making a strong recommendation. Practitioners are encouraged to look for updated guidelines periodically, as they will continue to be updated as the dynamic field of acute stroke care continues to evolve.
Management of AIS has undergone many changes in the past few years, with more patients receiving treatment to minimize long term disability. A critical advance has been the establishment of organized regional stroke systems of care that can quickly identify patients with stroke in the field and use decision support to get patients to the appropriate centers that can provide state of the art care for their condition. This includes doing the necessary clinical and imaging evaluation and interpretation of those results by clinicians with expertise in determining patients’ eligibility for rapid administration of intravenous thrombolytic therapy and endovascular thrombectomy. The devices available for endovascular thrombectomy continue to be improved, and the appropriate procedural and subacute management of patients continue to be refined. Presentation with acute stroke is also the time to begin acute measures aimed at preventing additional strokes in this high risk population. Appropriate application of the available treatments is crucial to optimizing outcomes of patients with stroke.
What are the best systems of care and protocols that can help emergency medical services to identify patients with large vessel occlusion and to determine to which level of stroke center to triage a patient?
Is endovascular thrombectomy efficacious in patients with acute ischemic stroke who have larger established core infarcts? What about those with low National Institutes of Health Stroke Scale scores and large vessel occlusion?
What is the best method of patient sedation during endovascular therapy (general anesthesia versus conscious sedation)? Would certain patients benefit from one approach versus the other?
What is the optimal blood pressure management during endovascular therapy and after successful recanalization of large vessel occlusion?
Is stem cell therapy effective for stroke recovery?
Glossary of abbreviations
ADAPT—a direct aspiration as first pass technique
AHA/ASA—American Heart Association/American Stroke Association
AIS—acute ischemic stroke
ASPECTS—Alberta Stroke Program Early CT Score
ASRH—acute stroke ready hospital
CPSS—Cincinnati Prehospital Stroke Severity Scale
CSC—comprehensive stroke center
CTA—computed tomography angiography
DWI—diffusion weighted imaging
EMS—emergency medical services
FDA—Food and Drug Administration
FLAIR—fluid attenuated inversion recovery
GOS—Glasgow Outcome Scale
ICA—internal carotid artery
LAMS—Los Angeles Motor Scale
LKW—last known well
LVO—large vessel occlusion
MCA—middle cerebral artery
MRI—magnetic resonance imaging
mRS—modified Rankin Scale
MSU—mobile stroke unit
NIHSS—National Institutes of Health Stroke Scale
PSC—primary stroke center
PWI—perfusion weighted imaging
RACE—Rapid Arterial Occlusion Evaluation Scale
RCT—randomized controlled trial
TIA—transient ischemic attack
TICI—Treatment in Cerebral Ischemia
Tmax—time to maximum of the residue function
TSC—thrombectomy ready stroke center
A patient treated for his acute ischemic stroke reviewed the draft of this manuscript and made suggestions and edits on the content and presentation of the article, specifically for the language (including spelling out acronyms), the inclusion of graphics, and the inclusion of clear summaries and recommendations
Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors
Contributors: MSP and CAC both did the literature search and prepared the initial draft of the manuscript. Both authors were substantially involved in the conception, drafting, and editing of the manuscript. Both authors have given final approval of the manuscript and are accountable for all portions of the manuscript. MSP is the guarantor.
Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: none.
Provenance and peer review: Commissioned; externally peer reviewed.