Acute care

A number of pre-hospital screening tools have been developed that are sensitive in detecting the majority of people with acute stroke that present with facial weakness, speech disturbance or unilateral limb weakness. The FAST test is accepted as the screening tool of choice for clinicians and the general public (Harbison et al, 2003). However, some people with symptoms of stroke will not be identified by the FAST test (e.g. sudden onset visual disturbance, lateralising cerebellar dysfunction) and thus stroke may not be suspected. The Working Party considers that community-based clinicians should continue to treat a person as having a suspected stroke if they are suspicious of the diagnosis despite a negative FAST test. Further evidence is required before the Working Party could recommend the use of other screening tools, e.g. forms of the National Institutes of Health Stroke Scale (NIHSS), Recognition of Stroke in the Emergency Room (ROSIER) that screen for non-FAST symptoms in the pre-hospital phase. [2016]

Community-based clinicians are likely to assess people whose neurological symptoms have already resolved before reaching hospital, suggesting a diagnosis of TIA rather than stroke. These people should be given antiplatelet treatment with aspirin immediately (Rothwell et al, 2016) and referred for urgent investigation in a specialist neurovascular clinic since the risk of subsequent stroke is substantial in the first few days. There was insufficient precision in the current evidence for the Working Party to make recommendations concerning risk stratification by community-based clinicians. When considering the diagnosis of TIA and making a direct referral, clinicians need to be aware that a person may have ongoing focal neurological deficits despite a negative FAST test – such people should be managed along the acute stroke pathway rather than a TIA pathway. [2016]

There is a general paucity of research evidence on the management of the person with suspected stroke before arrival at the hospital. The majority of recommendations are based on consensus and widely accepted practice in the pre-hospital management of people with suspected stroke or TIA. Pre-hospital brain imaging in other healthcare settings may reduce onset-to-treatment time (Ebinger et al, 2014), but cost-effectiveness and clinical outcomes have not been tested in UK/Ireland settings. [2016]

The training of primary care teams, other healthcare personnel and the general public in the recognition of the signs of possible stroke using the FAST involves an ongoing public health commitment requiring multiple approaches (Section 2.1 Public awareness of stroke). Patients in groups at high risk of stroke (e.g. older people with diabetes, hypertension or atrial fibrillation) and their family and/or carers should have training in the FAST test as part of their disease education. [2016]

A systematic review of observational studies of the risk of stroke within 7 days of confirmed TIA (Giles & Rothwell, 2007) showed a risk of stroke at 2 days of between 2.0 and 4.1%, and at 7 days of between 3.9 and 6.5%. The risk of completed stroke was much lower in studies of emergency treatment in specialist stroke services compared to non-urgent settings (0.9% v. 11.0%). Patients with suspected TIA should therefore have a full diagnostic assessment urgently without further risk stratification (Lavallee et al, 2007; Wardlaw et al, 2014). Risk predictive clinical tools such as ABCD2, ABCD2-I and ABCD3-I scores did not discriminate sufficiently between low- and high-risk patients in both the short-term and long-term follow-up (Ildstad et al, 2021) either for determining the urgency of assessment or subsequent treatment options (Amarenco et al, 2012) and their use is no longer recommended. Secondary prevention measures which can reduce the risk of recurrence should be promptly initiated (Rothwell, 2007). Additional risk may be conferred by the presence of atrial fibrillation or anticoagulant therapy, or with recurrent attacks, while patients presenting with symptoms more than a week ago can be considered at lower risk. [2023]

There is little evidence to guide the use of brain imaging in suspected TIA. The consensus of the Guideline Development Group is that imaging all people referred to a neurovascular clinic is not always appropriate or cost-effective given the high rate of stroke mimics in most clinics. Patients with suspected TIA should normally be assessed by a specialist clinician before a decision on brain imaging is made, and imaging should be used in those patients where the results are likely to influence management such as reducing diagnostic uncertainty, confirming the territory of ischaemia prior to making a decision about carotid artery surgery and prior to commencing dual antiplatelet therapy. When the exclusion of haemorrhage is the objective of imaging, early unenhanced computed tomography (CT) remains the most sensitive investigation (Wardlaw et al, 2014). The greater sensitivity of magnetic resonance imaging (MRI) to detect ischaemic lesions using diffusion-weighted imaging (DWI) makes it the modality of choice if positive confirmation of the presence or location of a lesion is the objective (Whiteley et al, 2022), but the significant false-negative rate with DWI precludes its use as a diagnostic tool in isolation from clinical assessment, particularly in unselected patients (Wardlaw et al, 2014). [2023]

Additional training of healthcare professionals and other primary care staff may be required so that they are able to appreciate the immediate risk in people presenting with a suspected TIA or minor stroke, advise immediate aspirin use where appropriate and expedite the process of referral for diagnostic assessment. Referrers and neurovascular clinics should discontinue the practice of triaging patients with suspected TIA according to risk stratification tools, and ensure that all patients with suspected TIA are assessed and diagnosed urgently 7 days a week. [2023]

Ischaemic stroke and TIA are similar manifestations of vascular disease and their treatment for the prevention of recurrent vascular events reflects this. Long-term risk factor management is reviewed in Chapter 5 on secondary prevention. In the acute setting, there is no evidence to support the use of anticoagulation for recurrent TIA for those in sinus rhythm. There are also no studies specifically addressing the clinical benefit of early anticoagulation for patients with cardioembolic TIA, but the consensus of the Working Party is that the balance of risk and benefit from the early secondary prevention of cardioembolism with a rapidly-acting anticoagulant is favourable in the majority of patients. [2016]

Recent evidence supports the early use of dual antiplatelet therapy in patients with TIA or minor ischaemic stroke. The CHANCE trial (Wang et al, 2013) showed in a Chinese population of patients with TIA or minor ischaemic stroke (NIHSS 0-3) that dual antiplatelet therapy started within 24 hours of onset for 21 days resulted in a significant reduction in ischaemic stroke from 11.7% (aspirin group) to 8.2 % (aspirin-clopidogrel group) with no significant difference in haemorrhagic stroke. A 300 mg loading dose of clopidogrel was used in this trial. The POINT trial (Johnston et al, 2018) showed in patients with TIA or minor ischaemic stroke (NIHSS 0-3) that dual antiplatelet therapy started within 12 hours of onset and continued for 90 days resulted in a significant reduction of ischaemic stroke from 6.5% (aspirin group) to 5.0% (aspirin-clopidogrel group). The trial was stopped early because of effectiveness and a significant increase in major haemorrhage in the dual antiplatelet group with an absolute risk increase of 0.5%. A 600 mg loading dose of clopidogrel was used in this trial. Pooled analysis of these trials demonstrated the benefit of dual antiplatelet therapy up to at least 21 days. The THALES trial (Johnston et al, 2020), showed in patients with a high-risk TIA or minor ischaemic stroke (NIHSS 0-5) that dual antiplatelet therapy with aspirin and ticagrelor started within 24 hours of onset resulted in a significant reduction in composite outcome of stroke or death from 6.6% (aspirin group) to 5.5% (aspirin-ticagrelor group) within 30 days. A 180 mg loading dose of ticagrelor was used. Severe bleeding was more frequent with ticagrelor with an absolute risk increase of 0.4%. In these three trials, high risk TIA was defined using an ABCD2 score of 4 or 6, despite evidence that the score performs poorly in successfully identifying TIAs previously identified as low risk which should be considered high risk (Amarenco et al, 2012). In these trials of TIA and minor stroke, patients underwent either CT or MRI to exclude the presence of cerebral haemorrhage or alternative explanations for their symptoms. [2023]

A substantial proportion of patients with TIA and stroke in some populations may be resistant to clopidogrel (Pan et al, 2017). Clopidogrel resistance and non-responsiveness is linked to a genetic polymorphism in the cytochrome CYP2C19 loss-of-function allele. Ticagrelor is not affected. In a selected Chinese population with a genetic polymorphism in the CYP2C19 loss-of-function allele, the CHANCE 2 trial (Wang et al, 2021c) showed that the risk of ischaemic stroke was significantly reduced from 7.6% in patients treated with clopidogrel (90 days) plus aspirin (21 days) to 6.0% in those randomised to ticagrelor (90 days) plus aspirin (21 days). These results are promising but not yet generalisable owing to the selected nature of the population. [2023]

Carotid imaging is essential for any patient, presenting with symptoms suggesting anterior circulation cerebral ischaemia, who might be suitable for intervention for carotid stenosis. There are two methods for reporting the degree of carotid stenosis that give differing results, derived from the North American Symptomatic Carotid Endarterectomy Trial (NASCET) and the European Carotid Surgery Trial (ECST). Both are valid but the Working Party considers that the NASCET method is preferred (Rothwell et al, 2003b). After carotid territory TIA or non-disabling stroke, the Working Party consensus was that carotid intervention should be performed as soon as possible. [2016]

A Cochrane review of carotid stenting for symptomatic carotid stenosis identified a higher risk of both short- and long-term stroke complications, especially in patients 70 years and older, but a lower risk of peri-procedural myocardial infarction and cranial nerve injury (Bonati et al, 2012). This section should be read in conjunction with Section 5.13 Carotid artery stenosis. [2016]

Local health economies and networks should establish clinical pathways designed to expedite referral of appropriate patients to vascular surgical centres, especially where reconfiguration has resulted in vascular services being at another site. [2016]

A protocol should be in place for the use of parenteral or direct oral anticoagulants (DOACs) in the setting of a neurovascular clinic, with a process to supervise the transition from acute to long-term anticoagulation. [2016]

The evidence supporting stroke unit care from the Stroke Unit Trialists in the 1990s has been updated in a 2013 Cochrane review, which found that patients with stroke who receive organised inpatient care in a stroke unit are more likely to be alive, independent, and living at home one year after a stroke (Stroke Unit Trialists’ Collaboration, 2013). [2016]

Imaging patients with suspected stroke immediately is cost-effective compared to other approaches because it enables emergency treatments directed at the pathology of stroke (Wardlaw et al, 2004), although MRI with DWI may be required when diagnostic uncertainty persists (Wardlaw et al, 2014). Interpretation of acute stroke imaging by trained non-radiologists is safe and effective (Spokoyny et al, 2014). Multi-modal imaging is becoming more widely available and quicker to undertake and interpret. It is recommended that if multi-modal imaging is required, all relevant imaging should be performed in the same session to avoid delays in decision making for acute treatments (National Optimal Stroke Imaging Pathway [NOSIP], (NHS England, 2021). The recommendations regarding reperfusion treatments contained in Section 3.5 Management of ischaemic stroke require CT angiography (with or without CT perfusion) to be performed immediately in potentially eligible patients, but image processing and interpretation should not delay intravenous thrombolysis if this is indicated. [2023]

These recommendations align with the recommendations in Chapter 2 concerning the organisation of acute stroke care and Sections 3.5 Management of ischaemic stroke and 3.6 Management of intracerebral haemorrhage regarding the management of acute stroke. Hyperacute stroke services will need to review their provision of specialist assessment and imaging policies for people with suspected acute stroke, which in many centres will involve a step-change in provision. [2023]

An updated Cochrane systematic review and an individual patient meta-analysis by the Stroke Thrombolysis Trialists’ Collaboration guide the use of intravenous thrombolysis without advanced imaging (International Stroke Trial Collaborative Group, 2012; Wardlaw et al, 2012; Emberson et al, 2014). These analyses emphasise the importance of rapid treatment. Patients who are over 80 years old with mild or severe stroke and those with early signs of infarction on initial brain imaging all benefit from treatment. [2023]

The Enhanced Control of Hypertension and Thrombolysis Stroke Study (ENCHANTED) of lower (0.6 mg/kg) versus standard dose alteplase showed a lower risk of intracerebral haemorrhage and early mortality with the lower dose, without conclusively demonstrating that the doses were of equivalent efficacy (Anderson et al, 2016). These findings suggest that there may be circumstances in which the treating physician and/or the patient wish to forgo some of the potential disability benefit from standard dose alteplase in order to reduce the early risk of intracerebral haemorrhage through the use of the lower dose. A meta-analysis of risk factors for intracerebral haemorrhage with alteplase (Whiteley et al, 2012) suggested a greater risk in people with atrial fibrillation, congestive cardiac failure, renal impairment, prior antiplatelet treatment, leukoaraiosis and visible cerebral infarction on pre-treatment brain imaging, but the extent to which any of these factors should influence dose selection for alteplase remains unknown. [2016]

Patients presenting with acute ischaemic stroke whilst taking a direct oral anticoagulant (DOAC) should be excluded from receiving thrombolysis unless, in the case of dabigatran, the prothrombin time and activated partial thromboplastin time are both normal. The use of reversal agents (idarucizumab or andexanet alfa) in order to then administer thrombolysis for an ischaemic stroke that has occurred during DOAC treatment is not recommended. [2016]

Patients more than 4.5 hours after stroke or with an unknown time of onset, and those with wake-up stroke, who have radiologically defined ‘penumbra’ benefit from alteplase. Participants in the WAKE-UP trial (Thomalla et al, 2018) were aged up to 80 years and had woken from sleep or could not report the time of stroke onset, were at least 4.5 hours from when last seen well, and had evidence of MRI DWI-FLAIR mismatch. Allocation to 0.9 mg/kg alteplase rather than placebo led to a higher proportion of patients (53.3% v. 41.8%) with an excellent functional outcome at 90 days. The THAWS trial (Koga et al, 2020) demonstrated no clear benefit from a lower dose of alteplase (0.6mg/kg). [2023]

In an individual participant data meta-analysis of EXTEND (Ma et al, 2019), ECASS-4 (Ringleb et al, 2019) and a subset of the EPITHET trial (Davis et al, 2008) patients with CT or MR perfusion imaging-defined penumbra between 4.5 and 9 hours after onset or with wake-up stroke (Campbell et al, 2019) gained benefit from intravenous alteplase (7% absolute increase in excellent outcome). No patients received mechanical thrombectomy in these trials. In a prespecified subset who had mismatch defined by a mismatch ratio greater than 1·2, a mismatch volume greater than 10 mL, and an ischaemic core volume less than 70 mL (EXTEND criteria), benefit with alteplase was also seen. Participants with and without a large artery occlusion appeared to benefit similarly. The TWIST trial (Roaldsen et al, 2023) comparing intravenous thrombolysis with tenecteplase given within 4.5 hours of awakening versus control (no thrombolysis) in patients with wake-up stroke, did not demonstrate benefit in patients selected with non-contrast CT imaging alone. [2023]

Tenecteplase is a single bolus thrombolytic agent with higher fibrin specificity and longer half-life than alteplase. Nine RCTs have compared tenecteplase with alteplase in people with acute ischaemic stroke (Haley et al, 2010; Parsons et al, 2012; Huang et al, 2015; Logallo et al, 2017; Campbell et al, 2018b; Bivard et al, 2022; Kvistad et al, 2022; Menon et al, 2022; Wang et al, 2023). No single trial in unselected patients has demonstrated that tenecteplase leads to greater recovery than alteplase. A 2019 meta-analysis (Burgos & Saver, 2019) concluded that tenecteplase was non-inferior to alteplase but this was confounded by the significant contribution of the large NOR-TEST study which used a higher dose of 0.4 mg/kg and included a substantial proportion of people with stroke mimics (Logallo et al, 2017). A subsequent trial of 0.4 mg/kg tenecteplase in patients with moderate-severe ischaemic stroke showed this higher dose led to higher rates of intracerebral haemorrhage than alteplase (NOR-TEST 2, part A; (Kvistad et al, 2022)), and this dose is no longer recommended. Tenecteplase (0.25 mg/kg) delivered in an MSU setting (TASTE-A; (Bivard et al, 2022)) led to better measures of imaging reperfusion than alteplase but the study was inadequately powered to test any difference in outcomes. Two large randomised trials have demonstrated that tenecteplase 0.25 mg/kg is non-inferior to alteplase for excellent clinical outcome when delivered within 4.5 hours of stroke onset (Menon et al, 2022; Wang et al, 2023). In patients with proven large artery occlusion prior to planned thrombectomy tenecteplase (0.25 mg/kg) may be superior to alteplase when given within 4.5 hours of onset (Campbell et al, 2018b). [2023]

Since the 2012 edition of the guideline, five RCTs have been published evaluating the effects of endovascular treatment in addition to thrombolysis, compared with standard treatment (intravenous thrombolysis alone administered within 4.5 hours) in ‘early-presenting’ patients (typically within 6 hours) with proven large artery occlusion stroke (Berkhemer et al, 2015; Campbell et al, 2015; Goyal et al, 2015; Jovin et al, 2015; Saver et al, 2015). In an individual patient meta-analysis of these five trials involving 1,287 patients (Goyal et al, 2016) endovascular therapy showed significant improvements in functional outcomes at 90 days. The number needed to treat for one additional patient to have reduced disability of at least one point on the mRS was 3. The trials were heterogenous in their patient selection (age, NIHSS score) and only included patients with pre-stroke mRS of 2 or less. There was also variation in imaging criteria, in particular whether the identification of salvageable brain tissue on neuroimaging was a trial inclusion criterion (EXTEND-IA, ESCAPE, SWIFT-PRIME, and REVASCAT beyond 4.5 hours) or not (MR CLEAN). Three trials included some patients for whom intravenous thrombolysis was contraindicated. The trials varied in onset to endovascular treatment from a maximum of 6 up to 12 hours, and it is pertinent that all the trials with an extended time window required imaging identification of the potential to salvage brain tissue prior to randomisation (see below). The SWIFT-PRIME trial had the fastest process times with a median time from hospital arrival to groin puncture of 90 minutes, and the median procedure time in the five trials was just under 60 minutes. An NIHSS score of 6 or more was an inclusion criterion for several trials with clear positive subgroup effects for NIHSS 6-19 (ESCAPE) and 6-17 (SWIFT-PRIME). Not all trials reported a positive effect on mortality. The Working Party concludes that mechanical thrombectomy is an effective acute stroke treatment for selected patients with proximal large artery occlusions as an adjunct to intravenous thrombolysis, and for those patients with contraindications to intravenous thrombolysis but not to mechanical thrombectomy (e.g. recent surgery, anticoagulant use). Centres that provide endovascular treatment should meet the professional standards set out by the joint societies’ working group (White et al, 2015; NICE, 2016a). There remain significant challenges to the full implementation of this treatment in the UK and Ireland. [2023]

Two RCTs have reported on patients with a large ischaemic core who have previously been ineligible for trials of thrombectomy beyond 6 hours. SELECT2 (N=352; Sarraj et al, 2023) demonstrated that patients aged 18-85 years with a pre-stroke mRS score of 0 or 1 presenting with a proximal large artery occlusion (ICA/M1) and an ASPECTS score of 3-5 or an infarct core of >50 mL benefitted from mechanical thrombectomy up to 24 hours after onset. Thrombectomy resulted in functional independence (mRS 0-2) in 20.3% of treated patients compared with 7% in the medical arm (NNT=8). ANGEL-ASPECT (N=456; Huo et al, 2023) also demonstrated significant benefit from mechanical thrombectomy in patients aged 18-80 years with a pre-stroke mRS score of 0 or 1 and an ICA/M1 occlusion and either ASPECTS 3-5 or an infarct core of 70-100 mL presenting within 24 hours of onset. Thrombectomy increased the proportion of patients with functional independence (mRS 0-2) to 30% compared with 11.6% in the medical arm (NNT=6). In both trials, perfusion imaging (almost all CT perfusion) was undertaken to confirm a large infarct core volume (median 80 mLs in SELECT2, 62 mLs in ANGEL-ASPECT) and benefits were observed despite a higher rate of intracranial haemorrhage with thrombectomy. It is noteworthy in these late-presenting groups that around two thirds of patients were still ineligible for treatment, and after thrombectomy the median mRS at 90 days was 4 and mortality ranged from 22-38%. Considering this evidence in conjunction with the results of DEFUSE-3 (Albers et al, 2018), in patients presenting between 12 and 24 hours after onset, there is robust evidence to select individuals for mechanical thrombectomy based on perfusion imaging (particularly perfusion mismatch > 15 mL). In patients presenting between 6 and 24 hours after onset with a pre-stroke modified Rankin score of 2 or more, or who are older than 85 years, there is still insufficient evidence, and the more selective DAWN/DEFUSE-3 radiological criteria should be considered as the primary means of selection for these groups. As yet there are no RCT data from patients with anterior circulation stroke presenting between 12 and24 hours with non-proximal M1/intracranial ICA occlusion. It should also be noted that the perfusion criteria applied in these trials of late-presenting patients (core defined by rCBF below 30% and penumbra by Tmax greater than 6 secs) were mostly based on the use of RAPID™ AI decision-support software from IschemaView (Stanford, USA) and direct extrapolation of these results to other AI systems should not be assumed as appropriate or equivalent to the referenced trials. [2023]

A recent meta-analysis (Turc et al, 2022b) of six RCTs (Yang et al, 2020; LeCouffe et al, 2021; Suzuki et al, 2021; Zi et al, 2021; Fischer et al, 2022; Mitchell et al, 2022) supports the administration of thrombolysis within 4.5 hours of onset in eligible patients prior to thrombectomy (bridging thrombolysis) given that no trial showed superiority and only two of the six trials, both judged at high or moderate risk of bias, showed non-inferiority for proceeding direct to thrombectomy (Yang et al, 2020; Zi et al, 2021). Furthermore, the meta-analysis indicated superior reperfusion rates, trends to improved clinical outcome and no statistical increase in adverse safety outcomes (mortality and symptomatic intracranial haemorrhage) with bridging thrombolysis. All the randomised trials recruited patients presenting directly to thrombectomy centres. No randomised trial has yet addressed the question of whether patients presenting initially to an acute stroke centre should proceed directly to mechanical thrombectomy without bridging thrombolysis. However, a meta-analysis of observational studies found better clinical outcomes for the bridging thrombolysis group, although direct mechanical thrombectomy was deemed safe (Turc et al, 2022b). There is now sufficient evidence to guide the selection of patients for both thrombolysis and thrombectomy presenting later than 4.5 hours after symptom onset or where the onset time is not known (Albers et al, 2018; Nogueira et al, 2018; Thomalla et al, 2018; Campbell et al, 2019; Albers et al, 2021; Jovin et al, 2022b; Tao et al, 2022). [2023]

Two RCTs (BASICS and BEST) were published in 2020 and 2021 addressing whether thrombectomy and best medical therapy was superior to best medical therapy alone in imaging-confirmed basilar artery occlusion (BAO)/vertebral artery (VA) occlusion (Liu et al, 2020; Langezaal et al, 2021). Neither trial had an NIHSS restriction on eligibility or systematic imaging exclusion criteria other than extensive bilateral brainstem ischaemia. Both trials were neutral on an intention-to-treat analysis for their primary endpoint of mRS of 0-3 at 90 days. Two further Chinese RCTs of the effectiveness of thrombectomy in BAO have been published (Jovin et al, 2022a; Tao et al, 2022). The ATTENTION trial randomised 340 patients in a 2:1 ratio to either thrombectomy and best medical therapy or best medical therapy alone, with additional eligibility criteria of an NIHSS of 10 or more, PC-ASPECTS of 6 or more and in a time window of up to 12 hours after stroke onset (Tao et al, 2022). Patients over 80 years of age additionally had to have PC-ASPECTS of 8 or more and a pre-stroke mRS of 0-1. The trial demonstrated superiority of thrombectomy with an absolute difference in mRS 0-3 of 23.2%. The BAOCHE trial randomised 218 patients aged 80 or younger between 6 and 24 hours after stroke onset in a 1:1 ratio if the NIHSS was 6 or more, PC-ASPECTS 6 or more and the Pons-Midbrain Index was 2 or more (Jovin et al, 2022a). It enrolled 82 patients (38%) in the 12-24 hour window and also showed superiority for thrombectomy with an absolute difference in mRS 0-3 of 22.1%. All three Chinese trials had much lower intravenous thrombolysis rates than was seen in BASICS. Despite the results from 4 RCTs, there are very limited data for patients over the age of 80 years, those with a baseline NIHSS of 6-9 and those presenting beyond 12 hours. [2023]

A systematic review and meta-analysis of five small trials randomising 498 patients either to general anaesthesia or heavy conscious sedation concluded that general anaesthesia resulted in superior functional independent outcomes, with only one of the five small RCTs showing statistical superiority for general anaesthesia over conscious sedation (Bai et al, 2021). A subsequent meta-analysis including non-randomised data from eight studies including 7,797 patients demonstrated that local anaesthesia without sedation was not significantly superior to either conscious sedation or general anaesthesia in improving outcomes (Butt et al, 2021). A subsequent larger multi-centre trial (GASS; Maurice et al, 2022) which included standardised approaches to blood pressure management did not demonstrate differences in functional outcome with anaesthetic method, despite a higher recanalisation rate with general anaesthesia. These data contrast with larger volume real-world registries or individual patient data level RCT post-hoc analyses, which tend to favour non-general anaesthesia over general anaesthesia in terms of functional outcomes (Campbell et al, 2018a; Powers et al, 2019). There is a paucity of randomised evidence comparing general anaesthesia with local anaesthesia or minimal conscious sedation approaches, but trials are ongoing. [2023]

The DESTINY II trial of decompressive hemicraniectomy for older patients with severe space-occupying MCA territory infarction has shown a substantial survival benefit for patients over the age of 60 years (Jüttler et al, 2014) akin to that seen in young patients (Cruz-Flores et al, 2012). Decisions to undertake major life-saving surgery need to be carefully considered on an individual basis, but patients should not be excluded from treatment by age alone. [2016]

These recommendations underpin the earlier recommendations concerning the organisation of acute stroke care (Section 2.2 Definitions of specialist stroke services, Section 2.3 Transfer to acute stroke services, Section 2.4 Organisation of inpatient services), with significant implications for the organisation of acute stroke services and referrals to tertiary neurosurgical and interventional neuroradiology services. Provision of hyperacute stroke care should be organised to minimise time to treatment for the maximum number of people with stroke, and in some areas this will require reconfiguration of hyperacute stroke services with some hospitals stopping providing acute stroke services altogether. [2023]

A global shortage of tenecteplase will limit the initial extent to which Recommendation 3.5A can be implemented, at least until the end of 2024. Individual nations and clinicians will need to consider these supply implications when planning a managed transition from alteplase to tenecteplase – for example, in England, the NHS expects a planned transition (licence permitting) from April 2025. [2023]

The effect of reducing blood pressure in the first few hours after the onset of ICH has been tested in two international RCTs. In INTERACT-2, the rapid lowering of systolic blood pressure (SBP) to a target below 140 mmHg within 1 hour in 2839 patients presenting within 6 hours of onset with mainly small, deep (thalamic or basal ganglia) ICH did not reduce haematoma expansion or improve the primary outcome of death or major disability (mRS 3-6), but secondary outcomes (ordinal shift analysis of the mRS and health-related quality of life measures) were improved. Death (12% at 3 months) and institutionalisation (9%) were not affected by intensive treatment (Anderson et al, 2013). In ATACH-2, pursuing a lower SBP target of 110-139 mmHg within 2 hours in 1,000 patients presenting within 4.5 hours of onset with small, deep ICH tended to reduce haematoma expansion but conferred no reduction in severe disability or death (mRS 4-6) compared to a target of 140-179 mmHg (Qureshi et al, 2016). Mortality (6.7%) was also not affected. More intensive SBP lowering was associated with more renal adverse events in the first 7 days. In comparing these two trials, it is noteworthy that a lower SBP was achieved in the acute phase in the standard treatment arm of ATACH-2 (all of whom received intravenous nicardipine) than in the intensive treatment arm of INTERACT2 (141 mmHg vs. 150 mmHg respectively). As the intensive arm of ATACH-2 led to a still greater SBP reduction, ATACH-2 can be interpreted as showing no additional benefit from a more aggressive blood pressure target than that tested in INTERACT2. More research is still needed to clarify the effect on clinical outcomes from hyperacute SBP reductions in people with acute ICH, including in large and/or lobar haemorrhages. [2016]

A systematic review and individual participant meta-analysis summarised the evidence regarding a much broader range of blood pressure-lowering strategies during the first 7 days after ICH (Moullaali et al, 2022). Active/intensive blood pressure-lowering interventions had no overall effect on mRS at the end of follow-up compared with placebo/standard (guideline-based) treatment but significantly reduced haematoma expansion. Subgroup analyses suggested that an intensive target should be preferred to a fixed dose of a specific agent and that other antihypertensive agents should be preferred over renin-angiotensin system blockers, though all data on renin-angiotensin system blockers in the meta-analysis came from a single trial of candesartan (Sandset et al, 2011). Post-hoc secondary analyses of INTERACT2 and ATACH-2 used as observational cohorts (i.e. no longer providing randomised data on the effects of the intervention) found that achieving early and stable target blood pressure over the first 24 hours was associated with better outcomes (Moullaali et al, 2019). Although there is sufficient evidence to support the safety of intensive blood pressure lowering as delivered in the INTERACT2 trial, further evidence is required to establish the optimal strategy for intensive blood pressure lowering. There is limited RCT evidence regarding intensive blood pressure lowering in patients after ICH with baseline systolic blood pressure greater than 220 mmHg or beyond 6 hours from onset, so no specific recommendations can be made regarding these patient groups although more careful acute blood pressure lowering may need to be considered. [2023]

Abnormalities of clotting, especially in patients taking anticoagulants, should be urgently reversed, using 4-factor prothrombin complex concentrate (PCC) to reverse vitamin K antagonists. A trial of idarucizumab in patients taking the direct thrombin inhibitor dabigatran has shown the agent to be safe, rapid in action and effective in reversing the anticoagulant effect (Pollack et al, 2017). Multiple low to medium quality, non-randomised, observational studies suggest that 4-factor PCC demonstrates haemostatic efficacy in ICH in patients taking factor Xa inhibitors (Jaspers et al, 2021), but there is insufficient evidence to support a significant benefit in terms of haematoma expansion, mortality, or functional outcome. Andexanet alfa has been shown in normal volunteers to reverse the anticoagulant effect of the factor Xa inhibitors apixaban and rivaroxaban (Siegal et al, 2015) and this has been replicated in patients with ICH (Demchuk et al, 2021). However, the lack of comparative trials examining the effects of andexanet alfa in improving clinical outcome needs to be addressed to inform treatment of this patient group (NICE, 2021d). [2023]

In contrast to the long-standing and clear role for neurosurgical intervention in posterior fossa haemorrhage, and following a neutral neurosurgical trial in lobar haemorrhage without intraventricular haemorrhage (Mendelow et al, 2013), the role of neurosurgery for supratentorial ICH remains unclear. Most patients with ICH do not require surgical intervention and should receive monitoring and initial medical treatment on an acute stroke unit. Such patients are those with small, deep haemorrhage; lobar haemorrhage without hydrocephalus, intraventricular haemorrhage or neurological deterioration; large haemorrhage and significant co-morbidities before the stroke; and those with supratentorial haemorrhage with a Glasgow Coma Scale score below 8 unless this is because of hydrocephalus. [2016]

The majority of spontaneous ICH is caused by cerebral small vessel disease, including arteriolosclerosis (also termed deep perforator arteriopathy or hypertensive arteriopathy) and cerebral amyloid angiopathy (CAA). In a minority of patients, bleeding is caused by a macrovascular abnormality such as an arteriovenous malformation, dural arteriovenous fistula, intracranial aneurysm or cavernous malformation or by cerebral venous thrombosis (CVT). Detection of these causes may be worthwhile because of the risk of ICH recurrence and progression, if preventative interventions are in the patient’s best interests. Intra-arterial digital subtraction angiography is the reference standard investigation for detection of all macrovascular abnormalities other than cavernous malformations, which require MRI for diagnosis; the angiography procedure is associated with a very small (0.5-1%) but important risk of serious vascular and neurological complications. A 2015 Cochrane review found that studies of non-invasive angiography (CTA or MRA) of the entire cerebral vasculature (i.e. not limited to the Circle of Willis) had good diagnostic accuracy in comparison to acute intra-arterial angiography. However, studies were of varying quality, with partial verification bias and retrospective designs being common (Josephson et al, 2015). Subsequently, the use of CTA or MRI/MRA has been investigated in the high quality, prospective DIAGRAM study (van Asch et al, 2015). In a selected cohort (aged 18-70 years but excluding patients over 45 years of age with hypertension and ICH in the basal ganglia, thalamus or posterior fossa because of the low probability of finding an underlying macrovascular cause), DIAGRAM demonstrated early CTA (within 48 hours of onset) to have a positive predictive value of 72% (60 to 82%), detecting 51 of 69 (86%) macrovascular causes identified after a systematic investigation pathway in a total study sample of 298 patients. Subsequent MRI/MRA at 4-8 weeks in 214 patients with a negative CTA identified another two of 69 (3%) macrovascular causes. Delayed intra-arterial angiography then identified the remaining 15 of 69 (22%) macrovascular causes in 97 patients where the CTA and MRI/MRA were normal or inconclusive. [2023]

For patients not meeting the DIAGRAM study criteria, clinicians will need to estimate the probability of an underlying macrovascular cause to decide on whether to proceed to CTA, MRI/MRA, and/or intra-arterial angiography. The DIAGRAM score (which includes as predictors younger age, lobar or posterior fossa ICH location and absence of cerebral small vessel disease markers) has been derived from the original study cohort and externally validated, with moderate performance for detection of a macrovascular cause in the validation cohort (Hilkens et al, 2018). Including the CTA result in the DIAGRAM score resulted in good performance in the external cohort. The secondary ICH score uses age, sex, hypertension and imaging features and has good performance in validation cohorts. High probability imaging features included enlarged vessels or calcifications along margins of ICH suggesting an AVM or hyperattenuation within dural venous sinus or cortical vein along the path of drainage of the ICH suggesting cortical vein or venous sinus thrombosis. Low probability features were basal ganglia, thalamus or brainstem ICH location (van Asch et al, 2013). Where CTA/MRI/MRA is normal or inconclusive, it is important to consider a subsequent intra-arterial angiogram so as not to miss a macrovascular cause amenable to treatment. [2023]

Poor therapeutic outcomes affecting the management of patients with acute intracerebral haemorrhage have been partially reversed by the findings of INTERACT2.  The resulting potential benefit of care associated with blood pressure reduction, in addition to the safety of blood pressure lowering, was the rationale for the strength of the Guideline Development Group’s recommendation being ‘should be considered’ despite the recent individual patient data meta-analysis showing no effect of blood pressure lowering on functional outcome.  All hyperacute stroke units should consider developing processes to quickly identify patients presenting with intracerebral haemorrhage within 6 hours of onset and consider rapidly and safely lowering SBP for the majority of these patients. As in Section 2.3 Transfer to acute stroke services, referral protocols will need to be developed or refined to specify the role of neurosurgery in ICH.  A systematic, multidisciplinary, evidence-based approach to investigation for a macrovascular cause of ICH will aid timely and equitable detection of underlying causes amenable to treatment aimed at reducing the risk of recurrence.  In implementing these recommendations, efficient resource use should be considered, such as undertaking CTA immediately after diagnostic non-contrast CT. [2023]

There have been a number of negative trials in SAH management. Statins (Liu & Chen, 2015), magnesium (Mees et al, 2012) and endothelin receptor antagonists (Guo et al, 2012) have all been shown not to improve clinical outcome after SAH. [2016]

The CADISS RCT in people with symptomatic carotid and vertebral artery dissection showed no significant difference between anticoagulant and antiplatelet treatment in the prevention of recurrent stroke or death (CADISS Trial Investigators, 2015). The incidence of either outcome was low, with a 2% stroke rate within 3 months and no deaths. This low rate may reflect greater diagnostic yield in patients previously classified as ‘cryptogenic’. The TREAT-CAD RCT did not show that aspirin was non-inferior to anticoagulation (Vitamin K antagonist) in the prevention of either new MRI lesions during follow-up, or clinical endpoints (acute ischaemic stroke, major intra- or extracranial haemorrhage, death) in a primary composite endpoint. The endpoint occurred in 21 (23%) patients in the aspirin group and in 12 (15%) in the VKA group (absolute difference 8% [95% CI, –4 to 21], non-inferiority p=0.55). There were no deaths in either group (TREAT-CAD Trial Investigators, 2021). [2023]

A meta-analysis combining the per-protocol results of CADISS and TREAT-CAD showed that at 3 months there was no significant difference between anticoagulant and antiplatelet treatment for the composite endpoint of ischaemic stroke or major haemorrhage (Debette et al, 2021). In patients randomised to anticoagulation, the odds of the composite endpoint was 0.35 (95% CI, 0.08-1.63), ischaemic stroke was 0.18 (95% CI, 0.03 to 1.10) and major bleeding was 3.28 (95% CI, 0.34 to 31.80). Overall, the two RCTs did not show a significant difference between the two treatment groups in the acute phase of symptomatic extracranial cervical artery dissection. [2023]

There is no evidence on which to base a recommendation regarding long-term antithrombotic treatment after cervical artery dissection as the intervention (anticoagulant or antiplatelet) in the two trials (CADISS and TREAT-CAD) was stopped at 3 months. [2023]

There is no evidence to suggest that thrombolysis carries any greater risk in patients with cervical artery dissection compared to stroke from other causes (Engelter et al, 2012). [2016]

Case series suggest that anticoagulation is the treatment of choice for CVT, even when haemorrhage is seen on brain imaging, with a reduction in death and dependency (Stam et al, 2002). A Cochrane review (Coutinho et al, 2011) identified two small trials of anticoagulation after CVT. Although not reaching statistical significance, there was a trend toward a positive benefit from anticoagulation for at least three months. DOACs are licensed for venous thromboembolism but not for CVT. There is no evidence to support the use of corticosteroids in the management of CVT; what information is available is likely to be affected by selection bias and does not support their use, and may even suggest some circumstances where their use may be harmful (Canhao et al, 2008). [2016]

Patients with acute stroke are at high risk of dehydration, malnutrition, infection, hypoxia and hyperglycaemia. Middleton et al (2011) showed that training stroke unit staff in the use of standardised protocols to manage physiological status can significantly improve outcomes. The management of blood pressure after acute ischaemic stroke remains an area with little evidence to guide practice (see Section 3.6 Management of intracerebral haemorrhage for the recommendation regarding blood pressure management in acute intracerebral haemorrhage). There is no evidence for the use of hyperbaric oxygen therapy in stroke (Bennett et al, 2014) nor for the use of supplemental oxygen in normoxic patients (Roffe et al, 2011) and from the evidence available, the Working Party recommends that mannitol for the treatment of cerebral oedema should not be used outside of a clinical trial. [2016]

There is very little trial evidence on which to base the management of hydration in acute stroke. A Cochrane review of the signs and symptoms of impending and current water-loss dehydration in older people (Hooper et al, 2015) concluded that there is little evidence that any one symptom, sign or test, including many that clinicians customarily rely on, have any diagnostic utility for dehydration. [2016]

There is good evidence that a multi-item dysphagia screening protocol that includes at least a water intake test of 10 teaspoons and a lingual motor test was more accurate than screening protocols with only a single item (Martino et al, 2014). There is good evidence from a systematic review (Kertscher et al, 2014) that the investigation of dysphagia with instrumental assessments providing direct imaging for evaluation of swallowing physiology help to predict outcomes and improve treatment planning. [2016]

In contrast to acute myocardial infarction, tight glycaemic control has not been shown to improve outcome in stroke (Gray et al, 2007) and studies have warned against aggressive lowering with insulin infusions due to the risk of hypoglycaemia. This has led the Working Party to recommend a broadening of the target range for blood glucose in acute stroke from 4-11 mmol/L to 5-15 mmol/L. [2016]

Two recent studies showed no clinical benefit from the prophylactic use of antibiotics in dysphagic stroke patients and thus routine antibiotic prophylaxis is not recommended (Kalra et al, 2015; Westendorp et al, 2015). [2016]

One systematic review (Olavarria et al, 2014) examined four small non-randomised trials of head position in acute patients with ischaemic stroke. The trials studied cerebral blood flow using transcranial Doppler but did not report on functional outcome. In the international multicentre cluster-randomised trial HeadPoST (Anderson et al, 2017) of 11,093 patients hospitalised with acute stroke, there was no significant difference between the lying-flat head position and the sitting-up position for the first 24 hours with respect to the primary outcome of disability at 90 days. There were also no significant differences in mortality or in the rates of serious adverse events, including pneumonia. [2023]

Recommendations have been changed as a result of an international RCT of over 2000 people with acute stroke (AVERT Trial Collaboration group, 2015). Although two small RCTs previously showed that very early mobilisation (beginning within 24 hours) was feasible in an acute setting, the AVERT trial showed that very early, more frequent, higher dose mobilisation focused on out-of-bed activities in addition to usual care was worse than usual care alone. Very early mobilisation led to greater disability at three months with no effect on immobility-related complications or walking recovery. The trial included people with previous stroke, severe stroke, intracerebral haemorrhage and those who were thrombolysed, if they required help to mobilise and were expected to remain in hospital for at least three days. It excluded those who were medically unstable or with significant previous disability. [2016]

To implement this evidence into practice it is important to understand the nature of the usual care and the other factors within this complex intervention. In AVERT’s very early intervention, 92% were mobilised within 24 hours of stroke onset (as opposed to admission) and 23% were mobilised within 12 hours. This was carried out by nurses or therapists an average of six times per day, and included an average daily amount of 31 minutes of mobilisation by a physiotherapist measured over 14 days or until transfer of care if earlier. Given the trial outcomes, such very early mobilisation cannot be recommended. [2016]

The more beneficial usual care was still early but slightly later, less frequent and at a lower dose. Almost everyone (93%) was mobilised within 48 hours of onset, 59% within 24 hours and 14% within 12 hours, by nurses or therapists an average of three times per day, and including an average daily amount of 10 minutes of mobilisation by a physiotherapist. A subsequent exploration of dose hypothesised that early mobilisation might be best delivered in short, frequent amounts (Bernhardt et al, 2016) but this requires further research. [2016]

The risk of symptomatic intracerebral haemorrhage outweighs the benefit from the prevention of venous thromboembolism (VTE) with routine anticoagulation with low-dose heparin (including low molecular weight heparin) following acute ischaemic stroke (Geeganage et al, 2013). It is also not possible to predict which patients with acute stroke may be at sufficiently high risk of VTE compared to haemorrhagic complications to inform the targeted use of heparin treatment in selected patients (Whiteley et al, 2013). The CLOTS 1 and 2 trials showed that graduated compression stockings were ineffective in preventing VTE or improving functional outcome in stroke (CLOTS Trials Collaboration, 2013). The CLOTS 3 trial showed that intermittent pneumatic compression (IPC) using sequential compression with venous refill technology in immobile patients in the first 30 days after stroke is an effective treatment for reducing proximal DVT and improves survival but not functional outcomes (CLOTS Trials Collaboration, 2014). In evaluating the cost-effectiveness of IPC in stroke, NICE recommended that healthcare professionals explain to the patient or their family members or carers that IPC reduces the risk of DVT and may provide an increase in survival, but it will not help them recover from their stroke, and there may be an associated increased risk of surviving with severe disability (NICE, 2015b). [2016]

If proximal DVT does occur in a patient with ischaemic stroke, the risk of PE is high and such patients should receive treatment-dose anticoagulation. If DVT occurs in a patient with ICH there are no randomised trial data to support any particular treatment, but single-centre case series have reported that in such cases a vena caval filter is probably safe and effective for the prevention of PE (Somarouthu et al, 2011). There is no evidence to guide the management of patients with ICH and PE, and the decision to use or to avoid the use of anticoagulant treatment can only be made on the physician’s individualised assessment of the balance of risk and benefit. [2016]