Long-term management and secondary prevention

The identification of risk factors for stroke and TIA should be part of the assessment during the acute phase.  Regular review of risk factors and secondary prevention in primary care may require additional resources. [2016]

The principal evidence for carotid endarterectomy for people with recent symptoms is from the European Carotid Surgery Trial (ECST) and the North American Symptomatic Carotid Endarterectomy Trial (NASCET) (Rothwell et al, 2003b). Only people with non-disabling stroke or TIA were included in these trials and the benefits of surgery cannot be assumed to apply to those with more disabling strokes. People with possible cardioembolism were also excluded. When allowance is made for the different methods used to measure stenosis from angiograms, the two trials report consistent findings. To avoid confusion regarding the degree of stenosis the technique used in NASCET should be used (the ratio of the diameter of the residual lumen at the point of maximum narrowing to that of the more distal internal carotid artery, expressed as a percentage). In a pooled analysis of the individual data from 6,092 patients, carotid endarterectomy reduced the 5-year absolute risk of ipsilateral ischaemic stroke by 16.0% in patients with 70–99% stenosis, and by 4.6% in patients with 50–69%. There was no benefit for patients with 30–49% stenosis and surgery increased the risk in patients with less than 30% stenosis. There was no evidence of benefit for patients with a near-occlusion. In these trials conducted in the 1980s the operative risk of stroke (ocular or cerebral) and death within 30 days of endarterectomy was 7%. [2016]

There is evidence of considerable heterogeneity in individual risk according to age, gender, degree of stenosis, presenting symptom, time from presenting symptom and presence of plaque ulceration (Rothwell et al, 2004). Prognostic models based on these characteristics have been derived which may be useful in the decision making process (Rothwell et al, 2005). These models are based on trial data which are now over 20 years old and with improvements in other treatments these models are likely to overestimate the absolute risk of stroke. Modified prognostic models incorporating corrections to allow for improvements in ‘best medical therapy’ have been developed (e.g. the Carotid Artery Risk score – www.ecst2.com/), but await validation. [2016]

In a systematic review of operative risks in relation to timing of surgery, no statistically significant difference for early versus late surgery was identified for patients with stable stroke (Rerkasem & Rothwell, 2009). In patients undergoing emergency surgery the pooled absolute risk of stroke and death was 20.2% for those with ‘stroke-in-evolution’ (fluctuating or progressive deficit) and 11.4% for those with crescendo TIA (more than 2 episodes in a week), significantly higher than for those undergoing non-emergency surgery (odds ratio [OR] 4.6). Such patients are likely to be at increased risk if surgery is not performed, but given these risks and the effectiveness of medical management it cannot be assumed that emergency surgery is beneficial in neurologically unstable patients. The outcome from carotid endarterectomy is not significantly influenced by whether the procedure is carried out under local or general anaesthesia (Vaniyapong et al, 2013), and if the person has a particular preference, this should be taken into account. [2016]

Compared to surgical endarterectomy, endovascular therapy involving carotid angioplasty and stenting is associated with an increased risk of stroke of any severity or death (Bonati et al, 2012). This increased risk is modified by age, with no difference in stroke or death when the comparison is confined to those below 70 years of age (International Carotid Stenting Study investigators, 2010). Long-term follow-up identifies an excess of procedure-related and non-disabling strokes with endovascular therapy (Bonati et al, 2015). By contrast, carotid endarterectomy is associated with an excess of cranial nerve palsy and myocardial infarction (Bonati et al, 2012). For endovascular procedures undertaken within the first few days after symptom onset there is an excess of disabling and fatal, as well as non-disabling strokes in comparison to carotid endarterectomy (Rantner et al, 2013). [2016]

There is no high quality evidence to guide decision making regarding the timing and indications for carotid revascularisation in patients presenting with ischaemic stroke who have been treated with intravenous thrombolysis. A number of case series have been reported with small numbers and few outcome events (Naylor, 2015). Activation of the coagulation system and fibrin formation occurs following alteplase therapy with changes peaking at 1 to 3 hours but detectable for up to 72 hours (Fassbender et al, 1999). It is not clear what impact if any these changes in the coagulation system may have on the balance of risks and benefits, but in the absence of high quality data it would seem reasonable to advise caution if considering surgery within 72 hours of intravenous thrombolysis. [2016]

Vascular surgery services should offer the option to perform carotid endarterectomy surgery under local or general anaesthetic.  Multidisciplinary teams should include a carotid interventionist able to advise on and deliver carotid artery angioplasty and stenting. [2016]

There is high quality evidence that BP reduction after stroke or TIA prevents further vascular events including recurrent ischaemic and haemorrhagic stroke (Progress Collaborative Group, 2001). In PROGRESS, giving two more BP-lowering medicines to people after stroke or TIA, 52% of whom were classified as normotensive, reduced BP by 12/5 mmHg and resulted in a 42% reduction in recurrent stroke and 35% fewer major coronary events. A net benefit was seen for those with baseline BP levels as low as 115/75 with the lowest risk of recurrent stroke seen in those achieving the lowest follow-up BP levels. There was no evidence of increased stroke risk at lower BP (Arima et al, 2006). A meta-analysis of 123 studies and 613,815 participants found BP-lowering treatment significantly reduced cardiovascular events and death in proportion to the magnitude of BP reduction achieved, with no differences in the proportional benefits between trials with lower (below 130 mmHg) or higher systolic BP at baseline (Ettehad et al, 2016). There was also no difference in the proportional risk reductions for major cardiovascular events by baseline medical conditions, but calcium-channel blockers were found to be superior to other medication classes in the prevention of stroke. Overall, a 10 mmHg reduction in systolic BP reduced the risk of cardiovascular disease by 20% and stroke by 27%. [2016]

In the Secondary Prevention of Small Subcortical Strokes (SPS3) trial targeting a systolic BP of below 130 mmHg in patients with recent lacunar stroke, reductions in the rate of all stroke, disabling or fatal stroke and the composite outcome of myocardial infarction or vascular death were not significant, but the rate of cerebral haemorrhage was reduced and the lower target was well tolerated (SPS Study Group, 2013). In addition to the PROGRESS trial (2001), further evidence to support the benefit of a SBP target of below 130/80 mmHg in reducing recurrent stroke was observed in a clinical trial and as a secondary outcome in another study (Mant et al, 2016; Kitagawa et al, 2019). Accompanying meta-analyses (Thomopoulos et al, 2015, 2016), have shown that lowering BP to below 130/80 mmHg, if tolerated, is beneficial in preventing stroke and major cardiovascular events. Two analyses with greater statistical power that included the Systolic Blood pressure Intervention trial (SPRINT research group, 2015) have shown that intensive treatment even to below 120 mmHg may be beneficial (Bangalore et al, 2017; Bundy et al, 2017). However, in SPRINT, patients with a history of stroke were excluded and the trial was not powered to assess a difference in stroke incidence (a secondary outcome). A U-shaped BP-cardiovascular risk relationship was observed both in the intensive and standard treatment groups in SPRINT, but randomised comparisons found a significant cardiovascular and mortality benefit for more versus less intensive treatment at each level of baseline BP (Beddhu et al, 2018). [2023]

There is uncertainty regarding the best time to start antihypertensive therapy following ischaemic stroke. Whilst BP can be successfully and safely reduced in the acute phase, there is no evidence that such early intervention results in long-term benefits (Bath & Krishnan, 2014). For patients admitted with stroke who are already taking antihypertensive medication, treatment can be safely withheld until patients are medically and neurologically stable and have suitable oral or enteral access (Bath et al, 2015). Unless there is severe hypertension or treatment is required for acute intracerebral haemorrhage (Section 3.6 Management of intracerebral haemorrhage) or to facilitate intravenous thrombolysis treatment (Section 3.5 Management of ischaemic stroke) antihypertensive medication should generally be initiated prior to the transfer of care out of hospital or at 2 weeks, whichever is the soonest, or at the first clinic visit for people not admitted. [2016]

Home BP monitoring may usefully contribute to hypertension management, but there are few trials of home BP monitoring in stroke. Kerry et al (2013) conducted a randomised trial of home BP monitoring with nurse-led support versus standard care in people with hypertension and a history of stroke. More changes to antihypertensive treatment were made in the intervention group than controls (60% v. 47%; p=0.02), but the fall in systolic BP from baseline did not differ significantly between the two groups. A post-hoc analysis showed a reduction of 6 mm Hg in systolic BP at 12 months in the intervention group (Kerry et al, 2013). A prospective cohort study found that half of people with stroke offered a BP monitor without nurse support continued monthly use for up to 18 months, but around a quarter of participants reported a disability that made it difficult to measure their BP unaided (Ovaisi et al, 2011). [2023]

Studies suggest that treatment based on home BP monitoring may reduce vascular risk, principally through improving long term compliance (Ovaisi et al, 2011; Kerry et al, 2013; Hanley et al, 2015). Further evidence is required to determine which people with stroke benefit, and the type of equipment (semi-automated v. automated), timing of measurement, ideal target and duration. Systematic reviews emphasise that attention should be given to correct measurement and a clear management strategy for the patient (and carer as appropriate). It is also recommended that the pulse rhythm is recorded at each clinical contact to detect non-symptomatic atrial fibrillation (Cappuccio et al, 2004; Bray et al, 2010; Glynn et al, 2010; NICE, 2021e). [2023]

There should be a move away from the concept of treating hypertension and towards the concept of modifying BP as a risk factor.  It is appropriate to lower BP in patients who previously would have been considered normotensive. [2016]

The benefit of lipid-lowering therapy with statins in reducing cardiovascular events and mortality has been confirmed in RCTs and meta-analyses both for individuals with cardiovascular disease and specifically those with cerebrovascular disease. The Cholesterol Treatment Trialists’ Collaboration (Baigent et al, 2005) showed that for each 1.0 mmol/L reduction in LDL-cholesterol with statin therapy there was a relative risk reduction of 12% in all-cause mortality and a 21% reduction in major cardiovascular events, including a 17% reduction in fatal or non-fatal stroke (Baigent et al, 2005). The SPARCL trial investigated the effect of atorvastatin 80 mg daily in patients with TIA or stroke in the preceding 6 months and demonstrated a relative risk reduction of 15% in stroke and 35% in major coronary events with treatment (Amarenco et al, 2006). More recently, the Treat Stroke to Target Trial (TSTT) demonstrated that patients with ischaemic stroke or TIA in whom investigation confirmed evidence of atherosclerosis had a lower risk of subsequent cardiovascular events with a target LDL-cholesterol of below 1.8 mmol/L than those with a higher target of 2.3-2.8 mmol/L (Amarenco et al, 2020). ‘Evidence of atherosclerosis’ in TSTT included intracranial or extracranial artery stenosis, aortic atheroma or known coronary artery disease. A meta-analysis (11 RCTs of 20,163 people with stroke, including TSTT) compared intensive versus less intensive LDL-cholesterol lowering for secondary prevention (Lee et al, 2022). Intensive lowering over 4 years reduced the absolute risk of stroke (of any type) by 1.2%, and major cardiovascular events by 2.8%, but increased haemorrhagic stroke by 0.4%. Although small numerical increases were seen in both intracranial haemorrhage and newly diagnosed diabetes in the lower-target group in TSTT, these were not statistically significant and were more than outweighed by the benefits in prevention of major cardiovascular events. [2023]

In primary and secondary prevention studies, lowering LDL-cholesterol by 1.0 mmol/L leads to a relative risk reduction of 21% in major cardiovascular events, 9% in total mortality, and 15% in stroke (of any type) irrespective of baseline cholesterol and gender (Fulcher et al, 2015). Thus the decision to initiate treatment should be determined by a person’s absolute cardiovascular risk rather than their cholesterol level. NICE clinical guideline CG181 for lipid modification (2021d) provides recommendations consistent with the findings from SPARCL and TSTT, and recommends high-intensity statin therapy with atorvastatin 80 mg daily with a lower starting dose for people at high risk of adverse events or medication interactions, with subsequent guidance identifying the role of newer agents (NICE, 2016c, d; NHS England Accelerated Access Collaborative, 2021; NICE, 2021b, d, 2022c). The Guideline Development Group emphasizes the importance of characterising the finding of atherosclerosis in communications to primary care about long-term risk factor management so the correct treatment strategy can be followed. The recommendations also reiterate NICE and specialist society guidance on when to suspect familial hypercholesterolaemia, or lipoprotein abnormalities requiring specialist lipid clinic referral. [2023]

One randomised trial of icosapent ethyl in people with elevated triglycerides on optimal statin therapy (Bhatt et al, 2019) showed a significant reduction in major cardiovascular events that was greater than accounted for by the observed reduction in triglycerides (NICE, 2022c). For cardiovascular prevention after ischaemic stroke the emphasis remains on achieving LDL-cholesterol reduction to target, but people with raised triglycerides should be reviewed for lifestyle interventions such as alcohol reduction, weight reduction and diabetes control, and those with persistent severe hypertriglyceridaemia (greater than 10 mmol/L) should be referred to a lipid clinic to consider treatment to reduce the risk of pancreatitis. [2023]

The Antithrombotic Trialists’ Collaboration (2002) demonstrated a 22% reduction in the odds of a vascular event (myocardial infarction, stroke or vascular death) in patients with a previous stroke or TIA treated with antiplatelet medication. Comparative trials such as CAPRIE (CAPRIE Steering Committee, 1996), ESPRIT (ESPIRIT Study Group, 2006) and PRoFESS (Sacco et al, 2008) show that aspirin plus modified-release dipyridamole and clopidogrel monotherapy are equally effective, with both options superior to aspirin monotherapy. [2016]

The combination of aspirin and clopidogrel has been compared to clopidogrel monotherapy in patients with recent TIA or stroke (Diener et al, 2004). The combination was not superior to clopidogrel alone, with evidence of increased adverse effects particularly bleeding. There is evidence that even in short-term use the combination carries an increased risk of bleeding, particularly in aspirin-naive individuals (Geraghty et al, 2010). The use of combination antiplatelet treatment within the first few weeks after stroke or TIA is addressed in Section 3.3 Management of TIA and minor stroke – treatment and vascular prevention and Section 3.5 Management of ischaemic stroke. [2016]

RESTART was a UK multicentre open label pilot phase RCT (RESTART Collaboration, 2019) that included 537 participants in the UK with spontaneous (non-traumatic) intracerebral haemorrhage (ICH) who were taking an antithrombotic medication (antiplatelet or anticoagulant) for the prevention of occlusive vascular events. Participants were allocated at random to either start or avoid antiplatelet treatment, with a median time from ICH symptom onset to randomisation of 76 days (IQR 29-146). The primary outcome was recurrent ICH. Over a median follow-up period of two years, 4% of those allocated to antiplatelet therapy had recurrent ICH compared with 9% of those allocated to avoid antiplatelet therapy [adjusted hazard ratio 0.51, 95% CI 0.25 to 1.03; p=0.06]. There was no difference in major occlusive vascular events, but a reduction in all serious vascular events, as defined by the Antithrombotic Trialists’ Collaboration, was reported. The results were consistent across subgroups including ICH location, time since symptom onset, type of antiplatelet medication, participant age, and history of atrial fibrillation, although the Guideline Development Group noted the lack of statistical power to test for these interactions. [2023]

After extended follow-up to five years (Al-Shahi Salman et al, 2021), ICH affected 22 of 268 participants (8.2%) allocated to antiplatelet therapy compared with 25 of 268 participants (9.3%) allocated to avoid antiplatelet therapy (adjusted hazard ratio, 0.87; 95%CI 0.49 to 1.55; p=0.64). A major vascular event affected 72 participants (26.8%) allocated to antiplatelet therapy compared with 87 participants (32.5%) allocated to avoid antiplatelet therapy (hazard ratio, 0.79; 95%CI 0.58 to 1.08; p=0.14). A brain imaging substudy of the RESTART trial (n=254 with MRI; Al-Shahi Salman et al, 2019) found no evidence of more frequent recurrent ICH in patients with two or more cerebral microbleeds compared to one or no cerebral microbleeds (adjusted hazard ratio 0.30 [95% CI 0.08 to 1.13] v. 0.77 [0.13 to 4.61]; p-interaction=0.41), nor in subgroups according to cerebral microbleed burden or strictly lobar location (that might indicate cerebral amyloid angiopathy [CAA]). [2023]

The Guideline Development Group considered the RESTART trial to be well conducted and of high quality although the small sample size limited statistical power and precision. There were also concerns regarding possible selection bias (average ICH volume of 4 mL smaller than most ICH seen in clinical practice) which might limit generalisability. There were few participants with probable CAA, a subgroup which appears to be at highest risk of ICH (Charidimou et al, 2017). The group also noted that the evidence from RESTART cannot be applied to patients very soon after ICH symptom onset (only 21 participants (4%) were randomised in the first week after ICH while most (74%) were randomised at or beyond 30 days), although RESTART appears to give some reassurance that restarting antiplatelet therapy after ICH may be safe with regard to recurrent ICH, the consensus was that the evidence was insufficient to provide a strong recommendation. The group also identified some observational cohort studies, but due to methodological limitations (including risks of selection bias and confounding by indication) none was judged to provide sufficiently strong evidence to inform a recommendation about antiplatelet therapy after ICH. Further randomised trials are ongoing or planned and clinicians and patients are strongly encouraged to participate in these wherever possible. [2023]

In practice clinicians will continue to make case-by-case decisions based on their understanding of the baseline risk of future events including recurrent ICH and vaso-occlusive events. For example, in observational studies the recurrence risk for lobar ICH is consistently higher than for deep ICH (Li et al, 2021b) while the risk of ICH recurrence in patients with probable CAA is substantially higher than in those without probable CAA (Charidimou et al, 2017). Haemorrhagic MRI biomarkers of CAA, including cerebral microbleeds and cortical superficial siderosis, are also associated with an increased risk of ICH recurrence. Clinicians are recommended to take these prognostic factors into account when deciding whether to recommend antiplatelet medication after ICH. [2023]

Anticoagulant treatment is not more effective than antiplatelet therapy in people with non-cardioembolic ischaemic stroke or TIA and carries a greater risk of bleeding (Mohr et al, 2001; Sandercock et al, 2009). Neither is there evidence of greater efficacy for anticoagulation in embolic stroke of uncertain source (ESUS; Hart et al, 2018; Diener et al, 2019). A Cochrane review found no evidence that early initiation of anticoagulation (within 2 weeks) in unselected patients with (all cause) ischaemic stroke reduced death or disability at or beyond 1 month. There was moderate grade evidence of large reductions in recurrent ischaemic stroke and pulmonary embolism with moderate increased rates of intracranial and extracranial bleeding (Wang et al, 2021b). [2023]

Intracranial haemorrhage (including haemorrhagic transformation of the acute infarct) should be assessed by brain imaging and taken into account when deciding whether and when to commence anticoagulation. Less severe degrees of haemorrhagic transformation may not necessarily be a contraindication to anticoagulation, an issue that will be clarified by ongoing randomised trials. In the case of patients with moderate-large volume infarction such as occurs often with cardioembolic stroke, there is concern that anticoagulation may increase the risk of haemorrhagic transformation of the infarct, and a delay of an arbitrary 2-week period has been recommended in previous editions of this guideline. Observational data have suggested a lower rate of the composite outcome of recurrent stroke, bleeding or symptomatic ICH when anticoagulation is started between days 4 and 14 compared to within 4 days of AF-associated stroke (Paciaroni et al, 2015). Cohort studies suggest that commencing anticoagulation with a DOAC may be safe between days 4 and 14 in patients with small- or medium-sized infarcts, at least (Gioia et al, 2016; Labovitz et al, 2021). Pooled data analysis of over 2,500 patients in European and Japanese prospective cohort studies showed no increase in ICH when a DOAC was started earlier (within 5 days) rather than later in people with AF-associated stroke (De Marchis et al, 2022). Such observational studies are vulnerable to selection bias and confounding by indication, but they suggest that for patients with minor stroke (e.g. NIHSS 0-3) and a lower risk of haemorrhagic transformation it may be appropriate to commence treatment sooner, at the discretion of the treating clinician. Recent evidence from the TIMING randomised trial showed early initiation of anticoagulation with a DOAC in people with AF-related stroke was non-inferior to delayed initiation, with no ICH in either group (Oldgren et al, 2022), although the majority of strokes in both groups were mild (median NIHSS of 4 in both groups). Definitive guidance, particularly in relation to moderate-severe stroke, must await the findings from RCTs addressing the issue of early versus late initiation in people with AF-related stroke that have either recently completed (e.g. ELAN: NCT03148457) or are ongoing (e.g. OPTIMAS: NCT03759938), and a planned individual participant data meta-analysis. [2023]

There is strong evidence for the use of anticoagulation for long-term secondary prevention of stroke in people with permanent AF (Saxena & Koudstaal, 2004), and the 12% risk of recurrent stroke per year (EAFT Study Group, 1993) substantially alters the balance of risk and benefit in favour of anticoagulation in almost every instance. In people with relative contraindications to anticoagulation identified through the use of a validated tool (e.g. HAS-BLED; (Pisters et al, 2010) or MICON-ICH; Best et al., 2021) it may be possible to intervene to reduce the bleeding risk through control of blood pressure, medication review, treatment of other conditions and multidisciplinary input to reduce risk of falls and improve medication adherence. Falls are associated with higher risk of injury and bleeding but the risk is very unlikely to outweigh the benefits of anticoagulation for stroke prevention (Man-Son-Hing et al, 1999) The safety and benefit of DOACs over VKA is maintained in people with a history of falls in the available RCT evidence (Steffel et al, 2016; Rao et al, 2018). Single-centre data also suggest a reduced risk of traumatic brain injury in people with a history of falls on DOACs compared with VKA (Scotti et al. 2019). Other imaging biomarkers that may influence the balance of risk and benefit from anticoagulation in individual patients are considered in Section 5.21 Cerebral microbleeds. [2023]

DOACs are rapidly replacing VKAs for secondary stroke prevention in people with non-valvular AF. These medications have a rapid onset of action, have fewer interactions with other medications and foodstuffs, do not require coagulation monitoring and are more patient-friendly. Meta-analysis of the four primary DOAC trials, RE-LY (Connolly et al, 2009), ROCKET AF (Patel et al, 2011), ARISTOTLE (Granger et al, 2011) and ENGAGE AF-TIMI 48 (Giugliano et al, 2013) involving over 70,000 patients, has shown significantly greater stroke and thromboembolic prevention (RR 0.81, 95% CI 0.73 to 0.91; p<0.0001), with a substantially reduced risk of intracranial bleeding compared to warfarin (Ruff et al, 2014). No age interaction was observed for efficacy or safety with any DOAC with the exception of standard dose dabigatran and excess GI bleeding (Eikelboom et al, 2011; Graham et al, 2015). Subsequent meta-analyses with almost three million patient-years of observation have shown consistently better safety with DOACs compared to VKA (Hirschl & Kundi, 2019; Xu et al, 2021) and possibly better efficacy overall (Makam et al, 2018; Shen et al, 2020). [2023]

Given the high attributable risk of recurrent stroke in people with AF, unmodifiable relative contraindications (e.g. age, history of stroke) should not dissuade prescribers from the use of anticoagulation, as these patients are also at greatest of recurrent stroke (Olesen et al, 2011). Older people (aged 65 years or older) with AF have a reduced risk of stroke and thromboembolism on anticoagulant treatment compared to no treatment (relative risk 0.59, 95% CI 0.51 to 0.76, I2 = 12.3%), and with a DOAC rather than a VKA (Bai et al, 2018). If, despite addressing modifiable risk factors for bleeding, the bleeding risk is still considered to be too high to use an anticoagulant safely, then aspirin cannot be regarded as a safer alternative, particularly among older patients (Mant et al, 2007). A current NICE guideline does not recommend the routine use of aspirin in these circumstances aside from when there are other indications unrelated to AF (NICE, 2021c). However, a recent systematic review and meta-analysis of antiplatelet use in patients with AF not treated with anticoagulation (Benz et al, 2022) identified an increased risk of major bleeding and intracerebral haemorrhage, and a reduced risk of myocardial infarction compared to no treatment, suggesting that in selected patients with a low or normal risk of bleeding and a higher risk of vaso-occlusive disease (such as in secondary vascular prevention), antiplatelet treatment may still be appropriate. There may be an emerging role for very low dose edoxaban in achieving an acceptable balance of overall benefit and risk in such patients (Okumura et al, 2020). [2023]

Bearing in mind that participants in all the original comparative trials of DOACs with warfarin had to be eligible for both treatments, the trials do not provide evidence regarding the safety or efficacy of DOACs in people for whom the bleeding risk is considered to be too high to safely use warfarin. However, such patients were included in the AVERROES trial comparing apixaban with aspirin (Connolly et al, 2011), and very low dose edoxaban (15 mg OD) in a high-risk elderly Japanese population appears to be effective and relatively safe (Okumura et al, 2020). [2023]

For selected patients with AF who cannot be treated with anticoagulation, it may be appropriate to consider a left atrial appendage occlusion device if the short-term use of antiplatelets/anticoagulation required following the procedure can be tolerated. In the PROTECT AF trial percutaneous left atrial appendage occlusion with a filter device (Watchman) was non-inferior to warfarin for stroke prevention in people with non-valvular AF (Holmes et al, 2009; Reddy et al, 2013). Device implantation was accompanied by warfarin anticoagulation for the first 45 days. No trials have compared left atrial appendage occlusion with DOAC treatment. [2023]

For people with mechanical heart valves or rheumatic heart disease-associated AF, VKA remains the anticoagulant of choice as DOACs have been shown to be inferior in these situations (Eikelboom et al, 2013; Connolly et al, 2022). There is evidence that combining antiplatelet medication with warfarin reduces the risk of thromboembolic complications, but with an increased risk of bleeding (Little & Massel, 2003; Dentali et al, 2007). Apart from some high-risk patients with mechanical heart valves and patients in AF requiring antiplatelet therapy after coronary stenting, there is no evidence that combining antiplatelet medication with warfarin is beneficial, but there is clear evidence of harm (Hart et al, 2005). [2023]

This guideline is likely to lead to an increase in the prescribing of DOACs, which are expensive but considered cost-effective by NICE and SMC, particularly when used for secondary prevention where the attributable risk of stroke is several times higher than in primary prevention.  Management of patients with TIA or ischaemic stroke in association with AF requires an interdisciplinary team approach to stroke prevention with close collaboration between stroke physicians/neurologists, cardiologists, general practitioners, pharmacists, specialist nurses. [2023]

AF may not be detected by a standard 12-lead ECG and may require more prolonged monitoring. In a systematic review involving 5038 participants with recent stroke or TIA who had at least 12 hours of monitoring, the detection rate for new AF was 11.5% (Kishore et al, 2014). Rates of detection were higher in selected patients (e.g. by age, stroke pattern, pre-screening risk scores). In general, the more prolonged the period of ECG monitoring the greater the likelihood of detection (Grond et al, 2013; Higgins et al, 2013; Gladstone et al, 2014; Sanna et al, 2014). A sequential approach to investigation, involving four incrementally more prolonged phases of monitoring, provided detection rates of AF ranging from 7.7% to 16.9%; the overall AF detection rate was 23.7% after all phases (Sposato et al, 2015). [2023]

Implantable loop recorders (ILRs) have been shown to be superior to short-term external monitoring in detecting AF and could lead to better stroke reduction through the greater uptake of anticoagulation following a stroke or TIA. The use of ILRs seems superior to external monitoring strategies to detect AF in stroke of unknown aetiology or ‘cryptogenic’ stroke – 15.3 vs 4.7% at 12 months (Buck et al, 2021), with AF detected in 22.8% at 12 months with ILRs in another study (Noubiap et al, 2021). In a meta-analysis of three trials using ILRs compared to conventional monitoring, the 12-month detection rate for AF was 13% v. 2.4% and stroke or TIA occurred in 7% of participants with an ILR-based strategy v. 9% with usual care (odds ratio 0.8, 95% CI 0.5 to 1.2; Ko et al, 2022). ILR-monitored patients were also observed to have a reduced relative risk of stroke or TIA (0.49, 95% CI, 0.30 to0.81) in one meta-analysis (Tsivgoulis et al, 2019), but a subsequent larger meta-analysis of eight studies (five RCTs, three observational; 2,994 participants) by the same group showed that while patients with an ILR were more likely to receive anticoagulation, a reduction in stroke was not seen in the RCTs (Tsivgoulis et al, 2022). It therefore remains uncertain whether an ILR-based investigation strategy reduces stroke in an unselected population of patients after stroke or TIA. [2023]

The rate of detection of AF with ILRs at 12 months varies widely in reported studies (range 12-25%) and compared to conventional strategies it may be more cost-effective in patients with cryptogenic stroke with no AF detected after 24 hours of external monitoring (Edwards et al, 2020). Monitoring using smart digital technologies such as wearables and smartphone recording may be superior to traditional ‘holter’ monitoring and may be an alternative method where an ILR is not feasible or desired by patients (Koh et al, 2021). [2023]

In selecting patients for prolonged cardiac monitoring and/or ILR insertion those with stroke of undetermined aetiology (‘cryptogenic’) are more likely to have PAF (Kishore et al, 2014). Likewise, certain patterns of ischaemic change seen on brain imaging increase the likelihood of an underlying cardioembolic source such as cortical/subcortical infarcts or multiple lesions in anterior and posterior circulations and/or both cerebral hemispheres (Kang et al, 2003). [2016]

These recommendations are likely to increase the number of patients requiring prolonged cardiac rhythm monitoring and ILR insertion with implications for cardiac procedural and monitoring resources. Paroxysmal AF after stroke or TIA warrants treatment with anticoagulation or consideration for left atrial appendage occlusion where anticoagulation is contraindicated.  Close collaboration between stroke physicians/neurologists and cardiologists, through a cardiology-stroke multidisciplinary meeting, can facilitate decision making in challenging cases, and agreement on local protocols for ILR implantation and explantation, and monitoring and alert-response procedures. [2023]

In some people with stroke or TIA, a PFO will be an incidental finding and optimal medical therapy and secondary prevention strategies alone should be used. There is no evidence that anticoagulation is superior to antiplatelet therapy in patients with stroke of undetermined aetiology in association with a PFO (Homma et al, 2002). Three multicentre RCTs did not show a significant benefit of PFO closure over medical therapy alone, based on an intention-to-treat analysis (Furlan et al, 2012; Carroll et al, 2013; Meier et al, 2013). However, three subsequent RCTs (Mas et al, 2017; Søndergaard et al, 2017; Lee et al, 2018), and extended follow-up data from the RESPECT trial (Saver et al, 2017) showed a significant reduction in recurrent stroke risk in patients allocated to endovascular PFO closure compared with medical therapy alone, most of whom were older than 60 years of age. It is important to note that these more recent trials which showed a benefit of PFO closure over medical therapy alone only included patients with a PFO and a large right-to-left shunt or an atrial septal aneurysm (Mas et al, 2017) or a PFO with a right-to-left shunt (Søndergaard et al, 2017; Lee et al, 2018). [2023]

Most of the RCTs described above excluded patients over the age of 60 years, so it is not known whether the observed benefit from PFO closure is applicable to older patients with ischaemic stroke of otherwise undetermined aetiology.  Furthermore, there is insufficient evidence to address whether oral anticoagulation alone (either with a VKA or DOAC) is inferior, equivalent or superior to PFO closure, and more research is needed in this area. [2023]

PFO closure is associated with an approximately five-fold increase in the risk of developing AF (5-6.6% following closure), although the duration and nature of the resultant AF is not fully understood.  Implantation is also associated with other peri-procedural risks, including inadequate PFO closure, implantation failure, pericardial effusion, and pseudoaneurysm formation. [2023]

These recommendations are likely to increase the number of patients referred for further investigation and treatment, and stroke services should establish regular multidisciplinary meetings with their colleagues in cardiology to consider and appropriately select patients for further investigation and consideration of endovascular device closure.  Further research is needed to clarify optimal secondary preventative strategies in patients with a PFO and a co-existent thrombophilia as such patients were excluded from the definitive trials of PFO closure. [2023]

A systematic review and economic evaluation sought to evaluate the cost-effectiveness of routine echocardiography in the assessment of individuals presenting with first-ever ischaemic stroke or TIA (Holmes et al, 2014). Clinically identifiable cardiac pathologies were excluded. Across a range of cardiac pathologies, transthoracic echocardiography (TTE) was found to be less sensitive compared with transoesophageal (TOE), with both demonstrating high specificity. In consultation with an expert panel it was determined that only the identification of left atrial and left ventricular thrombus by echocardiography would alter patient management. A median prevalence of 0.8% was reported for left ventricular thrombus and of 1.4% for left atrial thrombus. Considerable variability was found for the reported prevalence of atrial septal aneurysm (median 9.3%, range 0.4-28%) and PFO (median 17%, range 0.25-73%). Economic analysis in the UK concluded that TTE is a cost-effective use of healthcare resources compared with TOE, when clinicians deem it the most appropriate test, and might be applied primarily to people with stroke of undetermined aetiology if they are also candidates for oral anticoagulation. Certain patterns of ischaemic change seen on brain imaging increase the suspicion of a cardioembolic source such as cortico-subcortical infarcts or multiple lesions in anterior and posterior circulations and/or both cerebral hemispheres (Kang et al, 2003). [2016]

The open randomised phase 2 study VAST compared stenting with medical management in patients with recently symptomatic vertebral artery stenosis of more than 50% (Compter et al, 2015). The study was terminated early, after enrolment of 115 patients. There were no significant differences between the two groups, and based on the low stroke recurrence rate seen in the trial, a conclusive phase 3 trial would need to include 9,500 patients. A subsequent pre-planned individual participant data meta-analysis of three randomised trials including 354 patients did not show a significant benefit from extracranial or intracranial vertebral artery stenting (Markus et al, 2019). There is therefore no evidence to suggest that revascularisation for people with vertebral artery stenosis (stenting, endarterectomy or reconstruction/transposition) is superior to best medical therapy. [2016]

The recurrent stroke rate in people with intracranial artery stenosis is high; in the WASID RCT comparing aspirin with warfarin in people with greater than 50% stenosis of intracranial arteries, those on aspirin had a 22% risk of stroke or death during a mean follow-up of 1.8 years (Chimowitz et al, 2005). This trial confirmed an association between increasing degree of intracranial stenosis and stroke risk and showed that the development of an effective collateral circulation is protective (Liebeskind et al, 2011). Warfarin anticoagulation was no more effective than aspirin for stroke prevention in WASID, including for the subgroup enrolled with stroke whilst on antithrombotic therapy, but was associated with significantly more adverse events. [2016]

The SAMMPRIS trial (Chimowitz et al, 2011) compared angioplasty and stenting of intracranial stenosis of greater than 70% with medical management, including an initial 90 days of dual antiplatelet therapy with clopidogrel plus aspirin. The 30-day rate of stroke or death in the SAMMPRIS control group was lower than in WASID (5.8% v. 10.7%) suggesting dual antiplatelet therapy may be superior to aspirin alone. Targeted risk factor modification, particularly BP and LDL-cholesterol reduction, was also more frequently achieved in SAMMPRIS, and the trial found that medical management was superior to angioplasty and stenting with the difference maintained over a median follow-up of 32 months (Derdeyn et al, 2014). The VISSIT trial exploring the effect of balloon-expandable stents in people with symptomatic intracranial stenosis was halted following the results of SAMMPRIS, with analysis also demonstrating an increased risk of recurrent stroke with endovascular intervention (Zaidat et al, 2015). No comparison of dual antiplatelet therapy with clopidogrel monotherapy in this setting has yet been conducted. [2016]

No studies have assessed how the COC modifies the risk of recurrent stroke or TIA. Studies in women with no history of stroke or TIA indicate that there may be an approximate doubling of the relative risk of ischaemic stroke associated with use of combined (low-dose) oestrogen oral contraception. A Cochrane review of one cohort study and 23 case-control studies compared the risk of myocardial infarction or ischaemic stroke in users and non-users of COC (Roach et al, 2015). The relative risk of ischaemic stroke was 1.7 for COC users and the risk increased according to the dose of oestrogen. The risk was not influenced by the progestogen used. It has been estimated that for 10,000 women using a 20 μg oestrogen COC for 1 year, 2 will have an arterial thrombosis (Lidegaard et al, 2012). Pregnancy is associated with a risk of stroke of about 3 per 10,000 deliveries (James et al, 2005). [2016]

A meta-analysis of six case-control studies comprising 3,091 cases and 11,385 controls found no association between progestogen-only contraceptive (POC) use and stroke risk (Chakhtoura et al, 2009). The analysis provides limited support for use of the POC in situations where hormonal contraception is necessary, but the full range of contraceptive methods (hormonal and non-hormonal) should be considered. [2016]

Treatment decisions concerning HRT must balance clinical need (treatment of premature menopause or relief of menopausal symptoms) against a number of different risks. A Cochrane review of 19 RCTs (40,410 participants) comparing hormonal therapy with placebo or no treatment found no benefit in all-cause mortality, cardiovascular death, non-fatal myocardial infarction, angina or revascularisation (Boardman et al, 2015). An increased risk of stroke was found in primary prevention studies, but the effect was not significant in secondary prevention studies (5,172 participants in five studies). For the subgroup of women starting HRT within a mean of 10 years after the menopause or who were less than 60 years of age, treatment was associated with a significant all cause mortality benefit and coronary heart disease benefit compared with placebo, although with a persisting risk of venous thromboembolism and a trend towards increased risk of stroke (9,838 participants in three studies). In a nested case-control study of 15,710 cases of stroke matched to 59,958 controls from the UK General Practice Research Database, the risk of stroke was not increased with use of low-dose oestrogen patches (alone or with progestogen) when compared with no use in post-menopausal women (Renoux et al, 2010). The stroke rate was increased with use of high-dose patches. [2016]

People with stroke and OSA have been shown to have worse functional outcomes, longer hospitalisation and an increased risk of stroke recurrence (Kaneko et al., 2003, Rola et al., 2008).  Treatment with continuous positive airways pressure (CPAP) has been shown to favourably modify cardiovascular risk factors such as hypertension (Marin et al., 2012) and in a prospective observational study to reduce the risk of recurrent cardiovascular events (Martinez-Garcia et al., 2012).  Several small RCTs have failed to confirm a reduction in cardiovascular events with CPAP (Parra et al., 2011, Ryan et al., 2011, Hsu et al., 2006, Parra et al., 2015, Sandberg et al., 2001, Bravata et al., 2011).  Whilst uncertainty remains concerning stroke recurrence, there are other benefits from recognising and treating OSA and given the high reported prevalence, presentation with stroke provides an opportunity to screen patients for OSA.  As in the general population, patients with stroke and OSA may not declare a classical history of severe snoring and subjective daytime sleepiness.  The use of a simple clinical screening tool (such as the ‘STOP-BANG’ questionnaire) in people with stroke or TIA will identify those who are likely to benefit from further specialist assessment (Silva et al., 2011). [2016]

There will be resource implications for sleep services from an increased awareness of OSA among people with stroke or TIA. [2016]

There is uncertainty concerning the most effective strategy to prevent arterial thrombotic events in people with APS. Recommendations include long-term low-dose aspirin, low-, medium- and high-intensity warfarin and the combination of aspirin and warfarin. There is little RCT evidence and those RCTs that are available have either not shown an advantage for any particular strategy (Crowther et al, 2003; Levine et al, 2004; Finazzi et al, 2005) or have included only very small numbers (Okuma et al, 2010). Until better evidence becomes available, the Working Party recommends that treatment decisions should be made on an individual basis, ideally involving multispecialty input. [2016]

The PROactive study (Dormandy et al, 2005) assessed secondary prevention with pioglitazone in people with diabetes and prior vascular disease and was neutral in terms of the primary outcome of major vascular events, but did show a reduction in a secondary outcome that included stroke. A subsequent Cochrane review (Liu & Wang, 2015) found that glitazones might reduce recurrent stroke in people with stroke or TIA. The IRIS trial (Kernan et al, 2016) compared pioglitazone with placebo in 3876 non-diabetic participants with insulin resistance and a history of stroke or TIA, excluding those with diabetes, structural heart disease or congestive cardiac failure. The primary outcome (first fatal/non-fatal stroke or non-fatal MI) occurred in 9% with pioglitazone and 11.8% with control over 4.8 years of follow-up. Progression to diabetes was also reduced, but weight gain, oedema and bone fractures were all significantly increased with pioglitazone. Based on these results, for every 100 patients treated with pioglitazone for about 5 years, 3 fewer would suffer stroke or MI; 4 fewer would develop diabetes mellitus; 2 more would suffer bone fracture requiring hospitalisation; 18 more would gain >4.5 kg in weight, and 11 more would have new or worsening peripheral oedema. The study did not report quality of life outcomes and more evidence would be required before glitazone treatment can be recommended routinely for patients with insulin resistance. Targeting lifestyle modification, particularly exercise and diet, appears to be a safe and effective approach for reducing insulin resistance and progression to diabetes (Knowler et al, 2002; Lindstrom et al, 2006; Ivey et al, 2007). [2016]

Early diagnosis allows timely screening for secondary complications, treatment to delay renal and cardiovascular effects, lifestyle advice particularly in relation to smoking cessation, and genetic counselling. The diagnosis should be considered in people with any of the clinical features above, and can be confirmed by tests measuring α-galactosidase activity and/or with molecular genetic testing for mutations of the GLA gene. Treatment with α-galactosidase A enzyme replacement therapy has been available for some years, but long-term effectiveness in preventing cerebrovascular complications has not so far been demonstrated (Rombach et al, 2013; Anderson et al, 2014; Germain et al, 2015). [2016]

There are few directly relevant randomised trials relating to secondary prevention of recurrent ICH, and none specifically in people with CAA. In the PROGRESS trial (2001) which included 6,105 people with ischaemic stroke or ICH, treatment with 2-4 mg perindopril for all participants (plus 2-2.5 mg indapamide for those with neither an indication for nor a contraindication to a diuretic) reduced blood pressure (BP) by an average of 12 mmHg systolic and 5 mmHg diastolic, lowering the risks of first and recurrent ICH (adjusted HR 0.44 [95% CI 0.28 to0.69] and 0.37 [95% CI 0.10 to 1.38] respectively) (Progress Collaborative Group, 2001). Patients with prior ICH derived the greatest benefit, and the risk of stroke decreased with lower follow-up BP without evidence of a lower threshold for hazard. In a PROGRESS subgroup analysis, intensive BP treatment reduced the relative risk of probable CAA-related ICH by 77% (95% CI 19 to 93%), that of probable hypertension-related ICH by 46% (95% CI 4 to 69%), and that of unclassified ICH by 43% (95% CI -5 to 69%) (Arima et al, 2010). An observational cohort study in 1,145 people after ICH found that higher BP during follow-up was associated with increased risk of lobar ICH recurrence (HR 1.33 per 10 mmHg increase [95% CI 1.02 to 1.76]) and that the risk of recurrent ICH increased with systolic BP above 120 mmHg and diastolic above 80 mmHg (Biffi et al, 2015). However, such observational studies are subject to selection and severity bias and confounding by indication. Further RCTs of intensive BP reduction after lobar ICH probably or possibly due to CAA are required. [2023]

The RESTART trial in people after ICH associated with antithrombotic medication use did not find evidence for hazard of antiplatelet treatment in the small subgroup of participants with lobar ICH probably due to CAA, defined by the Edinburgh CT-based or Boston MRI-based criteria (Rodrigues et al, 2018; Al-Shahi Salman et al, 2019), but estimates in this subgroup were imprecise due to the limited sample size. In an observational study, resumption of oral anticoagulation (OAC) after lobar ICH attributed to CAA was associated with decreased mortality and a more favourable functional outcome, but such observational studies are subject to selection and severity bias and confounding by indication (Biffi et al, 2017). Two pilot-phase randomised trials (APACHE-AF and SoSTART) of OAC for AF after ICH (Schreuder et al, 2021; SoSTART Collaboration, 2021) did not find evidence of benefit or harm from OAC, but their sample sizes were insufficient to investigate the effect of OAC in the subgroup of patients with lobar ICH probably due to CAA, so further RCTs are ongoing (e.g. ENRICH-AF NCT03950076 and PRESTIGE-AF NCT03996772). Left atrial appendage occlusion with short duration antithrombotic therapy might be an alternative to OAC in ICH associated with CAA, and this is being investigated in an RCT (e.g. STROKECLOSE NCT02830152). [2023]

There are no proven disease modifying therapies for CAA. In a small RCT the monoclonal antibody ponezumab did not show benefit on vascular reactivity in patients with probable CAA (Leurent et al, 2019). [2023]

There are no RCTs examining secondary prevention approaches in people with CADASIL. Observational studies have shown that both smoking and hypertension are associated with an increased risk of stroke in this group (Peters et al, 2006; Adib-Samii et al, 2010). For this reason most clinicians, and the European Academy of Neurology (EAN) guidelines (Mancuso et al, 2020) recommend tight control of cardiovascular risk factors, particularly BP reduction and smoking cessation. Evidence is not available on other risk factors including exercise, obesity, and cholesterol, but most clinicians also actively treat these risk factors. These should be treated from the time of diagnosis, rather than waiting until the person has had a stroke. [2023]

There is no research to guide clinicians on whether antiplatelet agents should be given to patients with CADASIL. Most practitioners, and the EAN guidelines (Mancuso et al, 2020) recommend avoiding antiplatelet agents prior to the onset of clinical stroke, but treating with a single antiplatelet agent (aspirin or clopidogrel) rather than dual antiplatelet agents after a stroke or TIA. The presence of cerebral microbleeds is associated with an increased risk of ischaemic stroke and should not prevent administration of antiplatelets after a stroke or TIA (Puy et al, 2017). [2023]

In patients prescribed oral anticoagulants after cardioembolic stroke or TIA in the CROMIS-2 (Wilson et al, 2018) and HERO (Martí-Fàbregas et al, 2019) prospective multicentre cohort studies the presence of cerebral microbleeds was independently associated with an approximate three-fold increase in the risk of symptomatic ICH, but the absolute risk of ischaemic stroke was consistently higher than that of symptomatic ICH independent of the presence of cerebral microbleeds. The Microbleeds International Collaborative Network [MICON]) included 20,322 patients taking any antithrombotic medication(s) (Wilson et al, 2019) and found that the absolute risk of ischaemic stroke exceeded that of ICH even in patients with 10 or more microbleeds (annual rates: ICH: 2.7%; ischaemic stroke: 6.4%), or microbleeds in a strictly lobar distribution suggesting CAA (annual rates: ICH: 1.3%; ischaemic stroke: 4.8%). The MICON-ICH collaborative group also demonstrated that a cerebral microbleed-based risk score (MICON-ICH) improves the prediction of intracranial haemorrhage during follow-up for patients with ischaemic stroke or TIA taking antiplatelets, anticoagulants, or both compared to widely used clinical prediction instruments like HAS-BLED, ATRIA, or ORBIT (Best et al, 2021). A high predicted risk of ICH might lead clinicians to pursue assertive management of modifiable bleeding risk factors such as BP and alcohol intake, and review concurrent medication. A secondary analysis of the NAVIGATE-ESUS trial (Shoamanesh et al, 2021) did not find evidence that microbleeds modify the net benefit or harm of anticoagulant versus antiplatelet therapy. However, in the PICASSO trial, which enrolled Asian participants with previous ischaemic stroke and previous intracranial bleeding (symptomatic or radiological intracerebral bleeding or at least two cerebral microbleeds), in those with microbleeds only at baseline, cilostazol was associated with a lower risk of intracranial bleeding than aspirin (Park et al, 2021). More data are needed regarding the prognostic significance of cerebral microbleeds in patients taking direct oral anticoagulants (DOACs); moreover, the balance of risk and benefit is uncertain for patients with cerebral microbleeds taking combined antiplatelet and anticoagulant therapy and for patients with a large number of cerebral microbleeds beyond five years. [2023]

There are Cochrane reviews (English & Hillier, 2010; Saunders et al, 2013), other systematic reviews (Ada et al, 2006; van de Port et al, 2007; Veerbeek et al, 2014b) and one high quality, moderate-sized RCT , other systematic reviews (Ada et al, 2006; van de Port et al, 2007; Veerbeek et al, 2014b) and one high quality, moderate-sized RCT (English et al, 2015) on physical activity after stroke. Overall, the evidence shows that activity programmes have a positive effect on global disability, albeit in the predominantly ambulant stroke population (Saunders et al, 2013). Treatment benefits physical function and supports the use of aerobic exercise and mixed training programmes to improve gait (English & Hillier, 2010; Marsden et al, 2013; Saunders et al, 2013; Kendall & Gothe, 2015). Other studies also suggest positive effects on outcomes such as vascular function (Moore et al, 2015) and psychosocial benefits (Faulkner et al, 2015). [2016]

Cardiorespiratory training, especially when involving walking, appears to be the most effective for cardiorespiratory fitness (with a moderate effect size) but also for walking and balance. Mixed training has a slightly lesser effect. Resistance training is most effective to improve muscle strength and endurance (Saunders et al, 2020). Thus the type of exercise prescribed depends on the patient’s own goals and preferences. However, cardio-respiratory training involving walking has the greatest overall benefit which can persist into the long-term (Saunders et al, 2020). [2023]

There have been a large number of Cochrane reviews assessing a variety of interventions to promote smoking cessation in the general population (for the individual reviews, see https://www.cochranelibrary.com/). A beneficial effect has been demonstrated for nicotine replacement therapy, nicotinic receptor partial agonists (varenicline, cytisine), antidepressant medication (bupropion, nortriptyline), combined pharmacotherapy and behavioural interventions, financial incentives, motivational interviewing, e-cigarettes, exercise, print-based self-help, telephone counselling and brief physician and nurse interventions. The evidence for interventions to increase smoking cessation in people with stroke is limited. A systematic review identified only four studies involving a total of 354 patients (Edjoc et al, 2012). Meta-analysis was not possible and a simple summed cessation rate of 24% for those receiving an intervention compared with 21% for controls was reported. [2016]

NICE public health guidelines (NICE guideline [NG209] Tobacco: preventing uptake, promoting quitting and treating dependence; NICE, 2023) provide guidance on smoking cessation services for all smokers. Stopping in one step is recommended as the approach must be likely to provide lasting success. Recommended interventions include behavioural counselling, group therapy, pharmacotherapy (licensed nicotine containing products, varenicline or bupropion) and referral to statutory stop smoking services, alone or in combination. [2016]

Cardioprotective diet

A Cochrane review of the Mediterranean diet for the primary prevention of cardiovascular disease found very small reductions in total and LDL-cholesterol (Rees et al, 2013). Three of five RCTs showed a positive effect on BP, with reductions of 0.7-7.8 mmHg in systolic and 0.7-3.7 mmHg in diastolic BP. The Dietary Approaches to Stop Hypertension (DASH) diet lowered systolic and diastolic BP when followed for 8 weeks (Appel et al, 1997). Long-term follow-up for up to 24 years (Fung et al., 2008) demonstrated that adherence to a DASH-style diet is associated with a lower risk of coronary heart disease and stroke among middle-aged women. Effects on stroke recurrence and mortality are not known. A Cochrane review (Adler et al, 2014) of dietary salt reduction for the prevention of cardiovascular disease confirms that small reductions in BP can be achieved in normotensive individuals, and greater reductions in hypertensive individuals. Many of the component trials lacked sufficient detail to assess bias; any benefits in terms of cardiovascular mortality and morbidity were modest or non-significant and confined to hypertensive groups. [2016]

Weight-reducing diet

Overweight and obesity is a significant risk factor for the development of cardiovascular disease and is associated with an increase in all-cause mortality (Adams et al, 2006; Bazzano et al, 2006). Both overweight (body mass index [BMI] greater than 25 kg/m2) and obesity (BMI greater than 30 kg/m2) are associated with an increased risk of ischaemic stroke (Strazzullo et al, 2010). Customary advice has targeted a healthy BMI to reduce risk of stroke but high quality intervention studies to support this approach are lacking. Some observational studies have reported a paradoxical inverse relationship between BMI and mortality following stroke (Kim et al, 2011a; Vemmos et al, 2011) with overweight and obese people having reduced mortality, but how this observation might translate into an intervention to reduce recurrent stroke is unclear. [2016]

Alcohol

A meta-analysis of 27 prospective studies with 1,425,513 participants reviewed the dose-response relation between alcohol and risk of stroke (Zhang et al, 2014). In the majority of the component studies, alcohol intake was self-reported. Low alcohol intake (below 15 g/day) was associated with a reduced risk of total stroke, ischaemic stroke and stroke mortality with no significant effect on haemorrhagic stroke (one UK unit = 8 g of alcohol). Moderate alcohol intake (15-30 g/day) had little or no effect on risk of total stroke, haemorrhagic stroke, ischaemic stroke or stroke mortality. Heavy alcohol intake (above 30 g/day) was associated with an increased risk of total stroke. It is not known if a similar relationship would apply to people who have already experienced stroke. [2016]

Micronutrient supplementation

A Cochrane review (Marti-Carvajal et al, 2015) examined the use of B-vitamin supplementation to prevent cardiac events. No benefit was found for homocysteine-lowering interventions in the form of supplements of vitamins B6, B9 or B12 given alone or in combination, at any dosage. Two meta-analyses (Alkhenizan & Al-Omran, 2004; Bin et al, 2011) and one systematic review (Eidelman et al, 2004) examined the effects of vitamin E supplementation on stroke recurrence and mortality, with no benefit seen. Dietary calcium and/or vitamin D supplementation has not been shown to reduce cardiovascular risk (Elamin et al, 2011; Bolland et al, 2014) and in one systematic review it was associated with a modest increased risk (Bolland et al, 2010). Confining the analysis to people with vitamin D deficiency likewise showed no benefit. The impact of other nutrients, including plant stanols/sterols, antioxidants such as vitamins A and C or selenium in stroke prevention is unknown (Hookway et al, 2015). [2016]

The literature related to follow-up after the end of formal rehabilitation is diverse; it includes specific rehabilitation interventions already described in Chapter 4 and overlaps with other areas, including secondary stroke prevention (Chapter 5), Vocational rehabilitation (Section 4.15 Return to work), and Supported self-management (Section 4.4 Self-management). [2023]

An intervention aimed at maintaining motor function using individual coaching was compared with usual care in a Norwegian RCT of moderate quality (Askim et al, 2018; Døhl et al, 2020). This intervention was provided once a month for 18 months, initially delivered mostly face-face, gradually introducing telephone contact. Individual goals were set and monthly schedules were agreed, which comprised 45 to 60 minutes of exercise (including 2 to 3 periods of vigorous activity) once a week plus physical activity for 30 minutes daily. Participants were given various options to join groups in different settings to match their requirements. Findings showed that this coaching intervention was safe but it had no additional effect on any aspect of motor function, ADL, fatigue, mood, quality of life, or caregiver strain. The cost analysis (Døhl et al, 2020) indicated that the intervention added costs to usual care. [2023]

An extended stroke rehabilitation service (Rodgers et al, 2019b; Shaw et al, 2020), which aimed to maximise recovery and adjustment to residual disability in the context of everyday activities, was compared with usual care in a high quality, UK-wide RCT. It comprised five extra telephone reviews (including goal setting and action planning) with senior therapists over 18 months after the end of early supported discharge. Compared with usual care, this intervention did not improve extended ADL but it did improve health-related quality of life, anxiety and depression, and patient satisfaction (including whether needs had been met, and whether sufficient treatment had been received to improve mobility). The intervention may also be cost-effective due to the quality of life advantages and a tendency towards lower overall costs in primary care. Participants with stroke felt that the reviews were reassuring, thorough and comprehensive, and that the goal setting had motivated them, but they were unclear if actual outcomes had been improved. Therapists thought that the intervention addressed the unmet need for ongoing support for people living with stroke and that reviews were comprehensive, but they disapproved of having to deliver the intervention by telephone only. They also thought that the reviews were useful, primarily around emotional and social issues, and they valued being able to connect people with stroke to services, although they felt uncomfortable if they were unable to provide the necessary support or onward referral for an identified rehabilitation need. [2023]

An intervention which aimed to identify physical, cognitive and emotional problems in daily life, provide support and psycho-education and refer to further specialised healthcare as required, was compared with usual care in a Dutch study (Verberne et al, 2021; Verberne et al, 2022). It comprised face-to-face sessions of up to 45 minutes each in a primary care centre at 6 months after hospital discharge, with the number of sessions depending on the nurse’s judgement. This intervention did not result in any improvements in mood or social participation compared to usual care, whilst other outcomes were not reported. The authors concluded that the intervention was cost-effective (Verberne et al, 2021), but due to the low quality of the study design and differences in healthcare systems, no conclusions can be drawn to support a recommendation. [2023]

The Improving Longer Term Stroke Care (LoTS2Care) programme (Forster et al, 2021) was designed to develop and test a longer-term integrated stroke care intervention to improve quality of life for people with stroke and their family/carers by addressing unmet needs and enhancing participation in life situations. Based on service user experiences, systematic literature reviews, and an evaluation of existing service models, a novel intervention (New Start) was designed, refined and tested in a feasibility cluster RCT involving 10 stroke services across England and Wales, recruiting 269 participants. Further development work will be undertaken before a full trial evaluation, and findings will inform subsequent editions of this guideline. [2023]

Some people start or continue to improve many months or years after the event, and these people may benefit from further rehabilitation or other support at a later stage. Whilst limited, there is evidence to suggest that for some people improvements in communication, arm function, walking, physical fitness and ADL can be achieved with interventions more than 6 months after stroke (Palmer & Enderby, 2007; Duncan et al, 2011; Ferrarello et al, 2011; Lohse et al, 2014; Veerbeek et al, 2014b; Ward et al, 2019). To provide personalised care to this group and to identify other unmet needs, the consensus of the Guideline Development Group is that a comprehensive, structured needs reassessment should be undertaken at 6 months – or earlier, depending on the individual’s needs. Structured, planned follow-up may reduce some of the health inequalities between those with higher and lower levels of education (Irewall et al, 2019). [2023]

This review should consider physical, psychological and social needs (including relationships and work, where applicable) related to adjusting to life after stroke. The review should identify what matters most to the person with stroke, ensuring that any referrals are appropriately targeted. The role of the professional is to support the person with stroke to identify and prioritise their needs and goals and act as a facilitator, initiating timely referrals or providing guidance to appropriate service providers within health and social care or beyond, as and when required. Referrers should consider health and social care services, and services offered by other organisations, including stroke support groups (e.g. communication support provided by the Stroke Association, Chest Heart and Stroke Scotland, the Irish Heart Foundation or Different Strokes) and local councils (e.g. exercise referral schemes). [2023]

There is no ‘one size fits all’ with regard to onward referral; instead, given the diversity in recovery profiles, needs and timing, as well as the cultural contexts in which people with stroke lead their lives, a personalised approach is required. This is to ensure that the aims, content, mode of delivery (e.g. face-face, online) and timing of any services are aligned with the priorities, needs, goals and circumstances of the person with stroke and their family/carers where appropriate. [2023]

Primary care teams, in collaboration with hospital-based or community stroke teams, specialist community neurorehabilitation or brain injury rehabilitation teams, will need to consider the resource implications of implementing follow-up and annual review for people with stroke living in the community and make appropriate provision. [2023]

A metasynthesis of qualitative research identified several themes which, from the perspective of people with stroke, acted as barriers or facilitators to community reintegration (Walsh et al, 2015). As well as the primary effects of the stroke (impairments and fatigue), these comprised personal factors (perseverance, adaptability, emotional challenges, relevance of activities), social factors (sense of belonging versus stigmatisation, levels of support, environmental limitations) and interactions with professionals (levels of support, joint decision-making, relevance of rehabilitation to real world requirements). Anger, frustration and more challenging behavioural problems may present barriers to social and community integration but other than generic principles, there is limited evidence to guide management for people or families/carers with these problems. [2016]

A systematic review of leisure therapy including eight studies and 615 participants identified methodological shortcomings but nonetheless some evidence of short-term improvements in quality of life and mood as well as increased participation and satisfaction with leisure activities (Dorstyn et al, 2014). In a review of 24 studies (including 2042 people with stroke) that included measures of social participation as an outcome, a small beneficial effect was identified for interventions utilising exercise (Obembe & Eng, 2015). A community walking training programme in which people with stroke undertook walking therapy in a real-world environment resulted in greater improvements in walking function and social participation (Kim et al, 2014), but a Cochrane review found the available evidence insufficient to establish effectiveness (Barclay-Goddard et al, 2015). [2016]