6.1.1. Clinical introduction

Recent evidence underlines the importance of immediate assessment and treatment of patients with TIA who are at high risk of completed stroke.³⁰ Careful history taking and examination is essential to exclude other diagnoses (e.g. migraine, seizure, syncope, tumour) and to assess vascular risk factors including hypertension, diabetes and dyslipidaemia. See section 8.2 for the use of early aspirin and other preventative measures. Early carotid scanning is essential to exclude significant carotid stenosis in patients who would fulfil criteria for carotid endarterectomy (see section 6.3). Not all patients with TIA need brain scanning. The selection of patients for urgent scanning is dependent on clinical features; it is important that brain scanning does not delay the institution of optimum secondary prevention or the detection and treatment of significant carotid stenosis. MR scanning is very much more sensitive than CT, particularly if performed early and using diffusion-weighted imaging (DWI); CT perfusion can also be used to detect small ischaemic lesions that might not be visible on standard CT.

The clinical question to be addressed is which patients with suspected TIA should undergo brain imaging.

6.1.2. Clinical methodological introduction

One meta-analysis (N=19 studies) was identified that reported on the association between clinical and demographic factors and the presence of acute ischaemic lesions on DWI in patients with TIA.³¹ The analysis included studies on patients imaged up to 14 days post event (median delay to scan: 37 hours). Level 3

6.1.3. Health economic methodological introduction

No papers were identified.

6.1.4. Clinical evidence statements

The systematic review reported a positive association between a positive DWI and motor weakness, dysphasia, dysarthria, duration of symptoms ≥60 mins, atrial fibrillation and ipsilateral carotid stenosis ≥50%. There were no associations between positive DWI and age ≥60 yrs (NS), previous hypertension (NS), current raised blood pressure (NS) and diabetes (NS).³¹ Level 3

Of the studies reporting on patients who were scanned within 24 hours or less from the index event, a positive scan was significantly associated with motor weakness and dysphasia only.³¹ Level 3

6.1.5. Health economic evidence statement

It was noted that MRI scan is considerably more expensive than CT scan – £228 per scan compared with £78 per scan.³²

6.1.6. From evidence to recommendations

There is no evidence that specifically addresses the question of which patients with TIA should be referred for urgent brain imaging. The GDG noted that good clinical assessment is essential to detect stroke mimics and to establish the vascular territory involved where possible. Brain imaging is of potential value in the detection of stroke mimics and in establishing the diagnosis where this is in doubt. In addition, brain imaging may be of value in determining the vascular territory involved where this is not clear from the clinical assessment (examples where imaging may be helpful because diagnosis is in doubt or vascular territory may need to be determined, are illustrated in the section below entitled ‘Cases where brain imaging is helpful in the management of TIA’). The GDG extrapolated from the evidence presented in section 5.2 (scoring systems to identify patients with TIA at high risk) and agreed that the ABCD² score should be used to identify those patients in need of immediate assessment and management, including urgent scanning where required. These patients need MR with DWI (where there are no contraindications) within 24 hours to avoid delay in instituting secondary prevention and the detection and management of significant carotid stenosis. An expert consensus was agreed that patients with severe comorbidities may not be appropriate for scanning if the results would not change management.

6.1.7. RECOMMENDATIONS

R9.

People who have had a suspected TIA (that is, whose symptoms and signs have completely resolved within 24 hours) should be assessed by a specialist (within 1 week of onset of symptoms) before a decision on brain imaging is made.

R10.

People who have had a suspected TIA who are at high risk of stroke (for example, with an ABCD² score of 4 or above, or with crescendo TIA) in whom the vascular territory or pathology is uncertain should undergo urgent brain imaging^* (preferably diffusion-weighted magnetic resonance imaging (MRI)).

R11.

People who have had a suspected TIA who are at lower risk of stroke (for example, an ABCD² score of less than 4) in whom the vascular territory or pathology is uncertain should undergo brain imaging^** (preferably diffusion-weighted MRI).

6.2.1. Clinical introduction

In 2006, 78% of hospitals had neurovascular clinics, with a median time between onset and review of 12 days.³ The key purpose of the clinic is to confirm the diagnosis of TIA (and manage those patients with an alternative diagnosis) and to ensure timely and appropriate secondary prevention. There has been little clarity over the need for brain scanning, with wide variations between clinics in the proportion of patients with TIA routinely scanned. Many clinicians have used CT because of lack of access to MR but availability of MR is now improving rapidly across the UK. Brain scanning may be used to detect stroke mimic (e.g. tumour) but diagnostic yields are low, unless there are suggestive clinical features. Although CT is very sensitive to haemorrhage early after the event, bleeds may be missed if scanning is delayed. Brain imaging is of value in determining the presence of vascular lesions (which may be helpful if there is diagnostic doubt) and helping to establish vascular territory where this is not clear. MR scanning, especially with diffusion-weighted imaging/fluid-attenuated inversion recovery (DWI/FLAIR) performed early (ideally within 24 hours) has high sensitivity for the detection of small ischaemic lesions which may be missed on CT scan.³³

The clinical question to be addressed is in those patients with TIA who require brain imaging whether MR or CT provides the most information to guide treatment.

6.2.2. Clinical methodological introduction

For this question, we looked at studies that reported on the association between imaging findings and the subsequent risk of mortality or morbidity.

Five observational studies/case series were identified, all reporting on MR diffusion weight imaging (MR-DWI) findings. See table 6.1 below for a summary of the study characteristics.^34–38

Table 6.1

A summary of the study characteristics.

6.2.3. Health economic methodological introduction

No papers were identified.

6.2.4. Clinical evidence statements

The proportion of patients with MR-DWI abnormalities ranged from 25 to 58%.

Only the results of multivariate analysis are reported here:

• At 1 year, patients without a DWI abnormality were significantly more likely to have a subsequent TIA, but significantly less likely to have a subsequent stroke, than patients with a DWI abnormality (N=85).³⁴ Level 3
• Patients with a DWI abnormality were significantly more likely to have an in-hospital recurrent TIA or stroke than those without a DWI abnormality (N=146).³⁵ Level 3
• At 3 months, DWI abnormalities were a significant independent predictor of stroke (N=203).³⁶ Level 3
• The presence of a DWI abnormality in patients with TIA or minor stroke was significantly associated with an increased risk of 90-day stroke (N=120).³⁷ Level 3
• Symptoms greater than 1 hour and DWI abnormalities were significant independent predictors of further cerebral vascular events or any vascular event (follow-up mean 389 days) (N=83).³⁸ Level 3

6.2.5. From evidence to recommendations

The evidence reviewed did not specifically compare CT with MR after TIA. However, it is well established that MR is more sensitive than CT in the detection of vascular lesions particularly if performed early. The consensus of the GDG was that where brain scanning was felt to be necessary following TIA, MR with DWI within 24 hours should be performed. For those patients with contraindications or unable to tolerate MR, CT scanning should be used.

6.2.6. RECOMMENDATIONS

R12.

People who have had a suspected TIA who need brain imaging (that is, those in whom vascular territory or pathology is uncertain) should undergo diffusion-weighted MRI except where contraindicated,^*** in which case computed tomography (CT) scanning should be used.

6.3.1. Clinical introduction

Carotid imaging is required to determine the presence and severity of carotid stenosis in those individuals who may be appropriate for carotid endarterectomy, i.e. those with a TIA or minor or recovered stroke involving the anterior circulation who are fit and willing for surgery. Doppler ultrasound, MR angiography and CT angiography can be used in the screening for and assessment of carotid stenosis. The urgency of the carotid imaging depends on the individual’s risk of stroke (defined on clinical criteria: see section 6.4). Furthermore the value of carotid surgery decreases with time from the event, surgery ceases to be of value after 12 weeks of the event in trials for men and 2 weeks for women. Imaging should therefore be done rapidly if appropriate patients are to be assessed for surgery in a timely manner.

The clinical question to be addressed is which patients with suspected stroke/TIA should be referred for urgent carotid imaging.

6.3.2. Clinical methodological introduction

Four studies reported on the association between carotid stenosis and symptoms, demographics and comorbid conditions in patients who had undergone carotid duplex scanning (N=816);³⁹ (N=5,807 scans);⁴⁰ (N=726).^41,42 Two of the studies were retrospective;^39,40 and two were prospective;⁴¹ (N=305).⁴² One study was excluded⁴³ as all of the data were incorporated in a more recent study.⁴⁰

In two studies the populations were relatively homogenous, one was on patients with acute stroke admitted to hospital⁴² and the other on patients admitted to hospital or seen in an outpatient clinic with acute stroke, cerebral or retinal TIA or retinal strokes.⁴¹ Two studies reported on heterogenous populations, including for example patients with TIA, dizziness and dysphasia.^39,40

6.3.3. Health economic methodological introduction

One study was identified that modelled the cost effectiveness of different assessment strategies for carotid stenosis. ²⁷

The study was a NHS HTA report of a systematic review of the costs of less invasive tests, outpatient clinics, endarterectomy and stroke, along with a micro-costing exercise. A Markov model of the process of care following a TIA/minor stroke was developed, populated with data from stroke epidemiology studies in the UK, effects of medical and surgical interventions, outcomes, quality of life and costs. Both strokes and MIs were modelled. A survey of UK stroke prevention clinics provided typical timings of surgery. Twenty-two different carotid imaging strategies were evaluated for short- and long-term outcomes, quality adjusted life years, NHS cost and net benefit. The strategies varied according to:

• the choice and sequence of tests (which included ultrasound, computed tomographic angiography (CTA), magnetic resonance angiography (MRA), contrast enhanced MRA (CEMRA) and intra-arterial angiography (IAA)), and
• the level of stenosis at which surgery would be under-taken.

No attempt was made to assess whether carotid imaging is cost effective compared with no carotid imaging in any population.

6.3.4. Clinical evidence statements

6.3.5. Health economic evidence statements

In the cost-effectiveness model²⁷ based on contemporary UK timings, the strategies which prevented most strokes and produced greatest net benefit were those that:

1. allowed more patients to reach endarterectomy very quickly, and
2. where those patients with 50–69% stenosis would be offered surgery in addition to those with 70–99% stenosis.

This included most strategies with ultrasound as first or repeat test, and not those with intra-arterial angiography. However, the model was sensitive to less invasive test accuracy, cost and timing of endarterectomy. In patients investigated late after TIA, some tests are much less accurate and therefore contrast-enhanced magnetic resonance angiography should be used before surgery. The authors conclude that in the UK, less invasive tests could be used in place of intra-arterial angiography if radiologists trained in carotid imaging are available.

The cost effectiveness of carotid imaging compared with no carotid imaging could not be easily inferred from this study.

6.3.6. From evidence to recommendations

Carotid imaging is essential to identify those people who would benefit from carotid endarterectomy (CEA). The evidence does not identify any clinical sign that is pathognomonic for carotid stenosis although some (e.g. bruit) may be suggestive. The group therefore agreed that all people who are suitable for carotid interventions should have access to carotid imaging.

6.3.7. RECOMMENDATION

R13.

All people with suspected non-disabling stroke or TIA who after specialist assessment (see section 5) are considered as candidates for carotid endarterectomy should have carotid imaging within 1 week of onset of symptoms. People who present more than 1 week after their last symptom of TIA has resolved should be managed using the lower-risk pathway.

6.4.1. Clinical introduction

While the benefits of carotid intervention for symptomatic carotid stenosis of >50% according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria⁴⁴ and >70% according to the European Carotid Surgery Trial (ECST) criteria⁴⁵ have been clearly described elsewhere.²⁹ The benefit of early surgery (within 2 weeks of symptoms) may be outweighed by the risk of adverse events in patients with recent cerebral infarction, particularly those with significant neurological disability following a stroke or who have a high anaesthetic risk. However, patients with clinically defined high-risk TIA are clearly at highest risk of stroke within 2 days of the incident, implying that for some patients, very early endarterectomy might be most beneficial. Similarly, a case-series study reported no perioperative complications associated with early carotid stenting (<14 days) in patients with symptomatic carotid artery stenosis.⁴⁶ The non-randomised EXPRESS study²⁰ suggests that patients with TIA and minor stroke benefit considerably from a package of early medical interventions including antiplatelet agents, a statin and blood pressure treatment.

The clinical question is which patients with symptomatic carotid stenosis should be referred for early interventional procedures. It is of note that the lack of standardisation of the definition of significant carotid stenosis can be confusing. It is important that those reporting carotid imaging studies clearly state which criteria for diagnosis are being used.

6.4.2. Clinical methodological introduction

One systematic review was identified.⁴⁷ This reports that the data on CEA performed less than 1 week compared with 1 week or later (two studies, N=135 vs 1,492). The data from Rothwell et al.⁴⁸ are reported in both this systematic review and the pooled analysis.⁴⁴ However, the analysis is reported for different time periods and on different outcomes.

Two studies reported on pooled data from two large RCTs, namely the ECST and the NASCET (N=5,893).^45,44 ECST includes a measure of the normal lumen diameter at the site of the lesion based on a visual impression of where the normal artery was before the damage caused by the stenosis. NASCET includes a measure of the diameter of the visible portion of disease-free internal carotid artery (ICA) distal to the stenosis, or the stenosis was classified as 95% if the distal ICA had collapsed (see glossary). Patients with symptomatic carotid stenosis were randomised to medical treatment or to CEA. The data are reported according to time from last symptomatic ischaemic event to randomisation or surgery. 14.5% patients in the ECST and 25.9% patients in NASCET were randomised within 2 weeks.⁴⁵ Patients were included with TIA, non-disabling ischaemic stroke, or a retinal infarction, in the territory of a stenosed carotid artery. The two trials used different techniques to measure the degree of carotid stenosis and each trial made different recommendations regarding the degree of stenosis above which surgery was effective. However, when the angiograms from the ECST were re-measured in accordance with NASCET criteria, the outcomes of the two trials were comparable. Level 1++

No RCT studies were identified on carotid stenting in acute stroke.

The prospective case series (N=238) recorded data on all patients undergoing CEA after ipsilateral acute stroke performed within 1 month of symptom onset.⁴⁹ 55% of patients were operated within 2 weeks of symptom onset. All patients had stenosis of 50% or greater. Twelve patients underwent the procedure within 24 hours of symptom onset for stroke in evolution. According to NASCET criteria, of the 72% patients with available brain imaging, 35% were cortical infarcts, 16% small border zone infarcts, 13% deep infarcts and 36% no visible infarct. The degree of stenosis, or its statistical association with outcome, was not reported in this study. Level 3

6.4.3. Health economic methodological introduction

No papers were identified.

6.4.4. Clinical evidence statements

1.1. Clinical and demographic indicators

Table 6.2 below gives the absolute risk reduction with surgery in 5-year actual risk of ipsilateral carotid ischaemic stroke and any stroke or death within 30 days after trial surgery from the pooled analysis of the RCTs. This shows that the effects of surgery are modified by time since last event, gender and age such that the benefit statistically decreases as the time since last symptoms increases, and is significantly greater in males than females and in the elderly. These results are consistent across patients with 50 to 69% and 70% or more stenosis.⁴⁵ Level 1++

Table 6.2

The absolute risk reduction with surgery in 5-year actual risk of ipsilateral carotid ischaemic stroke and any stroke or death within 30 days after trial surgery from the pooled analysis of the RCTs.

Univariate analysis from the prospective case-series data showed that increasing lesion size on preoperative CT scan or MRI significantly increased the odds of permanent neurological deficit.⁴⁹ Level 3

6.4.5. From evidence to recommendations

No RCTs were identified which studied early vs late carotid interventions using the 2-week cut off for the definition of acute stroke. No evidence for early carotid stenting (within the 2-week time period of the guideline) was identified. The GDG recognised the need for further research in this area. There is a need for a randomised trial comparing the safety and efficacy of carotid stenting to CEA within 2 weeks of TIA or recovered stroke. The GDG noted the systematic review reported that there was no significant difference in outcome in perioperative stroke or death with patients undergoing CEA less than 1 week compared to greater than 1 week. The GDG also noted that only the pooled analysis reported the long-term absolute relative risk of stroke or death. The evidence for benefit of referral for early CEA was extrapolated from the two studies which reported on pooled data from two large RCTs. Differences in genders were also noted (women only benefit from surgery early while men continue to benefit from surgery for longer), the GDG therefore noted that in order for women to benefit from surgery all patients should receive surgery early. There is evidence that patients with unstable neurological symptoms (stroke in evolution, crescendo TIA) may be harmed by early surgery. In neurologically stable patients, there is no statistical difference in the incidence of postoperative neurological deficit after CEA performed at 1–4 weeks after stroke onset. There is clear evidence from pooled data which showed that the benefit of CEA decreases as the delay to randomisation increases for patients with >50% stenosis according to the NASCET criteria. It was therefore agreed by the GDG that patients should be referred for CEA within 1 week of onset. Patients who have carotid stenosis of <50% should receive best medical treatment.

6.4.6. RECOMMENDATIONS

R14.

People with stable neurological symptoms from acute non-disabling stroke or TIA who have symptomatic carotid stenosis of 50–99% according to the North American Symptomatic Carotid Endarterectomy Trial (NASCET) criteria, or 70–99% according to the European Carotid Surgery Trialists’ (ECST) Collaborative Group criteria, should:

• be assessed and referred for carotid endarterectomy (CEA) within 1 week of onset of stroke or TIA symptoms
• undergo surgery within a maximum of 2 weeks of onset of stroke or TIA symptoms
• receive best medical treatment (control of blood pressure, antiplatelet agents, cholesterol lowering through diet and drugs, lifestyle advice).

R15.

People with stable neurological symptoms from acute non-disabling stroke or TIA who have symptomatic carotid stenosis of less than 50% according to the NASCET criteria, or less than 70% according to the ECST criteria, should:

• not undergo surgery
• receive best medical treatment (control of blood pressure, antiplatelet agents, cholesterol lowering through diet and drugs, lifestyle advice).

R16.

Carotid imaging reports should clearly state which criteria (ECST or NASCET) were used when measuring the extent of carotid stenosis.