An assessment of the cost-effectiveness of magnetic resonance, including diffusion-weighted imaging, in patients with transient ischaemic attack and minor stroke: a systematic review, meta-analysis and economic evaluation.

BACKGROUND: Patients with transient ischaemic attack (TIA) or minor stroke need rapid treatment of risk factors to prevent recurrent stroke. ABCD2 score or magnetic resonance diffusion-weighted brain imaging (MR DWI) may help assessment and treatment.

OBJECTIVES: Is MR with DWI cost-effective in stroke prevention compared with computed tomography (CT) brain scanning in all patients, in specific subgroups or as 'one-stop' brain-carotid imaging? What is the current UK availability of services for stroke prevention?

DATA SOURCES: Published literature; stroke registries, audit and randomised clinical trials; national databases; survey of UK clinical and imaging services for stroke; expert opinion.

REVIEW METHODS: Systematic reviews and meta-analyses of published/unpublished data. Decision-analytic model of stroke prevention including on a 20-year time horizon including nine representative imaging scenarios.

RESULTS: The pooled recurrent stroke rate after TIA (53 studies, 30,558 patients) is 5.2% [95% confidence interval (CI) 3.9% to 5.9%] by 7 days, and 6.7% (5.2% to 8.7%) at 90 days. ABCD2 score does not identify patients with key stroke causes or identify mimics: 66% of specialist-diagnosed true TIAs and 35-41% of mimics had an ABCD2 score of ≥ 4; 20% of true TIAs with ABCD2 score of < 4 had key risk factors. MR DWI (45 studies, 9078 patients) showed an acute ischaemic lesion in 34.3% (95% CI 30.5% to 38.4%) of TIA, 69% of minor stroke patients, i.e. two-thirds of TIA patients are DWI negative. TIA mimics (16 studies, 14,542 patients) make up 40-45% of patients attending clinics. UK survey (45% response) showed most secondary prevention started prior to clinic, 85% of primary brain imaging was same-day CT; 51-54% of patients had MR, mostly additional to CT, on average 1 week later; 55% omitted blood-sensitive MR sequences. Compared with 'CT scan all patients' MR was more expensive and no more cost-effective, except for patients presenting at > 1 week after symptoms to diagnose haemorrhage; strategies that triaged patients with low ABCD2 scores for slow investigation or treated DWI-negative patients as non-TIA/minor stroke prevented fewer strokes and increased costs. 'One-stop' CT/MR angiographic-plus-brain imaging was not cost-effective.

LIMITATIONS: Data on sensitivity/specificity of MR in TIA/minor stroke, stroke costs, prognosis of TIA mimics and accuracy of ABCD2 score by non-specialists are sparse or absent; all analysis had substantial heterogeneity.

CONCLUSIONS: Magnetic resonance with DWI is not cost-effective for secondary stroke prevention. MR was most helpful in patients presenting at > 1 week after symptoms if blood-sensitive sequences were used. ABCD2 score is unlikely to facilitate patient triage by non-stroke specialists. Rapid specialist assessment, CT brain scanning and identification of serious underlying stroke causes is the most cost-effective stroke prevention strategy.

FUNDING: The National Institute for Health Research Health Technology Assessment programme.