Rapid Assay Diagnostic for Acute Stroke Recognition (RADAR): study protocol for a diagnostic accuracy study

Abstract

Introduction: Large-vessel occlusion (LVO) stroke is effectively treated by time-critical thrombectomy, a highly specialised procedure only available in a limited number of centres. Many patients with suspected stroke are admitted to their nearest hospital and require transfer to access treatment, with resulting delays. This study is evaluating the accuracy of a new rapid portable test for LVO stroke which could be used in the future to select patients for direct admission to a thrombectomy centre.

Methods and analysis: Rapid Assay Diagnostic for Acute Stroke Recognition (RADAR) is a prospective observational cohort study taking place in stroke units in England. Participants are adults with a new suspected stroke with at least one face, arm or speech (FAST) symptom(s) and known onset within 6 hours or last known to be well 6-24 hours ago. The index test ('LVOne test' (Upfront Diagnostics)), consists of two portable lateral flow assays which use fingerprick capillary blood to detect d-dimer and glial fibrillary acidic protein concentrations. Reference standards comprise independently adjudicated standard CT/MRI brain±CT/MR angiography with senior clinician opinion to establish: ischaemic stroke±LVO; intracerebral haemorrhage; transient ischaemic attack; stroke mimic. Analyses will report sensitivity, specificity and negative and positive predictive values for identification of LVO stroke. Powered using a primary analysis population (≥2 FAST symptoms and known onset within 6 hours), 276 participants will detect a test specificity of 92%. The broader total study population which allows evaluation of the test for milder symptoms and unknown onset times is estimated to be 552 participants.

Ethics and dissemination: Ethical (North East-Newcastle & North Tyneside 2 Research Ethics Committee (reference: 23/NE/0043), Health Research Authority and participating National Health Service Trust approvals are granted. Consent is required for enrolment. Dissemination of results will include presentations at conferences, publication in journals and plain English summaries.

Trial registration number: ISRCTN12414986.

Keywords: Research Design; STROKE MEDICINE; Triage.

Figure 1. Performing the LVOne test.

Figure 2. Decision tree for assigning the clinical outcome state. CTA, CT angiography; ICH, intracerebral haemorrhage; LVO, large-vessel occlusion; MRA, MR angiography; TIA, transient ischaemic attack.