GB2563414A - Improvements in stroke diagnostics - Google Patents

Abstract

A method of diagnosing or differentiating stroke comprises measuring the level of one or more of the biomarkers h-FABP, D-dimer, sTNFR1 and IL-6 in a sample taken from a patient who has suffered or is suffering stroke-like symptoms and comparing the measured level to a corresponding reference value, where the reference value is derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort. A method is also provided of determining whether to administer thrombolytic therapy to a patient who has suffered or is suffering stroke-like symptoms. Use of a stroke mimic or a combined stroke mimic and TIA patient cohort sample biomarker measurement value as a reference value in an assay for stroke diagnosis or differentiation is also disclosed.

Description

Improvements to Stroke Diagnosis

Background to the Invention

Stroke is the third leading cause of death worldwide and can be defined as the rapidly developing loss of brain function(s) due to interruption in the blood supply to the brain. According to the World Health Organisation, 15 million people per year suffer stroke worldwide, with 5 million dying and a further 5 million being permanently disabled. High blood pressure is estimated to be a contributing factor in 12.7 million of these 15 million stroke cases. In the UK, approximately 150,000 people have a stroke each year and stroke accounts for around 53,000 deaths per year. Stroke costs the economy an estimated £8 billion per year in England alone and stroke patients occupy approximately 20 per cent of all acute hospital beds and 25 per cent of long term beds. Stroke can be classified into three sub-types: i) Ischaemic stroke (IS) occurs when blood supply to the brain is decreased, resulting in brain damage. An ischemic stroke occurs when a blood vessel becomes blocked, usually via a blood clot. This clot may form locally at an atherosclerotic plaque (thrombotic stroke) or alternatively may occur due to a travelling particle or debris that has originated from elsewhere in the bloodstream (embolic stroke); ii) Transient ischaemic attack (TIA) is a ‘mini stroke’ that occurs when blood supply to the brain is temporarily decreased. A TIA is diagnosed if symptoms are quickly resolved (within 24 hours with the individual returning to normal health); and iii) Haemorrhagic stroke (HS) occurs when blood accumulates within the skull vault, usually when a weakened blood vessel ruptures. Haemorrhagic stroke can be classified into two major subtypes, namely intracerebral (within the brain tissue) and subarachnoid (around the surface of the brain and under its protective layer). IS and TIA account for approximately 85% of all stroke cases and HS accounts for 15%. In order to minimise neurological damage following stroke it is crucial that stroke patients are rapidly and accurately diagnosed, so that appropriate treatment can be administered. For example, in order to break down clots thrombolytic therapy such as tissue plasminogen activator (TPA) can be administered. However, such therapy is only warranted in IS and is detrimental in HS. The nature of TIA does not require such therapy and blood thinners such as warfarin and aspirin are prescribed in such cases. At present, if stroke is suspected, physical symptoms are evaluated and a computerised tomography (CT) scan is usually performed. A CT scan has good sensitivity for identifying HS patients (approximately 90% sensitivity) but poor sensitivity for identifying IS and TIA patients (approximately 20% sensitivity). In practice minimal or no tissue damage occurs for TIA due to its transient nature, therefore CT scanning is ineffective as a diagnostic technique. Magnetic Resonance Imaging (MRI) has improved sensitivity for IS diagnosis (up to approximately 80%) but increased time requirements, machine accessibility, and high cost have limited its use for stroke diagnosis. The poor sensitivity of CT scanning for the detection of IS and TIA means that a biological fluid-based diagnostic biomarker tests for detecting IS and/or TIA would be an invaluable tool to aid clinicians in the diagnosis of stroke sub-type. Biological fluid-based biomarkers have the potential to expedite and increase the accuracy of IS and TIA diagnosis; of particular importance is the diagnosis of IS, as this condition has immediate life-debilitating and life-threatening potential that can be counter-acted by the administration of clot-busting drugs such as TPA. Various candidate biomarkers have been proposed for the diagnosis of stroke and stroke sub-type delineation in the prior art. EP1238284, EP1671130, EP1902319, EP1664795, EP2341350, EP2357477, EP2786155, and EP3004382. Possible stroke biomarkers addressed in these prior art documents include heart fatty acid binding protein (h-FABP), protein deglycase DJ-1 (Parkinson disease protein 7 or PARK-7), interleukin 6 (IL-6), glial fibrillary acidic protein (GFAP), soluble tumour necrosis factor 1 (sTNFRI), nucleoside diphosphate kinase A (NDKA), glutathione S-transferase P (GSTPi) and D-dimer. Each uses methods that derive discriminatory patient sample biomarker levels of stroke by comparison of stroke patient samples to a reference value that is healthy patient samples.

Summary of the Invention

Improved methods of stroke diagnosis or differentiation are described which incorporate biomarker reference values of stroke mimic patients; the biomarker reference levels can, in addition to levels derived from stroke mimic patients, also comprise levels derived from TIA patients. Thus a first aspect of the invention provides a method of aiding the diagnosis of IS, comprising measuring the level one or more of h-FABP, D-dimer, IL-6 and sTNFRI in an ex vivo sample of a patient who has suffered, is suffering or is suspected of suffering from a stroke and comparing this level to a reference sample value in which a measured level in the patient sample that is greater than the reference value is indicative of IS characterised by the reference value of one or more of h-FABP, D-dimer, IL-6 and sTNFRI being derived from its or their level in a patient cohort of stroke mimics or combined stroke mimics and TIA.

Figure 1 shows the patient distribution for h-FABP and D-dimer for the various groups

Detailed Description of the Invention

Unless stated otherwise, all references herein to the term ‘stroke’ encompasses all three forms of stroke. As used herein, the term ‘ischaemic stroke (IS)’ refers to the type of stroke that occurs when blood supply to the brain is decreased, resulting in brain damage. An ischemic stroke occurs when a blood vessel becomes blocked, usually via a blood clot. This clot may form locally at an atherosclerotic plaque (thrombotic stroke) or alternatively may occur due to a travelling particle or debris that has originated from elsewhere in the bloodstream (embolic stroke). The term ‘transient ischaemic attack (TIA)’ refers to a ‘mini stroke’ that occurs when blood supply to the brain is temporarily decreased. A TIA is diagnosed if symptoms are quickly resolved (within 24 hours with the individual returning to normal health). The term ‘haemorrhagic stroke (HS)’ occurs when blood accumulates within the skull vault, usually when a weakened blood vessel ruptures. Haemorrhagic stroke can be classified into two major sub-types: intracerebral (within the brain tissue); and subarachnoid (around the surface of the brain and under its protective layer). A ‘stroke mimic’ is defined as an event that is stroke-like in its effect on a patient but is in fact an event other than a HS, IS or TIA. Traumatic brain injury, hypoglycaemia, hyperglycaemia, hyponataemias, migraine, fainting/loss of consciousness, numbness of hands and legs, and seizures were stroke mimic events encountered in this study. Other less commonly encountered stroke mimic conditions not encountererd in this study such as brain infection can be anticipated not to affect the statistical significance of the analyses incorporated herein and are also included under the term stroke mimic. References herein to ‘a patient who has suffered or is suffering a stroke or stroke-like symptoms’ include a patient who has been diagnosed as currently suffering from a stroke or who is or has recently had a stroke (within ~ 6 hours). That the stroke may have been a recent event initiated the process of the individual, or someone on behalf of the individual, seeking clinical help. The term ‘differentiating stroke’ (‘differential diagnosis’) means that specific stroke conditions TIA, IS and HS can be delineated when a patient presents having had or with a stroke condition. The terms “subject” and “patient” may be used interchangeably herein and refer to a mammal including a non-primate (e.g. a cow, pig, horse, dog, cat, rat and mouse) and a primate (e.g. a monkey and human). Preferably the subject or patient is a human. As used herein, the term ‘biomarker’ refers to a molecule present in a biological sample obtained from a patient, the concentration of which in said sample may be indicative of a pathological state. It is well understood in the art that biomarker normal or ‘background’ concentrations may exhibit slight variation due to, for example, age, gender or ethnic/geographical genotypes. As a result, the cut-off value used in the methods of the invention may also slightly vary due to optimization depending upon the target patient/population. The biological sample obtained from a patient is preferably a blood, serum or plasma sample. As used herein, the term ‘ex vivo’ has its usual meaning in the art and refers to a sample that has been removed from a patient’s body. When a blood sample is taken from the patient for analysis, whole blood, serum or plasma is analysed. Analysis of the blood sample can be by way of several analytical methodologies such as mass spectrometry linked to a pre-separation step such as chromatography. The preferred methodology is based on immuno-detection. Immuno-detection technology is also readily incorporated into transportable or handheld devices for use outside of the clinical environment. A quantitative immunoassay such as a Western blot or ELISA can be used to detect the amount of protein. A preferred method of analysis comprises using a multi-analyte biochip which enables several proteins to be detected and quantified simultaneously. A biochip is a planar substrate that may be, for example, mineral or polymer based, but is preferably ceramic. Probes are adsorbed on or chemically attached to the surface of the biochip. The probes can be any biomarker-specific probe or binding ligand. As used herein, the term ‘specific’ means that the probe binds only to one of the biomarkers of the invention, with negligible binding to other biomarkers of the invention or to other analytes in the biological sample being analysed. This ensures that the integrity of the diagnostic assay and its result using the biomarkers of the invention is not compromised by additional binding events. When identifying the various biomarkers of the invention it will be apparent to the skilled person that as well as identifying the full length protein, the identification of a fragment or several fragments of a protein is possible, provided this allows accurate identification of the protein. Similarly, although a preferred probe of the invention is a polyclonal or monoclonal antibody, other probes such as aptamers, molecular imprinted polymers, phages, short chain antibody fragments and other antibody-based probes may be used. Preferably, a solid state device is used in the methods of the present invention, preferably the Biochip Array Technology system (BAT) (available from Randox Laboratories Limited). More preferably, the Evidence Evolution and Evidence Investigator apparatus (available from Randox Laboratories) may be used to determine the levels of biomarkers in the sample. 2D Gel Electrophoresis is also a technique that can be used for multi-analyte analysis.

The term D-dimer as used herein refers to the fibrin degradation product. The term “sTNFRT’ as used herein refers to soluble Tumour Necrosis factor Receptor 1 (UniProt P19438). The term “IL-6” as used herein refers to interleukin 6 (UniProt P05231). The term “PARK7” as used herein refers to Parkinson disease protein 7, also known as DJ-1 (UniProt Q99497). The term “FABP3” as used herein refers to fatty acid-binding protein 3 (UniProt P05413). The term “GFAP” as used herein refers to Glial fibrillary acid protein (UniProt P14136). The term “GSTP” as used herein refers to Glutathione S-transferase P (UniProt P09211). The term “NDKA” as used herein refers to Nucleoside diphosphate kinase A (Uniprot P15531).

The skilled person will be aware of numerous suitable methods for developing statistical algorithms, and all of these are within the scope of the present invention. Examples of suitable classification algorithms include multinominal logistic regression, multilayer perceptron neural network (MLP), artificial neural networks, support vector machines, N Bayes classification and random forest classifiers. The present inventors have found that both multinominal logistic regression and Naive Bayes Classification achieve similar performance in the context of the present invention, suggesting the importance of the analytes (i.e. biomarkers) used in the methods of the invention, rather than the method used to generate the algorithmic model. However, in a preferred embodiment, the statistical algorithm includes Naive (N) Bayes classification. The accuracy of statistical methods used in accordance with the present invention can be best described by their receiver operating characteristics (ROC). The ROC curve addresses both the sensitivity, the number of true positives, and the specificity, the number of true negatives, of the test.

Therefore, sensitivity and specificity values for a given combination of biomarkers are an indication of the accuracy of the assay. For example, if a biomarker combination has sensitivity and specificity values of 80%, out of 100 patients which have stroke, 80 will be correctly identified from the determination of the presence of the particular combination of biomarkers as positive for stroke, while out of 100 patients who have not suffered a stroke 80 will accurately test negative for the disease. Sensitivity and specificity values are defined by the cut-off value assigned to the biomarker - depending upon the sensitivity or specificity required in an assay i.e. whether a ‘rule in’ or ‘rule out’ assay is required, the cut-off value assigned will vary. If two or more biomarkers are to be used in the diagnostic method a suitable mathematical model, such as logistic regression equation, can be derived. The logistic regression equation might include other variables such as age and gender of patient. The ROC curve can be used to assess the accuracy of the logistic regression model. The logistic regression equation can be used independently or in an algorithm to aid clinical decision making. Although a logistic regression equation is a common mathematical/statistical procedure used in such cases and is preferred in the context of the present invention, other mathematical/statistical procedures can also be used.

In a first aspect the invention describes a method of diagnosing or differentiating stroke comprising measuring the level of one or more of the biomarkers h-FABP, D-dimer, sTNFRI and IL-6 in a sample taken from a patient who has suffered or is suffering stroke-like symptoms and comparing the measured level to a corresponding reference value characterised by the reference value being derived from a stroke mimic patient cohort and/or a TIA patient cohort; in a preferred embodiment the reference value is derived from either a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort. Use of the words/terms ‘level’, ‘measured level’ and ‘reference value’ etc. in the singular form of noun in the various aspects of the invention that refer to ‘one or more of..”, also signifies the plural noun forms ‘levels’, ‘measured levels’ and ‘reference values’ etc. The reference value or control value is the level or concentration of the biomarker (h-FABP, D-dimer, sTNFRI or IL-6) measured in samples taken from (i.e. an ex vivo or in vitro sample) patients (subjects) of a cohort (group) who are suffering or have recently suffered from a stroke mimic event and/or TIA. It is preferred that the reference value is derived from samples taken from a combined stroke mimic and TIA patient cohort;

this allows ischaemic stroke to be differentiated from stroke mimic events and TIAs and supports a more informed decision on the administration of thrombolytic therapy (e.g. tissue plasminogen activator or ‘TPA’) to IS patients. In a further embodiment the reference value is derived from samples taken from a stroke mimic patient cohort; this enables the more accurate discrimination of IS and stroke mimic patients. In all of the biomarker-based methods the diagnosis of the patient’s condition the diagnosis can be improved through a combination of information derived from the biomarker measurements and clinician examination and the application of authoritative clinical stroke diagnosis guidelines (see Methods). The level of the biomarkers h-FABP, D-dimer, sTNFRI and IL-6 have all been found to increase in patients suffering from or have recently suffered IS compared to patients suffering from or have recently suffered a stroke mimic or TIA. As can be seen from Tables 1 (column 2) and 2, the levels of each of the biomarkers is increased in the samples of patients who are healthy (HC) compared to patients who have suffered or who are suffering a stroke mimic event or TIA (data not is shown but HC vs mimics + TIA also significantly differs for each of h-FABP, D-dimer, sTNFRI and IL-6). Thus by using a reference value from ex vivo samples derived either from a stroke mimic patient cohort, TIA or from a combined stroke mimic and TIA patient cohort a more accurate patient assessment of an IS event can be obtained. Thus in a further aspect the invention describes the diagnosis or differentiation of stroke comprising measuring the level of one or more of the biomarkers h-FABP, D-dimer, sTNFRI and IL-6 in a sample taken from a patient who has suffered or is suffering stroke-like symptoms and comparing the measured level or levels to a corresponding reference value which is derived from a stroke mimic patient cohort, a TIA patient cohort or a combined stroke mimic +TIA patient cohort, in which the reference values are about 40%, 60%, 63% and 6% greater for h-FABP, D-dimer, sTNFRI and IL-6 , respectively, than a corresponding HC value; the slight fluctuation of measured biomarker values due to factors such as assay platform, inter-run variability, interoperator variability etc. is well-known in the art. A further aspect of the invention is a method of diagnosing or differentiating stroke in a patient who has suffered or is suffering stroke-like symptoms comprising taking a sample from the patient and determining the sample concentration of one or more of h-FABP, D-dimer, sTNFRI and IL-6 in which the following concentrations (‘cut-off’ concentration) taken individually support the diagnosis of IS in the patient: h-FABP > about 4.70 ng/ml, D dimer > about 200 ng/ml, sTNFRI > about 2.30 pg/ml, IL-6 > about 5.90 pg/ml. These cut-offs are displayed in Table 2 and are based on the ROC curves derived using stroke mimics and stroke mimics + TIA populations as the control group. The problem using healthy controls as the reference population to derive the ROC curve is that the cut-off values are reduced resulting in an increase in the number of false stroke diagnoses i.e. an increase in stroke mimic and TIA patients being diagnosed as IS. In all methods and uses of the invention any suitable patient sample type such as urine, saliva, tears, hair, cerebrospinal fluid (CSF) can be used but preferably the sample type is blood, serum or plasma.

In another aspect the invention describes a method of determining whether to administer thrombolytic therapy to a patient who has suffered or is suffering strokelike symptoms comprising ruling out haemorrhagic stroke and measuring the level of one or more of h-FABP, D-dimer, sTNFRI and IL-6 in a sample taken from the patient and comparing it to a reference value in which an amount of one or more of h-FABP, D-dimer, sTNFRI and IL-6 which is greater than the reference value is indicative of an ischaemic stroke and therefore administering thrombolytic therapy, in which the reference value is derived from a stroke mimic patient cohort or a stroke mimic and/or a TIA patient cohort; a stroke mimic or a stroke mimic + TIA reference value is especially preferred. HS can be ruled out using one or more of a patient sample GFAP biomarker measurement, a neuroimaging technique such as a CAT scan or MRI and face to face diagnosis by clinician. If as a result of carrying out the method of the invention it is determined that the patient has not suffered, or is not suffering, an IS or HS, further investigations can be made and blood thinners such as warfarin and aspirin may be prescribed and administered if necessary i.e. if TIA is diagnosed. If as a result of carrying out the method of the invention it is determined that the patient has suffered, or is suffering, a HS then these patients would typically be sent to a surgical unit to repair the damaged blood vessels.

Use of a stroke mimic and/or a TIA patient cohort sample biomarker measurement value as a reference value in an assay for stroke diagnosis or differentiation, especially IS diagnosis or differentiation using a stroke mimic or stroke mimic + TIA patient cohort reference value is thus clearly outlined, especially in relation to the biomarkers h-FABP, D-dimer, IL-1 and sTNFRI.

Methods

Patient Group

The study consisted of 113 patients displaying stroke-like symptoms admitted to the emergency department of KAT general hospital, Athens, Greece. Only patients presenting with stroke for the first time were included. Patients with malignancies, recent history of traumatic brain injury, chronic or end stage liver and renal disease were excluded. Blood samples (serum and EDTA plasma) were taken at the time of admission and at 24, 48, 72, 96 hours and 7 days post admission. Serum and plasma were aliquoted and stored at -80 °C until tested. All patients underwent CT scan upon admission and at 72hrs after the stroke onset. Only patients admitted to the emergency department within 6 hours from the onset of neurological symptoms were included in the study. The mean time from the onset of stroke-symptoms to hospital admission was 3.2 hours. In each patient the presence of arterial hypertension, diabetes mellitus, smoking history, dyslipidemia and ischaemic heart disease were recorded. Stroke severity was measured at the time of admission with the National Institute of Health Stroke Scale (NIHSS). Functional outcome was measured with the modified Rankin scale (mRS) on day 7 and acute stroke patients were categorised into three severity groups (mild, moderate and severe) according to their mRS-score: mild (mRS-score:0-2), moderate (mRS-score:3-4) and severe (mRS-score:5-6). Clinical evaluation (using criteria highlighted in the background section above) and neuroimaging techniques identified 10 haemorrhagic stroke (HS), 53 ischaemic stroke (IS), 13 transient ischaemic attack (TIA) and 37 stroke mimics (M); 79 healthy subjects served as controls (C).

Sample Analysis

The following proteins were tested against EDTA plasma samples of blood obtained from the patients of the study group: h-FABP, D-dimer, GFAP, NDKA, IL-6, PARK-7, GSTPi and sTNFRI. The proteins were detected and quantified using multiplexed biochips incorporating biomarker-specific antibodies and the Evidence Investigator (Randox Laboratories Ltd, Crumlin, UK). The simultaneous immunoassays were performed according to manufacturer’s instructions. A nine-point calibration curve and three reference controls were assayed for each biomarker to allow validation of results.

Statistics

Analyses were effected using Graph Pad Prism and SPSS. Raw data was log transformed and an unpaired 2-tailed t-test effected for cohorts with non-significant variance differences, otherwise a Mann-Witney on raw data was effected for each biomarker to compare healthy patient samples and stroke mimic patient samples (column 2). ROC curves were effected to derive sensitivity (Sens), specificity (Spec) and AUC values for IS values vs mimic values and IS values vs mimic + TIA values if healthy control and mimic levels were significantly different. Randox forest was effected for the analysis of biomarker combinations.

Results

Table 1 Statistical analysis of healthy control cohort (HC) vs stroke mimic cohort (mimics) and ischaemic stroke cohort vs HC, mimics and TIA + mimics.

Biomarker	Bivariate anal. HC vs mimics	ROC IS vs HC	ROC IS vs mimics	ROC IS vs TIA + mimics
		Sens Spec AUC	Sens Spec AUC	Sens Spec AUC
h-FABP	P = 0.0032	85% 95% 0.9315	58% 89% 0.8389	56% 80% 0.7665
D-dimer	P < 0.0001	85% 95% 0.9667	69% 89% 0.8559	61% 80% 0.7922
GFAP	P = 0.5331	NA	NA	NA
GSTPi	NA	NA	NA	NA
IL-6	P = 0.0008	61% 95% 0.9126	56% 76% 0.7457	56% 74% 0.7074
NDKA	P = 0.0839	NA	NA	NA
PARK-7	P = 0.4143	NA	NA	NA
STNFR1	P = 0.0058	70% 95% 0.8881	63% 89% 0.8146	61% 80% 0.7587

Comparison of biomarker levels between healthy controls vs stroke mimics show that the stroke mimic cohort have significantly higher levels of h-FABP, D-dimer, IL-6 and sTNFRI than the healthy control cohort. ROC curve analysis for individual markers shows that using the stroke mimic cohort as the comparison group as opposed to healthy control the assay sensitivity is reduced; however, from a perspective of clinical utility, crucially, the number of false positives is reduced, resulting in unwarranted and potentially health-undermining thrombolytic drug therapy being avoided. The AUC of the assay using mimic or TIA + mimic as the reference value can be increased by using two or more biomarkers disclosed in Table 1 for IS diagnosis e.g. use of h-FABP + D-dimer gives an AUC of 0.8623, h-FABP + D-dimer + PARK-7 gives an AUC of 0.8930 and for IS vs mimic and h-FABP + D-dimer + PARK-7 gives an AUC of 0.8101 for IS vs mimic + TIA. GFAP is a specific biomarker of HS and can be incorporated in an analysis of stroke or suspected stroke, enabling HS to be ruled in or ruled out; alternatively, or in addition a brain scan can be used to rule in or rule out HS. Rule out of HS allows IS assessment incorporating a reference value derived from stroke mimic values and/or TIA values using one or more of the stroke biomarkers of Table 1. Although measurement of GSTPi, PARK-7 and NDKA individually using a reference value derived from stroke mimic values and/or TIA values as opposed to healthy controls (HC) shows no additional benefit (HC vs mimic levels P> 0.05 Table 1), for practical purposes use of the stroke mimic and combined stroke mimic + TIA patient cohort levels as the reference value(s) are preferred and enable their ready incorporation into a diagnostic biomarker group for improved IS diagnosis.

Table 2 Biomarker median values (units in first column) of HC, mimic and mimic + TIA cohorts and % increase (inc.) of HC vs mimic and HC vs mimic + TIA. Column 4 displays cut-off values used to derive the sensitivities and specificities highlighted in Table 1

Biomarker	HC	Median	HC	Median mim + TL	4 inc.	Cut-off
mimic	inc.	IS vs mim	IS vs mim+TIA
h-FABP (ng/ml)	1.57	2.24	43%	1.57	2.70	72%	4.78	5.00
D-dimer (ng/ml)	46	77	67%	46	102	122%	201	217.6
IL-6 (pg/ml)	1.00	1.71	71%	1.00	3.02	202%	5.98	6.07
sTNFRI (ng/ml)	1.48	1.59	7%	1.48	1.66	12%	2.37	2.55

Claims (11)

1. A method of diagnosing or differentiating stroke comprising measuring the level of one or more of the biomarkers h-FABP, D-dimer, sTNFRI and IL-6 in a sample taken from a patient who has suffered or is suffering stroke-like symptoms and comparing the measured level to a corresponding reference value characterised by the reference value being derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort.

2. The method of claim 1 in which the stroke being diagnosed or differentiated is an ischaemic stroke.

3. A method of either of the preceding claims in which the one or more biomarkers measured is h-FABP and its value is greater than a h-FABP reference value derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort.

4. A method of either of claims 1 and 2 in which the one or more biomarkers measured is D-dimer and its value is greater than a D-dimer reference value derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort.

5. A method of either of claims 1 and 2 in which the one or more biomarkers measured is sTNFRI and its value is greater than a sTNFRI reference value derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort.

6. A method of either of claims 1 and 2 in which the one or more biomarkers measured is IL-6 and its value is greater than an I L-6 reference value derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort. .

7. A method of any of the preceding claims wherein the sample taken from the patient is a blood, serum or plasma sample.

8. A method of determining whether to administer thrombolytic therapy to a patient who has suffered or is suffering stroke-like symptoms comprising ruling out haemorrhagic stroke and measuring the level of one or more of h-FABP, D-dimer, sTNFRI and IL-6 in a sample taken from the patient and comparing it to a reference value in which an amount of one or more of h-FABP, D-dimer, sTNFRI and IL-6 which is greater than the reference value is indicative of an ischaemic stroke and therefore administering thrombolytic therapy, the method characterised by the reference value being derived from a stroke mimic patient cohort or a combined stroke mimic and TIA patient cohort.

9. Use of a stroke mimic or a combined stroke mimic and TIA patient cohort sample biomarker measurement value as a reference value in an assay for stroke diagnosis or differentiation.

10. The use of claim 9 in which the biomarker is h-FABP, D-dimer, IL-6 or sTNFRI.

11. The use of either of claims 9 and 10 in which the assay is for ischaemic stroke diagnosis or differentiation.