Temporal Artery Biopsy vs ULtrasound in Diagnosis of GCA (TABUL)

Overview

Giant Cell Arteritis (GCA) causes inflammation and narrowing of blood vessels and can cause blindness in one third of patients. It is important that a prompt, accurate diagnosis of GCA is made and treatment given as steroids for two or more years. Currently there is no 100% accurate test for GCA. Patients usually have new headache and scalp tenderness, typically with an abnormal blood test. However, it can be difficult to distinguish non-serious forms of headache from GCA; infection produces similar abnormal blood results. If there is a suspicion of GCA, treatment with steroids is started straight away. To confirm a diagnosis, the patient will need a biopsy of a temporal artery (a minor procedure performed under local anaesthetic to remove a sample of one of the scalp arteries). However, up to 44% of patients will have a normal biopsy. Therefore it is difficult to know if a patient with a normal biopsy does or does not have GCA. Withdrawing steroid treatment may increase the risk of blindness. Continuing treatment in a patient without GCA increases the risk of side effects (e.g., weight gain, infection risk, osteoporosis and fracture risk, high blood pressure, diabetes, cataracts). It is important to improve diagnostic tests for GCA. Another test to help in diagnosing GCA is an ultrasound scan of the arteries in the side of the head and under the arms. Ultrasound does not involve surgery; it is a simple test which can be performed as an out patient. Gel is applied to both sides of the head and under each arm. A sound probe is placed over the artery at each site to produce the scan. The investigators' study will examine the role of ultrasound in diagnosis of 402 patients with suspected GCA. All patients will have an ultrasound examination in addition to biopsy within a week of starting steroids. Patients will be treated according to usual practice. After six months, the investigators will reassess the diagnosis. The investigators will look at the accuracy of ultrasound compared with or combined with biopsy. The investigators will look at how a doctor's knowledge of ultrasound results or biopsy results alone would affect the diagnosis and recommendation to continue or stop steroid treatment. The investigators will assess whether knowledge of both results together would alter the diagnosis and treatment. The investigators will collect information to estimate the costs of different ways of diagnosing GCA in relation to the impact on quality of life.

OVERALL DESIGN: The overall design consists of a cohort study of 402 participants with suspected GCA who will be followed up for 6 months; a cost-effectiveness study comparing ultrasound with temporal artery biopsy; a study of observer agreement in evaluating ultrasound and temporal artery biopsies; and an expert panel assessing the appropriateness of alternative strategies for diagnosing and treating patients with suspected GCA. PRIMARY OBJECTIVES: 1. To evaluate the diagnostic accuracy (sensitivity and specificity) of ultrasound as an alternative to temporal artery biopsy for the diagnosis of GCA in patients referred for biopsy with suspected GCA. 2. To evaluate the cost-effectiveness (incremental cost per QALY) of ultrasound instead of biopsy in the diagnosis of GCA. SECONDARY OBJECTIVES: 3. To evaluate inter-observer agreement in the assessment of ultrasound and temporal artery biopsy. 4. To elicit expert views on the appropriateness of performing a biopsy following ultrasound using clinical vignettes. 5. To evaluate the diagnostic accuracy (sensitivity and specificity) of the sequential diagnostic strategy from 4 as an alternative to temporal artery biopsy alone in the diagnosis of GCA. 6. To evaluate the cost-effectiveness (incremental cost per QALY) of the diagnostic strategy from 4 instead of biopsy alone in the diagnosis of GCA. DESIGN: A prospective cohort study to evaluate the impact of ultrasound or biopsy of temporal arteries on diagnosis of GCA and treatment decisions . The cohort study will use a paired design, i.e. all participants will have both US and biopsy, with diagnostic performance assessed against a composite reference standard diagnosis following final (6 months) assessment. To evaluate the impact of US/biopsy results on clinical practice and longer term outcomes, we will derive clinical vignettes based on cases recruited to the study, and present them to the treating clinician, along with either the ultrasound, or biopsy or both results, so that they can indicate diagnosis and proposed treatment. The main cost-effectiveness analyses will evaluate incremental cost per quality adjusted life year (QALY) on a long term (lifetime) horizon between diagnostic strategies. SETTING: Outpatient and inpatient rheumatology and ophthalmology departments in 25 National Health Service (NHS) trusts in the United Kingdom (UK) (also sites in Europe of required) with access to high resolution US. TARGET POPULATION: Patients with suspected GCA who would normally require an urgent temporal artery biopsy, i.e. referrals from primary care and suspected GCA identified in secondary care. Recruitment of participants will be restricted to patients for whom US and biopsy can be performed within 7 days of starting high-dose steroids. HEALTH TECHNOLOGIES BEING ASSESSED: Halo, stenosis, or occlusion assessed by high resolution US; presence of giant cells or granulomatous inflammation on temporal artery biopsy. MEASUREMENT OF COST AND OUTCOMES: Data collection at baseline, 2 weeks and 6 months will include clinical and laboratory markers of disease activity, resource use, health-related quality of life (HRQoL) using the EuroQol-5D (EQ-5D), and adverse events. Baseline assessment will include retrospective assessment of symptoms and erythrocyte sedimentation rate/C-reactive protein (ESR/CRP) results before starting steroids. The reference standard diagnosis will be made using a composite of American College of Rheumatology classification criteria, GCA-related events, and alternative diagnoses using data collected at all assessments. Proposed treatment data will be collected from participating clinicians after each test result is released, and classified as treatment for GCA (e.g. initiate/continue steroids) or not GCA (e.g. withdraw/rapid taper of steroids) to compare changes in proposed treatment and evaluate agreement with the reference diagnosis. Unit costs of resources used will be obtained from nationally published sources where available. Modelling will estimate the impact of diagnostic strategies on clinical outcomes (e.g. GCA complications and steroid related adverse events), their costs and impact on HRQoL beyond study follow-up and within study follow-up. Probabilities of events, their cost and impact will be obtained from study data, a systematic literature review, or in the absence of relevant data, by formal elicitation of expert opinion. Costs and benefits will be discounted at 3.5% (National Institute for Health and Care Excellence \[NICE\] guidance) and uncertainty (including modelling assumptions) subjected to probabilistic sensitivity analysis and scenario analysis. SAMPLE SIZE: A sample size of 402 patients provides 90% power at a 5% type I error rate to test the joint hypothesis that (i) US has greater sensitivity than biopsy (to detect an increase in sensitivity from 76% for biopsy (assuming a 0.24 false negative fraction based on 9-44% biopsy-negative GCA) to 87% sensitivity for US; and (ii) to detect specificity of US of no less than 0.83 based on an expected specificity of 0.96. This sample size will allow estimation of a one-sided rectangular confidence region for US false and true positive fractions, assuming 80% prevalence of GCA in patients having a biopsy for suspected GCA, with the sample size inflated (gamma=0.1) due to uncertainty in the proportion of cases/controls in a cohort design. We will actually recruit 430 cases to allow for possible drop-outs from the study. In addition we will recruit 270 individuals for training purposes, to allow each centre to learn the technique of temporal artery and axillary artery scanning. Each centre will collect 10 such individuals (training cases), who will be of similar age and gender as the study cohort, but who will not have temporal arteritis. This is very important in order to ensure that observers are trained to recognise the appearances seen in normal (non GCA arteries) especially patients with atherosclerosis. Further ultrasound training including a video exam (to identify images of both normal and abnormal features of GCA) and a 'hot' case assessment (scanning a patient with GCA) will be designed into the ultrasound training programme. In addition, we will provide adequate training days (e.g. 2 separate training days) when all observers will be trained formally by Dr Schmidt and other ultrasound experts to ensure adequate observer agreement. The first training day will be held prior to starting the recruitment of patients, and will be repeated after the first year. PROJECT TIMETABLE: Total: 48 months (UPDATED TO 60 MONTHS SEE BELOW) Month 1-6 Study materials/protocols prepared; Ethics and research governance approval complete; Centres trained, approved and ready to recruit. Month 7-12 Recruitment monitoring report; Quality control report. Month 13-18 Recruitment monitoring report; Quality control report; Additional centres (if required) trained, approved and ready to recruit; Month 19-24 Recruitment monitoring report; Quality control report. Month 25-30 Recruitment monitoring report; Quality control report; Web-based US and biopsy assessment developed. Month 31-36 Recruitment completed; Clinical vignettes (web-based) developed. Month 37-42 Follow-up completed; Inter-rater assessment of US and biopsy images analysis competed; Expert panel review of vignettes completed. Month 43-48 Database cleaned and locked; Analysis completed; statistical analysis and economic modelling, report drafting and preparation of papers. Final report completed. PLEASE NOTE: 12 MONTH EXTENSION WAS GRANTED BY THE FUNDER (NIHR HTA) IN SEPTEMBER 2012 TO EXTEND THE RECRUITMENT PERIOD FROM THE END OF DECEMBER 2012 TO END OF DECEMBER 2013 (A FURTHER 12 MONTHS). TH END OF STUDY WILL NOW BE DECEMBER 2014 The study will be overseen by a Trial Steering Committee (TSC) and an independent Data Monitoring Committee (DMC)at least 1 meeting per year. BIOBANK: We will develop a biobank of tissue, serum, DNA and ultrasound video imaging of blood vessels in GCA. An important benefit of the primary protocol is that we can use the accumulated material for a number of related projects.