Detecting and Diagnosing Atrial Fibrillation: enhanced opportunistic case-finding in high risk patients from general practice and evaluating three ways of irregular pulse detection. A cluster randomised trial with nested diagnostic studies.

FACTS The prevalence of atrial fibrillation (AF)doubles with each 10 years of life from 1% in people younger than 50 years old, to more than 10% in people over 80 years old. AF is a condition with important medical implications: increased mortality, reduced quality of life,an increased risk of heart failure and stroke (average 5x). A stroke caused by AF has a mortality 1,5-2x as high as a stroke not caused by AF and survivors have a poorer quality of life. Adequate antithrombotic treatment of AF decreases the risk of death by more than 25% and reduces the risk of stroke by half. In 20% of patients admitted to the hospital with a stroke, AF not previously discovered is diagnosed. The costs of care for stroke rank in the top ten of most expensive diseases, and comprise over 2% of total health care costs in the Netherlands. CHALLENGES Detection of AF is complicated by two factors: 25-60% of patients with AF are asymptomatic and AF often presents as paroxysmal disorder. Thus, symptoms are not always present. The chance of missing paroxysmal AF is large, especially in patients without symptoms. In paroxysmal AF, the risk of stroke is as large as for permanent or persistent AF and the effect of treatment also is equally large. It is estimated that one third to one half of the patients with AF is not known. Of the possible detection strategies, opportunistic pulse palpation - during consultations in patients over 65 years of age and in case of an irregular pulse performing a standard ECG - has been investigated in the SAFE study. The program was cost effective but had a relatively low yield. The 2010 ESC Guideline on AF as well as the 2009 Dutch NHG guideline recommend ‘opportunistic case finding’ but state there is insufficient evidence for this recommendation. Current guidelines state that the diagnosis AF should be confirmed with an ECG and if this is negative an ambulatory rhythm recording (Holter, automatic or patient-triggered event-recording) should be performed. The 'auto-triggered’ event recording (ATER) yields most extra arrhythmias and asymptomatic AF episodes, but no large studies on the effectiveness of any ambulatory rhythm registration with respect to the detection of AF in general practice have been published. This is acknowledged as a major evidence gap in the 2009 NHG and 2010 ESC guidelines. SOLUTIONS Innovative diagnostic methods for irregular pulse or AF detection have been developed. In this study we focus on an electronic sphygmomanometer that contains an algorithm allowing detection of AF during measurement of blood pressure (‘ eBPM-AF’), and on a hand-held device that captures a ‘one-lead' ECG (lead I). In (preliminary) studies both methods showed high sensitivity (97-100%) and specificity (90-100%). We combine these methods with special software built in the GP information system and the use of an ATER. STUDY PROPOSAL To bridge the evidence gaps, we perform a cluster-randomized trial to evaluate the extra yield in cases of AF of an intervention - special software indicating a patient at high risk for AF; a diagnostic protocol (pulse examination, standard ECG, if indicated ATER) - that enhances opportunistic pulse examination in high risk patients (hypertension, diabetes, or existing cardiovascular disease) and aged 65 years or older) who visit the GP’s office. Nested in this trial, three methods of irregular pulse detection - i.e. pulse palpation, eBPM-AF, and hand-held ECG - are compared for their diagnostic value to detect AF. The hand-held ECG device used by patients at home is compared with ATER. Finally, a cost-effectiveness analysis is performed. In Control Practices, delivering care ‘as usual’, all high risk patients are identified prior to the start of the study by the researchers. After one year the number of new cases of AF are determined (outcome) and diagnostic pathways are identified retrospectively (details: see ‘Strategy’). The primary outcome is the difference in detection rates of new cases of AF found in Intervention and Control practices, respectively. SAMPLE SIZE The estimated sample size is 11616 patients per group. Assuming a cluster size of 230 eligible patients/practice, we need 102 participating practice (details see‘Strategy'). COLLABORATION The study by the Departments of General Practice of MUMC (Maastricht) and the AMC (Amsterdam) builds on previous collaboration in the AMUSE projects (DVT and Pulmonary embolism) and guarantees a sufficient number of GP practices. Our combined GP networks have a potential of 180 fte GP. In both academic centres, an ‘AIOTHO’ (combined GP-traineeship and PhD) will be appointed for 6 years. The joint approach provides optimal opportunities for both candidates; alternating their activities and responsibilities in close alignment with their counterpart provides optimal continuity for the research project. The current application (€248.948) is to cover the costs of the Maastricht AIOTHO and part of the material costs.