Burden of atrial fibrillation risk and outcomes amongst patients attending cardiology and elderly medicine services: prospective in-hospital implementation of the FIND-AF algorithm

Abstract

Abstract Background Screening for silent AF is becoming increasingly important (1, 2), and the FIND-AF algorithms can use routinely-collected data to identify individuals at higher risk for undiagnosed AF (3). There may be a much higher prevalence of patients at risk for AF and its complications in a hospital environment compared with the community (4). Purpose We prospectively implemented the FIND-AF algorithm within cardiology and geriatric services to quantify the burden of AF risk amongst all-comers, and to understand their outcomes over the next 6 months. Methods We prospectively implemented the FIND-AF algorithm in the cardiology and geriatric departments of three UK hospitals for consecutive inpatient or outpatient attendances from February 1 to July 13th 2023. We assessed the number of individuals attending each service eligible for an AF screening protocol (exclusion criteria: age <30 years, history of AF, on a palliative care pathway, use of oral anticoagulation), and calculated the burden of AF risk in each service. Patients were followed up for 6 months for the primary outcome of incident AF, heart failure (HF), stroke/transient ischaemic attack (TIA)/systemic embolism, and death. Incident diagnoses were the first record of that condition in primary or secondary care records. Rate ratios (RRs) with 95% confidence intervals (CIs) for combined incident cardiovascular outcomes or death were derived through Poisson regression for predicted higher versus lower AF risk. Results Of 467 consecutive patients, 364 were eligible for AF screening. 66 were excluded due to a previous diagnosis of AF, 25 were already prescribed a long-term oral anticoagulant, 5 were on a palliative care pathway and 7 were aged <30 years. Baseline characteristics of each cohort are summarised in Table 1. Of patients eligible for AF screening, those attending geriatrics were at higher risk of incident AF than patients attending cardiology (median FIND-AF score: geriatrics, 0.015, 95% CI 0.008 - 0.024; cardiology, 0.005, 95% CI 0.001-0.015; p<0.05; distributions of risk summarised in Figure 1) and had a higher stroke risk. Absolute risk of AF was greater than 5% within the next 6 months for 11.6% and 9.1% of geriatric and cardiology patients, respectively. Of the eligible cohort, 328 had follow up at 6 months, with 46 experiencing the primary outcome (9, 7, 3, and 27 recorded cases of AF, HF, stroke/TIA, and death, respectively). The higher predicted risk cohort, compared to lower predicted risk, had a higher risk of combined incident AF, heart failure, or stroke/TIA/systemic embolism (Rate ratio: 5.59, 95% CI 1.13-101.01) and death (RR 9.68, 2.06-172.85). Conclusions The FIND-AF algorithm can be prospectively implemented in hospital records to identify cardiology and geriatric patients at risk of incident AF, heart failure, thromboembolic disease and death. There is an opportunity to conduct risk-guided screening in the context of hospital attendances.