Introduction: Atrial fibrillation (AF) disproportionately affects older people and the incidence of previously undiagnosed AF in long-term care (LTC) residents is high (approximately 14%). Prescription of oral anticoagulation (OAC) is integral to stroke prevention for AF but there is evidence of under-treatment in older people in LTC, likely due to clinicians’ concerns of iatrogenic harm and doubt over the net clinical benefit of pharmacological intervention. This systematic review reviews the prevalence and the net risk-benefit of OAC in the older population living in LTC.
Methods: Observational studies investigating the prevalence or outcomes of AF in LTC were identified from searching electronic databases (Ovid Medline, CINAHL, PsycINFO, Scopus, Web of Science) from inception to 31st October 2019. The OpenGrey repository was searched for unpublished literature/dissertations, complemented by hand-searching of two geriatric cardiology journals and Google Scholar. International Scientific Indexing conference proceedings were searched for conference abstracts and bibliographies of identified articles were reviewed for any additional relevant studies. Two authors independently identified relevant articles, performed data extraction and assessed the quality and risk of bias using the Newcastle Ottawa Scale. Disagreements were resolved by discussion with another reviewer. The protocol was registered with PROSPERO (CRD42020164963).
Results: After full-text review, 21 studies were identified which reported AF prevalence in LTC residents, ranging from 7.1%-38% (n=3 reported an AF prevalence <10%, n=13 a prevalence of 10-15%, and n=5 a prevalence >15%). There was no association between mean/median resident age (years) and prevalence of AF. The median resident age was 85 years [IQR 77-89] for the highest prevalence of AF (38%) and the mean resident age was 87.7 years [SD 6.5] for the lowest prevalence of AF (7.1%). Two studies reported on outcomes based on the prescription of OAC or not; one reported a reduction in ischaemic stroke event rate associated with OAC prescription (2.84 per 100 person years, 95% Confidence Interval (CI) 1.98-7.25 vs. 3.95, 95% CI 2.85-10.08, but a higher intracranial haemorrhage rate (0.71 per 100 person years, 95% CI 0.29-2.15 vs. 0.65, 95% CI 0.29-1.93). The second study reported a 76% lower chance of ischaemic stroke with OAC prescription after adjustment and a low incidence of bleeding (n=4 events) in residents on OAC. No studies examined the effect OAC type on outcomes.
Conclusions: Estimates of AF prevalence were inconsistent and varied extensively; this likely reflects heterogeneity in the methods of AF diagnosis, resident characteristics and type of LTC facility. The limited number of studies examining adverse outcomes of LTC residents prevents the drawing of any definitive conclusions. Whilst this observational data does provide some insight, in the absence of more rigorous study designs the risk-treatment paradox still needs addressing in this often-neglected population who are at high-risk of AF and adverse AF-related outcomes. We recommend more rigorous study designs augmented with routinely collected health and social care data.