Malini Madhavan (Department of Cardiology, Mayo Clinic, Rochester, Minnesota, USA) and others reported the results of a study in American Heart Journal that found that patients with atrial fibrillation who have cognitive impairment or are frail were less likely to be given oral anticoagulant treatment despite having a higher rate of mortality and being assessed at being at higher risk of stroke and bleeding.
Previous studies, such as those of Holt et al and Mohammed et al, have reported lower rates of oral anticoagulant prescribing in patients with atrial fibrillation who have dementia. However, the use of oral anticoagulation has been shown by Jacobs et al to reduce the incidence of cognitive impairment in patients with atrial fibrillation.
Similarly, Bertozzo et al found a high rate of doctors stopping warfarin in patients with atrial fibrillation who they perceived as being frail or having low life expectancy, despite high rates of mortality, bleeding and stroke in these patients after discontinuing anticoagulation. It was not known whether frailty affected the outcomes of anticoagulation sufficiently to recommend this practice of withholding treatment.
With this study, therefore, Madhavan and colleagues aimed to investigate whether the incidence of cognitive impairment and/or frailty impact therapy and outcomes in atrial fibrillation. The authors analysed the Outcomes Registry for Better Informed Care in AF for patients with atrial fibrillation who have cognitive impairment and/or frailty and examined the association with oral anticoagulation in determining outcomes.
Of a total of 9,749 patients with atrial fibrillation who qualified for the study, 293 (3%) were diagnosed with cognitive impairment, 575 (5.9%) were identified with frailty and 67 (0.7%) were diagnosed with both cognitive impairment and frailty. Persistent or permanent atrial fibrillation was more common in patients with frailty than those without (54.1% vs. 44.2%). Oral anticoagulants were prescribed to 7,445 patients (76.4%) at baseline: of these 6,965 were given warfarin and the remaining 480 were given dabigatran.
Using risk calculators, patients with cognitive impairment or frailty were assessed at being at higher risk of stroke and bleeding. However, oral anticoagulants were found to be less likely to be given to patients who were frail (68% vs. 77%, p<0.001) or to those who had cognitive impairment (70% vs. 77%, p=0.006).
The risk of dying was found to be higher in patients with cognitive impairment (HR: 1.34; 95% CI: 1.05–1.72; p=0.0198) and in patients who were frail (HR: 1.29; 95% CI: 1.08–1.55; p=0.006). However, there was no association identified between cognitive impairment or frailty and stroke, transient ischaemic attack or major bleeding.
When assessing the association between oral anticoagulation and different outcomes, there was found to be no interaction between either cognitive impairment or frailty and oral anticoagulant use in determining mortality, major bleeding or composite end point of stroke, non-central nervous system systemic embolism, TIA, myocardial infarction or cardiovascular death. The authors comment: “the treatment effect of oral anticoagulation on clinical outcomes is similar between frail and non-frail patients and similarly between cognitively impaired and non-cognitively impaired patients.”
The authors highlight a strength of this study: “a significant proportion of patients over the age of 75 years, a population that is underrepresented in randomised clinical trials.”
They also note some limitations of the study. As an observational study, it may be subject to unrecognised confounders so “these findings should be confirmed in future prospective studies”. The numbers of frail or cognitively impaired patients in the study are small so the results should be interpreted with caution and may not be generalisable to other populations. It is possible that cognitive impairment and frailty were underdiagnosed in the original patient population and/or that new diagnoses of cognitive impairment or frailty were not performed during the follow-up period. There was a small percentage of patients treated with NOACs, so the data are only applicable to patients treated with warfarin.
On considering the implications of the study, Madhavan and colleagues recommend that: “atrial fibrillation patients with cognitive impairment should be considered for oral anticoagulation based on their clinical factors.” They also suggest that “frailty should not be a contraindication to oral anticoagulation prescription” in patients with atrial fibrillation. They propose that future studies should consider whether routine assessment of frailty indices should be used to guide atrial fibrillation therapy.