Several studies on oral anticoagulation established the clinical benefit of warfarin anticoagulation over aspirin in patients at apparent intermediate risk of AF-related stroke and subsequently set the benchmark that under-treatment with oral anticoagulation was more harmful than overtreatment [44, 45]. In this context, the extended CHA₂DS₂-VASc score stroke risk index was developed by the Euro Heart Survey on AF which appeared to better differentiate truly low risk subjects who may not need antithrombotic therapy [46]. The CHA₂DS₂-VASc schema incorporated other known independent risk factors including the impact of female gender, documented vascular disease and the incremental risk per decade of age from age 65 years. The score allocates 2 points to high-risk factors of prior history of stroke, TIA or systemic embolism and also to age ≥75 years, with 1 point allocated to other risk factors. With clinical validation, the adjusted stroke rate per 100 patient-years ranged from 0 for CHA₂DS₂-VASc score of 0 to 15.2 in CHA₂DS₂-VASc score of 9 (see Table 1.2). The CHA₂DS₂-VASc score was incorporated into revised clinical guidelines for the management of AF from 2010 [47].

Despite evidence from randomised trials for the benefit of oral anticoagulation for at-risk patients with AF, multiple studies have shown that warfarin is prescribed in only about half of appropriate patients [48]. The risk of bleeding due to anticoagulation is cited among the most common concerns of physicians and the reasons for under-prescription [49]. A novel bleeding risk score was developed from multivariate analysis of a large number of European patients studied in the Euro Heart Survey on AF to better quantify bleeding risk in this population. The HAS-BLED score has been shown to have good predictive accuracy with a score of ≥3 associated with a high risk of major bleeding per 100 patient-years [50]. The score allocates 1 point each to risk factors shown to independently increase the 1 year risk of major bleeding events (defined as including intracranial, requiring hospitalisation, haemoglobin decrease >2 g/L and/or transfusion requirement). The risk factors include (H) uncontrolled hypertension (systolic >160 mmHg), (A) 1 point each allocated for abnormal liver or renal function, (S) prior history of stroke, (B) previous bleeding history or predisposition to bleeding, (L) labile INRs including unstable/high INRs or poor time in the therapeutic range, (E) elderly (age >65 years) and (D) drugs/alcohol concomitantly which allows for 1 point for concomitant use of drugs which increase bleeding risk, e.g. antiplatelet agents or non-steroidal anti-inflammatory drugs and 1 point for alcohol abuse (see Table 1.3). The HAS-BLED score was incorporated into revised clinical guidelines for the management of AF from 2010 [47].

Data from the Framingham Study showed that AF-related stroke was associated with increased stroke severity, poorer survival, greater disability among survivors and a higher recurrence rate of stroke as compared with other etiologies [51]. Other population studies have confirmed these findings.

The North Dublin Population Stroke Study found that AF was associated with a higher frequency of total and partial anterior circulation infarct syndromes and greater acute stroke severity, which mediated poorer functional incomes and greater disability at 90 days than non-AF strokes [52]. The Framingham Study found that by 3 months after stroke, 75 % of AF subjects remained moderately or severely dependent in ADLs (activities of daily living) [51].

AF-related stroke is also associated with higher mortality rates—the Framingham Study found a 30-day mortality of 25 % vs. 14 % for non-AF stroke and 1-year mortality of 63 % compared with 34 % for non-AF subjects. Similar findings were reported by an Italian population-based study with 32.5 % (vs. 16.2 % for non-AF) 30-day mortality and 49.5 % (vs. 27.1 % for non-AF) 1-year case fatality rates [53].

A large Japanese multicentre stroke study (J-MUSIC) published in 2005 which included data on thrombolytic therapy for superacute phase ischaemic stroke treatment also found a significantly higher 28-day mortality rate for AF-related stroke (11.3 %) than non-AF (3.4 %) [54]. During the study, 7.3 % of the AF group and 1.3 % of the non-AF group received thrombolytic therapy. Longer hospital stays were recorded in AF patients (mean 40.5 days) as compared with non-AF (mean 35.3 days). Stroke severity scores were significantly higher than in the non-AF group. After hospital discharge, 66.4 % of non-AF patients returned to their homes, whereas only 45.1 % of AF patients returned home and 54.9 % were sent to an institution for care.

One-year recurrence rates were also higher for AF-related stroke in The Framingham Study at 23 % vs. 8 % for non-AF [51], and in the Italian Registry 6.9 % vs. 4.7 % for non-AF [53], although the rates of anticoagulation following the initial stroke are not known for the relative studies.

It is also notable that AF is associated with a higher likelihood of presentation with intracerebral or intracranial haemorrhage—including spontaneous as well as associated with prescribed antithrombotic therapy. In the North Dublin Study, 9 % of all AF-associated strokes were hemorrhagic with an equal distribution occurring spontaneously and in patients on oral anticoagulation [52]. This observation was also confirmed by a recent USA Health Insurance Database analysis which found higher rates of intracerebral and intracranial haemorrhage in AF cohorts presenting with stroke [55]. Rates of anticoagulation at the time of stroke were 43.5 % for the AF coh ort [55].

The economic implications of AF stroke-related healthcare costs have been assessed in various studies around the world. The results from a 3-year Swedish Stroke Registry found that 3-year inpatient costs (including hospitalisation for index stroke and any stroke-related hospitalisation events including recurrent stroke) was 10,192 € for AF patients who survived the index stroke compared with 9374 € for non-AF subjects based on 2001 prices (exchange rate = US$0.90) [56]. Length of stay for the index event was longer in AF patients 22.4 days vs. 20.9 days for non-AF. A higher re-stroke rate was observed over the 3-year period for AF at 15 % vs. 13 % for non-AF. Patients with AF age <65 years were noted to generate significantly higher hospital costs (on average 4412 € higher) than similarly younger patients without AF—the difference in cost was noted to decrease with advancing age.

The Berlin Acute Stroke Study published data relating to AF—stroke-related use of all medical resources over a 12-month period (inpatient and outpatient) as well as indirect costs which included lost work productivity [57]. The findings showed that AF acute stroke patients consumed more medical resources than non-AF patients over the initial 12-month period 11,799 € vs. 8817 € based on 2005 prices (exchange rate = US$1.32), driven primarily by longer lengths of hospital stay and increased use of home nursing care. Indirect costs due to lost productivity from work absenteeism or early retirement were higher for the non-AF group, which was explained by the younger average age and higher employment rates 63.9 years of age with 30 % gainfully employed, as compared with the average age of AF patients 73.7 years and only 3 % employment rate.