Anticoagulation to prevent ischaemic stroke in patients with atrial fibrillation: a complex scenario including underdiagnosis, undertreatment, or underdosing of oral anticoagulants

This editorial refers to ‘Trends in preadmission oral anticoagulant use and clinical outcome in atrial fibrillation patients admitted with acute stroke in Denmark’, by L.K. Larsen et al., on page 112. The availability of non-vitamin K antagonists oral anticoagulants (NOACs), as an alternative to vitamin K antagonists (VKAs), for primary and secondary prevention of stroke in patients with atrial fibrillation (AF), constituted an important step for widening the possibilities of effective thromboembolic prophylaxis in daily clinical practice.1,2 The proportion of AF patients treated with oral anticoagulants increased in Europe after the approval of NOACs,3 with the result that, in recent observational studies, ∼85% of AF patients at risk of stroke were actually treated with a NOAC or a VKA.2 Clinical trials and observational studies evaluating NOACs and VKAs agents showed a residual yearly risk of ischaemic stroke for patients treated with oral anticoagulants between 1.0% and 2.5%, with a yearly risk of intracerebral haemorrhage between 0.2% and 1.5%.1,4 Following initiation of an oral anticoagulant, patient adherence to treatment and maintenance of the prescription by physicians becomes an important determinant of effective thromboembolic prophylaxis and this issue surely deserves more attention within the community of patients and physicians, since interruptions of anticoagulants, sub-optimal adherence, or shifts in treatments were not uncommon in observational studies at long-term follow-up.5 Larsen et al.6 report on the prevalence of oral anticoagulant treatment among patients admitted in Denmark with a first episode of acute stroke considering both patients with previously known AF and patients with AF detected at the time of admission. In this analysis, derived from the National Danish Registry, the authors focus also on stroke severity, length of hospital stay, and 30-day mortality. The results of this very interesting study are related to around 15 000 patients with AF admitted with an acute stroke between 2008 and 2016.6 In this patient population, the proportion of oral anticoagulant users was overall 44%, with a clear temporal trend of increasing use of anticoagulants in recent years, in parallel with availability in daily practice of NOACs. As a matter of fact, while the proportion of VKAs users was almost constant from 2008 to 2016, the net increase in oral anticoagulant users, especially among patients with a previous admission with AF, was actually due to the prescription of NOACs. It is noteworthy that the increased use along with the time of oral anticoagulants was associated with a net reduction in the proportion of patients inappropriately treated with an antiplatelet in monotherapy, in line with other observations from European data sets.3 Moreover, the proportion of strokes associated with AF and classified as severe declined along with time, in parallel with the relative increase in patients treated with oral anticoagulants, and it is reassuring that, among patients admitted with intracerebral haemorrhages, stroke severity and mortality rates did not increase for patients on treatment with oral anticoagulants.6 These findings have important implications for health economics because one of the consequences of the increased proportion of AF patients under anticoagulant treatment was that length of hospitalization, a major determinant of healthcare system costs, declined for patient with ischaemic stroke but did not increase for patient with intracranial haemorrhage. An additional meaningful finding was the reduction in 30-day mortality observed along with time, more pronounced in case of ischaemic stroke occurring in users of oral anticoagulants. In this study, the wider use of oral anticoagulants had no adverse impact on the 30-day mortality of patients presenting with intracerebral haemorrhage and this is surely an encouraging finding. The effect of oral anticoagulants in patients with AF at risk of stroke has been extensively studied through the most sophisticated approaches of evidence-based medicine and is one of the strongest benefits achieved in cardiovascular medicine.1 Consequently, a great interest emerged on the possibility to effectively translate the findings of evidence-based medicine into real-world practice with substantial benefits at the level of patient populations. A series of studies were also planned for assessing potential gaps in applying to AF patients the indications to oral anticoagulation validated by randomized controlled trials. In this perspective, the study by Larsen et al.6 allows to focus on a very specific high-risk population, promoting important clinical considerations. In the discussion of their findings, Larsen et al.6 propose that the proportion of patients presenting with acute stroke who result to be on pre-admission oral anticoagulant treatment should be used as a measure of appropriate care, but it is clear that this assessment can be applied only to patients with known clinically overt AF. According to the literature, in 11.5–24% of all the patients presenting with acute ischaemic stroke or transient ischaemic attack (TIA), the stroke/TIA is actually the first clinical documentation of AF,7 thus highlighting the importance of detecting asymptomatic AF. According to this background, it is clear that the scenario reported by Larsen et al.6 will substantially improve if the proportion of subjects in the community with undiagnosed AF, usually corresponding to asymptomatic or pauci-symptomatic AF, may actually decrease, through targeted initiatives of AF screening. In the literature, asymptomatic AF was more common in male subjects and in the elderly and was associated with adverse outcome implications at long term, in terms of risk of death8 and risk of stroke.9 Screening for AF in patients at risk has been proposed for identifying subjects with asymptomatic, or anyway undiagnosed AF, and opportunistic screening at the time of medical checks is currently recommended by the European Society of Cardiology guidelines for patients aged ≥65 years.10,11 Several methods can be employed for AF screening, such as pulse palpation, modified blood pressure monitors, handheld single-lead ECG, wearable devices, and smartphones.11 The goal of AF screening is to discover AF even in an asymptomatic form and to prescribe anticoagulants according to consensus guidelines; therefore, any screening initiative should include appropriate patient information and specific clinical pathways for a complete cardiovascular assessment, followed by oral anticoagulants prescription in patients at the risk of stroke in whom AF is definitely found and confirmed.11,12 The effect of oral anticoagulants in screen-detected AF has not been validated by randomized controlled studies, but incidentally discovered AF was found associated with a substantial risk of stroke and death in cohort studies, with a significant lower occurrence in patients treated with oral anticoagulants.13 The meaningful implications of community screening for AF are also supported by the positive outcome found in patients aged 75 years, in whom AF was discovered during a systematic screening initiative with intermittent ECG recordings for 2 weeks performed in Sweden.14 In this study, adherence to prescribed oral anticoagulants was high at 5-year follow-up and the occurrence of ischaemic stroke was significantly lower in the geographical area object of screening as compared with a surrounding control area.14 Screening for enhancing AF discovery in patients at the potential risk of stroke is currently a topical issue,7,11,12 and the article by Larsen et al.6 clearly outlines that it can be an important strategy for reducing the treatment gaps that still exist in the prevention of AF-related stroke. However, as the authors stress,6 even patients with well-known AF can be undertreated because, according to the Danish registry, >35% of the patients previously admitted to a hospital with AF were still not receiving an oral anticoagulant at the time of stroke in year 2016. Many factors or conditions, summarized in Table 1, can be associated with patients’ and/or physicians’ perceptions leading to a reduced attitude towards the use of anticoagulants in AF patients at the risk of stroke according to the CHA2DS2VASc score. Some of these factors clearly emerged also in the analysis by Larsen et al.6 where the use of antiplatelets and a high HAS-BLED were powerful predictors of the lack of use of oral anticoagulants. In many cases, perceptions rather than facts are at the basis of decisions, or lack of decisions on the institution of oral anticoagulants despite accepted indications, with the risk that the main driver corresponds to the fear of bleeding rather than the more balanced concept of predicted net patient benefit. Factors and conditions associated with patients’ and/or physicians’ perceptions leading to a reduced attitude towards use of anticoagulants in atrial fibrillation patients at risk of stroke according to CHA2DS2VASc score Factors and conditions associated with patients’ and/or physicians’ perceptions leading to a reduced attitude towards use of anticoagulants in atrial fibrillation patients at risk of stroke according to CHA2DS2VASc score An additional problem, not evaluated by Larsen et al.,6 is the use of NOACs at inappropriately low dosages, a practice not supported by the evidence of randomized controlled trials, where patients were treated with well-defined dosages and regimens for any specific tested drug.1 The use of NOACs at inappropriately low dosages is usually conditioned by the same factors reported in Table 1. A recent systematic review15 clearly outlines that underdosing of NOACs can be a frequent problem in daily clinical practice with the result of incomplete protection from the risk of stroke and increased rate of hospitalization, as shown by a series of observational studies. Many physicians may actually be in charge of the care of a patient with AF, commonly associated with many comorbidities, and a stricter collaboration between specialists and general practitioners can be a way to improve appropriateness in the prescription of oral anticoagulants in AF patients at risk. Moreover, this should be combined with appropriate patient education and engagement to maximize adherence to prescribed treatments and obtain the full spectrum of benefits currently achievable in the prevention of AF-related adverse outcomes. Conflict of interest: G.B. declares speaker’s fees of small amount from Boston, Biotronik, Boehringer, and Medtronic. The other authors have no conflict of interest to declare. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal – Quality of Care and Clinical Outcomes or of the European Society of Cardiology.