https://orcid.org/0000-0002-4593-3854
This editorial refers to ‘Predictors, time-course, and outcomes of persistence patterns in oral anticoagulation for non-valvular atrial fibrillation: a Dutch nationwide cohort study’, by M.M.A. Toorop et al., https://doi:10.1093/eurheartj/ehab421.
Barriers to achieving optimal treatment persistence with oral anticoagulants in patients with atrial fibrillation.
Atrial fibrillation (AF) is associated with an at least five-fold increase in the risk of ischaemic stroke, and ischaemic strokes related to AF are twice as likely to be fatal compared with non-AF strokes.1 , 2 Moreover, patients surviving an ischaemic stroke related to AF have been reported to have poorer functional status.3 Hence, stroke prophylaxis with oral anticoagulants (OACs) remains the cornerstone in the care of patients with AF. Vitamin K antagonists (VKAs) were the drug of choice until the advent of direct oral anticoagulants (DOACs) demonstrating at least non-inferior efficacy and safety while offering significant advantages over VKAs, including fixed dosing with a predictive onset of action, reduced drug–drug and food interactions, and, importantly, no need for continuous monitoring and frequent clinical consultations.4 An important aspect of treating chronic conditions such as AF is ensuring persistence to prescribed treatments. However, patterns of treatment persistence with DOACs vary greatly with rates of treatment discontinuation 1 year after treatment initiation, ranging from ∼10% to as high as 40% in some reports.5–9 Importantly, non-persistence has consistently been associated with increased risks of adverse outcomes including ischaemic stroke and mortality.9–11 Studies regarding persistence with OAC treatment are often based on retrospective historical datasets, and contemporary reports of persistence patterns and the associated clinical consequences are scarce. Current recommendations regarding the management of patients with AF tend to focus on identification of patients with thrombo-embolic risk factors in relation to initiation of treatment with OACs.12 The importance of ensuring subsequent persistence with the prescribed OAC treatment has received less attention.
In this issue of the European Heart Journal, Toorop et al. contribute with a study investigating persistence with treatment with OACs and the clinical consequences of non-persistence in a Dutch cohort of DOAC-treated patients with newly diagnosed AF.13 Using data from large administrative registries and databases, the authors report rates of non-persistence among 93 048 first-time initiators of DOACs between 2013 and 2018. Of note, the probability of discontinuation was highest during the early stages of treatment, with a persistence with OAC treatment of 88.1% after 1 year of treatment, declining to 72.0% at 4 years from DOAC initiation. Factors associated with better treatment persistence included female sex, advanced age, history of thrombo-embolic events, previous VKA experience, and, interestingly, being married. Finally, the authors report that being non-persistent with OACs was associated with an increased risk of a composite of ischaemic stroke and ischaemic stroke-related death [adjusted hazard ratio (HR) 1.79, 95% confidence interval (CI) 1.49–2.15], ischaemic stroke (HR 1.58, 95% CI 1.29–1.93), and all-cause mortality (HR 2.32, 95% CI 2.18–2.47).
AF is a major risk factor for thrombo-embolic events, and the attributable risk of AF for ischaemic stroke increases with age, reaching the level of hypertension in the oldest patients.14 Hence, persistence with OAC treatment is of utmost importance, and data pertaining to inappropriate treatment discontinuation and identification of patients at high risk of being non-persistent are central in efforts for improving outcomes for patients with AF.
The relatively high probability of non-persistence with OAC treatment combined with the robust association with ischaemic stroke and mortality in the study by Toorop et al. is concerning. Several other real-world studies using claims- and registry-based data have examined this research question, with great variation in the estimates of non-persistence with treatment. This is likely to reflect differences in methodology, definitions of non-persistence, and granularity of the source data used. In this context, it is noteworthy that the discontinuation rates of DOACs compared with VKAs are not consistently in favour of either class of drugs.8 , 9 , 11 , 15
Several important limitations must be mentioned when interpreting the reported findings. Serial medication refill is an accepted research methodology to estimate treatment persistence, but it does not inform about sporadic adherence. In addition, the data in the leveraged registries do not hold information on the number of days covered by a single DOAC prescription. Consequently, the authors define non-persistence as a treatment gap of 100 days following the last filled OAC prescription and do not consider treatment gaps prior to the last filled prescription. This is likely to lead to a less sensitive definition of non-persistence and a certain degree of underestimation of the true rates of non-persistence. Importantly, the authors have chosen a definition of non-persistence in which the patients cannot die during the first 100 days of being non-persistent. This immortality bias makes the reported risks of ischaemic stroke and mortality associated with non-persistence somewhat difficult to interpret as the risk for these outcomes is highest in the early period of drug discontinuation. Finally, the authors’ report contains no information as to why the patients stop their treatment, and causal conclusions should be drawn very cautiously. Attributing increased risks of adverse clinical outcomes associated with non-persistence to the absence of OAC treatment is not straightforward, as the reason for stopping treatment could in itself imply increased hazards (e.g. incident cancer, major surgery, or episodes of major bleeding). As such, the reasons for stopping OAC treatment are likely to be heterogeneous, encompassing factors pertaining to the patient as well as the treating physician and healthcare system in general. These factors could range from appropriate decisions relating to severe bleeding or rapid decline in kidney function, to potentially inappropriate circumstances such as general inconvenience, minor bleeding or bruising, overestimation of bleeding risk by the treating physician, overconfidence in restoration of sinus rhythm by the treating physician, or financial concerns.
The study by Toorop et al. reinvigorates several interesting topics of discussion concerning AF treatment in the DOAC era. The lack of need for continuous monitoring with DOACs is convenient from both a patient and physician perspective. However, it raises concerns regarding treatment persistence. Perhaps the convenience of DOACs over VKAs is a mixed blessing, and the frequent consultations associated with VKA treatment are in fact not only a nuisance but also a perpetual reinforcement of the rationale for treatment. Being persistent with a life-long pharmacological treatment with potential side effects can be cumbersome, requiring motivation and support. Moreover, patients achieving sinus rhythm could be under the impression that OAC treatment is no longer needed. Hence, continuous patient education and involvement in the discussion of treatment strategies are likely to be crucial elements in the process of optimizing persistence with OAC treatment. Integration of family members, and other central persons involved with the patient with AF, in these discussions could play a key role. One could speculate that this is also underlined by the authors’ findings that patients living alone have poorer persistence with OAC treatment than married patients. However, studies aiming at identifying interventions for improving adherence and persistence with OACs in patients with AF are unfortunately very scarce. The Assessment of an Education and Guidance program for Eliquis Adherence in non-valvular Atrial fibrillation (AEGEAN) study randomized patients to either isolated standard of care treatment or standard of care including an additional educational and motivational programme. The study found no differences between the intervention and the control group. However, standard of care in this study also included a regular consultation, which could potentially have been sufficient to improve adherence in the control group.16 Efficient interventions for improving treatment adherence and persistence with OAC treatment have the potential of reducing the risk of ischaemic stroke in AF patients, and this should be a topic for rigorous future research. The current stage of the digital age has several intriguing possibilities for electronic interventions, such as smartphone apps providing reminders and possibly also regular clinical consultations using telemedicine. Importantly, these patients are often followed in a primary care setting while AF is frequently diagnosed during acute hospital admissions. Initiatives such as automatic invitations for consultation in primary care after being discharged from the hospital with OAC treatment could also be a possibility. While it is unclear what specific initiatives and interventions could be beneficial in improving treatment persistence, it is obvious that the current rates of non-persistence with treatment are suboptimal, as stressed by the study by Toorop et al. Successful treatment of AF requires not only scrutiny of risk factors and identification of patients requiring stroke prophylaxis, but also subsequent recognition of patients at risk of being non-persistent with a potentially life-saving treatment. It is concerning that an inappropriately low persistence with treatment is still an obstacle in the quest to gain full benefit of prophylactic treatment with OACs in AF. Interventional studies aiming at improving treatment persistence should be a distinct priority in future AF research (Graphical Abstract).
Conflict of interest: E.H. reports research support from Abbott, Anthos Therapeutics, Bristol Myers Squibb, CryoLife, Janssen, and Medtronic; consultant fees from Bayer, Bristol Myers Squibb/Pfizer, Janssen, and Medtronic; and lecture fees from Boehringer Ingelheim and Bristol Myers Squibb/Pfizer. P.V.R. has no conflicts to declare.
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
References
