Stroke, atrial fibrillation, and the management of oral anticoagulation

Atrial fibrillation (AF) is the most common arrhythmia, becomes more prevalent with increasing age and is linked to neurological complications, including most notably ischemic stroke, but also cognitive dysfunction. The recent introduction of direct oral anticoagulants (DOAC) significantly advanced the management of patients with AF. Large-scale randomized controlled trials showed that – for most patients – DOAC are at least as effective as vitamin K antagonists (VKA) in preventing ischemic stroke but have the advantage of a lower risk for intracranial hemorrhage (ICH). However, clinically important patient populations were underrepresented or excluded in these trials, and more refined aspects which are important in clinical practice, including concomitant stroke etiologies or brain pathologies, medication adherence, neuroimaging characteristics or biomarker signatures were not addressed. Therefore, several research gaps and challenges remained for neurologists treating patients with AF, of which we selected 4 aspects to focus on in the following topics that comprise this PhD thesis.
The first topic focused on 3 high-risk subgroups of patients with AF treated with oral anticoagulants who were underrepresented or excluded from the large randomized trials. These were (i) stroke patients aged 85 years and older (“the oldest-old”), (ii) severely affected stroke patients dependent on the daily help of others, and (iii) stroke patients with concomitant cerebral small vessel disease. In the first main project of this PhD thesis we examined the performance of DOAC versus VKA in the oldest-old patients with recent stroke and AF in a large pooled analysis across 7 cohort studies. Facing the paucity of randomized evidence, many physicians have been reluctant to use DOAC in these patients. With this project, we provided new evidence that the benefits of DOAC over VKA are preserved in the oldest old with recent stroke, without any signal of a safety concern regarding risk of ICH. In an additional project from our local cohort, we showed that the favourable profile of DOAC over VKA was preserved also among patients with AF and recent stroke who were dependent on the daily help of others, a patient subgroup for which no data existed previously. Finally, we showed in the same cohort that concomitant cerebral small vessel disease in anticoagulated patients with AF and recent stroke was associated with an unfavorable clinical course, but the risk for ischemic stroke remained higher than the risk for ICH, even in the presence of small vessel disease. The latter two projects fall within the scope of this PhD thesis, but do not constitute its main body. Although both manuscripts were published in peer-reviewed journals, for the purpose of this thesis their presentation is restricted to abstracts. Our findings in the first topic of this PhD thesis advance the evidence for the use of anticoagulants in high-risk patient subgroups with AF and recent stroke.
The second topic examined ischemic stroke occurring despite anticoagulant therapy in AF patients. With the increasing use of oral anticoagulants, this scenario represents a growing challenge in everyday clinical practice, indicating the need to elucidate the underlying causes and – based on these – the optimal subsequent management strategies. We addressed this issue in the second main project of this PhD thesis in a large retrospective analysis pooling data of prospectively collected patients from 11 stroke centers. We found that the causes of stroke despite anticoagulation in AF patients were heterogeneous, but form three main clusters, all of which were comparably important. These included (i) competing stroke mechanisms other than AF-related cardioembolism, and (ii) insufficient anticoagulation due to prescription errors and nonadherence, suggesting that individualized treatment approaches to address these causes are necessary. Importantly, the third and most common cause was AF-related cardioembolism despite sufficient anticoagulation, indicating the need to develop novel preventive strategies beyond the currently available anticoagulants. Furthermore, in this project we were able to demonstrate that AF patients with stroke despite anticoagulation represent a high-risk patient population, with higher than expected rates of stroke recurrence and other unfavorable outcomes. Finally, we showed also in this population that subsequent treatment with DOAC was associated with better outcomes than VKA treatment. Interestingly, neither any specific switch between DOAC nor antiplatelets as add-on treatment to anticoagulation seemed to confer any benefit, although both approaches are often employed in clinical practice. This study advanced the evidence for the preferential use of DOAC over VKA in AF patients with stroke despite anticoagulation, for whom no data existed so far, while demonstrating the need for more individualized and novel treatment approaches in these high-risk patients.
The third topic of this PhD thesis was the adherence of stroke patients to DOAC. Unlike VKA, DOAC require no coagulation monitoring and have short half-lives, which has raised concerns about nonadherence in AF patients treated with DOAC. This is particularly pertinent to patients with stroke, as shown in the second topic of this PhD thesis. In order to examine the medication-taking behaviour and the effect of an adherence-enhancing intervention in patients with recent stroke, we designed, initiated, and undertook the MAAESTRO study. MAAESTRO has been a joint venture with the Pharmaceutical Care Research Group of the University of Basel and has used electronic monitoring as the main method to assess adherence, data on which have been scarce so far. MAAESTRO comprises an initial observational phase and a subsequent randomized controlled interventional phase. MAAESTRO successfully concluded recruitment in July 2021 with reaching the predefined goal of n=130 participants. The observational study phase has now been completed, but as follow-up in the interventional phase is ongoing, the main study results are not part of this PhD thesis. Still, we present the published study protocol and the first results from the observational phase on the patterns of DOAC-taking behaviour, as well as an exploratory analysis on how adherence was impacted by the COVID-19 lockdown as abstracts, as these publications fall within the scope of this PhD thesis, but do not formally constitute its main body.
The fourth and final topic of this PhD thesis focused on cognitive dysfunction as a neurological complication of AF. While stroke is a well-known consequence of AF, there is increasing evidence that AF is also linked to cognitive dysfunction independent of ischemic stroke, but the mechanisms underlying this association are unclear. To preserve cognitive function in the growing population of elderly AF patients, a better understanding of these mechanisms is needed. Using data from the multicenter Swiss-AF Cohort Study, in the third main project of this PhD thesis we investigated serum neurofilament light chain (sNfL), a novel blood-based biomarker of neuronal damage, as a tool to explore the mechanisms through which neurological disease occurs in AF. In a cross-sectional analysis, we showed that sNfL is inversely associated with brain volume and cognitive function, thereby demonstrating that it represents a relevant biomarker of brain health in AF patients. Furthermore, we showed that neuronal loss measured by sNfL is associated with age, diabetes mellitus, heart failure, blood pressure and vascular brain lesions, observations which provide mechanistic insights into the occurrence of neurological disease in AF. Finally, in the fourth main project of this PhD thesis, we additionally investigated in this elderly cardiovascular cohort how renal function and body mass index contribute to sNfL levels in order to gain insights into the homeostasis (i.e., clearance and distribution) of this biomarker in the blood compartment. A better understanding of this is necessary towards further establishing this neurological biomarker in cardiovascular and dementia research. We showed that both renal function and body mass index were strongly, inversely associated with sNfL, but only renal function explained a relevant proportion of its variance. With this project we provided evidence for the importance of accounting for renal function in future sNfL-based investigations in elderly cardiovascular populations, in whom chronic kidney disease is highly prevalent.