Epidemiology of anticoagulants prescribing in non-valvular atrial fibrillation patients in England: a cohort study exploring disease incidence and anticoagulants prescribing trends, comparing clinical outcomes, and risk of recurrent bleeding events

Date of Approval: 
2020-09-25 00:00:00
Lay Summary: 
Despite improvements in the assessment of the risks of stroke and bleeding in people with atrial fibrillation – a common type of irregular heartbeat, many people are still not optimally treated with anticoagulants (blood thinners to prevent blood clotting). It is believed that this may be due to overestimating patient’s risk of bleeding during anticoagulation therapy. The consequences of lack of anticoagulation could put patients at higher risk for stroke or venous thromboembolism (blood clot), which are associated with long-term disability and death. Using UK primary care data and hospitalization records from the Clinical Practice Research Datalink, we will determine how commonly atrial fibrillation is diagnosed in England and how it is treated. We aim to evaluate the factors that may influence the under-prescribing of anticoagulants in this sub-group of patients, including their medical conditions, frailty, bleeding risk and concomitant therapy. We will use statistical methods to compare the risk of stroke, bleeding, or death for this sub-group of patients with patients who received anticoagulant therapy. We will separately examine patients who experienced severe bleeding due to anticoagulants, and whether they continued on anticoagulants or not after their bleeding episode. This will allow the identification of factors related to these prescribing decisions and how anticoagulants influence patients’ clinical-outcomes. We will also assess and compare their risk for a stroke or another bleeding event using different clinical risk-assessment tools. We anticipate that our findings will help in improving prescribing decisions of anticoagulant therapy in patients with atrial fibrillation at risk for stroke.
Technical Summary: 
Background and Objective: Atrial fibrillation (AF) can significantly increase the risk of ischemic stroke. The optimal use of oral anticoagulants (OACs) and control of stroke-related risk factors, such as hypertension, is proven to substantially reduce stroke risk. However, some clinicians may not prescribe OACs to AF patients at risk for stroke, as they believe that the risk of bleeding outweighs the benefit of anticoagulation. We aim to assess and compare the impact of prescribing anticoagulants by examining OACs prescribing in a cohort of non-valvular atrial fibrillation (NVAF) patients in England. Methods and Data analysis: The study comprises of three phases; first is an epidemiological study that aims to delineate incidence of NVAF from the CPRD GOLD and Aurum databases between 2009 and 2019. The prescription of OACs will be assessed for all patients and relevant data will be extracted to identify potential predictors of OACs prescribing and further assessment using multivariable logistic-regression modelling. The second phase will be a comparative cohort study of incident NVAF patients who were classified as “OAC not prescribed” in phase one, by using a propensity score matching method to match each patient not prescribed OACs to up to 5 comparable patients treated with OACs. Then we aim to compare stroke, bleeding, and all-cause mortality associated with OACs prescribing versus no OACs. The third phase will focus on patients who developed major bleeding events in phase two while taking OACs. We aim to explore the variability of OACs prescribing after bleeding across patient groups and the different predictors for OACs resumption. We will explore the rate of bleeding reoccurrence, stroke, and mortality, and how they relate to OACs prescribing and timing of restarting OACs. We will also assess and compare patients’ risk for stroke and bleeding reoccurrence after the bleeding event using different risk assessment scores.
Health Outcomes to be Measured: 
Throughout the different phases of this project we aim to explore the following outcomes: • Phase 1 – The primary outcome is the prescription of OAC drugs including; vitamin K antagonists (VKAs) (acenocoumarol, phenindione or warfarin), or non-vitamin K oral anticoagulants (NOACs) (dabigatran, rivaroxaban, apixaban, or edoxaban). • Phase 2 - The primary outcome is ischaemic stroke and other thromboembolic events, or the occurrence of major bleeding (MB) defined according to the International Society on Thrombosis and Haemostasis (ISTH) criteria as (1) clinically overt bleeding accompanied by a drop-in haemoglobin level of least 2 g/dl and/or (2) transfusion of at least 2 units of packed red cells, occurring at a critical site (i.e., intracranial, intraspinal, intraocular, pericardial, intra-articular, intramuscular, and retroperitoneal) and/or (3) fatal bleeding.1 Secondary outcomes will include clinically relevant non major bleeding (CRNMB) that did not satisfy the ISTH criteria and led to hospitalization, medical or surgical treatment, or led to discontinuation of OACs. • Phase 3 – The primary outcomes include the reoccurrence of MB events or CRNMB for patients who did or did not continue on OACs, the rate of stroke and other thromboembolic events, cardiovascular related mortality, and all-cause mortality. The secondary outcome is the prescription of any OACs after the occurrence of the MB events. We aim to explore the timing of reinitiating OACs from the indexed MB event and weather patients have continued the same OAC before the index haemorrhage event, were switched to another OACs or received lone antiplatelet therapy. All outcomes will be identified over the course of follow-up through the identification of corresponding Read/SNOWMED codes in patients’ electronic health records. Hospitalization secondary to stroke or bleeding are to be identified through linkage to HES APC using the 10th revision of the International Classification of Diseases (ICD-10). ONS mortality data will be used to ascertain specific causes of death (i.e. bleeding related, cardiovascular, or all-cause death) according to ICD-10 codes.
Application Number: 

Evangelos Kontopantelis - Chief Investigator - University of Manchester
Alyaa Ajabnoor - Corresponding Applicant - University of Manchester
Darren Ashcroft - Collaborator - University of Manchester
Mamas Mamas - Collaborator - Keele University
Rosa Parisi - Collaborator - University of Manchester
Salwa Zghebi - Collaborator - University of Manchester

HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation