DeteCtiON and Stroke PreventIon by MoDEl ScRreenING for Atrial Fibrillation
DeteCtiON and Stroke PreventIon by MoDEl ScRreenING for Atrial Fibrillation
Atrial fibrillation (AF) is the most common clinical arrhythmia and the prevalence increases with age. AF increases the risk of ischaemic stroke fivefold and accounts for almost one-third of all strokes. As AF is often asymptomatic there are many undetected cases. It is important to find patients with AF and additional risk factors for stroke in order to initiate oral anticoagulation treatment, which can reduce the risk of an ischaemic stroke by 60-70%. Screening is recommended in European guidelines, however the most suitable population and the most suitable device for AF detection remain to be defined.
The main objective of this study is to test the hypothesis that AF screening with 14-days continuous ECG monitoring in high-risk individuals identified with a risk prediction model is more effective than routine care in identifying patients with undetected AF.
Effectively detecting AF among patients with risk factors for ischaemic stroke has the potential to decrease mortality and morbidity, stroke burden and costs for the society as a whole.
Objective(s): To compare the yield of atrial fibrillation (AF) using 14-days continuous ECG in a population aged ≥ 65 years with an increased risk for AF incidence according to the risk prediction model compared with standard of care in Region Halland.
Study Population: Residents in Region Halland age 65 and above.
Data Collection Methods: Electronic Health Records from Region Halland and 14-days continuous ECG recording using an ECG patch.
Study design:
Step 1:
To calibrate the BMS/Pfizer risk prediction model (RPM), we will extract two cohorts retrospectively: the AF cohort with an AF diagnosis (patients with a record of incident AF diagnosis between January 1, 2016, and December 31, 2019 as an observation period), and the control cohort without any AF diagnosis in their history. We will include patients ≥45 years of age at index date, which is the first date of an AF diagnosis recorded in the observation period and a random pseudo index date during the observation period for the control group, to follow the original study. Specifically, we are looking to calibrate the intercept (α) for the logistic regression where we already have the 13 odd ratios for the 13 risk factors from the original study. Then in the next step for the prospective study, applying the RPM on the RPM cohort, the at-risk group will be extracted for randomization step, using the recommended cut-off value.
Step 2:
The population in Region Halland aged 65 years and above and free from AF will be randomized into two halves, creating a general cohort and an RPM cohort. In the general cohort, 1480 individuals will be further randomized into two arms, general/control and general/intervention. In the RPM cohort, the risk of incident AF will be calculated according to the Pfizer/BMS RPM. Those with a predicted risk for incident AF above a pre-specified threshold will then be randomly extracted into two arms, RPM/control and RPM/intervention (figure 1).
Those randomized to the two intervention arms (general/intervention and RPM/intervention, n=740 each) will be invited to an AF screening intervention of 14-days continuous ECG using a patch device. Those randomized to the control groups (general/control and RPM/control, n=740 each) will not receive any information or intervention.
The primary endpoint will be the difference in yield of newly diagnosed AF between the RPM/intervention and the general/control arms, where the latter will represent standard of care. Participants with newly diagnosed AF in the intervention arms will be offered consultation aiming at AF work-up and initiation of oral anticoagulation treatment.
Inclusion Criteria:
Exclusion Criteria: