Hi, my name is Emma Sandgren and I'm from the Karolinska Institute in Stockholm, Sweden, and today I'm going to talk on this trial CONSIDERING-AF.

What is the importance behind the study?

Current European guidelines recommend considering systematic AF screening in individuals aged 65 years or older with additional stroke risk factors. And our thought was that some kind of enrichment aimed at target screening for individuals with highest risk of developing AF will potentially improve the diagnostic yield.

Therefore, our aim was to test whether the combination of our respiratory model developed using machine learning together with long- term ECG recording was superior to standard of care in detecting new cases of AF in a population aged 65 years or older.

What was the study design and patient population?

The trial design was randomized, control, sightless, non-blinded, superiority trail featuring four parallel arms: two intervention arms and two control arms. And the target population for the trial was residents in Region Halland, Sweden aged 65 years or older without a diagnosis of AF in their electronic health record. And exclusion criteria was having AF, having another indication for [illegible] treatment, having dementia or an implantable device.

What are the key findings?

Key findings of the trial was that in the [illegible] analysis, the incidence of AF after six months in the RPM intervention arm was significantly higher than the incidence in the general control arm, 3.8% versus 0.7%, which means an absolute increase of 3.1%, a risk ratio of 5.6 and a [illegible] of 32.

The result was also significant in the proprotocol analysis with an AF detection rate of 6.3% in the RPM intervention arm. And additionally, in the proprotocol analysis, both of the components, the risk prediction model and the long- term ECG recording, showed significant results indicating that both components were important.

Based on your findings, what would be your recommendation for the optimal duration of monitoring for AF screening?

The [illegible] yield depends on both the duration of the monitoring and the prevalence in the target population. And in this trial, we wanted to target individuals with highest risk of developing AF, and it's also these individuals who will benefit the most for ADHD treatment.

Interestingly, 76% of all cases of AF detected with the ECG patch were identified within the first seven days of monitoring, suggesting that these individuals, high risk individuals, have a significant AF burden which could indicate that seven days of monitoring may be enough in a population like this, but maybe rather repeat the screening in individuals where AF is not initially detected.

What are your take-home messages?

The combination of a risk prediction model and long-term ECG monitoring was superior to standard of care in detecting new case of AF in a population aged 65 years or older. The participation rate was high for a digital screen study. The overall participation rates in both intervention arms were 42%, but AF detection rate was low in the general non-enriched cohort.

What further research is needed, and what are the next steps?

Although target AF screening was efficient in finding patients with unknown AF, the benefits of finding these patients need to be verified with regard to clinical outcomes such as AF-related strokes.

Additionally, health economic consequences and the effect on health care consumptions need to be evaluated for target AF screening. Even though AF screening has the potential to reduce future AF- related strokes, finding these AF cases is related to our costs and at least temporarily increased healthcare consumption.