Early identification of AFib - PhD ceremony

Synapsis of the thesis:
Atrial fibrillation (AF) is a common arrhythmia, especially among the elderly. People with AF are at increased risk of stroke and heart failure. Therefore, it is important to identify AF in an early stage. In this thesis, several methods for early identification of AF were investigated in community and post-stroke/transient ischemic attack (TIA) patients.

In the first chapters, it was shown that several multivariable prediction models are able to predict risk of AF in community cohorts, albeit with varying degrees of applicability to primary care electronic health records (EHRs). In the following chapters, it was shown that frequent premature atrial contractions (PACs) on electrocardiogram (ECG) are associated with, AF, stroke and mortality, urging a revision of PACs as benign ECG findings. Then, a number of artificial intelligence (AI) applications were tested for their ability to facilitate early AF detection. A single-lead ECG device was shown to have excellent accuracy for diagnosing AF in Dutch primary care patients, when assessed by cardiologists. However, an AI-enabled algorithm to assess risk of underlying AF on short snippets of continuous ECG recording, was insufficiently accurate to substitute 14-day Holter recording in older primary care patients or people with recent stroke or TIA at relatively low cardiovascular risk. Finally, it was shown that increasing the duration of post-stroke continuous heart rhythm monitoring up to 14 days provided little additional benefit over the current standard of care (72 hours) in The Netherlands.

In the General Discussion, the current evidence on AF screening research was summarized. While risk-stratified AF screening had shown improved AF yield over including otherwise unselected elderly patients, few studies had shown actual improvement of long-term clinical outcomes from AF screening. A combined approach of stratified AF screening with integrated, holistic care for AF patients, coupled with more thorough use of available EHR data, was suggested as a next step towards improving prognosis after AF screening.

https://researchinformation.amsterdamumc.org/en/publications/early-identification-of-atrial-fibrillation-in-primary-care-and-p