App-based approximation of time spent with symptoms and atrial fibrillation in patients undergoing ECV: A TeleConvert-AF sub analysis
M Engels, B J M Hermans, J G L Luermans, S M Chaldoupi, J Hendriks, T A R Lankveld, M V M Kessels, B J G H Vorstermans, R J Beukema, K Vernooy, W Heesen, A N L Hermans, D K LinzAbstract
Background
Electrical cardioversion (ECV) is commonly used to restore sinus rhythm (SR) in patients with persistent atrial fibrillation (psAF). App-based intermittent photoplethysmography (PPG) monitoring could aid in assessing time spent with symptoms, time spent in AF, pulse characteristics and the association between symptoms and heart rhythm, but has not yet been prospectively evaluated in this population.
Methods
Patients were instructed to record pulse rate and heart rhythm, and symptoms via a smartphone-based PPG application three times daily from ≥2 weeks prior and 4 weeks after ECV or cancellation of ECV, and when experiencing symptoms. Recordings were classified as ‘normal’ (probable SR), ‘warning’ (SR with abnormalities such as tachycardia, extrasystoles, or bradycardia), or ‘urgent’ (possible AF). Patients with ≥14 recordings before ECV or its cancellation were included. PPG-derived pulse characteristics (pulse rate, heart rhythm, and pulse irregularity), time spent with symptoms, and time spent in AF were assessed (fig-1). "Load" is the proportion of time spent with symptoms or in AF relative to the total monitoring time; "density" only includes symptomatic or AF episodes with at least three consecutive recordings to capture the frequency of more sustained AF episode patterns. The symptom rhythm correlation index is the proportion of symptomatic AF recordings and asymptomatic non-AF (‘normal’ and ‘warning’) recordings relative to the total number of recordings during the monitoring period. The association between symptoms and PPG-derived pulse characteristics were studied.
Results
In total, 252 patients were included. The median symptom load prior to ECV or cancellation was 20.7% [1.4-61.2]. Female sex, absence of heart failure, and history of cardiac ablation were independent predictors for a higher symptom load. Symptom load was significantly correlated to AF load (Spearmans Rho=0.280, p<0.001). Symptom load and density, as well as AF load and density decreased significantly after ECV or its cancellation (all p<0.001) (fig-2). The overall symptom-rhythm correlation was low (0.34 [IQR 0.06-0.75]). Recordings with symptoms were identified as AF (61%)as compared with non-AF (39%)(p<0.001). Both pulse rate and irregularity were correlated to the presence of symptoms (both p<0.001). In multivariate analysis of all AF recordings, only pulse irregularity remained independently associated with the presence of symptoms.
Conclusion
App-based PPG monitoring offers valuable insight into symptoms and their relation to heart rhythm in patients scheduled for ECV for presumed psAF. Pulse irregularity emerged as the sole independent predictor of symptom presence during AF recordings, underscoring its clinical relevance in symptomatic AF episodes. Whether mobile rhythm monitoring can help to guide symptom control in patients with psAF warrants further studies.Fig-1:Algorithm clarificationFig-2:Evolution of paramaters