Symptom severity assessment in patients with atrial fibrillation after catheter ablation using a smartphone application
E Sandgren, K Betz, M Gawalko, A Hermans, Z Habibi, D Verhaert, J M Hendriks, D Den Uijl, S M Chaldoupi, J Luermans, T Lankveld, U Schotten, K Vernooy, D LinzAbstract
Background
In patients with atrial fibrillation (AF), greater symptom severity correlates with increased healthcare utilisation. Traditional follow-ups using spot-assessments lack the temporal precision to confirm symptom-rhythm concordance. Mobile health tools now enable longitudinal real-time concurrent evaluation of symptoms and rhythm status.
Purpose
To evaluate the feasibility of an on-demand smartphone application for longitudinal assessment of symptom severity in patients after AF ablation, to identify predictors of higher symptom severity, and to explore between- and within-patient variability in symptom reporting.
Methods
Patients were instructed to record heart rhythm, symptoms and symptom severity using a photoplethysmography (PPG)-based smartphone application three times daily for seven days on-demand at 3-, 6-, and 12-month follow-ups. Symptom severity during each PPG rhythm recording was assessed using the modified European Heart Rhythm Association symptom score. (Figure 1).
Results
In total 8575 PPG recordings were analysed from 140 patients. Median adherence was 67% (interquartile range (IQR) 56-76%) and median number of completed 7-day monitoring periods was 3 (IQR 2-3). A cumulative link mixed model showed that, compared with AF, extrasystoles and sinus rhythm were associated with lower symptom severity (odds ratio (OR) 0.25, 95% confidence interval (CI) 0.15-0.58; OR 0.13, 95%CI 0.06-0.32). Median mEHRA score during AF 2 (IQR 1-3), extrasystoles 1 (IQR 1-1), and sinus rhythm 1 (IQR 1-1), p<0.001. Female sex and higher heart rate were associated with higher symptom severity (OR 2.69, 95%CI 1.06-6.80; OR 1.05 95%CI 1.04-1.06). No significant effects were found for multimorbidity or age. The model included random intercepts for participant (σ² = 6.81) and monitoring period nested within participant (σ² = 1.34). Using the logistic link, approximate intraclass correlation indicated that 59% of symptom severity variation was between patients, 12% within patients across monitoring periods, and 29% unexplained, reflecting unpredictable measurement-level variation or unmodulated factors.
Conclusions
On-demand smartphone-based combined assessment of rhythm and symptom severity following AF ablation was feasible. AF recurrence, female sex and higher heart rate were associated with higher symptom severity. Most variability in symptom severity reflected differences between patients, with smaller contributions from within-patient fluctuations, underscoring the role of both individual factors and temporal variations in symptom assessment.