longest AF episode duration: a new indicator of AF burden?

Abstract Background/Introduction Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with increased morbidity and mortality, including elevated risk of stroke and worsening of heart failure. The AF burden, the time a patient spends in AF during a certain monitoring period, was recently shown to have significant effect on the thromboembolic risk of AF patients and may predict success of rhythm control therapies. However, continuous rhythm monitoring is required to determine AF burden, precluding its assessment in large-scale studies. Therefore, a surrogate parameter for AF burden that requires less monitoring time would be highly valuable. Purpose Using a patient-level Markov model simulating the occurrence and progression of AF episodes, this study evaluates 1) whether the longest AF episode duration can be used as a surrogate parameter for AF burden and 2) how much monitoring time is required to identify the longest AF episode. Methods The virtual AF model simulates the stochastic transition between sinus rhythm, AF, stroke, and death during the lifetime of a cohort of virtual patients with minute-level resolution. For each AF burden (3%, 6% and 9%) we considered two patient subpopulations: patients with low frequent long-lasting episodes (LFLE) and patients with high frequent short-lasting episodes (HFSE). We evaluated the correlation between the longest AF episode duration and AF burden. Furthermore, we assessed the monitoring time needed to detect the longest AF episode duration for a substantial proportion (80%) of patients during a follow-up period of one year. We also considered the monitoring time needed for episodes which are a good approximation of the longest AF episode duration: episodes exceeding 70%, 80% or 90% of the longest AF episode duration. Results (Figure 1 and 2) The longest AF episode duration and AF burden are highly correlated (Spearman’s : 0.77, P<0.001) in the virtual AF model. The longest AF episode duration can be detected for 80% of the patients after 311, 322 and 322 days for AF burdens of 3%, 6% and 9% in the HFSE subpopulation and after 314, 314 and 322 days in the LFLE subpopulation. An episode exceeding 70% of the longest AF episode duration can be detected for 80% of the patients after 161, 143 and 146 days for AF burdens of 3%, 6% and 9% in the HFSE subpopulation and after 238, 223 and 252 days in the LFLE subpopulation. Conclusion The longest AF episode duration correlated with AF burden. However, its feasibility as surrogate parameter depends on the episode frequency and the level of AF progression. In patients with the HFSE subtype and limited AF progression, longest AF episode duration can be estimated after relatively short follow-up periods. Further studies are required to establish cut-off values and to assess the monitoring time needed for the longest AF episode duration for specific patient subgroups.Figure 1 Figure 2