Background : The fibrillatory waves have various disorganized features. We hypothesized that mathematical analysis of arial fibrillary waves may be useful to predict recurrence of atrial fibrillation(AF) after electrical cardioversion.
Methods : We included 36 patients with symptomatic persistent AF ( 29 males, 58±8 years). All patients were assessed by 24hr holter monitoring and echocardiography within 3 months before cardioversion. AF recurrence was defined as AF episodes documented by EKG or holter monitoring within 3 months after cardioversion. We extracted only atrial fibrillary waves from V1 lead in 24hr holter data and measured 5 parameters including Shannon(Shn), approximate entropy(Aprx), Lyapunov exponent(Lya), Correlation dimension(CD) and dominant frequency(DF), in order to quantify entropy, chaotic activity, and frequency, respectively(day and night).
Results : All parameters which represented chaotic activity, entropy and frequency were represented by non-parametric analysis (median, 25th~75th order). (1) In chaotic activity, the patients with AF recurrence (n=22) had lower Lya scale(day) and higher CD sclale(day) than those without recurrence (n=14) (0.053,0.039~0.076 vs 0.078, 0.057~0.086, p=0.009),(1.828, 1.808~1.842 vs 1.815, 1.793~1.822, p=0.07). Whereas in entropy analysis, Aprx and Shn scale were not related to AF recurrence (p=0.539, p=0.733).
(2) In univariative Cox analysis, the patients with Lya scale(day)<0.07104(median) had more AF recurrence than those with Lya scale≥0.07104 (OR=7.65, 95%CI 1.635~35.798, p=0.01). In multivariate Cox analysis adjusted by age, sex and left atrium AP diameter(LAAP), the patients with Lya(day)<0.07104(median) had also higher AF recurrence within 3 month after cardioversion (OR=18.182, 95% CI 1.630~202.825,p=0.018)
Conclusions : Lya(day) scale which represented chaotic activity in fibrillary waves could predict AF recurrence within 3 month after cardioversion, but entropy scales were not related with early outcome of AF after cardioversion
|
