Abstract
BACKGROUND: Heart rate complexity, derived from nonlinear heart rate variability (HRV), has been shown to help predict the outcomes of various diseases. Changes in heart rate complexity before and after paroxysmal atrial fibrillation (PAF) events are unclear. OBJECTIVES: To evaluate changes in heart rate complexity through nonlinear HRV before and after PAF events. METHODS: We enrolled 65 patients (72 ± 12.34 years old, 31 females) with 99 PAF events who received 24-hour Holter recording, and analyzed nonlinear HRV variables including Poincaré plot analysis, sample entropy (SampEn), and multiscale entropy (MSE). HRV analyses were applied to a 20-minute window before the onset and after the termination of PAF events. HRV parameters were evaluated and compared based on eight different 5-minute time segments, as we divided each 20-minute window into four segments of 5 minutes each. RESULTS: SampEn and MSE(1~5) significantly decreased before the onset of PAF events, whereas SampEn, MSE(1~5) and MSE(6~20) significantly increased after the termination of PAF events. SD1 and SD2, which are nonlinear HRV parameters calculated via Poincaré plot analysis, did not significantly change before the PAF events, however they both decreased significantly after termination. CONCLUSIONS: Heart rate complexity significantly decreased before the initiation and increased after the termination of PAF events, which indicates the crucial role of nonlinear heart rate dynamics in the initiation and termination of PAF.