Distinguishing normal and abnormal heart rate variability using graphical and non-linear analyses

• Published in: Computers in Cardiology, 2004 pp. 205 - 208
• Main Authors: Stein, P.K., Hui, N., Domitrovich, P.P., Gottdiener, J., Rautaharju, P.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 2004
• Subjects: Cardiology; Doped fiber amplifiers; Fractals; Frequency domain analysis; Heart rate; Heart rate variability; Neural networks; Rhythm; Risk analysis; Sun
• ISBN: 0780389271; 9780780389274
• DOI: 10.1109/CIC.2004.1442908

Abnormal HRV could confound risk stratification. Method: Hourly PoincarP and FFTplots examined in 270 rapes from the Cardiovascular Health Study. Afrer 8 years, 63 subjects had died. Hourly short and longer-term oletrended fractal scaling exponent and interbeat correlations were calculated. Hourly HRV was scored as nom1 (a), borderline (0.5) or abnormal (1) from plot appearance and HRV values. Scores were summed by subject and normalized to create nn abnormalig score (ABN,O- 100%). Cox regression determined the relationship of ABN and mortality. Results: Increased ABN was associated with mortality, p=O.O0.5. After adjustment for age (p=O.OOI) and gender (p=O.OOS), ABN remained associated with mortality (p=O.OIS). When ABN was dichotomized at 57%. HR and SDNN were not diflerent, but higher ABN (N=67) had significantly increased short and intermediate-term XRV and mortaliry. Conclusion: Even with a relatively crude guantcation method, abnormal rhythms were associated with both mortality and increased HRV.