Efficacy of frequency domain parameters applied onto ultra-short ECG recordings in the diagnosis of definite cardiac autonomic neuropathy -Comparing Lomb-Scargle-Periodogram and Fast-Fourier-Transform

Cardiac autonomic neuropathy (CAN) is often, but not always, linked to diabetes and should therefore be part of a general health review. Analysis of heart rate variability (HRV), the beat-to-beat alterations in heart rate, is a potential screening tool in general practice, but optimal recording lengths and diagnostic parameters need to be established. We calculated the low frequency (LF) and high frequency (HF) frequency domain parameters for 11 patients with definite CAN (dCAN) and 76 controls from ultra-short ECGs of 10 to 300 seconds in length and their ability to distinguish dCAN from controls was assessed by estimation of an area under the receiver operating characteristic curve (AUC). Fast-Fourier-Transform (FFT) and Lomb-Scargle-Periodogram (LSP) have been used separately to allow for a comparison of methods. LF and HF allowed for good overall separation of groups from 300-second-recordings, which had the highest accuracy and has been the traditional recommended length for frequency domain analysis, with LF yielding good results down to 60 secs, The differences between values obtained through FFT and LSP were clinically negligible for all window sizes.