Renyi entropy in identification of cardiac autonomic neuropathy in diabetes

Heart rate variability (HRV) has been conventionally analyzed with time- and frequency-domain methods. More recent nonlinear analysis has shown an increased sensitivity for identifying risk of future morbidity and mortality in diverse patient groups. Included in the domain of nonlinear analysis are the multiscale entropy measures. The Renyi entropy is such a measure. It is calculated by considering the probability of sequences of values occurring in the HRV data. An exponent a of the probability can be varied to provide a spectrum of measures. In this work we applied the multiscale Renyi entropy for identification of cardiac autonomic neuropathy (CAN) in diabetes patients. Fifteen participants were identified with CAN (dCAN) using the five-test Ewing battery and 26 were control (nCAN). The multiscale Renyi entropy was measured from -5<a<+5. The best result was obtained with a=5, where the mean value for patients with CAN was 0.98 with standard deviation of 0.01, compared with a mean of 0.95 for controls with standard deviation of 0.02. The probability of the means being the same was p<0.0001, suggesting that a significant difference between these groups was found using the Renyi entropy. Other values of a also showed a significant difference. Different pathologies differ in their ECG and HRV and therefore no single HRV test should be expected to be ideal for all pathologies. However, this work shows that the multiscale Renyi Entropy provides a high level of discrimination and therefore should be considered as a neuroendocrine test for CAN.