It is now known that multiscale entropy has the potential to distinguish certain pathological time series clearly and reliably from the corresponding healthy series. However, the implications of this parameter for Heart Rate Variability (HRV) have not been studied extensively. Also, as reported by other studies, the Poincare plots of the R-R interval series of a human subject's ECG signal (which too function as a measure of HRV) could be more useful than the time-domain and frequency domain parameters of HRV for certain applications. Although the Poincare plots of healthy R-R interval series and corresponding PPG-based interval series have been investigated by many researchers, they do not seem to have been examined for unhealthy subjects. Our goal in this study has been to assess the extent to which PRV can substitute for HRV in the determination of these two nonlinear parameters. We perform multiscale entropy analysis (MSE) on Pulse Rate Variability (PRV) and HRV of 20 ICU patients. We also obtain Poincare plots associated with PRV from four PPG-based techniques and those characterizing HRV from the standard R-R interval technique. We then compare the resulting PRV data sets with their HRV counterparts. We observe that none of the PPG-based methods displays a satisfactory statistical agreement or even an acceptable statistical correlation with the standard ECG-based technique. Hence we conclude that as of now, one cannot rely on PRV as a convenient alternative to estimate the MSE values and Poincare plots of HRV. However, further investigation on the lines suggested in the paper might yield fruitful insights. © 2015 IEEE.