TY - JOUR
T1 - Investigation of glucose fluctuations by approaches of multi-scale analysis
AU - Lai, Yunyun
AU - Zhang, Zhengbo
AU - Li, Peiyao
AU - Liu, Xiaoli
AU - Liu, Yi Xin
AU - Xin, Yi
AU - Gu, Weijun
N1 - Publisher Copyright:
© 2017, International Federation for Medical and Biological Engineering.
PY - 2018/3/1
Y1 - 2018/3/1
N2 - Glucose variability provides detailed information on glucose control and fluctuation. The aim of this study is to investigate the glucose variability by multi-scale analysis approach on 72-h glucose series captured by continuous glucose monitoring system (CGMS), gaining insights into the variability and complexity of the glucose time series data. Ninety-eight type 2 DM patients participated in this study, and 72-h glucose series from each subject were recorded by CGMS. Subjects were divided into two subgroups according to the mean amplitude of glycemic excursions (MAGE) value threshold at 3.9 based on Chinese standard. In this study, we applied two types of multiple scales analysis methods on glucose time series: ensemble empirical mode decomposition (EEMD) and refined composite multi-scale entropy (RCMSE). With EEMD, glucose series was decomposed into several intrinsic mode function (IMF), and glucose variability was examined on multiple time scales with periods ranging from 0.5 to 12 h. With RCMSE, complexity of the structure of glucose series was quantified at each time scale ranging from 5 to 30 min. Subgroup with higher MAGE value (CloseSPigtSPi3.9) presented higher glycemic baseline and variability. There were significant differences in glycemic variability on IMFs3–5 between subgroups with MAGECloseSPigtSPi3.9 and MAGE OpenSPiltSPi = 3.9 (pOpenSPiltSPi0.001), but no significant differences in variability on IMFs1–2. The complexity of glucose series quantified by RCMSE showed statistically difference on each time scale from 5 to 30 min between subgroups (pOpenSPiltSPi0.05). Glucose series from subjects with higher MAGE value represented higher variability but lower complexity on multiple time scales. Compared with traditional matrices measuring the glucose variability, approaches of EEMD and RCMSE can quantify the dynamic glycemic fluctuation in multiple time scales and provide us more detailed information on glycemic variability and complexity.
AB - Glucose variability provides detailed information on glucose control and fluctuation. The aim of this study is to investigate the glucose variability by multi-scale analysis approach on 72-h glucose series captured by continuous glucose monitoring system (CGMS), gaining insights into the variability and complexity of the glucose time series data. Ninety-eight type 2 DM patients participated in this study, and 72-h glucose series from each subject were recorded by CGMS. Subjects were divided into two subgroups according to the mean amplitude of glycemic excursions (MAGE) value threshold at 3.9 based on Chinese standard. In this study, we applied two types of multiple scales analysis methods on glucose time series: ensemble empirical mode decomposition (EEMD) and refined composite multi-scale entropy (RCMSE). With EEMD, glucose series was decomposed into several intrinsic mode function (IMF), and glucose variability was examined on multiple time scales with periods ranging from 0.5 to 12 h. With RCMSE, complexity of the structure of glucose series was quantified at each time scale ranging from 5 to 30 min. Subgroup with higher MAGE value (CloseSPigtSPi3.9) presented higher glycemic baseline and variability. There were significant differences in glycemic variability on IMFs3–5 between subgroups with MAGECloseSPigtSPi3.9 and MAGE OpenSPiltSPi = 3.9 (pOpenSPiltSPi0.001), but no significant differences in variability on IMFs1–2. The complexity of glucose series quantified by RCMSE showed statistically difference on each time scale from 5 to 30 min between subgroups (pOpenSPiltSPi0.05). Glucose series from subjects with higher MAGE value represented higher variability but lower complexity on multiple time scales. Compared with traditional matrices measuring the glucose variability, approaches of EEMD and RCMSE can quantify the dynamic glycemic fluctuation in multiple time scales and provide us more detailed information on glycemic variability and complexity.
KW - Continuous glucose fluctuation
KW - Ensemble empirical mode decomposition
KW - Mean absolute glycemic excursions
KW - Multiple scales
KW - Refined composite multi-scale entropy
UR - http://www.scopus.com/inward/record.url?scp=85027892612&partnerID=8YFLogxK
U2 - 10.1007/s11517-017-1692-0
DO - 10.1007/s11517-017-1692-0
M3 - Article
AN - SCOPUS:85027892612
SN - 0140-0118
VL - 56
SP - 505
EP - 514
JO - Medical and Biological Engineering and Computing
JF - Medical and Biological Engineering and Computing
IS - 3
ER -