The lack of long-range negative correlations in glucose dynamics is associated with worse glucose control in patients with diabetes mellitus.

ABSTRACT

Glucose dynamics measured in ambulatory settings are fluid in nature and exhibit substantial complexity. We recently showed that a long-range negative correlation of glucose dynamics, which is considered to reflect blood glucose controllability over a substantial period, is absent in patients with diabetes mellitus. This was demonstrated using detrended fluctuation analysis (DFA), a modified random-walk analysis method for the detection of long-range correlations. In the present study, we further assessed the relationships between the established clinical indices of glycemic or insulinogenic control of hemoglobin A(1c) (HbA(1c)), glycated albumin (GA), 1,5-anhydroglucitol, and urine C-peptide immunoreactivity and the recently proposed DFA-based indices obtained from continuous glucose monitoring in 104 Japanese diabetic patients. Significant correlations between the following parameters were observed (1) HbA(1c) and the long-range scaling exponent α(2) (r = 0.236, P < .05), (2) GA and α(2) (r = 0.254, P < .05), (3) GA and the short-range scaling exponent α(1) (r = 0.233, P < .05), and (4) urine C-peptide immunoreactivity and the mean glucose fluctuations (r = -0.294, P < .01). Therefore, we concluded that increases in the long-range DFA scaling exponent, which are indicative of the lack of a long-range negative correlation in glucose dynamics, reflected abnormalities in average glycemic control as clinically determined using HbA(1c) and GA parameters.