Introduction & Objective: Detailed analysis of continuous glucose monitoring (CGM) data may identify novel risk markers for severe hypoglycemia (SH).
Methods: Correlations between established and novel CGM features (Table 1) derived from 30-day CGM data and the number of self-reported SH events in the previous 6 months were investigated in 163 adults with type 1 diabetes (T1DM).
Results: The number of SH events were significantly positively correlated with SD (p=0.0360), Mean (p=0.0446), TAR1 (p=0.0482), MODD (p=0.0448), 24-hour CONGA (p=0.0492), and significantly negatively correlated with TIR (p=0.0296), but was not significantly correlated with TBR2, TBR1, TAR2, CV, MAG, GVP nor MAGE (Table 1). The strongest CGM correlation with severe hypoglycemia was Shannon entropy (p=0.0161), a measure of randomness/disorder in the CGM data.
Conclusion: Surprisingly, the number of severe hypoglycemic event did not correlate with time below range. However, the novel CGM analytic feature of Shannon entropy had the strongest correlation with the number of SH events. These findings suggest that incorporation of CGM-derived Shannon entropy determinations may have clinical utility to identify T1DM adults at high risk for severe hypoglycemia.
X.D. Zhang: None. Y. Lin: None. R.D. Zhang: None. S.J. Fisher: None.
National Institutes of Health (UL1TR001998); National Institutes of Health (U01DK135111); National Institutes of Health (P30 DK020579); National Institutes of Health (OT2HL161847)
