Predicting factors associated with hypoglycemia reduction with automated predictive insulin suspension in patients at high risk of severe hypoglycemia: an analysis from the SMILE randomized trial.

BACKGROUND: This analysis from the SMILE randomized study was performed to identify predictive factors associated with the greatest reductions in hypoglycemia with the Medtronic MiniMed™ 640G Suspend before low feature in adults with type 1 diabetes at high risk of severe hypoglycemia. METHODS: Clinical and treatment-related factors associated with decreased sensor hypoglycemia (SH) were identified in participants from the intervention arm by univariate and multivariate analyses. RESULTS: The reduction in SH events <54mg/dL (<3.0 mmol/L) in the intervention group was significantly (p<0.0001) associated with the baseline mean number of sensor hypoglycemic events (MNSHE) <54mg/dL. When excluding CGM factors not readily available (MNSHE, duration of SH events, area under the curve [AUC], mean amplitude of glycemic excursions [MAGE]), only the baseline mean time spent <54mg/dL was found to be a significant independent predictor factor (p0.0001). Baseline HbA1c, mean self-monitoring of blood glucose (SMBG) and coefficient of variation (CV) of SMBG were significant, albeit weak, predictors in the absence of any CGM data. CONCLUSIONS: The greatest reductions in SH events achieved with the MiniMed 640G system with the Suspend before low feature were seen in participants with higher baseline MNSHE. Measuring these (usually uncollected) events can be a useful tool to predict hypoglycemia reduction.