Abstract
AIM: This study aims to predict risk factors for hypoglycemia in patients with type 2 diabetes mellitus (T2DM) using continuous glucose monitoring (CGM) and with time in range (TIR) > 70%. METHODS: Data from 111 patients with T2DM who underwent CGM with TIR > 70% were analyzed. A hypoglycemia episode was defined as CGM-detected glucose < 3.9mmol/L sustained for at least 5 min. Logistic regression analysis was performed to examine the relationship between hypoglycemia and mean blood glucose (MBG), glycemic variability (GV) metrics [including mean amplitude of glucose excursion (MAGE), largest amplitude of glycemic excursion (LAGE), mean of daily difference (MODD), coefficient of variation (CV), standard deviation (SD)], and low blood glucose index (LBGI). A nomogram model was constructed, and its diagnostic performance was assessed. Data were bootstrapped 1000 times for internal validation, and a calibration curve was drawn to evaluate the model's predictive ability. Decision curve analysis was performed to assess its clinical usefulness. RESULTS: Among the 111 included patients, 53 experienced hypoglycemic event during wearing CGM (47.75%). GV metrics were higher in hypoglycemia group, while MBG was lower. The multivariable logistic regression analysis showed that the MBG, GV metrics, LBGI were independently associated with hypoglycemia. The receiver operating characteristics (ROC) analysis indicated that the area under the curve (AUC) for the MBG-SD-LBGI model was 0.93 (95% CI = 0.88-0.97). The calibration curve showed good consistency between the predicted and observed probabilities. Decision curve analysis demonstrated strong clinical applicability. CONCLUSION: This study demonstrates a significant correlation between CGM metrics and hypoglycemia in patients with T2DM who achieved TIR > 70%. These findings suggest that CGM metrics can predict the risk of hypoglycemia in T2DM patients with a TIR > 70%, and the nomogram developed from these metrics holds strong potential for clinical application.