Diabetes & Metabolism · Jun 16, 2026

CGM-based metrics and mortality in older adults living with type 2 diabetes on insulin therapy: a secondary analysis of the HYPOAGE study

By Smati S, Boureau AS, Gourdy P et al.

AIM: Older adults living with type 2 diabetes represent a particularly vulnerable population. We investigated which continuous glucose monitoring (CGM)-derived targets are associated with all-cause mortality in this population. METHODS: HYPOAGE is prospective multicenter study including 141 insulin-treated older adults living with type 2 diabetes aged 75 and older, under insulin therapy for at least 6 months. All participants underwent standardized geriatric and diabetic assessments and wore an ambulatory blinded CGM (FreeStyle Libre Pro®) for 28 consecutive days. In this ancillary study, multivariable cox regressions were performed to identify factors associated with mortality after adjustment for age, sex, HbA1c, kidney function, geriatric status, and metformin use. RESULTS: At baseline, participants were 81.5 years old on average. After a median follow-up of 44 months, 58 of 141 patients had died. In adjusted model, higher percentages of level 1 time below range (TBR), level 2 TBR and glycemic variability assessed by the coefficient of variation (CV) were independently associated with an increased mortality risk (hazard ratio [95% CI] 1.51 [1.11; 2.06], 1.25 [1.02; 1.53], and 1.76 [1.21; 2.56] for an interquartile range (IQR)% increase of each parameter, respectively). When recommended CGM targets were considered, only glycemic variability (CV ≤ 36%), remained significantly associated with a lower risk of mortality (hazard ratio 0.57 [0.32; 0.99]), whereas TIR > 50% and TBR ≤ 1% were not. CONCLUSION: Among insulin-treated older adults living with type 2 diabetes, glycemic variability was independently associated with all-cause mortality, highlighting its potential relevance for clinical management in geriatric diabetes care.

doi.org/10.1016/j.diabet.2026.101775Open accessPubMed 42303126Populations