Diabetes overtreatment and flags for deprescribing glucose-lowering treatment in geriatric inpatients with type 2 diabetes: Prognostic insights from a longitudinal cohort.

Aims: The study investigated diabetes overtreatment (excessively intensive therapy) and flags for considering deprescribing glucose-lowering treatment (GLT) (clinical features identifying patients most likely to benefit from reducing treatment) in older adults with type 2 diabetes (T2D) and their association with five-year survival in geriatric inpatients.

Methods: Single-centre retrospective cohort study including geriatric inpatients hospitalised between 2008 and 2015, aged ≥75 years with T2D on GLT. Diabetes overtreatment, according to the 2019-Endocrine Society guidelines, was retrospectively defined based on the patient's health status, level of HbA1c and presence of inducing hypoglycaemia GLT. Flags for deprescribing were selected from a systematic review of recommendations for GLT deprescribing in older diabetic adults.

Results: Of the 447 inpatients (women: 52.8%; median age: 83 years), 89% had a poor health status, 55.2% were overtreated, and 99.5% had at least one deprescribing flag, and 44% of patients with ≥1 deprescribing flag were not overtreated. Over the five-year follow-up, median survival time after hospitalisation was 1.83 years [IQR, 1.56-2.09], with an 18% five-year survival [95%CI, 15%-22%]. Both overtreatment (HR [95%CI] 1.29 [1.04;1.58]) and accumulation of flags for deprescribing (HR [95%CI] 1.20 [1.11;1.29] per additional flag) were associated with increased mortality. Patients meeting both ≥4 flags for deprescribing and overtreated had the worst survival outcomes (10% five-year survival rate).

Conclusion: The cumulative burden of deprescribing flags provided complementary information beyond the current overtreatment definition. Patients classified as overtreated and combining several deprescribing criteria highlight a particularly vulnerable subgroup of patients with poor survival in which GLT use should be systematically reassessed.