Serious Hypoglycemia and the Risk of Dementia: A Population-based Cohort Study Using Exposure Density Sampling

Background: Hypoglycemia, a side effect of diabetes pharmacotherapy, has been linked to cognitive impairment.

Objectives: To assess the association between serious hypoglycemia and the risk of all-cause dementia using real-world data.

Methods: A population-based retrospective cohort study using administrative healthcare data (1996-2018) from British Columbia, Canada. From a cohort of individuals aged 40-70 years with incident type 2 diabetes, we created a dynamic sub-cohort of hypoglycemia exposed and unexposed individuals using exposure density sampling with replacement, in which those unexposed are eligible to be exposed. Four unexposed individuals were randomly selected within risk sets based on diabetes duration and age. Serious hypoglycemia was defined as ≥1 episode requiring hospitalization or a physician visit. A 2-year dementia-free latency period was required wherein the exposure index date was assigned as 2 years after the date of the first hypoglycemic episode for exposed and the equivalent date based on diabetes duration for those unexposed. Patients were followed until dementia diagnosis, death, emigration, or 31Dec2018. We used a high-dimensional propensity score to calculate the stabilized inverse probability of treatment weights (IPTW) to adjust for 530 potential confounders. We estimated the hazard ratio (HR, 95% CI) of dementia using weighted conditional cause-specific hazards risk models with death as a competing risk.

Results: Among 10,425 patients with incident type 2 diabetes, 2,085 experienced a serious hypoglycemic episode over a median (IQR) follow-up of 4.6 (5.7) years. The unadjusted incidence rate (95% CI) of dementia was 10.91 (9.46-12.94) per 1000 person-years for those exposed and 5.23 (4.59-5.97) per 1,000 person-years for those unexposed. IPTW resulted in well balanced groups across all the included potential confounders (mean age (SD)=64.3 (9.0); 45.4% females), with weighted incidence rates (95% CI) of 8.14 (6.53-10.14) per 1000 person-years for those exposed and 5.58 (4.83-6.44) per 1,000 person-years for those unexposed. The risk of dementia was twice as high among those exposed (vs not) to hypoglycemia (weighted HR, 95% CI = 1.98, 1.54-2.54).

Conclusions: Serious hypoglycemia episodes contribute to an increased risk of all-cause dementia.