Background
The pharmacotherapeutic guidelines for type 2 diabetes have changed considerably during the past decades. SGLT2 inhibitors and GLP-1 receptor agonists have emerged as first-line agents by preventing cardiovascular events within a few years of treatment. In contrast, sulphonylureas and insulin have been deprioritised due to less beneficial effects and the risk of hypoglycaemia—particularly in older people who are frail. We hypothesised that medications with a high risk of hypoglycaemia were used more often in older people compared with younger people.

Methods
In a nationwide cohort of people with type 2 diabetes in Denmark from 2019 to 2020, we described the use of specific glucose-lowering medications in relation to age and glycated haemoglobin A1C (HbA1c) by descriptive statistics and regression models adjusted for sex, socioeconomic factors, renal function, and several comorbidities.

Findings
Among 290 890 people with type 2 diabetes, glucose-lowering medication usage peaked at age 70 years. Increasing age was associated with relatively less use of metformin, GLP-1 receptor agonists, and SGLT2 inhibitors and more use of basal insulin, DDP-4 inhibitors, and sulphonylureas. When comparing 80-year-olds with 60-year-olds at similar HbA1c levels of 6·5% (48 mmol/mol), 80-year-olds used 8% (95% CI 7–10%) fewer glucose-lowering medications, were 55% less likely to receive GLP-1 receptor agonists or SGLT2 inhibitors (relative ratio 0·45, 95% CI 0·42–0·48), and 65% more likely to receive sulphonylureas (1·65, 1·54–1·76). Among 23 032 individuals aged 80 years or older with HbA1c levels of less than 6·5% (<48 mmol/mol), 2291 (10%) used sulphonylureas or insulin.

Interpretation
In Danish people with type 2 diabetes, the likelihood of using glucose-lowering medications with a high risk of hypoglycaemia (eg, sulphonylureas and basal insulin) increased with age, whereas the likelihood of using GLP-1 receptor agonists and SGLT2 inhibitors decreased. Some people aged 80 years or older with an HbA1c level of less than 6·5% (48 mmol/mol) were potentially overtreated with sulphonylureas or insulin. These findings emphasise the importance of frequently re-evaluating glucose-lowering treatments.