What the research found

Researchers examined treatment choices for over 58,000 adults aged 65+ with type 2 diabetes tracked between 2018 and 2026. They compared who received GLP-1 receptor agonists versus alternative drugs (SGLT2 inhibitors and DPP-4 inhibitors) and found striking patterns: GLP-1 recipients were younger (average 75 vs 78–80 years), and crucially, had substantially higher body weight—mean BMI of 34 versus 30 and 29 kg/m² respectively. About 72% of GLP-1 initiators had BMI ≥30, compared to roughly 35–42% in other groups.

Interestingly, blood sugar control measured by HbA1c was essentially identical across all three groups, sitting around 7.4–7.7%. GLP-1 recipients had fewer serious comorbidities—less dementia, heart failure, and kidney disease—and fewer hospitalizations in the year before starting treatment. When the researchers used statistical modelling to identify what actually predicted GLP-1 prescription, BMI above 27 kg/m² emerged as the dominant signal, with heart failure and prior hospitalisation working against it.

Why it matters for you

If you're tracking your weight and metabolic health on MyKine, this reflects real-world practice: clinicians appear to be selecting GLP-1 therapy primarily based on adiposity rather than blood glucose alone. This aligns with guideline recommendations for weight management alongside glucose control, but it matters for your expectations. If your main goal is optimizing HbA1c and your weight is already in a healthier range, this data suggests you may not fit the typical prescribing profile—even if you have diabetes.

The finding that healthier, younger patients with fewer comorbidities receive GLP-1s while those with existing heart failure or complex medical histories receive SGLT2 inhibitors instead also underscores that drug choice depends on your full clinical picture, not just one number. If you're considering or using GLP-1 therapy, you're likely someone whose doctors have identified weight as a meaningful therapeutic target alongside glucose control.

Caveats

• Observational only: Shows real-world patterns, not randomized evidence; cannot prove causation for why certain patients were chosen
• Regional healthcare system: Results may not reflect prescribing patterns elsewhere or across different insurance/healthcare models
• Older adults only: Findings don't apply to younger populations
• LLM-assessed cognition: The study used an AI language model to detect cognitive concerns from medical records—a method still being validated
• No outcome data: Study describes who gets what drug, not whether those choices led to better health outcomes