Insulin-treated older adults with Alzheimer's disease and related dementias (ADRD) who used a continuous glucose monitor (CGM) had a 43 percent lower rate of death and a 14 percent lower rate of any hospitalization over follow-up than matched patients who checked blood sugar with fingerstick tests, according to a retrospective cohort study published in JAMA Network Open on December 2, 2025. The hazard ratios were 0.57 (95% CI, 0.48 to 0.67) for all-cause mortality and 0.86 (95% CI, 0.76 to 0.96) for all-cause hospitalization, both statistically significant.

Researchers at the University of Florida drew a 15 percent random sample of national Medicare fee-for-service claims from January 2016 through December 2020, identifying 1,011 therapeutic CGM users and matching them, by propensity score, to 1,011 people using self-monitored blood glucose (SMBG, or fingersticks). All were 66 or older, on insulin, and diagnosed with diabetes and ADRD; people in hospice care were excluded. Mean age was 76.4. Current diabetes guidelines already tell clinicians to weigh a patient's cognitive status when choosing how tightly to manage blood sugar, but people with dementia have been routinely left out of the device trials that established CGM's benefits in other groups, leaving this comparison unanswered until now.

The outcomes CGM is specifically designed to prevent, however, told a less clear story. Hypoglycemia hospitalizations were lower with CGM but not significantly so (hazard ratio, 0.66; 95% CI, 0.40 to 1.08), as were falls (hazard ratio, 0.86; 95% CI, 0.68 to 1.08). Hyperglycemia crises actually trended higher among CGM users, also without reaching significance (hazard ratio, 1.38; 95% CI, 0.99 to 1.94). In subgroup analyses, men who used CGM did show a significant drop in hypoglycemia hospitalizations, and patients with type 1 diabetes saw significantly fewer falls and hospitalizations on CGM, but a significantly higher risk of hyperglycemia crisis. A negative control outcome, respiratory infections, showed no difference between groups, which the authors used as a check against gross unmeasured bias.

The study's authors, writing in the journal, characterized the findings as evidence of "potential long-term benefits" of CGM in this population, alongside "the need for clinical trials." They noted the design cannot separate a true effect of the device from differences between the kind of patient, or caregiver, who ends up prescribed a CGM instead of sticking with fingersticks.