Answers are grounded in this paper and cite their sources.
Retrospective cohort study finds HbA1c trajectories after SGLT2i initiation predict dementia risks in adults with diabetes, indicating intervention opportunities.
BackgroundType 2 diabetes mellitus (T2DM) is a major risk factor for dementia, yet the prognostic value of early dynamic glycemic changes following glucose-lowering therapy remains underinvestigated.ObjectiveTo determine whether early glycated hemoglobin (HbA1c) trajectories after initiating sodium-glucose cotransporter 2 inhibitors (SGLT2i) predict long-term dementia and mortality risks.MethodsThis retrospective new-user cohort study utilized electronic health records from the TriNetX Global Network. Adults with T2DM initiating their first SGLT2i were included. To mitigate immortal-time bias, a 1-year landmark design was applied. Patients were stratified by baseline HbA1c and classified into improved, stable, or worsened trajectories based on values 91-455 days post-initiation. Propensity-score matching was performed within strata. The primary outcome was all-cause dementia.ResultsAmong 172,050 matched patients, modest HbA1c worsening in those with Good baseline control (<7.0%) did not increase dementia risk (HR 0.93; 95% CI, 0.75-1.15). However, in patients with Fair baseline control (7.0-8.9%), worsening to Poor control significantly increased dementia (HR 1.39; 95% CI, 1.11-1.74) and mortality risks. Conversely, among those with Poor baseline control (≥9.0%), trajectory improvement conferred substantial neuroprotection, reducing the risk of dementia (HR 0.64; 95% CI, 0.51-0.79) and mortality (HR 0.71; 95% CI, 0.60-0.84). These findings provide clinically actionable evidence linking glycemic dynamics to neurodegenerative risk, particularly Alzheimer's disease and related dementias.ConclusionsDynamic HbA1c trajectories following SGLT2i initiation independently predict dementia risk. Integrating trajectory-based assessments into routine care provides an actionable, scalable biomarker to guide timely treatment intensification and mitigate diabetes-related neurocognitive decline.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: