Early glycemic control and subsequent risk of incident dementia among new metformin users

Melinda C. Power, Department of Epidemiology, George Washington University, Milken Institute School of Public Health, Washington, DC, USA. Electronic address: power@gwu.edu.
Morten Madsen, Department of Clinical Epidemiology and Center for Population Medicine, Aarhus University and Aarhus University Hospital, Aarhus, Denmark. Electronic address: morten.madsen@clin.au.dk.
Katrine Bødkergaard, Department of Clinical Epidemiology and Center for Population Medicine, Aarhus University and Aarhus University Hospital, Aarhus, Denmark. Electronic address: kani@clin.au.dk.
Scott C. Zimmerman, Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, USA; Department of Epidemiology, Boston University, Boston, MA, USA. Electronic address: Scott.Zimmerman@ucsf.edu.
Sarah F. Ackley, Department of Epidemiology, Boston University, Boston, MA, USA. Electronic address: sarah_ackley@brown.edu.
Paola Gilsanz, Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, USA; Kaiser Permanente Division of Research, Pleasanton, CA USA. Electronic address: Paola.Gilsanz@kp.org.
M Maria Glymour, Department of Epidemiology, Boston University, Boston, MA, USA. Electronic address: mglymour@bu.edu.
Victor W. Henderson, Department of Clinical Epidemiology and Center for Population Medicine, Aarhus University and Aarhus University Hospital, Aarhus, Denmark; Departments of Epidemiology and Population Health and of Neurology and Neurological Sciences, Stanford University, Stanford, CA, USA. Electronic address: vhenderson@stanford.edu.
Henrik Toft Sørensen, Department of Clinical Epidemiology and Center for Population Medicine, Aarhus University and Aarhus University Hospital, Aarhus, Denmark; Department of Epidemiology, Boston University, Boston, MA, USA; Departments of Epidemiology and Population Health and of Neurology and Neurological Sciences, Stanford University, Stanford, CA, USA. Electronic address: hts@clin.au.dk.
Reimar W. Thomsen, Department of Clinical Epidemiology and Center for Population Medicine, Aarhus University and Aarhus University Hospital, Aarhus, Denmark. Electronic address: rwt@clin.au.dk.

Document Type

Journal Article

Publication Date

11-4-2025

Journal

The American journal of medicine

DOI

10.1016/j.amjmed.2025.10.017

Keywords

Dementia; Diabetes mellitus; Glycated hemoglobin; Glycemic control; Metformin; Risk

Abstract

BACKGROUND: People with diabetes have elevated dementia risk. Optimal glycemic treatment targets for dementia risk reduction among individuals initiating metformin monotherapy are unknown. METHODS: We used Danish health registry data to identify metformin monotherapy initiators aged 50+ in Northern Denmark during 2000-2018. We examined the relation between early glycemic control, HbA1c reduction magnitude, and their combination with incident dementia using standard hazard models and propensity-score based restricted and weighted models, which better address confounding and mimic a clinical trial. We explored effects by age, calendar year of metformin initiation, and presence of cardiovascular disease. RESULTS: Among 46,332 metformin initiators (median age 65 years), 83% achieved an HbA1c level <7% (<53 mmol/mol) within one year after metformin initiation and 1,432 (3.2%) developed dementia over a median 5.6 years of follow-up. Both standard and propensity-score restricted/weighted analyses suggested elevated dementia risk at achieved HbA1c levels >7% (>53 mmol/mol), but little difference in dementia risk among those achieving tight (HbA1c of 6.5% to <7%/48 to <53 mmol/mol) or very tight glycemic control (HbA1c of 6% to <6.5%/42 to <48 mmol/mol or <6%/<42 mmol/mol). While P values for interaction were non-significant, point estimates suggested benefits of early glycemic control on dementia risk were limited to persons without established cardiovascular disease and take years to manifest. CONCLUSION: Our findings support established recommendations for early glycemic control (i.e., HbA1c <7%/<53 mmol/mol) in type 2 diabetes for dementia risk reduction, with little evidence of benefit for pursuing lower targets (i.e. <6.5%/<48 mmol/mol, <6%/<42mmol/mol).

Department

Epidemiology

Link to Full Text

Share

COinS