Background People with Type 2 diabetes mellitus (T2DM) are at increased risk of developing dementia. Evidence suggests that thiazolidinediones (TZDs) may be protective for dementia onset including Alzheimer’s disease and vascular dementia, compared to other second-line antidiabetic medications (SAMs). However, causality remains uncertain due to methodological limitations. We examined the effect of TZD on the risk of vascular dementia and all-cause dementia in T2DM, compared to other second-line treatments.

Methods We emulated a pragmatic randomised trial using UK primary care data, Clinical Practice Research Datalink Aurum, between 2003 and 2023 to estimate the comparative effectiveness of initiating a TZD, dipeptidyl peptidase-4 (DPP-4) inhibitors, sodium-glucose cotransporter-2 (SGLT2) inhibitors, or sulfonylurea (SU) against incident dementia in T2DM adults on metformin therapy. Patients were followed for up to 5 years from 180 days after their first SAM prescription. We used overlap weighting to adjust for baseline confounding and fitted double robust Cox models to estimate adjusted hazard ratios (aHRs).

Findings This study included 124,311 participants (mean age 63 years, 61% males, and 20% whites), of whom 595 developed vascular dementia and 1,678 developed all-cause dementia during follow-up. On top of metformin, 8,669 initiated TZD, 30,216 initiated DPP-4 inhibitors, 55,997 initiated SU and 29,429 initiated SGLT2 inhibitors. TZD were associated with a similar risk of vascular dementia compared with DPP-4 inhibitors (aHR 0.89;95% CI 0.36–2.23) and SU (0.58;0.24–1.42). SGLT2 inhibitors were associated with a lower risk of vascular dementia than TZD (0.29;0.09–0.94), DPP-4 inhibitors (0.25;0.10–0.64), and SU (0.17;0.07–0.40). Most patterns persisted in all-cause dementia: SGLT2 inhibitors vs DPP-4 inhibitors (0.51;0.26– 0.99) and SGLT2 inhibitors vs SU (0.35;0.18–0.67), with no difference observed between SGLT2 inhibitors and TZDs.

Interpretation Dementia risk was similar for TZDs, DPP-4inhibitors and SUs but was significantly lower for SGLT2 inhibitors, a finding that warrants further investigation. Considering potential cognitive effects when selecting therapies for T2DM is important in an ageing population.

Funding This work was funded by Alzheimer’s Research UK (grant: ARUK-PPG2023B-036) and the National Institute for Health and Care Research (NIHR) Greater Manchester Patient Safety Research Collaboration (GM PSRC).

Evidence before this study We searched for relevant studies in PubMed database with the search term (“thiazolidinedione” OR “TZD” OR “pioglitazone” OR “rosiglitazone”) AND (“diabet*” OR “T2D” OR “T2DM”) AND (“dementia*” OR “cognitive impairment” OR “Alzheimer*”) from inception until June 15, 2025. We identified 12 relevant cohort studies that examined the association between thiazolidinedione (TZD) and vascular dementia, as well as Alzheimer’s disease and all-cause dementia. The evidence from these studies is inconsistent; two of the studies suggested that TZD use was associated with a significant reduction in risk of vascular dementia, while one suggested that the risk was comparable. Current evidence suggests that TZD use were associated with a reduced risk of incident Alzheimer’s disease and all-cause dementia. Preclinical studies suggest that TZDs may reduce the risk of vascular dementia by improving cerebral perfusion, enhancing endothelial function, and reducing impairments in learning, memory, and blood–brain barrier permeability.

However, these findings are limited using a non-representative sample, insufficient adjustment for confounding, inadequate outcomes definition, and an increased risk of a false-positive result purely by chance.

Added value of this study To our knowledge, this is the first study in the UK to compare the effectiveness of TZDs with other second-line antidiabetic medications on the risk of vascular and all-cause dementia. To address methodological limitations in prior studies, we applied a rigorous causal inference framework and emulated a pragmatic randomised trial using a large, nationally representative primary care electronic health record dataset to minimise bias. We found that initiating TZDs as second-line therapy provided a comparable risk of vascular dementia and all-cause dementia as other second-line medications. Sodium-glucose cotransporter-2 (SGLT2) inhibitors were associated with a lower risk of vascular dementia than TZDs, and with a reduced risk of both vascular dementia and all-cause dementia compared with dipeptidyl peptidase-4 inhibitors and sulfonylureas.

Implications of all the available evidence Our findings do not support a cognitive benefit of TZDs when used as a second-line therapy for type 2 diabetes mellitus. In contrast, SGLT2 inhibitors may offer cognitive benefits, complementing their recognised cardiovascular and renal benefits, although further research is required to confirm their causal effects.

MKR reports consulting fees from Eli Lilly and modest stock ownership in GSK unrelated to this work. The remaining authors have nothing to declare.

This work was funded by Alzheimer's Research UK (grant: ARUK-PPG2023B-036) and the National Institute for Health and Care Research (NIHR) Greater Manchester Patient Safety Research Collaboration (GM PSRC). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.

Yes

The details of the IRB/oversight body that provided approval or exemption for the research described are given below:

The study was approved by the CPRD's Independent Scientific Advisory Committee (approval number: 24_003967). CPRD also has ethical approval from the Health Research Authority to support research using anonymised patient data (research ethics committee reference 21/EM/0265). Individual patient consent was not required, as all data were deidentified.

I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals.

Yes

I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).

Yes

I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable.

Yes

This study used pseudonymised patient-level data from the CPRD. To protect patient confidentiality, we cannot publish patient-level data. Other researchers can use patient-level CPRD data in a secure environment by submitting an application to the CPRD Independent Scientific Advisory Committee. Details of the application process and access conditions are provided by the CPRD at https://www.cprd.com/Data-access.

https://www.cprd.com/Data-access