Blood-based Neurological Biomarkers Following Dipeptidyl Peptidase-4 Inhibitors Exposure among Older Adults with Type 2 Diabetes
In older adults with type 2 diabetes, DPP-4i exposure is linked to lower plasma GFAP and NfL levels, suggesting a protective benefit for glial and neuronal health that merits further clinical investigation.
Itroduction:
Dipeptidyl peptidase-4 inhibitors (DPP-4i) are commonly prescribed antidiabetic medications with reported neuroprotective potential.
The association between DPP-4i exposure and measures of glial activation, neurodegeneration and Alzheimer's disease pathology remains unclear.
We examined the association between DPP-4i exposure and circulating levels of Glial Fibrillary Acidic Protein (GFAP), Neurofilament Light Chain (NfL) and phosphorylated tau217 (p-tau217) in older adults with type-2 diabetes (T2D).
Methods:
Data source - We utilized cross-sectional digital datasets from the Israel Diabetes and Cognitive Decline study and Maccabi Healthcare Services, a large non-profit health maintenance organization in Israel
Inclusion criteria: Individuals aged 65y and older; HbA1c ≥7.25% OR two separate glucose readings above 200 mg/dL, taken at least three months apart OR purchase of antidiabetic medication twice in the previous two months; At least three HbA1c measurements
Exclusion criteria:Individuals with dementia, MCI, cholinesterase inhibitor use, or significant cognitive-impairing illnesses; Missing all three blood-based biomarkers data; Missing covariate data
Exposure - The DPP-4i users' group was set as individuals who used DPP-4i as a single drug or in combination with other drugs for at least a year before their first clinic visit, based on electronic health records
Outcome - Log-transformed blood-based biomarkers levels of GFAP, NfL, and p-tau217
Statistical analysis - Available case analyses using generalized linear regression models; Models were adjusted for multiple potential confounders using a doubly robust approach that combined Inverse probability weights with multivariable regression adjustment; Accounting for 13 covariates: age at first clinic visit, years of education, sex, diabetes onset age, BMI, diabetes duration, HbA1c, triglycerides, LDL, systolic BP, CVD, eGFR, and apolipoprotein-E (APOE) genotype; Complete-case sensitivity analyses were performed using similar methods.
Results: N= 568; Median age 71.3 (Q1-Q3: 68.1-74.8); 38.6% women; n=59 (10.4%) in the DPP-4i group; The DPP-4i group had younger T2D diagnosis age, longer diabetes duration, and higher median HbA1c levels.
Main analysis:
GFAP: The main analysis demonstrated a statistically significant reduction of 10.5% (95% CI: -19.09 to -1.1). The unadjusted p-value was 0.030, which remained significant after False Discovery Rate (FDR) adjustment (adjusted p = 0.044).
NfL: Similarly, NfL showed a significant decrease of 11.0% (95% CI: -19.05 to -2.1) in the main analysis. The initial p-value of 0.017 remained statistically significant following FDR adjustment (adjusted p = 0.044).
p-tau217: The main analysis for p-tau217 indicated a 15.8% reduction (95% CI: -30.04 to 1.3). Because the confidence interval crosses zero, this result was not statistically significant (unadjusted p = 0.068, FDR-adjusted p = 0.068).
Conclusions: In this community-based study, DPP-4i exposure was associated with lower plasma levels of GFAP and NfL. Our findings lend tentative support to the potential repurposing of DPP-4i for neuroprotection
Limitations: Our cross-sectional observational study design restricts causal inference. The possibility of residual confounding cannot be ruled out. Our results do not generalizable beyond to individuals with diabetes with specific characteristics
