Premium
Relationships between oral hypoglycemic drugs and memory decline in people with type 2 diabetes: A stratified longitudinal observational study
Author(s) -
Wu CheYuan,
Ouk Michael,
Wong Yuen Yan,
Anita Natasha,
Edwards Jodi D.,
Yang Pearl,
Shah Baiju R.,
Kapral Moira,
Herrmann Nathan,
Lanctot Krista L.,
Macintosh Bradley J.,
Rabin Jennifer S.,
Black Sandra E.,
Swardfager Walter
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.041312
Subject(s) - medicine , confounding , type 2 diabetes , dementia , diabetes mellitus , poisson regression , metformin , thiazolidinedione , cognitive decline , demography , disease , endocrinology , population , sociology , environmental health
Background There is increased recognition that diabetes can elevate Alzheimer’s disease (AD) risk and exacerbate memory decline. However, few studies have compared memory decline among patients with diabetes using different oral hypoglycemic drug classes. Method Participants using any hypoglycemic medications from 2005 to 2019 were identified from the National Alzheimer's Coordinating Center database. Analyses were conducted separately in elderly with normal cognition (NC), amnestic mild cognitive impairment (aMCI), and dementia due to AD. Drug classes of interest included metformin, sulfonylureas, thiazolidinedione and dipeptidyl peptidase‐4 inhibitors (gliptins). Immediate and delayed recall outcomes were assessed using the Wechsler Memory Scale–Logical Memory test. Linear mixed‐effects models (in NC and aMCI) and mixed‐effects zero‐inflated Poisson or quasi‐Poisson regressions (in AD) were used to examine the associations between drug classes (time‐varying exposures) and memory performance over time. Confounding by indications were corrected by inverse probability treatment weighting. Models were adjusted for covariates including demographics, baseline MMSE, comorbidities and concurrent medications. In post‐hoc analyses, we examined whether ApoE genotype moderated the associations between drug class and memory. Result In NC (n=1,201, follow‐up years=3.43±3.27), metformin use was associated with better immediate (β=0.069 [0.011, 0.12]) and delayed (β=0.089 [0.032, 0.15]) memory over time, relative to no metformin use. No oral hypoglycemic drug class exhibited significant associations with memory in aMCI (n=671, follow‐up years=1.50±2.13). In AD (n=807, follow‐up years=1.90±2.21), DPP4 inhibitor use was associated with slower rate of decline in delayed memory (RR=1.22 [1.06, 1.40]), and thiazolidinedione use was associated with a faster rate of decline in immediate memory (RR=0.89 [0.82, 0.97]). In NC, DPP4 inhibitor use was associated with better delayed memory over time specifically among ApoE ε4 carriers (interaction: β=0.038 [0.0039, 0.072]). In AD, ApoE ε4 did not significantly interact with DPP4 inhibitor use, and thiazolidinedione use was associated with a faster rate of decline in immediate memory among non‐ε4 carriers (interaction: RR=1.26 [1.07, 1.47]). Conclusion Associations between memory and particular oral hypoglycemic medications may depend on the presence of AD symptoms. ApoE ε4 may further modify these relationships. Confirmatory analyses in other cohorts, particularly those with more complete clinical data, are required.