Association of gabapentin and neuropathological hallmarks in an autopsy cohort of older adults

Abstract

AbstractBackgroundPrevious studies of the association between gabapentin use and neurocognitive change have shown conflicting results. We examined the association between end‐of‐life gabapentin use and neuropathological features in older adults using National Alzheimer’s Coordinating Center Uniform Data Set and Neuropathology Data Set (9/2005‐3/2022).MethodsWe conducted a cross‐sectional study including participants with at least one visit within two years before autopsy and age at visit 65+ years. Gabapentin use was determined based on the two‐year window. Outcomes of interest included burden of tau (Braak stage), amyloid (Thal stage, diffuse plaques, and neuritic plaques), α‐synuclein (Lewy bodies), cerebrovascular disease (infarcts, microinfarcts, and arteriosclerosis), hippocampal sclerosis, cerebral amyloid angiopathy, TDP‐43, and atrophy (cerebral cortex and hippocampus). Collection of Thal stage, TDP‐43, and atrophy information occurred after January 2014; these outcomes were assessed in the subset with data. For each dependent neuropathological hallmark, two models were constructed (M1: adjusted for age, sex, APOE, center, and cognitive status; M2: adjusted for M1 confounders, agitation, depression, antiseizure drug, and use of opioids). To mitigate bias, joint stabilized inverse probability of treatment and censoring weights were used.ResultsWe included 213 gabapentin‐users (mean age[SD]: 80.9[8.7]; male 53.5%) and 4107 non‐users (82.0[8.7]; 54.7%). Gabapentin use was associated with reduced odds of Braak stage IV‐VI vs. 0‐III, moderate/frequent vs. non/sparse diffuse plaques, presence of Lewy bodies, infarct, and cerebral amyloid angiopathy in M1 and M2, and hippocampal sclerosis in M1 (Figures 1A and 1B). For the subset outcomes, 116 gabapentin‐users (mean age [SD]: 81.0 [8.6]; male 50.9%) and 1919 non‐users (81.8 [8.9]; 54.3%) were included. Gabapentin use was associated with increased odds of Thal phase 3‐5 vs. 0‐2 in M1 and M2, and amygdala TDP‐43, hippocampus TDP‐43, and cerebral cortex atrophy in M2; whereas gabapentin use was associated with decreased odds of hippocampus atrophy in M1 and M2, and amygdala TDP‐43, hippocampus TDP‐43, and cerebral cortex atrophy in M1 (Figures 1C and 1D).ConclusionEnd‐of‐life gabapentin use was not significantly associated with burden of neuropathologic features and there were no clear patterns. Further studies with increased sample sizes and better measurement of gabapentin use are needed.