The association of gabapentin initiation and neurocognitive changes in older adults with normal cognition

Background: Gabapentin is increasingly prescribed to older adults, which raises concerns about its potential to cause neurocognitive changes. Therefore, we aimed to examine the association of gabapentin use with neurocognitive changes (i.e., cognitive decline, functional status decline, and motor function change) in older adults. Methods: We conducted a retrospective cohort study using the National Alzheimer's Coordinating Center Uniform Data Set (UDS; September 2005-March 2021 data freeze). From the eligible sample (≥age 65 years), we identified cognitively normal new-users of gabapentin and the visit they initiated gabapentin (i.e., index visit). Initiators were matched to randomly selected nonusers on year of UDS enrollment and visit number from enrollment to index. Cognitive decline was defined as any increase in the Clinical Dementia Rating global score (CDRGLOB) and as a 1-point increase in CDR sum of boxes (CDR-SB). Functional status decline was defined as a 3-point increase in the sum of the Functional Activities Questionnaire (FAQ) and as 0.3-point increase in mean FAQ. Decline in motor function was defined as new clinician reports of gait disorder, falls, and slowness. To mitigate confounding and selection bias, we used joint stabilized inverse probability of treatment weights and stabilized inverse probability of censoring weights. All analyses were conducted comparing index to index+1 and index+2 visits. Results: From the eligible UDS participants (N = 23,059), we included 480 initiators (mean age [SD]: 78.7 [6.9]; male 34.4%); 4,320 nonusers (78.3 [7.0]; 34.4%). Gabapentin initiation was significantly associated with cognitive/functional status decline: worsening CDRGLOB at index+1 visit (odds ratio [95% confidence interval]: 1.55 [1.07, 2.25]); CDR-SB at index+1 visit (1.94 [1.22, 3.09]); and mean of FAQ at index+2 visit (1.78 [1.12, 2.83]). After excluding initiators with extant motor dysfunction (n = 21), we identified 459 initiators (78.7 [6.9]; 34.0%) and 4,131 nonusers (78.2 [6.9]; 34.7%); in this sample, gabapentin initiation was associated with increased falls at the index+2 visit (2.51 [1.19, 5.31]). Conclusion: Gabapentin initiation was significantly associated with deleterious neurocognitive changes among older adults with initially normal cognition. Further studies are needed to examine the risk/benefit of prescribing gabapentin in older adults.