Abstract
BACKGROUND AND OBJECTIVES: Antibiotics rapidly reduce intestinal bacterial diversity, leading to dysbiosis that persists for months to years. Although emerging evidence from retrospective and claims-based studies has linked dysbiosis to cognitive function, prospective data are lacking. We aim to examine the prospective association of antibiotics with cognitive aging among initially healthy older adults. METHODS: We leveraged data from prospective follow-up and observational extension of ASPirin in Reducing Events in the Elderly, a completed randomized trial of community-based Australian older adults. Among participants whose prescription records were available and without dementia during the first 2 years of follow-up, we identified any or repeated antibiotic use based on the Anatomical Therapeutic Chemical code (J01). We assessed the associations of antibiotic use during the first 2 years with longitudinal changes in standardized composite and domain-specific cognitive scores (global cognition, episodic memory, language and executive function, and psychomotor speed) using linear mixed models, and with incident, clinically adjudicated dementia (Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition criteria) and incident cognitive impairment, no dementia (CIND, without a dementia trigger but with significant, nontransient decline), using Cox proportional hazard models. RESULTS: Over a median of 4.7 years after the second follow-up visit, we documented 461 dementia and 2,576 CIND cases among 13,571 participants (mean age ± SD 75.0 ± 4.1 years, 54.3% female). Compared with nonuse, antibiotic use was not associated with increased risks for dementia (hazard ratio [HR] 1.03; 95% CI 0.84-1.25), CIND (HR 1.02; 95% CI 0.94-1.11), or subsequent declines in cognitive scores, after adjusting for sociodemographic, lifestyle factors, family history of dementia, baseline cognitive function, and medications known to affect cognition. There were also no associations according to the cumulative frequency of antibiotic use, long-term use, specific antibiotic classes (e.g., beta-lactams, tetracyclines, and sulfonamides), and subgroups defined by risk factors. DISCUSSION: Among initially healthy older adults, any or repeated antibiotic use was not associated with incident dementia, CIND, or accelerated cognitive decline. Although prescription data may not reflect the actual use, we examined the frequency of antibiotics within a defined period to capture the extent and duration of antibiotic exposure. Our results do not support an association between antibiotic-associated gut microbiome disruption and dementia risk.