Abstract
PURPOSE: To compare recoverable Enterobacteriaceae patterns in adults with epilepsy and controls across peri-urban and urban Ghana, and to explore associations with comorbidity and antiseizure medication (ASM) regimen. METHODS: This dual-site cross-sectional study enrolled 120 adults (70 people with epilepsy [PWE], 50 controls) from Tarkwa Municipal Hospital (peri-urban) and KNUST Hospital, Kumasi (urban). Stool samples were subjected to enrichment in Brilliant Green Bile Broth followed by selective/differential culture for Enterobacteriaceae. Recoverable CFU/g-equivalents were defined as colony-forming units recovered after enrichment and selective culture, back-calculated to the gram of stool input, and interpreted as a standardized recoverability measure (not absolute stool bacterial load). Isolates were identified by MALDI-TOF MS (Bruker) using manufacturer score thresholds (≥ 2.0 species-level; 1.70–1.99 genus-level). Richness was defined as the number of distinct MALDI-TOF–confirmed Enterobacteriaceae taxa recovered per participant. RESULTS: In the peri-urban cohort, PWE had lower setting-dependent recoverable burden than controls (11.20 ± 2.44 vs. 20.95 ± 3.95 [recoverable CFU/g-equivalents, post-enrichment]; p < 0.0001). In the urban cohort, PWE had higher setting-dependent recoverable burden than controls (17.48 ± 4.00 vs. 11.29 ± 1.86 [recoverable CFU/g-equivalents, post-enrichment]; p = 0.002). Site-dependent differences were observed for Citrobacter freundii, Proteus mirabilis, and Enterobacter hormaechei. Comorbidities were more frequent in urban PWE (50%) than peri-urban PWE (15%); among peri-urban PWE, E. hormaechei recoverable abundance was higher in comorbid versus non-comorbid participants (p = 0.03). In urban PWE, ASM polytherapy was associated with higher psychosis-related symptom scores than ASM monotherapy (p < 0.0001). CONCLUSION: Recoverable Enterobacteriaceae-associated patterns in epilepsy differed by setting under enrichment and selective culture conditions, supporting context-specific dysbiosis hypotheses in Ghana and motivating larger longitudinal, sequencing-based studies with environmental and standardized clinical phenotyping. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12866-026-04901-3.