The Interplay Between Ketogenic Diet and the Microbiota-Gut-Brain Axis in Drug-Resistant Epilepsy: A Systematic Review

Context The ketogenic diet (KD) is an established nonpharmacological therapy for drug-resistant epilepsy (DRE). Emerging evidence suggests the KD can modulate the microbiota-gut-brain axis (MGBA) and that this interaction may influence clinical response to the intervention. Objective To synthesize current evidence on MGBA alterations following a KD in the setting of DRE (DRE+KD) and on the associations between MGBA parameters and clinical outcomes after the diet. Data sources A systematic search was conducted of the PubMed/MEDLINE, Embase, Web of Science, Scopus, and Cochrane Library databases in July 2025. Data Extraction Two reviewers independently screened 799 records and extracted data from the eligible studies. We included original, peer-reviewed, longitudinal research. Findings were synthesized narratively, and the risk of bias (RoB) was appraised using the Methodological Index for Non-Randomized Studies. Data analysis Seven articles (n = 93 patients with DRE+KD) met the predefined eligibility criteria. We found a decreased abundance of the phyla Firmicutes and Actinobacteria and increased abundance of Bacteroidetes after KD. Functional routes involved in carbohydrate metabolism were diminished, and fatty acid-related pathways were increased. Levels of fecal short-chain fatty acids (SCFAs) and circulating inflammatory proteins also decreased, whereas those of ketone-related metabolites increased. While health-promoting MGBA features (eg, Bifidobacterium longum, B. breve, Faecalibacterium prausnitzii, SCFAs, plasmalogens) were associated with better seizure control, other taxa (eg, Ruminococcaceae, Escherichia coli, infant-type Bifidobacterium) were linked to poorer outcomes. Overall RoB was moderate to high. Conclusions Our findings suggest the KD induces recurrent MGBA-related changes in patients with DRE. Additionally, some taxonomic and metabolic profiles are associated with better seizure control. However, a moderate to high RoB, small samples, high heterogeneity, and predominantly hospital and pediatric cohorts limit the certainty and generalizability of the evidence. Nonetheless, the findings provide a rationale for further studying the interplay among KD, the MGBA, and DRE. Future research should prioritize including adults, larger samples, deep sequencing platforms, and explore potential microbiota-informed adjunctive therapies for DRE. Systematic Review Registration PROSPERO registration no. CRD420251085021.