Document resource
Background Metformin demonstrates antiepileptogenic (primary prevention) and antiseizure (symptom control) properties in animal models of epilepsy. It is unclear whether these effects translate to humans, particularly outside diabetes, where metformin is increasingly prescribed for weight loss for example. We therefore addressed two questions in people without evidence of diabetes: (1) following an incident brain insult (ischaemic stroke, intracranial haemorrhage, head injury or brain tumour), is metformin associated with reduced risk of first recorded seizure, epilepsy diagnosis or antiseizure medication (ASM) initiation? (2) In established epilepsy treated with ASMs, is metformin add-on associated with reduced risk of potentially seizure-related morbidity (coded seizures, emergency department attendances, hospital admissions, falls, injuries, burns or a composite of these) and all-cause mortality?Methods We undertook an international retrospective cohort study of real-world healthcare data in TriNetX. Propensity score matched cohorts compared metformin against licensed weight-management comparators. Cox proportional hazards models estimated HRs with 95% CIs for 10 year outcomes.Results 16 695 participants were included. Following brain insult, metformin was not associated with reduced risk of first recorded seizure, epilepsy diagnosis or ASM initiation. In established epilepsy, metformin add-on was associated with significantly lower hazards of coded seizures (HR 0.778, CI 0.624 to 0.971), emergency attendance or hospital admission (HR 0.696, CI 0.569 to 0.851), falls/injuries/burns (HR 0.677, CI 0.510 to 0.898) and a composite of these events (HR 0.729, CI 0.626 to 0.848; approximate number needed to treat ~6 at 10 years). Mortality did not differ (HR 1.039, CI 0.594 to 1.816).Conclusions Metformin could have repurposed benefits in epilepsy. This warrants further investigation.