OBJECTIVES Several commonly prescribed enzyme-inducing anti-epileptic drugs (EIAEDs) stimulate the synthesis of some hepatic enzymes responsible for drug metabolism. This synthesis can lead to complications by altering endogenous metabolic pathways or by affecting the elimination of concomitant drugs thus increasing healthcare costs. This study aimed to systematically review published estimates of direct and indirect costs associated with the use of EIAEDs compared with non-enzyme-inducing anti-epileptic drugs (nEIAEDs) in patients with focal and generalised seizures, and to evaluate methodological differences between the studies. METHODS Comprehensive electronic searches were undertaken using MEDLINE, EMBASE, Cochrane Library, EconLit, relevant conference proceedings and cost effectiveness analysis registries. All studies reporting any direct and indirect costs of AEDs for the treatment of patients with epileptic seizures were included. Study quality assessment was performed for every included study using a predesigned check list. RESULTS Thirty-seven full-length articles and two abstracts reporting costs were reviewed. Two studies reported AED costs, drug-specific adverse event costs and non-drug healthcare costs subsequent to the initiation of each individual AED (medical visits, MRI scans, etc.). Six studies reported specific AED costs and the overall subsequent non-drug healthcare cost without stratification by event.  Eighteen studies reported AED acquisition costs but did not report any other subsequent AED-related healthcare costs stratified by treatment. Thirteen studies reported the whole cost of illness with only a list of AEDs included. To date, no study has been specifically designed to compare the total costs between EIAED and nEIAED use, although some studies compared direct and indirect costs of several newer AEDs versus older AEDs. CONCLUSIONS Insufficient data and heterogeneity in methodology prevent valid comparisons being made between the total cost of EIAEDs and nEIAEDs. More research is required to identify if meaningful differences in the total cost of treatment exist between EIAEDs and nEIAEDs.