OBJECTIVE: Long-term antiepileptic drug (AED) use has been associated with bone disease, but many previous studies have been limited by inadequate control subjects. We used a cotwin affected sib-pair model to investigate this issue. METHODS: The authors studied 31 female twin (15 monozygous and 16 dizygous) and four sibling pairs (< 3 years age difference) aged 21 to 75 years, in which one member had > 12 months of AED treatment. Areal bone mineral density (ABMD, g/cm2) was measured at the lumbar spine (LS), total hip (TH), femoral neck (FN), and total forearm (FA). Three primary a priori defined subgroups were analyzed: a) use for > 2 years, b) use of enzyme-inducing AEDs, or c) age older than 40 years. RESULTS: For all pairs (n = 35), there were no significant within-pair differences in any ABMD measure. However, in Subgroup a (n = 27), there was a within-pair difference at the FA (0.513 vs 0.534, -3.9%, p = 0.016). In Subgroup b (n = 29), there was also a within-pair difference at the FA for AED user vs nonuser (0.508 vs 0.529, -3.8%, p = 0.010). In Subgroup c (n = 15), there were within-pair differences at the FA (0.492 vs 0.524, -6.1%, p = 0.017) and the LS (0.884 vs 0.980, -9.8%, p = 0.036). CONCLUSIONS: Patients using AEDs for > 2 years, in particular those taking enzyme-inducing AEDs and those older than 40 years, have significantly lower bone mineral density at clinically relevant fracture risk sites.