The contribution of glial cells, mainly astrocytes and microglia, to the pathophysiology of epilepsy is increasingly appreciated. Glia play a pivotal role in the initiation and maintenance of the central nervous system (CNS) immune response and neuronal metabolic and trophic supply. Recent clinical and experimental evidence suggests a direct relationship between epileptic activity and CNS inflammation, which is characterized by accumulation, activation, and proliferation of microglia and astrocytes. Concomitant glia-mediated mechanisms of action of several antiepileptic drugs (AEDs) have been proposed. However, their direct effects on glial cells have been rarely investigated. We aimed to investigate the effect of commonly used AEDs on glial viability, the gap junctional network, the microglial activation, and cytokine expression in an in vitro astroglia/microglia co-culture model. Primary astrocytic cultures were prepared from brains of postnatal (P0-P2) Wistar rats and co-cultured with a physiologic amount of 5%, as well as 30% microglia in order to mimic inflammatory conditions. Co-cultures were treated with valproic acid (VPA), carbamazepine (CBZ), phenytoin (PHE), and gabapentin (GBT). Viability and proliferation were measured using the tetrazolium (MTT) assay. The microglial activation state was determined by immunocytochemical labeling. The astroglial connexin 43 (Cx43) expression was measured by Western blot analysis. The transforming growth factor-β1 (TGF-β1) and tumor necrosis factor-α (TNF-α) cytokine levels were measured by the quantitative sandwich enzyme immunosorbent assay (ELISA). Astrocytes, co-cultured with 5% microglia (M5 co-cultures), showed a dose-dependent, significant reduction in glial viability after incubation with PHE and CBZ. Furthermore, VPA led to highly significant microglial activation at all doses examined. The antiinflammatory cytokine TGF-β1 release was induced by high doses of GBT and PHE. Astrocytes co-cultured with 30% microglia (M30 co-cultures) revealed a dose-dependent significant reduction in glial viability after incubation with PHE, accompanied by increased TGF-β1 and TNF-α levels. However, CBZ significantly reduced the amount of activated microglial cells and increased the total number of inactivated microglia. Finally, CBZ resulted in reduced viability at all doses examined. CNS inflammation is characterized by a disturbance of glial cell functions. Strong microglial activation, a typical hallmark of inflammation, was induced by VPA in M5 and continued in M30 co-cultures. With regard to the direct relation between CNS inflammation and seizures, VPA seems to be unsuitable for reducing inflammatory conditions. The reverse effect was achieved after CBZ. We noticed significant microglial inactivation, after incubation of the M30 co-cultures. In conclusion, we suggest that AEDs with antiinflammatory glial features are beneficial for seizures caused by persistent brain inflammation.