Antiepileptic drug therapy has significant inter-patient variability in response towards it. The current study aims to understand this variability at the molecular level using microarray-based analysis of peripheral blood gene expression profiles of patients receiving valproate (VA) monotherapy. Only 10 unique genes were found to be differentially expressed in VA responders (n = 15) and 6 genes in the nonresponders (n = 8) (fold-change >2, p < 0.05). PTGS2 which encodes cyclooxygenase-2, COX-2, showed downregulation in the responders compared to the non-responders. PTGS2/COX-2 mRNA profiles in the two groups corresponded to their plasma profiles of the COX-2 product, prostaglandin E 2 (PGE 2). Since COX-2 is believed to regulate P-glycoprotein (P-gp), a multidrug efflux transporter over-expressed at the blood-brain barrier (BBB) in drug-resistant epilepsy, the pathway connecting COX-2 and P-gp was further explored in vitro. Investigation of the effect of VA upon the brain endothelial cells (hCMEC/D3) in hyperexcitatory conditions confirmed suppression of COX-2-dependent P-gp upregulation by VA. Our findings suggest that COX-2 downregulation by VA may suppress seizure-mediated P-gp upregulation at the BBB leading to enhanced drug delivery to the brain in the responders. Our work provides insight into the association of peripheral PTGS2/COX-2 expression with VA efficacy and the role of COX-2 as a potential therapeutic target for developing efficacious antiepileptic treatment. Epilepsy, a multifactorial neurological disease, affects about 69 million people worldwide constituting nearly 1% of the world population 1. The treatment is primarily based on symptomatic pharmacological interventions i.e. antiepileptic drug (AED) therapy which controls the frequency of seizures in the patients. Despite the availability of appropriate therapy, there is significant inter-individual variability in AED response. Nearly 40-50% of the patients with epilepsy (PWE) fail to respond to their first AED monotherapy 2,3 with 30% cases showing refractoriness 4. This non-responsiveness to AEDs encouraged researchers to identify the predictors of poor response in PWE. Clinical factors such as early age at onset, higher pretreatment seizure frequency, cryptogenic epilepsy, brain neuroanatomic abnormality, etc. have previously been observed to be potential predictors of poor response to prescribed AEDs 3,5. Several studies also investigated the pharmacogenetics of AEDs and suggested the involvement of various drug-metabolizing enzymes (CYPs), drug transporters (ABC transporters), and drug target genes (ion-channels in brain) in deciding treatment outcome, however, remained inconclusive due to the inconsistency in their findings and failure of replication in different populations 6. One reason behind this failure could be the multifactorial nature of the disease which involves genome-environment interactions and therefore limits the independent use of genetics in studying AED response 6-8. Since any change in genetic or environmental factors