Abstract Objective Obsessive–compulsive disorder (OCD) is a severe psychiatric disorder characterized by its heterogeneous nature and by different dimensions of obsessive–compulsive (OC) symptoms. Serotonin reuptake inhibitors (SRIs) are used to treat OCD, but up to 40% to 60% of patients do not show a significant improvement with these medications. In this study, we aimed to test the impact of brain-derived neurotrophic factor (BDNF) Val66Met polymorphism on the efficacy of antidepressants in OCD overall, and in relation to the different OC dimensions. Methods In a 6-month prospective treatment study, 69 Caucasian OCD patients were treated with escitalopram for 24 weeks or with escitalopram for 12 weeks followed by paroxetine for an additional 12-week period. Patients were genotyped and assessed for treatment response. The main clinical outcomes were improvement of the Yale-Brown Obsessive–Compulsive Scale score and in different OC symptom dimension scores. Results The Val/Val group comprised 43 (62%) patients, the Val/Met and Met/Met group comprised 26 (38%) patients. Forty-two patients were classified as responders at 12 weeks and 38 at 24 weeks; no significant association was found between BDNF Val66Met and SRIs response at 12 and 24 weeks. In analyses of the different OC symptom dimensions, the Met allele was associated with a slightly reduced score in the aggressive/checking dimension at 6 months ( P = .048). Conclusions Our findings do not support the usefulness of BDNF Val66Met genotyping to predict overall response to treatment with SRIs in OCD; they did however suggest a better outcome at 6 months for the aggressive/checking symptom dimension for patients carrying the Met allele.