OBJECTIVES: To evaluate the health care and productivity costs in patients with generalized anxiety disorder (GAD) before and after the initiation of a SSRI (selective serotonin re-uptake inhibitor), a SNRI (serotonin norepinephrine reuptake inhibitor), benzodiazepine or pregabalin. METHODS: Patients with an ICD10-code F40-F41 and at least two prescription claims for either SSRI, SNRI, benzodiazepin or pregabalin were identified using data from the Danish National Patient Registry, the Psychiatric Central Registry, the Medicinal Registry and other registries (2006-2010). Patients identified with neuropathic pain were excluded. The index date was considered the first prescription for SSRI, SNRI, benzodiazepines or pregabalin. Descriptive assessments of healthcare and productivity costs were conducted 12 months pre and post the index date using the full dataset, whereas a sub-analysis focussed on F41.1 (GAD). To control for selection bias, a propensity score matched cohort controlling for age, gender, Charlson score, depression, alcoholism, socioeconomic and cohabit status, and healthcare resource use was also conducted. Statistical tests performed were Wilcoxon (α=0.05). RESULTS: A total of 18,357 (F40-F41) patients met the inclusion criteria (treatment courses included: 14,095 SSRI; 5,035 SNRI; 8,580 benzodiazepines; 1,628 pregabalin). Twelve months healthcare costs were only significantly reduced in the pregabalin group (€1,285; P<0.001). The three other groups resulted in significantly increased healthcare costs (€808-€1,548; P<0.001). Similar results were found focusing only on F41.1; however insignificant for pregabalin. Matched sub-analyses covering 1,588 patients in each group showed similar significant reduced 12 months healthcare costs in the SNRI and pregabalin groups (P=0.001). Across all four groups the ability to retain employment was significantly improved, whereas long-term sickness increased; however, insignificantly so in the pregabalin group. CONCLUSIONS: Health care costs 12 months after the initiation of the treatment with pregabalin were significantly reduced. Besides pregabalin, this was in matched analyses also the case for the SNRI group. Production loss did not differ between groups.