Currently, there are no data available comparing cost-effectiveness of two antidepressants in the primary care setting in the UK. Alongside a randomized, double-blind, 24-week study of mirtazapine and paroxetine, data were prospectively collected on patients' use of hospital and non-hospital services and days off work. Costs were estimated in each treatment arm from National Health Service (NHS) and societal perspectives, and were compared with selected outcome measures (numbers of 17-item Hamilton Rating Scale for Depression (17-HAMD) responders and changes in Quality of Life in Depression Scale scores between baseline and 24-week endpoint) to explore and compare relative cost-effectiveness. Mirtazapine treatment resulted in a statistically significantly greater improvement in quality of life than paroxetine at endpoint (P=0.021). Although the 17-HAMD response rates were higher for the mirtazapine users at endpoint, the difference (7%) was not statistically significant (P=0.31). However, mean total societal costs per patient were 375 pounds less with mirtazapine (1850 pounds) compared to paroxetine (2225 pounds; P=0.32). Mean total NHS costs per patient were also lower (120 pounds) with mirtazapine (1408 pounds) compared to paroxetine (1528 pounds). The advantage for mirtazapine remained present on all variables analysed after performing sensitivity analyses. The results suggest that mirtazapine may be a cost-effective treatment choice compared to paroxetine for depression in a primary care setting.