Background Antidepressant prescribing is widespread. Nonetheless, response to antidepressants is variable. If it was possible to predict response to medication and thus tailor treatment accordingly, this would not only improve patient outcomes but may also have economic benefits. Aims To test the hypothesis that individuals with more severe depression would benefit more from noradrenaline reuptake inhibitors (NARIs) than selective serotonin reuptake inhibitors (SSRIs) compared with individuals with less severe depression. Method Individuals recruited from UK primary care who met ICD-10 criteria for a depressive episode and scored 15 or more on the Beck Depression Inventory (BDI) were randomised to either an SSRI (citalopram 20 mg daily) or a NARI (reboxetine 4 mg twice daily). Randomisation was by means of a remote automated telephone system. The main outcome was depressive symptoms measured by the BDI total score 6 weeks after randomisation. (Trial registration: ISRCTN31345163.) Results In total, 601 participants were randomised (citalopram: n = 298, reboxetine: n = 303). Ninety-one per cent were followed up at 6 weeks (citalopram: n = 274, reboxetine: n = 272). There was little evidence to support an interaction between treatment and severity of depression (interaction term: 0.02, 95% CI —0.59 to 0.62, P = 0.96). Adjustment for potential confounders (age, gender, employment status, history of depression, number of life events and social support) did not affect the findings (interaction term: 0.06, 95% CI —0.54 to 0.66, P = 0.85). Conclusions Treatment with NARIs does not confer any advantage over SSRI treatment for outcome in those with more severe depressive illness presenting in primary care. Royal College of Psychiatrists