Serotonin reuptake inhibitors (SRIs) do not have to be administered continuously to be effective for premenstrual dysphoric disorder (PMDD), but can be given during luteal phases only. This is of practical importance, but also of theoretical interest since it suggests that the onset of action of SRIs is shorter in PMDD than in, for example depression. In this study, both continuous and intermittent SRI administration was compared with placebo, with the special purpose of analyzing if different PMDD symptoms respond differently depending on the treatment regimen. To this end, women meeting slightly modified DSM-IV criteria for PMDD (mean+/-SD age, 37+/-6.3 years) were treated for three menstrual cycles with paroxetine continuously, paroxetine during the luteal phase only, or placebo, the population completing at least one treatment cycle comprising 55-56 subjects per group. Continuous treatment with paroxetine reduced premenstrual symptoms effectively with a response rate of 85%. The effect size was highest for irritability (1.4) and lowest for lack of energy (0.5). Intermittent treatment was as effective as continuous treatment in reducing irritability, affect lability, and mood swings, but had a somewhat weaker effect on depressed mood and somatic symptoms. The study indicates that the response rate when treating PMDD with SRIs is high, and that irritability is a key target symptom. Symptoms such as irritability, affect lability, and mood swings appear to be more inclined to respond rapidly to SRIs, enabling intermittent treatment, than are, for example, the somatic symptoms.