Pharmacologic enhancement of serotonergic transmission by serotonin uptake inhibition has been suggested as one approach to improve upper-airway patency and thus nocturnal breathing in patients with obstructive sleep apnea (OSA). To test this hypothesis, we performed a double-blind, randomized, placebo-controlled crossover study testing the effect of paroxetine (20 mg od) on polysomnographic and psychopathologic outcomes in 20 male OSA patients (mean age 52.1 years, mean BMI 28.7 kg/m2, mean oxygen desaturation index on a previous screening 25.4/hour). The two treatment periods of 6 weeks and the separating washout period of 4 weeks were completed by 17 patients. Paroxetine reduced the apnea index during NREM sleep (-35%, p = 0.003), but not during REM sleep. No significant effect on hypopnea indices was found. With the exception of a previously described REM-postponing effect (p = 0.05), sleep architecture was not significantly influenced by paroxetine. Similarly, the effect of paroxetine on apnea was not associated with a significant overall alleviation of psychopathologic symptoms as rated on the Comprehensive Psychopathological Rating Scale or OSA-related daytime complaints assessed by visual analog scales. We conclude that enhanced serotonergic transmission improves breathing during NREM sleep in OSA. This effect is poorly related to effects on sleep architecture or daytime symptoms.