Any compound which has the potential to disrupt cognitive and psychomotor performance will exacerbate the impairments which are a feature of depressive illness. The consequences of this include compromising the patient in terms of recovery from depression, reducing the quality of life and increasing the risk of accident. The aim of the present randomised, parallel group, double blind study was to compare fluoxetine and dothiepin with respect to their effects of cognitive function and subjective ratings of sleep in depressed patients being treated in general practice.84 patients (mean age 43.8 years old) received either fluoxetine 20 mg mane and placebo nocte; or dothiepin 75 mg nocte (rising to 150 mg after the first week) and placebo mane for 6 weeks. Prior to treatment, then on days 7, 14, 28 and 42, patients were assessed on aspects of cognitive function, memory, sleep and daytime sedation. Patients were also scored according to the Hamilton Depression Rating Scale (HAM-D) before and after treatment.A significant reduction in the HAM-D scores from baseline to day 42 (21.1 and 10.2 respectively) was observed in those patients (n = 54) who completed the study (Wilcoxon Matched Pairs Test; n = 54, z = 6.36, p < 0.01). No differences were found between the two drugs. Two sets of analyses were performed on the test data: 1-intention to treat (i.e. all patients entered onto the study) and 2- completers who responded to medication. The alleviation of depression with both drugs was accompanied by improvement over time in many of the tests, however performance scores were generally superior with fluoxetine than with dothiepin in both sets of analyses. In the responders, critical flicker fusion scores were higher with fluoxetine than dothiepin (F (1.29)-5.01; p < 0.05) at all test points; this difference being evident using standard and alternative methodology (i.e. with and without artificial pupils). Fluoxetine also resulted in significantly better scores on a subset of the mental arithmetic test (p < 0.05), and non-significant trends showed that the fluoxetine group scored higher than dothiepin on Kim's Game and the Cognitive Failures Questionnaire. No significant differences were evident in subjective ratings of daytime sedation, as assessed by the Milford Epworth Sleepiness Scale. Subjective ratings of sleep were recorded on the Leeds Sleep Evaluation Questionnaire (LSEQ). No differences were observed in the getting to sleep component of the LSEQ, neither across time nor between treatments. Quality of sleep improved throughout the study with both drugs (F (3.84) = 6.14; p < 0.001) but no significant difference was detected between drugs. Dothiepin impaired ratings of feelings on awakening from sleep compared to fluoxetine, and this difference reached significance on day 7 (p < 0.05).The results from this study demonstrate that, although fluoxetine and dothiepin were similarly efficacious in relieving depression as assessed by clinical ratings (i.e. HAM-D), those patients in the fluoxetine group performed significant better on objective cognitive tests and felt better on awakening first thing in the morning. These factors should be taken into consideration in the management of depression in ambulant outpatients.