Rosenberg 1994.

Methods	Twenty‐two‐week, multicentre study. The primary treatment period was 6 weeks. However, patients who in the opinion of the investigator would benefit from further treatment could continue treatment under double‐blind conditions for a further 16 weeks, i.e. for a total of 22 weeks.
Participants	Depressed patients of either sex, who were assessed as being in need of antidepressant treatment and who had a total score of 14 or more on the Hamilton Depression Rating Scale (HDRS) Age range: 18‐65 years. Exclusion criteria: pregnancy, failure to use an acceptable contraceptive method, known alcohol or drug abuse within the past year, psychosis, serious somatic disease, treatment with MAO inhibitors within the last 2 weeks or with other antidepressants within the last week before inclusion and hypersensitivity to test drugs. Patients who required psychiatric in‐patient treatment were also excluded.
Interventions	Citalopram 10‐30 mg/day: 187 participants. Citalopram 20‐60 mg/day: 193 participants. Imipramine 50‐150 mg/day: 92 participants. Benzodiazepines or sedatives antihistamines could be prescribed for sleep disturbance, but other psychotropic drugs were not allowed.
Outcomes	Primary outcome: mean change in HDRS score from baseline to endpoint. Secondary outcomes: mean change in Clinical Global Impression (CGI) and Visual Analogue Scale (VAS) score, HDRS factors as depression, sleep disturbances, anxiety/somatization, retardation.
Notes	This study was funded by Lundbeck (citalopram manufacturer).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Patients were randomly allocated to one or two dose levels of citalopram or imipramine treatment. In each block of five patients one patient received imipramine and two pairs of patients each received one of the two citalopram dose". Randomization ratio 1:2:2.
Allocation concealment (selection bias)	Unclear risk	No data provided.
Blinding (performance bias and detection bias) All outcomes	Low risk	Quote: "This study was a double blind comparison (...) tablets of identical appearance were used".
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "All patients receiving at least one tablet constituted the Intention‐to‐treat (ITT) population. Patients who met the inclusion and exclusion criteria, had a compliance of 50% or higher and who completed at least 14 days of treatment constituted the Efficacy Population (EFF).
Selective reporting (reporting bias)	Unclear risk	No clear information reported.
Other bias	Unclear risk	Sponsorship bias cannot be ruled out.