Ansseau 1984.

Methods	Randomised controlled trial, 6 weeks
Participants	126 outpatients with primary affective non‐psychotic depression of at least moderate severity, required to have a Raskin Depression Scale (RDS) score of at least 8, at least 5 items on Feighner Depression Checklist (FDC), a Covi Anxiety Score (CAS) equal to or less than RDS, a HDRS score of at least 18 on the 21‐item HDRS, and an anxiolytic antidepressant was warranted 145 participants were enrolled of whom 19 were not available; 126 outpatients were therefore available Aged 18 to 70 years
Interventions	Alprazolam versus doxepin Placebo washout for 4 to 7 days Maximum dose for alprazolam 4.5 mg and for doxepin 225 mg Mean final doses 2.7 mg for alprazolam and 137.5 mg for doxepin
Outcomes	Primary outcome: Mean HDRS after 6 weeks: alprazolam 11 (initial 24.9), doxepin 11 (initial 25.1) Secondary outcomes: All‐cause withdrawals: 23% for alprazolam, 12% for doxepin Side effects: drowsiness (alprazolam 22%, doxepin 28%), dry mouth (alprazolam 3%, doxepin 36%), constipation (alprazolam 3%, doxepin 28%), lightheadedness (alprazolam 15%, doxepin 18%) 32 alprazolam patients (of 59) reported 142 side effects (mean of 4.4 per reporting patient), versus 45 (of 67) doxepin patients who reported 357 side effects (mean of 8.5 per patient)
Notes	Results reported in table 1 (change scores) and figure 1 do not match
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“Patients were randomly assigned to treatment.”  “... no statistical differences between the two treatment groups except for sex.” No further information provided
Allocation concealment (selection bias)	Unclear risk	No further information provided
Blinding (performance bias and detection bias) Patients	Unclear risk	“Double‐blind study” “Initial dose of 2 tablets...” They probably used identical tablets, but this is not described in the study
Blinding (performance bias and detection bias) Physicians	Unclear risk	They used a standard scheme for dose increase
Blinding (performance bias and detection bias) Investigators	Unclear risk	Not described how they managed to keep the medication separated for each patient so that the investigators were blinded for the therapy as well. No information is provided about whether they used an investigator to do the assessments
Incomplete outcome data (attrition bias) HDRS	High risk	Initially, “13 patients in the alprazolam group and 6 patients in the doxepin group were not available”. Thus, 19/145 (13%) of the enrolled patients enrolled were not available. They further describe that 7 patients in the alprazolam group and 4 patients in the doxepin group were not available at final assessment; but it is not clear why these are left out of the evaluation ”Concomitant medical events, which may or may not have been adverse reactions to treatment, prompted the investigators to discontinue treatment for 11 patients.”
Incomplete outcome data (attrition bias) Withdrawals/side effects	Low risk	Unclear why numbers do not seem to match
Selective reporting (reporting bias)	Unclear risk	No information provided
Other bias	Low risk	Dosing: Both drugs were given in the therapeutic range
Placebo washout	High risk	Yes