Study designs of evaluations included in the review
Double-blind randomised controlled trials (RCTs) and meta-analysis or reviews of RCTs. Most analysis was performed on an intention-to-treat (ITT) basis and included all randomised patients who had >/=1 post-baseline efficacy assessment and had received >/=1 dose of study medication. A number of studies stated that end-point analyses were performed using the last observation carried forward method.
Specific interventions included in the review
Mirtazapine (in the majority of trials dosage regimes were titrated according the patients' responses) given at an initial dosage of 5-20mg/day and maximum dosages of 35 to 80mg/day in all studies apart form one in which patients received higher dosages (40 to 100mg/day). Comparator drugs included tricyclic antidepressants (amitriptyline 30-280mg/day, clomipramine 50-200mg/day, doxepin 75-300mg/day, imipramine 38-450mg/day); selective seratonin reuptake inhibitors (fluoxetine 20-40mg/day, paroxetine 20-40mg/day); atypical antidepressant trazodone 40-450mg/day; or placebo.
Participants included in the review
Hospitalised patients and outpatients meeting American Psychiatric Association DSM-III, DSM-III-R or DSM-IV criteria for major depression were included. Patients had moderate or severe (17-item Hamilton Depression Rating Scale (HDRS) score 18 to 24 or >/=25, respectively) depression at baseline. Mean baseline HDRS scores were >/= 25 in a number of studies.
Outcomes assessed in the review
Outcomes were not defined a-priori. The main clinical assessment tool used in included studies was the HDRS. In general, patients were considered to be responders if they had a >/= 50% reduction in HDRS score from baseline. Remission and relapse rates (the proportion of patients with 17-item HDRS scores </=7 and >/= 16, respectively) at study end-point were also reported in a few studies. A number of other scales were also used to assess antidepressant efficacy. Information about any experienced adverse effects was also collected.
How were decisions on the relevance of primary studies made?
Inclusion of studies was based mainly on the methods section of the trials. When available, large, well controlled trials with appropriate statistical methodology were preferred. No further information was presented on how decisions on the relevance of primary studies were conducted.