Tsai 2011.

Study characteristics
Methods	Study design: parallel design Number of arms: 2 Experimental arm: milnacipran (SNRI) Control arm: matching placebo
Participants	Geographical location: Taiwan Setting: inpatient Stroke criteria: first ischaemic or recurrent stroke Method of stroke diagnosis: confirmed by imaging Time since stroke: 4 weeks before admission Inclusion criteria: 1) first or recurrent ischaemic stroke Exclusion criteria: 1) with past history of depression, substance abuse or psychosis; 2) taking antidepressants at least two weeks before stroke; 3) MMSE <15; 4) impairment of communication; 5) TIA; 6) had possible concurrent depression (HDRS > 10). Depression criteria: HDRS > 10 Total number randomised in this trial: 92 Number randomised to treatment group: 46 (65.2% men, mean age 61.0, SD 10.8) Number randomised to control group: 46 (63% men, mean age 64.9, SD 10.5) Total number included in final analysis: 56 Number included in treatment group for final analysis: 25 Number included in control group for final analysis: 31
Interventions	Treatment: milnacipran (SNRI), 50 mg/day titrated to 100 mg/day Control: matching placebo Treatment duration: 12 months Follow‐up: not reported
Outcomes	Primary outcomes Depression measured using the HDRS Secondary outcomes Quality of life measured using the QoL Index Disability measured using the London Handicap Scale
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The randomization was carried out by our pharmacy department, which has a professional team in charge of the clinical drug trials in our hospital." pp. 264 Comments: method of sequence generation not reported
Allocation concealment (selection bias)	Low risk	Quote: "They prepared the drug with an allocation number for each participant based on their random assignment." pp. 264
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The placebo was tailor‐made, and the appearance and weight (starch inside) were the same as that of the active drug, milnacipran." pp. 265
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "All the interviewers were blinded to the patient’s medication." pp. 265
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "Fifty‐six (60.9%) of 92 patients completed 12 months of study treatment: 25 (54.3%) of 46 in the treatment group and 31 (67.4%) of 46 in the placebo group, respectively." pp. 265 Comments: Figure 1 and Table 1 presents ITT analysis. All 92 participants were included in the analysis
Selective reporting (reporting bias)	Low risk	Quote: "The modified version of the 17‐item HDRS without item 14 (sexual behavior), NIHSS, and BI, were assessed in each of the follow‐up visits. The definition of PSD in the study was based on the diagnostic criteria of major depressive episode in DSM‐IV by professional psychiatrists."
Other bias	Unclear risk	Comments: treatment group appears to contain higher proportion of participants who have higher than secondary school education 34.8% compared to 20.0%