Seven RCTs were included in the review (336 participants, range 24 to 131). Trial quality ranged from good to excellent; Jadad scores were 5 (two RCTs), 4 (four RCTs) and 3 (one RCT). All RCTs described diagnostic criteria and used an objective depression scale and the Mini-Mental State Examination. Five used intention-to-treat analysis and three defined the primary outcome. None described their method of randomisation.
One of three RCTs that defined a primary outcome found significant evidence of efficacy. One of the four other RCTs found a significant benefit from antidepressants. When studies were pooled, the intervention had no significant effect on rates of response (six RCTs), remission (five RCTs), change scores (six RCTs) or rates of treatment discontinuation (either overall or due to adverse events; both seven RCTs). Heterogeneity was significant for response rate (Ι²=56%), moderate for remission rate (Ι²=49%) and low or absent for other outcomes.
No evidence of publication bias was found. The results of sensitivity analyses were also reported.