Eight RCTs were included in the review (n=560, range 30 to 162). All studies reported appropriate use of blinding and comparable groups at baseline. Six studies used an intention-to-treat analysis. For all studies it was unclear whether randomisation and allocation concealment were appropriate. Study duration ranged from six to 20 weeks.
Patients who used bupropion were 2.4 times more likely to have improved clinical outcomes compared with patients who used placebo (OR 2.42, 95% CI 1.09 to 5.36; five studies).
No statistical results were reported for other compounds; with the exception of desipramine (68% response compared with placebo; one study) they appeared to be inconclusive.