Sixteen observational studies (n=34,804) were included.
In terms of study quality, the major methodological flaw was the lack of reporting of relevant information. Seven studies used a validated depression rating scale. Most of the studies reported drop-outs and withdrawals. Ten studies assessed prognostic factors but none adjusted for confounders. Most of the studies did not describe the methods used to record adverse events and side-effects.
The studies showed physician-rated response rates ranging from 65 to 100% in the short term and 60 to 69% (based on two studies) in the long term.
The studies showed patient-rated response rates ranging from 63 to 98% in the short term. In the one study assessing long-term outcomes, 85% rated efficacy as good or excellent.
Drop-out rates ranged from 1.5 to 17.1% in the short term; the rate was 19.8% in the one study assessing long-term outcomes.
Drop-outs due to side-effects were rare, ranging from 0 to 2.8% in short-term studies and from 3.4 to 5.7% in the two long-term studies. The proportion of patients reporting side-effects was also low, ranging from 0 to 5.9%.
The most common side-effects were gastrointestinal symptoms, increased sensitivity to light and skin symptoms, although several studies reported a range of nervous symptoms. None of the studies reported any serious side-effects requiring hospitalisation.
The results from studies examining prognostic factors were also reported in the review.