In order to determine whether paroxetine was associated with any increase in suicidal thoughts or acts all controlled studies of paroxetine were examined in a series of metanalyses. Paroxetine showed an advantage in reducing suicidal thoughts in all analyses compared with placebo. On the MADRS there was a significant advantage compared with active controls at weeks 1, 3, 4 and 6 (P < 0.01). There were significantly fewer emergent suicidal thoughts on paroxetine compared with placebo in all analyses, and a significant advantage for paroxetine compared with active controls on the MADRS. A significant advantage for active controls compared with placebo was seen only on the HAMD. In the analysis of the data from controlled studies and open extension studies of paroxetine calculated by patient year of exposure there were 2.8 times fewer suicides in the paroxetine-treated group compared with active control and 5.6 times fewer compared with placebo.