The British Journal of Psychiatry
Discontinuation rates of SSRIs and tricyclic antidepressants: a meta-analysis and investigation of heterogeneity.
M Hotopf, R Hardy, G Lewis


BACKGROUND Previous meta-analyses suggest that individuals treated with serotonin-specific reuptake inhibitors (SSRIs) in randomised controlled trials (RCTs) are less likely to discontinue treatment than those on tricyclic antidepressants. This metaanalysis investigates whether this is due to the frequent use in RCTs of older reference tricyclics (imipramine and amitriptyline), which may have worse side-effects than more recent compounds.

METHODS A meta-analysis of RCTs comparing tricyclic and heterocyclic antidepressants with SSRIs in the treatment of depression.

RESULTS The overall odds ratio of discontinuation on tricyclic/heterocyclic antidepressants compared with SSRIs was 0.86 (95% CI 0.78-0.94). The odds ratio for reference tricyclics was 0.82 95% CI 0.72-0.23), newer tricyclics 0.89 (95% CI 0.74-1.06), and heterocyclics 1.02 (95% CI 0.78-1.35). The pooled advantage of SSRIs over tricyclics was maintained whether the population studied consisted of younger adults or only the elderly. No differences in discontinuation rates were detected between the SSRIs.

CONCLUSIONS The lower rate of discontinuation in patients on SSRIs may be due to the use of old tricyclics (which have worse side-effects) as reference compounds. The SSRIs do not show a statistically significant difference in discontinuation rates when compared with newer tricyclics or heterocyclics.