SSRI-suicide link expands to include all
tricyclic antidepressants

Dodgy data in previous studies leaves docs dazed and confused

By Owen Dyer

The reviews include one by Canadian scientists led by Dr Dean Fergusson of the Ottawa Health Research Institute. The Canadian team reviewed 702 studies on SSRIs that included 87,650 patients. Another was the work of a group from UK's Bristol University, headed by

Dr Carlos Martinez. These researchers conducted a meta-analysis of 477 pharmaceutical trials involving over 40,000 patients. The final, and largest study, was conducted by the same Bristol team, and reviewed the experiences of 146,095 first-time antidepressant users whose histories were recorded on the UK's General Practice Research Database.

The three mammoth studies all conclude that the link between SSRI use and suicide is real. However, the final study found that incidents of suicide, attempted suicide and self-harm were just as common among patients taking the older tricyclic antidepressants as they were among patients taking SSRIs. This provides the best insight yet into the properties of tricyclics, which have never been blamed for suicidal behaviour. The only distinction between SSRIs and tricyclics was seen among the under-19s. In this age group, there was evidence of extra risk from SSRIs. Similarly, there seemed to be no marked differences between different SSRIs except in the under-19 age group, who appeared to be more likely to resort to self-harm when taking paroxetine.

With the FDA issuing black box warnings and Britain banning SSRIs altogether in children and adolescents, it seems that the wheel has come full circle in the 10 years since Professor David Healy, now at McGill University, was widely dismissed as a crackpot for suggesting a link between SSRIs and suicidal behaviour. For the beleaguered family practitioner, however, the picture is as confusing as ever. Depressed patients still need treatment, and depression is itself a risk factor for suicide. But a return to tricyclics, it seems, cannot provide the answer.

The fact that none of the three BMJ papers end with a clear call to limit SSRI use is indicative of the confusion that surrounds the data. There's a consensus that children are more at risk than adults. Yet all 69 of the patients in the General Practice Research Database who eventually committed suicide were adults.

FAULTY DATA
But there's something not quite right about the data from the original drug company trials, according to the Bristol researchers. For one thing, the ratio of reported suicides to episodes of self-harm is about one to ten, while it's about one to thirty in the general population. This suggests that self-harm was underreported in these trials. Moreover, actual self-harm was reported as frequently as suicidal ideation, when normally it's only about one-fifth as common, suggesting that suicidal thinking was also underreported.

"A major concern in observational studies," the Bristol researchers conceded in their paper, "is when the outcome under study is itself an indication for which the drug might be prescribed." They added: "we did not deal with the question of whether people treated with SSRIs are at greater risk of self harm than those with equivalent morbidity who do not receive treatment."

Proving that antidepressants can cause suicides is one thing. It's quite another thing to prove that they cause more suicides than leaving depression untreated. "The debate is not yet done," said the BMJ's acting editor Kamran Abbasi, "but these papers crystallise arguments that have been drifting in the ether these past months. How many people who turned to 'happy pills' would not have done so if they had been fully aware of the potential harms?"

BMJ Feb 19, 2005;330(7488):396