Drugs Raise Risk of Suicide

Adolescent suicide rates have dropped 25 percent since the early 1990s, even as more than a million children were put on the drugs, Fassler said. Although no one can say for sure what the connection is between those two trends, Fassler said, "if the medications were significantly increasing the risk of suicide, it is unlikely we would be seeing this kind of decline over time."

The new analysis seems likely to deepen transatlantic divisions over how the drugs are perceived and prescribed: British authorities last year recommended that depressed patients who were able function at home and work should not be medicated right away, but should be counseled to try exercise, self-help, talk therapy -- or watchful waiting.

David Healy, a psychiatrist at the University of Wales who helped conduct the new study, said managed-care insurance companies in the United States were not likely to look kindly upon physicians who monitor patients without prescribing drugs.

Healy prescribes the drugs but has long raised red flags about them. He asked yesterday why scientists at the FDA and research universities had not previously conducted this analysis, given that the data have long been available.

"For whatever reason, an awful lot of people didn't want to think there was a risk," he said.

FDA officials have said they are conducting an independent analysis.

The new study was conducted by epidemiologist Dean Fergusson and his colleagues at the Ottawa Health Research Institute and included scientists from McGill University. It is published in the current issue of the British Medical Journal along with two related articles and a commentary by Geddes and University of Verona psychiatrists Andrea Cipriani and Corrado Barbui.

Scientists have long bemoaned the lack of high-quality studies on antidepressants and the risk of suicide. One of the new studies, by University of Bristol and University of London researchers, examined drug trials submitted by pharmaceutical companies to British regulators. It found some evidence of self-harm among patients taking antidepressants but was inconclusive as to whether the drugs increased the risk of completed suicides.

A third study found no difference in suicide risk between SSRI drugs and an earlier class of medications known as tricyclic antidepressants.

Complicating the picture is the fact that suicide is rare -- meaning that it takes very large studies to yield definitive results. Fergusson's meta-analysis pooled data from a large number of studies in the same manner as the recent FDA analysis involving children's trials.

The American Psychiatric Association's Fassler said such an approach could miscount the number of attempted suicides. He said depressed patients who were getting better as a result of medication might be more likely to tell doctors about a suicide attempt, thereby muddying the data.

While Fergusson agreed that better trials are needed, he said he was confident about the results of this analysis, because it relied on placebo-controlled trials -- considered the most definitive. And not all the patients getting drugs were being treated for depression, he said -- there was evidence of increased risk of suicide attempts for patients with all kinds of disorders.

Suicide "is uncommon but serious," Fergusson said in an interview. "The biggest concern is these drugs are widely prescribed. There are millions of people on the drugs, so even a risk of one in a 1,000 when you amplify it to the millions, it becomes a public health issue."