The risk of suicide among people taking antidepressants declines with age, according to U.S. researchers.

In Wednesday's online issue of BMJ, researchers with the U.S. Food and Drug Administration who reviewed 72 trials involving nearly 100,000 people concluded the risk of suicide was "strongly age dependent."

An increased risk for suicide was found among people younger than 25 taking SSRIs (selective serotonin reuptake inhibitors) from eight drug makers. There was no difference for those aged 25 to 64, and reduced risk in people 65 and older.

Since 2003, regulators in the U.S., Canada and Europe have warned about an increased risk of suicidal thoughts and behaviours among people under 18 who use antidepressant drugs.

In May 2007, the FDA extended the warning to people aged 18 to 24.

Study author Dr. Marc Stone and his colleagues said the review supports the thinking that antidepressants can have two distinct effects.

In some patients, the drugs promote suicidal thoughts and behaviour, a risk that seems to decrease with age.

In others, antidepressants provide relief from depression and help reduce the risk of suicide.

Risk varies with different medications

"Increased risk is probably restricted to younger people and varies greatly between individual medicines," John Geddes, a professor of epidemiological psychiatry at the University of Oxford in England, and his colleagues wrote in a journal commentary accompanying the study.

The commentary noted differences in the odds of suicidal thoughts and behaviours varied widely among different SSRI drugs. For example, studies cited in the report suggested sertraline (Zoloft) carried half the risk of suicidal thoughts and behaviour as some other SSRIs, Geddes said.

He also criticized the review for relying heavily on data from trials sponsored by drug companies, which compare drugs against a placebo rather than other drugs in the same class.

An essay published in Thursday's issue of the New England Journal of Medicine made the same argument, saying drug and device manufacturers should state how new medications compare with similar treatments on their labels, instead of better than a placebo.

For example, Dr. Randall Stafford, a professor of medicine at the Stanford Prevention Research Center in California, and his colleagues recommended that the FDA require new treatments to carry a label reading: "Although this drug has been shown to lower blood pressure more effectively than placebo, it has not been shown to be more effective than other members of the same drug class."

Changing drug labels could help counter the tendency of doctors and patients to want what is newest and more expensive, even though it may not be better, Stafford said.

The editorial authors reported receiving grant support and consulting fees from pharmaceutical companies.