Antidepressants tied to lower suicide rates in kids

Last Updated: 2006-11-02 16:11:01 -0400 (Reuters Health)

By Karla Gale

NEW YORK (Reuters Health) - In US counties with the highest rates of prescriptions for selective serotonin reuptake inhibitors (SSRI), a type of antidepressant, the rate of suicide among children ages 5 to 14 is lower than that observed in counties with lower rates, according to a new report.

"Three large population studies in adults and adolescents found an inverse relationship between SSRI use and increased risk of suicide," lead author Dr. Robert D. Gibbons told Reuters Health. "But there have been no such studies of younger children."

In 2004, the US Food and Drug Administration (FDA) advisory panel began requiring a black box warning on SSRI prescriptions, based on reports that fluoxetine (Prozac) increased suicidal thoughts in children and adolescents.

"Now we are evaluating whether suicide rates have increased since the black box warning was required," Gibbons told Reuters Health, "and to see if decreased use of SSRIs is associated with actual suicide."

"If we see that decreased prescribing of SSRIs brought about by the black box warning is leading to an increased completed suicide rate in kids, we want to get that message out as quickly as possible."

Gibbons, a biostatistician at the University of Illinois at Chicago, and associates analyzed data from the Centers for Disease Control and Prevention regarding suicides among children 5 to 14 years of age. The data covered the period 1996 through 1998, and was grouped by county. They then linked this data with SSRI prescriptions from a random sample of 20,000 pharmacies.

Among nearly 39 million children ages 5 to 14, there were 933 suicides between 1996 and 1998 (0.8 per 100,000 children per year).

Since the implementation of the black box warning, the rate of antidepressant prescriptions in children fell by about 30 percent, the authors note. Subsequently, the overall observed suicide rate was as high as 1.7 per 100,000 for counties with the lowest decile of SSRI prescriptions. At the highest decile, the rate was 0.7 per 100,000.

The association between SSRI prescriptions and completed suicides remained significant after adjusting for income and access to mental health care.

"Our findings counter results of randomized clinical trials, which may be due to their use of surrogates of suicide -- suicide attempts and suicidal ideation," Gibbons said, or to ascertainment bias.

Gibbons was one of the dissenting members of the 2004 FDA advisory panel that voted 15 to 8 in favor of a black box warning on SSRI prescriptions.

"The evidence provided by the FDA was a meta-analysis of about 25 randomized controlled trials," Gibbons explained, "but the data were a retrospective look back at charts of children, which again, may increase the likelihood of ascertainment bias."

In the investigators' paper, they report that, according to the FDA's analysis, the absolute rates of increases in suicidal thinking associated with SSRIs were only 0 percent to 2 percent higher than placebo, "which is only marginally significant," Gibbons said.

He noted that the only prospective study that evaluated suicide index -- and thus was not subject to ascertainment bias -- used the Hamilton Depression Rating Scale. "For the question about suicidal ideation and antidepressant use, the investigators found no effect whatsoever," Gibbons said.

"Which made me question the validity of (the FDA's) analyses," he added.

He noted that his group's findings have been replicated in data from the Netherlands, "and when we looked at insurance claims databases, we found no association between prescriptions of SSRI and suicide attempts in children."

Gibbons was also a member of the Institute of Medicine Committee on US Drug Safety. During his tenure there, he observed that MedWatch, the FDA's surveillance system to monitor reports of adverse drug reactions, "suicide was never once mentioned (during the FDA's hearing); or if it was analyzed, it was never reported by the FDA at advisory board hearings.

"That has prompted us to do a very extensive analysis of spontaneous reporting systems for all antidepressants."

SOURCE: American Journal of Psychiatry November 2006.