A new Food and Drug Administration (search) analysis backs earlier suggestions of a link between at least some antidepressants used by children and suicidal tendencies, prompting the agency to prepare new warnings.

Exactly what those warnings will say, and which drugs will be affected, hasn't been settled, according to FDA documents released Friday. The agency will take the debate to its scientific advisers next month for help in deciding.

"While there remains a signal of risk ... for some drugs in some trials, it is important to note that the data are not black-and-white in providing a clear and definitive answer," FDA psychiatric drugs chief Dr. Thomas Laughren wrote the advisory panel this week.

The controversy has simmered since last year, when British health authorities declared that a list of popular antidepressants might sometimes increase the risk of suicidal behavior in children and teenagers, and declared all but one -- Prozac (search) -- unsuitable for depressed youth.

It's a difficult issue to sort out because depression itself can lead to suicide, and studies show that antidepressants (search) have helped adult depressed patients recover. But, Prozac is the only drug FDA-approved to treat pediatric depression and there is little scientific evidence that any other antidepressant works in that age group, thus deepening concern about even potential risks.

In March, the FDA did urge doctors to monitor closely patients of all ages for warning signs of suicide, especially when they first start the pills or change a dose. The reason: They may cause agitation, anxiety and hostility in a subset of patients unusually prone to rare side effects. Or, in the time period when severe depression just starts to lift but isn't yet much better, patients who had been too lethargic to act on suicidal thoughts might gain the energy to do so.

Now the question is whether the drugs' labels need still more warnings - or, most costly to industry, specific instructions not to use certain ones in children and teenagers.

Among 25 studies involving 4,000 children and teens and eight antidepressants, there were no completed suicides.

FDA analyzed those studies in February, sent them for a second opinion from Columbia University researchers, and then reanalyzed them together with the Columbia data.

The result: Lump all the studies together, and patients given any of the antidepressants were about 1.8 times more likely to have suicidal thoughts or behavior than patients given dummy pills, FDA reviewer Dr. Andrew Mosholder wrote his bosses on Monday.

However, most suicidal tendencies showed up in studies with the highest proportion of patients who had a history of a suicide attempt or suicidal behavior before they were enrolled, noted a separate review by FDA's Dr. Tarek Hammad. Other studies didn't even admit those high-risk patients.

Further complicating the assessment, the risk varied widely not just from drug to drug but between studies of the same drugs, Hammad found. In general, patients in Prozac studies showed far fewer problems those on other drugs.

Still, "FDA will assure that the labels of the antidepressants used in pediatric patients reflect the most recent information obtained from these studies and analyses," said an agency statement issued Friday.