July 15, 2008

The FDA Says No Black Box Needed On Drugs That Increase Suicidality, But Still Needed For Those That Don't

Yes, you read that right. Drink a big glass of OJ and put away your blotter paper, you won't need 'em in here.

In a metanalysis of 199 studies covering 11 antiepileptics, the odds ration for suicidality was 1.80.

The first question that the committee had to decide was whether or not it agreed with the agency's overall finding of an increase in suicidality for the 11 AEDs that were analyzed.

The advisory committee, after much debate, determined unanimously, apart from 1 abstention, that yes, there is a signal that is statistically reliable.

Forget about whether these results are accurate or useful. Just follow their logic: they believe their data.

The second question asked whether...the findings of increased suicidality should apply to all drugs included in the analyses, despite the observation that the estimate for the odds ratio for 3 of the drugs was below 1.

Remember science? Remember classism? Remember not blaming everyone in a group for the behavior of a few? If the odds ratio is below one, then it doesn't have the risk.

The committee strongly agreed that the findings should apply to all drugs.

Awesome. Faith in egalitarianism and the supremacy of science restored.

Well, all right, D- for interpretation of science. They want to believe it causes suicide, so be it.

After this discussion, the committee members voted strongly against a black-box warning.

But I thought you guys said--?

One committee member pointed out the potential "hysteric reaction" that can accompany a black-box warning about suicidality. Other speakers emphasized the need to balance efficacy vs harm and commented that "in two-thirds of the trials this warning is irrelevant," "a black-box warning has a very negative connotation," and "the number needed to harm was 2 per 1000."

But then how do you possibly justify taking the box off drugs that show an increased risk, but leave the black box on drugs that have no such data? Why does Seroquel get the warning-- no actual risk of suicide ever found, despite looking-- but antiepileptics don't, and they do have the data?