GSK's Troy Hobbled His Future Employer From Reporting Tykerb Risks

It comes as no surprise that Daniel Troy, the current general counsel to GlaxoSmithKline, is at the center of the preemption mess unveiled by Rep. Henry Waxman today. It is surprising, however, that Troy's actions led to a three-month delay in getting a label change requested by GSK for Tykerb, when the company discovered new risks on that breast cancer drug, according to the documents obtained by Waxman.

"Preemption" is the notion that state consumer protection laws should be trumped by the FDA's jurisdiction over drugs, protecting drug companies from liability if their products injure patients because they were approved by the FDA. The issue is about to go before the Supreme Court.

In 2007, GSK was trying to do the right thing: alert authorities and the public that it had discovered a risk for irreversible lung disease in some Tykerb users. Previously, the FDA had taken the stance that it was OK for drug companies to tighten warnings on their own labels. But Troy, as FDA's chief counsel, had instituted a rule-change, which said that companies had to get FDA's permission to make their drug labels stricter. The policy change led to the delay.

In September of this year, Troy -- the Darth Vader of the drug business -- joined GSK as its general counsel.

Johnson & Johnson was also frustrated by the Troy-sponsored policy change. In 2007, J&J went to the FDA to inform it that it believed Levaquin, a J&J antibiotic, should have a greater liver-damage risk warning on its label than the FDA had required. The Troy policy kicked in and J&J was only allowed to add the warning three months later. Here's the new text:

Severe, and sometimes fatal, hepatotoxicity has been reported. Discontinue immediately if signs and symptoms of hepatitis occur.

"Excessive product safety" ... "excess investment in the avoidance of adverse drug events" ... "the FDCA, as administered by FDA, induces excessive product safety" ... "disclosing too much."

[M]uch of the argument for why we are proposing to invoke preemption seems to be based on a false assumption that the FDA approved labeling is fully accurate and up-to-date in a real time basis. We know that such an assumption is false. ... we know that many current approved drug labels are out of date and in many cases contain incorrect information (e.g., the overdose section) ...

I am not so comfortable with the whole argument in the document about preempting state liability cases against a manufacturer for "failure to warn" cases. I'm not sure why this falls within FDA's purview (are there any other examples of where FDA has promulgated regulations based primarily on a desire to protect sponsors from liability?) and I think the whole argument that liability concerns drive inaccurate labeling is false and misleading.

[T]here are continued references to sponsors "disclosing too much" risk information and its adverse impact. ... We rarely find ourselves in situations where sponsors want to disclose more risk information than we think is necessary. To the contrary, we usually find ourselves dealing with situations where sponsors want to minimize the risk information. ... the statement that 'FDA believes manufacturers should add risk information to labeling only after consulting with the agency ... is not true ... [W]e do not discourage sponsors from adding new information ...

Companies rarely press for meaningful risk information or additional warnings. And they always oppose black box warnings. Much of the discussion of what goes in the label centers around the sponsors [sic] wish to promote the drug fully and to not be handicapped by risk information that would have to be conveyed in ads. Sponsors do seek to increase the reactions in the Adverse Reactions section because this influences their reporting requirements (if it's in the label, it is an expected event and reporting requirements are less).