Data About Zyprexa's Adverse Effects Still Lost in the (Legal) Fog

We have posted (most recently here , here, here, and here) about allegations in a New York Times series that Eli Lilly and Company had used questionable marketing tactics to promote its atypical anti-psychotic drug Zyprexa (olanzapine), and especially that the company had suppressed information suggesting that the drug had more adverse effects than previously reported, and then how the company tried to restrict publication of the internal memos on which these allegations were based.

Last week, Alex Berenson in the NY Times reported again concerning the continuing controversy about whether Eli Lilly & Co suppressed data about the adverse effects of Zyprexa. Now the US Food and Drug Administration (FDA) seems also to have doubts about whether Lilly provided it was accurate data.

Per the Berenson article,

The Food and Drug Administration is examining whether Eli Lilly & Company provided it with accurate data about the side effects of the antipsychotic drug Zyprexa, a potent medicine that has been linked to weight gain and diabetes.

The F.D.A. has questions about a Lilly document from February 2000 in which the company found that patients taking Zyprexa in clinical trials were three and a half times as likely to develop high blood sugar as those who did not take the drug.

That document was not submitted to the agency. But a few months later, Lilly provided data to the F.D.A. that showed almost no difference in blood sugar between patients who took Zyprexa and those who did not.

The F.D.A. confirmed its inquiry in response to questions from The New York Times. The agency said it had not yet decided whether to take any action against Lilly.

'The F.D.A. continues to explore the concerns raised recently regarding information provided to the F.D.A. on Zyprexa’s safety,' Dr. Mitchell Mathis, a deputy director in the psychiatry division of the agency’s center for drug evaluation and research, said.

The Zyprexa document that has aroused the most interest at the F.D.A. is a Feb. 21, 2000, paper in which Lilly scientists discussed whether Zyprexa’s label should be changed to alert doctors of the risk of hyperglycemia, or high blood sugar, associated with the drug.

The paper showed that 154 of 4,234 patients, or 3.6 percent, who took Zyprexa in clinical trials developed high blood sugar. Only 1.1 percent of patients who took a placebo developed the condition.

Doctors have said that difference is worrisome because most patients in the clinical trials were taking Zyprexa for only a few weeks or months.

The data that Lilly provided to the F.D.A. was notably different from the results discussed in the February 2000 paper, with the gap between the two patient groups much narrower. The company told the agency that patients taking Zyprexa developed high blood sugar at a 3.1 percent rate, while those taking the placebo had a 2.5 percent rate.

[Lilly spokesman] Mr. [Phil] Belt said that after the February 2000 paper, Lilly performed a final quality check of the data and discovered that some patients had been incorrectly included in the analysis, while others had been excluded.

'The original data referred to in the memo was a preliminary analysis, not the final accurate analysis that was provided to the F.D.A., and is therefore very misleading,' he said.

The rebuttal by Lilly mirrors that found in a press release of December 21, 2006. The Lilly press release does not provide further explain what precisely was wrong with the data that apparently showed an excess of hyperglycemia associated with Zyprexa, or how the errors were rectified. Nor did it explain why the company produced the internal 2000 document detailing data that was so quickly proven to be preliminary and error ridden. The latter document may be in the cache of more than 300 documents that is still publicly available (here), but I cannot readily determine which one it is.

Thus we are in an evidence limbo. The very limited evidence that has appeared in public is insufficient to judge whether there was good reason to think in the year 2000 that Zyprexa caused hyperglycemia and diabetes, and whether Eli Lilly withheld data that would have been supported this conclusion.

Apparently evidence sufficient to clear up this conclusion exists. But it remains in the hands of Eli Lilly, which so far has not seen fit to release it. Instead, apparently Lilly is continuing to persue a legal case against the people who first released its documents to the public, thus ensuring that the evidence remains obscure. As reported by Nature this month [Giles J. Court case to reclaim confidential data. Nature 2007; 446: 838-9],

A public-health expert in the United States is facing a potential jail term or a hefty fine after distributing documents that contain safety data on a blockbuster drug. The material came to light during initial negotiations in an ongoing lawsuit over the drug, and its manufacturer says that the information should have remained confidential.

The case, which could result in charges of contempt being levelled against the expert, shines light on the murky debates that go on between lawyers before drug liability cases get to court.

The latest incident centres around David Egilman from Brown University in Providence, Rhode Island, who is a frequent critic of the pharmaceutical industry....

Egilman had access to 15 million pages of confidential documents containing clinical-trial data and details of marketing plans that Lilly had released to lawyers during pre-trial negotiations, known as the discovery phase. But because the cases were settled before they reached court, the material was bound by confidentiality agreements.

[An] attorney, James Gottstein of Anchorage, Alaska, subpoenaed Egilman for the Zyprexa documents and then forwarded them on to the [NY Times] reporter [Alex Berenson].

Although Egilman was subpoenaed to release the documents, he and Gottstein could face jail terms or be required to contribute to Lilly's legal bills if the contempt proceedings proceed and are upheld.

Thus, the evidence that would allow physicians (and patients) to decide whether Eli Lilly suppressed data about adverse effects of Zyprexa remains tied up in legal proceedings. The company has not made any effort to make any of this evidence public. Of course, continuing to suppress evidence about the suppression of research evidence can only raise further suspicion that the latter (and first-order) suppression of evidence did, in fact, occur.

And pharmaceutical executives seem to always be baffled about why the public may not trust them?

If Lilly wants to be better trusted, company executives should try a big dose of transparency. Why not clear up the fog, and fully release the February, 2000, document, and details of why its data was erroneous, how it was corrected, and how the corrections turned out?

Meanwhile, maybe the FDA will be able to figure it out for us.

NOTE: See further discussion of the latest wrinkles in the Zyprexa case here on the Clinical Psychology and Psychiatry blog.

NOTE (May 1, 2007): See this post on Furious Seasons which links to the February, 2000 document in question above. On quick review of this document, I found that it did summarize data that the rate of "treatment-emergent" hyperglycemia in the (apparently pooled) group of patients who were given Zyprexa in multiple clinical trials was 3.6%, versus the rate in the (apparently pooled) placebo group of 1.05%. However, the document did not give much explanation about where this data came from, how the trials were done, or how the data were pooled.