By Andrew Jack at The FT:
There are lies, damned lies and ... the results of clinical trials. That might be the conclusion from internal AstraZeneca documents made public during litigation on its antipsychotic drug Seroquel, which is now coming to a close.
They highlight apparent efforts by the group to communicate favourable data selectively to boost prescriptions. Their full context may never be fully scrutinised because of out-of-court settlements being finalised with US patients and prosecutors.
AstraZeneca said: “Selected documents produced in connection with the Seroquel product liability litigation do not provide a fair and accurate picture. The company has worked diligently with the US Food and Drug Administration to ensure that the safety profile of Seroquel is reflected appropriately in the prescribing information.”
While denying the allegations, it has reached a $520m (£335m) settlement with the Department of Justice and two whistleblowers, related to past marketing practices. It will soon pay $198m to thousands of patients who developed diabetes, and has won one case after arguing that it had not been proved that the drug caused the condition.
In 1997 AstraZeneca launched Seroquel, known generically as quetiapine, to treat schizophrenia and marketed it as causing fewer side effects than existing medicines, notably extrapyramidal symptoms (EPS) such as muscle spasms.
Through additional clinical trials to persuade regulators of wider-ranging efficacy and few side effects, and aggressive marketing to make a similar case to doctors, the drug became a $1bn-a-year “blockbuster” by 2002. It has generated more than $25bn in sales over the past decade and $4.5bn last year alone.
In February 1997, Richard Lawrence, Seroquel’s marketing manager, raised the results of “this cursed” Study 15, a clinical trial that was never published but showed no advantage of Seroquel over Haldol, a rival treatment. It was one of several studies that the company cited selectively in marketing materials.
“I am not 100 per cent comfortable with this data being made publicly [sic] available at the present time,” he wrote. “However, I understand that we have little choice ... Lisa has done a great ‘smoke-and-mirrors’ job!”
In 2000, an internal “meta analysis” comparing different studies on the efficacy of Seroquel and other antipsychotics showed no significant benefit over the alternatives. Geoffrey Birkett, an AstraZeneca executive, characterised the findings for Seroquel as “unsurpassed” – a word he defined to lawyers suing the company as “possibly better ... possibly equivalent”.
One of AstraZeneca’s pivotal marketing claims to doctors was that Seroquel resulted in less weight gain than other antipsychotics. Yet as the drug became more widely used, doctors and patients reported just such effects, resulting in diabetes.
As early as August 1997, Lisa Arvanitis, a researcher at the company, warned colleagues about going “ballyhoo” on Seroquel’s lower weight gain. “We know we have weight gain but is it limited to the short-term treatment?” she wrote in an e-mail. “I’m not sure there is yet any type of competitive opportunity, no matter how weak.”
In February 1999, Nick Hough, another executive, warned on one study suggesting patients on Seroquel lost weight: “We must not get too carried away with ‘weight loss’ when we know the rest of our data appears to point in the other direction.”
In December, employees debated whether they should sanction publication of a study called COSTAR, which showed statistical equivalence at best when comparing Seroquel with the rival drug Risperdal.
“I think that we are still not comfortable about communicating the overall results of this study,” wrote Martin Jones. “Are we sure that ... we can present the EPS data in isolation given the nature of the other results?”
John Tumas replied: “There has been a precedent set regarding ‘cherry picking’ of data ... Thus far, we have buried Trials 15, 31, 56 and are now considering COSTAR ... We must find a way to diminish the negative findings. But in my opinion we cannot hide them.”
Jim Galvin warned: “Selectively using (for example) the EPS data from COSTAR is pushing it too far ... and would destroy our current high standing in the publishing industry.”
Other AstraZeneca documents show that the company was cautious about research conducted by Michael Reinstein, a psychiatrist in Chicago, who was a strong advocate of Seroquel. But – recognising him as “possibly our most important single customer” – it paid him almost $500,000 over a decade in fees and costs to present his research at conferences, and circulated his findings to doctors.
As one AstraZeneca employee put it while dispatching a copy to a colleague in Sweden in 2005: “Goes to show you that if there’s a huge need for data to support a message, the data will find its way out (despite our guidance to the contrary).”
The same could be said for the internal e-mails – something likely to make users far more cautious in future as to what they commit to computer.