Governments have been stockpiling the antiviral drug oseltamivir as a defence against pandemic flu. Now the medical journal BMJ has claimed there is insufficient evidence that the drug prevents serious complications of flu to warrant the policy. In a series of articles, it says the Swiss pharmaceuticals giant Roche, which sells the drug as Tamiflu, has concealed the detailed data needed to support claims of its effectiveness.
鈥淎lthough billions have been spent on oseltamivir in the face of pandemic influenza,鈥� the journal states, investigators 鈥渇ound that the public evidence base for this global public health drug was fragmented and inconsistent鈥�.
But the doubts that BMJ casts over Tamiflu do not relate to the jobs it is meant to do in the pandemics for which it has been stockpiled. The journal has questioned the drug鈥檚 ability to prevent the bacterial infections that can follow ordinary seasonal flu, but in pandemics, it is intended to slow initial spread of the virus or to fight its direct attack on the lungs. For these jobs, it is vital.
Advertisement
This view is supported by the US Centres for Disease Control and Prevention, which says that Tamiflu is 鈥溾�� from pandemic H1N1. Any bacterial complications are treated with antibiotics.
Cochrane review
BMJ鈥榮 story begins in 2006, when the 鈥� which carries out independent reviews of treatments 鈥� that Tamiflu reduced the chance that an otherwise healthy adult with flu will also get bacterial bronchitis, sinusitis or pneumonia by 67聽per cent.
The Cochrane review pooled the results of studies in which otherwise healthy adults with confirmed flu had been given either Tamiflu or a placebo. The (PDF), carried out by Laurent Kaiser of the , Switzerland, and scientists employed by Roche, found a 55聽per cent reduction in bacterial complications in people given Tamiflu. The Kaiser study was itself a pooling of 10 smaller studies.
Last July, , a Japanese doctor complained to Cochrane that eight of the Kaiser studies were based on data held by Roche or published only as abstracts. Of the two that had been fully published, one found that Tamiflu (PDF) while the other .
Tom Jefferson of Cochrane asked the studies鈥� authors for the full data, but they said Roche had it. Roche was then contacted by a reporter for a British TV programme, Channel聽4 News, who asked for the data, explaining he was working on a story with Jefferson.
Data block
Roche did not release the data. The problem, says David Reddy, head of Roche鈥檚 pandemic taskforce, was that the data included patients鈥� initials and birthdates. Legally, regulatory agencies, doctors and Roche鈥檚 study managers could see it, but no one else unless they promised confidentiality.
鈥淲e were concerned because we didn鈥檛 understand why the TV channel was involved,鈥� says Reddy. 鈥淪o we asked Dr Jefferson to sign a confidentiality agreement.鈥� He refused, saying Cochrane reviews require published data. Roche sent him summaries of the data and offered to provide further detail if needed. In October, however, Jefferson said it was too late to include in new analysis.
闯别蹿蹿别谤蝉辞苍鈥檚 , published in BMJ on Tuesday, omits the Kaiser paper, and with it two-thirds of the patients in the original analysis.
The earlier Cochrane analysis found Tamiflu cut the risk of complications by 67聽per cent, but the new one cuts this to 45聽per cent, although the finding is statistically uncertain because of the much smaller sample size. Yet in his conclusion, Jefferson is less equivocal: 鈥淭he remaining evidence suggests oseltamivir did not reduce influenza related lower respiratory tract complications.鈥�
No clear conclusion
Other researchers say the study no longer contains enough participants for a clear conclusion to be drawn either way.
鈥淭he number of complications reported in the BMJ paper is small,鈥� says , a professor of clinical pharmacology at the University of Warwick, UK. 鈥淭he data do not show a significant reduction in complications with oseltamivir, but they are consistent with anything from a 78聽per cent reduction to a 35聽per cent excess in complications with drug vs placebo.鈥� What is needed, he says, is more data.
Roche has now and says it will make the detailed data available to researchers shortly.
Journal reference:
