Article Text

Download PDFPDF
Anti-viral drugs in flu: not that good?

Statistics from

Those of us who were on the front line during the 2009 H1N1 swine flu pandemic will remember giving antiviral neuraminidase inhibitors to, it seemed, almost every febrile child—particularly oseltamivir (Tamiflu), as guided by official advice. My impression at the time was that it didn't help much, and vomiting was common.

A controversy has raged in the media (, and has become a ‘cause célèbre’ for those who want pharmaceutical companies to make public all their trial data, not just those that favour the drug. The BMJ has been active in this campaign, and has now published two systematic reviews on effectiveness and adverse effects of oseltamivir and zanamivir (Relenza). These reviews were conducted by the Cochrane collaboration, after going to great lengths to obtain all trial data from the manufacturers. For the most part they separate data on adults from children.

For oseltamivir, they found 83 trials altogether, but were able to do a complete analysis on only 20 of them (Jefferson T et al. BMJ 2014;348:doi:10.1136/bmj.g2545). In children, treatment produced a reduction in time to alleviation of symptoms compared to placebo of mean 29 hours (95% CI 12–47 hours, p=0.001). Interestingly, no such benefit was seen when only children with asthma were included: perhaps because concurrent increase in acute asthma treatment eclipsed any effect of the antiviral. They were unable to show a significant reduction in hospital admissions. Likewise, there was no demonstrable effect in reducing incidence of ‘pneumonia’ (variably defined) or other secondary infections. Side effects were common, with vomiting in children reported in 5.3% (95% CI 1.7–10.3%). They also looked at its effectiveness in prophylaxis for exposed contacts across all age groups, and found that it reduced symptomatic flu by 55%.

The findings for the zanamivir systematic review were similar, but the reduction in time to alleviation of symptoms was not statistically significant in children (Heneghan C et al. BMJ 2014;348:doi:10.1136/bmj.g2547). Again, there was no benefit in preventing complications. As might be expected with an inhaled drug, side-effects such as vomiting were less prevalent. Effectiveness in prophylaxis was modest, as for oseltamivir.

All sorts of anomalies and inconsistencies emerged from these analyses: for example there was no consistent definition of ‘pneumonia’. Also, these studies were all done in the context of seasonal flu: in pandemic flu, the findings could differ.

In 2009, the UK government spent nearly £500 million stockpiling antiviral drugs. Decisions were made on the basis of data which at the time made these drugs appeared considerably more effective than these analyses would suggest. The commercial incentives are obvious. Not surprisingly, oseltamivir manufacturers Roche have contested the findings of these BMJ reviews, accusing the authors of using unorthodox statistics.

What no-one doubts is that vaccination is more effective than any drug in prevention: as long as the appropriate strain can be given in time.


  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.