Does Tamiflu actually work (and is it safe)?

Most of us will be familiar with the H1N1 (‘swine flu’) pandemic that never was. At the height of the hysteria here in the UK, individuals could phone up a help-line, answer a few questions, and get sent the drug Tamiflu (oseltamivir). According to a news report here, it turns out that only about 20 per cent of the people who were dispensed Tamiflu (over the phone) for H1N1 infection actually had H1N1 infection. Apparently, about 800,000 packets of Tamiflu (at about £15 a pop) were dispensed erroneously.

But the real story is not so much about this, I think, but the fact that it also turns out there really is no very good evidence that Tamiflu does much good. A review of the relevant evidence has this week been published in the British Medical Journal [1]. The authors of this study belong to an esteemed and generally trusted collection of researchers referred to as the ‘Cochrane Collaboration’.

One of findings of this study is that Tamiflu does reduce the duration of influenza symptoms, by about a day. The authors describe this as a modest benefit. I’m inclined to agree with them.

Another outcome assessed by this review related to the complications of flu. Flu is generally harmless, but it becomes less harmless should it lead to complications such as chest infections (lower respiratory tract infection). Now, previously there had been a meta-analysis (a conglomeration of relevant studies) that appeared to show benefit on this account. This was included in a previous review (2005) by the same authors. Now, though, the authors have decided to discount it. That’s because of the 10 studies that made up the meta-analysis, only two were actually published. According to the review authors the remainder were offered to us under conditions we thought unacceptable, and what was offered to us was insufficient to analyse properly. One wonders why the original meta-analysis was included in the previous review, seeing as it was of such dubious quality. But, I suppose even researchers from the esteemed Cochrane Collaboration can make mistakes.

The author also comment that The remaining data showed no benefit for oseltamivir against complications. In other words, there really is not good evidence that Tamiflu reduces the risk of complications.

And that’s not all, with regard to their assessment of the potential harms, the authors comment We therefore found under-reported evidence of varied quality, which could not answer concerns about the toxicity of neuraminidase inhibitors, especially oseltamivir. Governments should set up studies to monitor the safety of neuraminidase inhibitors.

So, in summary, what we have in Tamiflu is a drug of modest benefit and, largely, unknown safety. And this drug was apparently dished out to a million people, the vast majority of whom didn’t even have the infection Tamiflu was supposed to be treating. Excuse me if I don’t regard this as a shining example of ‘evidence-based medicine’. But, we all know we can trust out Governments to give us the best advice and healthcare, right?

References:

1. Jefferson T, et al. Neuraminidase inhibitors for preventing and treating influenza in healthy adults: systematic review and meta-analysis. BMJ 2009;339:b5106