Tamiflu and behavior "problems"

Over a year ago reports from Japan began to circulate that the influenza antiviral, Tamiflu, which is prescribed there often for uncomplicated seasonal influenza, was causing abnormal behavior, most worrisomely delirium and suicidal behavior in children. The drug is approved for adults and children over a year old. At that time the FDA decided the evidence was insufficient but planned to revisit the issue in a year. Now the year is up and FDA has apparently decided the evidence is stronger. They are now recommending patients who take Tamiflu be "monitored" for abnormal behavior. It isn't clear what it means to monitor someone except to be aware of the problem.

The new notice follows 103 reported cases of delirium, suicidal behavior and other problems between August 2005 and July 2006 (News-Medical.Net). Three cases were fatal:

Ninety five of the cases were apparently Japanese children treated with the drug.

The agency's review found that of the 12 deaths there was one suicide, four cases of sudden death and four cases of cardiac arrest.

There were 32 reports of psychiatric "events" among children, including hallucinations and abnormal behaviour and in two cases a 12-year-old and 13-year-old jumped out of second-floor windows of their homes after taking the medication.

The three deaths referred to are apparently the two defenestrations and a suicide. As of this writing FDA did not have any information on their website so the details remain a bit murky. Neither FDA nor European drug regulatory agencies are willing to say there is a causal connection between these reports and the drug. The drug's maker, Roche, denies there are any neuropsychiatric effects. They maintain the observed events are the result of influenza, not the drug. The Japanese already have a warning label similar to the one FDA will now recommend.

So add neuropsychiatric side effects as another urgent question about Tamiflu. Another is how effective it is against influenza viruses with N1 subtype, like H5N1, versus N2 subtype where it has been employed clnically in Japan. The experience with H5N1 is equivocal so far and the failure to collect systematically the clinical data that might shed further light on this is a major failing of WHO. At best Tamiflu shortens symptom recovery by a day or so when used therapeutically. Roche's data suggests it is more effective when used prophylactically, and in a pandemic one would expect that it would be used in this fashion for essential personnel and caregivers of infected people, professional and at home. When used in this fashion, with otherwise healthy people, the risk-benefit equation is changed. If tens of millions of people will be taking this drug from national stockpiles, then even relatively rare side effects can loom large in terms of numbers of people affected. Some resolution to the side effect problem is urgently needed. Even if effective for H5N1, the question of how readily resistance will develop also needs answering.

Unfortunately the list of urgent questions about influenza in need of quick answers is getting pretty long. Is someone in the federal government keeping track and directing resources to answer them?


More like this

Health Canada is following the US FDA in warning of adverse neurobehavioral effects of the influenza antiviral, Tamiflu. The drug is prescribed much more often in Japan, where it is used in seasonal flu, than in the US. The effects have been reported mainly in children and have included some…
First Tamiflu (oseltamivir), now Relenza (zanamivir): Health officials [in Canada] are investigating whether Relenza - a drug provinces have stockpiled in case of a pandemic flu outbreak - can be linked to fatal reactions or abnormal behaviour in children. [snip] The investigation is a response to…
Tamiflu side effects have been much in the news and we have concurrently been posting our mega-series on modeling antiviral resistance in influenza control. The two subjects are related in two ways, one obvious (Tamiflu is the main antiviral being stockpiled for influenza control) and one not so…
Press releases are the way a lot of scientific information is released today. Straight to the public, no peer review. This has advantages and disadvantages. The advantages are speed and directness. No filtering through reviewers, journal editors, colleagues. And of course that's the disadvantage,…

I agree with the more data thing. I always worry that negative side effects are well known, that the relative risks are understood, and that the actions people take are consistent with their perception of the risks. It is good news that the first bit of that is getting done here (with the info out there).

I don't know what the impact of the prophylactic use of Tamiflu would be during an epidemic (I do learn a lot from all the people here though). However, I often think about some of the social effects that are not always part of a narrow public health focus. An epidemic certainly will be a stressful time for everyone and health workers likely more so. I just wonder if the neuropsychiatric side effects will be greater under those stresses (small percentagewise but as you note still likely to effect large numbers of people).

By Floormaster Squeeze (not verified) on 15 Nov 2006 #permalink

The piece that is not here is how many doses does this represent? I understand that Tamiflu is commonly given in Japan -- are we talking thousands of doses? Millions? Hundreds of millions? I'd like to know because the frequency of side effects is important. I'd be far more worried to learn it was 100 events per thousand doses than if it's 100 events per 100 million doses. Do you know?

By Ellen Rice (not verified) on 15 Nov 2006 #permalink

Ellen: I think there were about 25 million prescriptions for Tamiflu in Japan last year. There is always the issue of rates versus numbers and each has its own significance. For some things rates are the most important factor, for others it's numbers. For example, you may be interested in the death rate for a city if you are comparing it to another city, but you are interested in the absolute number if you own a funeral home or are in administrative planning. We have discussed this issue elsewhere with respect to vaccines, where we have some of the same issues. If the vaccine causes sudden death in one in a million but you vaccinate 100 million people, 100 cases of sudden vaccine associated deaths can put a serious crimp in a vaccine program (or a Tamiflu prophylaxis program) even if the risk benefit equation is very favorable (consider the thimerosol debate).

So your question is pertinent but the post was discussing the significance of the finding for FDA and flu policy.

Ellen and Revere: although there have been 25 million prescriptions in Japan of Tamiflu last year, that isn't going to clear things a bit. Because just Tamiflu is a medicine that has been purchased not for an actual flu, but to be stockpiled for H5N1. Maybe in Japan they are so dedicated to work and schools that they won't miss one day there, and maybe they differ from some other countries and desperately take anything they can get to crawl back to their productivity. But it wouldn't surprise me if there hasn't been used more than one tenth of all Tamiflu there and the rest is stored for 'just in case'.
I myself had never heard of Tamiflu before I started to get interested in H5N1.

Taiwan is now backpedalling from the US satement and reassuring its populace that the effects observed in Japan were due to the flu being treated, and not the drug used to treat it.

By Lisa the GP (not verified) on 16 Nov 2006 #permalink