Tamiflu has the interesting property that the drug in the tablet is ineffective against the influenza virus. Tamiflu gains function inside the body due to chemical processing in the liver that reveals a reactive moiety. If you have an influenza infection, some proportion of the oseltamivir molecules will end up binding to neuraminidase, but the remainder will be excreted without further modification. This means that every time you are dosing a person with tamiflu you are also dosing the environment with it. That could be trouble, because some strains of influenza incubate in the wild, among ducks, for instance, who are likely to be swimming on the rivers or ponds where excreted tamiflu might be expected to end up. If these animals are constantly exposed to very low doses of tamiflu, it is quite conceivable that the viruses within them could evolve resistance to it without losing neuraminidase activity.
The question then becomes whether anything in the environment might degrade tamiflu before it reaches the ducks. We have one advantage in this regard: we generally treat our sewage before we let it go. In an experiment I do not envy one little bit, a Swedish team led by Jerker Fick tested actual sewage to see if oseltamivir carboxylate (the active form of the molecule) would break down in it (the article is open source on PLoS ONE). They took raw sewage, as well as samples of sewage at various stages of treatment, added oseltamivir, and incubated them with a similar temperature and duration to what they might experience in the treatment plant. The results were not encouraging: the amount of tamiflu they were able to recover from these samples was as much as or more than what they could recover from plain tap water incubated under the same conditions. Moreover, they found that the UV/visible light absorbance spectrum of tamiflu would not be conducive to photolysis in the environment. The conclusion from this is that we cannot expect any help from our water-treatment facilities in eliminating oseltamivir from water supplies that might eventually reach host organisms for influenza.
This puts us in a bit of a bind. It is impossible to use tamiflu without releasing it to the environment, even if it is used in accordance with rigorous guidelines designed to prevent resistance from arising in people. Yet, the more tamiflu that reaches the environment, the more likely it is that continued exposure among waterfowl will lead to the development of resistance. Although this is bad news for the future of tamiflu, it doesn’t necessarily herald the inevitable onset of some kind of superflu. For one thing, some of the mutations that confer tamiflu resistance also diminish the ability of the influenza virus to infect hosts. Moreover, there’s no guarantee that tamiflu in the environment will end up dosing avians — it might break down in their stomachs, even if it doesn’t in the sewage plant.
Also, some mutations that confer this kind of resistance leave neuraminidase vulnerable to other inhibitors such as zanamivir (Relenza). Granted, Relenza does not appear to be as effective as Tamiflu, but the fact that resistance is not necessarily shared is an encouraging sign that it will be possible to continue developing improved inhibitors even if the worst happens.
However, the results of this paper and the fact that tamiflu-resistant strains can be transmitted to people who have never used tamiflu indicate that we should be making a greater effort to prepare ourselves for the appearance of resistance. Much has been made about efforts to stockpile tamiflu in case of a pandemic, and while there is nothing wrong with that idea per se, it is foolish to assume that tamiflu is some kind of panacea. If a pandemic does occur, with the attendant explosion of evolutionary possibilities for the influenza virus, the chances of a rapid development of resistance in humans may be quite high. At that point it will be too late to start developing alternatives. In the near term, we should do as has been suggested — use tamiflu sparingly and aggressively to minimize the chances of developing resistance in humans. But we should also provide incentives or a mandate to develop a back-up plan in case tamiflu fails.