Tamiflu: Dylan's retort, part III (conclusion)

This is the last installment of reader "Dylan's" commentary on the Tamiflu issue (previous posts here and here). For readability I am not "block-quoting" and reducing font size as I usually do for quotes. Instead Dylan's remarks are full width and full font size, mine in block quotes and reduced font. Dylan's intention is to provoke discussion. In that spirit I have engaged him in a dialog. Whatever you think of his arguments, which are substantive, he has done us a service. He has my thanks.

Dylan concludes:

Since the level of lethality of H5N1 first became established at around 70%, there has not been any demonstrated reduction of this figure to date (disregarding the latest cluster in Thai Binh, about which any conclusions would be premature). There could be several reasons that this state obtains, but nothing about it says anything about Tamiflu. In the pantropic and neurotropic strains of H5N1 the NA segment of the virus is encoded for the binding and sequestering of plasminogen, and this changes the nature of the hemaglutinin cleavage at receptor sites. Instead of being confined to the upper respiratory tract, the virus can now fuse with the cells in most of the major organs of the body (pantropic), or with the cells of the central nervous system, including the brain (neurotropic). With the first symptomatic expression of nonpneumotropic H5N1 a victim's condition would almost certainly be beyond the scope of any prophylactic protection that Tamiflu might confer upon him/her. Research could certainly be conducted regarding Tamiflu's express capacity to confer limited "immunity" to infection by a pathogen of this nature, but I don't think that any has been conducted at this point.

Several ideas are expressed here. Let me try to unpack them. First, whatever changes in virulence the virus is or is not exhibiting, it is separate from the efficacy of Tamiflu. I agree, but I don't see the relevance of this to the discussion. Second, Dylan hypothesizes that pantropic strains get that way via an enhanced ability to cleave HA into two pieces in many tissues, not just lung. This has been suggested in the literature, but that same literature makes clear that virulence is a complex, multigenic affair and likely involves a complicated interaction with other altered genes (like NS and PB1). To settle on changes in NA, which is involved in the mechanism of action of Tamiflu, as the major or only factor is almost certainly not the case. Moreover, the active site of NA seems well conserved, so changes in NA that enhance virulence would likely not affect the active site. The third idea expressed here is that Tamiflu only works for pneumotropic viruses. Your comment that Tamiflu would not be an effective prophylactic after the first symptoms is a contradiction and I assume an oversight. Obviously, if there are symptoms then the use of the drug is therapeutic, not prophylactic. However your argument seems to just boil down to the claim that we don't know for sure Tamiflu would work for a neurotropic strain. We have covered this already.

Number three: If you are using Tamiflu to ward off the effects of H5N1 during the course of known exposure to the pathogen, then the fact that you are not displaying any symptoms does not mean that you are not capable of infecting another person (were the two nurses in Vietnam taking Tamiflu; I would think they should have been). It follows from this that all immediate family members in contact with any health care worker, first-responder, etc., who is exposed regularly to H5N1 will also have to be provided with Tamiflu, as a practical and an ethical matter.

Your statement is correct, but does not take into account the substantially reduced viral shedding that accompanies use of the drug. The main point is to reduce the average number of secondary cases each case produces. This will shorten the epidemic. The drug would plausibly do this, even if it isn't given to all contacts.

Assume, for the sake of argument that 60,000,000 ten dose courses are available when the pandemic arrives (we're probably dreaming here, right). Then if used as a preventative measure, which would require each person to take one dose a day for six weeks, at least, roughly 14,300,000 people world wide could be protected for six weeks. Three-million-six-hundred-thousand of those could be health care workers and first-responders, etc. Assuming three family members for each of these people, all 60,000,000 ten dose courses have been used up. And that these people are protected in the first place assumes that Tamiflu works perfectly. We have problem here, as I see it. How many nations around the world are going to be relying very heavily on Tamiflu when the pandemic arrives? What happens if Tamiflu doesn't work almost flawlessly? Anybody care to answer?

I agree with this argument. Tamiflu cannot stop a pandemic nor will there be sufficient supplies to make much of a dent if the event occurs in the next two years, which seems likely. I am not sure what we are supposed to take away from this argument, however. It is not an argument against the drug.

Number four: All human isolates of pneumotropic H5N1 tend to produce high cytokine response. This involves elements of the immune system that are the body's first response to any flu infection. They include interlukin, interferon, macrophages, tumor necrosis factor-a, and several others. I've encountered research that demonstrates that a single passage of H5N1 through mouse (not the best model, or the one of choice where flu studies are concerned, as I understand it) lung tissue was sufficient to evoke high cytokine response. I don't want to get too far into this -- I'm sure revere can explain it better than I could -- because it's very complicated, and it isn't necessary to make the point. Something in the NS segment of the viral RNA apparently interferes with the regulation of the cytokine response. When the response fails to neutralize the virus -- which is essentially immune to it -- it simply continues to produce progressively greater, and more "toxic," amounts of pro-inflammatory cytokines. The result is that the lungs are overwhelmed with the fluids that are produced. Most of the people who are dying because of H5N1 infection are probably dying as a result of this being the way that their body is responding to the pathogen. Tamiflu would have to prevent this at a very early stage of infection, because once begun it is rarely, if ever, reversible.

As you indicate, this is a complicated argument and the subject of a great deal of ongoing research. The NS (non structural) protein at both its C-terminal and N-terminal ends is capable of dampening cytokine response in infected cells. It is not clear whether this leads to a positive feedback situation with a consequent "cytokine storm" that is part of the problem or whether something else or many something elses are occurring. The way Tamiflu works is to prevent the virus from massive replication by preventing the virus from budding off infected cells and going on to infect other cells. It is likely that there is a balance between the various response mechanisms at different times and scales that can be tipped one way or another and probably turns on particular host characteristics. Tamiflu may or may not affect this complex chain of events. What we know so far, however, is that the drug is quite effective against influenza A infection if given prophylactically or early in infection (first 2 days). So far that is the bottom line. Arguments that some strains will be unaffected are hypothetical at this point. However to say that we may discover instances where the drug will not work is not an argument to say we shouldn't use it or that if we do use it it will be useless.

I'm sure that I have made some mistakes here, but on the other hand, I think that I have also proposed some fairly important questions.

Dylan, you have done a remarkable job in synthesizing a vast amount of information. To the extent I or anyone disagree, we should be required to make a counter argument, which I have tried to do, however briefly. In any event, I hope our conversation is useful and even interesting to some of our readers.

Many thanks to you.