Tamiflu resistance: digging under the headline

The news that H5N1 viruses isolated from an uncle and niece in Egypt who died in December has been found to carry a genetic change suggestive of resistance to the main antiviral drug oseltamivir (Tamiflu) headlined the H5N1 newswires yesterday. Specifically, WHO announced that genetic sequencing had found the N294S change in the isolates (explanation below the fold). It is not clear at the moment whether the change occurred during treatment of the pair with the drug or the virus carried the change when it infected them. It is also not clear what the clinical significance of the change is. We need to explain further.

Oseltamivir (which we will call by its trade name, Tamiflu) is a neuraminidase inhibitor. Neuraminidase (NA) is a protein that acts like an enzyme (it catalyzes a specific biochemical reaction) and it is present on the virus in the form of a glycoprotein, one of the two glycoproteins that stud the viral surface (for more on the basics of glycoproteins in lay language see our series that starts here). The other glycoprotein is hemagglutinin (HA). There are 9 immunological classes of NA and 16 of HA and the specific HA/NA combination of these types are what distinguish the various influenza A subtypes named for them (e.g., H3N2 or H5N1). These are also the main elements that our own immune systems recognize. It is the HA glycoprotein that becomes the basis for most of the influenza vaccines.

The HA protein plays an important role at two critical points in viral infection of a host cell. The first comes when it binds to a host cell receptor of the right kind. Currently we believe the main receptor is something called a sialylglycan (we have discussed details in the glycoprotein series linked above). Once docked to the receptor the virus is "taken into" the cell by being enveloped by the cell's membrane in a process called endocytosis. Think of a closed bag of liquid you push a small stone into the side of, making an indented pocket with a sort of "neck" to the outside. Now pinch off the neck part so that the stone is now inside the bag surrounded by its own little piece of the bag wall. It still isn't really inside the bag yet because of the surrounding coating of bag wall, but if it could get through that it would be bathed by the liquid in the bag and truly "inside." That's what the virus needs to do because it wants access to the cell's genetic and protein-making machinery so it can make copies of itself, its only real function. HA takes part in this process, too, helping the virus "fuse" with the surrounding little bit of bag wall and exit from its internal bubble into the cell proper.

That's HA. What about NA? Here's what is generally agreed upon. After the virus makes copies of itself and the parts fully assembled again to be budded from the surface of the host cell, it has to be cut loose from the same sialylglycans it used to attach, on entry. That is what NA does. But the mysteries of NA have not all been unraveled and at least two other functions have been suggested for it. Matrosovich et al. have suggested NA helps destroy the attachment of the virus to the many "decoy" sialylglycans found in the upper respiratory tract contained in mucus and other components not part of the cell membrane. In addition, Ohuchi et al. have recently put forward another role for NA. Once attached, it has been suggested that the virus needs to move around the surface a bit to find the right spot for the endocytosis, i.e., not every spot is suitable for the virus to invaginate into the cell, the next step in infection after attachment by HA. Without NA the virus is held to tightly on one spot and can't move around enough to find the "door" to get in. It's in the building foyer but still needs to find the front door and go through it. One of the mysteries about NA function is why it is good to break the sialylglycan bond on exiting but why thisisn't bad when entering. It is likely that the answer is that there is a balance between HA and NA. Too much NA activity is bad for entry and too little is bad. This might explain why viral infectiveness is not just a matter of whether the HA can find a receptor but also whether it is paired with the right NA. In any event, neuraminidase inhibitors like Tamiflu work by decreasing NA function, making it hard for them to exit after budding at the surface (this is accepted) and possibly making it hard for them to enter as well, by getting them stuck at the surface before they can get in (speculated). So the story is more complicated than first thought. No surprise.

Antiviral drugs like Tamiflu and Relenza (generic name zanamivir) work because they look like the sialylglycan the virus is trying to find for its initial attachment. They become like the decoy receptors in mucus. The drugs bind in a chemical recognition site, a sort of "pocket" on the NA glycoprotein. In order for the drug to bind, there must be a fairly specific set of attachment points in the pocket and this is determined by the sequence of amino acids that make up the NA protein. which in turn is determined by the genetic sequence in the viral RNA. So if that genetic material changes (a mutation) it can alter the attachment points the drug needs and produce an NA that no longer binds the drug. Of course that NA might also not work very well for its function, either, and there is some suspicion that the resistant mutant viruses might be less fit than the normal virus (called the "wildtype"). A resistant virus has been isolated from a patient prior to this from a young Vietnamese girl (who may have been infected while caring for her 21 year old brother). She recovered completely. She was found to have circulating viruses with varying genetic sequences, one highly resistant to Tamiflu (carrying the H274Y mutation), one with only slight resistance (H294N) N294S and one highly sensitive (wildtype, the same as the brother's virus). Thus the virus may have mutated while replicating within the patient. On the other hand, de Jong has reported two cases that ended fatally and also had H274Y.

What are these mutation designations? The wildtype condition is the first letter in H274Y and gives the one letter code of the amino acid that was there before the change. 274 is the position along the length of the NA protein. The second letter is the amino acide after the mutation. Here the amino acid histidine (whose one letter code is H) has been changed or mutated to the amino acid tyrosine (one letter code Y) at position 274. We don't know all the mutations that might confer Tamiflu or Relenza resistance, but based on experience the Japanese have had with the H3N2 virus and some experiments with human volunteers infected with H1N1 viruses, the mutations R292K, N294S and H274Y seem to be important to drug binding in the pocket. Much of the previous predictions were based on three dimensional structures of N2 type neuraminidase, the 3D structure of N1 not having been worked out. Last September the 3D structure of N1 was published and work is now proceeding using that information to understand better what is going on with resistance to neuraminidase inhibitors and possibly design new drugs for the purpose.

How do we know a virus is resistant to Tamiflu? It is done by detecting one of those three mutations in the genetic sequence. Since we have very little information on the clinical importance of those markers of mutation (the jump from sequence to biology is a big one), it is not clear what the significance of the N294S mutation is (it is an asparagine, N, to serine, S, mutation at position 294).

All this is explanation for this statement:

"What the resistance tests look for are markers associated with antiviral resistance," though finding the markers did not necessarily mean Tamiflu would not work, said Dr. Angus Nicoll, flu director at the European Centre for Disease Prevention and Control.

[Dr. Frederick] Hayden said the mutations found in Egypt were different from Tamiflu-resistant H5N1 viruses found in patients two years ago in Vietnam. The Vietnamese strains were definitely resistant to Tamiflu, whereas the Egyptian viruses have only proven they are not as susceptible to the drug, he said. (AP)

This sounds suspiciously like spin. But in fact it is based on what we know, and more importantly, what we don't know. Another good reason for never spinning. When you are telling it straight you want people to believe you.

More like this

CDC guidelines for antiviral therapy for swine flu infection: This swine influenza A (H1N1) virus is sensitive (susceptible) to the neuraminidase inhibitor antiviral medications zanamivir and oseltamivir. It is resistant to the adamantane antiviral medications, amantadine and rimantadine. (CDC)…
Since the antiviral agent oseltamivir (Tamiflu) has been touted as the global savior should a bird flu pandemic materialize the idea has been haunted by the specter of Tamiflu resistance. What if H5N1 becomes resistant to the drug? Is all lost? Now it is being reported in the media that the…
Via the Clinician's Biosecurity Network Report we learn of a new study from the Webster St. Jude laboratory in Memphis showing that H5N1 can mutate to oseltamivir (Tamiflu) resistance without any loss in genetic fitness. Tamiflu resistance has been seen but infrequent and there was considerable…
There have been three reported oseltamivir (Tamiflu) resistant isolates of H1N1 swine flu (added: and now a fourth in Canada) but with those exceptions all others have been sensitive to this oral antiviral. This is in marked contrast to the other H1N1 strain, the seasonal variety which is almost…

A remarkable piece of analysis and explanation.


Well it would have surprising if we'd gotten through the flu season without seeing more drug-resistant strains emerging somewhere. I susbscribe to the theory that the precondition for evolution is underdosing of patients, but what else might be necessary, who knows? Is it a spontaneous mutation, or are there a few strains out there circulating, and which already have the genetic changes? When I look back on older papers relating to zanamivir testing I think there was a bit too much optimism.

I wonder if you would mind clarifying one small point, you refer to the Vietmanese patient as having an H294N change but the link gives resistance as N294S is H or N the wildtype for this position or is this a secondary change?

Thanks for the explanation.

Why have we been talking about Tamiflu so much, anyway? It is after all administered by only a capsule. Wouldn't Relenza be better for far airway symptomic disease being an aerosol?

Sorry, one "far" too much! :)

The patient in Vietnam with N294S is a red herring. She is the 14 year old sister of the patient who holds the record for hospitalization with H5N1 (something like 85 days). He infected his sister and his nurse.

The sister developed Tamiflu resistance, but that was primarily due to H274Y. N294S was a minor component in the H5N1 isolated during treatment.

Egypt was quite different. Both patients had N294S in samples collected 2 days after start of treatment (2 X 75 mg Tamiflu). Sequencing was directly from clinical samples. There was no hint of wild type N at position 294 and no hint of H274Y.

N294Y in Egypt had nothing to do with the Tamiflu treatment of the two patients, who did not respond to the treatment and died (as did the third case, but H5N1 was not isolated).

Revere, this is impressive and so well-explained. I love how much I learn by coming to this site.

Did they test for HIV? Just wondered if the 85-day hospital patient was HIV+.

By crfullmoon (not verified) on 19 Jan 2007 #permalink

"there is some suspicion that the resistant mutant viruses might be less fit than the normal virus"

Can this mean also that (some forms of) resistance to Tamiflu might be a 'good thing' from our perspective?

Thanks for this post. Your descriptions paint a 3-D picture for me. I'm amazed I can come this close to understanding what's happening without having a science background. I really appreciate the time and effort you take to do it.

Thanks for the explanation... all the "have the such-and-so mutation" remarks that pop up periodically will make better sense to me now. :)

"there is some suspicion that the resistant mutant viruses might be less fit than the normal virus"

It would seem that the mutant in Egypt was quite fit...it killed both patients infected for a 100% CFR.

Authorities have a million reason why things aren't the way they are...except the facts of the matter.

Tom DVM: I don't remember my statistics class that well, but I remember the importance of a large enough sample size, and two people definitely isn't.

Say the "real" CFR of the Egyptian virus were 50%, or 0.5. That would yield a 0.25 or 25% chance that both patients would die (and also a 25% chance that only person A would die, a 25% chance that only person B would die, and a 25% chance that both would live). A "real" CFR of 40% would yield a 16% chance both would die; a "real" CFR of 25% would still yield a 6.25% chance both would die. That's not a large chance, but it's certainly non-trivial.

It's hard not to do the math when you're seeing reports of cases and deaths coming in, but I worry that people are extrapolating the deadliness of the virus in small clusters far beyond what the data support.

caia. Well said.

I did not mean to infer statistically significant...I was responding to a comment, implying resistance to Tamiflu is inversly related to virulence...

...which is a remarkable statement given the two patients being discussed died.

There is a clear history of rapid development of resistance...

...hope and spin after the fact, are not going to change the facts.

JJackson: Thanks for catching the mistake. Corrected. Late night tiredness and long day at pesky day job.

TomDVM: N294S does not confer much resistance, which suggests the function of NA is pretty much intact. I'm not sure why it would be a red herring. It is what it is. No attempts to deceive or mislead were made by WHO IMO. They reported the mutation. If they hadn't they would have been criticized (correctly). Oseltamivir looks very much like sialic acid. It is a reasonable thought that if NA won't bind oseltamivir well it also won't bind NA well and thus not perform its function well. Whether this is true or not remains to be seen. One patient with H274Y survived and two didn't, so we still don't know the answer. If you look at the graphs in Le's Figure 1, b, there is not not much difference between the sensitive and resistant clones in ferrets (Nature article by Le, Oct. 20, 2005), leading to the question whether H274Y was really resistant.

TomDVM: Exactly what are the facts you refer to? Two patients died in Egypt and they had N294S mutations. What is your evidence that N294S confers much resistance to H5N1? They are described as slightly resistant in Le's paper and this is clear when looking at the IC50s for neuraminidase inhibition comparing it to the highly sensitive wildtype clone versus H274Y. But even the latter clone shows little difference in viral titers in treated ferrets over time in Le's Figure referred to above. Facts, please.

caia (or anyone else who can help:) Speaking of statistics: I am still trying to find reasons for optimism, or at least reality based concern. One thing that would help is knowing what percent of a given population (say, in the United States, for example) is expected to come down with an ILI during flu season.

If I knew the percent, then I might be able to calculate about how many indonesians during any given month of flu season (for example, January) might be expected to have an ILI. From those figures, I might be able to calculate the expected number that would have pneumonia-like complications. And from those figures I might be able to infer whether this current spate of suspect cases is within the normal parameters, despite the seeming severity of symptoms in many.

When we look only at the numbers of very ill being reported, it is easy to be quite concerned that this current outbreak is statistically relevant. Especially since it is proven fact that the tests to determine presence of the virus in suspect cases are unreliable. So we need a norm to compare these cases with to really know for sure.

Is there anyone out there able/willing to do the math for the rest of us? Or if not, at least to provide the statistical average rates for flu and complications of flu, so one might attempt it oneself?

By mary in hawaii (not verified) on 19 Jan 2007 #permalink

Sorry for being dense - I suspect this is a matter of my mental oversimplifications combining with the necessary simplifications in your description...

I gather that the HA component finds and binds to the target sialylglycan, and that NA is involved in breaking that bond for release or reattachment. And I gather that the the decoy sialylglycan in mucus similarly grabs (or is grabbed by) the HA component, which will stop that virus if the virus' NA can't release it.

So, with the vast assumption that this is correct, how is Tamiflu working? You explained that Tamiflu acts like the decoy sialylglycan - which, by my understanding, means it binds the HA component. But then Tamiflu inhibits the NA by creating a pocket around important chemical recognition sections.

How did it go from binding with the HA to working on the NA end? I think I'm just having a hard time getting my head around the relative sizes. Is it simply that the molecule of Tamiflu is - err - big enough to grab an HA segment and still reach an NA section, too?

You also said that most vaccines attack the HA function. How is the action of Tamiflu different, then - at least on the HA end? Is it just because it also goes after the NA side of things, or do the other vaccines do something extra to the HA beyond simply acting as decoy sialylglycan?

By SmellyTerror (not verified) on 19 Jan 2007 #permalink


I'd suggest you take a look at the graphic near the bottom of the influenza primer part II on the FluWiki at http://www.fluwikie.com/pmwiki.php?n=Science.InfluenzaPrimerII#replicat…

As a new virus buds off from the cell, it has both HA and NA on its surface. Since the cell that the virus is leaving still has sialyglycans ("receptors" in the graphic) on its surface, the virus can get "stuck" on the cell it's leaving. The NA helps break off "cleave" these receptors so that the virus can break free to infect new cells.

The neuraminidase inhibitors (which resemble the sialoglycan receptors) seem to work by preventing the NA from doing its job by binding to the NA. With the NA inhbitors blocking the NA enzyme, the ability of the virus to replicate further in new cells is diminished.

The reason drugs like Tamiflu need to be taken early in an influenza illness is that they do not kill the virus, just slow down its spread. If you don't take the drug before lots of cells are infected, it may not have much effect on the course of the illness.

At least, that's how I understand it. Revere can correct any errors. Also read the FluWiki influenza primers if you are interested in the science. (I think Revere wrote them.)

Thanks for answering, bc!

...but that's the source of my confusuion. Tamiflu binds to the NA receptor. But how? You mentioned that it resembles the sialoglycan receptors - but don't these receptors bind with the HA end, not the NA? How is it getting to the NA?

To restate: If a function of NA is to cleave bonds with sialoglycan receptors, and a function of HA is to create bonds to sialoglycan receptors, then how is a drug that emulates the sialoglycan receptors binding to NA? Normal sialoglycan receptors don't (do they?). Is it attaching to the HA then attacking the NA (perhaps when the NA tries to break the bond)? Or is it skipping the HA and going straight for the NA?

If the latter, how is that anything like the activity of normal sialoglycan receptors?

By SmellyTerror (not verified) on 19 Jan 2007 #permalink

bc, ST: bc has it right. The NA also has to recognize the sialylglycan so it can break the bond between it and HA. They recognize it in different ways, but bc's explanation has it right.

MiH: I think you can assume about a 25% ILI per flu season in the US (the number is elusive but that's a good ballpark). In Indonesia there are about 500,000 pediatric pneumonias a year in West Java (according to data I got last year from someone familiar with the statistics).

Revere. I may be a simple farm animal veterinarian...but it seems to me the facts are that two people had the mutation and two people died.

If I had two dogs with the mutation and two dogs died, in a row, as a result of my protocol...I would be both examing the pathogen and the protocol.

Also, it is highly likely, that this mutation has shown up before but either the cases weren't identified as H5N1 or the virus wasn't isolated in identified cases; we have no way to tell because the WHO is complicit in a cover-up and given the potentials in the future, that is unethical...or maybe human medicine really is fundamentally different than animal medicine.

I wasn't talking about H274Y...that is a seperate issue involved in a geographical distinct cluster...and if I remember right, a greater number of Tamiflu resistant cases in Vietnam died as well.

Do you really think given the evidence in Japan and now the evidence in Vietnam and Egypt that antivirals are going to make any substantial (statistically significant /:0) difference to a pandemic with H5N1?

Or is the Relenza just SO ineffective that it isn't even worth thinking about?

Tom: And the fatal cases without the mutation? (most of the fatal cases being without the mutation means that being Tamiflu sensitive is bad, too?) The data show this mutation doesn't affect much in the way of neuraminidase inhibition. It's not that you are a simple farm animal veterinarian. It's that you aren't paying attention to the facts, which don't demonstrate what you say they do. They don't demonstrate anything. They are uninformative even though you insist on drawing conclusions from them.

Thin: Relenza is only effective to where it can get to (i.e., locally). It is not absorbed. If you can inhale it to where it needs to go it seems to work well. The Relenza vs. Tamiflu issue isn't what I was posting about. They are very similar chemically. Tamiflu is a little bulkier so one of the mutations that interferes with Tamiflu might not affect Relenza but pretty much they would be more or less the same when it comes to efficacy and resistance it looks to me.

Revere. Thanks. We will see.

revere, thank you for explaining so well not just the science, but the approach to scientific data.

We really don't know the significance of N294S, since the only time this mutation was reported was in this one case in Vietnam, where a) the patient recovered, b) the reduction of IC50 was slight, c) it was a mixed isolate and some clones had the much more potent H274Y mutation. Nor has there been any studies on this particular mutation (at least not any that I could find. Anybody who finds something do please let me know, thanks!)

So the only 'case' for tamiflu resistance in Egypt is based on the findings of the Vietnamese case, the clinical significance of N294S in that one being far from established.

I went on a rather circuitous route before I came to this observation. As usual, you are already there, and your level-headedness proves once again to be the way to go!

Ohhh, ok. Got it on a re-read.

Thanks guys.

By SmellyTerror (not verified) on 20 Jan 2007 #permalink

Thanks Revere. Looking at these recent cases in Indonesia with the understanding that there are 500K cases of pediatric pneumonia per year in that country from all causes tends to make things go back into perspective. I am curious about the 25% ILI figure you gave for US during flu season, however. You are saying that 1/4 of the population gets the flu each year, right? But then what do the CDC figures mean when they say that 2.1% of all people turning up for outpatient services per week are there for flu? First of all, there can't be all that high a percentage of the population turning up for outpatient help at any one time on the average, or our hospitals, doctors offices and urgent care clinics would have people in lines down the block. It certainly couldn't be more than 1% of the population at any given time, could it? and this is for ALL manner of things requiring medical care: broken bones, cuts, asthma, etc etc. So then, 2.1% (baseline...ranging up to a high of about 7%) of that 1% of the population is there for flu. That would be 0.021% of the population per week. I don't see how that adds up to 25% of the entire US population during the 4 months of flu season. Could you clarify, please?

By mary in hawaii (not verified) on 20 Jan 2007 #permalink

Jon Schultz

Also off topic. Can't help myself. Have to say hi to someone who can quote my favorite Alan Watts book so well...

Thinlina, Revere

Keeping it really simple. Take these two quotes from this NY Times article.

Then tell me why the massive bias towards Tamiflu.

( Especially when IV relenza is being worked on )

''A strain of avian flu that is resistant to the antiviral drug oseltamivir has been isolated from two family members in Egypt, the World Health Organization said yesterday.''

''The oseltamivir-resistant strain in Egypt was susceptible to zanamivir, which is sold as Relenza''


Think of the better systemic attack of IV relenza.

Also remember that Relenza as it is inhaled would attack the upper tract. The odds are that a easily transmitted H2H ''Bird Flu'' would originate from the upper tract. As was seen from the spanish flu of 1918.

Biota the creator of zanamivir is suing GSK for Over half a billion dollars for their lack of effort on promoting and developing zanamivir.

At least Biota and a japanese biotech are developing super relenza or LANI that will be up to 1000 times more effective than Relenza.

Re John Schultz's banned post: I have to say I've *almost* been moved to post on a related issue - that is, the members of the, um, pandemic community who seem to revel in the drama, who seem to have some ghoulish half-hope that the news of the day really is the harbinger of the apocalypse. You see the same people clearly reading too much into news pretty much all the time.

I don't mean to overstate this: it's an attraction to the exciting and the dramatic that most humans feel, even when that event is terrible. Disasters and tragedies can also be inspiring and empowering. I just like to think that people would have enough self-awareness to filter out their own excesses.

Calm and rational assessment may not be exciting, but it's a lot more useful when you're dealing with necessarily scarce input. The public is already reaching drama-exhaustion. Hyperbole is just self-defeating - and a little odious.

One reason I like the Revere(s) so much is an excellent and consistent adherence to scientific rigour. They don't get carried away with the drama, and that makes any pronouncement from them vastly more credible.

By SmellyTerror (not verified) on 20 Jan 2007 #permalink

MiH: No, you don't have it right, yet. 25% have ILI. ILI is influenza-like illness, not influenza. Most of it isn't influenza, except during outbreaks. It is a whole collection of respiratory illnesses from influenza, respiratory syncitial virus, adenovirus, coxsackie virus, etc. It is defined by fever above a certain point (101.5 degrees F.) amd one or more of cough, sorethroat, headache or muscle aches without another obvious cause (like strep throat, for example). Thus ILI and influenza are not the same thing. ILI is a clinical diagnosis, not a laboratory one. The positive predictive value of this symptom complex increases dramatically with high prevalence in the population (this is a simple consequence of Bayes Theorem), so that during an outbreak it has been estimated that 80% o ILI is influenza. At other times the figure is much smaller. A recent paper on ILI in Vietnam estimated only 2.5% was flu, overall, although this seems abnormally low to me. We don't have good figures for it, especially in subtropical climiates like Indonesia.

cpg: The Egyptian virus is not Tamiflu resistant. It has a mutation relaed to decreased neuramiidase activity which is not the same thing. The newspaper articles don't make this clear. See Le's article in Nature, Figure 1. That's an aside. Tamiflu is more important because it is much easier to administer (when inhaled, Relenza must get down to wherever the virus is, not just "breathed in" a little, so making sure the persopn got a dose is much harder) and for children and anyone in respiratory distress it is a more difficult drug to use. IV Relenza could be a valuable adjunct in therapy, but it requires medical help to administer. Taking a pill is much easier. In addition, the maker of Relenza, GSK, failed to market it for years and was sued for that reason by its developer, an Australian biotech company. Roche has been more aggressive in marketing Tamiflu, which is also easier to stockpile. There may be some differences in resistance development (Tamiflu is bulkier) but these two drugs are almost the same, chemically. The newer drug, peramivir (made by BioChryst), not yet available in the US, is quite different chemically and seems effective in trials and laboratory tests with seasonal flu. So that seems a new, viable alternative. But you can't make stockpiles of any of this quickly and Roche's product was better marketed and is more convenient. None of these drugs will stop a pandemic or make much difference in population mortality, IMO (hat's the usual case with individually oriented interventions), but can make a big difference in the therapy of individual patients.

Thanks, Susan, although in retrospect I think maybe "Snicklefritz" at FluTrackers was right that Watts would not have approved of my quoting him in the manner I did.

It's foolish to cast pearls before swine.

would it be possible to melt the existing Relenza
and give it IV , if necessary ?

anon: An IV form exists. It needs to be tested for safety and efficacy. Obviously this is only for treatment.

smelly terror: I don't know that the pandemic community is reveling in a ghoulish half hope that the news of the day is the harbinger of the apocalypse, as you claim. If we "read too much into the news" it is perhaps because there have been many many instances - not just with bird flu but with all sorts of world events - where the government controlled media has withheld vital information and/or twisted information to their own purposes (read "lied.")

Trust is low, and this particular situation is very very important. Should the H5N1 virus at its current virulence become a pandemic, it would be indeed be a world wide catastrophe such as mankind has never faced before. This can't be overstated. Since the government and the news media cannot be trusted to keep us legitimately informed, we are trying to keep ourselves informed. Since so little is known about the virus itself, how it mutates, what key triggers are needed to cause a pandemic etc, we try to "read the signs". Why? Survival, pure and simple. Not a desire for the drama of a catastrophe, a desire to survive it.

If you and Shultz want to make this into something negative, then you are mirroring yourselves. If you want to take the cynical view that the threat is overblown and unlikely, then feel free. But "reading all the signs" I think the threat is persisting and growing more probable, not going away. I want my family, my friends, my neighbors and my students to be able to survive this if it comes. I want to be able to warn them if it appears imminent, in time for them to prepare.

The adherence to scientific rigour you tout is very admirable, but one has to be careful in what science one places one's faith. If PCR tests on a person are negative 3 tests in a row, one with faith in those tests goes "great, no bird flu." But if one explores the data, one might find some serious outpoints with those tests. First, the sample of Viral RNA is very easily degraded. Taken early upon infection, there might not be enough of it to survive the testing protocols, and thus the first test could be negative. Fine, so after they get those results back they do a second PCR test. However in the meantime they immediate began treating the suspect victim with twice daily doses of tamiflu, preventing the virus in his blood from replicating further. So the second test, done after several days of treatment, also comes back negative several more days into treatment. So they do a third and final blood test, taking a sample from a patient who has now been treated for 10 or more days with tamiflu, and no surprise the final PCR test also comes back negative. (The surprise would have been if it came back positive!)

Scientific rigour says believe the three PCR tests. Ignore the fact that the patient had been exposed to poultry verified to have H5N1, or to other patients verified or suspected to have H5N1. Ignore the fact that they had to be put on a respirator to breath, their lungs were so congested. Believe the tests. Believe! It's science. Well, what is the clinical study then, what are the doctor's opinions and observations, what are the epidemiolgic data? Not science?

By mary in hawaii (not verified) on 21 Jan 2007 #permalink

I would like to hear your opinion about the Money Masters Video. What if this all is a "Wag he Dog" kind of a media hype just to collect the money?

Mary: Smelly and I were making different points. If you read my post at FluTrackers you'll see there's nothing in it that contradicts what you're saying.

Mary: Similarly, there's not a lot I disagree with in your post, either - at least, not the things you seem to think I disagree with. You've misunderstood.

I absolutely do *not* think the threat is overblown. That's not what I said at all. The threat is dire. This or some other pandemic is, in fact, inevitable. But I *do* think that some people frequently overblow the evidence we have at hand.

Don't people get sick of saying "this is it!" every time? We all know that innocuous data will probably be the first sign of a significant oubreak, but unless you have a good reason to think that this time is different, there's not much point jumping to conclusions. It doesn't add to our knowledge or response, just to the drama.

...and I think it's interesting you say "we" when I was, as I said, talking about a only one part of the community (one that is very much *not* - for example - represented by the Reveres). Do you consider yourself to be one who reads too much into the news? Why the big reaction? Why did you recognise yourself in my description?

Scientific rigour absolutely does *not* mean ignoring evidence just because it comes from outside a lab. If that's what you think science is, well, you're wrong. But scientific rigour also means you avoid getting swept up in the drama of the event. It means you DON'T go with instinct, or continually make bold predictions that are consistently wrong!

I can wake up every morning and say "today there will be an earthquake. Everyone panic!" One day I'll even be right. But that's not science, and it's not helpful.

By SmellyTerror (not verified) on 21 Jan 2007 #permalink

Just thinking out loud, this new tamiflu resistance mutation seen in Egypt (N294S) seems like a very 'smart' mutation. Unlike the H274Y mutation which seems to develop in Tamiflu treated patients and results in an apparently unfit virus, this one may be 'just right', thwarting our attempts at treatment but maintaining the in vivo function of the NA. I wonder if this could have developed by exposure of the virus to low levels of tamiflu possibly used in poultry or people.

kent: I don't think there is evidence it affects treatment much at this point. But there is a very interesting article by Marc Lipsitch I will post on one of these days that suggests that even very rare mutations where the virus is resistant can quickly become the dominant strain.