Everyone seems to have an opinion about whether bird flu will be the next terrible global pandemic. In current parlance “bird flu” means human infection with the highly pathogenic avian influenza/A subtype H5N1. There is no doubt that this is the 800 pound gorilla in the global health room at the moment, but not because it is more likely to become a pandemic (NB: pandemic by definition is a globally dispersed sudden increase in infection among humans; the same situation for animals is called a panzootic, and it is plausible to say we have an H5N1 panzootic for birds now). On the basis of biology humans are potentially susceptible to influenza viruses, although only the H1, H2 and H3 subtypes circulate or have circulated in human populations. But a handful of human infections have also been reported from H5, H7, H9 and H10 subtypes, although except for H5 the infections have been relatively mild. Out of 103 cases of the non-H5 subtypes there has been only one fatality (H7N7, one fatality with 88 non-fatal cases), whereas the current official tally for H5N1 gives 213 deaths in 346 cases (61%). Which is why the H5N1 gorilla weighs 800 pounds. This influenza/A variant is a remarkably virulent virus. Is it the most likely virus to cause the next influenza pandemic?
Recall the terminology here, as it makes distinctions that are important. A virus is pathogenic for a host if it is capable of infecting the host cell and cause disease in the host. So there are two criteria for pathogenicity: it can infect the cell; and the host organism gets a disease. This last criterion is vague. The “disease” state might range from very mild or hardly noticeable to fatal. The underlying idea is that the infection was in some sense parasitic, i.e., the virus reproduced itself with the host cell machinery at the expense of the host. That cost might be tiny or catastrophically large, but it wasn’t neutral or beneficial (symbiotic). The degree to which the cost of the typical infection is on the serious side is the virulence of the virus. Virulent viruses are by definition pathogenic, but in addition they cause serious disease. By this terminology the H7, H9 and H10 are human pathogenic viruses that aren’t very virulent (the most frequent disease symptom was conjunctivitis, “pink eye,” in the Netherlands H7N7 outbreak of 89 cases). There is a bit of special usage in the flu world here, which we should also clear up. The avian version of H5N1 exists in two main forms that differ in virulence, called confusingly low pathogenic and high pathogenic. To be consistent with the terminology used in the rest of infectious disease epidemiology they are both pathogenic viruses, one of low virulence, the other of high virulence. The standard test for high path was to see if it killed chicks. These days it is more common to look at a genetic feature, the presence of extra basic amino acids at the cleavage site of the hemagglutinin protein.
What bearing do these distinctions have on which of those influenza subtype is likely to become the next pandemic strain? None, really. They are all subtypes to which humans have had little prior exposure so there is no pre-existing immunity to the main antigen (the one designated “H”). It could also be a subtype not listed as having caused human infections, like an H6 or an H13. The thing that will decide whether a new subtype becomes pandemic, in addition, is whether it is easily passed from person to person, as is “ordinary” seasonal influenza (currently certain H2 H1 and H3 subtypes). This is still given by yet another term, transmissibility. On that list of non-H1 to H3 subtypes pathogenic for humans the most transmissible was H7N7, so on those grounds one might expect this to be the next pandemic. Why aren’t we more worried about it? Because a pandemic of “pink eye” is not the same thing as a pandemic that has a 60%+ case fatality ratio (CFR), the way H5N1 does. For comparison, the horrendous 1918 flu had an estimated CFR of under 3%.
All of these influenza subtypes mutate with ease. Most mutations are bad for the virus (in the sense that the mutated virus replicates less well than the unmutated version). So ease of mutation isn’t going to tell us which is the next pandemic subtype. We also don’t know the features that make an influenza virus easily transmissible. There are likely multiple combinations of features that can do this (i.e., many roads to the same end), but we don’t know what they are. We don’t know how these subtypes interact when they circulate simultaneously. For example, if another N1 is circulating, does this make it less likely a second N1 virus coupled with a different H will also gain a foothold? How likely are they to infect a host simultaneously? As if these questions weren’t themselves of key importance, we also don’t know the relationship between transmissibility and virulence. They are logically and biologically independent in general, but changes that affect transmissibility might also affect virulence. Because we don’t know any of these things (and lots more besides), it is not possible at the moment to quantify the probability of one or another (or any) subtype to “go pandemic.” Experts can make judgments (“subjective probability”) based on evidence, but at the moment the evidence isn’t strong enough to narrow the huge range of plausible judgments. The best we can do is make some assumptions, for planning purposes. Those assumptions are the source of most of the contentious arguments over the adequacy or lack thereof about pandemic preparedness.
Because perseveration is a characteristic and privilege of the aged, we will repeat again in this last bird flu post of 2007 what we have been saying here since late 2004. The best way to prepare for an influenza pandemic is to do those things which make for a robust community, especially building and strengthening the public health and social service infrastructure. This is like repairing the roof on your dwelling. It will help protect against heat, cold, rain, sleet or snow. It won’t keep you safe from an asteroid or a nuclear attack. A strong public health system also won’t protect you from a pandemic with a 30% attack rate and 60% CFR. But it will help with a hell of a lot of other things, including many of the most plausible candidate influenza pandemic viruses.