If the story was told in logical order, we would begin with sialic acid, and end with public health policy.  We choose instead an approach based on how a potential threat focuses interest, with increasing demand for detail as the threat evolves. So we begin with a synoptic view of the major questions.

Is a pandemic inevitable?  The evidence is highly suggestive of a yes. Despite all the advances in computational biology, the question cannot be answered like a mathematical proof. There are two types of suggestive evidence:

• Lab studies which have modified the genomes of H5N1 and similar viruses, pinpointing changes which enhance lethality and transmission in specific animal models. The  transfer of the result to a conclusion about the human adapted virus could be  frustrated by a fitness defect, which would nullify the proposed adaptation. Don’t bank on it.
• Widespread adaptation of H5N1 to  mammals has already occurred.  This implies that the genetic instability of H5N1, characteristic of an RNA virus, is actually engaged in Darwinian evolution on a massive scale. Why should  humans be the exception? Don’t bank on it.

Over the past 20 years, bird flu has killed half of the known human cases, which have been clustered in China and southeast Asia. Is this the kind of mortality we should expect?

Paracelsus, the first toxicologist, said “The dose makes the poison”.  He observed that how poisonous a substance is depends upon the quantity. Viruses are pathogens, which resemble poisons with the added ability to replicate within the body. To some extent, the quote of Paracelsus applies to pathogens. While some viruses are infective in near zero numbers, most infections require a minimum number of virus particles to cause disease, while severity often depends on quantity.

Unlike poisons, which are substances lacking the power of replication,  viral load grows roughly exponentially from the initial exposure, but only in the absence of immune system response. This is never the case. The immune system has two major logical divisions, the adaptive, and the innate. While the adaptive response takes days, weeks, or longer to respond,  the innate immune system responds immediately, reducing the growth of viral load, possibly allowing the individual to survive until the adaptive system kicks in.

This is why “The dose makes the poison” has relevance.  Asian cases likely involved the consumption or preparation of infected poultry. This may constitute much higher exposure  than human-to-human droplets and aerosols, which may imply much lower human mortality in  a pandemic. The 50% could also be confounded if H5N1 is common in these regions as a mild, under-diagnosed disease. Don’t count on it.

The complexity of the question is highlighted  on a Texas dairy farm. (CBS) Cats died after drinking raw milk from bird flu-infected cows.  The cows were sickened by H5N1, but most did not become critically ill. The farm had 24 cats, which were fed unpasteurized milk from the cows. Half of the cats died within several days. Despite all the analytic tools of modern biology, it is not feasible to answer, in the short term, whether humans resemble cats or cows.

Has surveillance been adequate? With so many mammalian vectors, it’s an  impossible problem. Cows eat hay, potentially contaminated by rodents, which caught it from birds…feel free to explore  the endless permutations.

Are pandemic preparations inadequate? (CNN) We aren’t doing enough about the risk of bird flu – but we can. Quoting,

Even today, the United States doesn’t know the extent of spread among animals or humans due to insufficient testing and tracking. The 7-1-7 target for outbreak detection and response should be our guiding principle…My organization’s latest Epidemics That Didn’t Happen report highlights how countries with responsive health systems prevent outbreaks from becoming epidemics through careful planning, early detection, and swift action.

I am extremely skeptical. Rather than niggle each point, note that:

• Excepting Finland’s H5N1, the diseases mentioned in Epidemics That Didn’t Happen have not manifested, in the past 50  years, as more than regional epidemics. They lack the adversarial capabilities of flu.
• Quoting,  “If one country responds to H5N1 well, that’s not enough. Microbes know no borders.” This is true, but human stupidity has unfathomable depth.
• H5N1 is now a worldwide, pan-animal-class zoonosis. It could jump to humans in myriad ways beyond any realistic, achievable monitoring.
• No influenza pandemic has ever been  stopped.

Can we be immunized now? Original antigenic sin is an obstacle for flu immunization in absence of the actual strain.  It makes this strategy questionable. Curiously, COVID seems to be unaffected by this phenomenon.

My recommendations:

• Quadruple cell-based vaccine manufacturing capacity. 140 million doses in six months is woefully inadequate. Diversify suppliers and technologies. Abandon egg-based production.
  
• Re-examine every aspect of public health messaging, which in the case of COVID resulted in a wide swath of irrational behavior.
• Hope for the best, but prepare for the worst.