1. The Virus We Were Watching
The Hong Kong H5N1 outbreak of 1997 was an important warning.
An avian influenza virus had demonstrated that it could infect humans directly.[5] The immediate concern was not simply that H5N1 could make an individual seriously ill. It was what might happen if such a virus acquired the ability to spread efficiently from one human to another.
That ability did not emerge. But H5N1 did not disappear either.
Over the following decades, H5 viruses continued circulating and evolving in birds, with intermittent infections in humans and other animals.
The warning light that came on in 1997 never completely went off.
2. The Virus That Comes in Eight Pieces
Influenza A has an unusual genetic architecture. Its genome consists of eight separate RNA segments.
This matters because if two different influenza A viruses infect the same cell, newly produced viruses can sometimes contain a mixture of genome segments from both parental viruses. That is reassortment.
It differs from the smaller genetic changes that accumulate as influenza replicates — antigenic drift. Reassortment can create a much larger genetic change in a single event.
Pigs have traditionally been described as influenza ‘mixing vessels’ because they can be infected by swine, avian and human influenza viruses. They remain important, but the biology is broader. Reassortment can occur in other susceptible hosts whenever compatible influenza viruses meet.
Wild birds, migratory birds, poultry, pigs, other mammals and humans therefore form an interconnected influenza ecology.
Occasionally, two viruses meet in the same host and the same cell. Influenza's eight-piece genome then gives evolution something resembling a genetic pack of cards to shuffle.
3. Most Experiments of Nature Fail
Reassortment is not synonymous with pandemic influenza.
Most new combinations do not acquire the characteristics necessary to become successful human viruses. Others may circulate in birds or mammals without ever becoming important human pathogens.
Even infecting a human is not enough. A potential pandemic influenza virus must overcome a succession of barriers: infect humans, replicate efficiently, transmit effectively from one person to another and encounter a population with insufficient immunity to stop it.
Occasionally, however, reassortment matters enormously. The 1957 H2N2 Asian influenza and 1968 H3N2 Hong Kong influenza pandemics each followed a major change in the virus's surface antigens. In 1957, both the haemagglutinin and the neuraminidase were new to humans. In 1968, the virus differed from its predecessor by its haemagglutinin but retained the N2 neuraminidase. Evidence suggests that pandemics marked by a change in haemagglutinin subtype arise from genetic reassortment with animal influenza A viruses.[6]
Reassortment therefore creates possibilities. Natural selection decides which possibilities survive.
4. We Were Watching H5N1. H1N1 Arrived.
For much of the 2000s, H5N1 dominated discussion about the next influenza pandemic.
Then, in April 2009, a novel H1N1 influenza virus appeared. Genetic analysis showed that it was a reassortant of influenza viruses circulating in North American and Eurasian pigs. Its eight segments carried a complicated evolutionary history ultimately involving swine, avian and human influenza lineages.[7]
But one characteristic mattered above all others. It could spread efficiently between humans. On 11 June 2009, WHO raised its pandemic alert level to phase 6.[8]
My Vads Corner H1N1 page went online on 27 April 2009 as the outbreak unfolded.[9]
The first influenza pandemic of the twenty-first century had arrived. It simply was not the influenza pandemic we had been expecting.
5. We Were Watching Influenza. A Coronavirus Arrived.
There was another lesson.
In March 2003, I began following a mysterious outbreak of severe respiratory disease that became known as SARS.[10] My archived SARS page still preserves something easily forgotten with hindsight: at the beginning of an outbreak, even the identity of the pathogen may be uncertain.
SARS was eventually shown to be caused by a previously unknown coronavirus.[11] It was controlled: WHO declared the global outbreak contained on 5 July 2003.[12]
Then, in late 2019, another coronavirus appeared. SARS-CoV-2 achieved what the original SARS coronavirus had not: sustained global human transmission. COVID-19 became the defining pandemic of our generation. On 11 March 2020, WHO characterised it as a pandemic, noting that it was the first caused by a coronavirus.[14] My COVID-19 archive followed that pandemic as it unfolded.[13]
We had spent years worrying about pandemic influenza. The pandemic that transformed the world was caused by something else.
That does not make influenza surveillance unnecessary. It teaches something more uncomfortable:
We may recognise threats without being able to predict which threat will succeed.
6. So Why Are We Still Watching H5N1?
Because H5N1 has not stood still. H5 viruses have spread geographically through wild birds and poultry and crossed repeatedly into mammals.
Yet one crucial distinction remains. Current WHO assessments state that H5N1 does not appear to transmit easily from person to person and that sustained human-to-human transmission has not been reported.[15]
An infected bird is not a pandemic. An infected mammal is not a pandemic. Even an isolated human infection is not a pandemic.
What changes the equation is a virus acquiring the ability to transmit efficiently and sustainably between humans. That remains the barrier we watch most closely.
7. Something Is Mixing in Asia
An October 2026 paper in Emerging Infectious Diseases brings attention back to South and Southeast Asia.[16]
Much international attention has understandably focused on the globally successful H5N1 clade 2.3.4.4b. But older H5N1 lineages have continued circulating in parts of Asia. And they have not remained genetically isolated.
Recent human H5N1 infections in Cambodia have involved reassortant viruses containing gene segments derived from long-established regional H5 viruses, newer 2.3.4.4b viruses and low-pathogenicity avian influenza viruses. Related reassortant viral backgrounds have also been associated with recent human infections in India and Bangladesh. In Cambodia, one such reassortant genotype has become dominant in poultry.[16]
This is important. But it requires careful interpretation.
The association between these reassortant viruses and renewed human infections does not prove that reassortment made the viruses more capable of infecting humans. Stronger surveillance since the COVID-19 pandemic might also partly explain why more human infections are now being detected. The authors emphasise the need for phenotypic studies to determine what these new genetic combinations actually do.[16]
A new genome is a signal to investigate — not a prediction of a pandemic.
8. H5N1 Is a Name, Not a Genome
When we hear ‘H5N1’, it sounds as though we are talking about one virus. We are not.
The H and N refer to two proteins on the viral surface: haemagglutinin and neuraminidase. Beneath them are six additional genome segments.
Two viruses can therefore both be called H5N1 while carrying importantly different internal genetic combinations.
Modern surveillance must consequently ask more than: Is this H5N1? It must ask: Which lineage? Which genotype? Where did each of the eight segments come from? What mutations have appeared? What does the virus actually do? And, most importantly: What is it doing epidemiologically in the real world?
9. What Would Actually Worry Me?
Not every report of H5N1 should trigger another prediction of an impending pandemic.
I would be much more concerned if several signals began appearing together: clusters of human infections without convincing animal exposure; increasing infections among mammals with evidence of onward transmission; genetic changes associated with mammalian adaptation; laboratory evidence of altered receptor binding, replication or transmission; significant antigenic change; and above all, sustained person-to-person transmission.
Modern genomics can tell us that a virus has changed. Laboratory experiments can tell us what some of those changes might do. Epidemiology tells us whether they actually matter.
The strongest warning comes when all three point in the same direction.
10. From Microscope to Genome
This is perhaps the greatest difference between 1997 and 2026.
When I first wrote about H5N1, surveillance depended heavily on clinical recognition, epidemiology, virus isolation, serology and much more limited genetic analysis.
Today, whole-genome sequencing can read all eight influenza segments. Phylogenetic analysis can reconstruct viral ancestry. Viruses collected thousands of kilometres apart can be compared. Reassortment events can be identified, and changes associated with adaptation can be investigated.
What once might simply have been called another H5N1 virus can now be recognised as a new genetic constellation assembled from different ancestors.
That is an extraordinary scientific advance. But it creates a temptation. Because we can see evolution in unprecedented detail, we may begin to believe we can predict where evolution is going. We cannot.
Seeing evolution is not the same as predicting its destination.
11. Malaysia — Are We Looking in the Right Places?
Malaysia has lived with this problem before.
In August 2004, H5N1 was confirmed in Malaysian poultry. At the time, I wrote that good surveillance and public-health preparedness were paramount.[3]
Malaysia maintains a formal veterinary framework for avian influenza. The Department of Veterinary Services' current Veterinary Protocol Malaysia — Avian Influenza, revised in 2024, sets out surveillance, diagnosis, reporting, control and eradication measures for the disease.[17]
There is also a swine dimension. A Malaysian study sampled 727 pigs from farms in six states in Peninsular Malaysia in 2005. Antibodies to H1N1 and H3N2 influenza A viruses were found in approximately 12% of pigs for each subtype. Importantly, however, neither influenza virus nor viral nucleic acid was detected in nasal samples.[18]
The study does not tell us what is circulating in Malaysian pigs today. It does remind us why pigs belong within influenza surveillance.
On the human side, the Institute for Medical Research serves as Malaysia's WHO National Influenza Centre.[19]
The modern question is therefore not simply whether individual sectors conduct surveillance. It is whether the pieces connect quickly enough.
If an unusual influenza virus appears in a wild bird, poultry flock, pig farm or hospital patient, can information move rapidly between veterinary laboratories, wildlife authorities, public-health laboratories, clinicians and epidemiologists? Can animal and human viral genomes be compared quickly enough to recognise something unusual before a pattern becomes obvious clinically?
I have not found enough current public information to say confidently how completely that genomic integration presently occurs in Malaysia. And perhaps that itself is a question worth asking.
That is where One Health stops being a slogan. The October 2026 Emerging Infectious Diseases paper argues for surveillance that connects animal, environmental and human data streams in near real time, with two-way genomic data exchange between animal and public-health authorities. Without such systems, it warns, reassortant events will continue to be detected retrospectively or with substantial delay.[16]
Viruses do not recognise the boundary between veterinary medicine and human medicine. Neither should surveillance.
12. What Can We Do as Individuals?
We cannot prevent influenza viruses from mutating or reassorting in animals. But that does not mean there is nothing we can do.
Keep seasonal influenza vaccination up to date according to current recommendations. It does not protect against H5N1. But CDC advises it for people who work with, or are frequently exposed to, infected or potentially infected animals, in part because it may in theory reduce the already rare possibility of simultaneous infection with human and animal influenza viruses — and therefore the opportunity for reassortment.[22]
Avoid unnecessary contact with sick or dead wild birds, poultry or mammals. Report unusual animal deaths to the appropriate veterinary or wildlife authorities rather than handling the animals yourself.
People whose work brings them into contact with potentially infected animals require appropriate personal protective equipment and occupational precautions.[20]
Food hygiene matters too. Poultry and eggs should be properly cooked, and raw or inadequately cooked animal products should be avoided where relevant outbreaks are occurring. For milk, WHO advises pasteurised milk; if that is not available, heating raw milk until it boils makes it safer.[21]
And tell your doctor about animal exposure.
A fever and cough usually mean something commonplace. A fever and cough after handling sick poultry may mean something quite different. The doctor cannot make that connection if the patient does not mention it.
If an unusual influenza virus eventually begins spreading efficiently between people, the familiar public-health measures return: staying home when ill, respiratory hygiene, good ventilation, appropriate masking when circumstances warrant it, testing and following public-health advice.
None of this is dramatic. That is precisely the point.
Pandemic preparedness does not begin when WHO declares a pandemic. It begins with ordinary surveillance, ordinary precautions and recognising when something is no longer ordinary.
Be aware. Be sensible. But do not be afraid.
13. What Nearly Thirty Years Have Taught Me
Looking back through the old Vads Corner and Cybermed pages is instructive.
1997: H5N1 taught us that an avian influenza virus could infect humans directly. 2003: SARS taught us that a new respiratory pathogen could emerge unexpectedly. 2009: H1N1 taught us that even when the next pandemic was influenza, it might not be the influenza virus we had spent years watching. 2020: COVID-19 taught us that the next great respiratory pandemic might not be influenza at all. 2026: genomic surveillance allows us to watch viral evolution and reassortment with a precision that would have been extraordinary in 1997.
Yet we still cannot reliably predict which viral experiment will succeed.
Perhaps we have been asking the wrong question. Instead of: ‘Which virus will cause the next pandemic?’ we should ask: ‘Will we recognise it quickly enough when it begins to acquire the characteristics that matter?’
That is a question surveillance can actually help answer.
14. The Next Pandemic May Already Be Mixing
Somewhere today, influenza viruses are replicating in wild birds, poultry, pigs and other animals.
Some will mutate. Some will reassort. Most will go nowhere. A few will cross species. Fewer still will infect humans.
One day, one virus may acquire the combination of biological characteristics and ecological opportunity necessary for sustained human transmission.
It may be H5N1. It may be another influenza virus entirely. The next pandemic may even come from another viral family.
That uncertainty is not an argument for complacency. It is the reason preparedness matters.
When I first put the H5N1 page on Vads Corner in December 1997, we knew enough to be concerned but had relatively limited tools with which to watch the virus.
Today we can read all eight segments of its genome, reconstruct its ancestry and sometimes watch reassortment almost as it happens.
Twenty-nine years of technology have made us much better at watching influenza. They have not made influenza predictable.
Perhaps that is the lesson. We do not need to predict the next pandemic perfectly. We need to be ready when the virus reveals itself.