Every flu season is a soup of viruses

It may surprise some to know that influenza isn’t caused by just one “flu virus”. There are multiple flu viruses which can all cause the flu and they are around each season. Every flu season is a soup of viruses.

We don’t host a pure colony or a clone of “a virus”. Our infectious passenger can replicate into a veritable soup of subtly different variants of that initial invader.
Photo by Navada Ra from Pexels

Viruses don’t hold meetings to decide who gets to be ‘the one”. Instead – in all their genetic variability and constant mutational glory – they come at us all year long and the winners are the ones with a genetic balance that best positions them to dodge our immunity, to last on surfaces and to transmit effectively. Grab a coffee because this is a long one…

Which Flu Will It Be This Year?

Recent mainstream media reports suggest the #flunami visited upon Australia since November 2018 may be due to the “3 flu strains” moving among us at the same time (co-circulating).

Comments like “unusually, both A strains A/H1N1 and A/H3N2 are being passed around” imply that it’s unusual for there to be more than one flu virus around at a time.

But those comments are all off the mark. Let’s look at why co-circulation of multiple flu viruses actually isn’t unusual, why the entire concept may underestimate how diverse any flu season is and why we still don’t know what’s driven our unprecedented Australian summer and autumn of flu.

Three Flu Friends For 2019?

The national flu report makes it look like we have 3 different viruses as shown in the graph below. And we kinda do.

A very basic snapshot of the “3 strains” of flu that are co-circulating in Australia this year. Each bar adds up to 100% of all the flu viruses testing positive in that week, nationwide. Is this number of flu viruses unique to 2019? Is it an accurate description of the number of viruses? Nope. Keep reading to find out why.

There is a flu type B virus which is infecting more people as the season goes on. These are cases identified in the green portion of each weekly bar above (more on flu B below).

There are lots of flu type A viruses in that chart that weren‘t subtyped (red bars), so we can’t say much more about those. We just assume they will break down in the same percentages as the flu A viruses that have been subtyped. And, ya know, cost.

Among the subtyped flu A viruses in that graph, there is a gradually shrinking amount of flu A/H1N1 (orange portion) and a sometimes steady/sometimes growing amount of A/H3N2 (yellow portion of each bar).

So we could say that there are 3 flu strains that can be grouped into a flu B and two subtypes of flu A.

But that is pretty basic. Let’s delve a little deeper.

Getting Nerdy With It

First up – “a virus” is a minimising description. That’s usually fine but sometimes we need to know more. An infected person is actually harbouring lots of subtle variations of “a virus”. They host an evolving soup of these virus variants all replicating to create the best version of themselves for the environment they find themselves in. The person may even have been infected by more than one distinct flu virus when they initially wore that face full of sneeze.

After the soup of viruses has been created in the infected person, the most victorious viruses get coughed out; a different soup to the one that went in.

What comes out can differ from what goes in

Virus-laden projectiles may contain only slightly mutated variants which are essentially what went in with some fine-tuning to help them avoid their last host’s antibody responses. The viruses that get picked up by the next sucker might also be those that were a bit more hardy and a bit better suited to the right environmental conditions. All of these fine-tuning evolutionary changes happen through mutation.

A virus factory functioining within specifications.
Image by Luisella Planeta Leoni from Pixabay

Alternatively, the coughed output could be very different – particularly if two distinct flu strains infected the same person. In that case, they could have mixed and matched their many gene segments to produce a child virus that is neither one of the parent viruses.

Flu clubs

The “virus” called A/H3N2 seems to be especially good at using us to mutate and to transmit. It’s been especially busy since 2012, but first appeared in a 1968 pandemic.

A/H3N2 viruses have settled out into a handful of groups which can be seen when we identify and compare the genetic sequence differences between viruses from different A/H3N2-infected people.

Taking a look at this variation using a spectacularly excellent website called nexstrain, we see those groups. Each group is comprised of A/H3N2 viruses that, as a group, distinctly differ from every other group.

A/H3N2 from down under

No – I don’t mean an “Aussie flu virus” 🙄 The likelihood that the “influenza strain that wreaked havoc in the northern hemisphere and Eastern States two years ago has hit WA ” is in fact the same virus, is vanishingly small. But science could tells us this for sure. Science is just…shy?

But I digress.

Each A/H3N2 group below is joined by lines that can be traced back to a common branch on a computer-generated, genetic sequence tree like the one below. All the virus group members that can be traced back to one common branch are said to represent a clade.

Humans like to put names and boxes around “things” so they are easier to visualise and talk about using normal English. Nature doesn’t care, but our brains do like the structure.

A/H3N2 viruses can be grouped into at least 5 genetically distinct clades using just the sequence of the hemagglutinin gene (H; see a quick overview of flu gene segments).

Genetic sequences of the hemagglutinin gene (H) from some of the Australian flu viruses circulating between November 2018 and April 2019. Each individual virus from an infected Australian case is marked with a highlighted coloured dot on the right-hand side of the tree. The most recent ones are the nearest to the right edge (see the date along the bottom). Most of 2019’s A/H3N2 viruses belong to a clade labelled 3c2.A1b/131K (red dashed oval). A tree like this presents the similarities and differences in their H-gene sequences in a way you can see them. The A/H3N2 vaccine component in this year’s southern hemisphere vaccine, A/Switzerland/8060/2017 (H3N2)-like virus; , is highlighted with a black cross (pink dashed circle) The tree uses a 6-month dataset. Source: nexstrain website.

In Australia, we’ve detected A/H3N2 strains in all clades from November 2018 to April 2019, but mostly from the excitingly named 3c2.A A1b/131K (or just A1b/131K; see the figure legend).

These genetic differences can result in protein differences. The clades don’t always represent this well, but they sometimes do. In particular, changes to the proteins that are recognised as not of us by our immune system; so-called antigens.

Changes in flu virus antigens lead to some of the clades being different enough from each other that no single A/H3N2 vaccine component is enough to protect us from all the clades. Because of this, we see the A/H3N2 vaccine component – indicated as black crosses in the picture above – change sometimes. This change is needed more often for the rapidly mutating A/H3N2 than A/H1N1 or the flu Bs.

We saw a mismatch between virus and vaccine happen in 2017. We may already be seeing some version of that in 2019 because a third (39 of 119) of Australian A/H3N2 viruses studied in 2019 have been classified as “low reactors“. In lab tests, these viruses didn’t react strongly to antibodies from the blood of animals exposed to the A/H3N2 vaccine virus.

Two flu Bs – and they’re both around

So I said earlier that it looked like we had one flu B.


The viruses we call flu B can also be grouped. In their case, the groups are called the Victoria lineage and the Yamagata lineage (think of them as clades; it’s easier). Sometimes both can co-circulate, and again, there can also be distinctiveness within each lineage.

The names Yamagata and Victoria come from the relevant reference viruses geographic place of laboratory characterisation.

Unfortunately few labs report the different lineages. It isn’t hard to do and there are PCR-based rapid methods (example of Yamagata and Victoria methods). The World Health Organization Collaborating Centre (WHO CC) for Reference and Research on Influenza and the Queensland Government Department of Health both report these details and so we can say from their reports that both B/Yamagata and B/Victoria (more dominant of the two) lineages are co-circulating in Australia in 2019.

Analysed type/subtype/lineage of samples tested at the World Health Organization Collaborating Centre (WHO CC) for Reference and Research on Influenza

What Is “A Flu Virus”?

All of this raises an interesting question. What makes one flu virus sufficiently different from another for it to be considered distinct? And do they stay that way or evolve away and become a different virus?

With other viruses that have only one-piece genomes (flu has 8 pieces to its genome), we can define a threshold or number of genetic mutations that contains all the variants that belong to one distinct genotype (genetic type) of virus.

For example, we might say any virus variants which share more than 97% (RPV…Randomly Plucked Value!) genetic identity are all members of the one genotype. Or, the other way round, if a virus variant is more than 3% genetically different, then we’d say it belongs to a different genotype (see the idea of drawn below).

A way to picture how we might define distinct viruses using a percentage cutoff value for genetic difference, or similarity. In this example, there are 3 genotypes (each distinct virus is coloured differently), each defined by a genetic “boundary”. All virus variants sharing that are 97% or more genetically identical are placed within the boundary and called members of the same genotype (“distinct virus”, or some might call it a strain). If variants share less than 97% identity (more than 3% difference), they fall outside that boundary and represent a different genotype. Each genotype is made up of a soup of slightly different variants. If you wanted to, you could imagine that those closest to the centre of each genotype share the highest levels of identity (maybe 99/9%) whole those nearest the boundary share the least (maybe 97.1%). Could we do this for flu viruses?

Does An Old Flu Virus Settle Down?

I think it would be useful to better define individual flu viruses rather than keep on referring to big virus buckets like strains. Especially when what’s being used in the media refers to a subtype. It might drive us to frame some things differently.

For example:

  • Does the host immunity really wane in 6 months, or do we actually mean that the viruses have changed sufficiently in that time that we now have a mismatch? Are the antibodies we produce in response to immunisation by the original vaccine components still protecting us….against those components?
  • Do each of the genetic clades within subtypes represent distinct viruses?
  • Would it be better to use whole genome genetic sequence rather than one segment’s sequence to define a flu virus?
  • Could distinct flu viruses “settle”, continuing to transmit but contributing only small amounts of illness and never again being able to drive a “flu season”?

I don’t have answers to any of those. But they interest me.

Can we really know much about a virus by looking at only 12.5% of its genome? Why are we so old-fashioned about influenza viruses?

Remember that tree earlier? We were looking intently at only an eighth of a flu virus’s genome. What happens in the other 7 genes is also likely to be important in defining both the identity of a flu virus genotype and how it behaves.

We tend to think of viruses as distinct if our immune system recognises the obvious parts of their jacket as such (see the protruding blue and yellow bits in the image above). But a virus like flu can theoretically change some or all of what lies underneath its jacket and it’s possible that we wouldn’t notice unless the virus behaved noticeably differently. What are we missing and are we missing it in 2018/19?

What’s The Virus Count In 2019?

Hopefully by now you can see how difficult it is to answer that question satisfactorily. We know we have:

  • 2 fluB viruses (with an unknown amount of distinctiveness)
  • An A/H1N1 virus (these also exist in clades but seem less diverse than the A/H3N2 virosystem)
  • At least 3 clades of A/H3N2 viruses co-circulating in Australia

Three A/H3N2 viruses is probably an underestimate. But if we have so many flu viruses co-circulating, are we in trouble this year?

2019: Is It Flumageddon?

We’re definitely experiencing an historic number of lab-confirmed flu cases. It’s not obvious that there’s more severe disease associated with those infections, just more of them. More sick people, more coughing in the cinema, fewer4 desks occupied in the office and more beds taken up by more cases of the flu

But it turns out that we see pretty much all the main flu virus players each year. Every flu season is a soup of viruses.

If we click through the gallery below, we can view the four seasons preceding this one. Despite the changing graph formats of national reporting, we have a mix of multiple flu viruses every season.

Every. Season.

I’m fairly sure that would pattern would be maintained if I went back further that 2015. Just a note: the vertical axis is a percentage in 2017 and 2018, so don’t think the number of cases is equal in all years – it isn’t. 2018 was a wimpy year while 2017 was massive one. 2015 was flu B dominated, the others flu A.

It’s hard to be sure exactly how many distinct viruses are around right now because we don’t think about flu viruses in this way and so we don’t look for them this way. Our surveillance system isn’t set up to do this. Neither is our vaccine effectiveness system. I think they should be. At least until we know whether additional detail would be worthwhile or not. We already spend a ton of public dollars on flu testing, we should do an equivalent amount of research. Right now we can say for sure is that we’ll know whether 2019 was flumaggedon…in 2020.

So What Is Driving flunami2019?

It’s got to be the viruses. We need to look at the viruses.

We’ve done enough talking. There’s nothing new about Australia seeing 3 distinct co-circulating viruses in a flu season. It’s not a change in the amount of travel or even travel itself. Travel is a well-known cause of the movement of viruses around the globe (also here and here and many more). It’s probably not the amount of testing that’s driving the flunami either.

Remember, we’re looking for something that’s changed this year.

I think the difference(s) and the answers will be found when we study the nature, sequence, structure and function of the 2018/19 flu viruses.

We need focussed research and the flexible funding to do this sort of public health research. This is a job for virology. Enough guessing. It’s beginning to look like we’ve made asking questions and guessing answers about flu virus behaviour an annual media spectacle.

Why is this season so big?

Why is that season so small?

Why was that season so long?

Why is our season so early/late?

Will your season affect our season?

And so on. Where’s the damn science??

It’s well past time that virology stepped up and got more serious about understanding flu viruses in public health and communicating that knowledge, in real time, to the public. Perhaps it needs some oversight. Or more funding. Or some selective pressure.

Visits: 1865

6 thoughts on “Every flu season is a soup of viruses”

  1. Ian,

    We continue to strongly disagree about the claims of “unusual” severity. I say this as a proponent of annual flu vaccination. Australia has indeed changed its surveillance procedures and has indeed rolled out rapid diagnostic testing at scale. Same thing happened in the US. We are drawing out reporting of a historically grossly underreported disease.

    “We saw a mismatch between virus and vaccine happen in 2017. We may already be seeing some version of that in 2019 because a third (39 of 119) of Australian A/H3N2 viruses studied in 2019 have been classified as “low reactors“. “

    There is a big difference between a true mismatch, meaning the wrong subtype strain was chosen for an upcoming seasonal vaccine versus drift that occurred during replication of the vaccine strain in eggs. Having a third of the viruses classify as “low reactors” is not uncommon with H3N2. If more than 3/4 of their viruses were not matched, then we would be concerned for a true mismatch. Let’s see if that occurs.

    H3N2-dominant seasons are typically associated with higher morbidity, mortality, and infrastructure strain compared to H1N1-dominant seasons. True vaccine mismatch with H3N2 has been associated with high severity. This scenario has happened several times in the last 30 years in Oz.

    James M Wilson V, MD FAAP
    Director, Nevada Medical Intelligence Center
    University of Nevada-Reno

    1. Yes, we disagree if you are saying this season is due to testing. Extra testing doesn’t raise hospital admission levels to those closer to traditional winter seasonal levels. Testing doesn’t change the rate of positivity to levels usually seen during winter seasonal levels. Testing doesn; cause deaths at levels not commonly to inter-seasonal levels. Testing doesn’t cause 20-30% of circulating H3N2 to be low reactors. And testing doesn’t move all this forward my more than 2 months.

  2. When I see all these data compiled into nice graphs, I have wet dreams: I’d just wish researchers on infectious diseases would disclose the raw data in a computer readable format so that we could all have a look at the raw data and manipulate it as we see fit.

  3. “The data for these graphs are available.”

    Perhaps. But where? That’s quite unclear to me.

Comments are closed.