The decline of H7N9 Wave 2: some thoughts on why it may be different from Wave 1...

Influenza virus and influenza the disease certainly give scientists a run for their limited money when it comes to predicting what either will do from year-to-year, country-to-country or outbreak-to-outbreak. 

And just when you think you know enough, things change. 

This morning my Twitter stream was fed by a sparkling rivulet of informed comment by @influenza_bio ("A biologist"; follow him if you don't already) on the subject of why H7N9 cases are falling. @influenza_bio groups together a few great points:

1. H7N9 cases are declining.
   Agreed, I think Wave 2 ended almost a month ago.
2. Overall, influenza-like illness (ILI) visits in China have declined.
   A clear parallel, but is it causal? ILIs provide a general guide to influenza circulation (general, because other viruses cause ILI which is basically fever + upper and or lower respiratory signs and symptoms - so a very broad but useful good guide

Percentage of Visits for ILI at Sentinel Hospitals in South China through 2/23/14. http://t.co/yKlqgzOkLg pic.twitter.com/IXKlWlltWX

— A biologist (@influenza_bio) March 12, 2014

3. Is the drop in human H7N9 cases linked to the end of a (silent) outbreak in birds (poultry, waterbirds, songbirds, both...)?
   Finding data on specific bird migration dates in the region is difficult. See here and here for some generalizations. Seems very reasonable.
4. Live bird market closures cannot be the only cause of a drop in H7N9 cases otherwise we'd expect to see cases in other areas continue to rise (presumably areas where markets are not closed).
   If we compare Zhejiang to Guangdong, then we can see that the delay in closing Guangdong's bird markets seems to have manifested as a delay in slowing of human cases; most recent H7N9 case acquisitions have indeed been in Guangdong (a major poultry producing area in southern China) whereas cases in Zhejiang which, like other eastern coastal regions shut their markets earlier and "permanently", generally speaking, have dried up.
5. If H7N9 human case decreases were linked solely to weather, then how could we explain the peak in 2013 which extended into late April whereas it looks to have peaked well before that, in early Feb, in 2014?
   Given that the seasons have not differed between the years (or have they?), I'd suggest we look more at the start of the 2 Waves; Wave 2 commenced earlier in 2014 than did Wave 1 in 2013, but the precipitous decline of both outbreaks of human notifications seemed to have been more closely tied to market closures than dates on a calendar. Of course markets are stocked with H7N9 infected birds and that which links to outbreaks at the supply end unless poultry acquired their infections at markets and then spread that between markets by bird movements which can extend right across southeast China. Why did it start earlier is a key question for me.

@influenza_bio finishes with the comment that...

Thus, wld seem possible that decline we're seeing in human #H7N9 may mostly be due to fall in bird #H7N9 cases. Need more bird surveillance.
— A biologist (@influenza_bio) March 12, 2014

As I've learned from @influenza_bio, many factors go into humans acquiring a particular influenza virus at a particular time/season, and probably no single thing is responsible for all events for any given outbreak. Phew. But that's why we don't have influenza infections all the time and it underpins why they peak at a certain time.

Human acquisition of influenza virus is related to:

• How a person is exposed to the virus (aerosol from upper respiratory tract coughs and sneezes or self-inoculation from contact with contaminated surfaces)
• Whether the virus survives long enough to be inhaled/self inoculated which is in turn linked to virus subtype and strain and environmental temperature and humidity (see some more on that in a guinea pig model here)
• The host and their immune state and general health, smoking, underlying diseases etc
• How much virus enters the host and where it "lands" and makes a footing in the host's respiratory tract
• The spaces we share with infected people and how fast and well the air is filtered/exchanged in those spaces
• The virus subtype in terms of what receptor it prefers and where those might be located throughout the respiratory tract.
• For avian influenza in humans there is also the type and length of exposure to the animal hosts and their environment

Not an all inclusive list I'm sure, but you get the point. Influenza viruses are a complex beast, made more so by the fact that any given subtype could be represented by a range of strains indicating a variety of stabilities, preferences for receptors, antiviral susceptibilities etc. 

So I complete agree with @influenza_bio, more bird surveillance would indeed be a very important step in understanding what is happening in and perhaps predicting the risk of, human outbreaks of this and other avian influenza viruses.