Wednesday, March 29, 2023

UK Technical Briefing (#3) On Risk To Human Health From HPAI H5N1



Although it still isn't clear whether avian H5N1 has the `right stuff' to spark a human pandemic, over the past 18 months we've seen it spread across two new continents (North & South America), spillover into dozens of mammalian species (including a few humans), and reinvent itself into more than 5 dozen new genotypes. 

Regardless of whether it will be successful, it is certainly sending out warning signs. 

Along the way we've seen - with increasing frequency - a number of detailed risk assessments, and technical documents, released by major public health entities like the WHO, CDC, ECDC, and the UK's HSA.   

A few in recent months include:

CDC Technical Report: Highly Pathogenic Avian Influenza A(H5N1) Viruses
New information on outbreaks in poultry, spillovers into mammals, new emerging genotypes, and sporadic human infections emerges practically every day, necessitating frequent updates. Today, for the third time in as many months, the UK's Health Security Agency has updated and published their risk assessment. 

This is a lengthy, and quite detailed report, and while much of it is a rehash of what we've seen in previous reports, in today's briefing they include 3 potential pandemic scenarios of differing severity. 

They had briefly mentioned the first two scenarios (a mild and a severe 1918-style pandemic) in their last report, but have fleshed it out to include a 3rd, even more severe scenario. 

You'll note, however, that the `tabloid-favored' 50% fatality rate is not considered, as that is based only on a very limited sample of hospitalized H5N1 cases reported over the past 20 years (mostly from places with limited healthcare delivery). 

For a more detailed look at why a 50% fatality rate is highly unlikely, you may with to revisit The H5N1 CFR (Case Fatality Rate) Debate.

I've only posted a few excerpts from the UK report, so follow the link to read it in its entirety.

Research and analysis
Investigation into the risk to human health of avian influenza (influenza A H5N1) in England: technical briefing 3
Updated 29 March 2023
Applies to England

The UK Health Security Agency (UKHSA) is working with the Animal and Plant Health Agency (APHA), the Department for Environment, Food and Rural Affairs (DEFRA) and the public health agencies of the 4 nations to investigate the risk to human health of avian influenza (influenza A H5N1) in England. This briefing is produced to share data useful to other public health investigators and academic partners undertaking related work. It includes early evidence and preliminary analyses which may be subject to change.

Data reported in the technical briefing is as of 15 March 2023 (or as specified in the text) to allow time for analysis.

Code associated with technical briefings can be found on GitHub.
Levels of human health risk related to the outbreak of avian influenza in England

Since the previous technical briefings, risk levels have been updated by the Technical Group. These adjusted levels synthesise epidemiological, virological and genomic indicators at every stage.

The following indicators are used for assessment of the risk level:
The avian influenza outbreak can be considered to fall into one of 7 potential levels of transmission:
  • detections in wild and farmed birds, avian host range and seasonal pattern
  • the frequency, nature and host range of mammalian spillover infections
  • exposures of confirmed human cases
  • the size of any clusters of mammalian or human infections, the implied mode of transmission (for example, evidence of long-range aerosol transmission), and estimated transmissibility where possible
  • genomic and phenotypic evidence of changes in receptor binding or other changes that may impact transmissibility (haemagglutinin)
  • genomic and phenotypic evidence of other adaptation for mammalian infection (non-haemagglutinin)
Level 0  Avian influenza circulating in birds, with normal epidemiology.
Level 1  Avian influenza circulating in birds, with abnormal epidemiology.
Level 2  Evidence of propensity to infect humans or other mammalian species. 
Level 3  Limited mammalian transmission (excluding human).
Level 4  Limited human-to-human transmission.
Level 5  Sustained transmission in mammals.
Level 6  Sustained human-to-human transmission.

Emerging influenza viruses may jump levels in this scheme (for example, from level 3 to 6), either because they emerge directly and successfully into humans, or because intermediate stages are not detected through surveillance.

The UK risk is assessed as level 3 (limited mammalian transmission, low confidence).

Influenza A(H5N1) clade can successfully infect mammals and humans with high levels of direct exposure (for example, scavenger mammalian species, humans working closely with birds).

Polymerase basic protein (PB2) mutations associated with mammalian adaptation, including E627K, are seen in mammalian infections. There is evidence of transmission in farmed mink and possible, though unconfirmed, transmission in populations of seals and sea lions.

There is limited mammalian surveillance.

The confidence level is assessed through the synthesis of epidemiology, genomic and virology data. 

Part 3. Planning scenarios
3.1 Scenario development: summary of assumptions

UKHSA has developed scenarios of early human transmission of influenza A(H5N1). It is not possible to know key parameters in advance, so these scenarios are not predictive. They are useful to explore the very early period of circulation of a new influenza virus in the UK, up to the time human-to-human transmission is detected, and the potential impact of pharmaceutical and non-pharmaceutical interventions.

The Technical Group reviewed and agreed the following parameters to be used in preliminary scenarios:


These scenarios assume that sustained human-to-human transmission has begun – by definition a reproduction number (R) above one. Past avian influenza outbreaks which have not led to sustained human-to-human transmission cannot be used as a guide to transmission parameters. Based on pandemic influenza data, we judge that looking at scenarios where R is between 1.2 and 2 (or where doubling times are roughly between 3 days and 11 days) is a useful initial assumption. 

Three scenarios have been developed, each assuming a different level of clinical severity associated with the disease. All 3 scenarios should be considered severe compared to the most recent influenza pandemic in 2009.

In scenario 1, the infection-fatality rate (IFR) is around 0.25%, similar to that of coronavirus (COVID-19) in autumn 2021, though considerably higher than the overall severity of the 2009 H1N1 outbreak (0.01% IFR) (Riley and colleagues, 2011; Wong and colleagues, 2013). There are very few studies with estimates of an infection-hospitalisation rate (IHR), and for this scenario it has been set at 1%.

Scenario 2 is more severe. We base our parameters on the 1918 influenza pandemic, which had an IFR of around 2.5% (Murray and colleagues, 2006). We also assume in this scenario an IHR of 5%. Although various H5N1 outbreaks have much higher fatality estimates, these are not examples of sustained human-to-human transmission and are therefore not directly comparable (for example see Li and colleagues, 2008; Poovorawan and colleagues, 2013).

In scenario 3, we choose a higher IHR of 10%, with IFR 5%.
This approximates the type of outbreak seen with SARS-CoV-1 in 2002 (9% reported case fatality rate). We note significant variation in reported CFR rates for SARS-CoV-1, suggesting the underlying IFR may have been even higher than the crude average of reported deaths as a fraction of reported cases.
          (Continue . . . )