Can Evidence Infiltrate 'Dread Reckoning'?
Thank you to those who conveyed their concerns to me about reports of 50% mortality for influenza A H5N1 (HPAI or avian/bird flu) in humans.
I begin this blog as the last, with Helen Branswell’s 2012 article in Scientific American entitled ‘Dread Reckoning: H5N1 Bird Flu May Be Less Deadly to Humans Than Previously Thought--or Not’.
The author conveyed the limitations of data available in 2012 that might lead to overestimated risk. Pooling the data from across the globe as though exposures, populations, and medical preparedness in every country were equal is dangerous business.
Risk analysts and many scientists are trained to question the representativeness of observational or experimental data for extrapolation to other populations and conditions. Let’s start by exploring the fallacy of assuming that these pooled mortality rates from across the world are representative of all individuals and populations at all future times.
Take another look at this paragraph from my previous blog and build on that evidence to assist our deliberations together.
The World Health Organization compiled evidence of 888 human cases, 463 fatal, for years 2003-2024. Since 1997, 909 human cases from 23 countries were associated with H5N1 (predominantly 2.3.4.4b clade). Thirteen of 28 human cases reported since 2021 are associated with clade 2.3.4.4b. This virus was documented in migratory birds in Africa, Asia and Europe in 2020, and spread to North America in 2021. Some asymptomatic cases were noted, and human symptoms range from no or mild illness to severe pneumonia with respiratory failure, encephalitis, and multi-organ failure.
If you click on the red text above, you’ll see a WHO table or a link to download WHO data listing portions of the data on 909 cases reported in 23 countries. Here is one more WHO table that is helpful in today’s deliberations about assuming the pooled 50% mortality rate applies to you and me, today and into the future.
Now let’s consider what we know and what we don’t know about mortality across countries and across time, from 1997 to present.
For illustration purposes, I selected two groups of countries and the period 2003-2021 to begin our deliberations on today. In parentheses for each country, you’ll see the percentages of 862 total influenza A H5N1 cases in this period and the percentages of the 455 total deaths.
China (6%; 7%); Egypt (42%; 26%); Indonesia (23%; 37%); and Vietnam (15%; 14%)
Chile, Ecuador, Spain, the UK, and the US (no reported cases or deaths in this period)
For the first group of four countries with cases and deaths in this period, here are estimated case-mortality percentages within each country for each year from 2003-2021. Note that the 100% values represent 3-5 fatal cases for given years for China and Vietnam.
Now, let’s look at the most recent 5 years of data for these four countries described by CDC here.
Next, let’s look at the data for cases in three recent years from the second group of 5 countries that had no reported cases in prior years.
So, given trends in numbers of cases and fatalities, plus percentages and other outcomes (asymptomatic and mild cases), over 2003-2021 and 2022-2024, here are some questions that I would pose.
1. Why do you think Table 1 values from 2003-2009 are so different from those in Tables 2 and 3 from 2022-2024?
2. Why do you think influenza A H5N1 caused asymptomatic and mild cases in Table 3 compared to single critical and fatal cases? Compared to fatalities from 2003-2009 in Table 1?
3. Does it look to you like the overall 50% mortality rate from 2003-2009 in Table 1 can be extrapolated directly to the future for any country and any individual, without additional information explaining the trends illustrated here?
4. What more do we need to know to assess the risk of infection, severe illness, and mortality for consuming raw milk that is PCR-positive for viral particles or viral fragments of influenza A H5N1?
If you read my first two blogs on influenza A H5N1, you may expect that mechanistic evidence of oral transmission to humans is tops on my list of unknowns for question 4.
The study by Gambotto and colleagues discusses viral transmission, clinical features, and host restriction and pathogenesis. Influenza A transmission to humans is well documented only via the respiratory system (inhaling droplets or dust particles contaminated by the virus) or direct contact with contaminated objects, then touching the eyes, nose, or mouth. These are the only two ‘theoretical’ transmission routes that CDC researchers are now exploring with experimental models (ferrets) as described in the May 10th update on H5N1. Katella provides additional perspective on influenza A in her post on Yale Medicine.
Remember, there is no evidence that influenza A passes through stomach acids, digestive enzymes, and human innate immune defenses that inactivate enveloped viruses by the oral route. The following viruses are documented in a recent review to be transmitted to humans by oral ingestion:
1. adenovirus
2. astrovirus
3. enteroviruses
4. hepatoviruses
5. norovirus
6. rotavirus
7. sapovirus
Influenza A and other influenza viruses are NOT documented to transmit to humans by ingestion. FDA and USDA had made this determination years ago, but they have been silent about their assessment, perhaps because of their long-standing fears about raw milk.
Second, the high mortality rates in Table 1 are primarily associated with respiratory infections after prolonged direct contact with sick, dying, and dead animals, typically wild birds and domestic poultry. Only in recent times is there awareness that respiratory protections, biosecurity, and safety training can likely prevent occupational and other exposures to influenza A H5N1. I see no evidence that consuming foods containing influenza A H5N1 (nucelic acid fragments or intact viral particles detectable by PCR) has caused illness or death in humans.
Please comment with questions and other evidence on influenza A H5N1 transmission to the respiratory tract, the eye, and the gut.