Cognitive Dissonance - Integrating All Evidence in Complex Disease Processes Is Basic to Scientific Progress

bioRxix preprint re: H5N1 infection in calves and lactating cows counters indirect evidence I've cited for a predominantly respiratory viral spread - maybe it's not a simple straight-forward question?

On August 9th bioRxiv posted a preprint of an international collaborative research effort between partners in the U.S. and Germany: Outcome of H5N1 clade 2.3.4.4b virus infection in calves and lactating cows. Work included intramammary transmission studies in lactating dairy cows in Germany utilizing both the U.S. H5N1 2.3.4.4b B3.13 clade shared by the U.S. and a contemporary European H5 clade. Transmission studies in calves in the U.S. at the Kansas State BSL-3 lab involved 6 intranasally infected calves, 3 sentinels for measuring signs of onward transmission, and 3 isolated negative controls. Please read the entire study and supporting documents in detail; however, here are the summary highlights from the authors:

Summary

In March 2024, highly pathogenic avian influenza virus (HPAIV) clade 2.3.4.4b H5N1 infections in dairy cows were first reported from Texas, USA. Rapid dissemination to more than 190 farms in 13 states followed. Here, we provide results of two independent clade 2.3.4.4b experimental infection studies evaluating (i) oronasal susceptibility and transmission in calves to a US H5N1 bovine isolate genotype B3.13 (H5N1 B3.13) and (ii) susceptibility of lactating cows following direct mammary gland inoculation of either H5N1 B3.13 or a current EU H5N1 wild bird isolate genotype euDG (H5N1 euDG).

Inoculation of the calves resulted in moderate nasal replication and shedding with no severe clinical signs or transmission to sentinel calves. In dairy cows, infection resulted in no nasal shedding, but severe acute mammary gland infection with necrotizing mastitis and high fever was observed for both H5N1 genotypes/strains. Milk production was rapidly and drastically reduced, and the physical condition of the cows was severely compromised. Virus titers in milk rapidly peaked at 108 TCID50/mL, but systemic infection did not ensue. Notably, adaptive mutation PB2 E627K emerged after intramammary replication of H5N1 euDG. Our data suggest that in addition to H5N1 B3.13, other HPAIV H5N1 strains have the potential to replicate in the udder of cows and that milk and milking procedures, rather than respiratory spread, are likely the primary routes of H5N1 transmission between cattle.

As those of you who have read my posts in the past realize, some of the conclusions are at odds with my thinking related to how H5N1 2.3.4.4b B3.13 is currently transmitting in the U.S. However, I must emphasize that I don’t deny the validity of these results, even if we differ on implications of the findings.

First, regarding the viral infusions (both H5 strains) into udders, results clearly show that the bovine udder is a superb incubator for these H5 Influenza A viral strains. Cows become extremely ill at administered doses, potentially resulting in death. I have to wonder if the usual lower field mortality rates could be related to a lower concentration of infecting virus. If so, the “natural” infectious dose, whether delivered by ascending infection from the teat canal, or via the blood-lymph stream through transient H5 viremia, could be quite low. And neither postulated route of infection is mutually exclusive! Within herd lactating cow udder quarter infections could certainly be plausibly explained through either ascending contamination or blood-lymph delivery to susceptible endothelial mammary duct cells in a percentage of lactating cows in affected herds.

Where the paper falls down in my opinion is in stating that milk and milking procedures, rather than respiratory spread, are likely the primary routes of H5N1 transmission between cattle. Unpublished serological work and practitioner feedback both provide good evidence of non-lactating animal infections in affected herds. Even if biosecurity issues can be identified, one must still identify a route for sufficient viable virus material to move from milk through a fomite(s) to other non-lactating animals in the herd. Almost by definition, those animals must be infected “by the head” - oral, eyes, or respiratory route(s) to acquire infection from infected cows exhibiting mastitis. Additionally, often the non-lactating cows are described as sick prior to the lactating cows exhibiting mastitis.

Between herd spread offers additional and even more challenging barriers to milking procedures as the route of transmission. Somehow, viable virus must survive on a fomite all the way to a lactating cow teat orifice on another farm, if “milk and milking procedure” is to be considered the primary route of infection. Multiple biosecurity breaches have certainly been identified in limited epidemiological studies in the U.S. However, those breaches must extend to activities that allow viral survival from the fomite to the cistern of the affected quarter - not a likely pathway! If the pathway involves a preliminary respiratory infection on the destination farm, then the “milk and milking procedure” is not the true route of infection.

Moving to the calf respiratory transmission work, this work produced similar results to what recent ARS research found in direct aerosol transmission studies. Neither study was able to utilize large numbers of animals, given the constraints of BSL-3 facilities. This study performed an additional step of exposing 3 negative sentinel calves to the 6 directly infected calves to explore transmissibility. In this study, the viral infection was NOT transmitted to any of the 3 sentinel animals, despite being found in nasopharyngeal swabs at levels sufficient for viral isolation at 7 days in infected animals. One difference between this study and the ARS work was that this work found no evidence of viral spread beyond the upper respiratory mucosal surfaces, while the ARS work found viral residue in lung sections, indicating a transient more systemic phase of H5N1 infection in directly infected animals.

Systemic infection dictates whether H5N1 virus might gain viremic passage to other parts of the body including lactating mammary tissue. H5N1 has accomplished this in other species, e.g. in cats where CNS infections are lethal. In looking at the 2 studies together, it’s not clear that this virus has clearly gained a reliable viremic pathway in young bovines in BSL-3 conditions; however, the ARS work would indicate that it is at least possible, and lung lesions described in the current work would raise the possibility, despite lack of viral residues described.

The failure of transmission to 3 sentinel calves certainly indicates that under BSL-3 conditions in small numbers, this virus did not demonstrate onward transmission. As with any experimental study, confidence comes with replications, which are complicated with select agents. I understand that ARS may be conducting transmission studies, as well as scientists in Canada, so we’ll likely have more replications soon. Another issue, certainly unproven and controversial, is the question of “superspreaders”. Is there a small percentage of individuals within a larger infected population that shed large volumes of virus and trigger onward transmission to herd mates? Alternatively, do the relatively sterile and stress-free conditions of BSL-3 existence lead to less shedding by infected animals and more natural resistance in sentinels than might be the case under field conditions? Do the study conditions sufficiently mimic the real world?

These questions will sort themselves out as we gain more experience and add more replications. Regardless, we can say that currently this H5 virus does not easily forward transmit under experimental conditions. We probably can predict that if it hangs around in cattle herds long enough, it will likely gain forward transmissibility through natural selection, unless we intervene with vaccines and other controls.

I used the term “Cognitive Dissonance” in the title to this thread. I had an interesting IM thread with a newly cherished colleague this week in which we discussed who we “believe” today, looking at science. After thinking a bit, I told her that I accept at face value everything presented to me and examine the evidence to the best of my abilities, asking as many questions as I can of those I respect. However, with a lifetime of experience, I don’t fully believe anyone, least of all myself! I gladly live with cognitive dissonance in many parts of my outlook on life. Despite that dissonance, I have preferred ideas and beliefs, formed through my own experiences. While I like to say I “follow the science”, deep down I kind of know where I want the science to lead, based on my own value systems, and my own biases and personal and professional rivalries, if I am totally honest with myself.

So as much as I’ve posted my concepts related to the applications of viral disease ecology to the H5N1 outbreak in dairy cattle this year, I’ve tried mightily not to ever believe my positions too fully! None of us should, because progress only comes when divergent hypotheses are openly tested in public dialogue with scientific evidence to support conclusions. This paper provides good pieces of evidence, some challenging my assumptions. Many more pieces of the puzzle will emerge. I can’t wait to keep being proven partly wrong, so that we can all get closer to the truth!

John