1918, I Love You
We should not be worried about a return of the 1918 flu. It has never left.
A new study reanimating the 1918 flu (not for the first time) has set off fresh hand-wringing about a return of the “deadly” virus behind the famous pandemic from both mainstream1 and alt-media2 sources.
This new round of naive alarm-ringing supplants a long tradition of stoking fears about a return of the 1918 virus or some sort of magical reboot of it in the form of a new strain, either from animal (“swine flu” and “avian flu”) or lab sources.
The problem is, the 1918 flu virus has never left.
The world of flu research, it turns out, is vast. As interpreting the evidence on the myth of “Original Antigenic Sin” requires being familiar with that world, I have immersed myself in its history and geography over these last six weeks. My audit of the research proceeded chronologically, so that modern beliefs that do not match the original evidence (like “OAS”) could be recognized.
This puts me in the position of being able to assess, however impudently, the case for whether humanity should be worried about a return of 1918.
The answer: No.
Outline:
Main:
Evidence that 1918 H1N1 was not extra-pathogenic vs. current model
Weak or Suspect Molecular / Lab Evidence for H1N1 Extra-Pathogenicity
Extras:
Case Studies:
Appendix:
Main
No actual argument in the pro-pandemic side
The argument that there is any possible “pandemic” threat from a novel or retro H1N1 flu emergence, absent the long-term removal of H1 from circulation, is patently weak. However, since the elements in play are complex and semi-esoteric - the full story of flu understood primarily by those with a self-interest in stoking confusion and fear - said story is persistently mischaracterized.
A case-study in this mischaracterization, “swine flu,” is presented in the extras.
Evidence that 1918 H1N1 was not extra-pathogenic vs. current model
Below, the lack of evidence for any molecular distinction between restored 1918 genomes and the current H1N1 model are presented. But the mere fact that the current human circulating H1N1 descends directly from 1918 is also evidence in favor of there being no such thing as a separate “pandemic” virus waiting to strike from without. In fact, 1918-derived H1N1 (1950 model) has already been reintroduced into circulation via a lab leak after a 20 year absence. Nothing happened (see Appendix 1 below), except that the virus has stuck around since.
Stronger evidence for non-pathogenicity comes from low death rates in older groups in 1918.
The evidence and intuitive pattern-fitting for prior circulation of H1-model flu in the 19th Century is routinely ignored or dismissed in epidemiological literature for no discernible reason (besides obvious self-interest). Namely: Adults alive before 1885 died at normal rates in 1918, whereas adults born after 1885 died at extra-high rates.
It is patently impossible that an intrinsically extra-pathogenic flu would not cause higher-than-normal-flu death rates in the oldest: They were protected by the same thing that protects them from a normal flu, memory immunity. (Protection is partial in the 35 to 55 groups; it may be the case that “vintage” H1 was dying out over a long period, so that those born after 1860 came into less frequent contact).
Natural conclusions stemming from such a pattern are that:
1918 was not extra deadly in humans who had immunity to a probably highly-distant H1 model, and therefore exactly or comparably “pathogenic” to any 1918-descendent H1N1 flu. Only the large pool of younger adult H1-immune-naive drove aberrant death rates in 1918.
19th-Century H1 was so non-deadly that its disappearance was not even noticeable except in the emergence of H2 and H3 in the void left behind (similar to 1918-descendent H1N1 flu).
As a note regarding death rates in children, there are two factors being considered here: The first is that innate immunity is more protective in this group vs. young adults; the second is that the case fatality rate still seems to exceed normal flu. But besides the under-reporting factor suggested in my graphic, there is the issue that normal “endemic” flu deaths have death rates discounted by survivorship bias: More cases are secondary exposures, and immune memory is already in place. Therefor, 1918 should not be considered extra-pathogenic in children vs. later H1’s. It was similar in nature (because all children experience flu as “novel” at some point) but different in context (because in 1918, all children were doing so in the same year).
Evidence that H1 is more pathogenic vs. H2 and H3?
H2 and H3 returned in 1958 and 1968 after 70 years, an even greater absence than that I have suggested for H1. And yet mortality in young adults was low. As H3 happened to be a reassortment virus that recycled the existing enzyme protein (NA), substantial protection was probably granted by this partial cross-immunity. For H2, note that the very elderly may have been protected by memory immunity; and other elderly by the successful Hilleman vaccine, but for 1958 at least this leaves young adults unaccounted-for. Additionally, antibiotics cannot apparently be credited, as post-influenza pneumonia was not rampant to begin with (regarding the trope that this was the primary etiology of deaths in 19183).
Does this means that a “Vintage H1 protection” theory assumes H1 generally is more pathogenic in humans than H2 and H3? In this, does it agree with a no-protection theory, suggesting that H1N1 is still intrinsically a “pandemic” model?
I would say, probably not.
First, there is the possibility that the recycling of “internal” H1N1 genes in 1958 (H2 was also a reassortment virus) and again in 1968 is enough to totally change the context in terms of immune protection, especially via tissue-resident T Cell recognition of the “internal” proteins corresponding to those genes once they are presented by infected cells.
Another thing to remember is that Case Fatality Rates, by describing a rare outcome, are well-suited for exaggerating marginal differences in pathogenicity. If 27 year-olds died at a rate of 3 per 100 infections; they survived at a rate of 97. This is almost identical to the normal survival rate. And once again, rates in children (who are more consistently protected by innate immunity) were comparable to normal flus for both death and survival. With this in mind, it seems more conservative to suppose that prior H1 avian-human crossovers may also have featured sub-100 survival rates in younger adults without prompting widespread cultural and historical awareness. Or, simply, it has not often been the case in history that flu HA proteins have cycled in a non-reassortment manner, if once again T Cell recognition of the internal genes is what is co-protective with anti-HA antibodies after childhood (so that only one of the two is required; and in 1918 older adults had one and younger adults did not).
Alternately, bacterial co-infection may have been a unique feature of the 1918 H1 crossover favored by contemporary environmental factors which were not repeated before or since (further discussed in Appendix 2); or innate immune dysregulation a unique feature favored in adults due to nutrition gaps in the early processed food era. And obviously, stress and movement due to a uniquely gruesome and visceral war could further explain why 1918 was different than 1958 and 1968 for young adults who were involved (with H3 nerfed in 1968); with the stark divergence in death rates by location a manifestation of the mood of different groups and communities.
As the next section will underscore, it seems unlikely that an “intrinsically” pathogenic H1 would have gone unnoticed in the past.
No genetic evidence for true 1918 H1 “novelty.”
Flu spike glycoprotein (HA) genes began to segregate evolutionarily into their distinct species very early in the virus’s history, per classic phylogenetic analyses. Although 1918 H1 was possibly introduced (along with all 7 other genes) from an immediate (American) avian version that may have been evolutionarily distinct from old world strains, this does not per se reflect any evolutionary "leap" preceding cross-over.4 (Though, more recent research from the House of Webster proposed a mammalian intermediary, but explicitly attached this analysis to an argument for "the need for high-throughput characterization of all 8 gene segments of human virus isolates," i.e. more funding.5)
Moreover, all genes in avian flu evince low or negative evolutionary pressure, with reversions often favored over additional residue changes (due to whatever mutations which do occur usually being detrimental to fitness). Unlike in humans, swine, and lab immune pressure experiments, HAs do not mutate rapidly in fowl. Thus, avian flu is not only a reservoir for flu genes, but a time capsule for ancient designs (emphasis added):
Each HA and NA subtype appears to be antigenically and phenotypically homogeneous and relatively genetically conserved when compared with the antigenic, phenotypic, and genetic distinctness of each subtype [note that this “relative” conservation is not absolute, since inter-subtype differences are extreme]. […] The very high level of conservation observed in proteins of avian viruses suggests that an adaptive optimum has been nearly achieved. The apparent evolutionary stasis of these proteins suggests, further, that within the normal avian host population, any modification of the protein sequence is likely to prove detrimental in the long run. [...] The very low levels of evolution observed for avian virus proteins suggest that many centuries have been required to generate the current genetic diversity and distinct separation of avian virus HA and NA subtypes.6
Since 1918 human H1 and all other genes came from an avian source, they came from a reservoir of stasis.
Therefore:
It is plausible that the H1 re-emerging into human circulation in 1918 was not genetically novel. It may even have been a “back-up” version of H1 after the circulating 19th Century human version went extinct.
It is implausible that H1, an ancient flu gene that appears highly compatible with humans (requiring only one amino acid change from the avian version), would not have crossed into humans before 1918. H1 would have existed well before the likely first transmission of influenza A into man.
Webster, et al. (1992.) Fig 6. Annotated with corrections* and counter-arguments to my “1918 H1N1 is not more pathogenic” theory. (*It’s not clear why the timing for H2 and H3 are occasionally reversed in this paper.) Compared to other influenza A HA’s, H1 is one of the three most “primitive” avian HA genes and the least mutated in its clade.
Webster, et al. (1992.) Fig 5. Annotated based on pixel count. For comparison, H1 is only 2x as different from the last common ancestor with H7/10/3/4/14 as the Omicron siblings are from B.1 SARS-CoV-2 Before the current era (where H1 is safeguarded from extinction by apparently frequent lab leaks), H1 (like H2 and H3) likely cycled in and out of human circulation repeatedly; but the mutations gained during their tours of the human ecosystem were invariably erased. Thus"H1" has been frozen in time in fowl for eons (with the same, singular residue change, E190D, required for later crossover7).
Again, implausible non-crossover is the reason why, above, H1 is assumed to only be marginally more "pathogenic" than H2 and H3.8
Overall, I would rate the evolutionary evidence for 1918 H1 “novelty” as wholly lacking. Certainly, the burden of proof is on epidemiologists to form an argument for 1918’s novelty out of the genetic evidence; they haven’t done so. As with the CFR pattern in 1918, they simply look the other way.
The best argument may in fact be the theory that internal gene recognition is protective during HA crossovers, proposed in the previous section. This, rather than anything special about the 1918 model of flu, might have implications for a future crossover, which is discussed in Appendix 4.
Weak or Suspect Molecular / Lab Evidence for H1N1 Extra-Pathogenicity
If 1918 H1N1 was more pathogenic than the current (1918-descendant) circulating version, it shouldn’t be beyond the powers of motivated scientists to determine why. Four partial and one complete copies of the 1918 (fall and winter) genomes were successfully sequenced in the early aughts. All of them reaffirmed the 1918 flu as the ancestor of all sequenced human H1N1 flus (and the internal genes of H2 and H3 flus), and none of them revealed any apparent “secret ingredient” explaining higher mortality in 1918 (emphasis and line-breaks added):
By the early 1990s, 75 years of research had failed to answer a most basic question about the 1918 pandemic: why was it so fatal? No virus from 1918 had been isolated, but all of its apparent descendants caused substantially milder human disease. […]
These efforts have now [as of 2006] determined the complete genomic sequence of 1 [preserved 1918] virus and partial sequences from 4 others.
The primary data from the above studies (11–17) and a number of reviews covering different aspects of the 1918 pandemic have recently been published (18–20) and confirm that the 1918 virus is the likely ancestor of all 4 of the human and swine H1N1 and H3N2 lineages, as well as the “extinct” H2N2 lineage.
No known mutations correlated with high pathogenicity […] have been found in the 1918 genome9
With no obvious molecular determinants of pathogenicity, experimenters began to compare current-recombinant and 1918-revived-recombinant H1N1 in tissue and animal models. This work was primarily carried out under the auspices either of the steadfast disciple of Robert Webster (the "pope" of flu), Yoshihiro Kawaoka, addressed as Yoshi by colleagues,10 or Jeffery Taubenberger. Neither man's scientific stature would be increased by results suggesting that 1918 was not intrinsically pathogenic.
And so naturally, any contemporary appraisal of these investigations should be wary of the replication crisis that has beset science in the last decade. Were any of the 1918-pathogenicity papers distorted by publication bias, and the context that researchers were looking for physical properties that weren’t in the 1918 genetic code?
The fruits of this early work on reanimated 1918 viruses:
Mouse infection models.
Wherein, current human H1N1’s were used as the “control.”11 However, it has been the case since the first flu strain recoveries in the 1930s that human flu is not pathogenic in mice until pre-passage in ferrets. Therefore, all a 1918 recomb. has to do, to outperform a modern human H1N1 in mice, is “be less human-specific.” Thus, any pathogenicity observed in mice should be interpreted as demonstrating lower human-specificity.
Meanwhile, pathogenicity of a perennial 1918-HA’d mouse-adapted strain (WS) is comparable to regular-HA’d WS, indicating little change between 1918 and 1933 in HA-based pathogenicity.
Tissue culture infection models
Here, Taubenberger et al. appear to demonstrate that the 1918 NA confers trypsin-free cleavage of the HA molecule by unclear means.12 This essentially converts HA into a highly fusogenic, high-pathogenicity mode. However, the use of mouse models to reinforce the findings merely recycle the flaws of the mouse-centric studies (as above). Additionally, given the modern history of avian H5 and H7 cross-overs, it is not clear in a human setting that intracellular cleavage can confer pathogenicity without sacrificing transmissibility. Overall, this work requires further in vivo and in vitro follow-up to grant more weight to, and that follow-up has not taken place.
Macaque infection models.
These involved seemingly inexplicably old (9 years +) macaques,13 raising the potential that adaptive immunity to currently circulating human strains was distorting results for the “control” recomb. virus. The authors did not provide any temporal serodynamics to rule out such immune protection (instead, the macaques were screened in advance with HI assay, which can lead to false negatives since antibodies fade; it is almost impossible that the immune systems of such aged macaques were truly naive to the version of H1N1 that had been in global circulation for their entire lives).
And so rather notably, the controversial paper published this month, using young macaques, completely refutes the results obtained for 1918 virus.14 Note the title.
Why did a reconstituted, “whole package” 1918 virus fail to cause illness in macaques this time when it did last time (despite using the same super-dosing methods as the previous paper)? Because of the robust function of the innate immune system in young mammals.
Emphasis added:
The animals in those studies [by Kobasa, D. et al., i.e. Yoshi and co.] came from a Canadian colony that was depopulating the older macaques, and the ages of the animals used in those studies ranged from 9 to 19 years. Out of the 12 animals used in our study, only 4 animals were between the ages of 9 and 12 years, while the rest were 3 years old and younger. Immune senescence in aged macaques has been described previously and suggests that the deterioration of both innate and adaptive immunity as animals age may influence how well they fight off infections
And so, the most plausible explanation for Yoshi and Co’s results is either that H1N1 is only slightly intrinsically more innate-immune-suppressive (certainly this is possible), or that the macaques were protected from similar outcomes in the control virus due to pre-exposure. The latter seems almost certain, and should have been ruled out more thoroughly at the time, before we can grant their study as evidence for the former.
Threat Mischaracterization Case Study: Swine flu
To see how the threat of another 1918 rests on misrepresenting H1N1 in general, take “swine flu,” centerpiece of the 1976 fiasco and forever after a media boogeyman. This bug is a 1918-derived virus which circulated in domestic swine throughout the rest of the 20th Century, with very little antigenic remodeling (at least at first) thanks to the lack of immune pressure on the HA and NA proteins (because of the yearly slaughter of immune pigs)15. In 1957, 1918-derived human H1 flus died out. Swine H1N1 began to cause swine-to-human transmissions 20 years later, thanks to immune debt.
Lack of immunity in the young, not a change in “swine flu,” drove outbreaks in 1976.
The “swine flu” that caused sporadic illness in 1976 was the same “1918 time capsule” humans had been coexisting with for six decades, without any threat. And in case there is any doubt that a 20 year gap is not enough to render H1’s return dangerous, in the very next year a lab leak resulted in the reintroduction of a 1950-vintage of (1918-derived) human H1. And so the return of H1N1 was accompanied with so little “oof” that the story if its lab origin could easily be memory-holed.
This is the line that has been circulating ever since (i.e., 1918-derived H1N1). (However, since 1979, novel avian H1 genes have co-circulated in swine, as well as H3; but the logic of whether these H1’s (triple-reassortment, etc.) should be monitored as threats always hinges politically and in the media on the link to 1918).
Non-swine-to-human-spread of symptomatic swine flu before 1976 demonstrates that human 1918-derived H1 immunity was still cross-protective against a “time capsule” of the 1918 model (it is not plausible that the immunity of those alive in 1918-23 alone could suppress swine flu between 1924 and 1957, and so “true 1918 immunity” is not what rapidly waned in the years afterward, but general H1 immunity).
The mischaracterization of the “swine flu” threat is par for the course with H1N1 pandemic fearmongering. (The case for a reassortment swine flu threat will be addressed further below.)
Evidence Dismissal Case Study: The “Paradox”
If there is a prima facie case for H1’s presence among humans before 1918,16 why isn’t it acknowledged? The answer cannot refer to any flaw in the evidence.
From the otherwise excellent review by Taubenberger and Morens, used to show the CFR curve above, behold the following foray into utter nonsense (emphasis added):
One theory that may partially explain these findings [aberrant high deaths in younger adults; and normal rates in children and older adults] is that the 1918 virus had an intrinsically high virulence, tempered only in those patients who had been born before 1889, e.g., because of exposure to a then-circulating virus capable of providing partial immunoprotection against the 1918 virus strain only in persons old enough (>35 years) to have been infected during that prior era (35). But this theory would present an additional paradox: an obscure precursor virus that left no detectable trace today would have had to have appeared and disappeared before 1889 and then reappeared more than 3 decades later.17
In what way is this a paradox?!
Since H2 and H3 cycled in and out of human circulation before H1, and then returned after, it makes perfect sense that H1 had also cycled before H2 and H3. That is how cycles work (before lab leaks distort them).
As for “detectable trace,” that trace in the case of H2 and H3 was blood samples from 1955. These allowed detection of pre-return immunity because the returns had not happened yet. If samples after 1958 and 1968 had been examined, there would have been no way to show beyond reasonable doubt that antibodies against the new HA models weren’t simply from recent exposure to the same. Post-return blood samples must all be regarded as mud that you have just seen someone step in - how could you possibly use that same footprint to infer that they had walked in that mud before?
And so, the reason why pre-1918 H1 immunity has no “trace” is because there are no frozen blood samples from before 1918 to look for it in.
A coherent paragraph with the logic of this one would have claimed, “This theory would present a paradox, because we have looked for pre-1918 immunity and have not found it.” It is not a “paradox” that we simply aren’t able to look. Not looking doesn’t mean the immunity isn’t there.
Indeed, as there is no other, more plausible account for the “dark matter” that protects older groups in 1918, it has arguably been staring us in the face for 100 years.
And so instead of coherent, the paragraph is sophistic. It is difficult to imagine an innocent motive for this rhetorical ploy.
Appendix 1: H1N1 Continuity after 1918
Now, let us confront the myth that there is some substantial disconnect between current circulating (human and swine) H1N1’s and the “deadly” 1918 version.
The 1918 version established persistent infection in pigs. This strain was recovered by Shope in 1930 and found to share serologically-relevant physical chemistry (antigens) with contemporary circulating human strains, along with not being lethal in pigs or humans. Additionally, later recoveries from swine featured no antigenic variation from this version, due to the yearly slaughter of (immune) pigs. And so current human strains were both related to the 1918 virus in design, not lethal to humans, and related to a non-lethal pig design that was free of attenuating evolutionary pressure (ie., swine flu would not have been penalized for having high mortality rate). Since swine flu spreads well in young pigs, it must feature non-attenuated innate immune suppression; and yet, its hosts (young pigs) are still rarely killed. This matches the mortality of the 1918 virus in the youngest humans (i.e., similar to normal flus).
The 1918-derived human H1N1 virus persisted in circulation until 1957. Substantial changes to the physical chemistry of H1 (“the antigen”) seem to have occurred in 1934 and (even more-so) in 1947, as would be expected with immune escape pressure; but the model in 1957 was still the genetic descendant of 1918 (as initially inferred from cross-reactivity to Shope’s 1930 swine flu recovery, and later from gene sequencing).18
This version was remodeled in 1958, with the arrival of “H2N2” - a reassortment in which a previously unrecognized avian glycoprotein and enzyme protein (HA and NA), as well as a new PB1 gene, circulated in a virus containing the other 5 “under the hood” genes from the 1918 bug. Another remodel occurred in 1968, ushering in the H3N2 model that still circulates today.
In 1976, only 19 years after H1 disappeared from human circulation, (1918-derived) swine H1N1 began to prompt human infection in geographically disparate locations (emphasis added):
In January, 1976, the Hsw1N1 swine strain, identical with those found in pigs in other parts of the U.S. in recent years, was isolated from a soldier who had died of influenza at Fort Dix. This strain was also recovered from five other cases of influenza among service personnel at Fort Dix, although at the same time influenza caused by the A/ Victoria/75 strain was infecting troops in the camp. Serological investigations showed that the Fort Dix swine strain had infected some 500 personnel at Fort Dix, but ap parently it did not spread further.
Previously human infection with Hsw1N1 had been observed only in 1961 in a farm worker in Czechoslovakia and in 1974 in a Minnesota farm boy who died of Hodgkin's disease (which inhib its the immunological protective system, so that the swine virus was proba bly an incidental infection). Just before the Fort Dix episode, however, Hsw1N1 infection in an eight-year-old boy and members of his family was noted in Wisconsin in the fall of 1975, coincident with an outbreak of swine influenza in pigs on the same farm. Since February, 1976, human infection with Hsw1N1 has been detected by isolating the virus from sick people on two farms in Wisconsin and one in Minnesota where pig infection was also present. […]19
This “gathering storm” of (1918-derived) swine-human H1N1 crossovers was interrupted, of all things, by the wholesale reintroduction, the year later, of (1918-derived) human H1N1 from a circa-1950 isolate, likely via lab leak.
In 1977 the Russian H1N1 influenza virus that had circulated in humans in 1950 reappeared and spread in children and young adults. This virus probably escaped from a laboratory and has continued to cocirculate with the H3N2 influenza viruses in the human population20
For a visualization of the raw evidence for this conclusion, here are the contemporary strains sorted by phylogenetic resemblance of their H1 genes:
This leak became co-fixed with H3N2 (meaning that neither has died out, and both seem to be preventing a return of H2 from the avian reservoir, for now). The result is direct continuity between the 1918 H1N1 spike gene and today’s H1N1 spike gene.
Here, a clear continuity between the avian-introduced 1918 H1N1 and the present strains is shown. Thus, the 1977 H1N1 reboot (in which the lab version became fixed in humans over the swine version) did not encounter an H1-naive young adult (20+) population, whereas the 1918 reboot did (20 to 33 and higher). However, as suggested above, higher mortality in 1918 (compared to the H2 and H3 re-emergences) may have still been more a result of contemporary host and environmental factors.
Additionally: Note that the “Russian” flu’s resemblance to an obsolete (1950) version of H1N1 is the primary argument for its lab origin; the same exact argument I made for the Omicron siblings in regards to “B.1” SARS-CoV-2.
Appendix 2: 1918 flu and age-group mortality
1918 H1N1 was likely a reintroduction (from an avian source) of the “H1” flu spike glycoprotein model after a disappearance circa 1885. This is why children and elderly experienced normal mortality, but some younger adults (who at that age begin to lose innate immune protection, and rely on memory immunity, according to individual health) were collectively bushwhacked.21
Additional support for the circulation of a protection-conferring H1 flu up until sometime in the mid-1880s is the high mortality experienced in remote islands in 1918. Alternately, this could be taken as an argument for the effect of cross-protection from post-1889 H2 and H3 exposure. In either case, adults in remote islands had little to no immune experience with flu and were therefore remarkably susceptible. To quote EO Jordan:
The description of what happened in Paperu on the island of Saparua (Nederlandsch Indie 1920, p. 45) must have been true of many other communities: “For some time, only eight of the 800 inhabitants were able to do their work. The others, all patients, were absolutely abandoned to their fate, and in the absence of any help they remained without food, drink and medicine.” A similar condition of misery prevailed in numerous places. Houses were closed, the streets were deserted; inside the houses the children cried for want of food and drink, and no one could help. Those who were on their legs again were too weak to provide so many with water and food. These were the days of the greatest misery. Even domestic animals died from hunger.” A death total of 800,000 or a rate per hundred thousand of about 1,600 [~4 to 5 times higher than Western nations] is probably a conservative estimate of the ravages of influenza in the Dutch East Indies.22
The Cytokineness of Strangers
The alternate, and perhaps currently reigning, explanation for the “W” mortality distribution in 1918 is that it was the robust immune response of younger adults which, via overreaction, turned their lungs into soufflé - afterward leaving them susceptible to opportunistic bacterial infection. This has been incorporated into models presuming an extra-pathogenic virus (by Burnet, in 194223) and a run-of-the-mill virus.
Burnet’s theory is notable for being influenced by his brilliant, early (re-)intuition of the importance of tissue immunity.24 However, he could not know that influenza A carries within its genes a whole toolkit of intracellular innate-immune suppressing proteins - therefor, all influenza A possesses the "extra" pathogenicity that he was intuiting as an explanation for its virulence in younger adults. Today we can assess the genome of the 1918 virus or replicate it in a lab to measure any difference in these qualities; and once again it scores, at most, slightly ahead of modern H1N1, with a need for better replication of certain findings. More fatally for any theory along these lines, extra-pathogenicity is unequivocably refuted by comparison of the mortality rate in the elderly with normal flu.
The best presentation of an “immune overreaction” theory featuring a normal-pathogenicity flu might be from Shanks and Brundage, helpfully highlighted by reader “Fla Mom” in a comment.25 In particular, the authors are provocative regarding the evidence for the protection afforded by isolation or long-term residency in locations where outbreaks occurred - the implication being that secondary pnuemonia was prompted by exposure to unfamiliar common pathogens in the wake of flu.
However, the problem once again resurfaces of why 1958 and 1968 didn’t replicate the same pathology - weren’t people moving around then, as well? Between all the war and protest, 1968 should adequately serve to demonstrate that the severity of primary infection is the difference. (Notably, this wasn’t as much of a “new” flu, since 5 genes including the antigenic enzyme protein (NA) were retained.)
As for why the “movement” problem is replicated in specific communities in islands, despite a lack of movement, the authors must substitute an alternate theory involving environmental and host factors besides movement. What seems more likely to be the case is that, once again, a very marginal difference in pathogenicity in the H1 model vs. H2 (and maybe H3) combined with a universe of confounding host vulnerabilities, some of which were environmental.
Finally, the supposition of immune overreaction in 1918 in modern literature is just that - a supposition. There is no primary evidence favoring “overreaction;” “cytokine storm” is a modern term whose application to the 1918 flu is an anachronism.
Shanks and Brundage quote the flawed macaques study as a modern lab demonstration of immune overreaction - but here, actually cite a contemporaneous commentary which interprets the results backwards. (Aged) macaques dosed with the 1918 virus experience immune suppression in the initial days; cytokines only surge after the virus has been allowed to replicate out of control (at which point there is no “over”-reaction, only proportional reaction to massive tissue damage).26
Lastly, the actual, essential relevance of bacterial invasion - as opposed to coagulation- or inflammation-based pulmonary edema - to the pathology of flu fatalities strikes me as questionable, given the inconsistent results of culturing.27
The example of SARS-CoV-2 shows us that what seals the fate of individual adults confronted with a novel respiratory virus is the promptness of their immune response,28 i.e. innate immunity (with the prompt initiation of novel memory responses being defined as part of innate immunity). For this reason, the young fare well, and adults not so much. The “individual” reasons can closely track to environmental factors, including the degree of spring and summer exposure (explaining the high performing medical staff highlighted by Shanks and Brundage).
Finally, previous H1 immunity offers a non-contrived explanation for why older groups are progressively rescued from mortality. In order to explain why even younger individuals - teens and the youngest 20s - do not “overreact” more than those older than them, Shanks and Brundage must resort to reference to sensitization with prior flus. However, an overlay of birth year the with dates of emergence of prior flus, as I have provided above, demonstrates that there is no apparent pattern here.
If 1889 exposure sensitizes, then those born circa 1885 should continue to show higher rates vs. those born in 1890. If 1900 exposure sensitizes, those born circa 1895 should show higher rates vs. those born in 1890 and 1885. Both exposures offer an inconsistent culprit for mortality rates and should be interpreted as nearly irrelevant, or perhaps protective (additive to the pattern already destined by innate immunity).
Appendix 3: A Case for recombinant swine H1 threat?
Avian H1 is an extremely evolutionary conserved protein. There is no basis for assuming that avian-to-swine-to-human H1 crossover will confer pandemic potential. Avian HA’s are a reservoir of stasis, not threat.
The 2009 swine flu scare, concerning triple-recombinant swine H1 flus with avian-derived genes, lacked even a bit of evolutionary or molecular evidentiary basis. Since everyone alive today has immune recognition of H1 and H3, neither should be considered to pose a threat. They are as tamed animals (even if it is only the trainer, not the animal, who changes).
What about H2, which was semi-harmless in 1958, but perhaps only because it was a reassortment with the H1N1 internal genes that were already familiar to the immune system? Is this the species of HA lurking in the avian reservoir that, if combined with novel internal genes (namely, a reassortment with the avian internal genes that are still circulating in pigs), would finally make the dreams of another 1918 come true?
It seems at least possible (though N1 and N2 immunity would still provide cross-recognition, even from an avian N1 or N2). But at least this could serve as a rubric for “rational” alarmism in the future: If it is about a currently human-circulating HA species in swine, it should not be regarded as threatening, whatever it is doing with other genes, because each HA species is a (separate) highly conserved design in the avian reservoir.
Appendix 4: What about the next long H1 absence?
And so it might finally be asked, if there was nothing “special” about 1918 H1 besides adult immune naivety… so what? Shouldn’t we worry that the same situation will happen again one day - that H1 will die out and take longer than 20 years to come back next time, or that it will bring with it a set of unfamiliar internal genes?
In the first place, it’s absurd to worry about a danger that requires 20 years of lead-up before that lead-up even begins. Unfortunately for flu researchers who potentially depend on government and media panics just to maintain baseline funding, they might need to get a different line of work in the meantime.
And yet ironically, this could be the thing that would make such a long absence possible. For as long as the current situation continues - with H1 flus perpetually unfrozen in labs and re-cultured, not to mention work on vaccines - it doesn’t seem possible that H1 can die out for more than 20 years. Just as in 1977, a lab strain will reemerge when immune debt begins to build, but before a new generation of adults is vulnerable.
If you derived value from this post, please drop a few coins in your fact-barista’s tip jar.
Salzberg, Steven. “Scientists Have Re-Created The Deadly 1918 Flu Virus. Why?” (2022, August 15.) Forbes.
Salzberg badly mischaracterizes the new study’s dosing method (it included high doses identical to earlier studies).
Renz, Tom. “Fauci is Now Performing Gain-of-Function on the Spanish Flu.” (2022, August 19.) Tom’s Newsletter.
This is further discussed in Appendix 2.
Webster, RG. et al. (1992.) “Evolution and Ecology of Influenza A Viruses.” Microbiol Rev. 1992 Mar; 56(1): 152–179.
Smith, G. et al. (2009.) “Dating the emergence of pandemic influenza viruses.” Proc Natl Acad Sci USA. 2009 Jul 14;106(28):11709-12.
This is an interesting document, as Webster is included in the author list, in a paper partially rebutting Taubenberger’s reaffirmation of the direct-transmission phylogenetic analysis presented by Webster in 1992.
Besides over-reliance on the molecular clock (which the authors are content to let propose that the 1977 , the apparent fact that future human H1N1’s appear to descend from a near-ancestor of the Brevig Mission sequence (and nearly identical partial sample sequences) is highlighted. Is this meant to be construed as an argument that the 1918 flu must have been “cooking” in an intermediary host between fowl and man? It is genuinely hard to tell, as here the authors let their algorithms do so much of the work of discernment that they seem to regard it unnecessary to actually analyze discrete findings.
But if so, this would be a logical non sequitur (of the described genetic non sequitur). Since flu had already spread beyond America in the spring of 1918 (and asymptomatically continued to spread) the lineages that became fixed in human and swine H1N1, were they not in the US during the spring and summer, obviously would not have featured mutations that occurred in the US during the same period (a period of faster mutations as the avian flu genes optimized for their new host). To illustrate the point, it is regrettably obligatory to depict the 1918 flu as a delicious avocado, gifted to humanity by a rubber ducky:
There is a limit, of course - at a certain number of lost mutations, it would certainly be convincing that “WS, etc.” were restored from a version of flu from before spring, implying reservoir in a non-avian source, which logically would be an intermediary. My going to bat for direct transmission is therefor somewhat trollish.
Reid, AH. et al. (2003.) “1918 Influenza Pandemic and Highly Conserved Viruses with Two Receptor-Binding Variants.” Emerg Infect Dis. 2003 Oct; 9(10): 1249–1253.
One possibility is that the previous circulating version of flu was human-adapted for more temperate and less immunogenic circulation in some unclear way, so that H1 was omnipresent and less-noticed in humans for centuries before 1885. One possibility I considered was that the prior version of flu was adapted for drinking water transmission (as avian flu still is), and went extinct due to sanitation measures. Thus in 1918 humanity gets a “backup” flu restored from an avian source. In this model, there’s no need for a prior “human” H1 at all, since immunity could derive from drinking avian H1 in drinking water.
However, this does not explain the apparent non-protection in tropical areas (millions are supposed to have died in India in 1918). A temperate, water-adapted influenza should have preserved H1 in human circulation well into the 20th Century thanks to the tropical world. So that theory can probably be tossed: Pre-existing immune protection was likely from a human, not avian H1, which was globally prevalent and died out.
Taubenberger, JK. Morens, DM. (2006.) “1918 Influenza: the Mother of All Pandemics.” Emerg Infect Dis. 2006 Jan; 12(1): 15–22.
Mandavilli, A. (2006.) “Profile: Yoshihiro Kawaoka.” Nature Medicine. volume 12, page 489 (2006)
Kobasa, D. et al. (2004.) “Enhanced virulence of influenza A viruses with the haemagglutinin of the 1918 pandemic virus.”
Kash, J. et al. “Genomic analysis of increased host immune and cell death responses induced by 1918 influenza virus.” Nature. 2006 Oct 5; 443(7111): 578–581.
Tumpey, T. et al. (2005.) “Characterization of the Reconstructed 1918 Spanish Influenza Pandemic Virus.” Science. 2005 Oct 7;310(5745):77-80.
Kobasa, D. et al. (2007.) “Aberrant innate immune response in lethal infection of macaques with the 1918 influenza virus.” Nature. 2007 Jan 18;445(7125):319-23.
Chan, M. et al. “Pandemic 1918 Influenza Virus Does Not Cause Lethal Infection in Rhesus or Cynomolgus Macaques.” J Virol. 2022 Aug 4;e0072822.
Overall rate of evolution for the H1 swine flu HA protein from 1930-1988 is found to be high by Webster, et al. in 1992; but it is not clear whether this combines the “classic” and avian strains, as the reference is to yet-unpublished in-house work.
I am setting aside Masurel and Heijtink’s absurd attempt to use (post-FM1-era) 1977 samples to infer a wave of “FM1”-like virus in the early century, “because OAS.” All results would be taken to flatly refute OAS, in any other context - and given that Thomas and Co. had published an entire series of studies alleging non-immunity against FM1 in older populations, the results from Masurel cast serious doubts on the validity of the work from both authors.
Kaplan, MM. Webster, RG. “The Epidemiology of Influenza.” (1977, December.) Scientific American.
p.226 Jordan, Edwin O. (1927.) “Epidemic Influenza: A Survey.” Archived online at https://quod.lib.umich.edu/f/flu/8580flu.0016.858
Burnet’s theory, apparently delivered in his 1942 monograph, is characterized by Andrewes a few years later:
Andrewes, CH. (1949.) “Influenza in Perspective.” Edinb Med J. 1949 Aug; 56(8): 337–346.
Before the end of the Flexner era, tissue (i.e. cellular) immunity was apparently still routinely proposed as more important than antibody (i.e. humoral) immunity. It is mentioned in:
Rivers, TM. (1927.) “Filterable Viruses A Critical Review.” J Bacteriol. 1927 Oct; 14(4): 217–258.
The degree of active immunity usually exibited by individuals recovered from virus diseases seems disproportionate to the amount of passive protection afforded by their sera. This fact has led many observers to believe that the protection against virus diseases is predominantly a tissue immunity rather than a humoral one.
Heresy, from the (then-) pope himself!
Shanks, GD. Brundage, JF. “Pathogenic Responses among Young Adults during the 1918 Influenza Pandemic.” Emerg Infect Dis. 2012 Feb; 18(2): 201–207.
Several key cytokine genes, including IL-8 and CXCL11, showed a delay in activation in 1918-virus-infected animals, although several chemokines important for the activation and recruitment of neutrophils, including CXCL6 and CXCL1, were preferentially upregulated. Strikingly, K173-infected animals showed a marked increase in expression of mRNAs for many type I interferons (IFNs) and a corresponding increase in mRNA expression of type-I-IFN-stimulated genes early in infection (Fig. 4c), coinciding with the greatest load of the virus (Fig. 1). This response was downregulated on days 6 and 8 post-infection, when the K173 virus was not detected. 1918-virus-infection, in contrast, induced much fewer IFN-α genes, suggesting that it caused an altered antiviral response in the bronchus. Accordingly, the 1918-virus-infected animals also showed differential activation of type-I-IFN-stimulated gene expression on days 3 to 8, despite viral titres in bronchi that were 10–5,000-fold higher than in K173-virus-infected animals.
An important pathway in activation of the antiviral response to influenza virus infection occurs through the activities of DDX58 (or retinoic-acid-inducible protein I) and IFIH1 (or melanoma differentiation-associated gene 5)20,21. Both genes were induced in the K173-virus-infected, but not the 1918-virus-infected, animals
The same dynamic of immune suppression followed by (proportionate) reaction to extreme viral destruction has been recognized in severe disease resulting from infection with SARS-CoV-2:
McGonagle, D. et al. “The Role of Cytokines including Interleukin-6 in COVID-19 induced Pneumonia and Macrophage Activation Syndrome-Like Disease.” Autoimmun Rev. 2020 Jun; 19(6): 102537.
Analogous to primary HLH, the loss of “front line” anti-viral defence mechanism may activate a “second wave” of more tissue aggressive immunity including exaggerated IL-6 production with a secondary cytokine storm supervening with increased tissue damage
As in O'Malley, JJ. Hartman, FW. (1919.) “Treatment of influenzal pneumonia with plasma of convalescent patients.” JAMA. 1919;72:3437.
“Out of forty cultures ten were found positive, showing pneumococcus Type I in 2 cases; Type II in 2 cases; Type IV in 4 cases, and Streptococus viridans in 2 cases.
It’s not clear that this exceeds what a random sampling of the same population would produce.
As the agents that were cultured were all common respiratory colonists, in Shanks and Brundage’s description, there is no self-evident way to distinguish their role as that of either instigator or bystander.
As with severe SARS-CoV-2, the extensive tissue destruction of the virus can plausibly set off coagulation and inflammatory positive feedback-loops that take days to lead to severe or fatal outcomes, even after the immune system has finally replied to the virus. The classic review by Morens, Taubenberger, and Fauci doesn’t articulate any rational for why their tissue samples justify a diagnosis of bacterial pneumonia as opposed to pneumonia from an inflammation cascade set by a virus now gone. Pathology in general suffers from the hazard of sorting out what it is that causes a dysfunction leading to death from the extreme second-order effects of dysfunction preceding death. Those effects will include disordered behavior by the microbiome, as immune reaction to the virus or to tissue debris can convert temperate bacteria to a pathogenic phenotype; that still wouldn’t mean the bacteria caused the immune reaction or the tissue destruction. Reference to circumstantial observations - that those who were isolated fared better - cannot replace a true pathological analysis. (I welcome an explanation for why the case for a true bacterial invasion etiology is actually strong in non-circumstantial terms.)
Overall, the ideal that bacteria played a “predominant role” seems to stem from human pattern-fitting in the face of a complex reality.
Morens, . Taubenberger, JK. Fauci, A. “Predominant Role of Bacterial Pneumonia as a Cause of Death in Pandemic Influenza: Implications for Pandemic Influenza Preparedness.” J Infect Dis. 2008 Oct 1; 198(7): 962–970.
See “Assay What,” in which I highlight among other things the following relevant results from the placebo arm of the Paxlovid trial:
Severe outcomes, under 65 (with co-morbidities):
46/909 (5.06%)
Severe outcomes, early seroconversion (with co-morbidities, including all ages):
8/528 (1.52%)
Prompt immune response prevents later “overreaction.” More importantly, it prevents the cause of that “overreaction,” the critical mass of cellular destruction that instigates an inflammatory and coagulation cascade.
Two of my former colleagues published this about 10 years ago, which is also relevant to reconsideration of the threat from the Great Influenza:
"Of the unexplained characteristics of the 1918–19 influenza pandemic, the extreme mortality rate among young adults (W-shaped mortality curve) is the foremost. Lack of a coherent explanation of this and other epidemiologic and clinical manifestations of the pandemic contributes to uncertainty in preparing for future pandemics. Contemporaneous records suggest that immunopathologic responses were a critical determinant of the high mortality rate among young adults and other high-risk subgroups. Historical records and findings from laboratory animal studies suggest that persons who were exposed to influenza once before 1918 (e.g., A/H3Nx 1890 pandemic strain) were likely to have dysregulated, pathologic cellular immune responses to infections with the A/H1N1 1918 pandemic strain. The immunopathologic effects transiently increased susceptibility to ultimately lethal secondary bacterial pneumonia. The extreme mortality rate associated with the 1918–19 pandemic is unlikely to recur naturally. However, T-cell–mediated immunopathologic effects should be carefully monitored in developing and using universal influenza vaccines."
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3310443/
The problem with GOF research is if the 1918 flu is not bad, they will make it bad.