The timing and location of the first cases of the 1918 influenza pandemic are still controversial, a century after the pandemic became widely recognized. Here, we critically review competing hypotheses on the timing and geographical origin of this important outbreak and provide new historical insights into debates within military circles as to the nature of putative pre-1918 influenza activity. We also synthesize current knowledge about why the 1918 pandemic was so intense in young adults. Although it is still not clear precisely when and where the outbreak began and symptom-based reports are unlikely to reveal the answer, indirect methods including phylogenetics provide important clues, and we consider whether intense influenza activity as far back as 1915 in the USA may have been caused by viral strains closely related to the 1918 one.
Moreover, the noteworthy influenza outbreak at Camp
Funston, Kansas, in March 2018, was mild, with many cases but
few deaths. This was after New York City already had evidence of
very high influenza mortality (higher than any other period except
for the fall of 1918, in fact) and increased mortality in young adults
[17]. The virus causing those cases in New York City in February
presumably predated the March outbreak at Camp Funston
To our knowledge, the only reasonably persuasive evidence we have of geographical
origin, which is by no means conclusive, comes from phylogenetic
analyses indicating that most of the avian-like genomic segments
in the 1918 human virus appear to be of Western Hemisphere and,
probably, North American origin.
"... and that the pandemic
virus’s HA gene was likely circulating in the human population
for many years prior to 1918.
A similar pattern can be observed in ongoing cases of avianorigin
infection by viruses including H5N1 and H7N9 [25]. As in
1918, when a generation of young adults exposed as children to a
group 2 HA suffered severe outcomes when later infected by a
group 1 HA virus, most cases of H5N1 (group 1) are observed in
those born after 1968, exposed in infancy to the human group 2
H3N2 virus. Conversely, most cases of H7N9 (group 2) are
observed in those born before 1968 and exposed as kids to either
H1N1 or H2N2, both of which are group 1 viruses. Hence, as a
general phenomenon, first exposure to an influenza virus appears
to determine the group of viruses to which a person develops a
life-long immunological imprint. Individuals who were born during
a period when their first exposure to IAV was to a strain of the
opposing phylogenetic HA group have immunologic imprinting
for the ‘wrong’ group of viruses and are therefore at increased risk,
possibly because they are less able to recall protective responses
to conserved epitopes of the HA stalk domain, which tend to be
shared within HA groups.
Although there is some evidence that the 1918 virus may have
elicited unusually strong innate immune responses—a so-called
‘cytokine storm’ [26]— we believe the imprinting hypothesis explains
the data better than the hypothesis that the young-adult
mortality in 1918 was a consequence of that cohort having
stronger immune systems, and hence suffering greater negative
effects than other age groups. First, we see no reason, under the
cytokine storm hypothesis, that the peak in mortality should have
been centred, and sharply so (Fig. 3), on 28-year-olds. Do 28-year olds
really have markedly stronger innate immune responses than
18-year-olds? Second, the primate experiments suggesting unusual
lethality of the 1918 virus included only a single modern
control strain. Perhaps a larger null distribution of other H1N1
strains would show that the 1918 virus was not so unusual in its
severity. Finally, because virtually all adults have had prior exposure
to IAV, experiments on immunologically naive animals [26, 27]
may generate misleadingly severe outcomes: if the same animals
had been previously infected by IAV at a young age, then secondarily
exposed to the 1918 virus as adults, it is unlikely they would
have experienced such severe symptoms. Nevertheless, results
indicating that the 1918 virus is unusually lethal in mice [27], along
with the fact that introduction of all new internal proteins in the
1918 virus could have played some role in its unusual virulence
(due, e.g., to absent cellular immunity to new T cell epitopes) [20]may be the only way to
resolve these questions and answer, finally, why this pandemic
was so catastrophic.
Returning to the 1917 paper by Hammond, Rolland and Shore
[5]: by describing a ‘symptom complex so distinctive as to constitute
a definite clinical entity’ they provided what may be the first
clear account of the pandemic disease that peaked in 1918 and
became known as ‘Spanish flu’. With great clarity the authors
described the clinical features and rapid progress of the disease,
which typically resulted in death by asphyxiation caused by pus-filled airways. They described their unsuccessful attempts at treatment
and alluded to their efforts to produce a vaccine. As medical
officers so close to the front in the British army’s biggest ever
overseas military hospital camp, their priority must have been
to attend to the thousands of soldiers with wounds.
Nevertheless they recognized an unexplained disease outbreak
as a scientific challenge that they should study, record and report.
The paper is thus of historical importance and a credit to the
authors and to The Lancet who saw fit to publish it. Astute observations
may have allowed these researchers to detect early embers
of one of humanity’s great calamities even in the midst of another.
More than a century later, their work, together with that of the
others who provided early clues about the 1918 pandemic, deserves
commemoration.
