Recently there’s been some talk (including in the comments on this blog) of a theory that high doses of aspirin played a key role in exacerbating the number of deaths in the 1918 flu pandemic:

The high case-fatality rate—especially among young adults—during the 1918–1919 influenza pandemic is incompletely understood. Although late deaths showed bacterial pneumonia, early deaths exhibited extremely “wet,” sometimes hemorrhagic lungs. The hypothesis presented herein is that aspirin contributed to the incidence and severity of viral pathology, bacterial infection, and death, because physicians of the day were unaware that the regimens (8.0–31.2 g per day) produce levels associated with hyperventilation and pulmonary edema in 33% and 3% of recipients, respectively. Recently, pulmonary edema was found at autopsy in 46% of 26 salicylate-intoxicated adults. Experimentally, salicylates increase lung fluid and protein levels and impair mucociliary clearance. In 1918, the US Surgeon General, the US Navy, and the Journal of the American Medical Association recommended use of aspirin just before the October death spike. If these recommendations were followed, and if pulmonary edema occurred in 3% of persons, a significant proportion of the deaths may be attributable to aspirin.

I don’t see that it’s especially relevant these days, because no one is recommending doses anywhere near that, and aspirin isn’t as widely used in general even in low doses (except for those who have had cardiovascular problems, who take aspirin in tiny doses to prevent the occurrence of a subsequent event). But putting that aside, I don’t think aspirin was a big deal in the death rate even in 1918, although it might have had some role.

Here’s why I say that:

The international characteristics of the pandemic make the salicylate hypothesis difficult to sustain as the primary explanation for the unusual virulence of the 1918–1919 influenza pandemic. Worldwide, an estimate of the mortality of the 1918–1919 pandemic is 50 million deaths, with a range of up to 100 million deaths. Taking the 50 million figure, this was about 2.5% of the world population. By contrast, in the United States, mortality was on the order of 0.5%. Clearly, the rest of the world was struck more severely, on average, than the United States.

India serves as a useful vignette. Mortality in India was staggering, with estimates of 18.5 million persons dead and higher. Indeed, the Indian peasant population was so severely affected that economics Nobel laureate Theodore W. Schultz used the pandemic as a natural experiment in per capita agricultural output. Given the huge number of deaths in India and the burden among subsistence agricultural workers, it is extremely implausible that salicylates played an exacerbating role in anything other than a trivial percentage of Indian mortality.

Thus, Starko’s intriguing hypothesis fails the test of dose-response. That is to say, in countries such as the United States, where salicylates were more available, mortality was much lower compared with regions where salicylates were less readily available. These observations are at the ecological level, and such comparisons are notoriously susceptible to confounding. However, if the salicylate hypothesis applies universally, then the ecological confounding would have to operate such that the salicylate-influenza connection is stronger in countries with less access to aspirin, which seems a priori unlikely. Indeed, the overwhelming majority of the millions of Indian peasants who were killed by the flu certainly had no access to salicylates whatsoever. If the salicylate hypothesis only works in the United States and in similar settings, then we question its validity given the worldwide scope of severe mortality in 1918–1919.

To summarize: the US had a much lower mortality rate for flu cases in 1918 than a lot of other areas. Of course, as the author of the above quote acknowledges, there may have been (and probably were) many factors that went into that lower mortality. But if in countries where aspirin use was basically nil, the mortality rates were much higher, then it does cut into the likelihood that aspirin was a big factor here, although it doesn’t eliminate it.

One would also have to know how often aspirin actually was used in the US in people who had the 1918 flu. The article speaks of certain spikes in deaths that occurred in waves. But those spikes and waves in which the death tolls were different at different times, even in the same country, are common around the world in flu epidemics. And what’s more, they also occurred in 1918 in other countries where aspirin usage doesn’t seem to have been practiced, much less megadoses of aspirin. It is often part of the natural history of such a disease for its lethality to wax and wane in waves, so attributing that waxing and waning in the US to aspirin usage seems quite iffy to me – especially since we don’t have a good sense of the ways in which aspirin usage waxed and waned in 1918.

And by the way, those mortality figures in the first paragraph of the first quote – 2.5% around the world and .5% in the US – illustrate something I wrote at length about yesterday, which is the confusion that sometimes arises in the reader because mortality percentages can be given in two different ways. To repeat: the figure can represent an illness’ mortality rate as a percentage of the entire population (in a city, a country, or the world). Or, the figure can represent an illness’ mortality rate among those infected (for infectious diseases, of course) or who suffer from it (for a non-infectious disease such as cancer). Quite obviously, the percentages of deaths in the afflicted are significantly higher than in the population as a while. The figures given in the excerpt here are for the percentage of the entire population – that is, 2.5% of the population of the entire world died, and .5% of the population of the US died.

And although the flu was especially bad in India, that country was not even the worst in terms of death rates:

Though other countries lost a higher fraction of their populations—Western Samoa (now Samoa) lost 22 percent, for example, compared to 6 percent in India—because of the larger size of the Indian population, that 6 percent translated into a staggering slew of death. Between 1918 and 1920, an estimated 18 million Indians lost their lives to influenza or its complications, making India the focal point of the disaster in terms of mortality. Asia as a whole experienced some of the highest flu-related death rates in those years, but the story of how the disease ravaged the continent is relatively unknown. The 1918 flu pandemic has been called the “forgotten” pandemic, and ironically the continent that seems to have forgotten it most thoroughly is the one that bore the brunt of it.

And here’s some information about the way that characteristic “wave” pattern I wrote about yesterday operated in 1918, as well as the contagiousness/lethality balance I also mentioned in that earlier post (emphasis mine):

Hunger weakens the immune system, and hunger was rife in many regions of the world in 1918, partly due to disrupted supply lines. Other infectious diseases, such as tuberculosis and typhus, had made inroads into human populations, capitalising on the disruption wrought by war and rendering their victims more vulnerable than usual to a new respiratory infection. Large numbers of people, both troops and refugees, were on the move, providing the ideal vehicle for disseminating that infection. Meanwhile, the very lack of mobility of one group may have helped brew a particularly lethal germ that year, or at least kept it lethal for longer. Once the virus reached the Western Front—the 16-kilometre-wide system of trenches that gashed France from the Belgian to the Swiss border—it encountered large numbers of young men who, packed into those trenches, did not go anywhere for weeks or months. Paul Ewald, an evolutionary biologist at the University of Louisville in Kentucky, has argued that under such exceptional conditions, the evolutionary pressure on the virus to moderate its virulence may have been relieved. It became the mobile one in the host-virus relationship, and it raced through the trenches, killing as it went.

Flu pandemics have a characteristic structure, engulfing the world in waves. The first wave, sometimes called the herald wave, is often quite mild, resembling a seasonal flu. This tends to be followed by a more deadly second wave, and in some cases, subsequent waves of varying severity. The flu pandemic of 1918, though unusually virulent, was no different in this respect. There was a mild herald wave in the northern-hemisphere spring of 1918, a much more lethal second wave in the latter part of that year, and a final recrudescence in the early months of 1919, which was intermediate in severity between the other two. The pattern was repeated in the southern hemisphere, but it was staggered in time with respect to the north, meaning that the waves tended to strike later there. The pandemic is conventionally considered to have been over by March 1920, although earlier this year, the epidemiologist Dennis Shanks and his colleagues at the University of Queensland in Brisbane reported that it dragged on in the Pacific islands for another year, with cases still being reported in New Caledonia in July 1921.

It’s rather complicated, and we’re still learning about it. But if aspirin played much of a role, I think it was a tiny one.