Pale Rider: The Spanish Flu of 1918 and How It Changed the World

by Laura Spinney

  First let’s address the issue of when. The Spanish flu occurred because a virulent viral strain acquired the ability, first to infect humans, and then to become highly transmissible between humans. It was this latter step that triggered the deadly autumn wave, and scientists now monitor strains in circulation in an attempt to predict when they might acquire that capability. One of the techniques they use is based, once again, on molecular clocks. The idea behind it is simple: as mutations accumulate over time, some of them may render a particular strain fitter or less fit than others. Those changes in fitness are reflected in the shape and branching of the flu family tree, because the fitter the strain, the more offspring it has. It is therefore possible, in theory, to predict when a particular strain might attain a level of fitness at which it has pandemic potential.

  Strains may in fact already have emerged that have that potential. These belong to the H5N1 subtype of influenza A–the subtype that killed the toddler in Hong Kong in 1997. Almost all human cases of H5N1 infection to date have been caught directly from birds, but a few have been transmitted between humans, and some fear it is only a matter of time before the virus becomes highly transmissible between people (another strain, H7N9, is under surveillance for the same reason). That hasn’t happened yet, and it may never happen, but since H5N1 also kills 60 per cent of those it infects, it is currently considered to rank among the world’s greatest pandemic threats.

  External factors–notably climate–might affect the timing of a pandemic. A study published in 2013 showed, for example, that prior to the Spanish flu and the three subsequent flu pandemics, the Pacific Ocean was in the La Niña phase of its temperature cycle.3 During La Niña–known as the ‘cold’ phase of the El Niño-Southern Oscillation (ENSO)–the region of the Pacific between the tropics of Cancer and Capricorn cools, while in the opposite phase, El Niño, those same waters warm. Ocean and air currents are linked, since both redistribute heat around the earth’s surface, and this has a knock-on effect on weather patterns around the globe, which is why meteorologists track ENSO so closely. (Could those portents seen prior to the Spanish flu–the withered roses, the owls turning up in new places–have been people’s heightened perceptions of real atmospheric phenomena?)

  El Niño (‘little boy’ in Spanish) episodes occur irregularly, but on average every two to seven years. They are sometimes, but not always, followed by La Niña (‘little girl’). La Niña tends to last longer than El Niño, however–between one and three years, as opposed to less than a year for the little boy–and both tend to coincide with the northern hemisphere winter. Nobody yet knows why La Niña should make a pandemic more likely, but it may have something to do with the effect those changes in air currents have on the paths taken by migratory birds–and hence the populations of domesticated birds with which they come into contact.

  Knowing that the world is about to enter a La Niña phase–as it did in August 2016–could therefore help us to predict the next pandemic, though only as part of a bigger, more complex puzzle. If we understood the relationship between bird migration paths and flu, however, we might also be able to determine how our burning of fossil fuels will impact the timing, and geographical origin, of any future pandemic. We have, after all, now entered the Anthropocene epoch, which is defined by the impact of humanity on earth–the trace that our cars, nuclear weapons and discarded chicken bones have left on the planet. The previous epoch, the Holocene, spanned the 12,000 years since the last ice age and–coincidentally–the farming revolution that marked the beginning of the story of flu as a human disease. In the Anthropocene, we have moved into uncharted territory. As palaeoclimatologist William Ruddiman put it, ‘we humans have now ended the 2.75-million-year history of northern hemisphere ice-age cycles for a time into the future that is beyond imagining’. In 2014, the Audubon Society of America found that birds had migrated further north by an average of sixty-four kilometres over the previous forty-eight years, as temperatures rose.4 Global warming could even be having a direct effect on the flu virus itself–we don’t know, but there are clues that this is the case. Cold, dry conditions tend to favour flu, but some strains currently in circulation appear to be adapting to a warmer world. There have been outbreaks of H5N1 in Asia in the summer months, for example.

  So much for when. What about how big? This is the $64,000 question, because many different factors determine the scale of a flu pandemic. If the strain that caused the Spanish flu were to emerge again today, it would likely cause a mild disease, since our immune systems are more or less primed to it. The danger is that a new strain appears in our midst, to which nobody alive has ever been exposed. Even then, it’s hard to predict what form a pandemic would take, because human beings have also moved on since 1918. The conditions that prevailed on the Western Front, and the massive displacement of people triggered by the First World War, are unlikely to be repeated. On the other hand, the globe is better connected. Transport, of humans and the germs that infect them, is faster, and we have fewer natural sanitary cordons in the form of geographic isolation. Our disease surveillance is better, and we have some effective medicines, including vaccines. But the world’s population has also aged. Though age weakens the immune system, the elderly also have immune ‘memories’ of more varieties of flu, and it’s not clear how those two effects would offset each other.

  In 2013, a company that specialises in catastrophe modelling, AIR Worldwide, tried to take account of all these factors, and came up with an estimate of between 21–33 million deaths worldwide, if a flu as dangerous as the 1918 strain struck. The global population has roughly quadrupled since 1918, so this represents a very much smaller disaster than the Spanish flu, but it’s still a staggering slew of death. It’s also on the low side with respect to some of the other estimates that have been put forward over the years, which range from fewer than a million, to upwards of 100 million dead. Reflecting that huge span, there are those who say there is nothing to fear from a future pandemic, and others who lament how woefully underprepared we are. The former accuse the latter of being alarmist, the latter accuse the former of burying their heads in the sand. The chasm between them illustrates how much we still have to learn about pandemics in general, and about flu pandemics in particular.

  Despite all the uncertainty, there are things we can do to prepare. The 2016 GHRF report called for governments and private and philanthropic bodies to stump up around $4 billion a year for pandemic preparedness, and it recommended that the money be invested in four main areas: a skilled and motivated public health workforce; robust disease surveillance systems; effective laboratory networks; and engagement with communities.

  The Spanish flu and subsequent pandemics demonstrated that, given the right incentives and training, health workers stay at their posts and honour their duty to treat, often at great risk to their personal safety. That workforce therefore needs to be supported as much as possible and cared for in the event of illness. The best way to support them is to arm them with effective methods of surveillance and prophylaxis, and to make sure that they are dealing with an informed, compliant public. All three areas have seen huge advances since 1918, but there is still room for improvement.

  At the moment, disease-surveillance agencies such as the CDC and the WHO take a good week to respond to a pandemic signal in the data. In 2009, two American researchers, Nicholas Christakis and James Fowler, set out to see if they could beat that, by identifying individuals who catch flu precociously in a pandemic, and who therefore act as ‘sensors’ of contagion. In an echo of Ronald Ross’s ‘theory of happenings’, they realised that the way that anything contagious spreads through a population–be it a virus or a meme–depends on the structure of human social networks.

  The key to their approach is something called the ‘friendship paradox’. This is the idea that, on average, your friends have more friends than you have, and it arises because of a bias inherent in the way we count our friends (essentially, popular individuals get counted more often t
han less popular ones because they crop up in more people’s social circles, so they inflate the average against which everyone compares themselves). For practical purposes, the friendship paradox means that if you pick a random person and ask them to nominate a friend, that friend is likely to be better connected than the person who nominated them. During the 2009 swine flu outbreak, Christakis and Fowler tracked infection in two groups–one randomly chosen group of Harvard undergraduates, and a second group whom the first had nominated as friends. They found that the friends fell sick on average two weeks earlier than their randomly picked counterparts–presumably because they were more likely to come into contact with carriers of infection.5

  If you could capture that early spike in flu cases, and mobilise a containment strategy up to two weeks earlier than is possible now, you could potentially save a great many extra lives. A lot of vulnerable people can be vaccinated in two weeks. But there is another way in which those sensors could help limit the impact of a pandemic–or even avert it entirely. If a high enough proportion of a population is vaccinated prior to a pandemic, they may confer what is known as ‘herd immunity’ on the rest. This is because they block the virus’s spread, meaning that the whole population is protected even though not everyone is immune. Christakis and Fowler showed that herd immunity could be achieved by vaccinating a smaller number of sensors than less well-connected people–again, because they were more likely to step into the paths of carriers of infection.

  What about prophylaxis? The annual flu vaccine is improving all the time, but it still has to be updated each year. Since 1973, the WHO has issued annual recommendations regarding which strains should go into it, depending on those that surveillance agencies indicate are currently circulating in the human population. It takes time to manufacture a new flu vaccine, however, so final decisions about the composition tend to be taken around February for a vaccination period that begins in October. And therein lies a problem: if a new strain surfaces between February and October, the vaccine will be partially effective at best. The molecular clocks might help prevent that happening, too, by making it possible to identify strains whose fitness is increasing, though they haven’t yet been detected as a threat.

  Meanwhile, work continues on a so-called ‘universal’ vaccine–one that will protect humans against flu without having to be updated each year. For some time now, vaccines have not made use of the entire flu virus, because exposure to this in the form of a vaccine can cause side effects that are sometimes more unpleasant than the flu itself. To stimulate a response from the human immune system, modern vaccines present it with the round, convoluted head of the H antigen alone. Unfortunately, it is precisely this that changes from year to year, so Taubenberger for one is pursuing an alternative approach.

  During his studies of the Spanish flu, he realised that another part of the H antigen does not change from year to year: the stem. That’s because it has to anchor the head in place, meaning it is subject to certain mechanical constraints. His group, among others, is now focusing on this essential but relatively unchanging component of the virus, in an attempt to develop a vaccine that protects, not only against all the flu strains that have caused pandemics in the past, but potentially also against those that could do so in future.

  In a future flu pandemic, health authorities will introduce containment measures such as quarantine, school closures and prohibitions on mass gatherings. These will be for our collective benefit, so how do we ensure that everyone complies? How, too, do we persuade people to get vaccinated each year, given that herd immunity is the best protection we have against a flu pandemic? Experience has shown that people have a low tolerance for mandatory health measures, and that such measures are most effective when they are voluntary, when they respect and depend on individual choice, and when they avoid the use of police powers. In 2007, the CDC issued guidelines for how to ensure maximum compliance with public health measures in a pandemic. Based partly on lessons learned in 1918, these recommended that measures only be made mandatory when the proportion of the sick who die rises above 1 per cent (remember that this proportion was at least 2.5 per cent for the Spanish flu). Using 2016 numbers, that means that more than 3 million Americans would have to die before the CDC would advise such a step–a measure of how counterproductive that organisation believes compulsion to be.

  But if disease containment works best when people choose freely to comply, then people must be informed about the nature of the disease and the risk it poses. This is one reason why it’s important to tell the story of the Spanish flu. It’s also one of the arguments used to justify films such as Outbreak. Presenting the worst-case scenario, the defenders of such films claim, is the best way to persuade people to get vaccinated and to keep funding scientific research via their taxes and private donations. It’s a controversial strategy, however, not only because of the danger that such films will provoke ‘apocalypse fatigue’, but also because the ability of scientists to predict the worst-case scenario depends on how well they understand the phenomenon in question. H5N1 might still turn out to be as dangerous as the fictional Motaba–we’ll have to wait and see. But in the early twentieth century, eugenics-inspired movies frightened people by showing them the supposed implications of their ill-advised reproductive choices for society–to wit, the proliferation of ‘defectives’–and eugenics has since been thoroughly discredited.

  Whatever the rights and wrongs of such shock tactics, the media clearly have a critical role to play in any future pandemic, and 1918 taught us a valuable lesson in this too: censorship and playing down the danger don’t work; relaying accurate information in an objective and timely fashion does. Information and engagement are not the same thing, however. Even when people have the information they need to contain the disease, they do not necessarily act on it. A few years ago, when the European Commission ordered the destruction of olive trees in the Italian region of Puglia, to prevent the spread of a dangerous plant pathogen, local people protested and challenged the decision in the courts. Olive trees have deep emotional significance in Italy, where families plant them to mark births over generations. The EC had not involved the olive-tree owners in their deliberations, and the owners rejected the scientific arguments it advanced.6 Trust broke down between the two parties–or rather, was never built up. But trust is not something that can be built up quickly. If it is not in place when a pandemic declares itself, then however good the information being circulated, it probably won’t be heeded.

  Another thing 1918 taught is that, sometimes, the reasons why people ignore advice are to be found deep in the past. In this century, President Thabo Mbeki of South Africa denied that AIDS was caused by a virus and appointed a health minister who recommended treating it with garlic, beetroot and lemon juice. Soon enough, AIDS patients were dying on the front lawns of hospitals because they were unable to get effective treatment. Mbeki’s behaviour seems impossible to understand, until you set it in the context of a long history of whites blaming blacks for disease in his country. The consequences of that blame have often been brutal and long-lasting for black South Africans, as they were in 1918. The pandemic forced action on an issue that had been under discussion for a decade: segregation of towns along colour lines. In 1923, the Natives (Urban Areas) Act was passed, and it wouldn’t be repealed for another sixty years.

  Through such secondary tragedies the Spanish flu cast a long shadow over humanity. Some of those tragedies could not have been avoided by good disease surveillance or a vaccine, but others could–the surge in post-viral depression, the creation of large numbers of orphans, the damaging of the life chances of the generation in the womb. That so much suffering is now preventable is a testament to the fact that Roscoe Vaughan, an anonymous woman in an Alaskan grave, and others whose tissues allowed Taubenberger and Reid to sequence the flu genome, did not lose their lives in vain. But we shouldn’t rest on our laurels, because the story doesn’t end here. Once people thought that flu was caused by the pull of dist
ant stars. Then they realised that something very small penetrated the body and made it sick. Finally, they understood that influenza is the product of an interaction between a host and an agent of disease. Over the centuries, humans came to perceive flu as an increasingly intimate dance with the Devil, and even as they add to their knowledge, man and microbe continue to shape one another.

  AFTERWORD: On Memory

  Sincerely Yours, Woodrow Wilson. Arthur Mole designed the portrait and photographed the scene after his partner, John Thomas, arranged the 21,000 soldiers on parade grounds at the US Army’s Camp Sherman, 1918.

  Whenever anyone asked Samuel about his parents, he would tell them that they had died of the Spanish flu. And if someone replied that this was quite impossible, since the Spanish flu epidemic had reached Brazil at the beginning of the twentieth century, he would respond: ‘Well, maybe it was Asian flu; I didn’t ask it for its passport.’

  Elias Canetti, Party in the Blitz

  Arthur Mole was a man of unusual vision. During the First World War, armed with a white flag and a megaphone, he choreographed tens of thousands of US soldiers into what he called ‘living photographs’. If you looked at the mass of men from the ground, or from directly above, they looked like a mass of men. But if you stood at the top of a twenty-five-metre-high viewing tower placed a certain distance away, you saw that they formed a patriotic image: the Statue of Liberty, Uncle Sam, the head of President Wilson.