The Chinese theory stood alone for a long time, but then in this century, two rival theories were proposed. According to one of them, patient zero or the index case–the first person to contract the ‘Spanish’ flu–did not fall ill in China, or even in the silent spaces of the Eurasian steppe, but a short train ride from the Western Front, in the heart of the European theatre of war.7
Between 1916 and the end of the war, Britain delivered a million or more fighting men to the Western Front–the sixteen-kilometre-wide system of trenches that gashed France from the Belgian to the Swiss border–but this feat presented them with certain logistical challenges. While the French, Germans and Russians had thousands of square kilometres in which to billet their reinforcements, stock their supplies and tend their sick and wounded, the British had to squeeze their entire support operation into the narrow strip of land between the front and the Atlantic Ocean. The solution they came up with was to build a camp at Étaples, a small fishing port just south of Boulogne-sur-Mer.
You can still see vestiges of the camp at Étaples. It starts at the northern edge of the town and runs up the coast–tens of square kilometres of land in which the remains of ammunition dumps occasionally break the surface. If you had flown over it in a military plane in 1916, you would have looked down on the River Canche, which flows into the English Channel at Étaples, and perhaps spied recruits being drilled in the wide dunes around it, or little huddles of deserters hiding out there. Heading north, you would have passed over the ‘Bull Ring’–the notorious exercise ground where men were pushed so hard they mutinied in 1917–shooting ranges, detention camps, and above all, row upon monotonous row of wooden barracks. Finally you would have come to the camp’s northern perimeter, and been impressed, or depressed, by the sight of a dozen hospitals lined up along it. Between them these boasted 23,000 beds, making Étaples one of the largest hospital complexes in the world at the time.
On any one day, this sprawling, makeshift city accommodated 100,000 men and women. Reinforcements arrived daily from the four corners of the British Empire, and nearby were camps for German POWs and French troops from Indochina. Fifty kilometres to the south, at Noyelles-sur-Mer, near the Somme estuary, the CLC had its headquarters and a hospital of its own (Number Three Native Labour General Hospital, to give it its correct name). In all, around 2 million human beings were camped out in this small corner of northern France. By 1916, Étaples had become an overcrowded holding pen for men who knew they were about to die. The British poet Wilfred Owen, who passed through it, described the ‘strange look’ peculiar to the camp, in a letter to his mother: ‘It was not despair, or terror, it was more terrible than terror, for it was a blindfold look, and without expression, like a dead rabbit’s.’8
Between July and November of 1916, during the Battle of the Somme, up to ten ambulance trains a night arrived at Étaples. Many of the wounded had been exposed to mustard gas, which causes the lungs to blister. In December–a whole year before Shansi’s bout of winter sickness, that is–something very like flu broke out at the camp. By the time the weather turned cold at the end of January, it had reached the proportions of a small epidemic, and it receded with the frost in March. A trio of British Army doctors led by Lieutenant J. A. B. Hammond described it in the Lancet medical journal in July 1917. They called it ‘purulent bronchitis’ and noted that it was characterised by a dusky blue hue to the face. They performed autopsies on some of the victims and found their lungs to be congested and inflamed–a signature, too, of the Spanish flu.9
Was purulent bronchitis a precursor of the Spanish flu? A British virologist called John Oxford thinks it was, and thanks to the assiduous record-keeping of military doctors during the First World War, he has built a persuasive case. A historian with whom he has worked, Douglas Gill, has studied the death records for British military hospitals in the French city of Rouen–a centre of hospitalisation that was almost as important as Étaples–and found that an epidemic passed through there too, at around the same time. An almost identical disease broke out in barracks at Aldershot, England, in early 1917.10
There is a problem with the Étaples theory, though: there are no records of outbreaks in the civilian population of northern France at that time. It seems odd that a dangerous infectious disease would erupt simultaneously at a number of military bases, while the civilian communities between them remained unaffected–especially since we know that the camp at Étaples lived ‘in osmosis’ with the town.11 British soldiers ‘fraternised’ with local women, and frequented the town’s shops, bars and brothels (the lady whose favours were most sought after called herself ‘the Countess’). But there may be a simple explanation for that: under the French civilian system in operation at the time, to protect individuals’ privacy, the cause of death was recorded separately from the announcement of that death. Though the public death registers have survived, often the doctors’ certificates mentioning the cause of death have not. There may have been outbreaks among civilians, in other words, but if there were, no record of them survives.12
Hammond produced a detailed description of purulent bronchitis, but he was no better equipped than Wu to isolate a virus, so the Étaples theory, too, remains conjecture. Having proposed such an early herald event, the onus is also on Oxford to explain why the pandemic proper took so long to erupt. His suggestion is that, although conditions in northern France were highly conducive to the emergence of a new pandemic flu strain in 1916, paradoxically, they also contained it. Travel was limited to the round trip from base to front and–if you were lucky–back again, or at most to a short hop across the Channel. In the year or more’s interval between the outbreak described by Hammond and the first recognised wave of the pandemic–in spring 1918–the virus may have maintained itself in small, localised epidemics while it acquired the molecular changes that would render it highly transmissible between humans.
What if the 1918 pandemic started not in China, or in France, but further west again–just down the road from the first recorded case? The third theory suggests that patient zero was not a gassed soldier recuperating at Étaples, nor a peasant farmer labouring among the cliffs and ravines of Shansi, but a peasant farmer labouring close to the geographical heart of America–in the ‘Sunflower State’ of Kansas.
Camp Funston drew recruits from a catchment area that included Haskell County, 500 kilometres to the east. Haskell was one of the poorest counties in Kansas at that time. Its inhabitants lived in sod houses, grew corn and raised poultry and hogs. In January 1918, they began to fall sick, and some went on to develop pneumonia and die. A local doctor, Loring Miner, was so alarmed by the severity of the outbreak that he reported it to the US Public Health Service, even though flu wasn’t a reportable disease in the US at that time. The epidemic receded in mid-March, and nobody might have given it another thought–besides the grieving inhabitants of Haskell County–except that by then, the infirmary at Camp Funston had been overrun by sick soldiers.
On the same day that the camp’s chief medical officer sent a telegram to the authorities in Washington DC about his outbreak, 30 March, a report of the earlier one in Haskell appeared in the public health service’s weekly journal. Almost nine decades passed before an American journalist, John Barry, suggested that the two might have been linked–that a young man hailing from Haskell, probably a God-fearing boy who had grown up on a farm and known no other life, unwittingly carried the virus into the midst of the American war machine, whence it was exported to the rest of the world.13
When you try to chart the progress of the pandemic’s spring wave from the first case in Camp Funston, eastwards to France, it seems at first pleasingly linear and one-directional. Then, however, you remember that large numbers of CLC labourers were being moved across North America that spring, in specially guarded trains. Though we have no reason to believe that they had any contact with the populations they passed through, it isn’t out of the question that a guard’s attention lapsed momentarily, or that he took pity on
a poor passenger and allowed him out to stretch his legs. His instructions were to keep the labourers moving eastwards as discretely as possible; he did not realise he was also defending a sanitary cordon. By April 1918, China was in the grip of yet another flu-like disease–an apparently new epidemic that nevertheless overlapped in time with the one that had started in Shansi the previous winter.14 This new epidemic, according to the consensus of the Chinese medical community, was definitely winter sickness and not plague. It wasn’t fatal and generally passed in four days (Wu disagreed; he was convinced it was the same disease that had broken out in Shansi, and that both were plague, but he was in a very small minority). The possibility exists, therefore, that it was the CLC that brought the flu to the eastern seaboard of North America. To confuse matters still further, there is evidence that New Yorkers were falling sick from late February 1918, before Gitchell checked himself into the infirmary at Camp Funston, prompting some to suggest that New York City received the infection from troops returning from France.
For the time being, therefore, all three theories of the origin of the ‘Spanish’ flu remain on the table. To choose between them would require a comparison of the flu strains that caused the putative precursor events, with the strain that was circulating in the autumn of 1918–something that hasn’t yet been possible. In the twenty-first century, scientists have produced a new kind of evidence that points to one of the theories being more likely than the other two–we’ll come to that–but this evidence, though tantalising, is not definitive. In 2017, therefore, there is only one thing we can say with something close to certainty: the Spanish flu did not start in Spain.
Note, for now, that if the Chinese-origin theory is correct, the pandemic cannot strictly be described as a product of the war. Patient zero was a poor farmer living in a remote village in the Chinese interior, who at the time that he fell ill was doing much the same thing his ancestors had done for generations before him, and who may not even have been aware that there was a war on. The same is true if it started on a farm in Kansas. Only if the origin was French can the pandemic truly be described as an outcome of the conflict, because in that case it was brewed in a camp where men were brought together (with some women) for the express purpose of killing other men. There is one final possibility: that none of the three theories is correct, and the real origin of the pandemic has yet to be proposed.
12
Counting the dead
How many had died? People wanted to know from the moment it was over, not only to gauge the pandemic’s impact on humanity, and to set the historical record straight, but also to extract lessons from it for the future. They had an idea of the scale of the previous flu pandemic, the Russian flu of the 1890s. It had killed around a million people. If the Spanish flu were in that ballpark, then perhaps a flu pandemic was simply something that happened periodically, and one had to learn how to manage it. If it were much larger, however, the conclusion would have to be different: something about that particular flu, or about the state of the world in 1918, or both, had created a deadly anomaly.
In the 1920s, an American bacteriologist named Edwin Jordan estimated that 21.6 million people had died from the Spanish flu. Right from the beginning, therefore, it was clear that it was in a league of its own. This was higher than the death toll of the First World War, and twenty times higher than the death toll of the Russian flu. We now know that Jordan’s figure was an underestimate, but it was one that stuck for close to seventy years, meaning that for a long time after the event, the human species had only a tiny inkling of its loss.
Jordan can be forgiven. Epidemiology was young in 1920. Diagnostic criteria for influenza and pneumonia were vague, and many countries didn’t count deaths in peacetime, let alone in the midst of a boundary-shifting, chaos-generating war. Where data were available, he could calculate excess mortality rates–a measure of the number of people who died over and above what might have been expected in a ‘normal’ or non-pandemic year–but these hid a multitude of diagnostic sins. There was no such thing as a ‘laboratory confirmed death’ from flu in 1918, because nobody knew that flu was caused by a virus. What’s more, flu pandemics don’t really start or stop. They invade the seasonal flu cycle, grotesquely distorting its morbidity (sickness) and mortality (death) curves, then recede until those curves reveal themselves again. Even now that the tools exist to differentiate seasonal and pandemic strains, defining a pandemic’s limits is an essentially arbitrary task.
In 1991, two American epidemiologists, David Patterson and Gerald Pyle, raised Jordan’s bid to 30 million–hinting at a bigger disaster, though still not one on the scale of the Second World War, which eliminated roughly twice as many souls. They incorporated new data that had come to light since Jordan’s day, but they only counted the death toll from the second, autumn wave. There were some areas of the world for which they had no better data than Jordan. They echoed his estimate of 450,000 dead for Russia, for example, along with his caveat that it was no more than ‘a shot in the dark’. ‘Little is known about the toll in China,’ they wrote, ‘but with some 400–475 million inhabitants the loss of life could have been enormous.’1 Russia and China were big, populous countries. Errors in the calculation of their death tolls would have a serious impact on any global tally, so it is worth examining Patterson and Pyle’s estimates for them in a little more detail.
The estimate of 450,000 deaths corresponds to roughly 0.2 per cent of the Russian population at that time. If that were correct, then Russia suffered the lowest flu-related mortality in Europe, which seems counter-intuitive in a country in the grip of a civil war, where the infrastructure of daily life had broken down. The Odessan case suggests, indeed, that it was not correct, and that the real number may have been higher. We know that Odessans were often infected with more than one disease at a time, and that the chances of misdiagnosis were high. Tizengausen, the pathologist at the Old City Hospital, found lung haemorrhage, a telltale sign of Spanish flu, in a larger number of corpses than had been diagnosed with it while alive. Tizengausen had a second job at the city morgue, and there he found the same signs in cases that had been wrongly diagnosed as cholera or, more vaguely, ‘plague’. He also discovered that some of those who had been correctly diagnosed with Spanish flu had been infected simultaneously with typhoid, dysentery, TB and other serious diseases.
Vyacheslav Stefansky, Yakov Bardakh’s former student who also worked at the Old City Hospital, noted that around 8 per cent of flu patients who were admitted to his hospital went on to die of the disease, and another doctor recorded a similar proportion at the Jewish Hospital. This compares to a global case fatality rate of 2.5 per cent.2 In the 1950s, a team of Russian epidemiologists led by V. M. Zhdanov estimated that 70,000 Odessans were sick with the Spanish flu in October 1918.3 If they were right, and if the case fatality rates calculated by Stefansky and his colleague at the Jewish Hospital were right, then around 6,000 Odessans died of ispanka in that month. That equates to 1.2 per cent of the population, or six times Patterson’s and Pyle’s estimated mortality rate for the country and the autumn wave as a whole.
Zhdanov felt that Odessa had suffered worse than any other major Russian city, so if Russia had been composed exclusively of cities, then we might have to revise that figure downwards. But Russia wasn’t composed exclusively of cities, of course. Urban folk were, in fact, very much in the minority, accounting for somewhere between 10 and 20 per cent of the population. And if the flu was bad in Odessa, it probably wasn’t any better in the surrounding countryside, where it was not uncommon for tens of thousands of people to depend on a single doctor with no drugs at his disposal. As we’ve seen, drugs didn’t work. But doctors themselves–and more importantly, nurses–could make a difference, and they were signally lacking. When the International Committee of the Red Cross sent French officer Ernest Léderrey to inspect the sanitary situation in Ukraine in 1919, he reported that some villages had lost 10 to 15 per cent of their inhabitants to typhus and Spa
nish flu the previous year, and dysentery had added to their woes (doctors noticed that Spanish flu often finished off the starving). With the onset of winter, what was left of the zemstvos–pre-revolutionary provincial councils–had tried to help by setting up temporary hospitals. ‘But what are fifty or sixty beds, when every house has at least one invalid who should be isolated,’ wrote Léderrey, ‘A drop in the ocean!’4 If we apply the 1.2 per cent mortality rate to the country as a whole, 2.7 million Russians died of Spanish flu.
China remains a conundrum, mainly because it is impossible to define the Chinese epidemic. In a country where a year did not pass without an epidemic, it is possible that Spanish flu came sandwiched between two visitations of pneumonic plague, that all three waves of respiratory disease–those of December 1917, October 1918 and December 1918–were caused by the influenza virus, or that another, as yet unidentified microbe was responsible for one or more of them.
America and Britain, wealthy countries, lost approximately 0.5 per cent of their populations to the Spanish flu. Extrapolating from poorer countries, but assuming that China suffered less badly than India (where the rate was ten times that in America), Patterson and Pyle came up with a range of between 4 million and 9.5 million deaths in China. But they had no Chinese data to work from, because there was no centralised collection of health data in China during the warlord period, and the missionaries who rode to the rescue of the ailing did not gather statistics systematically. The only parts of the country where some health statistics were gathered as a matter of routine were those that were under foreign control, and in 1998, a Japanese scholar, Wataru Iijima, made use of these to come up with a new estimate. Basing his calculations on foreign-controlled Hong Kong and southern Manchuria, and with many caveats, he estimated that only a million Chinese people died.5
Pale Rider: The Spanish Flu of 1918 and How It Changed the World Page 16