The term ‘Spanish Flu’ was coined because it was in Spain that the first reports of an epidemic during World War I were published. This happened only because Spain was a neutral country that did not censor its press, whereas other war-torn countries suppressed sensitive reports about disease outbreaks in the trenches (Spinney 2017: 63). At the end of World War I on 11 November 1918, no-one suspected that the death toll would continue to climb through the end of the year. Nor did anyone imagine that over the next three years, what is now called the 1918–20 influenza pandemic would kill three to five times more than the 19 million who had died from battle wounds during the four years of war. Estimates of global deaths due to that pandemic range from a conservative 50 million to double that number (Spinney 2017: 170–71).

Whereas a normal influenza outbreak will kill about 1 per cent of those exposed, the 1918 virus killed 2.5 to 5 per cent. Nearly half the world population was exposed to it and an estimated one-third was infected. The pandemic struck in three waves: the first outbreaks in the spring (March–June) of 1918; the period when it was most lethal, during the late summer and autumn (August–November) of that year, as people were returning home from the war arena; and the long dénouement through the winter of 1918 and spring of 1919 to 1921, when it ended in the Pacific Islands (Spinney 2017: 37–45; CDC 2018). The death toll by percentage of population varied widely globally, from the lowest in Odessa, Russia, of 1.2 per cent, to one of the highest, 40 per cent, in Bristol Bay, Alaska, for reasons that have not yet been fully assessed.

The lack of effective therapeutic interventions beyond good nursing care for the highly contagious people infected by this virulent influenza virus helps explain this catastrophic global loss of life. Also astonishing is the 50-million-people range of uncertain deaths: the virus was so infectious and deadly that healthcare infrastructure was challenged globally to a degree that people could barely deal with the sick, much less count the dead. Weariness from the war, moreover, contributed to people wishing to move on with life rather than dwell on the numbers who had died from influenza after the war. Furthermore, the first synthetic histories of influenza were focused on the United States (Crosby 1989) and Europe (Barry 2004) rather than fully accounting for the global reach of the virus. Fortunately, some scholarship has been addressing this problem over the past two decades (Killingray and Phillips 2003; Porras-Gallo and Davis 2014; Spinney 2017; Beiner 2022), including what happened in Japan (Hayami 2015).

Despite mainland China being at the centre of influenza-origin narratives since the 1918–20 pandemic itself, scholarship on what happened in the country at that time has only scratched the surface. The issue has been addressed in several English-language review essays (Iijima 2003; Langford 2005; Cheng and Leung 2007) but the authors largely relied on Western-language sources, such as contemporary reports of the Chinese Maritime Customs Service and articles in the China Medical Journal. More recently, one medical historian mined primary Chinese sources on influenza outbreaks in the entry-point cities of Beijing, Tianjin, and Shanghai in 1918 (Pi 2016a), as well as outbreaks across the country in 1919–20 (Pi 2016b). Another scholar made an exhaustive compilation of the historical records on influenza in the major newspapers of the time, Shenbao and Ta Kung Pao (Wu 2020, 2021). Their findings, however, were published in Chinese-language journals, which limits their reach beyond China. This essay seeks to examine China’s role in the influenza-origin debate, address why naming ‘influenza’ in 1918–20 in China remains problematic, and presents some evidence of how Chinese medical practitioners treated the virus. The resulting review of secondary scholarship hopefully will establish more solid ground from which further research can be launched and a more complete history can be written that does justice to the full complexity of the 1918–20 influenza pandemic in China.

Debates about the Origin of the Influenza Pandemic

There are four hypotheses for the origins of the 1918 influenza pandemic: in 1917 in British Army hospitals in France and England; in 1917 in Shanxi Province in mainland China; in a US military fort in Kansas in early 1918; or during the Russian Flu of 1889–94.

Since 1918, the creation of new scientific disciplines—such as virology, genetics, microbiology, and paleomicrobiology—has changed how we understand the identity of influenza. The isolation of human Influenza A in 1933 and the completion in 2005 of the genomic sequence of the 1918 H1N1 virus that caused the influenza pandemic have radically changed scientific understanding of the origin of the disease (Taubenberger et al. 2007). In the past decade, scientists have argued that the 1918 H1N1 virus most likely originated in North America (Worobey et al. 2014) or Europe (Oxford and Gill 2019). British virologist John Oxford (2001; Oxford and Gill 2019) posits that influenza originated in early 1917 in the war hospitals of the British Army—first, in December 1916 in Étaples, France, and then in Aldershot, England. Supporting evidence for this position comes from British medical officers who published two Lancet articles about their experiences treating what they called ‘purulent bronchitis’ in 1917. They considered it a new disease and described it as a small epidemic within the war context (Hammond et al. 1917; Abrahams et al. 1917). Historians Alfred Crosby (1989), David Patterson and Gerald Pyle (1991), and John Barry (2004), in contrast, argue that influenza started later, in March 1918, when a major epidemic occurred at Camp Funston in Fort Riley, Kansas.

However, the China-origin narrative for the 1918 influenza remains strong, reinforced retrospectively by each new influenza outbreak that has since originated in China, including the 1957 Asian Flu and 1968 Hong Kong Flu (Peckham 2022: 272). There has also been a longstanding argument that it was Chinese labourers who brought the disease from China, first, to North America, and then to France while on their way to engage in non-combat work for the British and French (Patterson and Pyle 1991: 8n12).

At the time, a Quaker physician contended that the pandemic influenza could be a modified form of a pneumonic plague that had ravaged Shanxi Province in 1917 and was brought to the arena of World War I by labourers from northern China (Cadbury 1920; Peckham 2022: 269). The China-origin argument thus dates from the period itself. Some historians have rejected this argument, pointing to the lack of evidence, and suggesting it was based on German propaganda (Patterson and Pyle 1991: 8). However, more recently, Japanese medical historian Wataru Iijima (2003) reiterated the hypothesis that the virus could have begun the previous year in 1917 Shanxi Province. Medical historian Cao Shuji (2006) instead argued that what was called the ‘pneumonic plague’ in Shanxi Province in 1917–18 (Wu 2014: 109) was complicated by the later presence of influenza. As mentioned above, some have even argued that the origins of influenza can be traced back to the 1889–94 Russian Flu pandemic (Peckham 2022: 268). Given new medical knowledge about coronaviruses, the Russia-origin hypothesis for influenza has been refuted (Berche 2022).

Some scholars have argued that, despite China’s poor population health and healthcare infrastructure, evidence of widespread but less lethal influenza in China in 1918–19 (Cheng and Leung 2007) supports the China-origin position because the lower lethality could be the consequence of greater immunity. This is more a hypothetical interpretation than a conclusive demonstration of what happened. Contemporary Chinese sources estimate that only 1.4 per cent of the Chinese population died from the influenza pandemic (Pi 2016b: 120n7). Although this figure is debatable due to scant and scattered primary sources, contemporary coverage within and outside China surmised that the Chinese did not experience the influenza pandemic as badly as elsewhere (Patterson and Pyle 1991: 8n11; Pi 2017: 125n43).

Christopher Langford (2005) took a middle position between the France-origin and China-origin arguments, suggesting a less virulent form of H1N1 could have played one of two roles in the Chinese epidemiological theatre: 1) it immunised the Chinese population before an antigenic drift (mutation) that occurred during World War I caused a more virulent version to return to China from Europe; or 2) it travelled with Chinese Labour Corps workers to France any time after mid 1916 and then underwent an antigenic drift in Étaples, France. As Oxford (2019 and Gill) first suggested, however, it is also possible that an antigenic shift (that is, a reassortment of viral components) created a completely new virus in France. According to Langford (2005: 494–95), only virological data can resolve these three possibilities.

In the past decade, Canadian historian Mark Humphries (2014) has revived the China-origin position by arguing that new records he found of the secret mobilisation of 94,000 Chinese labourers across Canada to work behind British and French lines starting in 1917 proved the disease’s Chinese origin. Shortly afterwards, Dennis Shanks (2016) refuted this with evidence from mortality lists of Chinese and Southeast Asian labourers and soldiers that demonstrate that these people died from influenza after other military units, so could not have been the primary source of the lethal form of influenza in Europe in 1918. Chinese medical historian Pi Kuoli (2016a, 2016b) systematically revisited the full range of primary sources on influenza in China. Some people at the time argued that influenza in China could have been the result of returning Chinese volunteer labourers and students from Europe as their waves of return overlapped in time and space with the outbreak of Spanish Flu in northern China (Pi 2016a: 61).

Although some historians have concluded that the existing textual evidence remains inconclusive (Spinney 2017), some scientists have used a ‘host-specific molecular clock approach’ to place the origins of the H1N1 influenza virus in the Western Hemisphere, in either North America (Worobey et al. 2014) or France (Oxford and Gill 2019).

The Trouble with Chinese Sources

In the era before bacteriology, microbiology, and virology, it was challenging for physicians and medical scientists alike to diagnose influenza with certainty. This was due to several problems: 1) the signs and symptoms were not specific; 2) the disease coexisted with other upper respiratory infections; 3) the outbreaks were sporadic and unpredictable; and 4) when mortality statistics first began to be gathered in developed nations in the mid 1800s, they showed high mortality only in the first few years of an epidemic, which supported the interpretation that influenza did not occur between epidemics (Taubenberger et al. 2007: 582).

In the context of China, due to scarce and uneven primary sources, it is even more difficult to establish with certainty how severe the influenza pandemic was across the country and, by extension, to compare China’s experience with that in other regions of the world. In addition to the unreliability of the numbers of those infected and deaths from influenza, there was no consensus in China at the time on how to define and diagnose influenza. Nonetheless, it is possible to find records on what Chinese physicians thought was influenza.

First, various syndromes within traditional Chinese medicine could indicate the main symptoms of influenza. Not one of them, however, covered all the modern criteria for symptoms (much less its aetiology) to such an extent that it is possible to identify with any certainty that the H1N1 influenza virus was involved. Moreover, in early twentieth-century China, the arguments for a climatic cause for influenza-like epidemics were deeply formalised in traditional medicine, adding to this complexity (Pi 2016a: 64n28; Hou 1920).

To make things worse, China was in great turmoil because of civil wars. There were no authorities who could effectively conduct the necessary public health investigations or were familiar with the medical infrastructure. Local governments and private organisations collected records of possible cases but only major cities benefited from this data-gathering (Zhang 2008: 42–49). The political turmoil prevented the outside world gaining a complete understanding of the situation in the country and foreign hypotheses about influenza infections and mortality rates in China were therefore often underestimates and incorrect.

However, clues could still be buried in local documents and collective memory. In 1918, memories of the 1910–11 Manchurian plague that killed roughly 60,000 people were still fresh in China and some have argued that Chinese people easily took cases of pneumonia and influenza to be another instance of pneumonic plague. After successfully leading the effort to control the 1911 Manchurian plague epidemic, the first ethnically Chinese medical doctor to graduate from Cambridge University, Wu Lien-teh (1879–1960), spent seven years running the new Manchurian Plague Prevention Bureau. Wu concluded that one of the origins of the now famous Shanxi influenza outbreak of 1917–18 was in fact pneumonic plague (Wu 2014: 104–9). If, as Wu wrote, only one of 16,000 people infected survived, pneumonic plague was a far more likely diagnosis than a highly contagious but less fatal influenza.

As noted, Cao Shuji (2006) recently concurred with Wu’s pneumonic plague diagnosis. However, he added the caveat that by late 1918 and through 1919, at least, influenza was also present in Shanxi Province. The lack of a clear definition of influenza distinct from other respiratory infections in traditional Chinese medicine, or for that matter in Western medicine, contributed to this lack of clarity, which in turn influenced the accuracy of mortality rates. The immediate fear of a relapsing pneumonic plague epidemic comparable with what happened in Manchuria in the early 1910s was also a reason behind the conflation of pneumonic plague with influenza.

Naming the Disease

For centuries, Chinese physicians possibly treated the common cold and influenza as the same syndrome due to their similar symptoms and course (Zeng 1998). As traditional Chinese medicine was the most common and popular therapy available when new epidemics like pneumonic plague, cholera, and influenza struck China in the early twentieth century, most medical practitioners and local authorities classified unknown epidemics using pre-existing Chinese knowledge of such sicknesses (Ji 2014).

At the time, for example, the Chinese-language press and journals variously referred to the surge of influenza-like cases with Chinese medical terms such as the ‘winter epidemic’ (冬疫) and the ‘infectious cold’ (流行性感冒).

One 1918 entry on ‘Influenza (?) in China’ in The China Medical Journal, the leading English-language medical journal in China at the time, notably featured a question mark and provided the multiple names given to the epidemic in different places across the country (Anonymous 1918: 608). These included a ‘five-day plague’ that was compared to cholera and a ‘severe and often fatal pneumonia’ that had all the symptoms of ‘typhoid fever’ or resembled ‘dengue fever’ and was also called ‘influenza’.

Some people in Japan called it the ‘malignant infection’ (惡性流感), emphasising its infectiousness with the characters for ‘transmission’ (流) and ‘infection’ (感). In China others called it the ‘malignant cold’ (惡性感冒), also emphasising its ‘malignant quality’ (惡性) and thus suggesting it was often fatal, but maintaining the conventional term ganmao (感冒) for the ‘common cold’ (Wang and Xu 2012: 71). Contemporary observers recognised the epidemiological complexity of simultaneous epidemics from different causes, so clearly differentiated the nearly 100 per cent fatality rate of the pneumonic plague from the ‘malignant’ but survivable common cold (Wang and Xu 2012). Early on, some observers even noted that Chinese mortality rates were lower than those in the United States and Europe (Alita et al. 2010).

More recently, other scholars have argued that this disparity in severity was because the Chinese population as an aggregate had developed immunity through previous exposure (Langford 2005). This hypothesis is linked to evidence from the severe outbreak of another influenza-like epidemic, in 1889 in Manchuria. According to contemporary understanding, this influenza erupted in China and was, because of its endemic origins, later transmitted to Europe. It subsequently re-entered China in 1892 from the Russian side of the border (Quan 1918). In fact, claims about Chinese immunity to influenza are often made in the literature to explain this discrepancy in fatality rates. Yet other scholars have argued that it was the then dominant use of traditional herbal medical remedies that played a role in lower mortality rates (Cheng and Leung 2007). A century later, we are unable to prove these hypotheses on population immunity and the efficacy of traditional therapies. It is nonetheless worthwhile to explore what the Chinese primary sources record about what Chinese physicians called influenza-like symptoms, at a time when modern germ theory was not well known, and the therapies they used to treat it.

Treatments

Acupuncture and massage were used as preventatives, but herbal medicine was the dominant therapy (Anonymous 1919). At least three types of herbal formulas were used to treat influenza-like symptoms at the time (Alita et al. 2010). First, the classic ‘cold-damage’ formula for clearing heat in the lungs was prescribed during the 1918 influenza pandemic in China (Cheng and Leung 2007). The formula’s Chinese name refers to its four primary ingredients—ephedra, apricot kernel, liquorice, and gypsum decoction (麻黃杏仁石膏湯)—and was first recorded in the classic Treatise on Cold Damage by Zhang Ji (circa second century CE).

Other formulas, such as the ‘mulberry chrysanthemum drink’ (桑菊飲) and ‘honeysuckle and forsythia pills’ (銀翹散, Yin Qiao San), may sound more like floral herbal teas but were considered frontline defences against the ‘malignant cold’ (Alita et al. 2010; Spinney 2017: 125). During the severe acute respiratory syndrome (SARS) pandemic of the early 2000s, some integrated-medicine physicians in mainland China used Yin Qiao San to treat symptoms (Hanson 2010). The ‘mulberry chrysanthemum’ formula was used for mild and relatively exterior cases that presented with predominantly a sore throat and symptoms that were still minor. The ‘honeysuckle and forsythia’ pills were for more severe cases that were still largely limited to the exterior of the body but manifested with a significant fever. Some contemporary practitioners of Chinese medicine refer to the latter as a form of ‘herbal antibiotics’. The Qing-era physician Wu Jutong (1758–1836) recorded both formulas in his 1813 Treatise on the Differentiation and Treatment of Warm Diseases (溫病條編) as treatments for clearing heat (that is, anti-febrile and anti-inflammatory) and overall detoxification (that is, anti-viral).

As the traditional Chinese term ganmao could apply at the time to both the common cold and influenza—two very different manifestations of lung-afflicting illnesses—further explanation is warranted. The one-to-one translation of ‘ganmao’ as ‘common cold’ is at best naive and at worse misleading. Chinese medicine distinguished between two types of what today is covered by the capacious English phrase ‘the common cold’: on one side was a type of ganmao due to wind-cold (the emphasis was on the external wind as a causative agent); on the other was a type due to wind-heat (with the emphasis on a wind that had penetrated further into the lungs).

The conventional cold-damage formula cited above – ephedra, apricot kernel, liquorice, and gypsum decoction – was used to treat wind-cold type of ganmao. Wu Jutong’s new ‘mulberry chrysanthemum drink’ and ‘honeysuckle and forsythia pills’ were used to treat the wind-heat type of ganmao and had a different kind of clinical logic (Bensky, et al., 2007, 34-38). Most of the formulas cited by the sources of the period focused on treating either a wind-cold or wind-heat type of ganmao.

Thus, one could argue that Chinese physicians did not recognise the Spanish Flu as a distinct disease because they did not separate its clinical presentation in patients from other symptoms of the common cold. Due to the different understanding of aetiology, pathology, and nosology (that is, disease classification), Chinese doctors understood this ‘malignant cold’ type of ganmao as a variation of a known ‘cold damage’ (傷寒). That is, they did not see this as a distinctly new disease, but rather as one that fit within an existing cold-damage framework. However, many Chinese doctors in the late 1910s and 1920s also declared that influenzas should be called ‘seasonal epidemics’ (時疫). This was a concept that emerged with new ideas about ‘warm diseases’ (溫病) as distinct from the older concept of cold-damage patterns of illness tied to climatic factors (Cao 1987: 747–48; Hanson 2011).

Although experts in Western medicine did not accept these Chinese medical interpretations, they were also confused by the variety of symptoms of influenza at that time. Another explanation for this confusion could be that, in the 1910s, several epidemics occurred in China, ranging from pneumonic plague and influenza-like viruses in 1918 (Watson 1919) to cholera in 1920. Typhoid fever was also possibly present. Furthermore, for those who had some Western medical knowledge, epidemic influenza was not only a serious potential threat for China but also a chance to reveal what they considered to be the greater capability of Western medicine to understand it, even though Western practice did not then have any better ways to treat it than did Chinese medicine.

Blaming, Naming, and Treating

An understanding of the full range of definitions of the pandemic influenza in China from 1918 to 1920 illuminates how infectious diseases were being redefined in Chinese culture and among practitioners of Chinese medicine. This case is thus a powerful lens to reveal significant gaps between governmental policies on infectious diseases and social responses to them as well as between what medical elites suggested and how laymen acted.

In sum, this essay has addressed the three broad themes of ‘blaming, naming, and treating’ in relation to the influenza pandemic in China in the late 1910s. First, in the historiography of the 1918–20 pandemic, there has been a persistent trope of blaming China for being the point of origin of the outbreak. As we have seen, there is significant evidence to the contrary, both old and new. Second, from the ‘purulent bronchitis’ in British war hospitals in Étaples, France, and Aldershot, England, to the ‘malignant infection’ in Japan and the ‘malignant cold’ in China, the process of naming ‘influenza’ as a distinct new disease entity was complex in the existing medical frameworks of both Western and Chinese medicine. Just as Western doctors struggled to determine whether and how it fit into existing understandings of respiratory illnesses, some Chinese doctors integrated it into known cold-damage patterns of the wind-cold and wind-heat types of the ‘common cold’. Yet others emphasised its outlier qualities as a seasonal epidemic.

Finally, although Chinese medical treatments included acupuncture and massage, herbal formulas to treat wind-cold or wind-heat types of ganmao that had penetrated the lungs were dominant. Furthermore, practitioners of Chinese medicine and integrated medicine in China have continued to use comparable herbal formulas to treat patients during both the SARS and the recent Covid-19 pandemics (Ochs and Avery 2020). Blaming, naming, and treating remain as relevant as themes for our current understanding of responses to Covid-19 as they are for undergirding what happened in China during the 1918–20 influenza pandemic.

We acknowledge Elaine Leong (co-creator of The Recipes Project) for encouraging one of the authors to follow up on her essay ‘How Best to Treat the Heat in 1793 Beijing’ (Hanson 2018) for a thematic series on heat, with an essay on ‘Treating the “Deadly Cold” in 1918–20 China’ for a thematic series on cold. We did not publish in that series, but because of Elaine’s invitation we had a draft we could finalise for Global China Pulse. We are also grateful for S.J. Zanolini’s clinical insights about the herbal formulas used in 1918 for treating influenza being both conventional cold-damage formulas to treat the wind-cold types of ganmao and newer wenbing formulas to treat wind-heat types. All errors remain our own.