Putting pandemics in perspective
Mark Honigsbaum |
- The World Health Organisation's declaration of a swine flu 'pandemic' in 2009 has been widely criticised for triggering contracts for the supply of billions of dollars worth of drugs and vaccines that, in the end, proved unnecessary because of the mild nature of the virus.
- Much of this criticism has focused on the WHO's changing definition of a pandemic, leading to the accusation that it somehow 'faked' the pandemic for the benefit of the pharmaceutical industry. However, surveys of the historical literature on influenza show that 'pandemic' has always been a malleable term.
- Today influenza epidemics and pandemics tend to be defined according to morbidity criteria (the number of confirmed cases of infection) and their geographical distribution. However, historically, the notion of 'excess mortality' was central to the definition of an epidemic, reflecting the role of Victorian epidemiology in making influenza more visible to public health.
- Although epidemiology continues to play a key role in monitoring unusual illness patterns, 'antigenic shifts' (which allow viruses to pass from one species to another) in pandemic viruses are increasingly defined according to genomic criteria that take little account of the severity of illness.
- The modern 'pandemic preparedness' industry is predicated on avoiding a repeat of the 1918-19 'Spanish' influenza pandemic, which killed an estimated 50 million people worldwide, but judging by the far lower mortality levels of subsequent pandemics, 1918 increasingly looks to have been an exceptional event.
- Despite recent advances in genomics, scientists are still no closer than they were a century ago to being able to predict when and where pandemic viruses will emerge and how they will impact human populations.
- The World Health Organisation should consider incorporating a severity test into its pandemic alert system so that only pandemic strains with case fatality rates in excess of 1 percent would trigger a 'phase six' pandemic alert.
- Health experts should acknowledge that while predicting pandemics is not a perfect science, advances in modern medical technology and virology give governments the ability to forestall and, in some cases, prevent the worst health and economic impacts of pandemics, resulting in benefits for everyone.
Writing in 1750, seven years after a devastating European-wide epidemic of influenza, the English country doctor and surgeon John Huxham characterised flu as the 'morbus omnium maxime epidemicus' or the 'greatest of all sicknesses'. For the next 200 years or so that served as a pretty good working definition. As the German disease geographer August Hirsch observed in 1883, the only difference between an epidemic and a 'true pandemic' was that in the latter case the sickness extended 'over the whole habitable globe'.
If only matters were so simple today. Despite predictions that Britain could be facing its worst winter flu season in years, uptake of flu vaccination among doctors and other health professionals remains alarmingly low (under 35 per cent according to recent reports), heightening the risk of infection among the over 65s who are amongst the heaviest users of the NHS. These low immunisation rates can be seen, in part, as a hangover from the 2009 swine flu pandemic and the widespread public distrust engendered by what turned out to be the government's needless spending on drugs and vaccines on the recommendation of the World Health Organization (WHO). As a report in May 2011 by the United Nation's International Health Regulations (IHR) Review Committee made clear, at the root of this distrust was the WHO's changing definition of a pandemic. For many years, the WHO had insisted that in order for a new strain of influenza to be considered 'pandemic' it had to cause 'enormous numbers of deaths and illness'. However, on 29 April 2009, shortly after the unexpected emergence of the swine-origin H1N1 virus in Mexico, the WHO deleted the severity requirement from its pandemic guidance papers and reverted to something closer to Huxham's and Hirsch's original definition. The result was that when, in June 2009, the WHO's director-general Margaret Chan called a press conference in Geneva to announce that the world was 'at the start of the 2009 pandemic', this definition was based only on sustained 'human-to-human spread' of the H1N1 virus in two countries in one WHO region plus 'community level outbreaks' in a country in another region.
That decision turned out to be something of a windfall for pharmaceutical companies, triggering 'sleeping' contracts for the supply of billions of dollars worth of vaccines and anti-viral drugs to national health providers. It was also a decision that boomeranged badly on Chan when swine flu turned out to be far milder than WHO experts had feared, prompting allegations in the Council for Europe that the WHO had somehow 'faked' the pandemic for the benefit of Big Pharma - an allegation Chan strongly denies.
It is not my intention to discuss the truth or otherwise of these allegations - for that I direct you to the IHR Review Committee's carefully worded report (for the record, it found no evidence of financial 'malfeasance' but criticized the WHO for not doing enough to dispel suspicions that it had something to hide). Instead, I wish to examine where the present-day notion of a pandemic comes from and how it has evolved over time. In particular, I wish to explore the role of epidemiology in the emergence of what I will call the 'modern' notion of a pandemic, before going on to consider the role of virology and new genomic technologies in the definition latterly adopted by the WHO.
The term 'pandemic' has always been somewhat slippery and malleable. Writing in 1997, the esteemed American virologist Ed Kilbourne compared the process to pornography: 'We all "know it when we see it," but the boundaries are a little blurred.' For Kilbourne, as for Hirsch, the difference between an epidemic and a pandemic of influenza was simply 'the occurrence of many cases throughout the world in a short period'.
Certainly, in the earliest recorded English usage severity was not part of the definition. Thus, writing in 1666, the English physician Gideon Harvey uses 'Pandemick' interchangeably with 'Epidemick' to refer to a malignant disease which 'do generally... haunt a Country.' Interestingly, the Victorian historical epidemiologist Charles Creighton eschewed the term altogether, preferring to refer to 'epidemics' of influenza even when they were extensively diffused over large parts of Europe and the world, as was the case with the epidemics, or pandemics, that visited England in 1836-37 and 1847-48.
So when did influenza first begin to be seen not only as an extensively diffused world epidemic disease, but also as a 'pandemic killer' and a potential object of panic? For the answer, we have to return to the successive waves of influenza that swept Europe, Britain, and North America in the early 1890s. Termed the 'Russian' influenza because the first reported outbreaks occurred in St Petersburg in the autumn of 1889, these were the first epidemics to be subject to close clinical and epidemiological surveillance on an international basis and the first to be studied with the 'modern' tools of the laboratory. The result of these investigations was a shift in the medical understanding of the causation involved that saw influenza re-classified from a miasmatic disease, whose prevalence was thought to be due to mysterious telluric (earthly) and celestial disturbances, to what the Medical Department of Britain's Local Government Board called an 'eminently infectious complaint', one that seemed to be intimately connected to new global transportation and communication technologies. At the same time, retrospective analysis of the death returns from the Russian flu pandemic (usually dated to 1889-93) made the high mortality due to respiratory complications such as pneumonia increasingly visible to doctors and public health officials, underlining the threat that influenza presented to a wide range of social classes and age groups. The result was that when, in the winter of 1892, the Duke of Clarence, Queen Victoria's 28-year-old grandson and the second-in-line to the throne, died from pneumonia following an attack of influenza, there were demands for the 'flu to be made a notifiable disease just like cholera and smallpox.
The mother of all pandemics and a continuing enigma
The crucial contribution, however, came from attempts by historical epidemiologists to divine a pattern in the recurrent waves of the disease. The Russian influenza was the first to be seen to occur in three distinct waves, with the first wave causing widespread morbidity but relatively few deaths, and the second and third waves proving highly mortal. However, it was the recurrence in 1918-1919 of this pattern of a mild 'premonitory' wave followed by deadly follow-on waves that fixed the notion that an influenza pandemic usually took the form of at least three successive waves of infection. Moreover, as retrospective analysis of the 1918-19 'Spanish' influenza proved, the death toll from these follow-on waves could be immense: 60 percent of the estimated 228,000 British casualties died in the second wave of the pandemic; the worldwide death toll from the Spanish flu has been estimated at a truly apocalyptic 50 million (at a time when the world's population was about a quarter its current size).
One result of this focus on the 1918-19 experience was to fix the notion that influenza pandemics result in 'enormous numbers of deaths'. As a WHO 'pandemic preparedness' document put it in 2005 at the time of mounting concerns over bird flu, the 1918-19 pandemic was 'the most deadly disease event in the history of humanity' - hence the importance of pandemic planners doing everything they could to avoid a reoccurrence. But what if 1918-19 was a unique, never-to-be repeated event? What if influenza does not conform to any set biological pattern or epidemiological cycle? What if instead of the past behaviour of influenza providing a guide to the future, it destabilizes the present, showing pandemics to be nothing more than a series of historically contingent events? That is both the enigma and challenge that influenza presents to pandemic planners.
The birds have it
To be fair, the WHO seems to have recognised the dangers of this historiographical bias towards 1918, acknowledging that while the 1918-19 pandemic resulted in 50 million deaths worldwide, the 1957 and 1968 pandemics each resulted in just one million deaths globally, a much lower fraction of a global population that had itself multiplied in size during the intervening decades. Moreover, by the late 1990s new insights into the virology and ecology of influenza viruses were forcing WHO scientists to revise their notions of how pandemic viruses emerge from 'hidden' animal reservoirs. Whereas previously pandemics had been defined largely according to clinical observations and epidemiological measures, by 2005 the WHO's expert advisors were urging that mutations to the virus's genes and surface proteins also needed to be taken into account.
According to the IHR Review Committee, these debates came to a head around bird flu. When the H5N1 virus first emerged in 1997 and then re-emerged in 2005, it proved highly lethal with a case fatality rate in excess of 70 percent. However, unlike true pandemic viruses, it was sluggish: human infections tended to occur in isolated family clusters and did not spread easily between non-blood relations. On the other hand, from analysis of the genome of the 1918 virus - the most avian-like of all pandemic viruses - scientists knew that H5N1 already had many of the characteristics of a pandemic virus and that with just a few mutations it could easily become highly infectious. The result was that in 2007, with growing evidence that bird flu was being 'seeded' worldwide and that early action could contain and prevent the virus before it mutated to cause 'enormous numbers of deaths and illness', WHO experts recommended lifting the severity requirement for classifying a pandemic and replacing it with the present definition.
Far from the definition change being made surreptitiously, as internet conspiracy theorists have suggested, the consultations were open and lasted seventeen months. According to the IHR committee, it was simply an accident of timing that the WHO deleted the phrase from its pandemic guidance documents in April 2009 just as H1N1 was emerging from H5N1's shadow, thereby fuelling claims that there was more to the definition change than met than eye.
Conspiracy theories and the 'miracle' of hindsight
Once they're up and running, conspiracy theories are very hard to put back in the box - and judging by the initial response to the IHR committee's findings this one won't be any different. According to Paul Flynn, the Labour MP for Newport who, as a member of the Council of Europe's socialist group brought the original complaint against the WHO, the issue was never 'malfeasance' but whether there was a 'conflict of interest' among the members of the WHO's 15-strong emergency committee that advised Chan on the timing of the pandemic, five of whom turned out to have declared financial ties to the pharmaceutical industry. Though the IHR committee found no evidence of 'attempted or actual influence by commercial interests', many people remained suspicious because of the manner in which the definition was altered at the very moment when the WHO's response to swine flu hung in the balance.
In fact, as the committee points out, it is only with the benefit of hindsight that we can say that swine flu, which to date has resulted in just 251 deaths in Britain and 18,448 deaths worldwide (or about half the number who die from seasonal flu in the US every winter), was not particularly deadly. By contrast, in June 2010 when Chan pushed the pandemic alert button, the mortality rate among patients hospitalised in Mexico City was running at 41 percent, and experts had good reason to fear the worst.
So what definition should we adopt in future? The first thing to say is that it is no use looking to historical epidemiology for simple answers. Yes, the 1918-19 pandemic was severe, but based on the recent run of mild pandemics in 1957, 1968 and now 2009, it increasingly looks like the exception. On the other hand, as the IHR committee points out, it would be rash to assume that pandemics are declining in severity simply on the basis of four observations spanning one hundred years.
What about research into the genetic markers of influenza viruses and other host factors? Are these likely to increase the accuracy of predictions? Again, for the moment, the answer appears to be no. As Jeffery Taubenberger and David Morens, two of the world's leading experts on, respectively, influenza virology and epidemiology, argue, despite the tremendous progress made in virology, microbiology, immunology, pharmacology, epidemiology, vaccinology, and preventive medicine over the last century, influenza researchers are still no closer to being able to predict when new pandemic strains will emerge or how they will impact human populations when they do. 'As our understanding of influenza viruses has increased dramatically in recent decades, we have moved ever further from certainty about the determinants of, and possibilities for, pandemic emergence,' they write. Indeed, for all that Chan boasted that the 2009 pandemic was the first to have been observed 'right from the beginning', Taubenberger and Morens point out that in fact: 'No one predicted the emergence of the 2009 H1N1 swine-origin pandemic virus; [and that] with current knowledge, we doubt anyone will be able to accurately predict any future pandemic either.'
While we may be no closer to being able to predict pandemics, we are surely far better equipped to track their progress and marshal an effective public health response. In 1890, the world had no way of monitoring obscure outbreaks in the Yucatan, Mongolia or other 'silent spaces' on the map. Today, for all its faults, the WHO has an unparalleled ability to keep tabs on novel pathogens thanks to its Global Outreach and Response Network engaging 300 technical partners worldwide, and electronic disease-reporting systems such as ProMED. Moreover, in 1890 and in 1918, there were no influenza vaccines or stockpiles of antiviral medications. Nor in most countries could doctors count on institutions like the NHS to deliver care to the elderly and the vulnerable.
The problem with medical progress is that advances in prevention and treatment present governments and taxpayers with new and difficult financial choices, especially when the drugs and vaccines must be ordered ahead of time and in sufficient quantities to be of any use. It is this - not differences over how to define pandemics - that best explains the lingering distrust over the WHO's response to swine flu and the poor uptake of vaccines. Given that pandemics are likely to occur again and given that, when they do, they are likely to spark differences of opinion, my prescription is for a little less panic and a little more commonsense. Rather than making rash predictions about likely death tolls, as was the case in 2009 when the Department of Health predicted that in a 'worst case scenario' as many as 65,000 Britons might die from swine flu, health officials would do better to err on the side of caution. Similarly, Margaret Chan ought to have waited for better serological data from Mexico - data that would have provided a more accurate gauge of the case fatality rate (CFR) from swine flu - before ringing the pandemic alarm bells (initial reports out of Mexico suggested a CFR of close to 7 percent which, if true, would have been unprecedented; once the true level of infections was known, this was revised down to 0.03 per cent, making swine flu 100 times less lethal than the Spanish flu). Indeed, before triggering such alerts in future, the WHO would be advised to set some sort of severity threshold - for sake of argument, a CFR in excess of 1 percent.
However, while such tests should surely be incorporated into the decision-making process, I can see little point in making them part of a formal definition. Instead, the WHO should acknowledge that while early warning disease-monitoring systems enable us to get a jump start on emerging viruses, predicting pandemics is not a perfect science. At the same time, advances in medical technology and virology give us the possibility of forestalling and, in some cases, preventing the worst health and economic impacts of influenza pandemics - something that wasn't possible a century ago.
Charles Creighton, A History of Epidemics in Britain, vol. II (Cambridge University Press, 1894).
Mark Honigsbaum, 'Russian influenza: lessons learned, opportunities missed', Vaccine, Vol 29, Supplement 2, 22 July 2011, pp. B11-B15.
Niall Johnson, Britain and the 1918-19 Influenza Pandemic: A dark epilogue (London; New York: Routledge, 2006).
'The handling of the H1N1 pandemic: more transparency needed', Report by Paul Flynn, Social Health and Family Affairs Committee, Parliamentary Assembly, Council of Europe. http://assembly.coe.int/CommitteeDocs/2010/20100604_H1N1pandemic_e.pdf
Report of the Review Committee on the Functioning of International Health Regulations in Relation to Pandemic H1N1 (2009). http://apps.who.int/gb/ebwha/pdf_files/WHA64/A64_10-en.pdf
Jeffery Taubenberger and David Morens, 'Influenza: The Once and Future Pandemic', Public Health Reports 125, 3 (April 2010): 16-26.
About the author
Dr Mark Honigsbaum is a Research Associate at the Institute and Museum for the History of Medicine, University of Zurich, and the author of Living With Enza: The Forgotten Story of Britain and the Great Flu Pandemic of 1918 (Macmillan Science, 2008). firstname.lastname@example.org