Summary

Influenza A (H5N1) has caused an unprecedented poultry pandemic and is threatening to cause a human pandemic which many believe would be at least as devastating as the "Spanish flu" of 1918-19. With the current global human population of 6.6 billion, hundreds of millions of lives may be lost. Nevertheless, compared to a century ago, we have good understanding of the influenza virus including detailed knowledge of its molecular biology. We are running surveillance systems on animals and humans that enable us to detect outbreaks early and institute effective measures to control it. We have drugs, vaccines, and modern healthcare infrastructure that can potentially reduce the damage should a pandemic occur. We should not be complacent, but never before have we been better equipped to fight the pandemic threat.

摘要

H5N1甲型流感已造成史無前例的家禽禽流感大流行。一旦發生人類的大流行，其殺傷力可能比1918-1919的「西班牙流感」更甚，按目前世界人口66億計算，全球死亡人數可能高達一億。可幸的是，相對1918年，我們現在對流感病毒包括其分子生物學已有校深的認知。全球監察流感機制使我們能儘早發現及控制小型禽流感爆發。倘若H5N1大流行不幸發生，藥物，疫苗及現代化的社會醫療架構可能減低其破壞力。雖然我們有更好的知識和裝備去面對禽流感大流行的威脅，但是，我們絕不能鬆懈。

---

Introduction

When Homo sapiens pursued a lifestyle of hunter-gatherers the limited food supply dictated that they were to live in small settlements. Infectious disease outbreaks likely occurred but with small isolated populations and limited travel, they were probably of very small scale. When the human race eventually adopted the practice of farming food became much more abundant. Large cities came into being and the progress of civilization took leaps and bounces, with infectious disease outbreaks likewise taking on a much larger scale. The major outbreak of a mysterious infectious disease in Athens, Greece, in 431 B.C. with a high mortality rate and enormous social disruption was the first of its kind to be chronicled in some detail.¹

In the latest millennium, fuelled by increasing global trading and exploration, means of transportation had progressively improved. Whereas previously each geographic location, particularly those isolated by inhospitable landscapes or stretches of ocean, had their own unique flora, fauna, microbes and pathogens, modernization of transport systems allowed alien species to set foot on unfamiliar niches, often with devastating biological consequences. For infectious disease pathogens this meant that they were no longer limited to individual villages and cities, but can penetrate every part of the Earth. The speed with which they do so paralleled the development of means of transport, and the damage paralleled the development of big cities. The Black Death originated in China in the year 1331 during a major civil war shortly into the Ming Dynasty, and together with war and famine halved the Chinese population. It slowly moved westwards along the trade routes, and reached the Crimea in 1346. The next year it entered Europe and killed one-third of the population in the next few years. The "Spanish flu" in 1918 travelled much faster and reached every part of the world within one year. In 2003, SARS was exported from Hong Kong to Vietnam, Singapore, and Toronto by commercial air flights within one day.

Pandemic influenza occurred about 3 times each century in the last 3 centuries, and each time caused major loss of life and societal disruptions. The "Spanish flu" of 1918-19 was particularly devastating and is considered the worst epidemic mankind has ever endured. Yet until recently influenza was seldom listed among the dangerous epidemic diseases. The complacency likely springs from the fact that influenza is a minor illness most of the time and people tend to forget that every few decades a novel strain of influenza A emerges and then influenza can turn really ugly. The first fatal case of human infection by influenza A (H5N1) in 1997 following a large scale poultry epidemic by the same virus served as a rude reminder that, rather than mere myth, influenza pandemics is something that may already be at our doorsteps.

A brief history of H5N1

H5N1 influenza virus was first isolated in poultry in Scotland in 1959. It left no doubts about its high pathogenicity when it caused an epidemic in common tern in South Africa two years later, killing thousands. Whereas H5N2 and H7 influenza A strains had repeatedly caused poultry outbreaks in different parts of the world in the last few decades, H5N1 did not do so until 1997. When it did it advertised its distinctiveness by immediately jumping the species barrier to infect humans, killing six of eighteen infected individuals. Transmission between humans was very inefficient however, and the poultry epidemic was quickly quelled by culling the entire poultry population of Hong Kong. Subsequent gene sequencing studies found that the haemagglutinin gene was similar to a goose H5N1 strain found in Guangdong in 1996, and the 6 internals genes came from a strain of H9N2 which was then commonly found in live poultry markets.²

True to Darwinism, Influenza A virus evolves constantly in response to survival pressure imposed by defense mechanisms of its hosts. It does so by mutations which are facilitated by the poor proofing mechanisms of RNA replication, as well as by reassortment between different influenza A strains which was well illustrated by the 1997 Hong Kong H5N1 strain. After 1997 H5N1 seemed quiescent for some years, but re-emerged in May 2001 in poultry markets in Hong Kong which necessitated another major culling. Waterfowl is the natural reservoir of all influenza A viruses, and infections are mostly asymptomatic. However, in December 2002, the virus took on its natural hosts and killed ducks, geese, egrets, herons and flamingos in Penfold Park and Kowloon Park.³ This was followed shortly in February 2003 by a father and son being infected by a closely-related strain, and the father died.^4,5

In December 2003, H5N1 poultry outbreaks occurred in 8 Asian countries, heralding a true poultry pandemic. The simultaneous appearance of human cases with high mortality in Vietnam and Thailand rekindled fears that a pandemic was imminent.⁶ There was some suspicion that this H5N1 strain, known as genotype Z, was spread by migratory birds,⁵ and the outbreak of bar-headed geese and other migratory birds in Qinghai Lake in May and June 2005^7,8 with subsequent appearance of H5N1 in the Middle East, Europe and Africa provided strong support for the suspicion. Others are unconvinced and believe that international trading of live poultry and poultry products (e.g. fertilizer for fish ponds) remains the most important means of spread.⁹ Virulence and host range continued to increase as evidenced by an outbreak among tigers and leopards in Thailand zoos. With more poultry and human cases coming up in Indonesia and China, it was soon apparent that there are 2 distinct clades of H5N1: the Vietnam/Thailand clade is resistant to Amantadine, whereas the China/Indonesia clade is sensitive.¹⁰ As of 4 July 2006, there were a total of 229 human cases since December 2003 with 131 deaths, giving a crude fatality rate of 57%.¹¹

Will H5N1 influenza virus cause a human pandemic?

H5N1 has proven its tenacity despite the culling of some 130 million poultry since early 2004 in attempts to eliminate it. Most experts believe that it is here to stay and will be an ongoing candidate pathogen for the next great human pandemic. The WHO has stipulated that in order for H5N1 to cause a serious human pandemic, there should be no herd immunity against it, there should be high virulence, and it should be able to transmit efficiently from human to human. Little need to be elaborated on the absence of herd immunity against H5N1 as unlike H1, H2, and H3 which infected humans in a cyclical pattern in the previous century, today's humans are virtually naive for H5. There is also little doubt about the very high virulence of current H5N1 strains on humans. Recent mutations involving the cleavage site of haemagglutinin has enabled the virus to infect multiple organs, not just the lungs, resulting in much increased lethality.¹² Other molecular determinants of virulence include mutations in the non-structural protein NS1 which interferes with cellular defense mechanisms including interferon production¹³ and mutations in the RNA polymerases PB1, PB2, and PA which are involved in many aspects of viral replication and host adaptation.^12,14

It appears therefore that whether or not we shall have a H5N1 human pandemic hinges on the ability of this virus to transmit efficiently from human to human in a sustained manner. The current low level of human to human transmission, which was also documented in the 1997 episode, is not a cause for concern. Recently results from two separate groups throw some light onto the mechanisms of this low-efficiency transmission. It is well known that the haemagglutinin of avian influenza viruses binds to receptors on the host cell surface with sialic acid linked to galactose by an a-2,3 linkage (SAa2,3Gal), whereas human influenza viruses prefer sialic acid linked to galactose by an a-2,6 linkage (SAa2,6Gal). Shinya et al demonstrated that in humans, SAa2,6Gal is dominant in the epithelial cell of nasal mucosa, nasal sinuses, pharynx, trachea, bronchi, terminal bronchioles and respiratory bronchioles. SAa2,3Gal on the other hand is only occasionally detected in the nasal mucosa, but is found in abundance in type II pneumocytes in the alveoli as well as in alveolar macrophages.¹⁵ This explains why human influenza viruses (H1N1, H3N2) infect people easily and is also easily passed onto other people because sneezing and coughing are common in upper respiratory infections. In contrast H5N1 need to get to the lower airways to initiate an infection and a much bigger infecting dose is required. Having established the infection transmission to another person is hampered by the deep location of the infection and scanty sneezing and coughing. Further laboratory data was provided by another group showing that the influenza virus A/Vietnam/1194/04(H5N1) attached predominantly to type II pneumocytes, alveolar macrophages and non-ciliated cuboidal epithelium of terminal respiratory bronchioles, with attachment getting progressively rarer towards the trachea.¹⁶

It is therefore rather disturbing news that according to studies based on the reconstructed genome of the 1918 H1N1 virus responsible for the "Spanish flu", which also is avian in origin, as few as two mutations of the haemagglutinin gene is needed for the virus to acquire the ability to bind to SAa2,6Gal.¹⁷ On the other hand, receptor specificity may not be the only factor in the emergence of a pandemic strain, as suggested by the finding that A/Hong Kong/213/2003(H5N1) which killed the 33 year old man in Hong Kong in February 2003 in fact recognized SAa2,6Gal as well as SAa2,3Gal,¹⁸ but no sustained human to human transmission occurred.

H5N1 pandemic: Spanish flu of the 21st century?

Many experts believe that if a H5N1-pandemic erupts and the case-fatality rate is only a fraction of the current death rate, the death toll is likely to be at least as horrendous as the "Spanish flu" of 1918-1919. But how bad was the Spanish flu?

The spring of 1918 in the Northern Hemisphere saw an influenza epidemic striking soldiers on both sides of "The Great War", afterwards known as World War I. The symptoms were typical of influenza, also known as "grippe" then, and fever typically lasted 3 days. Although many soldiers fell sick and some died, most were able to resume duty within a week, and the illness was restricted to the military. No one was unduly concerned and this preamble to subsequent waves was completely missed, which is hardly surprising since everyone was preoccupied with the progress of the war and the aetiology of influenza was unknown at that time so that the medical profession had no way of tracking the activity of the influenza virus.

The second and most deadly wave came in late August of 1918 simultaneously in France, Sierra Leone, and the Eastern seaboard of the United States. No country was prepared for an influenza pandemic which carried a many-fold increase in death rate. In Camp Devens just outside Boston, Massachusetts, there was an explosive increase in soldiers getting the flu and dying. A doctor stationed at the Camp wrote to a colleague: "These men start with what appears to be an attack of la grippe or influenza, and when brought to the hospital they very rapidly develop the most viscous type of pneumonia that has ever been seen. Two hours after admission they have the mahogany spots over the cheek bones, and a few hours later you can begin to see the cyanosis extending from their ears and spreading all over the face, until it is hard to distinguish the coloured men from the white. It is only a matter of a few hours then until death comes, and it is simply a struggle for air until they suffocate." By late September the daily death toll at this camp was upwards of 100 and rising, and an outrageous number of nurses and doctors were lost.

Civilians were quickly involved. In Philadelphia the influenza epidemic was becoming apparent by the third week of September, and by September 21 the Bureau of Health finally made influenza a reportable disease. Yet the officials were unwilling to admit that there was an epidemic, and allowed a Liberty Loan (war bond) parade involving 200,000 people to proceed on September 28, 1918. Within a few days there was a huge surge of influenza cases and deaths, and the city Government was forced to admit that epidemic conditions existed. Churches, schools, and theaters were ordered closed, along with all other places of "public amusement". The week ending 19th October 1918 saw 4,597 deaths from influenza and pneumonia in Philadelphia, and a similar number in New York. The latter city had a population of 3.2 million and registered 851 deaths from influenza and pneumonia in a single October day.¹⁹ No major city of the United States was exempt. October 1918 turned out to be the deadliest month in the nation's history as 195,000 Americans died of influenza. Within just one year, over 500,000 Americans died of influenza, and the number was more than American deaths in combat for World War I, World War II, Korean War, and Vietnamese War combined.

The situation was less well documented in other countries, but appeared to be at least as bad. In Frankfurt, Germany, the mortality rate was 27.3% for patients hospitalized for influenza. In Paris, France, mortality rate was stated as 10% for all cases of influenza and 50% for those with complications. In India, the mortality rate was 9.61% for British troops and 21.69% for Indian troops. Some estimated the overall mortality was around 18 million in this country alone.²⁰ In Gambia 8% of all Europeans died, and whole villages of 300-400 individuals were wiped out. Even Arctic villages could not escape the onslaught. Villagers would begin to fall ill from influenza a few days after a brief visit of the mail and supplies boat, and many then died of the disease or from lack of care.

Following a brief lull, when everyone was hoping that the Spanish flu was over and done, a third wave struck in December 1918 and continued over the ensuing Northern winter. All the mask-wearing and limitation of public gathering were re-instituted, but the death toll was not as dramatic as the autumn wave. The exception was Australia which was able to avoid the first and second waves by draconian maritime quarantine measures, but somehow allowed the flu to slip through in the third wave, with devastating results.

What about China and Hong Kong? There were few reliable central government statistics in China before the 1930s, and no reliable demographic sources for an analysis of the impact of the Spanish flu. Iijima examined the records of the Chinese Maritime Customs in the treaty ports, and noted that there were serious outbreaks in cities like Hankow, Tangshan, Chungking, Wenchow, etc, but appeared mild in Shanghai, Amoy (Xiamen), Wuchow, etc.²¹ He estimated the death toll in China to be in the region of 4-9.5 million, which was 0.8-2.0% of the total Chinese population of 472 million. In Hong Kong the government annual reports stated that "There were 485 admissions under the heading of influenza and seven deaths occurred".²² A special report by Dr. Olitsky of the US Rockefeller Institute on investigation of meningococcal outbreak in Hong Kong also mentioned in passing that "the epidemic was not severe.... but the distribution of these ailments was quite general".²³ The government data likewise did not show significant increase in death from all causes and from pneumonia,²⁴ but Iijima pointed out that mortality figures included only deaths which occurred in the Colony's hospitals, and many Chinese labourers were never taken to hospital,²¹ so it was quite possible that mortality from Spanish flu was underestimated in Hong Kong.

Globally, how many people were killed by the Spanish flu? In the 1920s it was estimated to be 21.5 million, but recent works showed that it is more likely to have been of the order of 50 million.²⁵ Even this vast figure likely underestimated the true mortality because of non-registration, missing records, misdiagnosis, and non-medical certification, and the actual death toll may be up to 100 million.²⁵ It is worth noting that the world population was 1.8 billion in 1918, and 50 million deaths was 2.8% of the global population. Today's world population is 6.6 billion and if the same proportion of people dies then the death toll would be at least 180 million. Furthermore, unlike seasonal influenza and other influenza pandemics, which mostly claim the lives of the very young and very old, deaths from influenza in 1918 physically and psychologically. The Hong Kong case in 2003 and others in South East Asia strongly suggest that H5N1 can do the same.

Is history going to repeat itself?

The old saying is "History always repeats itself". The world human population has increased by 3.5-fold since 1918 and still increasing rapidly, and this is paralleled by an alarming increase in the world poultry population. Together with the fact we the world is "overdue" for an influenza pandemic and the recent activities of H5N1, it seems the next pandemic is indeed imminent. Nevertheless, 2006 is 88 years from 1918, and medical sciences have made tremendous progress in the intervening years. Never before do we possess so much knowledge about the influenza virus. We have a fair understanding on infectivity, transmissibility, and pathogenicity of the virus in both clinical and molecular terms, and we are tracking its mutations closely and building phylogenic trees of it. We have even resurrected the Spanish flu H1N1 virus and studying it intensely to throw insight into our current predicament. We also begin to realize that human influenza is just a small part of the global biology of influenza, and to understand human influenza we have to study influenza in birds, pigs and other animals in detail. Any poultry outbreak of H5N1 is immediately quenched by culling, and human cases are isolated, so that opportunities for onward spread and adaptive mutations are greatly lessened. Effective medications in the form of neuraminidase inhibitors have been stockpiled by many countries, and if used in combination with other public-health measures may abort a pandemic at its outset.^26,27 The development of H5N1 prototype vaccines based on 2004 Vietnamese strains is another exciting area in the fight against the avian flu threat. Clinical trials in humans have been undertaken and the vaccines were shown to be safe and result in protective antibody levels.^28,29 There is ample evidence from animal models that cross-protection can be conferred to related influenza strains,^30-32 and the H5N1 prototype vaccines may likewise confer partial protection against the pandemic H5N1 strain at the early phase of its emergence and thus buy time for manufacture of the pandemic vaccine.

It does appear therefore that we are never better equipped to stop the influenza pandemic from occurring, or at least significantly minimize its damage by the many national and regional preparedness plans. Others however are less optimistic. Following a forum on avian flu hosted by Lancet in May 2006 the Editors felt that there are few truly effective measures against a potential pandemic and politicians and policy makers are over-confident in their ability to control it.³³ Nonetheless, armed with our new-found knowledge and the lessons from history, we can at least put up a good fight against the challenge and to avoid making the same mistakes a century ago.

Key messages

1. The world is having a poultry pandemic caused by Influenza A (H5N1) which has so far proved difficult to eradicate.
2. Humans have acquired H5N1 from poultry with a mortality rate of > 50%.
3. If H5N1 acquires the ability to transmit efficiently from human to human then the ensuing human pandemic may be worse than the "Spanish flu" of 1918, with a death toll of > 100 million.
4. We should prevent history from repeating itself by international collaboration on improving influenza surveillance on humans and animals, improved poultry farming and trading practices, timely culling of infected flocks, and timely isolation of human cases.
5. Our new-found knowledge on virology, our modern medical infrastructure, and lessons from previous pandemics may help reduce the damage in the event of a pandemic.

---

---

References