Select Committee on Science and Technology Fourth Report

CHAPTER 2: INFLUENZA

The influenza virus: background

THE DISEASE

2.1.  Influenza, or flu, is a viral infection caused by the influenza virus, affecting the respiratory tract. The virus was first identified in 1933, though the disease was familiar long before this. The symptoms include fever, runny or blocked nose, cough, sore throat, aching muscles and joints, tiredness, headache, vomiting and diarrhoea. However, similar symptoms, particularly the first few, can be caused by other respiratory viruses, and what is commonly described as "the flu" is usually caused by other viruses which cause upper respiratory illness, such as the common cold virus.

2.2.  For the purposes of contingency planning, influenza can be divided into three categories:

• "Ordinary" influenza circulates constantly in the human population, typically peaking in the winter months, and affecting around 10 percent of the population over the course of a year;
• "Avian" influenza is a disease of birds, caused by strains of the virus that are endemic in wild birds, and which do not normally infect man, though they periodically cause destructive epidemics among poultry;
• "Pandemic" influenza in humans, so called because of its global spread, typically infects around a quarter of the world's population within the space of a few months. It occurs on average around three times a century, when a strain of avian influenza acquires the ability to infect and pass efficiently between humans. As this is a new virus to humans, there is no immunity, and the disease spreads rapidly. Strains of pandemic influenza vary widely in virulence. Some past pandemic strains have in fact been less virulent than "ordinary" influenza; others (notably in 1918-19) have been significantly more virulent, with high mortality rates.

THE VIRUS: CLASSIFICATION

2.3.  The most severe infections, whether avian or human, are caused by "influenza A" type viruses, which are in turn divided into subtypes by reference to their 15 different haemagglutinin (H) and nine different neuraminidase (N) antigens.[4] These are protein spikes on the virus surface. Haemagglutinin allows the virus to bind onto healthy cells, while neuraminidase allows the virus to break out of infected cells, so that it can move onto other cells and spread infection throughout the body. H and N antigens are important in immunity to and treatment of influenza.

2.4.  Influenza A viruses are found in many different animals, including ducks, chickens, pigs, whales, horses, and seals—and of course humans. However, birds are the primary reservoir, and since it was established that bird flu was caused by the influenza A virus in 1955 more than 90 species of wild birds have been found to carry influenza A, excreting large quantities in their faeces, yet remaining apparently healthy. All known H and N subtypes have been identified in wild birds.

2.5.  Avian influenza A viruses are divided into two broad classes: High Pathogenicity Avian Influenza (HPAI) and Low Pathogenicity Avian Influenza (LPAI). In poultry LPAI can cause a range of illnesses, including a mild illness with ruffled feathers, reduced egg production and cough, while more virulent HPAI strains can kill an entire flock within 48 hours. Since 1959 there have been 24 outbreaks of HPAI recorded worldwide, of which 14 have occurred since 1995. Only one strain (known as H5N1) has spread internationally.

THE VIRUS: MUTATION

2.6.  Influenza viruses change continually as a result of two main mechanisms: "antigenic drift" and "antigenic shift". Antigenic drift occurs all the time as a result of the inherent instability of the virus: gradual mutations result in the virus envelope undergoing minor changes, resulting in a new strain that may not be recognized by antibodies to earlier influenza strains. As a result people or other animals that have already been exposed to influenza may be re-infected by a new, slightly different strain, although exposure to a previous strain may confer partial immunity, so that the severity of the illness is reduced. The instability of the virus means that a wide variety of different strains are present at any one time in different animal species. For example, H1N1 (the "Spanish flu" virus of 1918-19), H1N2 and H3N2 currently circulate among people; H7N7 and H3N8 circulate in horses.

2.7.  Antigenic shift occurs more rarely, and as the name implies represents a more dramatic change in the virus. It can occur as a result of the "recombination" or "reassortment" of two different strains affecting different species, and this appears to be one means whereby the virus can jump from one species (e.g. poultry) to another (e.g. humans). For instance, a human and an avian virus, both simultaneously infecting the same "mixing vessel" (an animal susceptible to both viruses, such as a pig), could combine. The result would be the emergence of a virus containing a high proportion of the genetic material of existing human strains, but incorporating a new haemagglutinin, against which humans have little or no immune protection.

2.8.  New research, however, shows that the effect of antigenic shift can also be achieved through mutation without recombination. An avian virus can by itself acquire the ability to infect humans and transmit from person to person without recombination in a "mixing vessel". Such a mutation seems to have led to the emergence of the "Spanish flu" virus in 1918.

Influenza pandemics

2.9.  We have already noted that a pandemic occurs when a new highly infectious strain of influenza appears, normally an avian virus which has undergone antigenic shift. It is thought that south east Asia is the most likely location for such a virus to develop, largely because of the number of people living in close proximity to very large populations of domestic birds and pigs.

2.10.  Roughly three influenza pandemics have been documented each century since the 1600s, occurring at irregular intervals of from 10 to 50 years. They usually spread to all parts of the world within less than a year, and affect more than a quarter of the world's population. The following graph illustrates the predicted course of a pandemic in the United Kingdom, starting from the introduction of one single case of influenza on Day 1:

FIGURE 1

Predicted course of an influenza pandemic in the UK (one import only)

Source: Professor Neil Ferguson. "Day" starts from introduction of first UK case; the modelling assumes a reproduction rate of 1.8. The curve derives from a classic "SEIR" model, classifying individuals as Susceptible, Exposed, Infectious or Recovered from the disease.

2.11.  In reality pandemics may not follow this simplified model exactly. Further waves, for instance, sometimes with more severe disease, generally follow the first wave, but the reasons for such repeated waves, and their likely timing and effect, are not well understood or modelled. The courses of the three pandemics of the 20th century illustrate both the common elements and the variations between pandemics that make any attempt at prediction so difficult.

"SPANISH FLU" (1918, H1N1)

2.12.  The 1918-19 pandemic, known as "Spanish flu", was one of the most deadly disease events in human history. Estimates of the number killed range from 20-50 million, but there is general agreement that within one year more people died from influenza than were killed in the first World War in the years 1914-18.

2.13.  The first outbreaks occurred in March 1918 in Europe and the United States. The infection then travelled via troopships and by land to Asia and Africa. The first wave was highly contagious but not particularly deadly. The second wave started at the end of August 1918 in France, Sierra Leone and the United States and saw outbreaks with a ten-fold increase in the death rate. The highest death rate was in fit young adults—unusually, as influenza commonly kills the very young and very old.

2.14.  Control measures included isolation, quarantine, personal hygiene, use of disinfectants, and the prevention of public gatherings. Many public institutions, including schools, were closed. The wide imposition of quarantine and isolation probably had little effect, although strict maritime quarantine delayed the arrival of the epidemic in Australia until the start of 1919. By this time the virus had become less lethal, and Australia had a milder, though longer, period of influenza activity than elsewhere. Even in Australia 60 percent of deaths occurred in persons aged 20 to 45 years.

2.15.  During the 1918-19 pandemic no part of the world was spared and 25-30 percent of the world population fell ill. The capacity to respond—for instance the supply of hospital beds for the sick, or of burial space and coffins for the dead—was overwhelmed.

"ASIAN FLU" (1957, H2N2)

2.16.  At the beginning of May 1957 reports of epidemics in Hong Kong and Singapore were received. Subsequently it became clear that epidemics had begun at the end of February in China, spread throughout that country in March, and reached Hong Kong by mid April. Since 1918 understanding of influenza had advanced considerably. The virus had been identified in 1933 and vaccines for seasonal epidemics had been developed. Antibiotics were available to treat complications. The World Health Organization (WHO) Global Influenza Surveillance Network—a virological monitoring and early warning system—had been in place for 10 years. By mid-May, laboratories had isolated the virus and identified it as a new subtype.

2.17.  By autumn 1957 every part of the world had experienced cases. In Europe the epidemic coincided with the September return to school. It peaked rapidly and was over by December. Mortality showed a similar pattern to that seen in seasonal epidemics, in that excess deaths were confined to infants and the elderly. Cases were concentrated in school-aged children and those crowded together such as in military barracks. A second wave followed one to three months later with very high rates of illness and increased fatalities.

2.18.  Quarantine measures were generally found to be ineffective, at best postponing the onset by weeks. Spread within some countries was associated with public gatherings, and the banning of such gatherings and the closing of schools were considered the only measures that could slow spread. Vaccines became available in the United States (August), United Kingdom (October) and Japan (November), but quantities were too small for wide-scale use. Total excess mortality globally has been estimated at more than two million deaths.

"HONG KONG FLU" (1968, H3N2)

2.19.  The 1968 pandemic was even milder than that of 1957. In mid-July a British newspaper published a story about widespread acute respiratory disease in China. During July the disease spread to Hong Kong, causing half a million cases within two weeks. The virus was rapidly identified as a novel subtype and on 16 August the WHO warned of possible pandemic.

2.20.  The disease spread more slowly rather than in previous pandemics, apart from in the United States. Here the epidemic, introduced by troops returning from Vietnam, began in September in California, and affected the whole country by late December. A significant increase in deaths, concentrated in the elderly, occurred during January. In Europe symptoms were mild and excess deaths negligible. In the United Kingdom the epidemic began in December, and demands on medical services were not excessive. Recorded deaths from influenza-like illness and pneumonia were actually lower than the year before.

2.21.  Vaccine manufacturing began within two months of the virus being isolated, but only 20 million doses were ready by the time the epidemic peaked in the United States. Global excess mortality was probably around one million. It is thought that the mildness of the 1968 H3N2 pandemic may have been because those exposed to the 1957 H2N2 strain enjoyed partial protection against the shared N2 subtype.

SUMMARY: PANDEMIC VIRUSES

2.22.  There are crucial differences between the viruses that caused the 1918 and the 1957 and 1968 pandemics. The full sequence of the 1918 H1N1 virus gene has recently been characterised.[5] Hitherto it has generally been assumed that pandemic strains emerge as a result of reassortment, but it now appears that the 1918 virus was a "pure" avian virus, which acquired through spontaneous mutation the features necessary to infect humans readily and transmit from person to person. It is not known if this occurred rapidly or took place over a number of years. It is thus possible that Spanish flu emerged as a result of a gradual process; it does not appear to have resulted from a one-off antigenic shift as a result of reassortment with a human virus.

2.23.  In contrast, analysis of viruses from the 1957 and 1968 pandemics shows that they are both reassortants (mixtures) of human and avian viruses. The 1957 H2N2 virus comprises three genes from an avian virus and the remaining five genes from the previously circulating human H1N1 strain (which in turn derived from that which caused the 1918 pandemic). The 1968 H3N2 strain also has three genes from an avian virus and the remaining five from the circulating human H2N2 strain (derived from 1957 pandemic virus). In both cases, it is suggested that the pandemic strains emerged as a result of reassortment in pigs, probably in those parts of Asia where large numbers of people live in close proximity to ducks and pigs, providing ideal conditions for the virus to cross between the species.

2.24.  The recent discoveries regarding the 1918 virus raise crucial questions. It appears that the exceptional virulence of that pandemic can be explained at least in part by the fact that it derived from a "pure" avian source, totally new to humans, with the result that people had no immunity (in marked contrast to the 1957 and 1968 viruses). The process of adaptive mutation in bird populations may also mean that it emerged in several locations more or less simultaneously, rather than in one "big bang". Worryingly, the current virus circulating in bird populations, H5N1, resembles the H1N1 virus responsible for Spanish flu, not only in its exceptional virulence, but also in its process of mutation. As the authors of the recent characterisation of the H1N1 virus note, "a number of the same changes [as occurred in the H1N1 virus] have been found in recently circulating, highly pathogenic H5N1 viruses".

The H5N1 virus

2.25.  Occasionally an avian influenza virus infects a person, normally as a result of close contact with infected birds—characteristically poultry workers and veterinarians are most at risk. There have been about 11 documented episodes since 1959, with from one to 133 people (in the current H5N1 outbreak) infected.

2.26.  Human infections with avian influenza viruses usually produce mild disease followed by full recovery. H5N1 has been the exception. The first outbreak of H5H1 avian influenza in poultry, in Hong Kong in 1997, caused 18 cases of human infection and six deaths. It was controlled by culling of the domestic poultry population, 1.5 million birds in total. Although this brought the initial outbreak to an end, sporadic cases of H5N1 in China during 2003 suggested the virus was still circulating in wild bird populations.

2.27.  The current outbreak of H5N1 avian influenza became apparent in December 2003, with reports of the deaths of large numbers of chickens in the Republic of Korea. As the disease spread laboratory tests identified the cause as the H5N1 virus. In January 2004 Vietnamese health authorities reported an unusual cluster of severe respiratory disease in 11 previously healthy children, of whom seven had died and two were in critical condition. On 8 January Vietnam confirmed that highly pathogenic H5N1 was the cause of the deaths in poultry, and on 11 January H5N1 was confirmed in samples from fatal cases of human infection in Hanoi.

2.28.  The scale of the outbreak soon became apparent: on 12 January Japan reported the detection of H5N1 in poultry; by 19 January five fatal cases had been confirmed in Vietnam; on 23 January Thailand reported H5N1 in humans and poultry; by the end of 2004 nine countries in south east Asia had experienced H5N1 outbreaks in birds and two countries (Thailand and Vietnam) had reported cases in humans. By 29 November 2005 three more countries, China, Cambodia and Indonesia, had confirmed human cases, with total reported cases numbering 133 and deaths 68.[6]

THE START OF A PANDEMIC?

2.29.  For a pandemic to start three conditions need to be met:

• A novel virus subtype must emerge to which the general population will have no or little immunity.
• The new virus must be able to replicate in humans and cause serious illness.
• The new virus must be efficiently transmitted from one human to another; efficient human-to-human transmission is seen as sustained chains of transmission causing community-wide outbreaks.

2.30.  The first two of these conditions have been met: we know that a novel virus (H5N1) has emerged. It has infected people and animals, and has shown that it can replicate and cause serious illness. However, there is no evidence at present of efficient or sustained human-to-human transmission of the virus.

2.31.  With a view to monitoring progress towards the fulfilment of these conditions more closely, the WHO has published a table summarising the various phases of pandemics, and this table is given below:[7]

TABLE 1

WHO Pandemic Phases

Inter-pandemic period
Phase 1	No new influenza virus subtypes have been detected in humans. An influenza virus subtype that has caused human infection may be present in animals. If present in animals, the risk of human infection or disease is considered to be low.
Phase 2	No new influenza virus subtypes have been detected in humans. However, a circulating animal influenza virus subtype poses a substantial risk of human disease.
Pandemic alert period
Phase 3	Human infection(s) with a new subtype, but no human-to-human spread, or at most rare instances of spread to a close contact.
Phase 4	Small cluster(s) with limited human-to-human transmission but spread is highly localized, suggesting that the virus is not well adapted to humans.
Phase 5	Larger cluster(s) but human-to-human spread still localized, suggesting that the virus is becoming increasingly better adapted to humans, but may not yet be fully transmissible (substantial pandemic risk).
Pandemic period
Phase 6	Pandemic phase: increased and sustained transmission in general population.
Post pandemic period: return to inter-pandemic period

2.32.  The world currently stands at Phase 3: cases of human infection have been reported, but no human-to-human spread has been confirmed, and certainly no significant clusters reported. This offers some comfort, given that the H5N1 virus emerged as long ago as 1997, and yet has still not succeeded in adapting to human-to-human transmission. Nor is it known to what extent the current outbreak of H5N1 influenza in birds increases the likelihood of a human pandemic. Thus many uncertainties remain.

2.33.  Even if it is assumed that H5N1 will at some point adapt to human-to-human transmission, the timing is impossible to predict: it could be in early 2006, or it might not be for many years. Nevertheless, on the basis of its summary of pandemic phases, the WHO warn that the world is closer now to an influenza pandemic than at any time since 1968. There are good reasons to treat the possibility of the development and spread of the H5N1 virus very seriously.

The likely impact of a pandemic

2.34.  One of the major difficulties faced by all governments in planning for a possible influenza pandemic is the huge uncertainty over its likely impact. Scientific and economic advances since 1968 should mean that the international community is better prepared for a pandemic than ever before. However, the example of the relatively small-scale outbreak of Severe Acute Respiratory Syndrome (SARS) in 2003 demonstrates the economic cost and social dislocation that can arise as a result of new diseases and, perhaps more significantly, of the control measures taken to prevent their spread.

2.35.  Moreover, the three influenza pandemics of the twentieth century demonstrate the enormous variability of the virus. The 1918 pandemic was among the most destructive and fast-moving ever known: previously healthy young adults dropped dead in the street. The mortality rate among those infected is thought to have been about three percent. In contrast, the 1957 and 1968 viruses were, though affecting very large numbers, of similar severity to normal seasonal influenza, with mortality rates in the range 0.2-0.4 percent.

2.36.  This raises a broader question: what causes mortality in influenza cases? In the case of ordinary seasonal influenza, the "at risk" groups are the elderly, the very young and those with chronic respiratory conditions or depressed immune responses. Members of these groups are susceptible both to direct attack by the virus on the respiratory system, and to secondary infections, such as pneumonia.

2.37.  However, in 1918-19, the highest mortality rates were seen in young adults. It may be that they did not die as a result of the direct action of the virus, but as a consequence of their own immune response. Proteins called cytokines, a normal part of the immune response, can be harmful if over-produced, damaging the lungs and other organs. This could explain why those with the strongest immune response had the highest mortality rates.

2.38.  It is impossible to say what form the H5N1 virus will take, if it acquires human-to-human transmissibility. However, it resembles the 1918 H1N1 virus in virulence, and has hitherto mutated in similar ways; if it continues to do so, remaining a "pure" avian virus, there is concern that it would both bring a significantly higher mortality rate than the 1957 or 1968 pandemics, and that it might kill a disproportionate number of young adults. Against the backdrop of these uncertainties, the Government's assumptions that there would be an attack rate of 25 percent, and a mortality rate of 0.37 percent, producing excess mortality in the United Kingdom of around 50,000, appear to be at the lower end of possible estimates.

2.39.  Similarly, the economic costs of a pandemic are shrouded in uncertainty. The Government's written evidence to this inquiry helpfully put some figures on possible costs to the United Kingdom:

• Illness-related absenteeism from work could take £3 billion to £7 billion off GDP;
• Excess mortality could take a further £1 billion (0.37 percent mortality) to £7 billion (2.5 percent mortality) off GDP;
• Loss of future lifetime earnings could cost in the longer term anything from £21 billion to £172 billion.

2.40.  Our own analysis bears out these figures. However, they are inevitably broad-brush, and subject to enormous uncertainty, for instance over the groups most at risk, who could be either the elderly or the young and economically active. Similarly, the direct consequences of excess mortality and absenteeism do not appear to take into account secondary effects, such as the impact of school closure and resulting absenteeism of working mothers; or the potential disruption to transport networks either as a direct result of illness, or of control measures.

2.41.  In a pandemic such costs would be replicated around the world. The World Bank, in a paper presented to a WHO conference in Geneva on 7-9 November, put the total cost of a pandemic to the world economy speculatively at $800 billion, representing about two percent of global GDP over a whole year (compared with a two percent loss to east Asian GDP as a result of SARS in the second quarter of 2003).[8]

Treatment and prevention

2.42.  In most healthy people, influenza will resolve in seven to ten days. Simple remedies to ease symptoms are all that is needed. However, in the case of a pandemic medical intervention either to prevent or treat infection will be required. Infection can be prevented by vaccination; or the disease itself can be treated by use of antiviral drugs.

VACCINATION

2.43.  The vaccine contains a killed version of the influenza virus, which stimulates the body to produce antibodies. The effect is not immediate—it takes about two weeks for the body to create enough antibodies to resist infection by the live virus.

2.44.  Vaccination against seasonal influenza is a well-established part of health-care in developed countries. The instability of the virus means that new strains emerge constantly, with the result that immunisation is provided every autumn, on the basis of an informed analysis of the likeliest strains that will circulate in the coming winter. Immunisation prevents most cases, although protection varies with age, health status, and how closely the virus strains contained in the vaccine actually match those that ultimately circulate. However, even if the vaccine does not prevent the disease the presence of antibodies can reduce the severity of flu symptoms and decrease the risk of complications.

2.45.  From the point of view of pandemic contingency planning, the crucial point is that influenza vaccines are specific to particular strains of the virus. Vaccination against the currently circulating strains of the H3N2 virus, for instance, will offer no protection against H5N1 strains, and only limited protection against newer strains of H3N2 itself. The Government has recently ordered over two million doses of H5N1 vaccine for at-risk and key workers, based on the current avian virus. However, while this may protect against the current circulating strain of "avian flu", there is no guarantee that it will offer protection against any pandemic strain of H5N1 that emerges in future.

2.46.  This need to tailor vaccines to particular strains means that only when a new pandemic strain has emerged can it be isolated and rendered safe, and only then will it be possible to commence vaccine production. According to Dr John Wood, of the National Institute for Biological Sciences and Control (NIBSC), the process before a viral strain can be made safe for distribution to vaccine manufacturers would take "about ten or eleven weeks". From that point the manufacturers would have to incubate the virus in hens' eggs, harvest, purify and prepare the vaccine, and distribute it to health services, with the result that, in the words of Dr Kevin Bryett of Chiron Vaccines, "from the receipt of the antigen to the first product being available is about four to six months". (QQ 33, 37)

2.47.  This means that under current conditions, from the first outbreak of an influenza pandemic to the availability (initially in small quantities) of vaccine, could take anything from seven to nine months. By this time the first wave of the pandemic, and possibly even the second wave, would have passed.

ANTIVIRALS

2.48.  For rapid response to a pandemic, antiviral drugs will be the most important weapon available to health services. Antivirals work by inhibiting the action of particular proteins on the virus. The four medications available to treat influenza can be divided into two classes:

• Amantadine and rimantadine, so-called "M2 inhibitors", which probably work by preventing the virus from entering the host cell;
• Zanamivir (sold as Relenza) and oseltamivir (sold as Tamiflu), so-called "neuraminidase inhibitors", which work by blocking the release of the virus from an infected cell.

2.49.  These drugs act directly on the virus, rather than stimulating the production of antibodies. They appear to be most effective when used prophylactically—that is, to prevent infection in those who have been exposed to the virus. They are also effective in treating the disease, as long as they are taken within 48 hours of the onset of symptoms—the earlier the better. If used in this way they can reduce the severity and duration of symptoms, and reduce the spread of the disease to others. However, antivirals work only so long as they are being taken: they confer no long-term immunity, and the protection ceases as soon as the course is completed. They are not a substitute for immunisation.

2.50.  In addition, many of the strains of the H5N1 avian influenza circulating in south east Asia show signs of resistance to amantadine. Therefore pandemic preparedness plans have focused on neuraminidase inhibitors, particularly oseltamivir, which, as it is available in tablet form, is easily stored and administered. The Government told us that they were "keeping a very close eye on" any indication of the emergence of resistance to oseltamivir. (Q 249)

OTHER PRECAUTIONS

2.51.  Finally, there are simple precautions which everyone can take to reduce, if not avoid, the risk of infection. Influenza spreads when an infected person sneezes or coughs. Infection occurs when droplets are inhaled, or when the virus is picked up from a contaminated surface and then passed to mouth or nose. The best way to prevent the spread of influenza is by limiting the spread of droplets (using a handkerchief, covering the mouth and nose) and by hand-washing to prevent contamination. The Contingency Plan, and other Government-issued guidance on pandemic influenza, explain these simple precautions fully.

2.52.  In addition, masks may have some value. The virus particles are too small to be caught by a normal surgical or dust mask, and there is no good evidence that the routine wearing of a mask will protect a wearer from catching influenza. However, the mask may slightly reduce the risk of an infected person passing the disease on to others. The mask should be a filtering face-piece mask, have a good fit and be changed regularly.

---

4   In addition each of the various H and N antigens has further subtypes, so that there is, for example, considerable variety within "N1" neuraminidase antigens. Back

5   Jeffrey K Taubenberger et al, "Characterization of the 1918 influenza virus polymerase genes", Nature 437 (5 October 2005), 889-93. Back

6   Source: WHO (http://www.who.int/csr/disease/avian_influenza/country/en/).  Back

7   Source: WHO global influenza preparedness plan (http://www.who.int/csr/resources/publications/influenza/WHO_CDS_CSR_GIP_2005_5/en/index.html). Back

8   Source: Paper presented to WHO conference in Geneva, 7-9 November 2005: http://www.who.int/mediacentre/events/2005/World_Bank_Milan_Brahmbhattv2.pdf.  Back