Influenza pandemics are global epidemics caused by novel influenza A viruses that spread readily from person to person and against which little or none of the population has protective antibody. Unlike seasonal influenza epidemics that occur almost annually, pandemics occur at irregular intervals and are generally associated with higher mortality rates. The most infamous, the 1918-1919 Spanish flu pandemic, resulted in more than 20 million deaths worldwide and a large increase in the U.S. mortality rate (Fig. I).1 Although timing is unpredictable, the occurrence of another pandemic is probably inevitable. In this article, we discuss the origins of pandemic viruses, the differences between pandemic and interpandemic influenza, and the evaluation of novel influenza A viruses, using the 1997 outbreak of influenza A (H5N1) in Hong Kong as an example.
Influenza viruses belong to the family Orthomyxoviridae. The three types (A, B, and C) are differentiated by the antigenicity of their matrix and nucleoproteins.1 Type C viruses cause clinically mild disease and are not discussed further.
Type A viruses are divided into subtypes on the basis of the antigenicity of the surface glycoproteins hemagglutinin and neuraminidase; 15 hemagglutinin and 9 neuraminidase subtypes have been identified to date. The virus genome consists of 8 single-stranded RNA segments with the genes encoding the hemagglutinin and the neuraminidase on different gene segments. The segmented nature of the genome allows the gene segments of two influenza viruses co-infecting the same host cell to reassort and produce influenza viruses with new combinations of genetic material and hemagglutinin and neuraminidase glycoproteins.
Type B viruses infect humans almost exclusively and are capable of causing seasonal epidemics. Type A viruses infect a number of species, including humans, swine, birds, seals, and whales. This makes type A viruses capable of causing both epidemics and pandemics in ways that will be explained. In the 20th century, type A viruses of just three hemagglutinin subtypes and two neuraminidase subtypes caused sustained outbreaks in humans: HlNl, H2N2, and H3N2. However, all known type A subtypes have been isolated from avian species.2 Infected waterfowl and shorebirds excrete high titers of virus in their feces and are asymptomatically infected. In contrast, domestic poultry such as chickens and turkeys can have high mortality rates when infected with certain highly pathogenic avian influenza viruses bearing H5 or H7 hemagglutinin genes.1,2 Avian species can thus serve as a reservoir for introducing viruses with novel hemagglutinin and neuraminidase genes into humans.
Among swine, influenza subtypes containing Hl and H3 and Nl and N2, which are antigenically distinct from human viruses of the same subtypes, have circulated. Thus, swine represent another potential source of new influenza viruses.2-4
The main protective immune responses to influenza infections are directed at hemagglutinin and neuraminidase proteins. The ability of these proteins to change is the hallmark of influenza viruses. This occurs in two major ways: drift and shift. Both types A and B undergo drift by antigenic changes in the hemagglutinin and neuraminidase proteins that arise from point mutations in their genes during viral replication. Antibody made in response to previous influenza infections is ineffective in protecting against a different virus type or subtype, or against a sufficiently drifted antigenic variant of the same type or subtype. Thus, current influenza vaccines contain viruses from each type (A and B) and subtype in circulation among humans. Currently, the trivalent vaccine includes a type B strain, a type A (HlNl) strain, and a type A (H3N2) strain. Strains that are antigenically similar to those anticipated during the upcoming influenza season are selected for the vaccine.1
Shift occurs when a type A virus containing a novel hemagglutinin or novel hemagglutinin and neuraminidase that is immunologically distinct from those that have been circulating in recent years is introduced. Such novel viruses can result in a pandemic in an immunologically naive population if they cause illness in humans and are readily transmitted from person to person. There are at least three possible mechanisms by which antigenic shift can occur: (1) a virus bearing a new hemagglutinin or neuraminidase can arise through genetic reassortment between a nonhuman and a human influenza virus; (2) a nonhuman influenza virus from another species (eg, birds or swine) can infect a human directly without reassortment; or (3) a nonhuman virus may be passed from one species (eg, birds) to humans through an intermediate animal host (eg, swine). The two most recent pandemic viruses were avian-human reassortant (mechanism I).2 Pigs have been proposed to be the "mixing vessel" for the generation of reassortant influenza A viruses between humans and birds because pigs can support replication of both avian and human influenza A viruses.2,3,5,6
Figure 1. Mortality rate from all causes per 100,000 persons, United States, 1900-1996, (Adapted from Armstrong C, Conn LA, Pinner RW. Trends in infectious disease mortality in the United States during the twentieth century. JAMA. 1999;281:61-66.)
Most documented introductions of novel type A subtypes have resulted in either isolated cases or limited outbreaks but not pandemics.2'4 The 1976 swine influenza outbreak among soldiers at Fort Dix, New Jersey, illustrates this.7 In that outbreak of swine HlNl influenza, an estimated 230 persons were infected, 13 were hospitalized, and 1 died. Concerns were raised about the pandemic potential of this virus because the 1976 swine influenza virus was thought to be antigenically similar to the 1918-1919 Spanish flu pandemic virus.1 However, person-to-person spread of the 1976 swine influenza virus was not sustained and illness did not spread to other populations.
Figure 2. Percentage of the total number of excess influenza and pneumonia deaths that occurred among individuals younger than 65 years during and after pandemics. The proportion of excess influenza-related deaths that have occurred among individuals younger than 65 years was highest in the first years after the emergence of new pandemic influenza A viruses and then gradually declined. Although influenza A (HlNl) began to cause illness in 1977 after having disappeared when influenza A (H2N2) emerged in 1957, influenza A (HlNl) has been associated with excess deaths in only 1 year since it reemerged in 1977. Most excess influenza-related deaths since 1968 have been associated with influenza A (H3N2) outbreaks. Influenza-attributed excess deaths are represented by squares for influenza A (HlNl), stars for A (H2N2), and diamonds for A (H3N2) epidemics. (Adapted from Simonsen L, Clarice M|, Schonberger LB, Arden NH, Cox NJ, Fukuda K. Pandemic versus epidemic influenza mortality: a pattern of changing age distribution. } infect Dis. 1998;! 78:53-60.)
EPIDEMIOLOGY OF INTERPANDEMIC VERSUS PANDEMIC INFLUENZA
During interpandemic periods, annual influenza activity generally begins to increase in the late fall and peaks during a 6- to 8- week period from December through March in temperate regions of the northern hemisphere. In temperate regions of the southern hemisphere, influenza activity peaks during June through August, and in tropical regions, influenza can occur any time of the year.1 Although the highest rates of illness occur among school-age children, the highest rates of influenza-related hospitalizations occur among children younger than 2 years8·9 and individuals 65 years and older.10 Overall, an average of approximately 114,000 hospitalizations are attributed to influenza annually in the United States.10 As outlined by Arden in this issue, individuals of any age with certain chronic medical conditions, including chronic heart disease, asthma or other pulmonary conditions, diabetes, renal failure, or immunocompromising illnesses, are at higher risk of influenza-related complications than are their age-matched counterparts.10 An average of 20,000 deaths are attributed to influenza annually, 90% of which occur among individuals 65 years and older.10
In contrast, pandemic influenza occurs infrequently and at irregular intervals. Three pandemics have been documented in the 20th century. During the Spanish flu (HlNl) pandemie of 1918-1919, more than 500,000 deaths occurred in the United States. Nearly half were individuals between 20 and 40 years of age, and a case fatality rate of 30% among pregnant women was reported. The 1957-1958 Asian flu (H2N2) and the 1968-1969 Hong Kong flu (H3N2) were associated with an estimated 34,000 and 70,000 deaths each in the United States alone.10 Mortality rates among young persons were not as striking during the two most recent pandemics. The proportion of total deaths that occurred among individuals younger than 65 years in all three pandemics was highest during the first few years after the pandemics and then gradually declined (Fig. 2).11
Influenza infection rates in all age groups are generally higher during pandemics than during annual epidemics. However, in both pandemic and epidemic influenza, school-age children play an important role in the spread of disease.12
All three pandemics of the 20th century spread throughout the world within a year of initial detection. Outbreaks of Asian flu (H2N2) were first reported in late February 1957 in China, and spread to Hong Kong by early April and to other Asian countries during April and May.13 Quarantine efforts did not curtail the spread. By June, outbreaks among ship passengers and crew bound for ports in the United States and Europe were reported. Small outbreaks in closed populations were detected here in June and July and community-wide outbreaks occurred in August. School openings in the United States and Canada were followed by widespread dissemination of Asian flu in those countries during September and peak mortality occurred by late October. This rapid spread highlights the need for extensive and timely international virologie surveillance to detect the emergence of a new subtype as early as possible and set prevention and control strategies in motion, including the development of appropriate vaccine. During the 21st century, global spread of a highly communicable novel influenza A virus could progress even faster because of increased international travel.
Currently, 110 national influenza laboratories in 83 countries participate in influenza virologie surveillance through the World Health Organization (WHO).14 The purpose of the network is to detect the emergence of novel influenza A subtypes in human populations and the emergence of new drifted variants of known circulating types and subtypes. This information is used to develop recommendations for annual vaccine strain selection. Four WHO Collaborating Centers for Reference and Research on Influenza in Atlanta, London, Melbourne, and Tokyo serve as influenza reference laboratories and conduct detailed molecular and antigenic analysis of viruses from around the world. The WHO Collaborating Center at the Centers for Disease Control and Prevention in Atlanta also develops antisera for inclusion in WHO reagent kits. These kits are sent to WHO national laboratories to assist them in the identification of circulating influenza A subtypes. If a national laboratory is not able to subtype an influenza virus using available reagents, the WHO recommends that the isolate be shipped to one or more collaborating centers for further analysis.
This system was tested in 1997 when the WHO national laboratory in Hong Kong isolated an influenza A virus from a sample collected from a 3-year-old child that could not be subtyped. Virus samples sent to influenza reference laboratories were identified by the WHO national influenza laboratory in Rotterdam, The Netherlands, and confirmed by the Centers for Disease Control and Prevention to be influenza A (H5N1), a subtype that previously had been detected only in birds.15-16
EVALUATION OF A NOVEL INFLUENZA A VIRUS
In 1999, the WHO Influenza Pandemic Preparedness Plan14 outlined the role of the WHO in pandemic preparedness and response, recommended guidelines for preparedness at national and regional levels, and listed the steps for the evaluation of novel type A viruses. When a single novel virus is first identified, the plan recommended: (1) evaluating the possibility of laboratory error or specimen contamination as the source; (2) re-isolation of the virus from original clinical material at a reference laboratory into a substrate acceptable for developing vaccine seed virus; (3) sequencing the viral genome; and (4) evaluating the sensitivity of the new isolate to available antiviral medications. In addition, the WHO recommends that laboratories heighten surveillance and expedite sending viruses that are difficult to subtype to WHO Collaborating Center laboratories. Epidemiologie investigations concerning the potential source of the virus and the possibility of person-to-person transmission should be conducted in conjunction with laboratory studies.
In May 1997, a previously healthy 3-year-old child had a febrile respiratory illness. Despite aggressive medical therapy, the child died of complications, including acute respiratory distress syndrome and Reye syndrome. A trachéal aspirate grew influenza. The virus was identified as type A by immunofluorescence, using highly cross-reactive antibodies against the influenza nucleoprotein. However, the virus could not be subtyped using WHO reagents directed against circulating human type A viruses. Antigenic analysis of the hemagglutinin by hemagglutination inhibition assays indicated that the hemagglutinin gene belonged to the H5 subtype. Molecular analysis confirmed that the hemagglutinin gene was related to previously characterized avian H5 hemagglutinin genes. The neuraminidase gene was related to Nl neuraminidase genes from avian type A viruses.15'17 Susceptibility of this H5N1 virus to the antiviral medications amantadme and rimantadine was established in the laboratory.16
The influenza viruses that caused the 1957 (H2N2) and 1968 (H3N2) pandemics were reassortant viruses deriving some gene segments from avian viruses and others from previously circulating human type A strains.2 Before 1997, it was believed that nonreassortant (or wholly) avian type A viruses did not replicate well in humans. Therefore, an important virologie question was whether the H5N1 virus was a reassortant virus. Molecular analysis established that each of the 8 gene segments was most closely related to the genes of avian influenza A viruses. In other words, the H5N1 virus isolated from the 3-year-old child was an avian influenza A virus that had crossed the species barrier to infect a human directly, without réassortaient and without an intermediate host.16'17
Figure 3. Patients with influenza A (H5N1) illness by age group and illness severity, Hong Kong, 1 997. Patients 1 5 years or older were more likely to have severe illness requiring mechanical ventilation or to die of complications of their illness compared with younger individuals.
Epidemiologie and laboratory studies were conducted to estimate the level of immunity in the Hong Kong population against this new H5N1 strain. Outbreaks of highly pathogenic avian influenza A (H5N1) had occurred among domestic poultry on Hong Kong farms in the 2 months prior to the child's illness. However, before the investigation, it was not known whether individuals could be infected after contact with H5Nl-infected poultry and it was hypothesized that H5-seroprevalence in the general population would be low. A serosurvey was conducted to determine whether individuals in the general population had been infected with H5N1. Among the 419 individuals tested, none were H5-antibody positive.18
An investigation was also conducted to evaluate the origin of the child's infection. The child had a history of exposure to chicks prior to his illness, but an investigation of the school and home did not conclusively identify the source of infection.16
To evaluate whether person-to-person transmission had occurred, serologie testing of individuals potentially exposed to the child, including classmates, household members, neighbors, and health care workers, was done. Among 476 individuals tested, 5 (1.1%) were H5-antibody positive and only 1 of the 5 had been exposed to the child, but not to poultry.18 Although the case appeared to be sporadic and person-to-person transmission was thought unlikely, surveillance among humans and poultry for H5N1 was enhanced in Hong Kong and reagents to detect H5N1 were prepared by the Centers for Disease Control and Prevention and distributed to WHO national laboratories in anticipation of possible future cases.
When two or more human infections with a novel influenza virus have been confirmed, the WHO Pandemic Plan recommends additional steps.14 These include (1) enhancing surveillance and diagnostic capacity in the affected country and in other regions with considerable travelrelated contact; (2) initiating special studies to evaluate transmissibility of the novel virus; and (3) initiating efforts to evaluate vaccine candidates. In November and December 1997, 17 additional cases of human H5N1 illness were detected in Hong Kong. Among the 17 patients, 7 (41%) were male and 10 (59%) were younger than 18 years (age range 1-60 years). All 17 were hospitalized, 7 required mechanical ventilation, and 5 died. Older age was associated with more severe illness; 6 of the 7 patiente who required mechanical ventilation or died were 18 years or older (Fig. 3).19 The patients' illnesses occurred coincident with outbreaks of highly pathogenic avian A (H5N1) influenza among live poultry in Hong Kong markets. Virologie surveillance found that nearly 20% of these birds were infected with H5N1 and their mortality rates were high.18
Partial molecular characterization of all available isolates from humans was undertaken to determine whether genetic réassortaient had occurred between the avian H5N1 virus and circulating human influenza viruses. This analysis revealed that all viruses were avian in origin and no réassortaient had occurred. The Hong Kong H5N1 viruses were closely related to viruses isolated from live birds in Hong Kong markets.20'21
Epidemiologie studies were conducted among exposed family members, social contacts, and health care workers and individuals unlikely to have been exposed to the patients. A cohort study of exposed and unexposed health care workers suggested that limited person-to-person transmission had occurred in the health care setting.22 However, no evidence of person-to-person transmission was found among coworkers in an office setting, among persons exposed during a combined bus and airplane trip, or among a patient's day care classmates19 (Centers for Disease Control and Prevention, unpublished data, June 1995). Person-to-person transmission was difficult to assess among household members because they had environmental exposures similar to those of the patients. Among the household members of 51 patients, 6 (12%) had antibody against H5, but 5 of the 6 also had recent exposure to poultry. Only 1 of the 6 had a history of respiratory illness. This person, one of the cases, was a 2-year-old child who had fever and respiratory symptoms 2 days after exposure to a 5-year-old cousin with culture-proven H5N1 infection. Both were exposed to poultry.19 Although limited person-to-person transmission may have occurred, the primary risk factor for illness was exposure to live poultry.18
On December 29, 1997, the Hong Kong government initiated a large culling operation in which 1.5 million chickens were slaughtered. This was based on surveillance data that suggested widespread H5N1 infection among Hong Kong poultry. Whether the culling operation prevented emergence of an H5 virus that was more easily transmitted from person to person and thus prevented a pandemic cannot be known with certainty. However, since completion of the culling operation, no additional H5N1 human cases have been detected, despite continued enhanced surveillance among both humans and poultry.18
Because H5N1 viruses or their progenitors are likely to continue circulating among wild bird populations, enhanced surveillance among both humans and poultry is ongoing.18 In 1999, two human cases of avian type A (H9N2) illness were detected in Hong Kong by this system and were investigated.23 Enhanced surveillance for H5N1, H9N2, and other potential novel type A viruses remains in effect.
It has been estimated that the next pandemic could result in 20 to 47 million illnesses, 18 to 42 million outpatient visits, 314,000 to 734,000 hospitalizations, and 89,000 to 207,000 deaths in the United States alone.24 Early detection of novel influenza viruses is essential to have any effect on limiting the impact of the next pandemic. Close collaboration between public health sectors and veterinary and agricultural authorities at all levels is essential in this effort, including cooperation among WHO collaborating centers, WHO national laboratories, and the WHO Collaborating Center for the Study of Animal Influenza Viruses.4-14
The importance of early detection for vaccine development is highlighted by the H5N1 Hong Kong example. Because the virus was highly pathogenic for poultry and humans, higher levels of biologic containment were required and the standard method for producing virus for vaccine, involving growing the virus in embryonated hens' eggs, could not be used.4 Thus, antigenicalIy similar viruses for use as seed for vaccine and alternative methods of vaccine production had to be evaluated. In the year 2000, 3 years after H5N1 was first detected in humans, vaccine candidates are still being developed. Although few novel influenza viruses may go on to cause widespread outbreaks among humans, evaluation of these viruses can yield important information regarding needs for enhanced surveillance, innovation in vaccine development, and enhancement of pandemic planning efforts. This is essential for pandemic preparedness.
1. Cox NJ, Kawaoka Y. Orthomyxoviruses: influenza. In: Collier L, Balows A, Sussman M, eds. Topley and Wilson's Microbiology and Microbial Infections, 9th ed. New York: Oxford University Press; 1998:355-433.
2. Webster RG, Bean WJ, Gorman OT, Chambers TM, Kawaoka Y. Evolution and ecology of influenza A viruses. Microbiol Rev. 1992^6:152-179.
3. Brown IH, Harris PA, McCauley JW, Alexander DJ. Multiple genetic reassortment of avian and human influenza A viruses in European pigs, resulting in the emergence of an H1N2 virus of novel genotype. / Gen Virol. 1998;79:2947-2955.
4. Snacken R, Kendal AP, Haaheim LR, Wood JM. The next influenza pandemic: lessons from Hong Kong, 1997. Emerg Infect Dis. 1999;5:195-203. Available at: www.cdc.gov/ncidod/eid/vol5no2/snacken.htm.
5. Kida H, Ito T, Okazaki K, Kawaoka Y, Webster RG. Avian influenza viruses as the origin of pandemic strains: perpetuation in nature and potential for transmission to pigs. In: Brown LE, Hampson AW, Webster RG, eds. Options for the Control of Influenza. Oxford, England: Elsevier Science; 1996:529-536.
6. Ito T, Couceiro JN, Keim S, et al. Molecular basis for the generation in pigs of influenza A viruses with pandemic potential. J Virol 1998;72:7367-7373.
7. Top FH, Russell PK. Swine influenza A at Fort Dix, New Jersey (January-February 1976): IV. Summary and speculation. J Infect Dis. 1977;136(suppl):S376-S380.
8. Izurieta HS, Thompson WW, Kramarz P, et al. Influenza and the rates of hospitalization for respiratory disease among infants and young children. N Engl } Mea. 2000; 342:232-239.
9. Neuzil KM, Mellen BG, Wright PF, Michel EF Jr, Griffin MR. The effect of influenza on hospitalizations, outpatient visits, and courses of antibiotics in children. N Engl } Med. 2000;342:225-231.
10. Centers for Disease Control and Prevention. Prevention and control of influenza: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWK. 2000;49(RR03):l-38.
11. Simonsen L, Clarke MJ, Schonberger LB, Arden NH, Cox NJ, Fukuda K. Pandemic versus epidemic influenza mortality: a pattern of changing age distribution. J Infect Dis. 1998;178:53-60.
12. Glezen WP. Emerging infections: pandemic influeriza. Epidemial Rev. 1996;18:64-76.
13. Dunn FL. Pandemic influenza in 1957: review of international spread of new Asian strain. JAMA. 1958;166:11401148.
14. World Health Organization. Influenza Pandemic Preparedness Plan: The Role of WHO and Guidelines for National and Regional Planning. Geneva, Switzerland: World Health Organization; 1999. Available at: www.who.int/ emc-documents/influenza/whocdscsrcdc991c.html.
15. de Jong JC, Claas EC, Osterhaus AD, Webster RG, Lim WL. A pandemic warning? Nature. 1997;389:554.
16. Subbarao K, Klimov A, Katz J, et al. Characterization of an avian influenza A (H5N1) virus isolated from a child with a fatal respiratory illness. Science. 1998;279:393-396.
17. Claas EC, Osterhaus AD, van Beek R, et al. Human influenza A H5N1 virus related to a highly pathogenic avian influenza virus. Lancet. 1998;351:472-477.
18. Lee SY, Mak KH, Saw TA. The avian flu (H5N1): one year on. Public Health & Epidemiology Bulletin. 1999;8:l-7.
19. Katz JM, Lim W, Bridges CB, et al. Antibody response in individuals infected with avian influenza A (H5N1) viruses and detection of anti-H5 antibody among household and social contacts. / Infect Dis. 1999,180:1763-1770.
20. Suarez DL, Perdue ML, Cox N, et al. Comparisons of highly virulent H5N1 influenza A viruses isolated from humans and chickens from Hong Kong. / Virol. 1998; 72:6678-6688.
21. Shortridge KF, Zhou NN, Guan Y, et al. Characterization of avian H5N1 influenza viruses from poultry in Hong Kong. Virology. 1998;252:331-342.
22. Bridges CB, Katz JM, Seto WH, et al. Risk of influenza A (H5N1) infection among health care workers exposed to patients with influenza A (H5N1), Hong Kong. / Infect Dis. 2000;181:344-348.
23. Peiris M, Yuen KY, Leung CW, et al. Human infection with influenza H9N2. Lancet. 1999;354:916-917.
24. Meltzer MI, Cox NJ, Fukuda K. The economic impact of pandemic influenza in the United States: priorities for intervention. Emerg Infect Dis. 1999;5:659-671. Available at: www.cdc.gov/ncidod/eid/vol5no5/meltzer.htm.