Intermittent influenza epidemics have been chronicled by clinicians and medical historians since at least the 15th century. The impact of influenza can be detected by excess mortality rates and increased rates of hospitalization or outpatient visits during most influenza seasons. However, influenza can also cause an illness that is clinically indistinguishable from that of other respiratory pathogens, especially among children,
In its classic form, influenza infection is characterized by sudden onset of fever (often with chills) accompanied by respiratory symptoms such as dry cough, sore throat, and coryza. Systemic symptoms include headache, myalgia, arthralgia, and extreme fatigue. Ocular signs and symptoms are common, including conjunctival injection, photophobia, tearing, and periorbital pain. In younger children, influenza may present as the typical illness or as a nonspecific febrile illness, croup, bronchitis, bronchiolitis, or pneumonitis. Younger children are more likely than older children and adults to have high fevers, nausea, vomiting, or diarrhea. Influenza in infants can be difficult to recognize, and may present with fever and nonspecific signs, including lethargy and poor feeding. Respiratory signs may be mild or absent.1'2
THE INFLUENZA VIRUSES
Influenza viruses consist of segmented, singlestranded RNA, which is referred to as "negative sense" in that it serves as a template for transcription of complementary viral mRNA. The RNA segments, which are surrounded by a lipid envelope, each encode one or two viral proteins.
The influenza viruses are classified into three groups: types A, B, and C. Types A and B each have eight gene segments, whereas type C has seven. Type A viruses are further classified into subtypes based on the antigenic properties of the two surface glycoproteins, hemagglutinin and neuraminidase. Hemagglutinin is responsible for virus attachment (to cell receptors bearing terminal sialic acid residues), adsorption, and fusion. Neuraminidase is an enzyme that removes sialic acid, allowing release of new virus progeny from infected cells, and helps to prevent these viruses from clumping after release. Type C has only one type of surface glycoprotein that performs both receptor recognition and cleavage.3
Types A and B can also be distinguished from type C by epidemiologie characteristics. Types A and B both cause recurrent epidemics. Type C can cause respiratory illness, is usually mild, and does not occur in epidemic form. Although epidemics during which type B predominates do not result in excess mortality rates as often as those in which type A (H3N2) predominates, type B infection can cause serious illnesses and complications and cannot be distinguished clinically from type A infection in an individual patient.2'4
ANTIGENIC DRIFT AND SHIFT
Type A and B viruses cause frequent, recurrent epidemics based on two kinds of antigenic changes in hemagglutinin and neuraminidase surface proteins. The first is antigenic drift, which occurs in both type A and B viruses. Antigenic drift is due to an accumulation of hemagglutinin and neuraminidase point mutations that gradually change the amino acid structure of parts of these molecules, which changes their antigenic properties. Over time, hemagglutinin and neuraminidase are no longer recognized by antibody produced by an earlier antigenic variant, so the host is again susceptible to infection. This kind of change is responsible for the epidemics that occur almost every year.
The second change is referred to as antigenic shift, and is unique to type A viruses. Antigenic shift is the abrupt appearance of a new subtype of type A, which contains a novel hemagglutinin, neuraminidase, or both. There is evidence that this has generally occurred as a result of genetic reassortment among human and nonhuman influenza virus strains (see the article by Subbarao and Bridges in this issue for additional information). When antigenic shift takes place, a global epidemic, or pandemic, can occur, with high attack rates in human populations throughout the world.1,4,5
PANDEMIC AND INTERPANDEMIC INFLUENZA
During the 20th century, influenza pandemics occurred in 1918-1919, 1956-1957, and 1968-1969. In each pandemic, a novel influenza A subtype emerged and the previously circulating subtype disappeared. Each was also associated with increased mortality compared with interpandemic periods (see the article by Subbarao and Bridges in this issue for additional information).
In 1977, an influenza virus of the A (HlNl) subtype appeared that was virtually identical to one that had circulated in the early 1950s, taking influenza researchers by surprise and upsetting some previous theories. Adding to this surprise was the fact that the type A (H3N2) virus that had emerged in 1968 continued to circulate rather than disappearing, as previous type A viruses had done following an antigenic shift.4 Both type A (HlNl) and A (H3N2) subtypes continue to co-circulate and undergo antigenic drift. However, since their reemergence, type A (HlNl) viruses have been the predominant strain during only three influenza seasons in the United States. Adults have also been less likely to become ill when exposed to type A (HlNl) viruses compared with type A (H3N2), perhaps because many had experienced type A (HlNl) as their first influenza infection of childhood. Thus, in epidemics since 1977 where type A (HlNl) viruses predominate, pediatrie attack rates are disproportionately high compared with those of adults.
The recent discovery and identification of the human A (H5N1) infections in Hong Kong came about as a result of global surveillance. The identification, investigation, and containment of this virus before it spread illustrates the value of global influenza surveillance with international cooperation and collaboration. This system is described in more detail in the article by Subbarao and Bridges in this issue.
Influenza virus nomenclature was developed to facilitate global surveillance. The first part of the name refers to the type of virus. This is followed by its geographic origin, the strain number, and the year of isolation. For type A viruses, the subtype is also included. For example, A/New Caledonia /20 /99 (HlNl) was the 20th influenza A virus isolated in New Caledonia during 1999.2
Our current system for national influenza surveillance in the United States also includes systems to identify and monitor the prevalence of circulating strains, detect new strains, and assess the progress of influenza during each season. Thus, surveillance can help identify outbreaks, assess vaccine effectiveness, and assist in disease control. In collaboration with state health departments, the incidence of influenza-like illness and die virus strains that are circulating in each state are assessed and reported weekly from October through May. The impact of influenza on mortality is estimated using vital statistics data from 122 major metropolitan areas. More detailed analyses of influenza-associated mortality are conducted after each season using mortality data from the entire United States.2-6
VIRUS TRANSMISSION AND AGE-SPECIFIC ATTACK RATES
Influenza virus is spread from person to person by virus-laden respiratory secretions. There is evidence that small-particle aerosols are the primary mode of transmission. Large-particle droplets can also be infectious, but experimental studies have shown that substantially higher doses are needed to cause infection compared with inoculation by small-particle aerosols. The relative efficiency of small-particle aerosols can also be demonstrated by strong evidence that a single infected person can transmit virus to a large number of susceptible people in a closed environment.4
Once virus is deposited in the upper respiratory tract of a susceptible person, the principal site of replication is the columnar epithelium. Depending in part on the dose of the inoculum, the incubation period varies from 18 hours to 5 or more days, but is most commonly 2 to 3 days. Viral titers often continue to rise following illness onset, and virus shedding varies from person to person but usually lasts from 3 to 7 days. However, viral titers are generally higher in young children than in older children or adults, and virus shedding can last 10 days or longer.1
In view of their relatively high attack rates and their larger quantity and duration of virus shedding, it is not surprising that children play an important role in transmission within households and tìie community. Many studies have shown that the highest influenza attack rates occur among school-age children during most epidemics. Although generally lower than those among older children, high attack rates have also been found among preschool children. Influenza illness attack rates have been shown to be higher among persons of all ages who live in households with children who attend school or day care compared with those who do not.4'7
INFLUENZA IN INFANTS
Attack rates are more difficult to quantify among infants, in part because it is more difficult to recognize influenza in this group. Consequently, many epidemiologie studies have not reported separate attack rates among infants compared with children older than 1 year. Although some studies have separated data for children younger than 1 year, few have made distinctions among different age groups within the first year of life. One such study found that rates of infection were approximately twice as high during the second 6 months of life compared with the first, and that younger infants who became infected experienced illnesses that were milder than those experienced by older infected infants.7
These two phenomena are thought to result from maternally derived antibody. Another study showed that infants born to mothers with detectable serum antibody to influenza A had higher antibody titers than did those born to seronegative mothers. Although infants with anti-influenza antibody observed in this study did not have a lower incidence of influenza infection than seronegative infants, they did have a shorter mean duration of illness. However, maternally derived antibody usually wanes to very low levels by the age of 4 months.8
Rates of influenza-associated hospitalization are higher among infants and younger children compared with older children and healthy young adults. Estimates of age-specific rates of influenza-associated hospitalization have varied considerably in different studies, partly because these involved different age groups in different influenza epidemics and different methods. Among those 4 years and younger, estimated rates have ranged from approximately 100 per 100,000 healthy children to 500 per 100,000 population in the same age group with medical conditions that put them at higher risk for influenzaassociated complications.9
A recent study of those without high-risk conditions in the birth to 4 years group has concluded that infants younger than 6 months have by far the highest risk of influenza-related hospitalization, at approximately 1,040 per ???,???.10 This and other studies suggest that infants younger than 6 months have at least as high a risk of hospitalization as individuals who are 65 years and older. The latter have an estimated risk ranging from approximately 200 to 1,000 per 100,000 population during different influenza epidemics. Estimated rates of influenza-associated hospitalization among healthy children younger than 2 years and children between 2 and 4 years have been approximately 200 per 100,000(10,11) and 8 to 136 per 100,000, respectively. Estimated rates among high-risk children 5 to 14 years old are fivefold to tenfold higher than rates among healthy children in the same age group (200 per 100,000 vs 20 to 40 per 100,000) and higher than estimated rates among high-risk individuals 15 to 44 years old (40 to 60 per 100,000(9); Table).
MORTALITY ASSOCIATED WITH INFLUENZA
Individuals 65 years and older and those 50 to 64 years with high-risk conditions have the highest influenza mortality rates. Some researchers have concluded that excess influenza-associated mortality rates have not been demonstrated among infants and children in recent decades.1 However, other evidence suggests that influenza is associated with fatal illness relatively frequently in children, but this has not been apparent in excess mortality calculations because other respiratory viruses that circulate outside of the influenza season also cause life-threatening infections in children (eg, respiratory syncytial virus and parainfluenza types 1 and 3). Thus, deaths due to viral respiratory infections and their complications in children occur throughout the year and, therefore, an excess death rate is not detected during the influenza season.12
Estimated Age-Specific Rates of Influenza- Associated Hospltallzatlon In the United States Among Persons With and Without High Risk Conditions
A variety of techniques will diagnose influenza infection. The two standard methods are virus isolation in cell culture or embryonated hens' eggs (using nasopharyngeal or throat swab specimens, or, less commonly, nasal wash specimens) or demonstration of a rise in strain-specific antibodies between acute and convalescent serum specimens. Serum antibody may be detected as early as 4 to 7 days after onset of illness and will reach peak levels after 2 to 3 weeks. Because some proportion of the population usually possesses specific or cross-reactive antibody to prevalent influenza strains, both acute and convalescentphase specimens are essential for a serologie diagnosis. Serodiagnosis can be more difficult in young children because a significant number may not have a detectable rise in antibody after initial influenza infection.2,13
Neither serologie diagnosis nor traditional virus isolation yield results timely enough to provide guidance in treatment. A number of commercially available assays can now rapidly detect influenza from nasopharyngeal or throat swab specimens in the clinical setting. One is an enzyme immunoassay that uses a membrane filter and monoclonal antibodies to detect antigen; a purple triangle develops on the membrane to indicate a positive result, but this test will detect only type A.14 The other rapid diagnostic tests for use at the care site detect types A and B, but do not distinguish between them. One is a colonmetric enzyme assay that detects the influenza neuraminidase enzyme.15 Another is an optical immunoassay that detects antibody-antigen interactions on a slide.16 The newest test is an immunoassay using monoclonal antibodies. A color indicator on a test strip is used to determine whether influenza virus antigens are present in throat swab, nasal wash, or nasal aspirate specimens.17
The use of rapid diagnostic testing can greatly facilitate early detection and provide guidance for the treatment of patients with acute febrile respiratory illness. This is particularly useful now that there are antiviral agents effective against types A and B oseltamivir and zanamivir), as well as others effective against only type A (amantadine and rimantadine). The article by Subbarao and Bridges in this issue provides more information on these agents.
It is not necessary to test every patient once a community outbreak of influenza has been confirmed. Knowledge of national, state, and local influenza surveillance data disseminated by the Centers for Disease Control and Prevention and state and local health departments can help determine which antiviral agents or rapid diagnostic assays to use in a given influenza season. During most seasons, only one virus type predominates in a given geographic area at a given time. More unusual seasons may see both types A and B circulating simultaneously, so two tests may occasionally be needed to determine the etiology of a given illness. When suspected or confirmed outbreaks occur, especially in hospitals or residential long-term-care facilities where high-risk patients will be exposed, specimens should also be sent to state or local health departments for viral culture. This is necessary to characterize the virus and determine how similar it is to vaccine strains, or to test for other agents if influenza is not detected. This also contributes to the virologie data needed for annual influenza vaccine strain selection.2-9
1. Van Voris LP, Young JF, Bernstein JM, et al. Influenza viruses. In: Belshe BB, ed. Textbook of Human Virology. Littleton, MA: PSG Publishing; 1984:267-297.
2. LaForce FM, Nichol KL, Cox NJ. Influenza: virology, epidemiology, disease, and prevention. Am ] Prev Med. 1994;10(suppl):31-44.
3. Ruigrok RWH. Structure of influenza A, B and C viruses. In: Nicholson KG, Webster RG, Hay AJ. Textbook of influenza. Oxford, England: Blackwell Science; 1998:29-42.
4. Noble GR. Epidemiological and clinical aspects of influenza. In: Beare AS, ed. Basic and Applied Influenza Research. Boca Ratón, FL: CRC Press; 1982:1277-1282.
5. Kilbourne ED. Influenza. New York: Plenum Medical; 1987:255-282.
6. Simonsen L, Clarke MJ, Williamson GD, Stroup DF, Arden NH, Schonberger LB. The impact of influenza on mortality: introducing a severity index. Am ] Public Health. 1997;87:1944-1950.
7. Taber LH, Paredes A, Glezen WP, Couch RB. Infection with influenza A /Victoria virus in Houston families, 1976. Journal of Hygiene, Cambridge. 1981;86:303-312.
8. Reuman PD, Ayoub EM, Small PA. Effect of passive maternal antibody on influenza illness in children: a prospective study of influenza A in mother-infant pairs. Pediatr Infect Dis J. 1987;6:398-403.
9. Centers for Disease Control and Prevention. Prevention and control of influenza: recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR. 2000;49(RR03):l-38.
10. Neuzil KM, Mellen BG, Wright PF, Mitchel 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.
11. Izurieta HS, Thompson WW, Kramarz P, et al. Influenza and the rates of hospitalization for respiratory disease among infants and young children. N Engl J Med. 2000;342:232-239.
12. Glezen WP. Consideration of the risk of influenza in children and indications for prophylaxis. Rev Infect Dis. 1980;2:408~420.
13. Shaw MW, Arden NH, Maassab HF. New aspects of influenza viruses, din Microbici Rev. 1992;5:74-92.
14. Directigen Flu A [package insert], Cockeysville, MD: Becton Dickinson Microbiology Systems; 1996.
15. Zstat Flu [package insert]. Oklahoma City, OK: ZymeTx Inc.; 1999.
16. FLU OLA [package insert]. Boulder, CO: Biostar, Inc.; 1999.
17. QuickVue [package insert]. San Diego, CA: Quidel; 2000.
Estimated Age-Specific Rates of Influenza- Associated Hospltallzatlon In the United States Among Persons With and Without High Risk Conditions