Biology of Influenza
Influenza A and B are the two types of influenza viruses that cause epidemic human disease. Influenza A viruses are categorized into subtypes on the basis of two surface antigens: hemagglutinin and neuraminidase. Since 1977, influenza A (H1N1) viruses, influenza A (H3N2) viruses, and influenza B viruses have circulated globally. Influenza A (H1N2) viruses that probably emerged after genetic reassortment between human A (H3N2) and A (H1N1) viruses also have been identified in some influenza seasons. Both influenza A subtypes and B viruses are further separated into groups on the basis of antigenic similarities. New influenza virus variants result from frequent antigenic change (i.e., antigenic drift) resulting from point mutations that occur during viral replication (13).
Currently circulating influenza B viruses are separated into two distinct genetic lineages (Yamagata and Victoria) but are not categorized into subtypes. Influenza B viruses undergo antigenic drift less rapidly than influenza A viruses. Influenza B viruses from both lineages have circulated in most recent influenza seasons (13).
Immunity to the surface antigens, particularly the hemagglutinin, reduces the likelihood of infection (14). Antibody against one influenza virus type or subtype confers limited or no protection against another type or subtype of influenza virus. Furthermore, antibody to one antigenic type or subtype of influenza virus might not protect against infection with a new antigenic variant of the same type or subtype (15). Frequent emergence of antigenic variants through antigenic drift is the virologic basis for seasonal epidemics and is the reason for annually reassessing the need to change one or more of the recommended strains for influenza vaccines.
More dramatic changes, or antigenic shifts, occur less frequently. Antigenic shift occurs when a new subtype of influenza A virus appears and can result in the emergence of a novel influenza A virus with the potential to cause a pandemic. New influenza A subtypes have the potential to cause a pandemic when they are able to cause human illness and demonstrate efficient human-to-human transmission and there is little or no previously existing immunity among humans (13).
Clinical Signs and Symptoms of Influenza
Influenza viruses are spread from person to person primarily through large-particle respiratory droplet transmission (e.g., when an infected person coughs or sneezes near a susceptible person) (16). Transmission via large-particle droplets requires close contact between source and recipient persons, because droplets do not remain suspended in the air and generally travel only a short distance (<1 meter) through the air. Contact with respiratory-droplet contaminated surfaces is another possible source of transmission. Airborne transmission (via small-particle residue [<5µm] of evaporated droplets that might remain suspended in the air for long periods of time) also is thought to be possible, although data supporting airborne transmission are limited (16—21). The typical incubation period for influenza is 1—4 days (average: 2 days) (13). Adults shed influenza virus from the day before symptoms begin through 5—10 days after illness onset (22,23). However, the amount of virus shed, and presumably infectivity, decreases rapidly by 3—5 days after onset in an experimental human infection model (24,25). Young children also might shed virus several days before illness onset, and children can be infectious for >10 days after onset of symptoms (26). Severely immunocompromised persons can shed virus for weeks or months (27—30).
Uncomplicated influenza illness is characterized by the abrupt onset of constitutional and respiratory signs and symptoms (e.g., fever, myalgia, headache, malaise, nonproductive cough, sore throat, and rhinitis) (31). Among children, otitis media, nausea, and vomiting also are commonly reported with influenza illness (32,33). Uncomplicated influenza illness typically resolves after 3—7 days for the majority of persons, although cough and malaise can persist for >2 weeks. However, influenza virus infections can cause primary influenza viral pneumonia; exacerbate underlying medical conditions (e.g., pulmonary or cardiac disease); lead to secondary bacterial pneumonia, sinusitis, or otitis media; or contribute to coinfections with other viral or bacterial pathogens (34—36). Young children with influenza virus infection might have initial symptoms mimicking bacterial sepsis with high fevers (35—38), and febrile seizures have been reported in 6%—20% of children hospitalized with influenza virus infection (32,35,39). Population-based studies among hospitalized children with laboratory-confirmed influenza have demonstrated that although the majority of hospitalizations are brief (<2 days), 4%—11% of children hospitalized with laboratory-confirmed influenza required treatment in the intensive care unit, and 3% required mechanical ventilation (35,37). Among 1,308 hospitalized children in one study, 80% were aged <5 years, and 27% were aged <6 months (35). Influenza virus infection also has been uncommonly associated with encephalopathy, transverse myelitis, myositis, myocarditis, pericarditis, and Reye syndrome (32,34,40,41).
Respiratory illnesses caused by influenza virus infection are difficult to distinguish from illnesses caused by other respiratory pathogens on the basis of signs and symptoms alone. Sensitivity and predictive value of clinical definitions vary, depending on the prevalence of other respiratory pathogens and the level of influenza activity (42). Among generally healthy older adolescents and adults living in areas with confirmed influenza virus circulation, estimates of the positive predictive value of a simple clinical definition of influenza (acute onset of cough and fever) for laboratory-confirmed influenza infection have varied (range: 79%—88%) (43,44).
Young children are less likely to report typical influenza symptoms (e.g., fever and cough). In studies conducted among children aged 5—12 years, the positive predictive value of fever and cough together was 71%—83%, compared with 64% among children aged <5 years (45). In one large, population-based surveillance study in which all children with fever or symptoms of acute respiratory tract infection were tested for influenza, 70% of hospitalized children aged <6 months with laboratory-confirmed influenza were reported to have fever and cough, compared with 91% of hospitalized children aged 6 months—5 years. Among children who subsequently were shown to have laboratory-confirmed influenza infections, only 28% of those hospitalized and 17% of those treated as outpatients had a discharge diagnosis of influenza (38).
Clinical definitions have performed poorly in some studies of older patients. A study of nonhospitalized patients aged >60 years indicated that the presence of fever, cough, and acute onset had a positive predictive value of 30% for influenza (46). Among hospitalized patients aged >65 years with chronic cardiopulmonary disease, a combination of fever, cough, and illness of <7 days had a positive predictive value of 53% for confirmed influenza infection (47). In addition, the absence of symptoms of influenza-like illness (ILI) does not effectively rule out influenza; among hospitalized adults with laboratory-confirmed infection in two studies, 44%—51% had typical ILI symptoms (48,49). A study of vaccinated older persons with chronic lung disease reported that cough was not predictive of laboratory-confirmed influenza virus infection, although having both fever or feverishness and myalgia had a positive predictive value of 41% (50). These results highlight the challenges of identifying influenza illness in the absence of laboratory confirmation and indicate that the diagnosis of influenza should be considered in patients with respiratory symptoms or fever during influenza season.
Health-Care Use, Hospitalizations, and Deaths Attributed to Influenza
In the United States, annual epidemics of influenza typically occur during the fall or winter months, but the peak of influenza activity can occur as late as April or May (Figure 1). Influenza-related complications requiring urgent medical care, including hospitalizations or deaths, can result from the direct effects of influenza virus infection, from complications associated with age or pregnancy, or from complications of underlying cardiopulmonary conditions or other chronic diseases. Studies that have measured rates of a clinical outcome without a laboratory confirmation of influenza virus infection (e.g., respiratory illness requiring hospitalization during influenza season) to assess the effect of influenza can be difficult to interpret because of circulation of other respiratory pathogens (e.g., respiratory syncytial virus) during the same time as influenza viruses (51—53).
During seasonal influenza epidemics from 1979—1980 through 2000—2001, the estimated annual overall number of influenza-associated hospitalizations in the United States ranged from approximately 55,000 to 431,000 per annual epidemic (mean: 226,000) (7). The estimated annual number of deaths attributed to influenza from the 1990—91 influenza season through 1998—99 ranged from 17,000 to 51,000 per epidemic (mean: 36,000) (6). In the United States, the estimated number of influenza-associated deaths increased during 1990—1999. This increase was attributed in part to the substantial increase in the number of persons aged >65 years who were at increased risk for death from influenza complications (6). In one study, an average of approximately 19,000 influenza-associated pulmonary and circulatory deaths per influenza season occurred during 1976—1990, compared with an average of approximately 36,000 deaths per season during 1990—1999 (6). In addition, influenza A (H3N2) viruses, which have been associated with higher mortality (54), predominated in 90% of influenza seasons during 1990—1999, compared with 57% of seasons during 1976—1990 (6).
Influenza viruses cause disease among persons in all age groups (1—5). Rates of infection are highest among children, but the risks for complications, hospitalizations, and deaths from influenza are higher among persons aged >65 years, young children, and persons of any age who have medical conditions that place them at increased risk for complications from influenza (1,4,5,55—58). Estimated rates of influenza-associated hospitalizations and deaths varied substantially by age group in studies conducted during different influenza epidemics. During 1990—1999, estimated average rates of influenza-associated pulmonary and circulatory deaths per 100,000 persons were 0.4—0.6 among persons aged 0—49 years, 7.5 among persons aged 50—64 years, and 98.3 among persons aged >65 years (6).
Among children aged <5 years, influenza-related illness is a common cause of visits to medical practices and emergency departments. During two influenza seasons (2002—03 and 2003—04), the percentage of visits among children aged <5 years with acute respiratory illness or fever caused by laboratory-confirmed influenza ranged from 10%—19% of medical office visits to 6%—29% of emergency departments visits during the influenza season. Using these data, the rate of visits to medical clinics for influenza was estimated to be 50—95 per 1,000 children, and to emergency departments 6—27 per 1,000 children (38). Retrospective studies using medical records data have demonstrated similar rates of illness among children aged <5 years during other influenza seasons (33,56,59). During the influenza season, an estimated 7—12 additional outpatient visits and 5—7 additional antibiotic prescriptions per 100 children aged <15 years has been documented when compared with periods when influenza viruses are not circulating, with rates decreasing with increasing age of the child (59). During 1993—2004 in the Boston area, the rate of emergency department visits for respiratory illness that was attributed to influenza virus based on viral surveillance data among children aged <7 years during the winter respiratory illness season ranged from 22.0 per 1000 children aged 6—23 months to 5.4 per 1000 children aged 5—7 years (60).
Rates of influenza-associated hospitalization are substantially higher among infants and young children than among older children when influenza viruses are in circulation (Figure 2) and are similar to rates for other groups considered at high risk for influenza-related complications (61—66), including persons aged >65 years (59,63). During 1979—2001, the estimated rate of influenza-associated hospitalizations, using a national sample of hospital discharges of influenza-associated hospitalizations in the United States among children aged <5 years, was 108 hospitalizations per 100,000 person-years (7). Recent population-based studies that measured hospitalization rates for laboratory-confirmed influenza in young children documented hospitalization rates that are similar to or higher than rates derived from studies that analyzed hospital discharge data (33,35,36,38,65). Annual hospitalization rates for laboratory-confirmed influenza decrease with increasing age, ranging from 240—720 per 100,000 children aged <6 months to approximately 20 per 100,000 children aged 2—5 years (38). Hospitalization rates for children aged <5 years with high-risk medical conditions are approximately 250—500 per 100,000 children (56,58,67).
Influenza-associated deaths are uncommon among children. An estimated annual average of 92 influenza-related deaths (0.4 deaths per 100,000 persons) occurred among children aged <5 years during the 1990s, compared with 32,651 deaths (98.3 per 100,000 persons) among adults aged >65 years (6). Of 153 laboratory-confirmed influenza-related pediatric deaths reported during the 2003—04 influenza season, 96 (63%) deaths occurred among children aged <5 years and 61 (40%) among children aged <2 years. Among the 149 children who died and for whom information on underlying health status was available, 100 (67%) did not have an underlying medical condition that was an indication for vaccination at that time (68). In California during the 2003—04 and 2004—05 influenza seasons, 51% of children with laboratory-confirmed influenza who died and 40% of those who required admission to an intensive care unit had no underlying medical conditions (69). These data indicate that although deaths are more common among children with risk factors for influenza complications, the majority of pediatric deaths occur among children of all age groups with no known high-risk conditions. The annual number of deaths among children reported to CDC for the past four influenza seasons has ranged from 84 during 2004—05 to 84 during 2007—08 (CDC, unpublished data, 2008).
Death associated with laboratory-confirmed influenza virus infection among children (defined as persons aged <18 years) is a nationally reportable condition. Deaths among children that have been attributed to co-infection with influenza and Staphylococcus aureus, particularly methicillin resistant S. aureus (MRSA), have increased during the preceding four influenza seasons (70; CDC, unpublished data, 2008). The reason for this increase is not established but might reflect an increasing prevalence within the general population of colonization with MRSA strains, some of which carry certain virulence factors (71,72).
Hospitalization rates during the influenza season are substantially increased for persons aged >65 years. One retrospective analysis based on data from managed-care organizations collected during 1996—2000 estimated that the risk during influenza season among persons aged >65 years with underlying conditions that put them at risk for influenza-related complications (i.e., one or more of the conditions listed as indications for vaccination) was approximately 560 influenza-associated hospitalizations per 100,000 persons, compared with approximately 190 per 100,000 healthy elderly persons. Persons aged 50—64 years with underlying medical conditions also were at substantially increased risk for hospitalizations during influenza season, compared with healthy adults aged 50—64 years. No increased risk for influenza-associated hospitalizations was demonstrated among healthy adults aged 50—64 years or among those aged 19—49 years, regardless of underlying medical conditions (64).
Influenza is an important contributor to the annual increase in deaths attributed to pneumonia and influenza that is observed during the winter months (Figure 3). During 1976—2001, an estimated yearly average of 32,651 (90%) influenza-related deaths occurred among adults aged >65 years (6). Risk for influenza-associated death was highest among the oldest elderly, with persons aged >85 years 16 times more likely to die from an influenza-associated illness than persons aged 65—69 years (6).
The duration of influenza symptoms is prolonged and the severity of influenza illness increased among persons with human immunodeficiency virus (HIV) infection (73—77). A retrospective study of young and middle-aged women enrolled in Tennessee’s Medicaid program determined that the attributable risk for cardiopulmonary hospitalizations among women with HIV infection was higher during influenza seasons than it was either before or after influenza was circulating. The risk for hospitalization was higher for HIV-infected women than it was for women with other underlying medical conditions (78). Another study estimated that the risk for influenza-related death was 94—146 deaths per 100,000 persons with acquired immunodeficiency syndrome (AIDS), compared with 0.9—1.0 deaths per 100,000 persons aged 25—54 years and 64—70 deaths per 100,000 persons aged >65 years in the general population (79).
Influenza-associated excess deaths among pregnant women were reported during the pandemics of 1918—1919 and 1957—1958 (80—83). Case reports and several epidemiologic studies also indicate that pregnancy increases the risk for influenza complications (84—89) for the mother. The majority of studies that have attempted to assess the effect of influenza on pregnant women have measured changes in excess hospitalizations for respiratory illness during influenza season but not laboratory-confirmed influenza hospitalizations. Pregnant women have an increased number of medical visits for respiratory illnesses during influenza season compared with nonpregnant women (90). Hospitalized pregnant women with respiratory illness during influenza season have increased lengths of stay compared with hospitalized pregnant women without respiratory illness. Rates of hospitalization for respiratory illness were twice as common during influenza season (91). A retrospective cohort study of approximately 134,000 pregnant women conducted in Nova Scotia during 1990—2002 compared medical record data for pregnant women to data from the same women during the year before pregnancy. Among pregnant women, 0.4% were hospitalized and 25% visited a clinician during pregnancy for a respiratory illness. The rate of third-trimester hospital admissions during the influenza season was five times higher than the rate during the influenza season in the year before pregnancy and more than twice as high as the rate during the noninfluenza season. An excess of 1,210 hospital admissions in the third trimester per 100,000 pregnant women with comorbidities and 68 admissions per 100,000 women without comorbidities was reported (92). In one study, pregnant women with respiratory hospitalizations did not have an increase in adverse perinatal outcomes or delivery complications (93); however, another study indicated an increase in delivery complications (91). However, infants born to women with laboratory-confirmed influenza during pregnancy do not have higher rates of low birth weight, congenital abnormalities, or low Apgar scores compared with infants born to uninfected women (88,94).