Chapter 195: Influenza

The term influenza represents both a clinically defined respiratory illness accompanied by systemic symptoms of fever, malaise, and myalgia and the name of the orthomyxoviruses that cause this syndrome. Although this term is sometimes used more generally to denote any viral respiratory illness, many features distinguish influenza from these other illnesses, most particularly its systemic symptoms, its propensity to cause sharply peaked winter epidemics, and its capacity to spread rapidly among close contacts. The morbidity and mortality associated with influenza epidemics are documented closely in the United States by the Centers for Disease Control and Prevention (CDC), which records clinical cases of influenza-like illness, cases of virologically documented influenza, and excess deaths due to pneumonia and influenza combined.

Three influenza viruses occur in humans: A, B, and C. These viruses are irregularly circular in shape, measure 80–120 nm in diameter, and have a lipid envelope and prominent spikes that are formed by the two surface glycoproteins, hemagglutinin (H) and neuraminidase (N) (Fig. 195-1). The hemagglutinin functions as the viral attachment protein, binding to sialic acid receptors on the cells that line the superficial epithelium of the respiratory tract. The neuraminidase cleaves the virus from the cell membrane to facilitate its release from the cell and prevents self-aggregation of viruses. Influenza A viruses have eight single-strand negative-sense RNA segments in their genomes that encode hemagglutinin and neuraminidase as well as internal genes, including polymerase, matrix, nucleoprotein, and nonstructural genes. The segmented nature of the genome allows gene reassortment; an analogy for reassortment is the shuffling of a deck of cards. Reassortment takes place when a single cell is infected with two different strains.

Among the influenza viruses, the A viruses are of paramount importance for several reasons: (1) the plasticity of their genomes, which enables them to react to the prevailing immunity in the community by modifying their immunogenic epitopes, particularly on the hemagglutinin surface protein (antigenic drift); (2) the segmentation of their genomes, which allows genes coding both surface and internal proteins to be reassorted between influenza A variants (antigenic shift); and (3) their extensive mammalian and avian reservoirs, in which multiple variants with distinct hemagglutinin and neuraminidase genes lie in wait. As a result of all of these factors, influenza A virus has the ability, particularly after an antigenic shift, to cause a worldwide epidemic (pandemic). The most severe influenza A pandemic in modern history took place in 1918; ~50 million deaths were attributed to the culpable influenza A H1N1 virus in the years surrounding 1918.

The influenza A viruses are further classified by their surface glycoproteins (H and N), the geographic location of their isolation, their sequential number among isolated viruses, and their year of isolation. Thus, the influenza vaccine for the 2017–2018 season in the Northern Hemisphere was formulated to provide protection against influenza A/Michigan/45/2015 (H1N1)pdm09–like virus, influenza A/Hong Kong/4801/2014 (H3N2)–like virus, and two lineages in the influenza B family: B/Brisbane/60/2008–like virus (Victoria lineage) and B/Phuket/3073/2013–like virus (Yamagata lineage).

Influenza virus causes outbreaks during the cooler months of the year and thus has a mirror-image season in the antipodes compared with that in the Northern Hemisphere. The circulation of strains in the Southern Hemisphere has some predictive value for vaccine composition in the Northern Hemisphere, and vice versa. This information is important as the degree of antigenic drift is one determinate of vaccine efficacy. Vaccine composition typically must change in at least one component yearly in anticipation of the predicted circulating strains.

A typical outbreak begins in early winter and lasts 4–5 weeks in a given community, although its impact on the country as a whole will be of considerably longer duration. When excess mortality occurs, an influenza outbreak is classified as an epidemic. Influenza’s impact is reflected in increased school and work absenteeism, increased visits to emergency rooms and primary care physicians, and increased hospitalizations, particularly of elderly patients and individuals with underlying cardiopulmonary disease. The impact often is most easily recognized in the pediatric population, whose school absenteeism quickly peaks. Despite efforts to limit influenza spread through vaccination, cohorting, use of masks, and hand washing, long-term-care facilities house another sentinel population, including many elderly patients who are at increased risk of complicated disease.

Influenza is largely spread by small- and large-particle droplets; spread is undoubtedly facilitated by the coughing and sneezing that accompany the illness. Within families, the illness is often introduced by a preschool or school-aged child.

Influenza’s global spread and causative strain(s) in a given year are well documented by the surveillance networks of the World Health Organization (WHO) and the CDC. The severity of an epidemic depends on the transmissibility and virulence of the viral strain, the susceptibility of the population, the adaptation of the virus to its human host, and the degree of antigenic match to the recommended vaccine. None of these parameters is totally predictable for influenza A.

Influenza A Viruses

When a major shift in the hemagglutinin and/or the neuraminidase occurs, with introduction of a new serotype from an animal or avian reservoir, an influenza A strain has the potential to cause a pandemic. In modern influenza history, such shifts occurred in 1918 (H1N1), 1957 (H2N2), 1968 (H3N2), 1977 (H1N1), and 2009 (H1N1pdm) (Table 195-1). On the basis of seroarchaeology (the analysis of serum antibody profiles in the elderly), epidemics that took place in the 1890s have been attributed to H3N2 and H2N2 viruses. Epidemics typical of influenza have been documented throughout recorded history.

In some epidemics, a younger age group proves especially susceptible. This is the case with current H1N1 epidemics, where individuals born before 1968 had likely been exposed to related viral strains and thus were relatively protected against the current strain. The 1918 epidemic was striking in this regard: the most severely infected individuals were infants and previously healthy young adults—the latter being a group not typically found to have high influenza mortality (Fig. 195-2). The 1918 epidemic increased all-cause mortality and led to more deaths than all military losses in World War I. In spite of the attention paid to the risk and impact of pandemic disease, it is generally appreciated that—with the exception of 1918—cumulatively more illness occurs during yearly epidemics combined than in pandemics.

All of the annual influenza A epidemics in the past 50 years have been caused by H1N1 and/or H3N2 strains. H2N2 strains circulated between 1957 and 1968, and H1N1 strains circulated prior to that, including in 1918. However, potentially pandemic viruses continue to emerge, mostly in Asia, with higher-numbered hemagglutinins (e.g., H5, H6, H7, H9) reflecting some of the 16 distinct H and 9 distinct N subtypes in avian reservoirs. Most cases of these potentially pandemic illnesses have occurred in individuals who have had direct contact with domesticated birds or who have visited live-bird markets, which are common in Asia. In addition to the global aeronautic movement of infected people, bird migration is one mechanism for rapid global spread. It is not clear why higher-numbered avian hemagglutinin strains have not acquired the degree of transmissibility necessary to cause pandemic disease.

Avian and Swine Influenza Viruses

The full panoply of influenza viruses is found in domestic and migratory wild birds. It is postulated that epithelial cells in the swine respiratory tract may play a specific role as a “mixing vessel,” allowing the reassortment of genes from avian and human sources and thereby permitting the transition of avian viruses to humans. The nature of the sialic acid receptors for influenza virus hemagglutinin partially accounts for host preference. Humans have largely α-2,6-galactose receptors, while birds have α-2,3-galactose receptors. Swine have both types of receptors on their respiratory epithelial cells—hence their postulated role in facilitating reassortment and host adaptation of avian strains to growth in humans. Strains such as 2009 H1N1pdm (pandemic) had genes of avian, swine, and human origin. Some avian strains—notably H5 strains—are highly pathogenic in humans, as was the 1918 strain. The reasons for the high pathogenicity of certain strains are not entirely clear. Virulence and transmissibility often appear to be separate genetic traits.

After the sequencing of the 1918 virus recovered from the lungs of bodies buried in the Arctic permafrost, the virus was genetically reconstructed under carefully controlled isolation conditions. In animal studies of this viable 1918 virus, both the hemagglutinin and the ribonucleoprotein contributed to high levels of replication accompanied by an abnormally enhanced innate immune response characterized by proinflammatory cytokines. Perhaps this “cytokine storm” is the best explanation for the enhanced illness occurring in young, immunologically vigorous individuals in the 1918 pandemic. Sequencing demonstrated that the 1918 virus was of avian origin. Although the 1918 virus was first identified in military camps in the United States, its impact cannot be attributed to the disruption of war: the illness was well documented in countries such as Iceland that were not directly involved in World War I.

The same concerns about a “cytokine storm” have been raised with regard to the H5N1 viruses that first emerged in Hong Kong in 1996. These viruses exhibited high pathogenicity in individuals who had direct contact with domestic fowl, with mortality rates close to 50%, but also displayed poor human-to-human transmissibility. Pathogenicity appears to be a function not just of the viruses’ surface proteins, but also of an optimal gene constellation including all eight segmented influenza genes. However, unlike the 1918 strain, the H5N1 viruses have, to date, caused only sporadic disease, as have other limited clusters of a highly pathogenic H7N9 virus.

Influenza B and C Viruses

The influenza B viruses are more genetically stable than the influenza A viruses and have no animal reservoirs. Two lineages of influenza B have circulated for the past 40 years (B/Yamagata-like and B/Victoria-like viruses), and it has proven very difficult to predict which strain will be dominant in a given year. This issue has led to the incorporation of representatives of both influenza B lineages plus influenza A/H1N1 and H3N2 viruses into a quadrivalent vaccine.

Influenza C viruses cause intermittent mild disease and have attracted little attention. These viruses have been the subject of fewer than 10 publications annually since the year 2000.

Influenza-Associated Morbidity and Mortality

Although epidemics vary in severity and in the age groups most affected, certain high-risk groups are seen in all epidemics (Table 195-2). These groups are assigned the highest priority for vaccination and other preventive and therapeutic measures. Their caregivers and close contacts are also prioritized targets of interventions. A generalization is that the relative impact of an epidemic is seen in the youngest age group with the least prior exposure—and therefore the least immunity—to influenza. The impact of influenza can be depicted as a pyramid of illnesses, medical visits, hospitalizations, and deaths (Fig. 195-3).

Pneumonia and influenza mortality, reported as excess over the anticipated sine-wave curve of deaths during the year, is seen in the CDC’s data for 2012–2017 (Fig. 195-4). In addition to excess respiratory deaths directly attributed to influenza, an increase in circulatory deaths also occurs during an influenza epidemic.

At a cellular level, influenza virus binds to sialic acid receptors and enters the epithelial cell through receptor-mediated endocytosis. The virus then enters an endosome, where acidification promotes proteolytic cleavage of the hemagglutinin, exposing a fusion domain. The influenza hemagglutinin undergoes a marked structural reorganization in this cleavage step. Hemagglutinin cleavage may be one of the factors that restrict viral growth to epithelial cells, as a unique protease in the respiratory milieu is required for this cleavage to occur. The fusion domain allows the viral RNA to enter the cytoplasm. The nucleoprotein is transported into the nucleus of the cell, where transcription to a positive-sense RNA and replication take place. Viral proteins then assemble on the apical surface of the infected cell and, after incorporation of cellular membrane, bud from the membrane back into the mucosal milieu.

Influenza infection is initiated in the upper respiratory tract via aerosolized virus. The cells infected with influenza virus are primarily the ciliated cells of the respiratory tract. Denudation of the superficial epithelium probably accounts for much of the symptomatology and may predispose to secondary bacterial infections. The onset of symptoms follows an incubation period that, for a viral illness, is very short: 48–72 h. The infection spreads to the lungs but, even there, remains confined to the epithelial layer.

Uniquely among respiratory viruses, influenza virus is associated with systemic symptoms of fever, malaise, and myalgia. These manifestations are presumed to be mediated by cytokines, and excess cytokine production has been implicated in the acute toxicity of H5N1 and other highly pathogenic influenza viruses.

The immune response to influenza virus occurs at the systemic and mucosal levels and involves both T and B cells. The B cell responses are directed primarily toward antigenic epitopes on the two surface glycoproteins—i.e., hemagglutinin and neuraminidase. At a structural level, the four recognized epitopes on the hemagglutinin are largely confined to the globular head of the protein, which collectively constitute the targets for hemagglutination inhibition (HAI) antibodies. HAI and neutralizing antibodies are highly correlated; HAI antibody levels are used as a measure of susceptibility to clinical infection and thus as a measure of vaccine-induced protection. In a child or an adult without prior vaccination or with the emergence of a distinctly new strain, serum HAI antibody is a surrogate for protection. However, in individuals with both vaccine-induced and natural immunity, the protective efficacy of a vaccine based on serum HAI antibody is more difficult to predict.

Studies with improved collection methods and assays that more sensitively and reproducibly measure mucosal antibody suggest that mucosal neutralizing IgA antibody more accurately reflects susceptibility to infection. Perhaps the patterns of immune protection are best shown in a murine model, in which passively administered IgA antibody to influenza virus protects animals from initiation of infection and epithelial damage in the upper respiratory tract, while infused IgG antibody to the virus is protective in the lungs.

There is now considerable research interest in the induction and protective role of broadly neutralizing antibodies that recognize less antigenically variable regions on the stalk of the hemagglutinin. The results of these studies have led to talk of a universal influenza vaccine, although no such vaccines are yet available in clinical practice.

The role of T-cell immunity, which primarily recognizes internal protein epitopes, remains unclear in humans. However, T-cell immunity is thought to play a role in clearance of an influenza infection that quite reproducibly develops 8–10 days after exposure. A role for T cells in protection against acquisition of infection has also been proposed.

Influenza is primarily a respiratory illness causing rhinorrhea, sore throat, conjunctivitis, and cough. The illness has a sudden onset and is epidemiologically linked to close contact with persons who have similar symptoms and often to community-wide respiratory illness. What distinguishes influenza from other respiratory illnesses is the degree of accompanying fever, fatigue, myalgia, and malaise. The symptoms typically begin within 48–72 h of exposure. The constellation of symptoms caused by an H3N2 viral strain, A/Port Chalmers 1/73, was followed prospectively in young seronegative children. Although these data involve children and a viral strain circulating 45 years ago, they present a representative picture of influenza today except that irritability in a young child is more specifically recognized as malaise, myalgia, and headache in an adult (Table 195-3).

Respiratory symptoms, particularly recurrent cough, persist well beyond the 2–5 days of systemic symptoms. There is a postinfectious delay in return to normal levels of activity. Pulmonary function is persistently decreased after acute influenza. Persons with a regular exercise routine (e.g., runners) note a decrease from their prior level of performance that typically lasts for a month or more. In the elderly, the respiratory presentation may be less prominent, but there is often a decline in baseline activity and a loss of appetite.

The variation among epidemics in symptom severity must be stressed. The circulation of an early H1N1 virus, A/USSR/77, in young children followed prospectively was detected only by comparison of pre- and post-season serum samples, which documented that 31% of 195 children had experienced infection with this virus without developing associated clinical illness. In the same population in the same year, influenza A/Texas/H3N2 virus infected 37% of children and was clearly associated with a typical influenza illness. In recent years, when H1N1 and H3N2 viruses have co-circulated, the latter has caused more severe epidemics.

On physical examination, the patient with influenza appears ill and rheumy, with sweating, coughing, nonpurulent conjunctivitis, and diffuse pharyngeal erythema. Pulmonary examination typically reveals nonlocalizing scattered rales, rhonchi, and wheezes. When present, localized pulmonary findings suggest relatively complicated pneumonia with a bacterial component. Muscle pain may be elicited by pressure, particularly in the calves and thighs. There are rare gastrointestinal findings. No rash is associated with influenza.

Complications of influenza occur most commonly in persons >65 years of age, those with underlying cardiopulmonary disease, those with immunosuppression, and women who are in the second or third trimester of pregnancy. In recent years, there has been mortality attributable to influenza among often previously healthy children <5 years of age in the United States, with ~100 deaths per year.

Respiratory Complications

Pneumonia characterized by progressive air hunger, localized pulmonary findings on physical examination, and radiographic findings of infiltrates and consolidation is the most common complication of influenza. Pneumonia in influenza can be primary influenza viral pneumonia, secondary bacterial pneumonia, or mixed viral and bacterial pneumonia. Primary viral pneumonia is characterized by increasing dyspnea, persistent fever, and—in more severe cases—cyanosis. Primary influenza pneumonia was typical in the 1918 pandemic and occurs with H5N1 virus, as initially described in Hong Kong in 1997. Pathologically, a marked inflammatory reaction in the alveolar septa is characterized by infiltration of monocytes, lymphocytes, and macrophages, with variable numbers of neutrophils. Destruction and hemorrhage are seen in the respiratory epithelium. Large amounts of virus can be recovered from the lungs.

In secondary bacterial pneumonia or mixed viral and bacterial pneumonia, illness may be biphasic, with evidence of recovery from the primary influenza illness followed by recrudescence of fever and pulmonary symptoms. Localizing findings may be detected on pulmonary examination and/or x-ray. The development of secondary bacterial infection is not surprising, as influenza de-epithelializes the airways and destroys ciliary function, allowing bacterial contamination. Another proposed mechanism for bacterial/viral enhancement is the production by Staphylococcus and Pseudomonas of proteases that enhance cleavage of the influenza hemagglutinin and thereby facilitate viral replication. The risk of secondary bacterial disease is greatest in elderly patients and those with chronic obstructive pulmonary disease.

Some influenza strains cause laryngotracheobronchitis or croup in children. Otitis media—a common accompaniment to influenza in children—may also be due to a combination of influenza virus and bacteria.

Extrapulmonary Complications

Although influenza is believed to spread only rarely beyond the respiratory epithelial cells, where unique endogenous proteases facilitate hemagglutinin cleavage and productive infection, this disease causes not only prominent systemic complaints but also a variety of extrapulmonary manifestations. The most common extrapulmonary manifestation of influenza is myositis, which is seen more often in influenza B and is characterized by severe muscle pain, elevated creatinine phosphokinase levels, and myoglobinuria that can lead to renal failure. The muscles are extremely tender to touch. Myo/pericarditis is seen less frequently and has been reported only in selected epidemics, notably the pandemic of 1918. However, a consistent epidemiologic link exists between influenza epidemics and excess cardiovascular hospitalizations.

Postinfectious acute demyelinating encephalomyelitis can follow influenza as well as other viral infections. The literature is mixed on the benefit and reliability of efforts to establish a polymerase chain reaction (PCR)–based diagnosis in this condition. MRI shows distinctive multifocal, symmetric brain lesions affecting the thalamus, brainstem tegmentum, cerebral periventricular white matter, and cerebellar medulla. Encephalitis and transverse myelitis may accompany influenza infection. Guillain-Barré syndrome can develop after influenza and was reported after a widespread influenza vaccination effort in the fall of 1976 that was undertaken in anticipation of a swine influenza epidemic (which never materialized). Until aspirin was recognized as a co-factor in its precipitation, Reye syndrome, an acute hepatic decompensation, was seen commonly in children and adolescents with influenza, particularly those infected with influenza B virus. Subsequently, the use of aspirin for fever control and symptom relief in children with viral infections was strongly discouraged, and Reye syndrome has virtually disappeared from clinical practice.

There is a strong argument for establishing a microbiologic diagnosis from both an individual-patient and a public-health perspective. This information is particularly valuable early in the season, when the extent of influenza and the precise circulating strain(s) are uncertain; in the management of complicated cases in hospitalized patients; and in settings such as long-term-care facilities and hospitals, where the institution of specific infection-control measures is appropriate.

Influenza virus is most easily recovered from nasopharygeal specimens. These samples are most effectively collected with a flocked swab that is inserted 1–2 inches into the nose (following the course of the inferior meatus), twirled, placed in viral transport medium that supports viral viability, and transported on ice to the laboratory as promptly as possible. The available rapid tests based on antigen detection vary in complexity and cost. Some are point-of-care tests, and others require laboratory support. These tests are highly specific but have a sensitivity of only 50–70%. Their sensitivity is strongly dependent on sample collection early in the course of illness—ideally within 48 h of the onset of symptoms. Traditionally, viruses have been isolated in tissue culture or with the use of embryonated eggs, and infection has been confirmed with hemadsorption or hemagglutination.

The most useful clinical approach today is to use a PCR-based molecular probe that amplifies specific segments of the influenza genome. Not only is this the most sensitive and specific method; it also provides opportunities to identify the strain with some specificity. Testing by multiplex PCR can simultaneously identify multiple respiratory pathogens—an advantage in the ill hospitalized patient. Serologic confirmation of infection is also possible but requires paired serum samples, with the convalescent-phase sample obtained 2 weeks after infection. Mucosal antibody assays that are now being developed can detect strain-specific antibodies in paired mucosal specimens and yield insights into the importance of mucosal immunity in protection against influenza.

Other laboratory tests are of limited value. Mild leukopenia is seen in influenza, and a white blood cell count above 15,000/μL suggests a secondary bacterial component in influenzal pneumonia.

The distinctive nature of influenza is such that clinical diagnoses by experienced pediatric physicians are 85% concordant with microbiologic etiologic confirmation. Respiratory syncytial virus often co-circulates with influenza virus; it particularly affects the youngest children, causing bronchiolitis, but it can also infect the elderly, leading to an influenza-like nonspecific respiratory illness and a decline in mobility, nutrition, and pulmonary function, with resultant hospitalization.

The major intervention to limit influenza illness is vaccination, which is conducted on a yearly basis because of variation in the predicted circulating strains for the coming year and which enhances existing immunity in advance of the influenza season. The decision about vaccine composition must be made ~10 months before the seasonal peak in influenza virus circulation; this decision is made by committees at the WHO and the CDC. This timing can result in a mismatch of vaccine composition with the viral strains that are actually prevalent in the upcoming season. The overall accuracy of the prediction is at least 70% for all strains in the recommended vaccine. Influenza vaccine is unique in being given seasonally in the months immediately preceding an outbreak. In the United States, vaccine is typically available from September.

The currently available vaccines are all based on purified subunit inactivated virus produced in eggs, in tissue culture, or through a baculovirus-expressed hemagglutinin protein. They are all calibrated to hemagglutinin content. Depending on the vaccine, they are administered intramuscularly or intradermally. Some influenza vaccines include MF-59 as an adjuvant or have elevated hemagglutinin content for enhanced immunogenicity in the elderly. The recommendations for use, the approved age range of each product, the route of administration, and the anticipated side effects are published annually by the CDC (https://www.cdc.gov/vaccines/hcp/acip-recs/vacc-specific/flu.html). The protection (and licensure) of all influenza vaccines depend on their stimulation of antibodies to the hemagglutinin; all are 50–75% effective at preventing clinical influenza. No effort is made to standardize the neuraminidase content. The contraindications to inactivated influenza vaccine administration are limited to individuals who have experienced a Guillain-Barré reaction within 6 weeks of a prior influenza vaccination. Egg allergy is not considered a contraindication to vaccination.

In recent years, a live, attenuated, intranasally administered vaccine (LAIV) has been used in children and has exhibited an efficacy exceeding that of injected inactivated vaccines. However, beginning in 2014–2015 and continuing in the following season, LAIV had no demonstrable efficacy assignable to the vaccine’s H1N1 component. This change led advisory committees in the United States not to recommend the use of LAIV despite its prior effectiveness, ease of administration, and theoretical advantage of stimulating mucosal immunity by the topical route. Active research is aimed at discerning the reason(s) for this lack of efficacy.

In the United States, the recommendation is that all individuals >6 months of age receive inactivated influenza vaccine yearly and that two doses of vaccine be given to children <9 years of age who are getting their first or second yearly vaccination. Groups at special risk of experiencing or transmitting influenza for whom influenza immunization is a particularly high priority are listed in Table 195-2.

Especially in hospital settings, considerable attention is paid to hand washing and the use of masks by persons with respiratory symptoms and those who are at particular risk of acquisition (typically, immunocompromised patients). Studies have demonstrated the benefit of face masks and hand hygiene in the hospital setting.

Additional Online References