The hidden burden of influenza: A review of the extra‐pulmonary complications of influenza infection

Severe influenza infection represents a leading cause of global morbidity and mortality. Although influenza is primarily considered a viral infection that results in pathology limited to the respiratory system, clinical reports suggest that influenza infection is frequently associated with a number of clinical syndromes that involve organ systems outside the respiratory tract. A comprehensive MEDLINE literature review of articles pertaining to extra‐pulmonary complications of influenza infection, using organ‐specific search terms, yielded 218 articles including case reports, epidemiologic investigations, and autopsy studies that were reviewed to determine the clinical involvement of other organs. The most frequently described clinical entities were viral myocarditis and viral encephalitis. Recognition of these extra‐pulmonary complications is critical to determining the true burden of influenza infection and initiating organ‐specific supportive care.


| Search strategy and selection criteria
The authors undertook a comprehensive literature review by querying MEDLINE (PubMed database) for articles in English pertaining to extra-pulmonary complications of influenza infection published between January 1, 1959, and July 1, 2016, using organ-specific search terms determined by author consensus (Table 1)

| Cardiovascular complications of influenza
Cardiovascular disease and influenza have long been associated due to an overlap in the peak incidence of each disease during winter months. Epidemiologic studies have also noted an increase in cardiovascular deaths during influenza epidemics indicating that cardiovascular complications of influenza infection, including exacerbation of heart failure, acute ischemic heart disease, and less often acute myocarditis, are important contributors to morbidity and mortality during influenza infection. [5][6][7] Additionally, studies from vaccine and antiviral therapeutic trials also highlight an important association and suggest that specific pharmacologic strategies may prevent or reduce the risk of many of the cardiovascular complications of influenza. 8-10

| Myocarditis
Clinically diagnosed myocarditis, based on a combination of symptoms, elevated cardiac enzymes, and echocardiographic findings, has been reported in approximately 0.4%-13% of hospitalized adult patients with documented influenza. 11,12 Myocarditis may, however, be a more common feature in fatal influenza infections as classic histopathologic findings, including cellular infiltration and myocyte necrosis, have been found in 30%-50% of patients at autopsy despite cardiac involvement not being clinically suspected. [13][14][15] Additionally, influenza-related myocarditis often occurs in the absence of more severe respiratory complications as only 40% (6/15) of patients with myocarditis in one study also had documented pneumonia. 16 There have been approximately 44 cases of influenza-associated myocarditis in adult patients described in case reports and case series (Table 2). 15, Among those reported, 52% (23/44)  The clinical course of influenza-associated myocarditis is variable, with the majority of patients experiencing an acute onset of symptoms related to cardiac dysfunction, including chest pain, dyspnea, syncope, hypotension, and arrhythmia, between days 4 and 7 following the initial symptoms of viral infection. 43 Most patients, with admission symptoms reported, presented to the hospital with cardiac symptoms (97%; 34/36), suggesting that it was the cardiovascular manifestations that were responsible for the patient seeking medical attention rather than typical respiratory symptoms. Of the 37 cases that documented time to onset of cardiac symptoms, only 3 (8%) developed symptoms late in their course as defined by greater than 10 days after initial viral symptoms. 15,20,36 The severity of influenza-associated myocarditis spans a wide spectrum ranging from asymptomatic to severe disease. Although recognition of influenza-associated cardiovascular complications occurs primarily in patients with cardiac symptoms, there is some suggestion that a significant proportion of patients with influenza infection may suffer clinically unrecognized, asymptomatic myocardial injury. A Japanese study of 96 patients found that 11% of patients who were infected with influenza A (H3N2) had elevated myosin light-chain I concentrations, a marker for myocardial injury. 44 However, the degree of injury is likely mild as another study of 152 slightly younger patients in the UK found no elevation of cardiac troponin I and T levels, which are more sensitive markers of cardiac injury. 45 A third study followed 30 previously healthy young adults diagnosed with influenza infection for 28 days after enrollment with serial electrocardiograms, echocardiograms, and cardiac enzymes and found that no enrolled patients had clinically significant changes in their cardiac studies or cardiac enzymes. 46 At the other end of the spectrum, a number of cardiac-specific complications have been described in the setting of influenzaassociated myocarditis including heart failure, arrhythmias, pericardial effusion, and cardiac tamponade. Congestive heart failure, as diagnosed by regional or global hypokinesis on echo/MRI, is the most common complication and is seen in 84% (37/44) of patients with influenza-associated myocarditis. More than half (23/37; 62%) of patients with heart failure from influenza-associated myocarditis required advanced cardiac support therapies. Thirty-eight percent (14/37) 29 Another patient had resolution of his influenza-associated myocarditis followed by recurrent myocarditis 22 days after initial presentation. 29,33 The underlying pathophysiology of influenza-associated myocarditis remains unclear. In a study of endomyocardial biopsies from 38 patients with viral and non-viral mediated myocarditis, the viral etiology was determined by PCR in 20 patients including 2 with influenza. 47 In another study of 29 patients with fatal influenza B infection with cardiac tissue samples available at autopsy, only one patient had influenza B detected by RT-PCR. 13 Of the case reports described above, the pericardial fluid of one patient was found to be positive for influenza A(H1N1)pdm09by PCR,. 24 Another patient had evidence of influenza A(H1N1) in her myocardium identified by both immunofluorescence and viral culture after her heart was replaced by a Jarvik artificial heart on hospital day 4; in this case, influenza A was also detected in her blood. 32 Evidence of virus in these cases suggests a potential role for direct viral invasion as the underlying pathogenic mechanism, although PCR positivity alone does not provide definitive evidence. Alternatively, the increased incidence of myocarditis among patients with more severe influenza infection may implicate an exaggerated immune response in the pathogenesis, as increased serum cytokines are commonly found in severe influenza. 48 and was higher in prevalence than generally reported in patients with acute respiratory distress syndrome (ARDS) (10-25%). 55

| Ischemic heart disease
Several large epidemiologic studies in Russia, the United States, the United Kingdom, and Hong Kong have revealed a temporal association between the circulation of influenza viruses and an increase in hospitalizations and deaths due to ischemic heart disease (IHD). 5,[56][57][58] Additionally, a significant increase in the number of IHD deaths has also been detected during epidemic periods. 57 Two large studies utilizing self-controlled case series analysis found that rates of a first MI were highest in the first 3 days following an acute respiratory infection with a reduction in the effect over time. 6,58 9 However, a separate retrospective study of 1378 patients who had survived a prior MI found no association between influenza vaccination and reduction in a subsequent MI. 64 Taken together, these studies strongly suggest that influenza vaccination is associated with a reduction in initial MIs, while larger studies on secondary prevention are needed.
Only one retrospective study has evaluated the effect of neuraminidase inhibitors on the recurrence rate of cardiovascular disease following influenza infection. 10 Patients with a prior history of cardiovascular disease and acute influenza infection were prescribed oseltamivir and those who filled with prescription within 2 days suf- Influenza-associated IHD is posited to be driven by inflammation which is known to have a critical role in the development of acute coronary syndrome. 65 In apolipoprotein E-deficient mice, an animal model of atherosclerosis, animals that were infected with influenza A virus developed subendothelial and smooth muscle inflammatory infiltration with overlying platelets and fibrin strands in atherosclerotic plaques. 66  influenza IgG seropositivity is not independently associated with IHD alone, patients who were seropositive for multiple pathogens were more likely to have IHD suggesting that cumulative infection burden may be related to development of atherosclerosis. 71

| Stroke
Similar to ischemic cardiac complications, the risk of ischemic cerebral vascular accidents (CVA), or strokes, appears to be significantly increased in the days after a respiratory tract infection. 58 However, only two cases of ischemic CVA in the setting of influenza infection have been reported. One case of multiple strokes occurred in a young woman with disseminated intravascular coagulation who was critically ill. 72 The other case occurred in a 50-year-old woman with ARDS secondary to influenza A(H1N1) who developed multiple strokes in the territory of the right middle cerebral artery. 73 There may be indirect evidence of an association between influenza and CVAs, but data from influenza vaccination and neuraminidase inhibitors are conflicting. 74 77 When stratified by type of stroke, the reduced risk was most significant for ischemic stroke with a trend toward protection in hemorrhagic stroke and no protection against transient ischemic attacks (TIA). In contrast, a large study of over 23 000 patients from two prospective cohorts (OPTIC and ARMISTAD) and a randomized trial (PERFORM) did not find any association between influenza vaccination and cerebrovascular accidents, although patients in this study had higher levels of known CVA risk factors including dyslipidemia and hypertension. 78 In this study, propensity score matching was used to compensate for potential healthy user bias that may explain the significant benefit seen in previous studies. Lastly, there is evidence that treatment with the neuraminidase inhibitor oseltamivir in the setting of acute influenza infection reduces the risk of stroke. 79 The protective effect remained significant in patients under the age of 65 for 6 months after infection, whereas in patients 65 and older, the association was significant only in the first month after infection. 79 Cardiac complications associated with influenza infection, including myocarditis, ischemic heart disease, and stroke, can have a significant impact on influenza mortality and morbidity. Early recognition of these complications is critical to initiate organ-specific supportive care Additionally, the protective role of influenza vaccines against an initial MI and early initiation of neuraminidase inhibitor with a decrease in combined cardiovascular end points highlights important preventative and therapeutic opportunities to reduce influenza-associated cardiovascular complications.

| Neurologic complications of influenza
Influenza infection can lead to a variety of neurologic complications including a number of specific clinical entities grouped together as influenza-associated encephalitis or encephalopathy (IAE), as well as a separate syndrome known as post-influenza encephalitis, GBS, Reye's syndrome, and Parkinsonian symptoms. While neurologic complications are more frequently noted in pediatric populations, they are also increasingly being recognized in adults. 80 The neurologic manifestations of influenza have been primarily reported in Japan, which may be due to greater recognition in that region. However, a study of influenza-associated neurologic complications in the USA did find that a disproportionate number of Asian/Pacific Islanders were affected (12.79/1 000 000) compared with white, non-Hispanic patients (3.09/1 000 000) suggesting a possible underlying genetic predisposition. 81

| Influenza-associated encephalitis/ encephalopathy
Influenza-associated encephalopathy or encephalitis (IAE) is a rapidly The spectrum of influenza-associated encephalitis can also be expanded to include posterior reversible encephalopathy syndrome (PRES), which is associated with areas of edema on MRI and can occur days to weeks after initial viral symptoms, and acute hemorrhagic leukoencephalopathy (AHLE), characterized by a rapid and fulminant demyelination and inflammation of the white matter. However, these entities are primarily described in the pediatric literature as IAE is more frequently recognized and reported in children less than 5 years of age. The incidence of IAE in adults has been reported in up to 4% of hospitalized adults, although case reports rarely apply the terminology of specific clinical syndromes beyond IAE. 83 There have been 28 reported cases of IAE in adults with 7 further differentiating the clinical syndromes including 3 cases of PRES, 2 cases of MERS, and 2 cases of AHLE (Table 4) The pathogenesis of influenza-associated encephalopathy and encephalitis in adults remains undefined. Demonstration of viral RNA, as detected by rRT-PCR, in brain tissue and CSF suggests direct viral invasion of the CNS. [106][107][108][109] Patients with IAE more frequently experience concurrent hepatic and renal function dysfunction, which could suggest a component of metabolic encephalopathy coexisting with severe influenza illness rather than as a direct effect of the virus itself. 83 A dysregulated immune response has also been posited to drive neurologic complications in influenza. Serum levels of cytokines IL-6, TNF-alpha, and IL-10 were found to be significantly elevated in pediatric patients with IAE as compared to influenza-infected patients without neurologic involvement. 110,111 Similarly, CSF levels of IL-6 were also found to be significantly elevated in pediatric patients with IAE compared to children with neurologic disorders and not infected with influenza. 112 Transcriptomic profiling has also demonstrated elevated levels of IL-6, IL-10, and TNF-alpha mRNA in patients with IAE compared to those with influenza-associated febrile seizures in the setting of equivalent viral loads. 111 One study demonstrated significantly greater serum levels of soluble CD163, a scavenger receptor for hemoglobin-haptoglobin complexes expressed by monocytes and macrophages in patients with more severe forms of IAE such as ANE compared to those with milder disease suggesting a dose-response association. 113 Although IAE has been described in case reports from around the world, there appears to be a higher incidence in East Asian populations, suggesting the possibility of a genetic predisposition. In a study of 29 Japanese patients with AESD or ANE, a higher frequency of several single nucleotide polymorphisms in the carnitine palmitoyl transferase II gene was found compared with healthy controls. 114 Another study identified an association between AESD and a genetic variant of the adenosine A2A receptor (ADORA2A) which may alter cyclic AMP signaling. 115   with an influenza-like illness. 128 In a study of 405 French patients with GBS, peaks in the incidence of cases without an identified etiologic agent are also temporally associated 1-2 months following peaks in the incidence of influenza-like illnesses. 120 In an additional analysis of

| Other neurologic complications of influenza infection
Narcolepsy is a sleep disorder characterized by excessive daytime sleepiness with nighttime disturbance and can be associated with cataplexy, the sudden loss of muscle tone triggered by strong emotions. It is caused by loss of neurons in the hypothalamus that secrete hypocretin, neuropeptides involved in regulation of arousal and wakefulness.
However, upper respiratory infections have also been loosely linked to disease presentation given a temporal association between infection and narcolepsy symptom onset, including Streptococcus pyogenes infection. 138  The neurologic complications of influenza infection are extremely variable in presentation and strength of association. In particular, IAE is poorly described in adults and requires MRI for diagnosis in 50% of cases despite being associated with death in 21% of reported cases.
Low threshold for diagnostic evaluation and early initiation of antiviral treatment and supportive care remains critical.

| Musculoskeletal complications of influenza
While myalgias are a common complaint among individuals with many viral infections, the development of rhabdomyolysis represents a less common but more serious complication. In cases of virusassociated rhabdomyolysis, influenza is identified as the most common etiology. 151,152 There have been approximately 27 reported cases of rhabdomyolysis in the setting of influenza infection, although this likely represents a small fraction of the total cases as myositis and/or rhabdomyolysis have infrequently been reported prior to the 2009 A(H1N1) pandemic (Table 5)

| Ocular manifestations of influenza infection
Influenza-associated ocular disease can result from direct conjunctival invasion by influenza virus and presents most commonly as a conjunctivitis although retinopathy, uveal effusion syndrome, and optic neuritis have also been reported. 173 Avian influenza viruses of the H7 subtype in particular, with the exception of zoonotic A(H7N9) display a prominent ocular tropism compared to other strains of influenza.
The human conjunctiva expresses the alpha 2-3 sialic acid residue which is preferentially recognized by avian influenza viruses but lacks the alpha-2,6 sialic acid residue that is classically recognized by human influenza viruses. 174  were found in murine ocular challenge models in which oseltamivir was shown to reduce A(H7N7) and A(H7N3) viral replication in both ocular and respiratory tissue. 180 Conjunctivitis has also been reported in association with the A(H1N1)pdm09 virus. 174 In a study of 89 patients with H1N1 infection, 58 (65%) presented with conjunctivitis, 7 patients (8%) presented with uveal effusion syndrome-a unilateral red painful eye associated with severe visual loss-and another 3 (3%) presented with optic neuritis. 173 Both the uveal effusion syndrome and the optic neuritis responded to treatment with corticosteroids. More severe ocular manifestations have been reported in association with influenza A(H1N1) pdm09 infection-two cases of acute retinitis and a case of bleeding follicular conjunctivitis. [181][182][183] All three patients recovered their vision although one did have persistently impaired color perception. In the patient with bleeding follicular conjunctivitis, viral RNA was detected by RT-PCR in the affected eye and in one of the retinitis cases, the patient had vitreous antibodies to A(H1N1).
The ophthalmologic complications of influenza are reported variably and may be strain-specific based on the evidence reviewed.
However, understanding of these complications is likely limited by under-recognition and under-reporting.

| Renal complications of influenza
Influenza infection can also affect renal function through a number of complications including acute kidney injury (AKI), acute glomerulonephritis, minimal change disease, and acute tubulointerstitial nephritis (ATN).

| Acute kidney injury (AKI)
Observational studies suggest the incidence of influenza-associated AKI ranges from 18% to 66% in patients cared for in an ICU setting. 184 four of those patients also had myoglobin pigment in the tubules consistent with rhabdomyolysis, and a fifth had evidence of thrombotic angiopathy. 191 Direct viral injury to the kidneys is another postulated mechanism of disease, but there is limited evidence to support this.
Influenza A(H1N1)pdm09 virus has been detected in the cytoplasm of glomerular macrophages in the kidneys of 4 of 5 patients in one post-mortem study; however, this may suggest circulating virus or simply the presence of viral genetic material in macrophages rather than direct invasion. 190 There is only one report of a patient with evidence of viral shedding in urine. 192 Renal complications in the setting of influenza infection, whether an exacerbation of an underlying condition or a novel consequence of the infection, remain poorly defined and warrant further investigation.

| Hepatic complications of influenza
Hepatic complications of influenza are rarely reported, yet recent reports suggest that liver injury occurs in the setting of influenza infection. [193][194][195] In an uncontrolled human challenge model of influenza infection, 4 of 15 human subjects experimentally inoculated with influenza developed elevated blood transaminase levels (greater than threefold the upper limit in two subjects). 196 197,198 However, liver damage may be a marker of disease severity as liver function tests including transaminases, bilirubin, and GGT were also associated with duration of hospitalization, hypoxia, and CRP. 193 Similarly, transaminase elevations have been more commonly reported in severe cases of influenza including 60% (6/10) of patients infected with highly pathogenic avian A(H5N1) and in 66% (73/111) of patients with A(H7N9) infection. 198,199 Post immediately following infection and demonstrated a delayed response to corticosteroids. 203 Liver injury due to influenza infection, possibly secondary to systemic inflammation mediated by viral infection, appears to be present in a percentage of cases suggesting that liver enzymes should be monitored closely.

| Thromboembolic disease
Post-mortem studies of patients who died from influenza infection have offered conflicting evidence about the association between influenza infection and thromboembolic disease likely due to differences in study design. In eight cases of fatal influenza A(H1N1), a higher incidence of pulmonary thrombi (75%) was found in comparison with the reported frequency in intensive care unit patients at autopsy (5%). 204

| Hemolytic-uremic syndrome (HUS) and thrombotic thrombocytopenic purpura (TTP)
Hemolytic-uremic syndrome (HUS) and thrombotic thrombocytopenic purpura (TTP), two overlapping thrombotic microangiopathic syndromes defined by non-immune-mediated hemolytic anemia, thrombocytopenia, AKI, and neurologic abnormalities, have been rarely associated with influenza infection in adults. Only four patients with influenzaassociated HUS or TTP have been described including one case that occurred in a renal transplant patient and one patient who had prior TTP and experienced a recurrence in the setting of influenza infection. [208][209][210][211] It has been postulated that the viral neuraminidase protein unmasks the Thomsen-Friedenreich cryptoantigen, which has previously been implicated in atypical HUS. 208,212 Alternatively, one patient with influenzaassociated TTP developed high titer antibodies against ADAMTS13 (a disintegrin and metalloproteinase with a thrombospondin type 1 motic, member 13), a protein that cleaves von Willebrand factor (vWF) multimers. 209 The association between influenza and TTP, however, is based on a paucity of case reports and thus remains speculative.

| Hemophagocytic syndrome
Hemophagocytic syndrome (HPS) is a clinical condition characterized by activated macrophages and histiocytes leading to secretion of an extraordinary amount of cytokines and an uncontrolled phagocytosis of platelets, erythrocytes, and lymphocytes. It is an uncommon condition thought to be triggered by specific autoimmune conditions, infections, or malignancy. However, the pathologic finding may be more common than thought as one post-mortem review of six patients who died of influenza A(H5N1) infection all had evidence of hemophagocytosis in the bone marrow on autopsy. 213

| Limitations
The quality of the literature presented is limited as this review relies primarily on case reports, case series, and observational stud- case-control studies, 3 basic science papers, 1 case series, and 1 case report. One of the case-control studies also addressed optic neuritis as a complication, and there were three additional case reports regarding acute retinitis and bleeding follicular conjunctivitis.
Ten articles were reviewed discussing the association of acute kidney injury in influenza infection, including 6 observational cohort studies, three autopsy studies, and one case report. Although the quality of the evidence is reasonably strong, the conclusions drawn are inconsistent.