Influenza Virus-Related Critical Illness

Prevention, Diagnosis, Treatment

Eric J. Chow; Joshua D. Doyle; Timothy M. Uyeki


Crit Care. 2019;23(214) 

In This Article


Persons with uncomplicated influenza typically experience acute onset of respiratory symptoms (cough, rhinorrhea, congestion), myalgias, and headache with or without fever. During influenza season, clinicians should also consider influenza when there is only fever present or in patients who are afebrile and have respiratory symptoms.[49] Complications of influenza vary by age, underlying comorbidities or high-risk conditions such as pregnancy, and immune function; elderly and immunocompromised persons may not always manifest fever. Critically ill patients may be admitted with respiratory or multi-organ failure, exacerbation of an underlying condition such as chronic lung disease,[50,51] heart failure,[52] or other extrapulmonary complications including stroke, encephalopathy, or encephalitis.[30,49,53]

Influenza testing is recommended for all patients requiring hospitalization with suspected influenza, including those admitted to the ICU during influenza season with acute respiratory illness and community-acquired pneumonia, without a clear alternative diagnosis. Furthermore, all individuals requiring critical care outside of influenza season should be tested for influenza if there is a possible epidemiological link to an individual with recent influenza, such as travel to areas with influenza activity or exposure to an institutional influenza outbreak. Special consideration should be given to elderly and immunocompromised patients, as influenza virus infection may not present with typical acute respiratory illness signs and symptoms (e.g., absence of fever). The Infectious Diseases Society of America (IDSA) 2018 Influenza Clinical Practice Guidelines also recommend influenza testing for patients at high risk of complications such as exacerbation of chronic cardiopulmonary disease.[49] Diagnosis of influenza should be made as soon as possible in critically ill patients, and initiation of antiviral treatment should not be delayed while awaiting results of diagnostic tests. Studies have reported an increase in mortality of ICU patients with influenza A(H1N1)pdm09 virus infection when diagnosis was delayed,[54] and shorter hospital length of stay when antiviral treatment was initiated within 6 h of admission.[55]

Several kinds of influenza diagnostic tests are available in clinical settings with variable sensitivities and specificities, including antigen detection assays, and molecular assays (nucleic acid detection) using respiratory tract specimens (Table 3). Within each of these testing categories, there is a wide range of available tests with varying diagnostic accuracy, and understanding the limitations of each diagnostic tool will allow the clinician to properly interpret their results. Most studies of influenza diagnostic accuracy have been conducted on specimens from patients with uncomplicated influenza, and few have assessed the performance of influenza tests in critically ill patients. The IDSA guidelines recommend molecular influenza assays for testing respiratory specimens from all hospitalized patients with suspected influenza because of their high sensitivity, specificity, and time to results (15 min to several hours).[49] The use of rapid influenza molecular diagnostic testing can result in better outcomes for patients and reduce the amount of resources required to care for patients in the emergency room.[57] Serology and viral culture are not recommended for clinical decision making, because timely results will not be available to inform clinical management. Serology requires collection of appropriately paired acute and convalescent sera performed at specialized public health reference laboratories, and results based upon a single serum specimen are not interpretable.[49] Although viral culture can confirm the presence of infectious virus with very high sensitivity and specificity, it must be performed at public health laboratories and requires 3–10 days to yield results.

A recent meta-analysis reported that influenza antigen detection tests that produce rapid results had very high specificities (> 98%), but sensitivities were highly variable compared with RT-PCR.[58] Rapid influenza diagnostic tests (RIDTs) without an analyzer device had only moderate sensitivity (53–54%), RIDTs that utilize an analyzer device (digital immunoassays) had moderately high sensitivity (77–80%), and rapid influenza molecular assays (nucleic acid detection) had high sensitivity (92–95%).[58] Low sensitivity of RIDTs for detecting influenza virus in ICU patients has been reported.[59] Recently, a systematic analysis of rapid influenza molecular tests from 29 studies reported pooled sensitivity and specificity of 87.9% and 97.4%, respectively.[60] Therefore, antigen detection assays, such as rapid influenza diagnostic tests and immunofluorescence assays, are not recommended for hospitalized patients with suspected influenza because of their lower sensitivities, unless molecular assays are not available.[49] Negative results for influenza based on tests with low sensitivity (e.g., RIDTs, immunofluorescence assays) should not be used to make clinical decisions. Instead, negative test results should be followed up with reverse transcription polymerase chain reaction (RT-PCR) or other influenza molecular assays to confirm results, and antiviral treatment should continue until results are available.

Preferred respiratory specimens for influenza testing in hospitalized patients without lower respiratory tract disease include nasopharyngeal, mid-turbinate nasal, or combined nasal-throat swabs. Collection of lower respiratory tract specimens should be considered in hospitalized patients with suspected influenza if upper respiratory tract specimens are negative and a positive test would result in a change of clinical management,[61] because viral replication in the lower respiratory tract may be ongoing and prolonged after virus is no longer detectable in the upper respiratory tract.[24,25] Influenza A(H1N1)pdm09 virus in particular has been shown to have affinity for infecting the lower respiratory tract.[24,31] In hospitalized patients receiving invasive mechanical ventilation in whom influenza is suspected, but not yet diagnosed, influenza testing should be performed on endotracheal aspirate specimens instead of those collected from the upper respiratory tract.[61] Molecular testing, including RT-PCR for influenza viruses can also be performed on bronchoalveolar lavage (BAL) fluid if collected for the testing of other pathogens. Blood, plasma, serum, cerebrospinal fluid, urine, and stool samples have very low diagnostic yield and are not recommended for influenza testing.[49] Diagnostic test results on specimens collected from non-respiratory sites should not be used for clinical decision making even for patients with extra-pulmonary complications of influenza.

Novel influenza A viruses are typically of animal origin, differ antigenically and genetically from currently circulating seasonal influenza A viruses (including H1N1pdm09 and H3N2 subtypes) and have infected at least one person. Novel influenza A viruses can cause a wide clinical spectrum of illness, ranging from asymptomatic infection, uncomplicated illness, to fulminant pneumonia, ARDS, and multi-organ failure[62] and human infection with a novel influenza A virus is of public health concern. In the U.S., human infection with a novel influenza A virus is nationally reportable to the Centers for Disease Control and Prevention; globally, under the International Health Regulations, countries are required to report such human cases to the World Health Organization. A major concern is the risk of novel influenza A virus transmission among humans; depending upon the prevalence of pre-existing immunity in the population, novel influenza A viruses may have pandemic potential. Patients suspected with novel influenza A virus infection should be investigated for a possible epidemiological link, i.e., a history of recent exposure to poultry or pigs or close contact to an individual with suspected or confirmed novel influenza A virus infection. Novel influenza A virus infection cannot be distinguished from seasonal influenza A virus infection by clinical findings or testing at clinical laboratories and therefore requires specific molecular testing of respiratory specimens by RT-PCR at public health laboratories.[63] Cases of suspected novel influenza A virus infections should be discussed with appropriate local and or national public health and laboratory staff to coordinate the testing of appropriate respiratory specimens.