Long-term Shedding of Influenza A Virus in Stool of Immunocompromised Child

In immunocompromised patients, influenza infection may progress to prolonged viral shedding from the respiratory tract despite antiviral therapy. We describe chronic influenza A virus infection in an immunocompromised child who had prolonged shedding of culturable influenza virus in stool.

In immunocompromised patients, infl uenza infection may progress to prolonged viral shedding from the respiratory tract despite antiviral therapy. We describe chronic infl uenza A virus infection in an immunocompromised child who had prolonged shedding of culturable infl uenza virus in stool.

H uman infl uenza A virus infections are typically lo-
calized to the respiratory tract, and viral presence in the gastrointestinal (GI) tract is rarely observed. Isolation of infl uenza virus from the stool is most commonly documented in human infections of avian infl uenza subtype H5N1. In these infections, virus in the stool may be related to a disseminated infection with exceptionally high viral titers, atypical of seasonal, circulating infl uenza infections (1). Interestingly, infections with the infl uenza A pandemic (H1N1) 2009 virus can be associated with a high rate of GI symptoms (2). Although seasonal infl uenza RNA in stool has been described, we report here the culturing of H1 virus from stool.

Case Report
The patient was a 4.5-year-old boy who had received a bone marrow transplant for Wiskott-Aldrich syndrome 4 years earlier; he also had chronic graft versus host disease of the GI tract. In January 2008, fever and respiratory symptoms developed in the patient, and direct fl uorescent antibody (DFA) testing of a nasopharyngeal swab specimen showed infl uenza A infection. Over the next several months, he had additional infl uenza A-positive DFAs and was given oral and aerosolized ribavirin for chronic infection. Despite this aggressive antiviral treatment, >24 respiratory specimens were positive for infl uenza A by DFA or culture over the next year.
When the patient was ≈3 months into an extended hospitalization at the Lucile Packard Children's Hospital (complicated by pseudomonal and enterococcal bacteremia, as well as disseminated aspergillosis), he experienced multiple daily episodes of nonbilious, nonbloody emesis (≈2-3/day) and loose stools without blood or mucous (up to 12/day). The clinical team suspected a graft versus host disease fl are but also conducted a work up of the patient to identify an infectious process. Results of blood cultures, stool culture, Clostridium diffi cile-toxin B cytotoxicity assay, and stool examination for ova and parasites were all negative. Results of a urine culture were positive for enterococcus. Culture of stool samples was enterovirus positive and over the subsequent 2 weeks, enterovirus was isolated from 2 additional viral stool cultures. These cultures were not evaluated for the presence of infl uenza virus. Three weeks after the initial stool testing, results of viral stool culture and repeat stool studies were negative. However, the diarrhea and emesis persisted, and ≈8 weeks after the onset of symptoms, another stool specimen was sent for viral culture.
After 3 days, the culture demonstrated cytopathic effect on primary RhMK cells but not on human foreskin fi broblasts, MRC-5 fi broblasts, or A549 lung carcinoma cells, which is consistent with enterovirus infection. However, results of immunofl uorescent staining of the RhMK cells (by using a panenterovirus blend of monoclonal antibodies) were negative. Staining results were also negative with serotype group-specifi c reagents, including the coxsackie virus B blend, echovirus blend, enterovirus 70 and 71, poliovirus blend, coxsackie virus A9, and coxsackie virus A24 (all enterovirus reagents from Millipore/Light Diagnostics, Billerica, MA, USA). An astute technologist associated the pattern of cells showing a cytopathic effect with the patient's concurrent infl uenza A-positive respiratory specimen and long history of infl uenza infection. She then set up a standard respiratory virus DFA panel (Millipore/Light Diagnostics), which included fl uorescein-conjugated antibodies for the detection of infl uenza A and B; respiratory syncytial virus; parainfl uenza 1, 2, and 3; and adenovirus. Strikingly, the specimen was strongly positive for infl uenza A virus and showed obvious hemadsorption with guinea pig erythrocytes. Two months later, infl uenza A was again isolated from the patient's stool, which suggested persistent infection of the GI tract with infl uenza A virus. Subsequent nucleic acid testing revealed that this chronic infl uenza A infection was caused by the seasonal, circulating subtype H1N1 virus. Overall, the patient shed infl uenza A from respiratory secretions for >1.5 years and from stool for >2 months.

Conclusions
Because viral stool cultures from patients with respiratory infections are infrequently ordered, the true occurrence of infl uenza virus in stool is unknown. The few studies to date have considered viral RNA in fecal specimens as a marker of GI infl uenza infection, which may not ac- curately refl ect the shedding of intact virus or the capacity for transmission. One study of 4 children with respiratory symptoms and confi rmed infl uenza infection showed that half had infl uenza RNA in stool (3). In contrast, 6 (<1%) of 627 patients with GI symptoms had detectable infl uenza RNA in fecal samples (4). Similarly, infl uenza RNA was detected in 21 (2.9%) stool samples from 733 children in Indonesia who had concurrent diarrhea and infl uenzalike illness (5). Notably, in this study, culturable infl uenza B virus was isolated from the stool of 1 patient. Future studies will be required to ascertain the incidence of infl uenza virus in the stool of children and adults with infl uenzalike illness and respiratory infl uenza infection. Furthermore, given the importance of this issue for infection control and the limited number of laboratories that perform stool viral culture, additional work will be necessary to correlate infl uenza RNA in feces with the presence of infectious virus.
The lack of a fully intact immune system likely predisposed our patient to chronic infl uenza infection and spread of the virus to the GI tract. This patient had received a bone marrow transplant for primary immunodefi ciency as well as immunosuppressive therapy for graft versus host disease with methylprednisolone, tacrolimus, sirolimus, and daclizumab. Prolonged viral shedding from the respiratory tract and the development of antiviral resistance is well documented in immunocompromised patients, including patients who have received bone marrow transplants (6). Although bone marrow transplant patients appear more susceptible to lower respiratory tract disease, in particular, during infl uenza outbreaks (7), infl uenza virus in stool samples from this patient population has not been well studied. Notably, our patient was not treated with either of the common classes of anti-infl uenza drugs, the neuraminidase inhibitors or adamantanes, but rather received a long-term course of ribavirin. Although clinical cases of ribavirin-resistant infl uenza virus infection have not yet been reported, genotypic and phenotypic analysis of this patient's isolate may show resistance or other virus-specifi c factors associated with chronic infl uenza and the presence of virus in stool. Although culturable infl uenza A virus was isolated from the stool of our patient, whether it played a causative role in the patient's gastroenteritis could not be determined.
While infl uenza virus likely spread to the patient's GI tract after a primary respiratory infection, the route of dissemination remains unknown. One possibility is direct GI inoculation by swallowing respiratory secretions. Because infl uenza viruses enter the cell through acid-activated fusion with the endosomal membrane (8), a low pH environment, for example in the human stomach, is thought to render most infl uenza viruses noninfectious by prematurely inducing an irreversible conformational change in the viral hemagglutinin (9). However, the sensitivity of infl uenza virus to low pH inactivation appears dependent on strain and subtype (10). Our patient was on a proton pump inhibitor, which would reduce gastric acidity. Another possibility is that the virus reached the GI tract hematogenously, as is suspected in human cases of avian infl uenza (1).
Whether the infl uenza subtype that infected this patient is capable of local GI replication in humans is also unclear. In the GI tract, the virus likely encountered the proteases necessary for hemagglutinin cleavage and activation (11). However, the H1 hemagglutinin has relative specifi city for α 2,6-linked sialic acid, a cell-surface glyco-conjugate not normally found on mucosa of the colon or small intestine (12,13). Nevertheless, this binding specifi city is not absolute (14), and 2,3-linked sialic acids are abundantly expressed on colorectal epithelial cells (13). Future studies should assess the ability of infl uenza viruses to replicate in the human intestinal epithelium.
Early epidemiologic study of the pandemic (H1N1) 2009 virus suggested that it produced diarrhea, vomiting, or both, in ≈25% of case-patients, more often than the previous seasonal, circulating infl uenza viruses (2). Consistent with the GI symptoms of human infection, experimental respiratory inoculation of ferrets with human isolates of the pandemic strain results in high infl uenza virus titers in the intestinal tract of infected animals (15). Because knowledge of transmission of this novel virus is limited, the Centers for Disease Control and Prevention recommends that all bodily fl uids, including the diarrheal stool of infected persons, be assumed to be infectious and handled with precautions. With the emergence of this pandemic (H1N1) 2009 strain known to produce GI symptoms, further research addressing the presence of infl uenza virus in stool could have major consequences for both infection control and disease management.