Fecal detection of influenza A virus in patients with concurrent respiratory and gastrointestinal symptoms

1. Background 

Infection caused by seasonal influenza virus is characterized by fever and respiratory tract symptoms.1 However, gastrointestinal (GI) symptoms such as diarrhea, vomiting, and abdominal pain are sometimes observed, especially among young children.2, 3 Yet fecal excretion of seasonal influenza virus has seldom been studied and only recently described in a few pediatric cases.4

2. Objectives 

This study aims at investigating the presence of influenza A virus in stool specimens collected from a large cohort of patients from all age groups presented with symptoms of gastroenteritis using polymerase chain reaction (PCR). Here we report virological and clinical findings on 6 cases who had seasonal influenza A virus detected in their stool specimens.

3. Study design 

Retrospective analysis was performed using archived stool specimens collected in our previous study on the etiology of gastroenteritis.5 This cohort was comprised of 651 stool specimens collected from 627 patients (age<1–97 years [median: 46; interquartile range: 12–67]; male 43.5%) presented with symptoms of gastroenteritis during a 12-month period (December 2004–November 2005). Specimens from both hospitalized patients and out-patients were included. All stool specimens were initially tested for common enteric bacterial pathogens (Salmonella, Campylobacter, and Vibrio species) at the time of collection, and then tested for enteric viral pathogens (human norovirus, group A rotavirus, and sapovirus) by virus-specific PCR as we described.5

In this study, we further evaluated the presence of influenza A virus in the 651 stool specimens blinded to clinical and virological information related to respiratory tract infection. Influenza A viral RNA was detected by conventional PCR using primers targeting the matrix (M) gene as described.6 Virus subtyping was performed by multiplexed H1- and H3-specific PCR.7 Fecal concentration of influenza A viral RNA was estimated by TaqMan®-based real-time PCR.6 Copy number of viral RNA was determined against 10-fold serial dilutions (from 2×108 down to 2 copies) of external plasmid standards prepared by cloning corresponding amplicon into vector pCR2.1-TOPO (Invitrogen, USA). The lower detection limit of the assay was equivalent to 4.9×103 PCR copies per gram of stool.

The period of local influenza season during the 12-month study period was defined as from February 2005 through July 2005,8 based on the monthly number of laboratory-confirmed influenza cases detected in our territory-wide sentinel surveillance system.9 The primary influenza virus subtype circulating during the study period was H3N2 which accounted for 77.5% of laboratory-confirmed cases, followed by influenza B virus.9 Clinical records of patients with detectable influenza A viral RNA in their stool specimens were reviewed.

4. Results 

Of the 651 stool specimens, 131 (20.1%) were tested positive for commonly known enteric agents. These included enteric viral (human norovirus [n=54, 8.3%],5 group A rotavirus [n=34, 5.2%], and sapovirus [n=8, 1.2%]) and bacterial (Salmonella [n=25, 3.8%], Campylobacter [n=14, 2.2%], and Vibrio species [n=8, 1.2%]) pathogens. Co-infection with two or more enteric pathogens was identified in 8 (1.2%) specimens.

Seven (1.1%) specimens collected from 6 patients were tested positive for influenza A virus by PCR. Stool consistency was either loose or watery. The median viral concentration was 3.2×104 PCR copies per gram of stool (range 4.9×103 to 8.0×107 copies). Virus subtyping (H1/H3) was successfully performed in 5 patients’ specimens. They all belonged to the H3 subtype, the primary subtype circulating during the study period. No co-infection with other enteric viral or bacterial pathogen was observed.

All 6 patients presented between March 2005 and July 2005, which fell within the local influenza season as defined above. PCR positive detection rates for influenza A viral RNA during the local influenza season (February 2005–July 2005) was 2.2% (7/312) among the total number of specimens collected and 2.0% (6/300) among the total number of patients recruited. Influenza A viral RNA detection in stool specimens was significantly associated with the occurrence of influenza season (p=0.0056; two-tailed, Fisher's exact test).

The demographics, virological data, and clinical features of the 6 patients are summarized in Table 1. All patients were at their extreme of ages (<5 or >60 years) or had underlying comorbidities (lymphoma, post-chemotherapy, and diabetes mellitus). They all presented with GI symptoms including diarrhea, nausea/vomiting and abdominal pain; and all had concurrent respiratory symptoms. Nasopharyngeal aspirates were obtained from 3 patients who had predominant symptoms of respiratory tract infection to investigate their viral etiologies and all were confirmed to be the H3N2 subtype as we described.6 Two adult patients with underlying lymphoma had the highest viral concentrations in their stool specimens (8.0×107 and 2.6×106 PCR copies per gram of stool). All 6 patients required hospitalization. One lymphoma patient died of complications of pneumonia and organ failure during the same hospital stay.

Table 1. Demographics, virological data, and clinical features of 6 patients with concurrent respiratory and gastrointestinal symptoms and tested positive for seasonal influenza A virus by reverse transcription-polymerase chain reaction in their stool specimens.

	Patient	Sex/age (year)	Month of Collection in 2005	Consistency of stool specimen	Viral concentration (PCR copies/g stool)	Underlying medical conditions	Presenting symptoms	Virus subtyping from stool	Virus isolation from NPAa	Outcome
	1	M/4	March	Watery	3.2×104 (4.9×103 in 2nd specimen)	Glucose-6-phosphate dehydrogenase deficiency, history of orchidopexy	Fever, vomiting, diarrhea, severe abdominal pain, and minimal URTI symptoms; laparoscopic appendicectomy performed and showed normal results	H3	H3N2	Survived
	2	F/4	April	Loose	1.7×104	None reported	Persistent fever for 5 days, nausea, diarrhea, abdominal pain, decreased oral intake, and URTI symptoms (cough and runny nose); CXR clear	N/A	H3N2	Survived
	3	M/42	June	Watery	2.6×106	Relapsed B-cell lymphoma post-chemotherapy; hypertension, fatty liver	Fever, mild diarrhea with watery stool, and URTI symptoms (dry cough and runny nose); CXR clear	H3	NDb	Survived
	4	M/64	May	Watery	2.8×104	History of open cholecystectomy and peptic ulcer	Severe abdominal pain, vomiting, diarrhea and minimal URTI symptoms; laparotomy performed and showed adhesions and intestinal obstruction; CXR clear	H3	NDb	Survived
	5	M/86	July	Loose	5.0×105	Diabetes mellitus on oral hypoglycaemic agent	Fever, diarrhea, cough, shortness of breath, and pneumonia; treated with oseltamivir after influenza A confirmed	H3	H3N2	Survived
	6	F/92	March	Loose	8.0×107	Diffuse large B-cell lymphoma, palliative chemotherapy given	Fever, mild diarrhea with loose stool, pneumonia, and congestive heart failure	H3	NDb	Died

Note: CXR, chest X-ray; N/A, not available; ND, not done; NPA, nasopharyngeal aspirate; URTI, upper respiratory tract infection.

a NPA was subjected to immunofluorescence assays for influenza A and B viruses, respiratory syncytial viruses, and parainfluenza viruses, followed by virus isolation. b Virus isolation from NPA was not performed due to lack of clinical suspicion for influenza.

5. Discussion 

In this study, we detected seasonal influenza A (H3N2) viral RNA in stool specimens from 6 patients with concurrent respiratory and GI symptoms. Other common enteric viral and bacterial causes were excluded. It is noteworthy that these patients were either at the extreme of ages or had underlying comorbidities. Furthermore, our 2 patients who had underlying lymphoma and chemotherapy exhibited the highest viral RNA concentrations in their stool specimens, in which one died. It is possible that influenza naivete or immunologic immaturity in young children or immunodeficiency in adults has predisposed them to the development of serious influenza infection, leading possibly to extrapulmonary, disseminated manifestations that may also include GI symptoms.3, 4, 10, 11

The biology and clinical significance behind fecal excretion of seasonal influenza A virus remain elusive. However, in avian models, influenza virus has been shown to infect GI tract and shed in large quantity into feces.1 Moreover, GI manifestation in human infection of highly pathogenic avian influenza A (H5N1) virus has also been recently documented12, 13 and the virus has been detected in human intestinal tissues14, 15 and feces.16, 17 In this regard, our findings raise question on whether seasonal influenza virus would likewise infect human GI tract and excrete in feces, at least in high-risk influenza patients. Such knowledge may have implications for the clinical management of severe seasonal influenza as well as on aspects related to infection control. Further investigation on the GI involvement of seasonal influenza is required.

This study is limited by its retrospective design and that virus culture was not performed. However, it should be noted that in human influenza patient care and study, only respiratory specimens are routinely collected for the purpose of diagnosis and virus characterization. Therefore, study on the fecal excretion of seasonal influenza viruses has been difficult due to the lack of appropriate specimen type. Here we used an alternative cohort type to show that seasonal influenza A virus can be detected from stool specimens of patients with concurrent respiratory and GI symptoms. Of note, this cohort was blinded to clinical and virological information related to respiratory tract infection during sampling, thus avoiding potential bias. While virus isolation in cell culture was not performed, our PCR results are unlikely to be false positives as they were detected by two PCR assays targeting the M gene (detection) and the hemagglutinin gene (subtyping) of the virus genome separately. Moreover, the same virus subtype (H3N2) was also detected in patients’ parallel respiratory specimens whenever available.

In conclusion, we report the fecal detection of seasonal influenza A (H3N2) virus in 6 high-risk influenza patients with concurrent respiratory and GI symptoms. Further investigation on the GI involvement of seasonal influenza is required.

Conflict of interest 

None.