Swine Influenza (H3N2) Infection in a Child and Possible Community Transmission, Canada

Seropositivity to the same strain was demonstrated in the child and in multiple other community members.

I nfl uenza A is endemic in a broad range of species, with avian and swine strains having the greatest potential for transmission to humans. Pandemics of infl uenza A occur when a major change occurs in the proteins of circulating strains of the virus. During the pandemics of the past century, this antigenic shift resulted from reassortment of human and avian strains or adaptation of avian viruses to facilitate person-to-person transmission (1). Avian infl uenza preferentially binds to sialic acid-galactose receptors with an α-2,3 linkage that is abundant on duck intestinal epithelium; human infl uenza preferentially binds to sialic acid-galactose receptors with an α-2,6 linkage that is abundant on human respiratory epithelium. The respiratory epithelium of swine contains both types of receptors and can potentially be simultaneously infected with avian and human infl uenza (2). Human infection with avian infl uenza subtype H5N1 is of great concern, with 194 deaths of 321 cases reported worldwide through August 16, 2007 (3). Swine infected with avian subtype H5N1 have been identifi ed in Vietnam (4), raising the possibility that swine could act as the "mixing vessel" that allows avian infl uenza (H5N1) to reassort with a human infl uenza strain, resulting in a virus with high pathogenicity and a high potential for person-toperson spread.
Another theoretical mechanism for the origin of an infl uenza pandemic would be the adaptation of a swine strain that results in effi cient person-to-person transmission, although cross-protection by antibodies to recently circulating human strains may prevent this from occurring with swine infl uenza virus (SIV) H1 and H3 strains. Infection of humans with SIV was fi rst recognized in 1974 with an H1N1 strain (5); the solitary outbreak occurred in military recruits at Fort Dix, New Jersey, USA, in 1976 (6). Human infection with SIV subtype H3N2 was fi rst described in Europe in 1993 (7). The fi rst reported case of probable infection of a person in North America with a non-H1N1 subtype of SIV occurred in Ontario, Canada, in 2005 with an H3N2 strain detected in the respiratory tract of an adult with no serologic evidence of infection (8). We describe a case of SIV (H3N2) infection in a Canadian infant, confi rmed by viral isolation and serologic testing.

Swine Infl uenza (H3N2) Infection in a Child and Possible Community Transmission, Canada
Case Report A 7-month-old boy was admitted to the hospital on September 10, 2006, with a 3-day history of fever, rhinitis, and cough. He had had no previous contact with ill persons. The child was born at term and was hospitalized for 21 days at 5 weeks of age when he received ventilation for 6 days for pneumonia due to respiratory syncytial virus. He lived on a communal farm (90 occupants) with horses, cows, swine, sheep, dogs, cats, turkeys, geese, ducks, and chickens but had no direct contact with the animals. The swine were contained in barns and did not mix with the other animals. His household contacts did not work directly with animals, but his father occasionally spent time in the barns, and his uncle, who lived next door, worked in the swine barns. On admission, the child was afebrile with a heart rate of 120 beats/min, respiratory rate 56/min, and oxygen saturation of 85% on room air. Diffuse wheeze was noted. Chest radiograph results were unremarkable. Direct fl uorescent antibody testing on a nasopharyngeal aspirate was positive for infl uenza A, and the virus was isolated in rhesus monkey cell culture. The isolate was sent to the National Microbiology Laboratory for infl uenza subtyping as a requirement of the Canadian infl uenza surveillance program, where it was subsequently designated A/Canada/1158/2006. The child stayed in the hospital for 2 days and then made an uneventful recovery at home. A cough and rhinitis developed in his 19-month-old brother on the day the index patient was admitted to the hospital, but the brother was not assessed by a physician.

Antigenic Analysis
For the antigenic characterization of A/Canada/ 1158/2006, hemagglutination-inhibition (HI) assay was performed by using 4 hemagglutination units of virus, 0.7% v/v guinea pig erythrocytes, and postinfection fowl serum specimens for the currently circulating human

Molecular Characterization
All 8 RNA segments of A/Canada/1158/2006 were amplifi ed by reverse transcriptase-PCR (RT-PCR) and sequenced. A universal primer set for the full-length amplifi cation of all infl uenza A viruses was used for the RT-PCR (10). Viral RNA was extracted from 100 μL of tissue culture fl uid with the RNeasy Mini Kit (QIAGEN, Mississauga, Ontario, Canada). Viral RNA was amplifi ed in a OneStep RT-PCR reaction (QIAGEN) following the manufacturer's recommendations. Briefl y, 5 μL RNA was added to the RT-PCR mixture containing 2 μL QIAGEN OneStep RT-PCR enzyme mix, 10 μL 5× QIAGEN On-eStep RT-PCR buffer, 400 μmol/L dNTP, 0.6 μmol/L of each primer, and 10 μL Q-solution in a fi nal volume of 50 μL. The conditions used for the Gene Amp 97700 (Applied Biosystems, Streetsville, Ontario, Canada) thermocycler were as follows: 50°C for 30 min for reverse transcription, 95°C for 15 min for the activation of the HotStart DNA polymerase; then 35 cycles of 94°C for 20 s, 58°C for 30 s, 72°C for 4 min, followed by an extension of 10 min at 72°C. The PCR products were purifi ed by using QIAquick PCR purifi cation kit (QIAGEN) and sequenced on an ABI 377 Sequencer, using a fl uorescent dye-terminator kit (Applied Biosystems). The DNA sequences were assembled and analyzed with SEQMAN, EDITSEQ, and MEGALIGN programs in Lasergene (DNASTAR, Madison, WI, USA). Phylogenetic trees were generated by the neighbor-joining method using the MEGA program (11).

Serologic Testing
Once it became evident that A/Canada/1158/2006 was closely related to swine infl uenza viruses, HI was performed on serum specimens collected from the index patient, the symptomatic sibling, and both parents 29 days after the hospitalization. To further investigate the spread of SIV to humans, approval was then granted by the Health Research Ethics Board of the University of Alberta to obtain information and serum specimens from other members of the communal farm. The study team visited the farm 3 months after the hospitalization of the index patient and explained the study to the occupants. Serum specimens were then collected from the other 4 siblings of the index patient and 46 other occupants who lived in a total of 17 households. Participants provided the following data: age, exposure to swine (none, <1 hour/week, or >1 hour/week), and history of infl uenza-like illnesses (ILI; defi ned as cough and fever) in the preceding year. Serum samples were tested by using an HI assay against the currently circulating human strains A/New Caledonia/20/99 (H1N1), A/ Wisconsin/67/2005 (H3N2), and the isolate from the index patient, A/Canada/1158/2006. HI titers were defi ned as the reciprocal of the highest dilution of serum that completely inhibited hemagglutination of a 0.7% solution of guinea pig erythrocytes. Specimens were considered seropositive for infl uenza virus at a titer of >32.

Swine Investigation
The purpose of these investigations was to determine the extent of recent swine infl uenza in swine on the farm and to look for evidence of infection with the SIV strain isolated from the index child. The history of infl uenza or unexpected respiratory illness in the swine on the farm was obtained. Nasal swabs were obtained from grower pigs (4 to 16 weeks of age) and processed by RT-PCR for infl uenza A matrix gene. Serologic testing for infl uenza, using an ELISA for H1N1 and H3N2 strains and HI for A/Canada/1158/06, was performed on samples from grower-fi nisher pigs (12 weeks to 6 months of age). Five grower pigs that were doing poorly were killed and pulmonary autopsies were performed. All swine used in these investigations were on the farm at the time the index child was ill.

Antigenic and Molecular Characterization of A/Canada/1158/06
Initial HI testing showed that the isolate was not inhibited by antiserum against recent (A/Wisconsin/77/2005 and A/New Caledonia/20/99) and past (A/Panama/2007/99 and A/Nanchang/933/95) human infl uenza A strains but was inhibited by antiserum against A/swine/Texas/4199-2/98 (H3N2) virus with HI titer of 128. These fi ndings indicate that the A/Canada/1158/06 virus was antigenically related to SIV ( Table 1). The results also indicate that the assay is specifi c because no cross-reactivity was observed between the human reference strain antiserum and the swine infl uenza viruses (Table 1) Nucleic acid identity between the HA and NA genes of A/ Canada/1158/06 and the current vaccine strain A/Wisconsin/67/05 was 90.9% and 94.6%, and the aa identities were 90.2% and 94.5%, respectively.

Serologic Testing
Seropositivity (HI titer >32) to A/Canada/1158/2006 was demonstrated in the index patient, the symptomatic sibling, 1 asymptomatic sibling, and both parents ( Table 2, household A). Three other siblings were seronegative. Four children from 2 other households were also seropositive ( Table 2, households B and C); the father from household B, 1 other child from household B, and the mother from household C were seronegative. The father from household C worked in the swine barn but was unavailable for testing. History of ILI within the preceding 12 months in seropositive participants was reported only for the index patient and for a 3-year-old girl from household C who was not hospitalized or tested for infl uenza virus during her illness. Seronegative results were obtained from another 20 adults (14 women and 6 men) and 19 children (8 girls and 11 boys) from 14 different households. For these households, swine exposure was reported as none for 9 adults and 7 children, <1 hour/week for 11 adults and 8 children, and >1 hour/ week for 4 children including 3 teenagers who worked in the swine barns. When serum samples from the 54 participants in the study were tested for HA-specifi c antibodies to the current human infl uenza A virus H3N2 and H1N1 subtypes, one of the patients who was seropositive for SIV at a titer of 32 had an identical titer for A/Wisconsin/67/2005 (H3N2) ( Table 2), and one of the adults who was seronegative for SIV had a titer of 32 for A/New Caledonia/20/99 (H1N1) (data not shown). All other persons tested were seronegative for the 2 human strains of infl uenza. Infl uenza (H3N2) was last documented in the swine herd in September 2005. The herd received breeding animals from a Manitoba herd, where swine infl uenza of an unknown subtype had recently been documented. Nasal swabs collected from 25 grower pigs ≈3 weeks after the index child was ill were negative for SIV. Serum specimens obtained from 10 grower-fi nisher pigs were all negative by ELISA for swine infl uenza (H1N1), but 4 were positive for swine infl uenza (H3N2) strains, with 1 of these 4 strains being seropositive for A/Canada/1158/2006 by HI assay (HI titer 32). Results of the lung autopsies all showed evidence of subacute bronchointerstitial pneumonia, varying from mild to moderate. Lesions typical for swine infl uenza were not noted, but an initial insult due to SIV could not be excluded.

Discussion
We describe an infant with virologic and serologic evidence of infection with SIV (H3N2) and an ILI. Serologic evidence of infection with the same strain was found in 4 of 7 household members and in 3 of 46 nonhousehold contacts, with only 1 of the seropositive patients having a history of an ILI within the preceding year, which demonstrated unrecognized human infection with SIV. This relatively high seroprevalence is in contrast to a recent outbreak of avian infl uenza (H7N3) in which seropositivity was not documented in 91 persons exposed to infected poultry, including 2 poultry workers from whom the virus was isolated (12). The difference in the apparent incidence of infection may be explained in part by the fact that culling of infected poultry occurred immediately; in our study, infection of swine was not recognized and long-term human exposure may have occurred.
Infection of swine with human infl uenza viruses has been recognized for decades (2); in a recent US study, 22.8% of pigs were seropositive for human infl uenza viruses, although some may have had vaccine-induced im-munity (13). Swine infl uenza (H3N2) emerged in 1998 in the United States, where subtype H1N1 viruses had predominated for 60 years (2). The isolate from this current study is closely related to triple reassorting genotype viruses that spread rapidly throughout the US swine population and have HA, NA, and RNA polymerase (PB1) genes of human infl uenza virus lineage; nucleoprotein, matrix, and nonstructural genes of classic swine infl uenza (H1N1) lineage; and RNA polymerase (PA and PB2) genes of North American avian virus lineage (8). However, triple reassortant SIV was not documented in swine in Canada until 2005 (8), which makes it unlikely that human cases occurred before that year and that seroreversion had occurred in any of the persons in the current serosurvey.
A previous study showed cross-reactivity in HI assay between the vaccine strain A/Panama/2007/99 reference antiserum and the triple reassortant A/swine/Minnesota/593/99, which is not unexpected since the HA gene of the triple reassortant viruses is a descendant of human viruses that circulated in 1995 (14,15). However, no cross-reactivity was observed between the reference human strain antiserum and the isolate from this study, which suggests that the seroconversion observed was indeed due to infection with swine infl uenza (H3N2) and not to cross-reactive antibody to human infl uenza (H3N2) infection. The low rate of seropositivity to recently circulating strains of human infl uenza in the study is likely explained by the fact that the farm is a relatively closed community. The child who was seropositive for both human and swine infl uenza viruses was likely exposed to both viruses.  strains currently circulating in North America. This region of the protein has been assigned to antigenic sites (17) and has been associated with adaptation to growth in eggs (18 (19). The spectrum of pathogenicity of SIV infection ranges from asymptomatic infection (6) to death; 7 of these 50 patients died (5,(20)(21)(22)(23)(24). Laboratory-confi rmed swine infl uenza in humans may be "the tip of the iceberg." Diagnosis of the current case was serendipitous because typing was performed only because the case occurred outside of infl uenza season.
The mode of spread of SIV in humans is not established. Because of his young age, the index patient was not likely to have had unrecognized direct contact with swine. That aerosolization of infl uenza virus occurs is increasingly recognized (25), but the child was reportedly never in the barns that housed the swine. However, other members of the farm reported that infants were sometimes taken for walks through the barn. The child also may have acquired the virus from person-to-person spread or from fomites. All 13 patients in the Fort Dix outbreak and 15 of 37 previously reported civilian case-patients also had no swine contact (19,20).
The Fort Dix outbreak of SIV in humans lasted only 21 days and never spread outside the military base. The calculated basic reproductive rate (R 0 ) was only 1.1 to 1.2. This suggests that person-to-person spread of the implicated H1N1 strain was not effi cient enough to produce a major epidemic (26). However, future strains of SIV could have a higher R 0 , and documentation of a case of swine infl uenza (H3N2) in a child with unrecognized transmission within the community adds another possible mechanism by which major epidemics of infl uenza could arise. Swine infl uenza infection in humans most commonly results in either no symptoms or a self-limited illness (6). However, routine surveillance for cases among swine workers may enable early detection of a strain with the potential for personto-person transmission, prompting institution of infection control measures and vaccine development.