Detection of influenza A(H3N2) virus in children with suspected mumps during winter 2014/15 in England

Influenza A(H3N2) virus was detected in oral fluid from 16/107 children (aged 2 to 12 years) with a clinical diagnosis of mumps, who were sampled between December 2014 and February 2015 in England, during the peak of the 2014/15 influenza season. Sequence analysis of an A(H3N2) virus from a child with suspected mumps showed the virus was similar to other circulating A(H3N2) viruses detected in winter 2014/15, which were antigenically drifted from the A(H3N2) vaccine strain.


Detection of influenza A(H3N2) virus in children with suspected mumps during winter 2014/15 in England
Influenza A(H3N2) virus was detected in oral fluid from 16/107 children (aged 2 to 12 years) with a clinical diagnosis of mumps, who were sampled between December 2014 and February 2015 in England, during the peak of the 2014/15 influenza season.Sequence analysis of an A(H3N2) virus from a child with suspected mumps showed the virus was similar to other circulating A(H3N2) viruses detected in winter 2014/15, which were antigenically drifted from the A(H3N2) vaccine strain.
During winter 2014/15, clinical parotitis in children with confirmed influenza A(H3) infection was reported in North America [1].In contrast, however, neither clinical nor virological surveillance for influenza-like illness (ILI) through sentinel general practitioners (GPs) in England detected an association between parotitis and influenza virus infection in the 2014/15 winter.In light of the observations from North America, gingival crevicular fluid (oral fluids) submitted for mumps surveillance in England were examined for influenza virus.Analysis of samples from 107 children (aged 2 to 12 years) with a clinical diagnosis of mumps but negative for mumps virus, sampled between December 2014 and February 2015, during the peak of the 2014/15 influenza season, showed that 16 (15%) were infected with influenza A(H3N2) virus.

Testing oral fluid samples from mumps surveillance for the presence of respiratory viruses
Of 116 residual oral fluids tested for mumps virus, a panel of 107 samples that were mumps virus negative was selected for testing for respiratory viruses based on age (2-12 years), sample taken between 1 and 5 days post onset of clinical parotitis between December 2014 and February 2015.Anonymised specimens were screened for respiratory viruses including influenza A(H1N1)pdm09 and A(H3N2), influenza B, respiratory syncytial virus A and B (RSV-A and RSV-B) and human metapneumovirus A and B (hMPV-A and hMPV-B) by real-time reverse transcription (RT) PCR using previously described assays [2][3][4] (Table 1).A total of 16 (15%) oral fluids tested were positive for influenza A(H3N2) virus in the mumps virus negative cohort (Table 2).One sample was found to contain a coinfection of influenza A(H3N2) virus and RSV-A.
As a control group, nine mumps virus positive oral fluids were screened using the same protocols and no respiratory viruses were detected.A further 63 control oral fluids were selected from the same age group (2-12 years) presenting with a different syndrome (suspected measles): only two (3%) of the measles virus positive samples were found to be positive for influenza A(H3N2) virus (Table 2).Two further separate coinfections of measles virus with RSV-A and hMPV-A were found.

Phylogenetic analysis of isolated influenza A(H3N2) virus haemagglutinin gene
Nucleotide sequencing of the haemagglutinin (HA) and neuraminidase (NA) genes of A(H3N2) viruses from influenza virus positive oral fluid samples by PCR (cycle threshold (Ct) values < 33 (4/16) was undertaken (primer sequences available on request).Full-length HA and NA sequence was obtained from one oral fluid, Sample 69 (A/England/Sample69/2015), and an HA phylogenetic tree constructed as described previously [5] (Figure ), using HA sequences from A(H3N2) viruses isolated in the 2014/15 influenza season in the United Kingdom, selected to be representative of the range of patient's age (2-12 years) and date of sample collection (between December 2014 and February 2015).Other reference virus sequences were obtained from the Global Initiative on Sharing All Influenza Data (GISAID) EpiFlu database (Table 3).
Analysis of the HA gene from A/England/Sample69/2015 showed that it fell into the A(H3N2) HA 3C.2a clade, as did the majority of A(H3N2) influenza viruses genetically characterised in the United Kingdom (UK) in 2014/15 (Figure) [6,7].Similarly, phylogenetic analysis of the NA gene from A/England/Sample69/2015 showed that it clustered with other 3C.2a clade viruses from 2014/15 (data not shown).Viruses from this clade are typically antigenically drifted from the 2014/15 northern hemisphere vaccine virus A/Texas/50/2102 and are more similar to A/Switzerland/9715293/2013, the A(H3N2) virus selected for the 2015/16 northern hemisphere influenza vaccine [8].

Discussion
Mumps is a notifiable disease in the UK, which is characterised by bilateral or unilateral inflammation of the parotid salivary glands.Mumps virus is the most common viral agent associated with parotitis, although other viral aetiologies have been investigated [9][10][11].Surveillance for mumps virus infection is carried out widely in primary care through oral fluid testing of suspect cases of parotitis [12].All oral fluid samples from throughout England are sent to the Public Health England (PHE) Virus Reference Department, Colindale, for testing.Recent infection is confirmed by detection of mumps virus-specific IgM antibodies and/or by detection of viral RNA by RT-PCR.Laboratory-confirmed positive cases are followed up to determine vaccination status.
An association between acute parotitis and influenza A virus has only rarely been made during previous epidemic periods in North America [13,14] and influenza is not often considered as part of differential diagnosis for this syndrome, and as a consequence oral fluid is not tested in England for the presence of influenza viruses [9,15].
In the presence of a well-performing measles-mumpsrubella vaccination and surveillance programme in England, mumps infection is rarely confirmed in children aged under 12 years, with less than 10% of samples submitted from this age group being positive.During the last quarter of 2014, only 2% of samples submitted to the PHE Virus Reference Department were positive.
Systematic surveillance for influenza-associated parotitis using optimised specimen types, sensitive RT-PCR technology and appropriately timed studies during periods of heightened influenza virus circulation has, to the best of our knowledge, not hitherto been performed.This study, using oral fluid, provides evidence that a proportion of children presenting with suspected mumps during winter 2014/15 had influenza A(H3N2) virus infections.Mumps virus negative oral fluids were tested as a proxy for sampling for parotitis due to nonmumps aetiology.Although the oral fluids were not specifically collected for respiratory virus detection, the study selection criteria included only those that were taken within five days of symptom onset, to optimise detection of respiratory viruses during the period of greatest virus shedding [16].Furthermore, analysis of oral fluid samples that were from patients without parotitis (i.e.measles virus positive) also found evidence of coinfections with respiratory viruses including influenza A(H3N2), RSV-A and hMPV-A.
Influenza A(H3N2) was the predominant virus circulating during the 2014/15 influenza season in the UK [6].Genetic characterisation of an A(H3N2) virus  from a child with parotitis found the viral HA and NA sequences were from the 3C.2a genetic subgroup, which is antigenically distinguishable from the vaccine strain, and similar to other A(H3N2) viruses detected through ILI surveillance [7].Previous studies have found very limited sporadic evidence of seasonal influenza A(H3N2) or zoonotic virus infection with swine influenza A(H3N2) in children with parotitis [14,17] or occasionally in adults [18].Former seasonal influenza A(H1N1) or A(H1N1)pdm09 viruses in patients with parotitis has not been described, to the best of our knowledge, although influenza A subtyping has not always been performed [9,15].
Information on influenza vaccination status in our study population was not collected so it is not possible to determine whether this could have affected the positivity rate for detection of influenza virus.The recent introduction of routine live-attenuated influenza vaccine (LAIV) for children aged 2-4 years in the UK may change the influenza positivity rate compared with that in an unvaccinated population [19] and further work to determine LAIV vaccine effectiveness against parotitis could be considered in future seasons.
Influenza virus as an atypical cause of acute viral parotitis should be considered especially during epidemic The haemagglutinin gene sequence from the child's sample (A/England/Sample69/2015, marked with ) was analysed with HA sequences from A(H3N2) viruses isolated in the 2014/15 influenza season in the United Kingdom, selected to be representative of the range of patient's age (2-12 years) and date of sample collection (between December 2014 and February 2015).These representative gene sequences were deposited in the EpiFlu sequence database of the Global Initiative on Sharing All Influenza Data (GISAID) (accession numbers provided in Table 3).Other reference virus sequences were obtained from the GISAID EpiFlu database (Table 3).The phylogenetic tree was constructed with a neighbour-joining algorithm available in the Mega 6 software [5,24].Branch lengths are drawn to scale.[20].Both mumps and influenza viruses bind sialic acid receptors on cells in the upper respiratory tract [21].The unusual clinical presentation of parotitis during the 2014/15 season in conjunction with a change in virus receptor binding properties warrants further investigation.
In England, oral fluid sampling is routine for suspected mumps.Young adults (between 15 and 30 years) are the expected demographic for true mumps infection [22,23].The target population for our study was younger children, where non-mumps parotitis is likely to be more common than true mumps infection.
Oral fluid sampling for detection of influenza virus should be further explored with appropriately timed studies in different age groups.Systematic collection of oral fluids alongside conventional nose/throat swabs during sentinel ILI surveillance would be a useful source of oral fluids for validation of detection of respiratory viruses.This would be a mechanism both for exploring less invasive sampling for influenza virus infection and also investigation of the incidence of influenza virus infection in uncommon clinical presentations such as parotitis.Oral fluids are not typically used as clinical samples for the detection of influenza and their utility for conducting influenza virological surveillance remains uncertain.The specific detection of influenza A(H3N2), and not influenza B or A(H1N1) pdm09 in this sample set, in combination with the clinical syndrome reporting from North America, suggests that influenza should be considered as part of the differential diagnosis for parotitis at the time when influenza virus is circulating, but this conclusion requires evaluation with different circulating influenza virus strains.The question of whether 3C.2a A(H3N2) influenza virus strains have an unusual tissue distribution compared with other A(H3N2) viruses or whether a subset of children infected with any influenza strain experience parotitis remains to be determined.Our conclusions would be strengthened by evaluation of further studies during future influenza seasons where oral fluids are taken in parallel with nasal swabs from children with influenza, with and without parotitis.Nevertheless, greater awareness of influenza virus as a potential cause of parotitis especially during epidemic periods associated with a drifted A(H3N2) strain is an important clinical and public health message.

Figure
Figure Phylogenetic analysis of influenza A(H3N2) virus haemagglutinin gene sequence from an oral fluid sample from a child with suspected mumps, England, December 2014-February 2015

Table 1
Oral fluid samples from mumps and measles surveillance selected for respiratory virus study, England, December 2014-February 2015 (n = 179) a Number of days symptomatic before sampling unknown for one sample.

Table 2
Detection of respiratory viruses in oral fluid samples from children with suspected mumps or measles, England, December 2014-February 2015 (n = 179)

Table 3
Origin of influenza A(H3N2) virus haemagglutinin sequences used in the phylogenetic analysis