Journal of Infection (2014) xx 1e8

wwwelsevierhealthcomjournalsjinf

Influenza C virus high seroprevalence ratesobserved in 3 different population groups

Nicolas Salez a Julien Melade a Herve Pascalis bSarah Aherfi d Koussay Dellagi b Remi N Charrel adFabrice Carrat c Xavier de Lamballerie ad

aUMR_D 190 Emergence des Pathologies Virales Aix-Marseille Univ IRD French Institute of Researchfor Development EHESP French School of Public Health Marseille FrancebCentre de Recherche et de Veille sur les maladies emergentes dans lrsquoOcean Indien Sainte-ClotildeLa Reunion Francec Epidemiologie des Maladies Infectieuses et Modelisation (UMR-S 707) INSERM-Universite Pierre etMarie Curie Paris Franced IHU Mediterranee Infection APHM Public Hospitals of Marseille Marseille France

Accepted 13 March 2014Available online - - -

KEYWORDSInfluenza CEpidemiologyHaemagglutinationELISAVirus NeutralisationTestReal-time RT-PCR assay

Corresponding authorE-mail addresses salezumr190y

0163-4453$36 ordf 2014 The British Infehttpdxdoiorg101016jjinf2014

Please cite this article in press as SaInfect (2014) httpdxdoiorg101

Summary The epidemiology of Influenza C virus (FLUCV) infections remains poorly charac-terised Here we have examined the age- and location-specific seroprevalence of antibodiesagainst FLUCV in 1441 sera from metropolitan continental France (Marseille) South-West In-dian Ocean French territories (Reunion Island) and United-Kingdom (Edinburgh) using a combi-nation of haemagglutination inhibition virus neutralisation and ELISA assays Our results showthat immunity to FLUCV is common in all locations studied (global seroprevalence values gt50)and that the first immunising contacts generally occur early in life (ie in the 0e4 year-oldage group) The latter item is further supported by the detection of FLUCV RNA by RT-PCRin naso-pharyngeal samples collected in patient attending the Emergency Room of the Publichospitals of Marseille France with a large majority of children under 10 years-old 17 (607) inchildren 3 yo 10 (357) in the 4e10 yo age group and 1 (36) in an adult (49yo) The tem-poral distribution of cases was atypical with regard to influenza (a large proportion of casesoccurred in spring and summer) and the clinical presentation was diverse including but beingnot limited to classical Influenza-like-Ilnesses Altogether our results indicate an intense cir-culation of FLUCV in the different study areas and an early occurrence of infection in humanlife Flu C appears to be a widely under-diagnosed and under-studied human paediatric diseasethat obviously deserves further clinical and epidemiological characterisationordf 2014 The British Infection Association Published by Elsevier Ltd All rights reserved

ahoofr saleznicolasyahoofr (N Salez)

ction Association Published by Elsevier Ltd All rights reserved03016

lez N et al Influenza C virus high seroprevalence rates observed in 3 different population groups J016jjinf201403016

2 N Salez et al

Introduction

Influenza C virus (FLUCV family Orthomyxoviridae genusInfluenzavirus C ) was first isolated in 1947 from respiratorysamples collected from 4 individuals 17e21 year-old in theUSA1 Diagnosis and epidemiological studies rely on alimited number of available techniques (i) the virus canbe isolated from biological samples by inoculation of theamniotic cavity of 10-day-old embryonated henrsquos eggs2 orby propagation onto MDCK (canine kidney) cells (ii) theviral genome can be detected by molecular biologyassays3e7 (iii) specific antibodies to the virion proteinscan be detected using a variety of techniques such as ELISAhaemagglutination inhibition (HI) or seroneutralisation as-says The clinical and epidemiological aspects of FLUCVinfection in humans remain globally poorly characterisedand mainly rely on a few studies in paediatric populationsIt is generally admitted that FLUCV infection can be associ-ated with mild influenza-like illness8 The first serologicalstudies were conducted in the early 1950s9e11 and evi-denced the widespread distribution in human populationsSince there have been a few publications describing thesero-epidemiology in the population of Europe12e15

Americas16e19 and Asia2021 Similar global seroprevalenceresults (50e66) obtained in these different studies havesuggested that FLUCV infection is widely distributedthroughout the world and that the majority of humans ac-quire antibodies to the virus early in life22

In the current study we investigated the presence ofantibodies to FLUCV in human sera collected from mainlandEurope (in South-eastern France (Marseille) and Scotland(Edinburgh)) and in Reunion Island (French overseas terri-tory in the South West Indian Ocean) using a combination ofELISA HI and seroneutralisation tests In addition nasalssamples collected in Marseille France were tested for thepresence of FLUCV using a specific molecular technique

Materials and methods

Serum samples

First we examined a series of sera from Reunion Island(French overseas territory in the south-west Indian ocean)initially collected in order to assess at the community levelthe magnitude of the pH1N12009 influenza A pandemicand the extent of the herd immunity acquired after passageof the epidemic wave Sera were collected during aprospective population serosurvey conducted in ReunionIsland during the 2009 austral winter season (CoPanFlu-RUNcohort)23 The protocol was conducted in accordance withthe Declaration of Helsinki and French law for biomedicalresearch (N ID RCB AFSSAPS 2009-A00689-48) and wasapproved by the local Ethics Committee (Comite de Protec-tion des Personnes of Bordeaux 2 University) Every eligibleperson for participation provided written informedconsent

A total of 2164 individuals from 772 households wereenrolled between weeks 30 and 44 (July 21st to October31st 2009) allowing the collection of 1932 sera at inclusionThe selected households represented a wide range ofgeographic locations in order to minimize the repartition

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

bias Among these inclusion sera we randomly selected 973sera for the influenza C study The distribution in agegroups of our sample was 0e3-year old (yo) 1234e10 yo 1470 11e30 yo 2641 31e50 yo 2693gt50 yo 3073

Second we examined a collection of sera collected in2007e2008 in Marseille (South-east metropolitan France)This archival material was originally randomly selectedfrom the serum library of the Public Hospitals of Marseillesin order to investigate the sero-epidemiology of influenzainfections in the pre-pandemic period24 A series of 331sera were randomly selected amongst the 1693 samples ofthe original anonymous panel The distribution in agegroups was 0e3 year-old (yo) 1118 4e10 yo 123911e30 yo 2387 31e50 yo 2810 gt50 yo 2447

In addition 95 Scottish left-over archival sera werekindly provided by Pr P Simmonds and Dr C Sharp (theUniversity hospital of Edinburgh UK) The distribution inage groups of our sample was 31e50 yo 1368 gt50 yo8632

Nasal samples

673 nasal swabs were randomly selected in the collection ofthe Microbiology and Infectious Diseases department of thePublic Hospitals of Marseille amongst samples collectedbetween April 2011 and February 2012 from in- and out-patients with acute respiratory symptom and a negative RT-PCR result for Influenzavirus A amp B Human Respiratory Syn-cytial Virus A amp B and Human Metapneumovirus A amp B(referred to as Flu AampB hRSV and hMPV respectively inthe current article) The distribution in age groups was0e3 yo 6790 4e10 yo 1486 11e30 yo 55031e50 yo 342 gt50 yo 832

Laboratory procedures

D virus propagation and purification

The CLeningrad23283 strain of Influenza C virus waskindly provided by Dr MY Eropkin (Research Institute ofInfluenza StPetersburg Russia) and propagated ontoMDCK cells at 32 C in a 5 CO2 atmosphere in MinimumEssential Medium (MEM) supplemented by 2 mM L-gluta-mine 100UIml Penicillin-100 mgml Streptomycin 3 mgmL trypsin and 10 heat-treated foetal calf serum (all re-agents from GIBCO)

Cell culture supernatant was collected at day 5 pi andclarified by centrifugation (2000 g 5 min) dilutions ofthis viral antigen were used directly for HI and virusneutralisation techniques

For producing the ELISA viral antigen 4 volumes ofclarified supernatant were mixed with 1 volume of a 25 NNacl-50 PEG6000 solution (SigmaeAldrich St Louis MOUSA) incubated overnight at 4 C and centrifuged at5000 g for 3 h in an angular rotor (Thermo ScientificFiberlite F14-6x250LE) The pellet was re-suspended in aminimal volume of PBS to obtain a w100-fold viral concen-tration sonicated (360 s 30 W 46 Hz) loaded on a sucrosegradient (15e60 wv) and centrifuged at 30000 g for3 h in a swinging bucket rotor (Beckman Coulter SW 32 Ti)

seroprevalence rates observed in 3 different population groups J

Influenza C virus high seroprevalence rates 3

The purified viral suspension was aliquoted for conservationat 80 C

DSerum treatment

Sera were heat-inactivated at 56 C for 30 min

DHaemagglutination Inhibition assay (HI)

No receptor destroying enzyme (RDE) treatment wasperformed according to previous demonstration that thistreatment does not affect HI results in the case of FLUCV19

Viral antigen was prepared by diluting in PBS a non-inactivated FLUCV cell culture supernatant (see above)The haemagglutinating titre of viral antigen was deter-mined extemporaneously All tests were performed usingU-shaped microplates Twofold dilutions of sera in PBS (110 to 11280) and preparation of the HI mix were per-formed with an Eppendorf epMotion-5070-CB devicelocated in a biological safety cabinet 25 mL of viral antigendilution (corresponding to 533 haemagglutinating units)was added to 25 mL of each serum dilution and incubated50 min at 4 C Following addition of 25 mL of a 1 humanOthorn red blood cells suspension in PBS and another incubationof 30 min at 4 C the HI titre was determined as the highestdilution providing clear inhibition of haemagglutination Allexperiments included the same negative and positive con-trols and a serum agglutinating activity control

DELISA assay

Maxisorp plates (Nunc USA) were coated overnight at 4 Cwith a 1400 dilution in PBS of the sucrose gradient-purifiedFLUCV prepared as described above Plates were washedwith PBS Tween 005 (3 times between each subsequentstep) They were saturated with 1 Bovine Serum Albuminduring 2 h Serum samples diluted to 1400 in PBS thorn1milk were incubated 1 h at 37 C Peroxydase-conjugatedAffini Pure F(abrsquo)2 Fragment Goat Anti-Human IgG(H thorn L)(dilution 17500 Beckman Coulter) was added and incu-bated at 37 C for 1 h Tetramethylbenzidine was used assubstrate of peroxydase (Eurobio Life sciences) and aftera 4-min contact the reaction was stopped by addition of1 N citric acid Absorbance was read at 450 nm (SunriseTecan)

Each plate included a series of 6 blank wells (no serumadded) to determine the reaction background noise whichwas subtracted from ELISA optical density (OD) results

A series of 345 negative sera (ie providing negative re-sults in both HI and VNT techniques) was used to calculatethe cut-off value The OD limit of non-negativity for eachmicroplate (L) was calculated using the formulaL Z m thorn 2 SD where m and SD represent the mean ODand OD standard deviation values respectively calculatedfrom negative sera

DVirus Neutralisation Test (VNT)

Twofold serial dilutions of sera were mixed with a PBSdilution of culture supernatant corresponding to 50 TCID50of FLUCV (total volume 100 mL) and incubated for 1 h at

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

37 C Confluent MDCK cells (grown onto 96-well plates)were then infected with each dilution After a 2-h incubation the inoculum was removed and fresh EMEM[containing 3 foetal calf serum glutamine (2 mM) anti-biotics (penicillin 100UIml streptomycin 100 mgml) andtrypsin (10 mgmL)] was added In the absence of adetectable CPE produced by FLUCV virus growth wasdetected by qRT-PCR at day 5 pi (see below) The VNTtitre corresponded to the last serum dilution in which thevirus was not detected

DReal time RT-PCR

Viral RNA extraction was performed using the BioRobot EZ1Workstation and the EZ1 Virus Mini Kit V20 (Qiagen) Nu-cleic acids were extracted from 400 mL of sample andeluted in a 90 mL final volume

One-step RT-PCR was carried out in a 30 mL final volumeaccording to the manufacturerrsquos instructions using theldquoiScript One-Step RT-PCR for probesrdquo kit (Bio-Rad)8 pmol of each primer 04 pmol of probe and 10 mL ofRNA extraction The thermocycling profile included 50 Cfor 15 min 5 min at 95 C 95 C for 15 s (39 cycles) 60 Cfor 45 s The amplification was carried out using CFX96Real-Time PCR Detection System (Bio-Rad LaboratoriesInc)

Primers and probe sequences (matrix gene) were4

FluC_M-F CAT AAT TGA ACT TGT CAA TGG TTT TGTFluC_M-R TTC AGG CAT AAT TGT GGT CTT TAT ATC T probeFluC_M-FAM FAM-CTC GGC AGA TGG GAG AGA TGG TGT G-TAMRA

Results

Serological results

Performances of the different assaysAfter testing all 973 samples from Reunion Island using theHI and ELISA methods sera were randomly selected foreach HI titre (titre 5 4 sera 10 11 sera 20 9 sera 40 11sera 80 11 sera 160 11 sera 320 9 sera total 66 sera)and further tested using the VNT technique

The relationship between HI and VNT results is shown inFig 1a Using VNT as a reference test (samples with aVNT titre 10 were considered positive) an HI cut-offvalue at 80 (ie samples with a titre 80 were consid-ered positive) was associated with high Specificity (86)and Positive Predictive (90) values and a lower Sensi-tivity (80)

The relationship between ELISA and VNT results isshown in Fig 1b The cut-off value calculated asdetailed in the Methods section was 0797 Using VNTas a reference test ELISA was associated with highSpecificity (86) and Positive Predictive (91) valuesand a modest Sensitivity (66)

When HI and ELISA results from the 973 Reunion islandsera were compared a strong linear correlation be-tween the HI titre and the corresponding log of theELISA OD median value was observed (see Fig 2)

seroprevalence rates observed in 3 different population groups J

Figure 1 a Relationship between Haemagglutination Inhibi-tion titre and results of the Virus Neutralization Test in 66 serarandomly selected from the Reunion Island CoPanFlu cohortSamples with a VNT titre 10 were considered positive bRelationship between Haemagglutination Inhibition titre andresults of the Virus Neutralization Test in 66 sera randomlyselected from the Reunion Island CoPanFlu cohort Confirma-tion of ELISA results by VNT assay according to cut off

Figure 2 Relationship between ELISA median DO values andHI titres in 973 Reunion Island sera Of note the 0797 OD value(used as cut-off value for the ELISA) is associated with an HItitre close to 180

4 N Salez et al

Importantly the abovementioned 0797 ELISA OD cut-off value was associated with an HI titre at 80

Overall this evaluation of the different serologicalassays indicated that results provided by our ELISA werevery similar to those provided by HI using a cut-off at 80Using VNT as a reference both tests were associated withsimilar and high Specificity and Positive Predictive valuesand globally underestimated the number of true positivesThis implicates that ELISA-based prevalence numbers ob-tained in the current study constitute robust but lowestimates of the actual FLUC seroprevalence

Seroprevalence results

In Reunion Island HI and ELISA results from 973 samplesprovided a very similar distribution of positives in agegroups (Fig 3A) The seroprevalence was around 30in children under the age of 4yo and around 50 in olderage groups A similar trend was observed for the HI geo-metric mean titres (GMT)

In metropolitan France (Marseille) 331 sera weretested by ELISA The distribution in age classes(Fig 3B) also showed a rapid progression of seropreva-lence between children under the age of 4yo (w30)

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

and those in the 4e10yo age group (w60) The sero-prevalence reached w75 in individuals older than50 yo Because these results were similar to those ob-tained in Reunion Island the panel of sera studiedwas not further extended

In the UK (Edinburgh) 95 sera from patients older than30yo were tested by ELISA In the 31e50 yo and gt50 yoage classes results were similar with those observed inMarseille (Fig 3B)

Molecular biology results

A total of 28 influenza C cases were identified from 673nasal swab samples tested using a real-time RT-PCR assaySeventeen cases were diagnosed in children under the ageof 4 yo and ten in the 4e10 yo age group One additionalcase was detected in a 49-year old patient The genderratio (MF) was 115

The temporal distribution of cases was atypical withregard to influenza infections (Fig 4)

Clinical information was available for 11 children andthe adult patient (Table 1) 6 presented with an Influenza-like illness (fever thorn cough) one with isolated fever one (in-fant) with respiratory syndrome without fever one withgastroenteritis two with neurological signs The adult pa-tient presented with respiratory insufficiency

Discussion

The current study provides novel information regarding thedistribution of seroprevalence in age-groups in mainlandand overseas (Reunion Island South West Indian Ocean)European territories It confirms the widespread exposureto Flu C in the human populations studied with globalseroprevalence values over 50 in all locations in accor-dance with a number of previous studies in France14

Spain15 and England12but also the United-States1619 Ja-maica17 Brazil18 Japan20 and Philippines21

Of note all serological techniques tested providedconvergent results In particular as previously observed25

IHA and ELISA tests provided very similar results The ELISA

seroprevalence rates observed in 3 different population groups J

Influenza C virus high seroprevalence rates 3

The purified viral suspension was aliquoted for conservationat 80 C

DSerum treatment

Sera were heat-inactivated at 56 C for 30 min

DHaemagglutination Inhibition assay (HI)

No receptor destroying enzyme (RDE) treatment wasperformed according to previous demonstration that thistreatment does not affect HI results in the case of FLUCV19

Viral antigen was prepared by diluting in PBS a non-inactivated FLUCV cell culture supernatant (see above)The haemagglutinating titre of viral antigen was deter-mined extemporaneously All tests were performed usingU-shaped microplates Twofold dilutions of sera in PBS (110 to 11280) and preparation of the HI mix were per-formed with an Eppendorf epMotion-5070-CB devicelocated in a biological safety cabinet 25 mL of viral antigendilution (corresponding to 533 haemagglutinating units)was added to 25 mL of each serum dilution and incubated50 min at 4 C Following addition of 25 mL of a 1 humanOthorn red blood cells suspension in PBS and another incubationof 30 min at 4 C the HI titre was determined as the highestdilution providing clear inhibition of haemagglutination Allexperiments included the same negative and positive con-trols and a serum agglutinating activity control

DELISA assay

Maxisorp plates (Nunc USA) were coated overnight at 4 Cwith a 1400 dilution in PBS of the sucrose gradient-purifiedFLUCV prepared as described above Plates were washedwith PBS Tween 005 (3 times between each subsequentstep) They were saturated with 1 Bovine Serum Albuminduring 2 h Serum samples diluted to 1400 in PBS thorn1milk were incubated 1 h at 37 C Peroxydase-conjugatedAffini Pure F(abrsquo)2 Fragment Goat Anti-Human IgG(H thorn L)(dilution 17500 Beckman Coulter) was added and incu-bated at 37 C for 1 h Tetramethylbenzidine was used assubstrate of peroxydase (Eurobio Life sciences) and aftera 4-min contact the reaction was stopped by addition of1 N citric acid Absorbance was read at 450 nm (SunriseTecan)

Each plate included a series of 6 blank wells (no serumadded) to determine the reaction background noise whichwas subtracted from ELISA optical density (OD) results

A series of 345 negative sera (ie providing negative re-sults in both HI and VNT techniques) was used to calculatethe cut-off value The OD limit of non-negativity for eachmicroplate (L) was calculated using the formulaL Z m thorn 2 SD where m and SD represent the mean ODand OD standard deviation values respectively calculatedfrom negative sera

DVirus Neutralisation Test (VNT)

Twofold serial dilutions of sera were mixed with a PBSdilution of culture supernatant corresponding to 50 TCID50of FLUCV (total volume 100 mL) and incubated for 1 h at

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

37 C Confluent MDCK cells (grown onto 96-well plates)were then infected with each dilution After a 2-h incubation the inoculum was removed and fresh EMEM[containing 3 foetal calf serum glutamine (2 mM) anti-biotics (penicillin 100UIml streptomycin 100 mgml) andtrypsin (10 mgmL)] was added In the absence of adetectable CPE produced by FLUCV virus growth wasdetected by qRT-PCR at day 5 pi (see below) The VNTtitre corresponded to the last serum dilution in which thevirus was not detected

DReal time RT-PCR

Viral RNA extraction was performed using the BioRobot EZ1Workstation and the EZ1 Virus Mini Kit V20 (Qiagen) Nu-cleic acids were extracted from 400 mL of sample andeluted in a 90 mL final volume

One-step RT-PCR was carried out in a 30 mL final volumeaccording to the manufacturerrsquos instructions using theldquoiScript One-Step RT-PCR for probesrdquo kit (Bio-Rad)8 pmol of each primer 04 pmol of probe and 10 mL ofRNA extraction The thermocycling profile included 50 Cfor 15 min 5 min at 95 C 95 C for 15 s (39 cycles) 60 Cfor 45 s The amplification was carried out using CFX96Real-Time PCR Detection System (Bio-Rad LaboratoriesInc)

Primers and probe sequences (matrix gene) were4

FluC_M-F CAT AAT TGA ACT TGT CAA TGG TTT TGTFluC_M-R TTC AGG CAT AAT TGT GGT CTT TAT ATC T probeFluC_M-FAM FAM-CTC GGC AGA TGG GAG AGA TGG TGT G-TAMRA

Results

Serological results

Performances of the different assaysAfter testing all 973 samples from Reunion Island using theHI and ELISA methods sera were randomly selected foreach HI titre (titre 5 4 sera 10 11 sera 20 9 sera 40 11sera 80 11 sera 160 11 sera 320 9 sera total 66 sera)and further tested using the VNT technique

The relationship between HI and VNT results is shown inFig 1a Using VNT as a reference test (samples with aVNT titre 10 were considered positive) an HI cut-offvalue at 80 (ie samples with a titre 80 were consid-ered positive) was associated with high Specificity (86)and Positive Predictive (90) values and a lower Sensi-tivity (80)

The relationship between ELISA and VNT results isshown in Fig 1b The cut-off value calculated asdetailed in the Methods section was 0797 Using VNTas a reference test ELISA was associated with highSpecificity (86) and Positive Predictive (91) valuesand a modest Sensitivity (66)

When HI and ELISA results from the 973 Reunion islandsera were compared a strong linear correlation be-tween the HI titre and the corresponding log of theELISA OD median value was observed (see Fig 2)

seroprevalence rates observed in 3 different population groups J

Figure 1 a Relationship between Haemagglutination Inhibi-tion titre and results of the Virus Neutralization Test in 66 serarandomly selected from the Reunion Island CoPanFlu cohortSamples with a VNT titre 10 were considered positive bRelationship between Haemagglutination Inhibition titre andresults of the Virus Neutralization Test in 66 sera randomlyselected from the Reunion Island CoPanFlu cohort Confirma-tion of ELISA results by VNT assay according to cut off

Figure 2 Relationship between ELISA median DO values andHI titres in 973 Reunion Island sera Of note the 0797 OD value(used as cut-off value for the ELISA) is associated with an HItitre close to 180

4 N Salez et al

Importantly the abovementioned 0797 ELISA OD cut-off value was associated with an HI titre at 80

Overall this evaluation of the different serologicalassays indicated that results provided by our ELISA werevery similar to those provided by HI using a cut-off at 80Using VNT as a reference both tests were associated withsimilar and high Specificity and Positive Predictive valuesand globally underestimated the number of true positivesThis implicates that ELISA-based prevalence numbers ob-tained in the current study constitute robust but lowestimates of the actual FLUC seroprevalence

Seroprevalence results

In Reunion Island HI and ELISA results from 973 samplesprovided a very similar distribution of positives in agegroups (Fig 3A) The seroprevalence was around 30in children under the age of 4yo and around 50 in olderage groups A similar trend was observed for the HI geo-metric mean titres (GMT)

In metropolitan France (Marseille) 331 sera weretested by ELISA The distribution in age classes(Fig 3B) also showed a rapid progression of seropreva-lence between children under the age of 4yo (w30)

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

and those in the 4e10yo age group (w60) The sero-prevalence reached w75 in individuals older than50 yo Because these results were similar to those ob-tained in Reunion Island the panel of sera studiedwas not further extended

In the UK (Edinburgh) 95 sera from patients older than30yo were tested by ELISA In the 31e50 yo and gt50 yoage classes results were similar with those observed inMarseille (Fig 3B)

Molecular biology results

A total of 28 influenza C cases were identified from 673nasal swab samples tested using a real-time RT-PCR assaySeventeen cases were diagnosed in children under the ageof 4 yo and ten in the 4e10 yo age group One additionalcase was detected in a 49-year old patient The genderratio (MF) was 115

The temporal distribution of cases was atypical withregard to influenza infections (Fig 4)

Clinical information was available for 11 children andthe adult patient (Table 1) 6 presented with an Influenza-like illness (fever thorn cough) one with isolated fever one (in-fant) with respiratory syndrome without fever one withgastroenteritis two with neurological signs The adult pa-tient presented with respiratory insufficiency

Discussion

The current study provides novel information regarding thedistribution of seroprevalence in age-groups in mainlandand overseas (Reunion Island South West Indian Ocean)European territories It confirms the widespread exposureto Flu C in the human populations studied with globalseroprevalence values over 50 in all locations in accor-dance with a number of previous studies in France14

Spain15 and England12but also the United-States1619 Ja-maica17 Brazil18 Japan20 and Philippines21

Of note all serological techniques tested providedconvergent results In particular as previously observed25

IHA and ELISA tests provided very similar results The ELISA

seroprevalence rates observed in 3 different population groups J

Figure 1 a Relationship between Haemagglutination Inhibi-tion titre and results of the Virus Neutralization Test in 66 serarandomly selected from the Reunion Island CoPanFlu cohortSamples with a VNT titre 10 were considered positive bRelationship between Haemagglutination Inhibition titre andresults of the Virus Neutralization Test in 66 sera randomlyselected from the Reunion Island CoPanFlu cohort Confirma-tion of ELISA results by VNT assay according to cut off

Figure 2 Relationship between ELISA median DO values andHI titres in 973 Reunion Island sera Of note the 0797 OD value(used as cut-off value for the ELISA) is associated with an HItitre close to 180

4 N Salez et al

Importantly the abovementioned 0797 ELISA OD cut-off value was associated with an HI titre at 80

Overall this evaluation of the different serologicalassays indicated that results provided by our ELISA werevery similar to those provided by HI using a cut-off at 80Using VNT as a reference both tests were associated withsimilar and high Specificity and Positive Predictive valuesand globally underestimated the number of true positivesThis implicates that ELISA-based prevalence numbers ob-tained in the current study constitute robust but lowestimates of the actual FLUC seroprevalence

Seroprevalence results

In Reunion Island HI and ELISA results from 973 samplesprovided a very similar distribution of positives in agegroups (Fig 3A) The seroprevalence was around 30in children under the age of 4yo and around 50 in olderage groups A similar trend was observed for the HI geo-metric mean titres (GMT)

In metropolitan France (Marseille) 331 sera weretested by ELISA The distribution in age classes(Fig 3B) also showed a rapid progression of seropreva-lence between children under the age of 4yo (w30)

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

and those in the 4e10yo age group (w60) The sero-prevalence reached w75 in individuals older than50 yo Because these results were similar to those ob-tained in Reunion Island the panel of sera studiedwas not further extended

In the UK (Edinburgh) 95 sera from patients older than30yo were tested by ELISA In the 31e50 yo and gt50 yoage classes results were similar with those observed inMarseille (Fig 3B)

Molecular biology results

A total of 28 influenza C cases were identified from 673nasal swab samples tested using a real-time RT-PCR assaySeventeen cases were diagnosed in children under the ageof 4 yo and ten in the 4e10 yo age group One additionalcase was detected in a 49-year old patient The genderratio (MF) was 115

The temporal distribution of cases was atypical withregard to influenza infections (Fig 4)

Clinical information was available for 11 children andthe adult patient (Table 1) 6 presented with an Influenza-like illness (fever thorn cough) one with isolated fever one (in-fant) with respiratory syndrome without fever one withgastroenteritis two with neurological signs The adult pa-tient presented with respiratory insufficiency

Discussion

The current study provides novel information regarding thedistribution of seroprevalence in age-groups in mainlandand overseas (Reunion Island South West Indian Ocean)European territories It confirms the widespread exposureto Flu C in the human populations studied with globalseroprevalence values over 50 in all locations in accor-dance with a number of previous studies in France14

Spain15 and England12but also the United-States1619 Ja-maica17 Brazil18 Japan20 and Philippines21

Of note all serological techniques tested providedconvergent results In particular as previously observed25

IHA and ELISA tests provided very similar results The ELISA

seroprevalence rates observed in 3 different population groups J

Figure 3 Seroprevalence results in age groups A Results inReunion Island (973 sera) The figure provides prevalence re-sults obtained using the Haemagglutination Inhibition (cut-offtitre 80) and ELISA methods It also includes HI GMT valuesB ELISA-based seroprevalence data in Reunion Island (973sera) metropolitan France (331 sera) and the UK (95 sera)

Influenza C virus high seroprevalence rates 5

method that was used for the analysis of all groups wasassociated with high Positive Predictive values but a modestsensitivity (using virus sero-neutralisation as a reference)This means that the actual proportion of the populationthat underwent Flu C virus infection is somewhat higherthan indicated based on our seroprevalence values

Results in metropolitan France globally conform withdata produced by Manuguerra and collaborators14 from asample of 301 sera collected in 1988 despite minor differ-ences in the distribution of positives in age-groups thehighest rates for positives were identified in Manuguerrarsquosstudy in the 16e30 age class (similar to Andrewsrsquo study12)whilst in our study the seroprevalence peaked in the4e10 age group (similar to Gerth13) slightly decreased inyoung adults (11e30 age group) and then steadily increasedin older groups The same trend was identified in ReunionIsland regardless the method (HI or ELISA) and in Scottishsamples (limited to the study of adults) Such fluctuationsof seroprevalence values in age groups observed indifferent trans-sectional studies brought to publicationmay reflect different histories of previous FLUCV circula-tion in the populations concerned

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

The analysis of seroprevalence in children from ReunionIsland and metropolitan France showed in the current studya strong increase between those under the age of 4 and the4e10 age group from 306 to 598 (p lt 00005) This isevocative of a common contact with the virus during thefirst years of life as previously suggested by serolog-ical10121319 and incidence studies52226 This epidemiolog-ical pattern is further supported by our retrospective PCRanalysis of respiratory samples collected in 2011e2012 atthe Virology department of the Public Hospitals of Marseilleand that tested negative for Flu A Flu B hRSV and hMPV Itidentified a majority of cases under the age of 4 (1728)The clinical presentation of Flu C infection has been re-ported to be undistinguishable from other influenza infec-tions22 with cases attracting medical attention beingmore frequent in children under the age of 2522 Alto-gether the conjunction of seroprevalence data and thecharacteristics of incident cases reported in the literatureimplicate a variety of presentations with a significant pro-portion of a- or pauci-symptomatic cases mild ILIs but alsomore severe cases eg low respiratory tract infec-tions522 Here clinical files were available for 12 patientsWhilst it is not claimed that all symptoms were necessarilya consequence of FLUCV infection it is interesting to notethat a majority of mild ILIs was observed but also thatbronchiolitisbronchitis was noted in 4 cases vomitinggastroenteritis in 2 cases skin rash in 1 case and neurolog-ical symptoms in 3 cases

Robust incidence values for FLUCV infections of medicalimportance are difficult to provide It is important tomention that the value of RT-PCR tests for the detectionof FLUCV has not been investigated in depth Whilst it hasbeen repeatedly reported that molecular detection wasmore sensitive than virus isolation52728 the limited numberof sequences available in databases (only 1 completegenomic sequence of human origin) has not permitted thedevelopment and comparison of a variety of assays Thisspecific aspect deserves further investigations since the cir-culation of antigenicgenotypic variants has beenreported1328e30

In previous studies using a combination of PCR andculture for detection whatever the Flu AampB hRSV andhMPV status the rate of FLUCV diagnosis was 3931 or4227 amongst children under the age of fifteen and10 amongst children under the age of six5 Hereamongst children in the same age group (lt6 yo) anextrapolated estimate of 33 (whatever the Flu AampBhRSV and hMPV status) was positive for FLUCV RT-PCRThis possibly slightly underestimates the actual infectionrate in the period (it was hypothesised that no co-infections with Flu AampB hRSV and hMPV occurred) butas for other respiratory viral infections significant year-to-year variations are possible and the high rate observedmay also over-represent the general rate of FLUCV infec-tions if the study was performed during a transitory dis-ease spillover

Very little is known regarding the temporal distributionof Flu C cases Matsuzaki et al (2007) reported a Flu Coutbreak in Japan that occurred between January and June2004 with a peak in May Pabbaraju et al (2013) reportedin Canada cases occurring between December and AprilGouarin et al (2008) reported in metropolitan France a

seroprevalence rates observed in 3 different population groups J

Figure 4 Temporal distribution of influenza A B amp C and human respiratory syncytial virus cases in Marseille metropolitanFrance

Table 1 Clinical information relating to 12 patients with biologically confirmed Influenza C virus infection

Age (years) Admission date Gender Symptoms (keywords)

040 Mar-11 M Coughrhinorrhoeabronchiolitisvomiting067 Mars-11 M Fever thorn seizureunconsciousness089 May-11 M Fever thorn coughrhinorrhoea392 May-11 M Fever144 May-11 F Fever thorn coughrhinorrhoea248 Feb-12 M Fever thorn coughbronchiolitis218 May-11 M Fever thorn coughrhinorrhoea400 Feb-12 M Fever thorn coughrhinorrhoea574 May-12 F Fever thorn gastroenteritis625 Jun-11 F Fever thorn coughbronchitisskin rash685 Jul-11 M Bronchiolitishemiparesisdrowsiness4969 Jun-12 M Respiratory insufficiency

6 N Salez et al

Please cite this article in press as Salez N et al Influenza C virus high seroprevalence rates observed in 3 different population groups JInfect (2014) httpdxdoiorg101016jjinf201403016

Influenza C virus high seroprevalence rates 7

majority of cases occurring during the winter season butalso a few cases identified in spring and summer Antonet al (2011) reported in Spain cases occurring during all 4seasons with the highest numbers observed in summer Inthe current study FLUCV circulation was observed duringthe winter flu season but interestingly a majority of caseswas observed out of the periods of circulation of Flu A Flu Band hRSV A large proportion of cases occurred duringspring and cases were also observed during the summerseason suggesting that the epidemiological circulationpattern of Flu C that occurred during the study periodmay have been more similar to that of hRSV (yearlypersistence with seasonal spillovers) than to that of flu A(yearly re-introduction of new variants) This is in agree-ment with evolutionary analyses by Anton et al (2011) whoshowed that strains circulating in Catalonia during spring2010 were phylogenetically closely related with strainscirculating in the same region in September 2009 Thiscertainly deserves further investigations with a refinedevolutionary analysis of circulating strains during prolongedperiods

In conclusion the current state of our knowledgeregarding Flu C is quite paradoxical Whilst during thelast decade viruses newly identified in respiratory sampleshave attracted considerable attention from the medicaland scientific community (even though their medicalimportance and implication in acute respiratory infectionswas sometimes limited or disputable) FLUCV which hasbeen identified in humans more than 65 years ago and isrecognised to be a frequent pathogen in children (espe-cially under the age of 4) and can be associated with severepresentations is the object of few studies We suggest thatthe detection of FLUCV should be included in respiratoryvirus testing panels28 and should be further evaluated andthat more globally obtaining more precise informationregarding epidemiological clinical and molecular aspectsof FLUCV infections would be of significant medical and sci-entific interest

Acknowledgements

We are grateful to Dr Mikhail Yurievich Eropkin (Head ofthe laboratory of evolutionary variability of influenzaviruses Research Institute of Influenza St PetersburgRussia) for kindly providing us with an Influenza C strainand advice regarding Influenza C propagation in cellculture

References

1 Taylor RM Studies on survival of influenza virus between epi-demics and antigenic variants of the virus Am J Public HealthNations Health 1949 Feb39(2)171e8

2 Ritchey MB Palese P Kilbourne ED RNAs of influenza A B andC viruses J Virol 1976 May18(2)738e44 [PubMed PMID944790]

3 Bellau-Pujol S Vabret A Legrand L Dina J Gouarin S Petit-jean-Lecherbonnier J et al Development of three multiplexRT-PCR assays for the detection of 12 respiratory RNA virusesJ Virol Methods 2005 Jun126(1e2)53e63

4 Faux C Influenza type C PCR for clinical microbiologySpringer SciencethornBusiness Media BV 2010 pp 311e2

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

5 Gouarin S Vabret A Dina J Petitjean J Brouard J Cuvillon-Nimal D et al Study of influenza C virus infection in FranceJ Med Virol 2008 Aug80(8)1441e6

6 Kimura H Abiko C Peng G Muraki Y Sugawara K Hongo Set al Interspecies transmission of influenza C virus betweenhumans and pigs Virus Res 1997 Apr48(1)71e9

7 Peng G Hongo S Muraki Y Sugawara K Nishimura H Kitame Fet al Genetic reassortment of influenza-C viruses in man JGen Virol 1994 Dec753619e22

8 Katagiri S Ohizumi A Ohyama S Homma M Follow-up study oftype C influenza outbreak in a childrenrsquos home Microbiol Im-munol 198731(4)337e43

9 Gerber P Woolridge RL Seal JR Ziegra SR Epidemic influenzaB and C in navy recruits during winter of 1951e52 Proc Soc ExpBiol Med 1952 Dec81(3) Unknown

10 Hilleman MR Werner JH Gauld RL Influenza antibodies in thepopulation of the USA an epidemiological investigation BullWorld Health Organ 19538(5e6)613e31

11 Minuse E Quilligan Jr JJ Francis Jr T Type C influenza virus IStudies of the virus and its distribution J Lab Clin Med 1954Jan43(1)31e42

12 Andrews BE McDonald JC Influenza virus C infection in En-gland Br Med J 1955 Oct 222(4946)992e4 [PubMed PMID13260642]

13 Gerth HJ Bauer KH Steinitz H Is there evidence for antigenicdrift of influenza C virus Zentralbl Bakteriol Parasitenkd In-fekt Hyg 197623147e56

14 Manuguerra JC Hannoun C Aymard M Influenza C virus infec-tion in France J Infect 1992 Jan24(1)91e9

15 Manuguerra JC Hannoun C Saenz Mdel C Villar E Cabezas JASero-epidemiological survey of influenza C virus infection inSpain Eur J Epidemiol 1994 Feb10(1)91e4

16 Dykes AC Cherry JD Nolan CE A clinical epidemiologic sero-logic and virologic study of influenza C virus infection ArchIntern Med 1980 Oct140(10)1295e8

17 Jennings R Respiratory viruses in Jamaica a virologic andserologic study 3 Hemagglutination-inhibiting antibodies totype B and C influenza viruses in the sera of Jamaicans Am JEpidemiol 1968 Mar87(2)440e6

18 Motta FC Luiz MO Couceiro JN Serological analysis revealscirculation of influenza C viruses Braz Rev Saude Publica2000 Apr34(2)204e5

19 OrsquoCallaghan RJ Gohd RS Labat DD Human antibody to influ-enza C virus its age-related distribution and distinction fromreceptor analogs Infect Immun 1980 Nov30(2)500e5

20 Kaji M Hiromatsu Y Kashiwagi S Hayashi J Oyama SKatagiri S et al Distribution of antibodies to influenza C virusKurume Med J 198330(3)121e3

21 Nishimura H Sugawara K Kitame F Nakamura K Sasaki HPrevalence of the antibody to influenza C virus in a northernLuzon Highland village Philippines Microbiol Immunol 198731(11)1137e43

22 Matsuzaki Y Katsushima N Nagai Y Shoji M Itagaki TSakamoto M et al Clinical features of influenza C virus infec-tion in children J Infect Dis 2006 May 1193(9)1229e35

23 Dellagi K Rollot O Temmam S Salez N Guernier VPascalis H et al Pandemic influenza due to pH1N12009 vi-rus estimation of infection burden in Reunion Island througha prospective serosurvey austral winter 2009 PLoS One20116(9)e25738

24 Delangue J Salez N Ninove L Kieffer A Zandotti C Seston Met al Serological study of the 2009 pandemic due to InfluenzaA H1N1 in the metropolitan French population Clin MicrobiolInfect 2012 Feb18(2)177e83

25 Troisi CL Monto AS Comparison of enzyme-linked immunosor-bent assay and hemagglutination inhibition in a seroepidemio-logical study of influenza type C infection J Clin Microbiol1981 Nov14(5)516e21

seroprevalence rates observed in 3 different population groups J

Influenza C virus high seroprevalence rates 7

majority of cases occurring during the winter season butalso a few cases identified in spring and summer Antonet al (2011) reported in Spain cases occurring during all 4seasons with the highest numbers observed in summer Inthe current study FLUCV circulation was observed duringthe winter flu season but interestingly a majority of caseswas observed out of the periods of circulation of Flu A Flu Band hRSV A large proportion of cases occurred duringspring and cases were also observed during the summerseason suggesting that the epidemiological circulationpattern of Flu C that occurred during the study periodmay have been more similar to that of hRSV (yearlypersistence with seasonal spillovers) than to that of flu A(yearly re-introduction of new variants) This is in agree-ment with evolutionary analyses by Anton et al (2011) whoshowed that strains circulating in Catalonia during spring2010 were phylogenetically closely related with strainscirculating in the same region in September 2009 Thiscertainly deserves further investigations with a refinedevolutionary analysis of circulating strains during prolongedperiods

In conclusion the current state of our knowledgeregarding Flu C is quite paradoxical Whilst during thelast decade viruses newly identified in respiratory sampleshave attracted considerable attention from the medicaland scientific community (even though their medicalimportance and implication in acute respiratory infectionswas sometimes limited or disputable) FLUCV which hasbeen identified in humans more than 65 years ago and isrecognised to be a frequent pathogen in children (espe-cially under the age of 4) and can be associated with severepresentations is the object of few studies We suggest thatthe detection of FLUCV should be included in respiratoryvirus testing panels28 and should be further evaluated andthat more globally obtaining more precise informationregarding epidemiological clinical and molecular aspectsof FLUCV infections would be of significant medical and sci-entific interest

Acknowledgements

We are grateful to Dr Mikhail Yurievich Eropkin (Head ofthe laboratory of evolutionary variability of influenzaviruses Research Institute of Influenza St PetersburgRussia) for kindly providing us with an Influenza C strainand advice regarding Influenza C propagation in cellculture

References

1 Taylor RM Studies on survival of influenza virus between epi-demics and antigenic variants of the virus Am J Public HealthNations Health 1949 Feb39(2)171e8

2 Ritchey MB Palese P Kilbourne ED RNAs of influenza A B andC viruses J Virol 1976 May18(2)738e44 [PubMed PMID944790]

3 Bellau-Pujol S Vabret A Legrand L Dina J Gouarin S Petit-jean-Lecherbonnier J et al Development of three multiplexRT-PCR assays for the detection of 12 respiratory RNA virusesJ Virol Methods 2005 Jun126(1e2)53e63

4 Faux C Influenza type C PCR for clinical microbiologySpringer SciencethornBusiness Media BV 2010 pp 311e2

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

5 Gouarin S Vabret A Dina J Petitjean J Brouard J Cuvillon-Nimal D et al Study of influenza C virus infection in FranceJ Med Virol 2008 Aug80(8)1441e6

6 Kimura H Abiko C Peng G Muraki Y Sugawara K Hongo Set al Interspecies transmission of influenza C virus betweenhumans and pigs Virus Res 1997 Apr48(1)71e9

7 Peng G Hongo S Muraki Y Sugawara K Nishimura H Kitame Fet al Genetic reassortment of influenza-C viruses in man JGen Virol 1994 Dec753619e22

8 Katagiri S Ohizumi A Ohyama S Homma M Follow-up study oftype C influenza outbreak in a childrenrsquos home Microbiol Im-munol 198731(4)337e43

9 Gerber P Woolridge RL Seal JR Ziegra SR Epidemic influenzaB and C in navy recruits during winter of 1951e52 Proc Soc ExpBiol Med 1952 Dec81(3) Unknown

10 Hilleman MR Werner JH Gauld RL Influenza antibodies in thepopulation of the USA an epidemiological investigation BullWorld Health Organ 19538(5e6)613e31

11 Minuse E Quilligan Jr JJ Francis Jr T Type C influenza virus IStudies of the virus and its distribution J Lab Clin Med 1954Jan43(1)31e42

12 Andrews BE McDonald JC Influenza virus C infection in En-gland Br Med J 1955 Oct 222(4946)992e4 [PubMed PMID13260642]

13 Gerth HJ Bauer KH Steinitz H Is there evidence for antigenicdrift of influenza C virus Zentralbl Bakteriol Parasitenkd In-fekt Hyg 197623147e56

14 Manuguerra JC Hannoun C Aymard M Influenza C virus infec-tion in France J Infect 1992 Jan24(1)91e9

15 Manuguerra JC Hannoun C Saenz Mdel C Villar E Cabezas JASero-epidemiological survey of influenza C virus infection inSpain Eur J Epidemiol 1994 Feb10(1)91e4

16 Dykes AC Cherry JD Nolan CE A clinical epidemiologic sero-logic and virologic study of influenza C virus infection ArchIntern Med 1980 Oct140(10)1295e8

17 Jennings R Respiratory viruses in Jamaica a virologic andserologic study 3 Hemagglutination-inhibiting antibodies totype B and C influenza viruses in the sera of Jamaicans Am JEpidemiol 1968 Mar87(2)440e6

18 Motta FC Luiz MO Couceiro JN Serological analysis revealscirculation of influenza C viruses Braz Rev Saude Publica2000 Apr34(2)204e5

19 OrsquoCallaghan RJ Gohd RS Labat DD Human antibody to influ-enza C virus its age-related distribution and distinction fromreceptor analogs Infect Immun 1980 Nov30(2)500e5

20 Kaji M Hiromatsu Y Kashiwagi S Hayashi J Oyama SKatagiri S et al Distribution of antibodies to influenza C virusKurume Med J 198330(3)121e3

21 Nishimura H Sugawara K Kitame F Nakamura K Sasaki HPrevalence of the antibody to influenza C virus in a northernLuzon Highland village Philippines Microbiol Immunol 198731(11)1137e43

22 Matsuzaki Y Katsushima N Nagai Y Shoji M Itagaki TSakamoto M et al Clinical features of influenza C virus infec-tion in children J Infect Dis 2006 May 1193(9)1229e35

23 Dellagi K Rollot O Temmam S Salez N Guernier VPascalis H et al Pandemic influenza due to pH1N12009 vi-rus estimation of infection burden in Reunion Island througha prospective serosurvey austral winter 2009 PLoS One20116(9)e25738

24 Delangue J Salez N Ninove L Kieffer A Zandotti C Seston Met al Serological study of the 2009 pandemic due to InfluenzaA H1N1 in the metropolitan French population Clin MicrobiolInfect 2012 Feb18(2)177e83

25 Troisi CL Monto AS Comparison of enzyme-linked immunosor-bent assay and hemagglutination inhibition in a seroepidemio-logical study of influenza type C infection J Clin Microbiol1981 Nov14(5)516e21

seroprevalence rates observed in 3 different population groups J

8 N Salez et al

26 Katagiri S Ohizumi A Homma M An outbreak of type C influ-enza in a childrenrsquos home J Infect Dis 1983 Jul148(1)51e6

27 Matsuzaki Y Ikeda T Abiko C Aoki Y Mizuta K Shimotai Yet al Detection and quantification of influenza C virus in pedi-atric respiratory specimens by real-time PCR and comparisonwith infectious viral counts J Clin Virol 2012 Jun54(2)130e4

28 Pabbaraju K Wong S Wong A May-Hadford J Tellier RFonseca K Detection of influenza C virus by a real-time RT-PCR assay Influenza Other Respir Viruses 2013 Feb 27

29 Anton A Marcos MA Codoner FM de Molina P Martinez ACardenosa N et al Influenza C virus surveillance during the

Please cite this article in press as Salez N et al Influenza C virus highInfect (2014) httpdxdoiorg101016jjinf201403016

first Influenza A (H1N1) 2009 pandemic wave in CataloniaSpain Diagn Microbiol Infect Dis 2011 Apr69(4)419e27

30 Matsubara K Sakano T Takao S Daikoku K Influenza C virusisolated in Hiroshima Prefecture during the 19992000 winterseasonea clinical and epidemiological study KansenshogakuZasshi 2004 Jun78(6)470e5

31 Matsuzaki Y Abiko C Mizuta K Sugawara K Takashita EMuraki Y et al A nationwide epidemic of influenza C virusinfection in Japan in 2004 J Clin Microbiol 2007 Mar45(3)783e8

seroprevalence rates observed in 3 different population groups J