Equine influenza (EI) is one of the most important respiratory diseases of horses. Beyond the welfare issue induced by this infectious disease, the potential impact for the equine industry could be devastating, as clearly illustrated in 2007 when EI reached for the first time the naïve population of Australia, it infected over 76,000 horses [1
]. The overall cost for the Australian economy to control this epizooty and to regain its EI-free OIE status was estimated to have reached A$
Due to the density of the equine population and a very high transmission capacity, prevention methods such as vaccination are essential to prevent or control EI outbreaks. The efficacy of vaccines against equine influenza virus (EIV) infection has been demonstrated through numerous clinical and field studies [2
] but little or no information is available about EI immunity in the field and vaccine coverage in horse populations. In humans, most of the studies that investigate vaccination coverage are based on an epidemiological questionnaire and showed that the influenza vaccination coverage is often less than 60% [6
]. As for equine influenza, vaccine coverage is well known to be essential to prevent human influenza, as recently illustrated by Uchida et al. [10
], who showed that high vaccine coverage was significantly and negatively correlated with the level of influenza epidemic among elementary school units. A postal questionnaire survey of randomly selected horse owners in Great Britain has reported a frequency of vaccination against EI and tetanus of 71.3% [11
]. Through a retrospective mathematical modeling study of the 1971 Japanese EI outbreak in racehorse facilities, Satou and Nishiura calculated that 50% to 80% of an equine population should be vaccinated with a completely effective EI vaccine to achieve protection against an EIV strain with a relatively low to moderate rate of transmission, respectively. Taking into account that sterilizing immunity is rarely achieved by EI immunization and that EI vaccines efficacy wanes with time, the author also provided a more realistic vaccine coverage threshold of 86.5% [12
The French horse population regroups 1.106 million horses (as recorded in 2016), which represents 15% of the European equid population with France placed third in terms of number behind Germany and the United Kingdom. Despite active disease surveillance that involves around 800 equine veterinary practitioners taking part in the French equine pathology epidemiological surveillance network (RESPE), EI has not been detected in France between mid-2015 to early December 2018 [13
]. During this period, there was evidence of EIV circulation in other European countries and worldwide [14
]. Since early December 2018, several European countries (Belgium, France, Germany, Ireland, the Netherlands, Sweden and the United Kingdom) have reported unusual levels of EIV circulation. The number of EI outbreaks is staggering in some countries, with more than 200 outbreaks reported in the United Kingdom between January and October 2019, leading to a 6 day shutdown of horse racing in February (the last shutdown linked to EI dated from 1979) and more than 174 racing stables placed in lockdown with mandatory EIV testing. This situation, unseen in Europe since the late 1970s and 1980s, reminds us of the impact of the 5 months long EI in Australia in 2007, when over 76,000 horses were infected with EIV and with an economical cost reaching A$
1 billion [1
]. The H3N8 EIV at the origin of the current European outbreaks belongs to the Florida Clade 1 (FC1) sub-lineage, which was usually circulating in North and South Americas [17
] with only occasional appearances in Europe from time to time [18
]. FC1 EIV strain was not isolated in France since 2009 [13
The current report aims to present and discuss the results from a large-scale sero-epidemiological study designed to provide a picture of the immunological status against EI of the French horse population in late 2017 and the first incursion in a decade of an FC1 EIV. These results should provide information about vaccine coverage and an explanation about the absence of reported EI outbreaks between mid-2015 and late 2018. Unlike previous studies on vaccination coverage against human and equine influenza A viruses, this study is based on an evaluation of antibody levels measured by single radial hemolysis assay (SRH), a well-recognized correlate of protection against EIV infection and by an enzyme-linked immunosorbent assay (ELISA), which detects antibodies against the viral nucleoprotein (NP) of type A influenza viruses and could be used as a DIVA (differentiating infected from vaccinated animals) marker in specific situations [1
]. Due to the large-scale EI epizooty affecting France and other European countries since late 2018/early 2019, details of the French EI outbreaks will also be reported. A phylogenetic analysis of the EIV strains involved and the detail of amino acid substitutions in the hemagglutinin (HA) protein sequence and antigenic sites [15
], which are primary targets for virus-neutralizing antibody response, will be presented. These elements will be discussed in relation to EI vaccination in the field and the risk of vaccine breakdown.
Recent data indicate that influenza vaccination coverage in the US is often below 60% in adult humans and below 45% in children [6
]. This present study uses a serological test to evaluate the level of specific antibodies to EIV unlike equivalent studies in humans that favor questionnaires and epidemiological surveys. The overall analysis of immune coverage in horses from the four French regions studied highlights that only 12.4% of horses were considered as unprotected, based on their SRH antibody level at the time of sampling. A large proportion of horses have reached the virological protection threshold (60.3% ≥ 154 mm2
). Overall immune coverage is estimated at 87.6%, a level described by Satou and Nishiura (2006) [12
] as sufficient to protect an equine population against an equine influenza virus presenting a relatively low rate of transmission (retrospective study based on the 1971 EI outbreak in a Japanese racehorse facility). Such immune coverage in the French horse population studied could provide an explanation for a lack of detection of EIV and/or the disease between mid-2015 and late 2018, despite extensive surveillance by the RESPE and associated diagnostic laboratories (several hundreds of nasopharyngeal swab samples analyzed every year). Prior to this period, the last French EI outbreaks were reported in 2014 and 2015 but were limited in size and number (six and four, respectively) [13
]. Other European countries with large equids population report EI cases every year [15
]. While no or limited information is available, it would be interesting to investigate actual EI vaccine use and coverage in these countries to determine if more frequent and recurrent EI outbreaks recorded through the years may be explained by lower EI vaccination (e.g., anecdotal information based on the number of EI vaccine doses sold per year in relation to the number of horses may be used to calculate rough estimate of EI vaccine coverage).
This study, which aimed to provide an idea of the immune/vaccine coverage against EI in the French horse population, used a serological test to evaluate the level of EIV-specific antibodies unlike equivalent studies in humans or horses that used questionnaires, epidemiological survey and provide vaccine coverage results based on owners and veterinarians confirmation of vaccination. The SRH assay was selected for this study due to the well-known and described correlation between SRH antibody titers and protection against EIV infection, development of clinical signs of EI and virus shedding. The EIV strain used as SRH antigen was selected to be close to the FC2 recommended vaccine strain A/eq/Richmond/07, which is contained in the HA recombinant canarypox-based EI vaccine predominantly used in France. At the time of the serological study, EIV strains circulating in Europe were mostly belonging to the FC2 sub-lineage, which supported the SRH antigen choice. Cross-reactivity and cross-protection have also often been documented in recent vaccine clinical trials. Preliminary data indicates that serums used in this study also show cross-reactivity when tested against the FC1 EIV strain A/eq/Paris/1/2018, but further studies are warranted to explain the current EI epidemic in Europe. The current study aimed to be representative of the French horse population. However, some limitations were inevitable and should be taken into account for the interpretation of results: (i) four geographical regions that represented 57.5% of all French equids breeding centers were selected, with a known bias for Normandy due to the localization of the serum archive (i.e., LABÉO). When adjusted on the number of breeding centers per region (cf. Supplementary Figure S1
), overall results were not significantly different (p
-value = 0.085), with 86.8% above the 85 mm2
(when compared with 87.6% without adjustment, cf. Figure 1
). (ii) Some horse populations are never visited by veterinary practitioners, for multiple reasons (e.g., economic, cultural, etc.). These populations, which are usually missed by the questionnaire and epidemiological surveys, are probably not represented in the current study. Obviously, such equids populations that are usually not vaccinated either, represent an important reservoir for pathogens and the weak link in any strategy of prevention.
Results for each French region selected in this study follow the same trend with protection rates reaching 84% to 87% (i.e., horses considered as vaccinated with clinical protection and reduced virus shedding in case of EIV infection). Some differences were measured between each sample’s categories. Results showed that “sale” and “diagnosis” categories were significantly lower when compared with other categories. Results indicate that after the exclusion of horses less than 2 years old at the time of sampling (born in 2016 and 2017), the immune coverage increased significantly from 79% to 96% for the “sale” category (in Normandy). A possible hypothesis is that horses born in 2016 and 2017 were sampled during their primary course of EI vaccination and may have not yet developed a complete and robust EI humoral immune response at the time of sampling. The results are correlated with a previous field study that highlighted a large frequency of Thoroughbred foals displaying low or negative SRH antibody titers during their primary EI immunization and up to 5 months after the third EI immunization (V3) [27
] and other clinical studies highlighting the immunity gap frequently observed in the weeks preceding the third EI immunization [28
]. In the field study, the frequency of seronegative foals (i.e., SRH = 0 mm2
) was greater than 25% at different sampling time points during the primary EI immunization (at the time of first immunization, two weeks and three months after the second immunization and two days after the third immunization) [27
]. After V3, the frequency of seronegative foals remained below 20% up to three months after this immunization [27
]. In the present report, the frequency of horses located in Normandy and born in 2016 and 2017 with negative or low SRH titers was 49% and 42% for “sale” and “diagnosis” categories, respectively.
The categories “breeding” and “exportation” present high protection rates that reach 87% at least (excepted for the “exportation” category from Pays de la Loire). This observation is not surprising because these categories are submitted to mandatory EI vaccination in France. For the covering season of stallions, some studbooks require that EI vaccination be carried out in accordance with the EI vaccine manufacturers recommendations (i.e., primary vaccination with two immunizations 4–6 weeks apart and a third dose 5 to 6 months after the last immunization, then annual boost immunization afterward) otherwise breeding cannot be authorized. Similarly, for mares, studbooks and horses, breeding centers require EI vaccination. The regulation to export horses is also strict. In most cases, the sanitary measures imposed by importing countries require that horses have received two immunizations prior to movement. The information available does not provide an explanation for the higher percentage of negative and <85 mm2 samples in the “exportation” group for Pays de la Loire.
The use of an EIV NP-specific enzyme-linked immunosorbent assay (ELISA) allows differentiating infected animals from horses vaccinated with the HA recombinant canarypox-based EI vaccine (ProteqFlu®
; Mérial, Lyon, France) [1
]. Horses infected with EIV or vaccinated with a whole inactivated EI vaccine produce antibodies against all viral proteins, including the EIV NP. These horses have a positive serological status for the NP (NP+) unlike horses immunized with the recombinant canarypox EI vaccine that seroconvert to HA only and therefore have a negative serological status for the NP (NP−) [31
]. The use of an EI vaccine with DIVA ability has proved very useful in past EI epidemics. Emergency vaccination was implemented during the 2007 Australian EI outbreak and this specific vaccine was selected for its successful use in South Africa in 2003 and its DIVA ability, amongst other reasons. The possibility to differentiate infected animals from vaccinates was of a great importance for the Australian’s EI outbreak management because it allowed to monitored EIV transmission and disease spread in light of ring EI vaccination [1
]. The use of a DIVA assay to evaluate EI vaccination coverage in an endemic situation depends on the type of EI vaccine available and used in the field (i.e., recombinant or whole inactivated EI vaccines). In France, based on anecdotal discussions with EI vaccine manufacturers, the HA recombinant canarypox-based EI vaccine is predominantly used. In our current study, the use of a DIVA test confirmed that the EI seropositivity of young horses (≤3 years) was primarily due to EI vaccination using this specific EI vaccine. The DIVA analysis showed that 83% of samples from young horses born in France and seropositive by SRH assay were EIV NP-seronegative compared to old horses for which only 52% were EIV NP-seronegative. This observation is consistent with the last EI outbreak registered in 2015 (RESPE) and the predominant use of this specific HA recombinant EI vaccine in France. Before 2015, EI cases have been occasionally observed, which could explain higher frequency of older horses seropositive for EIV-NP (48%). The high level of young horses seronegative for EIV-NP supports the RESPE surveillance data that reports no EI clinical case in France between mid-2015 and late 2018. While the use of whole inactivated EI vaccines may explain EIV-NP seropositivity for the other 17% of young horses born in France, a low level of EIV circulation in France could not entirely be ruled out. Only 53% of young horses born abroad were EIV-NP seronegative. These results could be explained by a broader diversity of EI vaccines used and possible EIV circulation in other countries. The use of the DIVA test revealed that 96% horses with SRH antibody titers equal to 0 mm2
were also seronegative by NP-ELISA. In the absence of non-H3 equine influenza circulation, the remaining 4% (SRH negative and NP-ELISA) may be explained by a difference of sensitivity between the two assays. Hopefully, veterinary vaccine manufacturers will incorporate DIVA markers in their future EI vaccines, which will greatly benefit disease surveillance, would help to identify poor-vaccine responders and contribute to reducing fraudulent vaccination in horses in the long term (i.e., “pen vaccination”).
Despite high EIV-specific immune coverage and the use of an EI vaccine fully updated according to the last OIE recommendation on EI vaccine strain composition (i.e., EI vaccines should contain representative EIV strains of both FC1 and FC2 sub-lineages), the absence of clinical EI in France came to an end in early December 2019. Numerous EI outbreaks were reported in several European countries (Belgium, Germany, Ireland, the Netherlands, Sweden and the United Kingdom). The scale, number and duration of this epidemic had not been experienced in Europe since the late 1970s and 1980s. Sequencing results revealed that H3N8 EIV strains at the origin of the 2018–2019 French outbreaks (and EI outbreaks in other European countries, OIE ESP communication) belong to the FC1 sub-lineage, which was not isolated in France since 2009 [13
] and was usually circulating in North and South Americas [17
]. While several amino acids mutation were identified in the HA, results from the hemagglutination inhibition assay using mono-specific ferret sera and the associated antigenic cartography analyses indicate that FC1 EIV strains the origin of the 2018–2019 EI outbreaks were still antigenically closely related to the recommended FC1 EIV strains for inclusion in EI vaccine [33
]. As a consequence, this epidemic was not considered by the OIE ESP to be linked to a mismatch with the EI vaccine strains and the EI vaccine strain recommendation remained unchanged in 2019. The introduction of more pathogenic strains could be an alternative explanation. The FC2 EIV strains isolated in Europe in recent years were of mild and decreasing pathogenicity [34
], which may provide an explanation for the absence of EI in France in recent years (results from this study indicate that EI immunity/vaccine coverage is close to the level described by Satou and Nishiura [12
]). However, a few anecdotal reports of EI induced mortality in the field in several EU countries and observation of mild but noticeable clinical signs of disease in EI vaccinated horses raise a question about the pathogenicity of the current FC1 EIV strain. At the time of this report, the current FC1 EIV strain has not been used in a controlled experimental infection, which prevents any strong assumption about its pathogenicity. Surveillance results reported in the current study highlight a larger amount of French EI outbreaks involving vaccinated horses, which is not entirely surprising when the high EI vaccine coverage measured here is taken into account. For the few cases where serum were obtained at the onset of disease, it appears that infection could be explained in half of the cases by a lower than expected SRH antibody titer at the time of contact with EIV (irrespective of the time since last immunization). With the exception of horse #6, which had an SRH antibody titer above the 154 mm2
threshold, other vaccinated horses had average titers (between 93 and 129 mm2
), probably high enough to significantly reduce the clinical signs of disease but insufficient to induce sterilizing immunity, which is rarely measured [35
], even in optimal study conditions.
The overall immune coverage in 2018, which was estimated at 87.6%, was sufficient to prevent EI clinical cases induced by FC2 EIV strains circulating in the EU between 2015 and late 2018. Regretfully, such an immune coverage (based on historical SRH protection thresholds) was not high enough to prevent the 2018–2019 FC1 EIV strains circulation. However, it is very important to note that all known field and veterinary reports indicate that clinical signs of disease observed in EI vaccinated horses were clearly reduced when compared with unvaccinated animals, which continue to support the benefit of EI vaccination. Vaccination has probably slowed the spread of EI in France, which provided invaluable time to Equine Veterinary Practitioners.