Epidemiological investigation of a tularaemia outbreak after a hare hunt in Bavaria, Germany, 2018

In November 2018, a tularaemia outbreak occurred in Bavaria, Germany, among participants of a hare hunt and butchery employees handling the hares. We conducted an epidemiological outbreak investigation, including a retrospective cohort study among hunting participants, to identify likely transmission routes and activities associated with infection. Twelve of 41 participants were antibody‐positive for Francisella (F.) tularensis (attack rate: 29%). Cases reported influenza‐like symptoms (n = 11), lymphadenopathy (n = 1) and conjunctivitis (n = 1). Infection only occurred in those hunting participants present while hares were processed, while risk of infection was highest when directly involved (RR = 10.0; 95%CI: 2.6–392). F. tularensis was isolated from 1/4 hares. Only two individuals reported using some of the recommended personal protective equipment (PPE). Occurrence of mainly non‐specific symptoms, likely due to early treatment, was not indicative of a specific transmission route. Transmissions via direct (skin/mucosa) contact and by inhalation of contaminated aerosols seem plausible. Promoting and increasing appropriate use of PPE among people processing hares is crucial to prevent future outbreaks.


Tularaemia is a bacterial infection caused by Francisella (F.) tularensis.
The broad host spectrum includes small mammals (hares, rabbits, mice, wildlife and pets), birds and amphibians, while different arthropods mainly play an important role as vectors (deer flies, mosquitoes, ticks) (Hestvik et al., 2015;World Health Organization, 2007).
Although fatality through infection with F. tularensis ssp. holarctica, the only documented cause of human tularaemia in Europe (Maurin & Gyuranecz, 2016), is very low in humans (ECDC, 2017;Kohlmann et al., 2014;Robert Koch Institute, 2016), complications are frequent and may prolong the course of disease (Maurin & Gyuranecz, 2016;World Health Organization, 2007). In most human cases, the incubation period is between 3 and 5 days, although depending on the route and dose of infection, it can range from 1 to 21 days; rarely, it may take up to several weeks (Robert Koch Institute, 2016;World Health Organization, 2007). Early clinical disease mainly presents with non-specific, influenza-like symptoms. Progressed clinical forms depend on the route of infection. Most common are ulceroglandular or glandular (contact to contaminated animal material/ water via skin lesions/mucous membrane, stings/bites of infected arthropods), oculoglandular (touching eye after contact to contaminated material/infected animal), oropharyngeal (oral intake of contaminated food/water) and pulmonal/respiratory forms (inhalation of contaminated dust/aerosols). Human-to-human transmission has not yet been described (Robert Koch Institute, 2016;World Health Organization, 2007).
In Germany, the annual incidence is <0.1/cases per 100,000 inhabitants (Faber et al., 2018;Kohlmann et al., 2014). Between 2015 and 2019, 253 cases were reported (annual average 51, range 34-72) (an der Heiden et al., 2019;Beermann et al., , 2017Robert Koch Institute, 2020;Sin et al., 2018). Nationally reported tularaemia infections are mainly sporadic single cases or small clusters, acquired in Germany (Faber et al., 2018). Known exposure in autochthonous cases could be linked to vectors (ticks or mosquitoes) in some and to hares/rabbits or meat products in most cases (Kohlmann et al., 2014). Most outbreaks in the past were associated with contact to infected hares, including two of the three largest ones in Germany since 2001, affecting between 6 and 11 people respectively. (Burckhardt et al., 2018;Hauri et al., 2010;Sin et al., 2013). Hunters and game butchers, who are frequently in contact with hares, represent a main risk group. The Federal Research Institute for Animal Health recommends avoidance of dust and aerosol formation, the use of gloves, a dust-tight breathing mask and safety goggles when handling game as well as no further disassembling of suspected game (Friedrich-Loeffler-Institut, 2015). The last officially reported fatal tularaemia infection in Germany was in 2017 (Sin et al., 2018). In the German state of Bavaria (~13 million inhabitants), the annual number of reported tularaemia cases ranged from 2 to 13 between 2013 and 2017. The local health authority conducted the epidemiological outbreak investigation in cooperation with the LGL, initiated immediate control measures and performed active case finding. Eight hares (Lepus europaeus) were shot during the hunt. Four of those were heavily damaged and therefore disposed at a rendering plant. The hunters processed the remaining four carcasses, hung them in a slaughter room at one hunter's home and provided them to a wild game butchery afterwards, where the meat was processed and sold to customers.
We had the opportunity to investigate this outbreak in a timely and detailed manner. Our aim was to assess and describe the extent of the outbreak, identify activities associated with infection and most likely transmission routes. Here, we report the epidemiological outbreak investigation. A manuscript focussing on laboratory testing in this outbreak has been published elsewhere (Jacob et al., 2020).

| Study design
All persons with potential exposure to F. tularensis during the hunt, through further processing of game or contact to potentially contaminated or infected hunting dogs were contacted by the local health authority and invited to participate in the epidemiological study. Our study included the collection of questionnaire information (self-administered) and at least one blood sample. The simultaneous collection was organized by the local health authority. Aiming to identify risk factors for infection, we conducted a retrospective cohort study including a subset of study participants, namely hunting participants. The investigation was also extended to the participating hunting dogs and hunted hares.

Impacts
• Game (including hares) infected with Francisella tularensis poses high risk of infection for persons getting in close contact.
• Early suspicion and rapid treatment and detection of tularaemia infection may help to prevent more severe courses of disease.
• In order to prevent further outbreaks, use of recommended PPE is necessary, which may be targeted through increased awareness of tularaemia infection and potential transmission routes, especially among risk groups handling hares.

| Study population
We defined a possible case as a participant with no antibodies against F. tularensis detected at the last sampling point, sampled at least two weeks after exposure, and who reported any of the following symptoms 1-21 days after exposure: fever, chills, headache, limb pain, sudden sweating episode(s), sore throat, nausea, vomiting, stomach pain, cough, stomatitis, pharyngitis, tonsillitis, difficulties breathing or shortness of breath, pneumonia, conjunctivitis, swelling of the eyelid(s), skin rash, skin inflammation or ulcer, swollen lymph node(s). We defined a confirmed case as a participant with antibodies (IgG, IgM and/or IgA) against F. tularensis, indicative of acute infection. We focussed on confirmed cases in the analysis; if not explicitly stated otherwise, 'cases' refer to confirmed cases.
The outbreak investigation was conducted as part of the authoritative, official task of the county health departments and the state health department (LGL), and was therefore exempt from institutional review board approval. In addition, all participants gave written consent for blood sampling and use of questionnaire information. We included all contacted persons who consented to participate in our study. In the outbreak investigation, we address the four identified exposure groups separately (hunting participants, butchery employees, veterinary staff, family members).

| Questionnaire
We developed a questionnaire for the cohort of hunting participants (Appendix, supplement 1), collecting information on symptoms, details about the hunting event and other risk factors. We adapted the questionnaire for butchery employees, veterinary staff and family members.

| Clinical information
Self-reported clinical information was complemented by the collaborating local hospital (Klinikum St. Marien, Amberg) with details about medical examinations performed and treatment applied. Eleven of 12 hospitalized participants, including the initial group of hunting participants, were admitted to and treated at the local hospital.

| Human testing
The collaborating local hospital performed diagnostic tests at their laboratory among admitted persons, using blood cultures, swabs and sera. Additional oropharyngeal swabs were tested for influenza A, B and respiratory syncytial virus (RSV). Sera were tested for Leptospira IgG and IgM in an external laboratory using an enzyme-linked immunosorbent assay (ELISA).
Every participant provided at least one blood sample, taken at least two weeks after the initial exposure, allowing this as a minimum time window for antibody development and detection (Jacob et al., 2020;Koskela & Salminen, 1985). The Consiliary Laboratory (CL) for Tularaemia at the Robert Koch Institute (RKI) tested all samples, including those initially tested by the local clinic. The CL used this outbreak as an opportunity to add to the scarce knowledge regarding seroconversion after contact to F. tularensis. Therefore, the CL obtained further samples of the initially hospitalized group of hunting participants and closely monitored their serology (Jacob et al., 2020).
For direct detection of F. tularensis by specific real-time PCR and inoculation on culture media, the CL used blood cultures (n = 56), throat swabs (n = 10) and one eye swab (case 12). 76 human blood sera (including repeated samples) were analysed for antibodies against the lipopolysaccharide (LPS) of F. tularensis. Serological testing was done by an ELISA for screening of antibodies against F. tularensis, and findings were confirmed by Western blot (WB).

| Animal testing
Organs and muscle tissues of four secured hares were analysed at the LGL department of animal pathology and bacteriology and PCRtested for F. tularensis genus and species holarctica. Differential diagnosis included cultural and molecular analysis for Pasteurella, Yersinia, Leptospira and Brucella. Cultures were further investigated, for example for the identification of the subspecies, at both the LGL and the CL.
Additionally, a local veterinary practice collected samples of involved dogs. The CL tested throat swabs and EDTA blood via specific real-time PCR and serum for the detection of antibodies against F. tularensis (Jacob et al., 2020).

| Statistical methods
The outbreak description included all persons for whom we identified a potential risk of infection and whom we aimed to interview, as well as involved hares and dogs. We displayed the temporal course, including important events of the outbreak and its investigation, and described cases by date of disease onset. We described attack rates by exposure group (hunting participants, butchery employees, veterinary staff and family members).
We conducted risk factor analysis for the cohort of hunting participants. We combined the five activities 'skinning', 'opening up', 'disembowelling', 'rinsing with hose' and 'handling of hares afterwards' as 'processing' of hares. We considered anyone who carried out at least one of those listed activities to have taken part in the processing. We calculated relative risks (RR) with 95% confidence intervals (CI) using a log-binomial model and taking laboratory confirmed acute infection with F. tularensis as outcome of interest. Inclusion of further variables did not yield in an improved multivariable model based on the Bayes information criteria (BIC). As the model with more than one explanatory variable did not perform better than the one with only one explanatory variable, we only report crude risk estimates, in order to avoid overfitting in this small study group.
Due to the small number of cases, we report p-values testing categorical data according to Fisher's exact test. For testing trends using categorical variables, we use an extension of the Wilcoxon rank-sum test. The significance threshold was set at 0.05. We only considered available data for description and analyses. We double entered questionnaire data in EpiInfo (version 4.4.1.0) and performed data analysis in Stata (version 16).

| Study participants
We identified 42 people with potential exposure to F. tularensis.
Case 12 belonged to two exposure groups (hunting participants and family members) and is therefore listed separately in both groups throughout ( Figure 1). Identified persons invited to participate in the investigation included hunting participants (n = 35), butchery employees who handled the hares after the hunt (n = 4), family members of a hunter's household where hares hung to bleed out (n = 2) and employees of a veterinary practice (n = 2; veterinary assistant and veterinarian) where several involved dogs were tested. Hare parts were sold vacuum packed. All customers were successfully contacted, and the meat was returned before anyone had opened the packaging. Thus, customers who had already purchased parts of the hunted hares were not included in further investigations.
We included 41 of the 42 invited persons (97.6%) in our outbreak investigation ( Figure 1); the veterinarian did not participate.
Questionnaire and laboratory information was available for all 41 participants; laboratory information was available for all four available hares and ten of the 11 involved dogs. We identified 11 possible and 12 confirmed cases. All confirmed cases reported having had symptoms ( Figure 2). First detection of antibodies was two weeks (n = 2), last 21 weeks (n = 1) after exposure (no testing between three and 21 weeks after exposure); most remaining cases were first confirmed following sampling three weeks after exposure (Jacob et al., 2020). Maximum time windows for serological testing after possible exposure among those with no evidence for an infection with F. tularensis were 14 days (n = 1), 18 days (n = 1), 19 days (n = 1), 26 days (n = 20), 34 days (n = 2), 37 days (n = 3) and 21 weeks (n = 1).

| Clinical and treatment information
All confirmed cases reported mainly influenza-like symptoms, except case 12 (Table 1). Headache (8/10), limb pain (8/10), chills (7/10) and sudden sweating episode(s) (5/10) were most often reported. A one-sided conjunctivitis (case 12), a pharyngitis and swollen lymph nodes were each reported once. None of the patients reported any F I G U R E 1 Tularaemia outbreak in Bavaria, Germany, 2018: Overview of persons potentially exposed to Francisella tularensis following a hare hunt by exposure group, including participation in the investigation (questionnaire and lab-test availability), case definition (possible and confirmed cases) and attack rate skin condition, such as skin inflammations or ulcers. The case with latest disease onset (case 12) was potentially exposed later; we address this in more detail under 'Family members'. The hunt, as the day of exposure, took place on October 27, and symptom onsets were reported between November 1 and November 20 ( Figure 2).
All but case 12 reported having consulted a physician and to have received antibiotic treatment.
According to the local hospital, nine hunting participants were admitted to their clinic as a precautionary measure for medical monitoring after the self-reported suspicion of tularaemia infection. Eight of them confirmed as cases. All nine reported influenza-like symptoms at time of admission, including headache, limb pain, fever and chills, general weakness and reduced general state of health as well as night sweats. Additionally, chest pain was reported in three cases, dizziness and dry cough in one case and painful swelling of lymph nodes in the left upper limb/axilla in one case. No one reported or was clinically diagnosed with ulcers or eschars, stomatitis, tonsillitis, pharyngitis or pneumonia.
No gastrointestinal symptoms were reported during their hospital stay. All nine hunters received an electrocardiography (ECG) and a chest X-ray. Chest X-rays did not reveal pulmonary infiltrates. Four  Two participants stated wearing gloves during processing activities (one case). Several participants used a cloth, water or water and soap to clean their hands following activities with direct hare contact (Table 2). One person who was involved in the processing of hares and tested negative had worn glasses (any type); no one had worn a mask.
Apart from those hunting participants directly involved in the processing of hares, 11 further participants were present, of whom two became cases, among those also case 12. Of the 14 remaining participants who were not present at the time of hare processing, none became a case. Of those, four reported direct contact to hares; all had carried them, during which one had worn gloves; one had additionally emptied the bladder without wearing gloves.
Eight of the ten dogs that joined the hunt had hare contact. Eight participants had subsequent contact to dog saliva or rabbit blood attached to the dog; four of them became cases. Three of the four had also been directly involved in the processing of hares, whereas the fourth, case 12, was only present during the processing.
None of the cases reported having been exposed to or having noticed other known risk factors for tularaemia infection during the hunt (e.g. contaminated dust/water, tick/insect bite).

| Butchery employees
Three butchery employees reported hare contact (e.g. touched, washed, disassembled); one person was not sure. However, all four specified the days of hare contact and the activities they had carried out. No one had worn PPE during any of the processing activities (Table 4). Two butchery employees tested positive (attack rate 50%).

| Family members
The four hare carcasses hung in the slaughter room of a hunter's home overnight to bleed out before providing them to the game butchery. The hunter himself, who was also involved in the processing of hares, confirmed as a case. One of the two family dogs had participated in the hunt, whereas both dogs had licked hare blood in the slaughter room. Two further family members, of whom one had also participated in the hunt as a beater and confirmed as a case (case 12), did not report any direct contact to the hares or hare blood, neither during the hunt nor at home, but reported frequent contact to the dogs. Case 12 developed a conjunctivitis on one eye on November 20, as the only sign of infection. The other family member who did not participate in the hunt tested negative.

| Veterinary staff
As a precaution, ten of the 11 involved dogs were tested. Staff of the veterinary practice was informed about the tularaemia suspicion beforehand. One veterinarian and one veterinary assistant treated the dogs. Both used PPE during all procedures, including facemask, gloves and surgical gown. General examination and blood sample collection was done between November 6 and 14.
The veterinary assistant reported multiple symptoms, starting on November 11. Tularaemia suspicion was not confirmed, and her test revealed a negative result for antibodies against F. tularensis; symptoms could be attributed to another cause.

| Hare samples
The LGL laboratory isolated F. tularensis from an abdominal lymph node of one hare. Results were confirmed by specific PCR at the CL; F. tularensis subspecies holarctica was identified through culture.
Details on laboratory findings and phylogenetic analysis of the F. tularensis strain are described elsewhere (Jacob et al., 2020).

| Dog samples
None of the dogs who had potential contact to the hares or their body fluids showed any symptoms. Culture and PCR testing of all 10 tested dogs revealed negative results. Serological testing provided tularaemia specific IgG antibody detection in four dogs, with no increase in paired samples within eight days.

| D ISCUSS I ON
Comprising 12  All confirmed cases reported having had symptoms. The majority showed non-specific, influenza-like symptoms. Approximately

15% of tularaemia cases notified in Germany between 2002 and
2016 showed non-specific symptoms, for example fever (Faber et al., 2018). This difference to our observation may be both due to younger age of affected persons compared to other findings (Hauri et al., 2010;Mailles & Vaillant, 2014) and due to the early communicated suspicion of tularaemia infection shortly after symptom onset.
Thus, antibiotics specifically recommended for treatment of tularaemia infections and specific testing were immediately initiated. In  (Hauri et al., 2010). Infectious aerosols generated by washing contaminated products were also responsible for tularaemia outbreaks in sugar beet factories (Puntigam, 1960). In our cohort study, for one of the two cases with no direct contact to hares, no alternative explanation is available for infection, other than through inhalation.
The other case was a hunter's family member, whose property was used to hang hares to bleed out. Although this family member also took part in the hunt as a beater, the case had no direct hare contact, but reported close contact to the two family dogs, which both had licked hare blood and tested positive for IgG antibodies, although the results were indicative of a past infection.

ACK N OWLED G EM ENTS
We would like to thank all participants of this outbreak investigation for their contribution to these findings. Moreover, we thank

CO N FLI C T S O F I NTE R E S T
None.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared.