Risk Factors for and Seroprevalence of Tickborne Zoonotic Diseases among Livestock Owners, Kazakhstan

Crimean-Congo hemorrhagic fever (CCHF), Q fever, and Lyme disease are endemic to southern Kazakhstan, but population-based serosurveys are lacking. We assessed risk factors and seroprevalence of these zoonoses and conducted surveys for CCHF-related knowledge, attitudes, and practices in the Zhambyl region of Kazakhstan. Weighted seroprevalence for CCHF among all participants was 1.2%, increasing to 3.4% in villages with a known history of CCHF circulation. Weighted seroprevalence was 2.4% for Lyme disease and 1.3% for Q fever. We found evidence of CCHF virus circulation in areas not known to harbor the virus. We noted that activities that put persons at high risk for zoonotic or tickborne disease also were risk factors for seropositivity. However, recognition of the role of livestock in disease transmission and use of personal protective equipment when performing high-risk activities were low among participants.

particularly Germany, Austria, and Slovenia (10,11). Lyme disease is the sixth most commonly reported notifiable infectious disease in the United States (https://www.cdc.gov/lyme). In addition, incidence of Lyme disease and the range of tick vectors have been increasing in Europe and Asia (10,11), where Lyme disease is found in western Russia, Mongolia, northeast China, and Japan.
CCHF, Q fever, and Lyme disease are endemic to the southern Kazakhstan region of Zhambyl, but their true burden is largely unknown because few serologic surveys have been conducted in Kazakhstan and central Asia. The Zhambyl region covers >55,000 km 2 and has a population of ≈1.2 million. The region is characterized by diverse ecology, containing both desert steppes and mountainous pastures, and elevations of 213-4,115 m. The region has 363 villages and 4 cities. Livelihoods are largely pastoral or agricultural, and common occupations involve a high degree of animal contact, placing humans at increased risk for zoonotic infections.
Among the 3 diseases, only CCHF is a reportable disease in Kazakhstan. During 2000-2013, the Zhambyl region had 73 reported human CCHF cases, the second highest case-count among regions in Kazakhstan (12). However, data on human prevalence of CCHF in Kazakhstan are limited to reported clinical cases, even though studies show >80% of infections are subclinical (13). Although Q fever was detected in Kazakhstan in the 1950s, the lack of surveillance or serologic studies obscure our understanding of Q fever or Lyme disease incidence in the population (14). Quantifying seroprevalence of these diseases in humans can help identify areas of pathogen circulation and areas where humans could be infected.
For this study, we aimed to determine the seroprevalence of antibodies against CCHFV, C. burnetii, and B. burgdorferi in humans who interact with livestock in the Zhambyl region. In addition, we sought to assess the population's knowledge of risk factors for disease transmission and how frequently they engage in activities that increase or reduce risk for infection.

Methods
In June 2017, we conducted a knowledge, attitudes, and practices and risk factor survey (KAP/risk factor survey), along with serosurveys for CCHF, Q fever, and Lyme disease, in 30 rural villages in the Zhambyl region. Participants could enroll in the KAP/risk factor survey, the serosurvey, or both. Eligible participants were >18 years of age, residents of the village for >2 months, and residents of a household containing a sheep or cow of >1 year of age.

Sample Size
Sample selection was based on concern about CCHF as a nationally reportable disease. We conducted surveys in households in which sheep and cattle serosurveys simultaneously were conducted; sample size calculations were based on expected seroprevalence of sheep and cattle. We calculated a target sample size of 561 households with sheep and 473 households with cattle. We based the sample size on an α of 0.05, power of 80%, a design effect of 2, and an expected response rate of 90%. We assumed CCHF seroprevalence of 24% in sheep and 19% in cattle, on the basis of a meta-analysis of previous serosurveys (15).

Participant Selection
We stratified the 363 villages in the region by known (CCHF-endemic) or unknown (non-CCHF-endemic) recent circulation of CCHF. We defined recent circulation as 1 confirmed human case reported in hospital-based surveillance from Zhambyl Oblast Health Department (Taraz, Kazakhstan) or 1 CCHF-positive tick confirmed in the previous 5 years and reported in annual tick surveillance data from the Ministry of Agriculture of Kazakhstan. We identified 66 (18.2%) villages that met the definition for having known, recent CCHF circulation.
We selected 15 villages from each stratum; probability of selection was proportional to the number of sheep and cattle in the village. We obtained livestock counts from reports by village veterinarians to the Ministry of Agriculture. Elevation of the 30 villages was 220-2,590 m (mean 781 m, median 488 m). Mean elevation was 513 m for villages with known CCHF circulation and 1,049 m for villages without known circulation.
Local veterinarians provided information on livestock-owning households in each village. To verify, data collectors conducted a census of 5 villages and mapped households containing sheep or cattle. The veterinarian registry was accurate except for 2 instances in which the household recently had sold animals. Survey teams randomly selected 35 households from these registries and 1 adult per household for study participation.

KAP/Risk Factor Survey
We adapted our KAP/risk factor survey from one conducted in Georgia during a 2014 CCHF outbreak (16). We translated the survey into Russian and Kazakh, the 2 most common languages in the region. Survey teams pilot tested the questionnaire in an eligible village not selected for the study.
The survey team administered the questionnaire verbally at each respondent's residence. Survey questions covered demographics; occupations; history of animal and tick interactions; illness in the previous 4 months or fever and hemorrhaging; and knowledge of CCHF transmission routes, symptoms, and risk factors. The survey did not contain questions specific for Lyme disease or Q fever.

Serosurvey
After answering the questionnaire, respondents were asked to go to their local health clinic to provide a blood sample on the same day. Each village had a health clinic within walking distance of participants. Nurses drew 5 mL of blood from each participant and stored it in a serum separator tube. Blood samples were kept on ice, centrifuged within 6 h, and transported within 24 h to the Zhambyl Regional Laboratory for Especially Dangerous Pathogens in Taraz, where laboratorians aliquoted serum into 4 samples/ participant and stored serum at −20°C until analysis.
Laboratorians analyzed samples for evidence of recent CCHF exposure, indicated by presence of IgM, by using VectoCrimean-CCHF-IgM Kits (Vector-Best, https://vector-best.ru) and for evidence of past CCHF exposure, indicated by IgG, by using VectoCrimean-CHF-IgG Kits (Vector-Best). Laboratorians assessed past exposure to C. burnetii, indicated by presence of IgG, using ELISA-Anti-Q Kit No. 1 (Pasteur Institute of Epidemiology and Microbiology, http://www.pasteur-nii.spb.ru), and exposure to Borellia spp., indicated by presence of IgG against B. afzelii, B. garinii, or B. burgdorferi, by using LymeBest-IgG Test Kits (Vector-Best). All testing was performed with commercially available ELISA kits, according to manufacturer instructions (17,18).

Data Analysis
We analyzed data by using R version 3.4.3 (19). We weighted results for each participant by calculating the inverse probability of selection and applying a poststratification adjustment to each stratum to account for nonresponses. We stratified KAP/risk factor answers specific to CCHF according to whether the health department recognized the village as having known, recent history of CCHF. We used χ 2 test in bivariate analysis to compare frequencies between these 2 strata. We used logistic regression models to test associations between risk factors and seropositivity. We defined risk for zoonotic or tickborne disease as participation in >1 of the following activities: handling ticks with bare hands; working with livestock; working in a healthcare setting; being a veterinarian; or herding, birthing, shearing, slaughtering, or milking animals.

Ethics Review
Each participant provided written, informed consent. No personal identifying information was collected. The Institutional Review Board in Almaty, Kazakhstan, through the Committee for Public Health Protection, approved the study. The protocol was reviewed according to the US Centers for Disease Control and Prevention human subjects review procedures, which determined the agency was not engaged in the study because the Zhambyl Departments of Health and Agriculture owned and collected the data.

KAP/Risk Factor Survey
We selected 969 households; 948 persons completed surveys, a 98% response rate. Reasons for nonresponse included 4 households that were not visited, 2 that were abandoned, and 1 that was not found. In addition, 12 persons did not consent: 4 did not want to participate in the serosurvey, 1 did not have time, 1 distrusted the data team, and 6 gave no reason. Further, 2 persons were excluded from analysis because information on their village of residence was missing and they could not be analyzed according to survey design.
Of the 17 persons seropositive for CCHFV, median age was 54 years; 58% were male (Table 6). No persons reported previous CCHF diagnosis or illness with fever and hemorrhaging in the previous 5 years or a tick bite or handling ticks with bare hands in the previous 4 months. Occupations among the 17 seropositive persons were farmer or herder (n = 2), healthcare worker (n = 1), office or indoor worker (n = 1), homemaker (n = 5), retired (n = 3), unemployed (n = 4), and other (guard; n = 1).
Of 5 participants with evidence of recent exposure to CCHFV, 4 reported participating in >1 highrisk activity in the previous 4 months: 3 milked animals, 2 helped birth animals, 1 sheared animals, and 1 slaughtered animals. One participant reported experiencing an illness with joint pain in the previous 4 months. Three were from non-CCHF-endemic villages, which could suggest a wider range of virus circulation than previously thought.

Discussion
We conducted a serosurvey to update data on the prevalence of CCHF, Q fever, and Lyme disease in Kazakhstan. Because little is known about the seroprevalence of these diseases in central Asia, this study will increase regional awareness. Cases go undetected because of subclinical infections, nonspecific diagnostic methods, or poor surveillance. Our serosurvey identified persons exposed to these pathogens who might have been missed by existing surveillance platforms.
We found a weighted seroprevalence of 1.2% of CCHF in the study region, comparable to findings from studies in Turkey, Iran, and Bulgaria that reported seroprevalences of 2.3%-2.8% (20)(21)(22). We found a seroprevalence of 3.4% in villages classified as CCHFendemic, similar to findings from studies in Bulgaria, China, Georgia, Kosovo, and Turkey that reported seroprevalences of 3%-4% (16,(22)(23)(24)(25)(26)(27)(28). Most CCHFV serosurveys have been conducted in the Middle East, with a few in Asia, and prevalence estimates range widely, even in the same country.
We found moderate seroprevalence (2.4%) for B. burgdorferi compared with findings for other countries in the region. For instance, a serosurvey in Ukraine found seroprevalences of 25%-38% in a healthy population depending on the ecologic zone (43). However, seroprevalence could be caused by other Borrelia species in that region and might not be specific to the Lyme disease group of Borreliae. In addition, <3 of 35 persons tested in some villages were seropositive (43).
We also found a lower weighted seroprevalence for Q fever (1.3%) than most reports. Our findings more closely approximate the 3.1% seroprevalence reported in the United States (44). However, as with CCHFV, prevalence of past infection varies widely by location. For instance, reports from Turkey demonstrate ≈4% seroprevalence in urban areas but 19% in rural areas (45). As we saw with Lyme disease, some villages in our sample had <3 of 35 participants who tested seropositive. Previous studies identified higher seroprevalence for Q fever in butchers (46,47), and our study showed 30.6% of participants seropositive for Q fever had butchered animals or handled raw meat.
A limitation of our study is that the Lyme disease assay was designed for broader reactivity and was not analytically specific to a single agent. This assay likely also reacts with relapsing fever Borreliae, which  17) reported an analytic sensitivity of 80% for CCHF IgG Vector-Best kits and 88% for the IgM kits, with a specificity of 100% for both, so the true seroprevalence could be underestimated. The same is true for Lyme disease; a study reported a sensitivity of only 68.8% for the Vector-Best Lyme IgG kit (18).
We were surprised by the few reports of tick bites, considering that ≈90% of respondents listed ticks as a major problem and about one third had found ticks on their livestock. A previous survey identified crushing ticks with bare hands as common and a risk factor for CCHF (16). However, most respondents in our study reported crushing ticks with an object, suggesting contact with livestock could be a more common route of exposure among participants. This possibility could be problematic because <20% of respondents identified infected animals as a potential source of transmission. In addition, nearly half did not wear PPE when slaughtering animals, an exceptionally high-risk activity. The low recognition of the role of livestock in CCHF transmission is seen in other regions (48,49), but targeted educational campaigns have improved knowledge of transmission routes (50).
Our results have been translated into direct public health action. For instance, the serosurvey revealed that CCHFV is circulating in areas previously unknown to have CCHF activity. Because such areas were not prioritized for educational activities, knowledge of CCHF and modes of transmission was low compared with areas of known transmission. In addition, whereas the KAP/risk factor survey revealed that most respondents understood the risks posed by ticks and many took precautions against tick bites, most did not understand the role animals play in these zoonoses, nor did they wear proper PPE when performing high-risk activities. We helped the health department clarify their pamphlets to state specific high-risk activities and describe which PPE should be worn during each activity. Formative research into the availability and affordability of PPE, as well as the cultural perceptions of PPE when performing activities that may have ritualistic significance, such as slaughtering, is warranted.
A One Health approach that recognizes the interconnectedness of animal, human, and environmental health is needed for effective zoonotic and vectorborne disease control. This study incorporated personnel from the Kazakhstan Ministry of Health, Zhambyl Oblast Public Health Protection Department, and the Ministry of Agriculture. Additional studies in the region will analyze blood and ticks collected from livestock for evidence of past zoonotic infection. Combining the results of the human serosurvey with results of the animal and tick surveys will permit more in-depth investigations into the role of environmental factors, such as climate and elevation, in the transmission of these pathogens. Ticks transmit a variety of different pathogens, including bacteria, protozoa, and viruses, which can produce serious and even fatal disease in humans and animals. Tens of thousands of cases of tickborne disease are reported each year, including Lyme disease. See the EID Lyme Disease Spotlight. Lyme disease is the most wellknown tickborne disease. However, other tickborne illnesses such as Rocky Mountain spotted fever, tularemia, babesiosis, and ehrlichiosis also contribute to severe morbidity and more mortality each year.
Symptoms of tickborne disease are highly variable, but most include sudden onset of fever, headache, malaise, and sometimes rash. If left untreated, some of these diseases can be rapidly fatal.