Seroprevalence and Risk Factors Associated with Chlamydia abortus Infection in Sheep and Goats in Eastern Saudi Arabia

Chlamydia abortus (C. abortus) is intracellular, Gram-negative bacterium that cause enzootic abortion in sheep and goats. Information on C. abortus seroprevalence and flock management risk factors associated with C. abortus seropositivity in sheep and goats in Saudi Arabia are scarce. The objectives of this study were to (i) estimate the animal, flock, and within-flock seroprevalence of C. abortus among Eastern Province sheep and goat flocks and (ii) identify the flock management and animal risk factors associated with C. abortus seropositivity in Eastern Province, Saudi Arabia. A cross-sectional study with a two-stage sampling process was carried out in the Eastern Province, Saudi Arabia, between 2015 and 2016. A total of 1717 sheep and 1101 goat serum samples were collected from 21 sheep and 14 goat flocks, then were tested for C. abortus antibodies using a commercial ELISA Kit. In addition, vaginal swabs and aborted tissue samples were collected from sheep (n = 48) and goats (n = 15) with recent history of abortion for detection of C. abortus pmp gene using PCR. A questionnaire was constructed to collect information about flock management and animal risk factors possibly associated with C. abortus infection in sheep and goats. The true sheep and goat-level seroprevalences were 11.1% (95% CI: 9.7–12.7) and 10.6% (95% CI: 8.8–12.5), respectively. The true flock-level seroprevalence was 100% for both sheep and goats. However, the average within sheep and goat flocks true seroprevalences were 9.6% (95% CI: 1.8–22.9) and 9.3% (95% CI: 1.8–19.5), respectively. Multivariable logistic regression revealed that introduction of new sheep to the flocks (OR = 2.6; 95% CI: 1.5–4.4), type of breeding system (OR = 1.8; 95% CI: 1.0–3.4), flocks allowing females in (OR = 1.9; 95% CI: 1.1–3.3) or females out (OR = 2.2; 95% CI: 1.1–4.3), and sheep age 1.4–2.8 years (OR = 1.9; 95% CI: 1.3–2.9) were potential risk factors for C. abortus seropositivity in sheep flocks. However, in goat flocks, the introduction of new goats to the flocks (OR: 1.9; 95% CI: 1.2–3.0) was identified as a risk factor, whereas good farm hygiene (OR: 0.3; 95% CI: 0.2–0.7) was identified as a protective factor. C. abortus pmp gene was identified in 45 (93.8%) and 15 (100%) of samples collected from sheep and goats, respectively. These results could be used to implement efficient management measures to prevent and control C. abortus infection in sheep and goats in Eastern Province, Saudi Arabia, but also could be used to reduce the risk of C. abortus infection in sheep and goat flocks with similar management practices in other regions.


Introduction
Chlamydia is an obligate intracellular, Gram-negative bacterium that causes various diseases in animals and humans [1,2]. However, Chlamydia abortus (C. abortus) and Chlamydia pecorum (C. pecorum) are the two Chlamydia species that cause serious infection in sheep and goats [3,4]. C. abortus is one of the main causes of abortion in sheep and goats worldwide [5,6] and usually transmitted through ingestion or inhalation of dust, food, and/or water contaminated with aborted animal uterine discharge, placenta, and fetus [3,7]. C. abortus cause enzootic abortion disease in sheep and goats which characterized clinically by abortion in the last 2-3 weeks of pregnancy [3]. Furthermore, C. abortus infection may cause stillbirth, premature lambing, and delivery of weak lamb/kids that generally die immediately [3,8]. In human, C. abortus infection is considered a zoonotic occupational disease and most of the infections were reported in veterinarians, livestock breeders, butchers, abattoir, and farm workers [3,9]. In humans, common symptoms in men and women are characterized by mild influenza-like illness or urethritis, however in severe cases, pregnant women are also at risk of abortion [10][11][12][13].
Several studies have investigated the seroprevalence of C. abortus in sheep and goats [14][15][16]. However, limited studies have assessed the flock-level risk factors associated with C. abortus seropositivity in sheep and goat flocks, including flock size, reproductive practices, husbandry, and farm sanitation [17][18][19]. Previous studies have reported that the real drivers of C. abortus infection in sheep [16,17], goats [19,20], camels [21], and cattle [22] are intrinsic farm factors such as production system and management practices, but a clearer understanding of these factors is needed. Knowledge of flock management practices and factors associated with the risk of C. abortus infection would prevent transmission and improve control strategies. In addition to the quality and effectiveness of the C. abortus diagnostic tests, C. abortus diagnosis requires either detection of bacteria/antigen by microscopic, immunofluorescence, ELISA, detection of the bacterial DNA by PCR and microarray [23,24], or detection of antibodies against C. abortus by ELISA [24,25].
In Saudi Arabia, few seroprevalence surveys have been conducted across different ruminants, including camels [26,27], sheep, and goats [28]. However, data on farm management risk factors for C. abortus infection in sheep and goats in Saudi Arabia is limited. Determination of C. abortus seroprevalence and risk factors associated with infection could help in the control of C. abortus in both animals and humans in Saudi Arabia. Therefore, the objectives of the present study were to (i) estimate the animal, flock, and within-flock seroprevalence of C. abortus among sheep and goat flocks in Eastern Province and (ii) identify the flock management and animal risk factors associated with C. abortus seroprevalence among sheep and goat flocks in Eastern Province, Saudi Arabia.

Risk Factor Analysis
The flock and animal risk factors associated with sheep and goats seropositivity in a univariable analysis at P < 0.20 are presented in Tables 1 and 2, respectively. In sheep flocks, results showed a significant positive association (risk factor) between C. abortus seropositivity and larger flock size (P = 0.075), season (P = 0.003), recent history of abortion (P = 0.000), the introduction of new sheep to flock (P = 0.000), history of previous treatment (P = 0.001), type of breeding system (P = 0.000), sheep exchange during breeding (P = 0.007), and sheep age (P = 0.003). Whereas, farm hygiene (P = 0.002) was negatively associated (protective factors) with C. abortus seropositivity (Table 1). Similar results were found in goat flocks, except recent history of abortion (P = 0.969), the history of previous treatment (P = 0.654), and type of breeding system (P = 0.855) were not significantly associated with C. abortus seropositivity ( Table 2).   Table 3 showed flock and animal level factors associated with sheep and goats seropositivity in a multivariable analysis at P < 0.05. In both sheep and goat populations, the flock size variable could be a potential confounder based on the change in the β-coefficient (>20%). Flock size was kept as a continuous variable in the sheep model, given that assumption of linearity was respected. However, this variable was kept as a categorized variable in the goat model, given linearity was not respected.   For goat flocks, two factors were retained as being significantly associated with C. abortus seropositivity in the final multivariable logistic regression model. The odds of C. abortus seropositivity were higher for flocks introducing/purchased new goats (OR = 1.9; 95% CI: 1.2-3.0) than flocks that did not introduce new animals to flocks. However, results showed that flocks with bad hygienic measures were three times (OR = 3.0; 95% CI: 1.5-6.0) more likely to be seropositive for C. abortus compared to flocks with good hygienic measures. The Pearson's chi-square (P = 0.4) and the Hosmer-Lemeshow (P = 0.9) fit statistics suggested a reasonable fit of the model.

Seroprevalence of C. abortus
In this study, the true seroprevalence of C. abortus antibodies at sheep-level was 11.1%, which was higher than the 7.5% reported previously in sheep in Riyadh, Saudi Arabia; similar to the 10.6% reported in sheep in India; but lower than the 40.1% and 24.5% reported in sheep in China and Algeria, respectively. The true seroprevalence at goat-level in this study was 10.4%, which was similar to the 11.4% reported in Jordan and lower than the 34.5% reported in Saudi Arabia and 33% reported in goats in Spain [19].
Our results revealed that 100% of the sheep and goat flocks had at least one C. abortus seropositive. This flock-level seroprevalence was similar to the 100% reported in both sheep and goat flocks in Jordan but was higher than the 70.4% of sheep flocks in Algeria and 78.6% of goat flocks in China [29]. The average within sheep and goat flock seroprevalences in this study were 9.6% (ranging: 1.8 to 22.9) and 9.3% (ranging: 1.8 to 19.5), respectively, indicating the spread of C. abortus infection within sheep and goat flocks in Eastern Saudi Arabia. This result was similar to the range of within flock seroprevalence 3.7% to 25.0% reported in sheep in Costa Rica [30] and the 0.0% to 29.9% reported in goats in China [29]. Such variation in C. abortus seroprevalence is likely due to differences in sheep and goat breeds, husbandry practices, sanitation, time of sampling, and serological test used. In addition, other reasons for the variations of seroprevalence may be related to differences in climate conditions, including temperature, rainfall, and altitude [31]. However, this study showed no significant differences in the seroprevalence of C. abortus antibodies between sheep and goats at the animal-and flock-level. This may attribute to the pasture restriction in the study area, which allows sheep and goat flocks to graze in a common pasture. Previous studies have reported that grazing on contaminated pasture is a source of C. abortus transmission among animals [6,32].
In Saudi Arabia, vaccination against chlamydiosis is not applicable. Thus, the relatively high seroprevalence of C. abortus in the current study is suspected and could be explained by the fact that the study area (Eastern Province) has borders with five C. abortus endemic countries [33][34][35]. Furthermore, sheep and goat flocks are seminomadic and moving for long distances every day during the grazing season, which might contribute to the horizontal spread of C. abortus to a large number of flocks.

Risk Factors
Various risk factors were assessed to be associated with the seroprevalence of C. abortus in sheep and goat flocks in Saudi Arabia. In the present study, the highest sheep flock (22.9%) and goat flock (19.5%) seroprevalence was reported in large size flocks (>200), which was confirmed by the univariable analysis. This finding is consistent with previous studies from Tunisia and Spain, which indicated that larger size flocks are associated with higher C. abortus seropositivity compared to smaller size flocks. In contrast, other studies from Iran, Algeria, and Brazil [17] have reported no association between flock size and C. abortus seropositivity. The possible explanation that sheep/goats overcrowding in large size flocks may influence animal welfare and hygienic measures, which increase the risk of C. abortus transmission [4,6]. In addition, the association between a higher C. abortus seropositivity and the increased flock size may also be related to the larger number of visitors to the farms by veterinarians, feed suppliers, and farm workers [14,36]. In this study, larger size flocks were correlated with introduction of new animals to the flock (0.77), open breeding system (0.78), and lower hygienic scores (−0.38).
Climates play a role in C. abortus infection [31]. The warm and humid conditions are favorable for the survival of Chlamydia [37]. In the present study, the seroprevalence of C. abortus in winter was significantly higher in sheep and goat flocks than in summer. This result may be due to the climate of Eastern Province is desert which mainly dried hot in summer and mild humid in winter. Similarly, previous studies have reported that season/climate is a risk factor for C. abortus infection in sheep [38] and goat [29] flocks in China.
Several studies have pointed out that the seroprevalence of C. abortus in small ruminants varies based on the management system operated on the flock [19,39,40]. Therefore, we could assume that the flock management practices in Saudi Arabia play a significant role in the C. abortus seroprevalence reported in the present study. The univariable risk factor analysis of the flock management revealed several factors to be significantly associated with C. abortus seroprevalence. In sheep flocks, results revealed that flocks with a recent history of abortion, history of treatment, introduced new sheep to the flock, exchanged rams during breeding were identified as risk factors for C. abortus infection. However, in goat flocks, only the goat age, introduction of new goats to flock, and goat exchange during breeding were identified as risk factors for C. abortus infection. Furthermore, it was revealed that the seroprevalence of C. abortus in the present study significantly decreases in sheep and goat flocks, where good hygienic measures were applied. Similar results were previously reported in sheep flocks in China [38], Egypt, Algeria, Brazil, Costa Rican, and goat flocks in China [29]. In contrast, a recent study in Tunisian sheep flocks has reported no association between seroprevalence of C. abortus and the number of ewes with abortion history, the exchange of breeding male, and the frequency of disinfection [15].
The risk of sheep and goat flocks acquiring C. abortus infection seems to be complex. The multivariable analysis in the present study confirmed that four different variables were associated with C. abortus seropositivity in sheep flocks. The multivariable model showed that flocks using open breeding system, adding new sheep and exchanging sheep during breeding are at higher risk of C. abortus seropositivity. Similar to our results, Mamlouk et al. reported that the absence of previous control of sheep, when introduced to the flock, is a risk for C. abortus seropositivity. Furthermore, Barkallah et al. observed that quarantine had a significant relation with C. abortus infection in sheep. The observed high seroprevalence of C. abortus in older sheep in this study was in agreement with previous studies, which reported an association between the increase of C. abortus seroprevalence and sheep age [41,42]. This result may be due to earlier infection or the high chance of exposure to the source of infection with age increase [38].
In goat flocks, results showed a significant relationship between C. abortus seropositivity and the introduction of new goats to the flock. Some farms purchased new goats without pre-testing and history taking, leading to an increased risk of infection. This result is consistent with Santos et al. [17], who reported that sharing goats for breeding between farms contributed to the spread of C. abortus infection from one flock to another. On the other hand, the risk of C. abortus seropositivity in goat flocks with good hygienic measures is lower than flocks with bad hygienic measures, which contrasts with Santos et al., who reported no association between use of disinfectant and risk of C. abortus seropositivity.

Study Limitations
The ELISA test was used in the present study to assess the seroprevalence of C. abortus antibodies in sheep and goat serum samples. Although the sensitivity and specificity of the used ELISA are high but indeed not 100%, so some misclassification errors may therefore occur, which may add some noise to the observed relationship between dependent and independent variables. Furthermore, the antigenic cross-reactivity between C. abortus and C. pecorum in sheep may result in an overestimation of seroprevalence [43,44]. Because of the relatively small number of goat flocks included in the study, only variables that were strongly associated with C. abortus could be evaluated in the final model. The absence of a particular variable from the final model may be due to the limited sample size. However, the identification of at least one PCR positive animal in each flock decreases the impact of these limitations on this study results.

Study Area
Eastern Province (22 • 30 N, 51 • 00 E) is located in the eastern side of Saudi Arabia, at 390 km from the capital Riyadh and shares the borders with five countries, including Oman, United Arab Emirates, Qatar, Kuwait, and Iraq ( Figure 3). The Eastern Province is the third most populated province in Saudi Arabia, with diverse climatic conditions ranging from semi-desert to desert. According to the General Authority for Statistics for 2015, sheep and goat populations in Eastern Province represent 13% and 2.5% of the total sheep and goat populations in the kingdom.
The production system of small ruminants is mainly of the seminomadic type as sheep and goats move out from their raising farms for grazing during early March until August when pastures are mostly available. During the fall and winter months, sheep and goats are housed and fed on concentrates. Sheep and goats in Saudi Arabia are mainly raised for meat production.

Study Design
A cross-sectional study with a two-stage sampling process was carried during 2015-2016. The total number of sheep and goat flocks to be sampled was determined using the 'epi.ssclus2estb' function of the 'epiR' package with 50% an expected seroprevalence, 5% absolute precision, and 95% confidence interval (CI). The required number of flocks was determined to be 12 flocks. For the sheep population, the number of flocks was increased to 21, and for the goat population increased to 14 flocks to improve the precision of the estimated seroprevalence in sheep and goats [45]. Selected sheep flocks size ranged from 40 to 800 (median = 350) sheep per flock. However, selected goat flocks size ranged from 40 to 500 (median = 290) goats per flock. The number of sheep and goats sampled per flock was calculated according to Cannon and Roe, using an expected seroprevalence of 8% with a 95% CI to detect at least one seropositive. Eastern Province (22°30′ N, 51°00′ E) is located in the eastern side of Saudi Arabia, at 390 km from the capital Riyadh and shares the borders with five countries, including Oman, United Arab Emirates, Qatar, Kuwait, and Iraq ( Figure 3). The Eastern Province is the third most populated province in Saudi Arabia, with diverse climatic conditions ranging from semi-desert to desert. According to the General Authority for Statistics for 2015, sheep and goat populations in Eastern Province represent 13% and 2.5% of the total sheep and goat populations in the kingdom. The production system of small ruminants is mainly of the seminomadic type as sheep and goats move out from their raising farms for grazing during early March until August when pastures are mostly available. During the fall and winter months, sheep and goats are housed and fed on concentrates. Sheep and goats in Saudi Arabia are mainly raised for meat production.

Study Design
A cross-sectional study with a two-stage sampling process was carried during 2015-2016. The total number of sheep and goat flocks to be sampled was determined using the 'epi.ssclus2estb' function of the 'epiR' package with 50% an expected seroprevalence, 5% absolute precision, and 95% confidence interval (CI). The required number of flocks was determined to be 12 flocks. For the sheep population, the number of flocks was increased to 21, and for the goat population increased to 14 flocks to improve the precision of the estimated seroprevalence in sheep and goats [45]. Selected sheep flocks size ranged from

Sample and Data Collection
A two-stage random sampling process was carried out. Flocks were selected in the first stage and animals within flocks were selected in the second stage. In total, blood samples were collected from 1717 sheep in 21 flocks and 1101 goats in 14 flocks, with an average of 80 animals sampled per flock. In addition, vaginal swabs and aborted tissue samples were collected from a number of sheep (n = 48) and goats (n = 15) with recent history of abortion. All samples were labeled with flock and animal ID and sampling dates and then transported to the laboratory for serological analysis.
A questionnaire was constructed to collect information about flock management and animal risk factors possibly associated with C. abortus infection in sheep and goats. The questionnaire consisted of 11 closed and open-ended questions written in Arabic and available from the corresponding author upon request. The questionnaire was completed by face-to-face interviews with the flock owner on the day of sampling. Questions covered the flock characteristics and management practices, including flock size (small (≤200 animal/flock) and large (>200 animal/flock)), the season of sampling, the recent history of abortion (Yes and No), the introduction of new animals to flock (Yes and No), mixed breeding (Yes and No), farm hygiene (Bad and Good), history of previous treatment (Yes and No), type of breeding system (Open and Closed), animal exchange during breeding (No, Female in and out), and type of the used vaccine. In addition to the animal age, sex, and breed, this information is presented in Supplementary Table S1.

Serological Examination
Serum samples were tested for antibodies against C. abortus infection using commercial indirect ELISA kits (IDEXX Chlamydiosis Total Ab Test, IDEXX Laboratories, Broomfield, CO, USA) according to manufacture instructions. Briefly, serum samples and controls were diluted at 1:400 and then tested in duplicates. Microplate ELISA reader measured the optical densities (OD) at 450 nm. The average OD was estimated for duplicate samples and controls, then used to calculate the OD% by the following equation: OD (%) = (OD sample − OD negative control )/(OD positive control − OD negative control ) × 100. (1) Any sample with OD% >40% was considered seropositive; if the OD% was between ≥30 and ≤40, the result was considered doubtful, while any sample with an OD% <30% was classified as seronegative. Regarding the diagnostic performance of the IDEXX ELISA's kites used in this study, Wilson et al. [46] evaluated the accuracy of ELISA in the absence of a gold standard and reported a 98.1% sensitivity (Se) and 100% specificity (Sp) in sheep and goats.

Molecular Identification of C. abortus
DNA was extracted from vaginal swabs and aborted tissue samples using the Qiagen QIAamp DNA mini kit (Qiagen, Courtaboeuf, France) according to the manufacturer's instructions. Purified DNA was amplified by PCR for detection of the pmp gene of C. abortus using the primer (CpsiA (5 -ATGAAACATCCAGTCTACTGG-3 ) and CpsiB (5 -TTGTGTAGTAATATTATCAAA-3 )) previously described by Greco et al. [47]. Briefly, 2 µL sample of each purified genomic DNAs was amplified in 20 µL of the final volume of a 2X HotStartTaq Plus Master Mix (QIAGEN, Germantown, MD, USA) containing 1.5 mM MgCl2, 200 µM of each dNTP,1-unit HotStartTaq Plus DNA polymerase, and 10 µM of each forward and reverse primers. The PCR assay was performed using DNA thermal cycler I (BIORAD, CA, USA) according to Greco et al. [47]. The amplified PCR products were electrophoresed in 1.5% agarose gel stained with ethidium bromide and documented using an ultraviolet gel documentation system (BIO-RAD, Hercules, CA, USA).

Statistical Analysis
Epidemiological data and results of the serological analysis were introduced into Stata Statistical Software v. 15 (Stata Corp, College Station, TX, USA) and R software (R Core Team, 2019; version 3.5.3) for descriptive and statistical data analysis with results considered significant at P-value < 0.05. The apparent seroprevalence of C. abortus at the individual animal-level was estimated as the ratio of seropositive sheep/goat to the total number of sheep/goats examined. However, the true seroprevalence of C. abortus at the animal-level was estimated according to Thrusfield [48] using the 'epi.prv' function from the 'epiR' package in R [49]. The flock-level seroprevalence was estimated from the ratio of positive flocks to the total number of flocks tested. Flocks that contain at least one seropositive sheep/goat were considered positive. The flock seroprevalence was adjusted for Se and Sp of ELISA test used in the study to obtain the true flock seroprevalence using the following formula [50]: where THP and AHP are the true and apparent flock seroprevalence, respectively. HSe and HSp refer to the flock-test sensitivity and specificity, respectively, based on animal-test results, calculated using 'epi.herdtest' function from Package 'epiR'. Analysis of the risk factors potentially associated with C. abortus seropositivity was evaluated in two steps, using univariable and multivariable logistic regression models. Firstly, univariable screening of all risk factors for association with C. abortus seropositivity at liberal P-value < 0.20 using unconditional associations. Significant variables in the univariable analysis were checked for collinearity using Spearman correlation coefficients and were considered collinear when coefficient >0.8 [51]. Retained variables were used to conduct multivariable analysis, and non-significant variables were removed sequentially using backward elimination at P-value < 0.05. The fit of the final model was evaluated using Hosmer-Lemeshow goodness-of-fit statistics [52].

Conclusions
The present study revealed that C. abortus seroprevalence in sheep (11.1%) and goats (10.4%) is relatively high in Eastern Province, Saudi Arabia. Consequently, some management practices must be implemented, especially the risk factors identified in this work, such as introducing new animals to the flock and improving farm hygienic measures. Since C. abortus is a zoonotic disease that can cause health and economic loss in humans and livestock, thus the use of a chlamydial vaccine is strongly recommended. In addition to the implementation of integrated control and efficient management measures to prevent and control C. abortus infection in sheep and goats in Eastern Province, Saudi Arabia, it also could be used to reduce the risk of C. abortus infection in sheep and goat flocks with similar management practices in other regions. Institutional Review Board Statement: The Taif University Ethics Committee has approved the study protocol (TURSP-2020-57).

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.