Prevalence of Mycobacterium bovis in deer in mainland China: a systematic review and meta-analysis

Background Deer tuberculosis is a chronic zoonotic infectious disease, despite the existence of socio-economic and zoonotic risk factors, but at present, there has been no systematic review of deer tuberculosis prevalence in mainland China. The aim of this meta-analysis was to estimate the overall prevalence of deer TB in mainland China and to assess possible associations between potential risk factors and the prevalence of deer tuberculosis. Methodology This study was searched in six databases in Chinese and English, respectively (1981 to December 2023). Four authors independently reviewed the titles and abstracts of all retrieved articles to establish the inclusion exclusion criteria. Using the meta-analysis package estimated the combined effects. Cochran’s Q-statistic was used to analyze heterogeneity. Funnel plots (symmetry) and used the Egger’s test identifying publication bias. Trim-and-fill analysis methods were used for validation and sensitivity analysis. we also performed subgroup and meta-regression analyses. Results In this study, we obtained 4,400 studies, 20 cross-sectional studies were screened and conducted a systematic review and meta-analysis. Results show: The overall prevalence of tuberculosis in deer in mainland China was 16.1% (95% confidence interval (CI):10.5 24.6; (Deer tuberculosis infected 5,367 out of 22,215 deer in mainland China) 5,367/22215; 1981 to 2023). The prevalence in Central China was the highest 17.5% (95% CI:14.0–21.9; 63/362), and among provinces, the prevalence in Heilongjiang was the highest at 26.5% (95% CI:13.2–53.0; 1557/4291). Elaphurus davidianus was the most commonly infected species, with a prevalence of 35.3% (95% CI:18.5–67.2; 6/17). We also assessed the association between geographic risk factors and the incidence of deer tuberculosis. Conclusion Deer tuberculosis is still present in some areas of China. Assessing the association between risk factors and the prevalence of deer tuberculosis showed that reasonable and scientific-based breeding methods, a suitable breeding environment, and rapid and accurate detection methods could effectively reduce the prevalence of deer tuberculosis. In addition, in the management and operation of the breeding base, improving the scientific feed nutrition standards and establishing comprehensive standards for disease prevention, immunization, quarantine, treatment, and disinfection according to the breeding varieties and scale, are suggested as ways to reduce the prevalence of deer tuberculosis.


Introduction
Deer tuberculosis (Deer TB) is a chronic bacterial disease caused by Mycobacterium bovis (1), which is mainly manifested in the formation of tuberculous nodular granulomas comprising necrotic foci and abscesses (2).Deer tuberculosis is an important zoonotic infectious disease, which not only affects deer breeding, but also threatens public health security (3).According to the global tuberculosis report released by the WHO in 2020, Mycobacterium bovis is the most common cause of bovine tuberculosis and zoonotic tuberculosis worldwide (4).At present, the transmission of bovine TB among deer mainly occurs by indirect oral transmission between cattle and deer (5); host transmission systems of common cattle-deer-wild boar in Europe (6); a rat-deer transmission system in New Zealand; and kinship transmission between individuals in the same population (7).Domesticated deer are important economic animals in several regions, including Asia (5).In recent years, with the increase in deer breeding density and the frequent domestic and foreign trade, once bovine TB occurs, it will spread widely, posing a threat to the health of deer and humans (3).At the same time, the homogeneous distribution of TB-infected wild deer and TB cattle has deprived many European countries (e.g., Italy, Portugal, and Spain) of the opportunity to obtain official tuberculosis-free (OTF) status (6).Many developed countries have implemented effective control and eradication strategies, and strict food safety standards for zoonotic TB caused by M. bovis, while in developing countries and economically less developed regions, zoonotic TB caused by M. bovis might have higher actual infection rates and cause more severe economic losses in human and animal populations than the available data suggests (8).To date, China has introduced a series of policies and assistance funds to encourage and support intensive farming, at the same time, the awareness of farmers regarding the prevention and control of deer TB was strengthened, and disease eradication programs were implemented on farms (9).Some articles have reported the epidemic situation of deer TB in certain areas of China; however, these studies are fragmentary and cannot explain the epidemic status and influencing factors of deer TB in mainland China.In addition, as far as we know, there is no national summary report on the prevalence of deer TB in mainland China, and no article on the systematic evaluation of deer TB has been published.Therefore, in this study, we summarized cross-sectional studies on deer TB in mainland China from 1981 to December, 2023 and conducted a systematic review and meta-analysis to estimate the overall prevalence of deer TB in mainland China during this period.We also analyzed various risk factors that affect deer TB prevalence (climate, altitude, annual average sunshine radiation, annual average precipitation, annual temperature, average temperature, geographical distribution, age, species, and detection techniques).The data are intended to assist in assessing the prevalence of deer TB, and to assess the potential risk factors for M. bovis infection in deer in mainland China.Our analysis also provides a basis for developing future rational disease control strategies and the accurate assessment of the association between the economy, health, and disease prevalence.
2 Materials and methods

Article retrieval strategy
The study was conducted according to the PRISMA (2020) Checklist item (Supplementary Table S1) (10).The languages were restricted to English and Chinese.We searched PubMed, ScienceDirect, Web of Science, Chinese Web of Knowledge (CNKI), WanFang, and Chongqing VIP databases for studies reporting bovine tuberculosis (bTB) infection in deer in mainland China from1981 to December 25, 2023.
We established retrieval formulas based on the retrieval patterns of six databases, and to obtain more comprehensive research data, we also conducted supplementary searches (Supplemental searches increase the number of duplicates, but our review of all included articles began with duplicate exclusion) (Supplementary Table S3), with no time limit for the publication in the searched journals, and included synonym extensions in search processes.When we conduct a full-text review of articles that met the inclusion criteria, we also reviewed each reference one by one to identify other studies that were not found during the database search process.We did not attempt to verify unpublished reports.

Literature screening
Endnote (Clarivate, London, UK) was used for the statistical collation of the retrieved articles.Four authors (DNL, XYL, YHS and YXJ) independently reviewed all retrieved articles, firstly by comparing titles, author information and abstracts to exclude duplicate studies; secondly by reviewing titles and abstracts to exclude articles that did not meet the criteria for this study (e.g., studies of tuberculosis in non-Chinese mainland deer, reports of conferences, case reports, etc.); and lastly by downloading the full text of the included articles and reading the full-text content carefully, and further excluded studies that did not meet the inclusion criteria based on the inclusion and exclusion criteria, and all the studies that met the criteria were cross-sectional studies (Figure 1).

Data extraction
The data were extracted into Microsoft Excel (Microsoft Corp., Redmond, WA, United States).Data from each study were independently extracted by authors DNL XYL and DL.During the data extraction process, any opinions and uncertainties in the data extraction were discussed with the main author of this study (DL), and the data were extracted based on the discussion results.We extracted nine data items from each study (Sampling years, Province (Region), Variety, Detection method, Breeding mode, Season, Gender, Age, Score applied to this study) (30-32).At the same time, we obtained data on altitude, climate, annual rainfall, annual average temperature, and annual maximum sunshine hours based on the sampling time and location of each research data (data source: CMDC).The effect measures extracted for each study were the number of positives, the total sample size and the positivity rate.

Quality assessment
We evaluated the quality of the included studies based on the GRADE method (33).Authors (D-NL, X-YL, and DL) scored each study during independent extraction of data from each study based on whether the study had a clear detection method, whether it had a clear sampling year, whether it had randomized sampling, whether it had a clear method of sample collection, and the presence of four or more risk factors (Supplementary Table S4), which were ultimately used to evaluate the risk of bias of the included studies.According to the scoring criteria, the articles included in the study were awarded 0-5 points (Table 1).A score of 4-5 was considered a high-quality article; 0-1 points indicated a low-quality article (32) (The scoring criteria are only applicable to this meta-analysis and do not serve as a quality assessment of the original study).

Statistical analysis
The extracted data were analyzed using the R program, using the "meta" data package to estimate the model (31,34,35).The positive rates of various studies were subjected to Poisson-Lognormal (PLN) analysis and the combined effects were estimated using a metaanalysis package (Table 2).Cochran's Q-statistic was used to analyze heterogeneity, while Higgin's statistic analyzed the differences in heterogeneity (I 2 > 50% heterogeneity was considered significant) (36,37).By analyzing the forest map, the estimated values included in the study data were summarized and the sources of heterogeneity were shown.Based on the estimated heterogeneity, we selected the random-effects model (RE Model) for overall effect estimation and subgroup analysis.To further determine publication bias or small data volume bias, we analyzed funnel plots (symmetry) and used Egger's test; and used the trim and fill analysis method for verification and sensitivity analysis.To further analyze the potential sources of heterogeneity, we conducted subgroup and meta regression analysis.We analyzed and evaluated subgroups such as regional location, province, sampling time, positive diagnosis method, sampling season, feeding mode, and age in the research data (Table 3).To further analyze and evaluate the potential sources of heterogeneity, we also conducted an analysis and evaluation of geographical factor subgroups, including longitude and latitude, and annual average temperature (Table 4).The code used for R program statistical analysis is shown in Supplementary Table S5.

Publication bias and sensitivity analysis
We performed PLN conversion on the positive rate to ensure that the combined effect size data was closer to a normal distribution (Table 2), and the results showed high heterogeneity in the included studies (I 2 = 100%, p = 0; Figure 2).A funnel plot can qualitatively identify publication bias, and the asymmetry of the scatter distribution indicated the existence of publication bias and/or small sample size bias in the study (Figure 3).The Egger linear regression method further validated the existence of publication bias (p < 0.001; Figure 4; Supplementary Table S6).To confirm the reliability of the results, sensitivity analysis showed that the results remained unchanged when any study was removed from the analysis (Figure 5).This further proved that the meta-analysis was reliable.The trim and filling analysis showed that after the scatter distribution was symmetrical, the estimated value of the overall effect size did not change significantly, thus the effect of publication bias on the results was not obvious (Figure 6).

Discussion
The spread of deer tuberculosis has seriously restricted the development of deer industry.To understand, prevent and control the spread and prevalence of deer tuberculosis, it is essential to identify and control risk factors.The meta-analysis of deer tuberculosis showed that the average prevalence of tuberculosis in deer in mainland China between 1981 and 2023 was 16.1% (95% CI:10.5-24.6;5367/22215), which was lower than that in free-range deer herds in southern Spain (18.5%, 2008) (38, 39), and higher than that in New Zealand (0.18%) (40), Italy (0.19%) (41), and Minnesota, United States (0.25%) (42).This reflects the fact that the control of deer TB in mainland China remains a huge challenge.There is no official vaccine to control deer TB; therefore, countries around the world mainly use the "quarantinecull" approach to control TB in other animals, which is difficult to achieve for wild deer herds (43).Therefore, many developed countries have introduced policies related to deer TB, and in countries such as Australia and the United States, herding measures and targeted programs have been mainly implemented to control deer TB (44).New Zealand has introduced a program for the diagnosis and slaughter of TB-infected deer, comprising laboratory testing of suspect animals, necropsy of slaughtered animals, and ongoing surveillance of all carriers, under which all infected animals can be dynamically controlled, which reduced the prevalence of approximately 1.5% in 2004 to less than 0.2% by 2012, the prevalence of deer TB was further reduced to 0.18% by the end of 2013 (40,45).In 2012, China issued the "Standard for Prevention and Control of Tuberculosis and Brucellosis in Deer, " which stipulates the management requirements, quarantine, immunization, disease reporting, and treatment of sick deer and pens to prevent the transmission of tuberculosis and brucellosis in deer (46).Considering the above successful control experience of deer TB in the United States and Canada, China could also consider long-term continuous surveillance of suspected infected deer and pay special attention to their isolation from cattle.
In the subgroup of breeding model, the prevalence was significantly higher in free range deer (25.9, 95% CI:4.7-100.0;73/170) than in captive deer (15.3, 95% CI:10.0-23.4;5294/22045), which is caused by many factors, such as population size, transmission by other susceptible animals in the vicinity, and the environment and morphological structure of the stocking area (47,48).Observations of infected free-range deer in New Zealand showed a significant reduction in the prevalence of deer TB following control of TB-infected possums, suggesting that deer might be a frequent spillover host for possum infection (48).In Switzerland, the disease has been eradicated from free-range deer because of the successful reduction in the prevalence of TB in cattle (49).Ingestion of contaminated feed or water might play an important role in transmission of M. bovis among animals.Indeed, M. bovis infections frequently occur via the oral route (50).Therefore, free-range deer exposed to an M. bovis-contaminated environment for a longer period are more likely to come into contact with infected possums or cattle, with a greater chance of infection.To reduce the spread of deer TB in concentrated breeding areas, some regions have issued prevention and control guidelines (3).Since China's accession to the World Trade Organization, the awareness of breeding management and deer breeding protection among breeders has gradually increased, making China's deer breeding industry more standardized (43).Implementing scientific breeding techniques is an important guarantee to promote the healthy development of the deer breeding industry and is an important means to improve economic efficiency.Overall, the intensive breeding model still has shortcomings.Therefore, we recommend that when developing large-scale intensive deer farming, standardizing breeding management, improving the environment of the livestock sheds, and improving animal welfare, might produce better prevention and control effects (51).
Analysis showed that in the age subgroups, the prevalence of TB in young deer (13.4,95% CI:10.5-16.3;72/529) was lower than that in older deer, which might be related to the consciousness of breeders to strengthen the feeding management of young deer.Many deer farms have adopted the preventive measures of disinfecting and isolating fawns immediately after birth, and artificially nursing and raising healthy fawns, which effectively controls the transmission of TB to fawns and lays a good foundation for the elimination of TB from deer herds (52).In this study, the highest prevalence was found in breeding deer (17.8, 95% CI:13.9-21.7;195/1068) and the lowest in adult deer (9.5, 95% CI:2.7-16.4;275/2101), which is inconsistent with previous studies showing that the prevalence of infection increases with age (30).The reason for this might be that most of the studies on the disease in breeding deer and young deer in the articles screened for this study were from areas with a high prevalence of deer TB, such as Liaoning province (20.8, 95% CI:1.0-100.0;1203/4656) and Jilin province (13.2, 95% CI:6.9-25.3;880/8423).
The analysis results showed that in the seasonal subgroups, the prevalence in winter to spring (28.9, 95% CI:12.2-68.5;63/224) was significantly higher than that in summer to autumn (7.5, 95% CI:3.6-15.6;317/5069).Deer are climate-sensitive ungulates (53), and temperature changes during the spring and winter seasons might be the main reason for the high prevalence.This result was further analyzed in the context of geographical factors: The spring and winter seasons are colder in the Northeastern part of China, and studies have  shown that the growth and reproduction of northern deer species are closely related to summer nutrition; for example, the population dynamics of moose are influenced by the availability of nitrogen in summer food (54).In addition, researchers generally agree that winter nutrition limits the growth of ungulates (55).Klein's theory suggests that the feeding status of large herbivores in spring and summer largely influences their growth, body weight, and sustained fecundity in autumn (56).Related studies have fleshed out Klein's theory, showing that breeding animals take in more food in the summer and autumn to increase their internal storage capacity (57).This feeding behavior is mainly used to increase protein storage to cope with the severe winter cold period.Protein is an important nutrient for animal survival and reproduction, and deer in summer and autumn have high protein storage capacities and are more resistant to disease; therefore, this might be one of the reasons for the low prevalence of TB in deer in summer and autumn (58).This viewpoint is also supported by the analysis of the annual temperature subgroups.TB prevalence correlated positively with temperature rises below 10°C, and negatively with temperature rises above 10°C, with the highest prevalence occurring over an average temperature between 5 to 10°C (12.1, 95% CI:8.7-5.8;718/6930).M. bovis does not readily survive in hot and dry environments (59).Therefore, we speculated that a temperature of 5� 10°C might promote the survival or transmission of M. bovis in deer.In the humidity subgroup analysis, deer raised in a humidity of 70-85% had the highest prevalence of TB (10.9, 95% CI,5.1-18.5125/706),which is consistent with the trend in the metaanalysis of dairy cattle TB (31); therefore, we speculated that a humid environment also favors TB spread.The sample size in the winter to spring (number of studies: 224) group was much smaller than that in the summer and autumn (number of studies:5069) group; therefore, the research effect is small.Further research is needed to investigate the relationship between the two.In the regional subgroup analysis, during the study period, the prevalence of deer tuberculosis in Central China (17.5, 95% CI:14.0-21.9;63/362) was higher than that in the other regions, followed by Northeast China (16.2, 95% CI:8.3-31.6;3640/17370).The latitude and longitude subgroups showed the highest prevalence in the longitude range 120-125° (15.5, 95% CI:0.0-59.2;1160/2465) and latitude range 40-45°(15.0,95% CI:5.1-28.9;1848/9260).Latitude was significantly associated with seasonal peaks of TB: the greater the latitude, the greater the magnitude of the seasonal peaks (60,61).Different latitudes result in different winter temperatures and sunlight UV levels.Northeast China is a region with high latitude, and in the spring and winter, the short daylight time and large temperature difference between day and night, will reduce the time deer spend outdoors in deer farms, thus decreasing the opportunity for deer to supplement vitamin D through sunlight.While skin exposure to solar UV radiation is the main source of vitamin D, another small portion comes from dietary intake (62).A positive correlation between the severity of tuberculosis in deer and the concentration of vitamin D was found in the serum of M. bovis-infected animals (63), and this vitamin enhances the ability of macrophages to kill mycobacteria (64).
In the province subgroup analysis, the highest prevalence of deer TB was found in Heilongjiang province (26.5, 95% CI:13.2-53.0;1557/4291), followed by Liaoning province (20.8, 95% CI:1.0-100.0;1203/4656), which corresponds to the high prevalence of deer tuberculosis in northeast China.This is probably caused by the temperate monsoon climate in northeastern China, which is ).At present, the detection of tuberculosis in mainland China mainly uses the bovine Mycobacterium purified protein derivative (PPD) method for the SICT, and the positive animals detected are isolated or eliminated.Although SICT is less specific, it has high sensitivity, and universal detection of the whole deer herd using SICT is beneficial for the early detection of TB infection in areas where TB is endemic.However, in areas where TB is controlled or shows stable control, SICT alone tends to produce more nonspecific interference, which is not conducive to tuberculosis purification (75).The prevalence of deer tuberculosis varies in different regions, environments, and conditions of deer breeding in mainland China; Forest plot of prevalence of tuberculosis in deer amongst studies conducted in mainland China.Funnel plot with pseudo 95% confidence interval limits for the examination of publication bias.This study is the first to analyze the prevalence of tuberculosis in deer in mainland China.The sample range of the study is wide and the method was rigorous, including a comprehensive analysis of various influencing factors, thus providing an effective reference for the prevention and control of tuberculosis in deer in mainland China.However, some limiting factors might make some of the results of this meta-analysis unstable.There were few articles examining deer TB, although we used several different search formats based on six databases to retrieve more eligible studies.However, only 20 articles ultimately met our screening criteria.This left the analysis under-powered for certain subgroups and might have led to inconsistent results.

Conclusion
Our research showed that deer tuberculosis is widespread in mainland China (16.1,95%CI:10.5-24.6;5367/22215; 1981 to 2023).Geographical distribution, seasons, and detection methods are factors affecting the assessment of the prevalence of deer  Funnel plot with trim and fill analysis of the publication bias.

FIGURE 1 Flow
FIGURE 1Flow diagram of eligible studies for searching and selecting.The inclusion and exclusion criteria only to this study.

FIGURE 4 Egger
FIGURE 4Egger's test for publication bias.

TABLE 1
Included studies of tuberculosis in deer in mainland China.

TABLE 2
Normal distribution test for the normal rate and the different conversion of the normal rate.

TABLE 3
Pooled prevalence of tuberculosis in deer in mainland China.

TABLE 4
(52,raphical factors prevalence of tuberculosis in deer in mainland China.bysimultaneousrainandheat, a dense network of rivers, sufficient water supply, fertile soil, and high vegetation cover in the black soil area, which can provide sufficient food and water for deer.Moreover, the northeastern region is sparsely populated, so it is very favorable for deer breeding (65).Since the beginning of this century, because of the financial crisis and other reasons, China's deer breeding industry has been in a downturn.The slaughter volume around the rapid rise of TB, and the high rate of culling sick deer, have played a role in the control of tuberculosis, and since 2008, it has been influenced by the national policy of supporting the breeding industry and the rise of international antler prices, consequently, the volume of deer breeding in the northeast has grown significantly; however, the development of quarantine work for TB is slow, and the contradiction between the two is an important reason for the increased prevalence of deer TB in Northeast China(52), and has contributed to the high prevalence of the disease (2006, 64/563; 2010,210/1086)(52, 66).Field research by the China Animal Husbandry Association Deer Branch in Jilin province, Heilongjiang province and other places concluded that the local government of Heilongjiang Province needs to further increase its emphasis on deer breeding and scientific research, and increase funding and policy support to establish breeding standards and feed nutrition standards to meet the needs of high-level development of deer factories (67).However, there were few studies in multiple provinces included in the present meta-analysis, which might have affected the stability of the results.Therefore, we suggest that the relevant departments of provinces and cities should strengthen the monitoring of deer TB to clarify the regional differences in deer TB in mainland China.In the meta-analyses of deer TB prevalence, the different assays used were the main source of heterogeneity.
(73)acterized is an in vitro immunological method recommended by WHO to diagnose M. tuberculosis infection, and China added TB-IGRA to the industry standard for tuberculosis diagnosis (WS 288-2017) in 2017(72).TB-IGRA detects tuberculosis mainly using the proteins encoded by the RD1 and RD16 regions of M. tuberculosis, which have good specificity(73); however, their sensitivity is limited (71, *: Annual average temperature; Temperature 2 *: Annual minimum temperature; Temperature 3 *: Annual maximum temperature.

TABLE 4 (
Continued) Li et al. 10.3389/fvets.2024.1333975Frontiers in Veterinary Science 14 frontiersin.orgtuberculosis.Based on the results of our study on the epidemic factors of deer TB in mainland China, we recommend the establishment of scientific breeding bases, scientific feeding programs, strengthening technical training, increasing the attention of regulatory authorities to deer breeding research, and improving deer feeding and disease diagnostic techniques according to different local feeding methods, geographical factors, and climatic conditions.This is especially important in the northeast region, where breeding is concentrated.In addition, deer tuberculosis control and epidemic prevention systems should be established in all breeding areas.Researchers should carry out epidemiological investigations in more breeding areas to further refine the risk factors for deer tuberculosis epidemics, with the aim of providing a solid foundation for the prevention and control of deer tuberculosis in mainland China.