Bovine Tuberculosis in Cattle in the Highlands of Cameroon: Seroprevalence Estimates and Rates of Tuberculin Skin Test Reactors at Modified Cut-Offs

The aim of this study was to obtain epidemiological estimates of bovine tuberculosis (TB) prevalence in cattle in the highlands of Cameroon using two population-based tuberculin skin test (TST) surveys in the years 2009 and 2010. However, prior to the TST survey in 2010, blood was collected from already chosen cattle for serological assay. Anti-bovine TB antibodies was detected in 37.17% of tested animals and bovine TB prevalence estimates were 3.59%–7.48%, 8.92%–13.25%, 11.77%–17.26% and 13.14%–18.35% for comparative TST at ≥4 mm, ≥3 mm and ≥2 mm cut-off points and single TST, respectively. The agreement between TST and lateral flow was generally higher in TST positive than in TST negative subjects. The K coefficients were 0.119, 0.234, 0.251 and 0.254 for comparative TST at ≥4 mm, ≥3 mm and ≥2 mm cut-off points and the single TST groups, respectively. Chi square statistics revealed that strong (P < 0.05; χ 2 > 48) associations existed between seroprevalence rates and TST reactors. The study suggested that using lateral flow assay and TST at severe interpretations could improve the perception of bovine TB in Cameroon. The importance of defining TST at modified cut-offs and disease status by post-mortem detection and mycobacterial culture of TB lesions in local environments cannot be overemphasised.


Introduction
Bovine tuberculosis (TB) is a zoonotic disease with severe public health significance but it is neglected in Cameroon. The tuberculin skin tests (TSTs) are currently the best available techniques for international field diagnosis of bovine TB in live animals [1,2] and it is based on delayed hypersensitivity reactions [3]. The single intradermal comparative cervical tuberculin (SICCT) test involving the intradermal injection of bovine tuberculin (BT) and avian tuberculin (AT) at separate sites in the skin of the neck gives more specific results than the single intradermal tuberculin (SIT) test which uses only BT [4,5]. TST can effectively detect early stages of M. bovis infection in cattle and allows for rapid removal of infected animals, limited transmission, and fast eradication of bovine TB [6]. There are OIErecommended cutoff points of the increase in skin thickness for SICCT-BT and SIT-BT to be positive [3]. However, the OIE-recommended cutoff values were established mainly in developed countries for Bos taurus cattle, and different cutoff 2 Veterinary Medicine International values are applied according to a particular country's disease status and objective of its disease control programme [4,[7][8][9].
The performance of TST could be affected by environmental factors, host factors (status of immunity, genetics), and nature of the tuberculin used [1,4,5,9]. A perfect cutoff point in a specific geographic area may not be so useful in another environment [1,4]. Also, the ability of the test to predict positive disease status depends on its sensitivity and specificity and prevalence of the disease in tested population [1]. Anergic animals, animals exhibiting reactions to both avian and mammalian tuberculins, animals in advanced stage of disease, periparturient cows, and animals with confined infection notably in the udder and with localised infection often in the lymphatic glands that has become inactive (latent) have been reported to be poor responders to TST [10]. However, severe interpretations are done in regions or herds where M. bovis infection has been confirmed, and SIT-BT reactors may also be subjected to an SICCT-BT test, based on the discretion of the veterinarian [4]. Veterinarians continue to play pivotal roles in inspections of animal (antemortem and postmortem) and animal products, diagnosis of M. bovis infected cattle, and impacting of cattle producers in bovine TB eradication programs [11]. Postmortem detection of TB lesions and other bovine TB diagnostic techniques (e.g., gamma-Interferon, ESAT-6 tests, serologic and fluorescence polarization assays) have been used to determine the ability of TST in the diagnosis of bovine TB in cattle in different environmental conditions around the world, including parts of Africa [1,2,6,7,9,[12][13][14][15][16]. However, TST-negative animals at slaughter with evidence of encapsulated lesions confirmed as caused by M. bovis have also been reported [10].
TST may demand physical exertion in the field but it is also simple and relatively inexpensive and offers reliable means of screening cattle populations in an entire region [4,6]. Ancillary tests are being used and/or currently being validated to improve diagnosis and reduce the number of false positive results following TST [1,2,6,7]. Also, rapid and simple immune-chromatographic assays for the serodiagnosis of bovine TB have been developed [17,18] and proposed as additional tests to the TST for antemortem diagnosis [2,19,20]. These chromatographic immunoassays employ unique cocktails of selected M. bovis antigens as both qualitative captures and detectors of specific antibodies against M. bovis in plasma, serum, and whole blood [17,21]. MPB83, ESAT-6, 14-kDa protein, CFP-10, MPB70, MPT63, MPT51, MPT32, MPB59, MPB64, Acr1, PstS-1, M. bovis purified protein derivatives, ESAT-6/CFP10 fusion protein, 16-kDa alpha-crystallin/MPB83 fusion protein, and M. bovis culture filtrate have been identified as the common seroreactive antigens in bovine TB [17,18,22]. The bound antibodies are visualized with the naked eye as colour band at the test device within some minutes of application [17,21]. The assay requires no specific expertise or equipment, and the test kit may be kept without the need for refrigeration [17,18,21].
There are scanty reports of bovine TB prevalence in Cameroon, modifications of the OIE standards of TST applied elsewhere have been used to estimate the disease status in cattle in the country, and the findings have varied widely, even for the same sites [23][24][25][26][27]. This study was carried out to investigate bovine TB prevalence in cattle in the highlands of Cameroon through seroprevalence estimations, rates of TST reactors at modified cutoff points, and the epidemiological usefulness of the proposed screening algorithms. TST data of tested cattle in the years 2009 (n = 2, 853) and 2010 (n = 1, 381) were reanalyzed, and the epidemiological implication for applying TST at various cutoff points for a predominantly Zebu cattle population was discussed.  [25] in the Western highlands of Cameroon was used to estimate the number of cattle required to detect ≥1 positive reactor with a desired 95% confidence and precision of 5% as previously described [28]. The selection of cattle herds was done by the random-number generation method of cattle keeping communities, cattle owners, and locations of herds from records of annual livestock vaccination campaigns (contagious bovine pleuropneumonia, pasteurellosis, black quarter) at the Regional Delegations of MINEPIA (Ministère de l'Elevage, des Pêches et des Industries Animales (Ministry of Livestock, Fisheries and Animal Industries)). All animals within selected herds were tested except recently calved cows (within 2 months postpartum) and calves less than 6 months old because of immunosuppression in lactating cows and high maternal antibodies in calves that desensitizes them to tuberculin [29,30].

Materials and Methods
During March to September 2009, a total of 2,853 cattle (84 herds) were tested in five administrative divisions in the Northwest regions of the Western highlands (Donga and Matung, Menchum, Bui, Mezam and Boyo) and one division in the Adamawa plateaux (Vina) of Cameroon ( Figure 1). Similarly, 1,381 cattle (40 herds) were tested during May to September 2010 in Mezam and Bui divisions in the Western highlands which showed high bovine TB prevalence rates in the previous survey and also in the Vina division in the Adamawa plateaux. However, 30-60 minutes prior to the TST carried out in the year 2010, blood was collected from 807 cattle in 20 randomly selected herds of the 40 already chosen herds (1,381 cattle) to extract serum for lateral flow assay of antibovine TB antibodies (Antibovine TB Ab).
Risk assessments were done to avoid hazards to all persons and animals involved in the project. The project approval and ethical clearances were obtained from the required authorities in Cameroon including the National Ethics Committee, regional delegations of MINEPIA in the Northwest and Adamawa regions. The purpose of the study was explained to the targeted participants usually with   the assistance of resident veterinarians, local community leaders, and trusted intermediaries. A herd was tested after an informed consent was given by the owner. Apart from minor jugular vein puncture for blood collection, intradermal injections of AT and BT, and procedural restraining manipulations for safety purposes, the animals were not subjected to suffering.

Tuberculin Skin Tests and Classification of Reactors.
TSTs were carried out in the selected cattle (2,583 in the year 2009 and 1,381 in the year 2010 including the 807 blood donors but after blood collection) by intradermal injections of 0.1 mL each of AT (2500 IU/dose) and BT (3000 IU/dose) in two sites, at 12 cm apart in the right neck region. A correct intradermal injection was confirmed by palpating a small grain-like swelling at each injection site. The skin thickness was measured prior to and 72 hours after injecting the tuberculins using a digital calliper. The OIErecommended ≥4 mm cutoff point of increase in skin fold thickness [3] as well as ≥3 mm and ≥2 mm cutoff points was assessed for SICCT-BT reactor status. The corresponding ranges ≥1 mm to <4 mm, ≥1 mm to <3 mm, and ≥1 mm to <2 mm were classified as doubtful responses, respectively. SICCT-BT was noted as negative if the skin response was <1 mm. SIT-BT interpretations were done using skin fold thickness of ≥4 mm, ≥2 mm to <4 mm, and <2 mm for positive, doubtful, and negative responses, respectively [3]. These cutoff points were assessed against the demonstrated circulating antibovine Tb antibodies status and classified as adapted from Martrenchar et al. [23] to determine the cutoff zone and risk group of TST reactors for consideration ( Figure 2).

Data Management and Statistical
Analysis. The lateral flow assay results and TST data at the ≥2 mm, ≥3 mm, and ≥4 mm cutoff points for individual cattle were entered into Microsoft Excel (Microsoft Corporation, USA) and also exported to SigmaPlot (Systat Software Inc, USA) for further analysis. The seroprevalence estimates, rates of TST reactors in the tested cattle population, and agreement between both methods at the predefined cutoff points were assessed [28].
The predictive values and diagnostic likelihood ratios of TST at the various cutoff points were compared against the antibovine TB Ab assay [28]. With sensitivity and specificity values obtained by Ameni et al. [9] and Pollock et al. [12], the observed prevalence rates were corrected using the Roganand-Gladen formula [28,31]. The kappa statistics was used to estimate the degree of agreements between both tests while Chi-square techniques were applied to compare individual and herd prevalence of reactors in the different variables [28,32].

Comparison of Tuberculin Skin Tests at Modified Cutoff Points and Lateral Flow Assay in Cattle
Reactors. The predictive values and likelihood ratios of SICCT-BT at various cutoff values and SIT-BT in cattle reactors against the antibovine TB Ab assay are shown in Table 3. Strong associations were noted between the seroprevalence estimates and rates of TST reactors irrespective of the TSTS cut-off value (P < 0.05; χ 2 > 48) in this study. However, decreasing the cutoff points revealed inverse relationships with test predictive values and diagnostic likelihood ratios. The ability of SICCT-BT to produce no false negative result increased with increase in cutoff point (nonsignificant differences were noted between the ≥2 mm versus ≥3 mm and ≥3 mm versus ≥4 mm cutoff points). The findings also suggested that prediction of disease status improved with severe interpretation of TST (decreasing cutoff point). The study indicated that using antibovine TB Ab assays as ancillary diagnostic tests to SICCT-BT in cattle could significantly improve diagnosis of bovine TB cases. Statistically, the best all round SICCT-BT performance was realized at the ≥3 mm cutoff point. However, the ≥2 mm cut-off value showed the highest positive predictive value and a comparable positive diagnostic likelihood ratio to the others.

Discussion
There is gross inadequacy in the implementation of the existing bovine TB control policy in Cameroon. Culling of TST reactors as part of a national animal disease control policy is not a routine practice due to political, economic, and social limitations. However, veterinarians continue to identify bovine TB lesions in slaughtered cattle across the country [33][34][35]. TB lesions have been detected in TST reactors at cutoff points less than the OIE-recommended optimal 4 mm cut-off [8,9,15] and TST negative reactors [10]. TB lesions were also observed in TST doubtful and negative reactors in Mezam Division in the present study. Lack of knowledge on the actual magnitude and distribution of the disease, inadequate laboratories and field expertise, and politicoeconomic deficiencies are common factors that limit bovine TB control in most of Africa [36]. The current control approach in Cameroon is based on controlling animal movements, culling suspected bovine TB cases and  carcass condemnation (partial or whole) at meat inspection [37]. Apparently, the strategies were designed to reduce the general prevalence and monitor spread of the disease in livestock. TST is presently a passive component of Cameroon's government strategy to control bovine TB which is of major concern to the veterinary and medical services. Maximum detection of bovine TB in cattle populations in Cameroon is vital to understand its epidemiology and zoonotic potentials and also achieve significant reduction and control of the disease in livestock. Cell-mediated immune responses develop early after bovine TB infection in cattle while antibody responses may not become obvious until later and at advanced stages of the disease, when cell mediated reactions (TST reactions) are waning [38][39][40]. TST can boost antibody responses in M. bovis infected cattle and emphasizes the importance of timing of collection of blood samples on the interpretation the test [38]. In this study, the antibovine TB antibody detection (Anigen lateral-flow assay) that employed recombinant M. bovis MPB70 antigen as capture and detector was conducted prior to TST. This antibovine TB antibody test kit has a sensitivity of 90% against bovine TB confirmed by bacterial isolation and a sensitivity of 85.1% and specificity of 98.6% against TST [41]. Also using the Anigen lateral-flow assay, Whelan et al. [42] achieved a sensitivity of 84% and a specificity of 84.2% for serological diagnosis of M. bovis infection in cattle. Similar and relatively high sensitivity (86.5% and 84.6%) and specificity (83.8% and 91.4%) have been reported with other lateral flow techniques (CervidTB STAT-PAK and DPP VetTB assays, resp.) for the rapid diagnosis of bovine TB in farmed Red deers [43]. Furthermore, a sensitivity of 89.6% and specificity of 90.4% were achieved in the diagnosis of M. bovis infection in Eurasian wild boar using the DPP VetTB assay (based on combining two separate test antigens) [44]. However, the specificity of these test kits could be affected by cross-reacting members of the M. avium complex [43,44], and high false positive results were observed when a commercial multiantigen lateral flow assay was performed in dairy cattle [45]. Nonetheless, significantly higher specificity of 98.4% and sensitivity of 93.1% in the diagnosis of bovine TB in cattle have been obtained for multiplex immunoassay based on a combination of antigens compared to those of assays based on a single antigen [22,42]. The TST accuracy against postmortem detection of TB lesions revealed a sensitivity of 86% and specificity of 90% for SIT-BT [12], while sensitivity values of 69%, 65%, and 59% at SICCT-BT ≥2 mm, ≥3 mm, and ≥4 mm cutoff points and a specificity of 97% at these cutoff points have been reported [9]. The lack of a well-established gold standard in this study was a key problem in calculating the sensitivity and specificity of the lateral flow assay and TST at the modified cutoff points.
The findings of this study suggest that TST at any cutoff point could be used to detect bovine TB in cattle and the test accuracy increased with increase in cut-off value. Cattle presenting differential SICCT-BT skin thickness of less than 4 mm in Cameroon should therefore not be excluded that they are negative for bovine TB. These animals may be infected but low reacting or not reacting at all if their immune systems were not stimulated enough for a positive response at the ≥4 mm cutoff point [46,47] due to conditions such as stress that compromise immune function [48]. Also, the animals may have been sensitized to environmental mycobacteria [38]. Furthermore, delayed hypersensitivity to tuberculin may not develop for a period of 3-6 weeks following infection [3,10]. Delaying TST of a herd/animal suspected to have been in contact very recently with infected animals in order to reduce the probability of false-negatives has been suggested [10] since it is unlikely that the control and eradication of TB from a herd will be achieved with only a single tuberculin test [3]. In this study, maximum positive prediction values and negative likelihood ratio were observed at the SICCT-BT ≥2 mm cutoff point and maximum negative prediction and positive likelihood ratio at the ≥4 mm cutoff point. The findings also revealed that 31 cattle (over 3.84%) considered as SICCT-BT doubtful reactors at the ≥4 mm cutoff point could be identified as positive bovine TB cases at the ≥3 mm and ≥2 mm cutoff points. The poor to fair agreements recorded suggested that severe interpretation of TST (i.e., decreasing skin response cut-off values) improved the agreement between TST and the lateral flow assay to detect TST positive reactors. The prevalence rates at the modified cutoff points could have influenced the estimated Kappa values. However, low kappa values have been obtained between good diagnostic and negatively correlated tests [28]. The poor correlation between comparative TST at the ≥4 mm cutoff point and antibovine TB antibody test results in the study was not unexpected. Therefore, the importance of determining appropriate localised TST cut-off values supported by validated methods in Cameroon cannot be overemphasized.
Though it is essential that tuberculin of sufficient potency to produce a reaction in the maximum number of infected animals is essential, a tuberculin of potency greater than that to which the majority of infected animals will respond has been proposed in TST [10]. However, Good and Duignan [10] had warned that highly potent tuberculin tends to increase the frequency of reactions associated with crosssensitisations arising from other organisms such as the human and avian types (M. tuberculosis and M. avium, resp.) and other (nonpathogenic) mycobacteria. Nonspecific responses in TST due to atypical or environmental mycobacteria have been widely reported [2,3,[49][50][51]. Indeed, Lesslie et al. [52][53][54] recorded hypersensitivity responses to avian tuberculin that was equal or higher than responses to bovine tuberculin in cattle naturally infected with M. bovis and presenting visible lesions at slaughter. Therefore, severe interpretations of TST reactions should be employed when EU-and OIE-recommended tuberculin preparations are used in bovine TB endemic regions and environments where multiple mycobacteria are coexisting. The findings of this study agree with Martrenchar et al. [23] who reported high frequency of atypical mycobacteria which severely limited the reliability of SIT-BT and SICCT-BT results at the OIE-recommended 4 mm cutoff point in Northern Cameroon. Severe interpretations of TST results in the study revealed that many SIT-BT positive and SICCT-BT doubtful responses at ≥4 mm cutoff point could be appropriately identified as bovine TB cases at reduced cutoff points (some Excess D4 and Excess D3 reactors). The high detection of TST and antibovine TB antibody positive herds irrespective of TST cutoff point and findings of circulating antibovine TB antibody could suggest that the cattle were widely exposed to and affected bovine TB and other mycobacterial infections.
In this study, reducing the cutoff point from ≥4 mm improved the ante mortem detection of bovine TB in cattle using SICCT-BT and antibovine TB Ab tests. Overall, the maximum test ability was realized at ≥3 mm cutoff point and the best SICCT-BT positive predictive value was at ≥2 mm cutoff point. These findings revealed that interpreting SICCT-BT at the ≥2 mm cutoff point, and not at the ≥3 mm or ≥4 mm cutoff points, was beneficial from a public health perspective. However, there would be concrete risk of unnecessarily identifying more cattle at severe TST interpretations. This study cannot exclude that some SICCT-BT doubtful reactors at the ≥3 mm and the ≥4 mm cutoff points were infected cases detected at the ≥2 mm cutoff point. The application of the SICCT-BT ≥2 mm cutoff point should be considered in cattle in the agro-ecological highland environments of Cameroon for greater detection of bovine TB. Severe TST interpretation would be vital to effective control of the disease and reduction of its zoonotic risks to public health and food safety in the country.

Conclusion
The TST and antibovine TB antibody tests when used in parallel offered improved detection of bovine TB compared to individual tests. Bovine TB was detected at all the cutoff points and there were strong associations between both methods in the highlands of Cameroon. The best test performance was realized at the ≥3 mm cutoff point. However, interpreting SICCT-BT at ≥2 mm cutoff point was more strategic from a public health context since more affected cases would be predicted. The study revealed that the prevalence of bovine TB was high and atypical mycobacteria infection was widespread in the regions. Bovine TB-infected cattle which maybe anergic due to age, malnutrition, and/or suffering from concurrent diseases such as internal and external parasitosis (common scenarios in the study regions) could be detected at severe SICCT-BT interpretation. Their delayed hypersensitivity responses to tuberculin would be limited and cannot express the full OIE-recommended ≥4 mm cutoff point. However, it is important to investigate the performance of TST at modified cutoff points against 12 Veterinary Medicine International defined bovine TB status confirmed by postmortem examination and culture of TB lesions in reacting animals in the Cameroon environments.