Research in Veterinary Science 95 (2013) 1107–1113

Contents lists available at ScienceDirect

Research in Veterinary Science

journal homepage: www.elsevier .com/locate / rvsc

Ante-mortem diagnosis of caprine tuberculosis in persistently infectedherds: Influence of lesion type on the sensitivity of diagnostic tests

0034-5288/$ - see front matter � 2013 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.rvsc.2013.10.003

⇑ Corresponding author. Address: Departamento de Anatomía y AnatomíaPatológica Comparadas, Universidad de Murcia, 30100 Murcia, Spain.Tel.: +34 868884336; fax: +34 868884147.

E-mail address: abuendia@um.es (A.J. Buendía).

Antonio J. Buendía a,⇑, Jose A. Navarro a, Jesus Salinas b, Jim McNair c, Lucia de Juan d, Nieves Ortega b,Paulina Cámara e, Pilar Torreblanca e, Joaquin Sanchez a

a Departamento de Anatomía y Anatomía Patológica Comparadas, Facultad de Veterinaria, Campus Regional de Excelencia Internacional, Campus Mare Nostrum, Universidad deMurcia, Spainb Departamento de Sanidad Animal, Facultad de Veterinaria, Campus Regional de Excelencia Internacional, Campus Mare Nostrum, Universidad de Murcia, Spainc Veterinary Sciences Division, Agri-Food and Biosciences Institute, Stormont, Belfast, Northern Ireland, United Kingdomd Centro de Vigilancia Sanitaria Veterinaria (VISAVET), Universidad Complutense de Madrid, Spaine Servicio de Salud Animal, Consejería de Agricultura y Agua, Región de Murcia, Spain

a r t i c l e i n f o

Article history:Received 11 February 2013Accepted 3 October 2013

Keywords:GoatsTuberculosisDiagnosisELISASkin test

a b s t r a c t

Caprine tuberculosis is a major health problem for goats, and an important zoonosis. Eradicationprogrammes using the comparative skin test are being developed to aid in decreasing the prevalenceof infection. However, persistent tuberculosis infections are found in herds subjected to eradicationprogrammes. In the present study a commercial IFN-c release assay and an experimental ELISA basedon MPB70, were evaluated as potential ancillary tests to detect infection. The relationship between thethree techniques (skin test, IFN-c release and ELISA) and histopathological lesions was analyzed in 162goats from herds with persistent tuberculosis infection. The presence of related pathogens (paratubercu-losis and pseudotuberculosis) was also studied. The IFN-c release assay and the ELISA had a higher sen-sitivity than the comparative skin test (65.3% and 66.3% vs 44.5%) using as a gold standard a combinationof histopathological analysis and isolation. Used in combination, ELISA and the skin test detected 89.1% ofgoats with tuberculosis while a combination of IFN-c assay and skin test detected 78.2%. The types ofmacroscopic and microscopic lesions reflected the results of diagnostic tests.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

The extent of caprine tuberculosis has been underestimated fordecades, but recent studies have pointed to an increased incidenceof the disease and a new species, Mycobacterium caprae, has beenincluded in the Mycobacterium tuberculosis complex (Aranazet al., 2003). The disease causes serious economic losses due tothe removal of infected animals and imposes limitations to trade(Seva et al., 2002; Daniel et al., 2009). In the Region of Murcia(south-eastern Spain) an eradication programme has been in pro-gress since the mid-1990s, using the comparative tuberculin skintest, which involves the intradermal injection of bovine purifiedprotein derivative (PPD) and avian PPD. However, although theprogramme has managed to reduce the prevalence of the disease,there are still a number of herds with persistent tuberculosis. Thisdifficulty in eliminating the disease from herds with a long history

of tuberculosis justifies research into new ancillary techniques thatwill aid in the eradication of the disease.

Cell-mediated immunity is of fundamental importance to thecontrol of tuberculosis in the host and has therefore been the focusof the most widely used diagnostic tests, including the tuberculinskin test and the measurement of IFN-c released after mycobacte-rial antigen stimulation (Bezos et al., 2012). In contrast, the lowsensitivity of antibody-based tests has prevented their use on alarge scale in the diagnosis of tuberculosis in both cattle and goats(Pollock et al., 2001). Recently however, the use of serological testshas shown promising results for the diagnosis of tuberculosis inboth goats and cows (Marassi et al., 2009; Whelan et al., 2010).

A relationship between the results of diagnostic techniques andthe type or severity of tuberculous lesions has been observed incattle (Vordermeier et al., 2002). A previous study carried out byGutiérrez et al. (1998), in goats that were not included in a tuber-culosis eradication programme, showed an association betweenthe results of the antibody- and cell-mediated based techniquesand the progressive development of tuberculosis in goats. Forexample, the ELISA showed low sensitivity in animals with minormacroscopic lesions, although sensitivity increased in animals withsevere macroscopic lesions, while the cell-mediated immunity-

1108 A.J. Buendía et al. / Research in Veterinary Science 95 (2013) 1107–1113

based techniques were reported to be more efficient in goats withminor lesions (Gutiérrez et al., 1998). Furthermore, the local andsystemic immune responses of goats with granulomatoustuberculous lesions was substantially different from that observedin animals with cavitary lesions (Caro et al., 2001; Sanchez et al.,2011), suggesting that the type of disease progression disclosedby histopathological examination might influence the result ofthe diagnostic technique being used. However, the influence ofthe phase of the disease and the type of lesion on the performanceof the diagnostic tests has not been described in animals fromherds subject to eradication programs.

Another aspect that might influence the diagnosis ofcaprine tuberculosis is the presence of pathogens capable ofcausing cross-reactions, leading to misdiagnosis, for exampleMycobacterium avium subs. paratuberculosis (Álvarez et al., 2008),or the presence of the tuberculosis-like lesions caused by infectionwith Corynebacterium pseudotuberculosis (Sharpe et al., 2010).

The aim of the present study was to identify those pathologicalfactors that might influence the ante-mortem detection of tubercu-losis in caprine herds with a long term history of infection.

2. Materials and methods

2.1. Herds and animals

The Murciano-Granadina goats used in this study were obtainedfrom four different herds, from the Region of Murcia. Spain. Theherds were selected according to their history of persistent infec-tion with Mycobacterium spp. in the last four years and herd size(medium size herds, with 100–200 reproductive animals). The fourselected herds were included in the regional caprine tuberculosiseradication programme based on results from the comparativetuberculin skin test. Using the skin test, each herd had been testedat least three times per year over the previous four years, with po-sitive animals in each test. These herds had not been vaccinatedagainst paratuberculosis. The number of animals tested in thestudy in all the four herds was 498. Blood samples were taken priorto the skin test for the ELISA and specific IFN-c release assay. Atotal number of 162 goats were sacrificed (41 from herd A, 39 fromherd B, 39 from herd C and 43 from herd D). The ‘‘positive’’sacrificed animals had all returned positive results in one or morediagnostic tests (n = 126 animals). The study was completed with anumber of ‘‘negative’’ animals (n = 36) randomly selected amongthe 372 remaining goats that were negative to the three diagnostictests performed. This ‘‘negative’’ group contained animals from allfour herds. All the 162 sacrificed animals were older than one yearand female except for three males.

2.2. Cervical comparative tuberculin skin test

The comparative tuberculin skin test was conducted by theVeterinary Services of the Regional Government of Murcia,according to national and regional legislation (R.D. 1328/2000and regional directives 30/03/1995 and 26/07/2000) following asevere interpretation: animals were classified as positive whenthe increase in skin fold thickness in the side of bovine PPDwas greater than 2 mm and exceeded the increase on the avianPPD side. The skin fold thickness was measured and recorded bythe same veterinarian. Two intradermal injections were per-formed in each animal on both sides of the neck. The animalswere inoculated with 0.1 ml of bovine or avian PPD (CZ Veterina,Pontevedra, Spain), the dose containing 0.1 mg of bovine PPD or0.05 mg of avian PPD. The skin test was conducted every 2–3 months in the problem herds, with the average period betweentwo tests being 69.8 days.

2.3. Specific IFN-c release assay

The specific IFN-c release assay was carried out in 24-wellplates, where 1 ml of blood was placed in each well. Stimulationwas carried out with avian or bovine PPD following the methoddescribed by Casal et al. (2012). Wells treated with PBS served asnegative controls. Plasma samples were analyzed using the Bovi-gam assay (Bovine Gamma Interferon Test, Prionics AG, Schlieren,Switzerland). Animals were considered positive if the opticaldensity (OD) of a sample stimulated with bovine PPD minus theOD of PBS was greater than 0.05 and greater than the OD of thesample stimulated with avian PPD.

2.4. Elisa

The recombinant protein MPB70 from Mycobacterium bovis waspurified according to Lightbody et al. (1998) and used as antigen inan ELISA to detect specific antibodies. Briefly, the protein was di-luted to a concentration of 0.5 lg/ml in a carbonate/bicarbonatecoating buffer (pH 9.6) and used to coat the 96-well plates (Nunc,Roskilde, Denmark). The plates were allowed to stand overnight at4 �C and then washed three times with 250 ll/well of PBS contain-ing 0.05% Tween 20 (Sigma, Madrid, Spain) (PBST). Goat seradiluted to 1/100 in an ELISA Diluent Solution (eBioscience, SanDiego, CA, USA) were then added to each well of the plate. Afterincubation for 60 min at 37 �C, the wells were washed four times,and a peroxidase conjugated anti-goat polyclonal IgG (AbD Serotec,UK) was used at a dilution of 1/5000 in the ELISA Diluent Solution(100 ll/well). After a second incubation for 60 min at 37 �C, theplates were again washed four times. The reaction was revealedwith 100 ll/well of Super AquaBlue ELISA Substrate (eBioscience)for 15 min. Results were expressed as optical densities (OD) afterreading the plates at 405 nm in an ELISA plate reader (DigiScanwith DigiWin Software, ASYS Hitech, Austria). A cut-off valuewas calculated using 10 sera obtained from 10 goats in a tubercu-losis-free herd (Experimental Farm, School of the VeterinaryMedicine, Murcia). These control negative animals previouslytested negative for tuberculosis by skin test and the IFN-c releaseassay. The formula used to obtain the cut-off had been describedpreviously (Salinas et al., 2009), and calculated from the meanvalue of the 10 negative sera plus three standard deviations.

2.5. Postmortem analysis

Goats were killed by intravenous injection with pentobarbital,following the guidelines of Bioethics Commission of the Universityof Murcia. Necropsies were performed by two certified veterinarypathologists. Based on the presence and distribution of lesionscompatible with tuberculosis, all animals were classified into threegroups: Group 0, animals without macroscopic lesions; Group 1,animals with single lesions; Group 2, animals with multiple lesions(in the same organ or in various organs). Animals with a singlelesion in an organ (i.e. the lung) and a lesion in the regional drain-ing lymph node (i.e. tracheobronchial or mediastinal lymph nodes)were included in Group 1. Samples from each lesion compatiblewith tuberculosis (found both in organs and lymph nodes) were ta-ken in duplicate for histopathological analysis and mycobacterialculture. For animals identified in Group 0, the mediastinal, tracheo-bronchial and retropharyngeal lymph nodes were taken forhistopathological and bacteriological studies.

2.6. Histopathological study

Tissues were fixed in 4% formaldehyde for 48 h before process-ing and embedding in paraffin wax. Serial sections at 4 lm were

A.J. Buendía et al. / Research in Veterinary Science 95 (2013) 1107–1113 1109

stained with haematoxylin and eosin (HE) for routine histopathol-ogy and with Ziehl-Neelsen for detection of acid-fast bacilli.

Based on microscopic histopathology four types of lesion weredefined:

(i) Small granulomatous lesions (SG) were characterized by theaccumulation of epithelioid macrophages with low numbersof lymphocytes and neutrophils. Langhan’s multinucleatedgiant cells were sometimes present, but necrosis was absentor minimal. The granulomas may, or may not, have been sur-rounded by a thin connective tissue capsule. The number ofacid-fast bacilli in the lesions was variable.

(ii) Large granulomatous lesions (LG) were characterized by athick fibrotic capsule surrounding irregular multicentricgranulomas, with prominent caseous necrosis andnumerous foci of dystrophic mineralization. Epithelioidmacrophages and Langhan’s multinucleated giant cells sur-rounded necrotic areas, with particularly dense clusters oflymphocytes near the peripheral fibrotic capsule. Acid-fastbacilli were scarce in the lesions. Large granulomatouslesions were classified into large granulomatous granulomasin single lesions (LGs) or in multiple lesions (LGm) in orderto establish the relationship between microscopic lesionsand the diagnostic results.

(iii) Cavitary lesions (CL) admixed with large granulomas andtuberculous pneumonia. The histopathological findingsassociated with cavitation included striking cellular accu-mulation on the luminal surface of the cavity, consisting ofnumerous intact or degenerate neutrophils and some epi-thelioid macrophages adjacent to or mixed with liquefiedcaseous necrotic material. The peripheral zone was formedof granulation tissue infiltrated by a number of macrophagesand lymphocytes and surrounded by a thick fibrotic capsule.Acid-fast bacilli were abundant in the cavitary lesions.Microscopically, tuberculous pneumonia was characterizedby large uncalcified areas of necrosis associated with a dif-fuse irregularly distributed leucocytic infiltrate. Langhan’smultinucleated giant cells were scarce.

(iv) Fibrocalcified lesions (FC) were associated with a combina-tion of mineralization and/or collagenous material thatreplaced caseous necrotic areas. Acid-fast bacilli were scarce.

2.7. Immunohistochemistry

Immunohistochemistry was carried out as previously described(Sanchez et al., 2011). Briefly, paraffin sections were dewaxed,rehydrated and treated with 3% H2O2 in methanol for 15 min toblock endogenous peroxidase activity. The primary antibody usedwas a rabbit polyclonal antibody against M. tuberculosis (AbDSerotec, Kidlington, UK). Positive control sections of previouslyconfirmed cases of tuberculosis were included in the technique.The binding of primary antibody was detected using the avidin–biotin–peroxidase complex (ABC Vector Elite, Vector LaboratoriesBurlingame, CA, USA). Positive labeling was detected using3,3-diaminobenzidine tetrahydrochloride (Dako Cytomation,Carpinteria, CA, USA) and sections were counterstained withMayer’s haematoxylin.

2.8. Isolation of M. tuberculosis complex bacteria

Samples for bacteriological analysis were collected and storedat �80 �C until required. Isolation was carried out from all lesionsor from pooled lymph nodes (mediastinal, tracheobronchial andretropharyngeal) of animals without lesions. Bacteriological isola-tion was carried out as previously described (Álvarez et al.,2008). Briefly, samples were homogenized and decontaminated

with 0.35% hexadecylpyridinium chloride for 30 min, centrifugedat 1068g and cultured on Coletsos and 0.2% (w/v) pyruvate-en-riched Löwenstein-Jensen media (BioMérieux, Madrid, Spain).Isolates were identified by Ziehl-Neelsen staining and PCR amplifi-cation of the Mycobacterium genus-specific 16S RNA fragment andMPB70 sequence (Wilton and Cousins, 1992). Isolates were charac-terized using spoligotyping (Kamerbeek et al., 1997).

2.9. Detection of infection with related pathogens

The presence of paratuberculosis or pseudotuberculosis in goatswas analyzed and compared with the outcome of diagnostic tests.Antibodies against M. avium subsp. paratuberculosis were detectedusing a commercial kit (Parachek�2, Prionics), which uses a serumadsorption step to remove antibodies that cross-react withMycobacterium phlei, a procedure that increases the specificity ofthe technique. The assay was performed according to the manufac-turer’s instructions and interpreted as prescribed by the kit insert. C.pseudotuberculosis was isolated from typically tuberculous lesionsdisclosed in lymph nodes, using Columbia blood agar(BioMérieux). The plates were incubated for 48 h at 37 �C in aerobicconditions. Further biochemical identification was achieved byusing the commercial API Coryne (version 2.0) system (BioMérieux).

2.10. Statistical analysis

Sensitivity of the different diagnostic techniques was calculatedusing as gold standard the isolation and/or microscopic detectionof Mycobacteria in the lesions. The agreement between the diag-nostic techniques was established by the Kappa index. The associ-ation between the results of the diagnostic techniques and thedifferent parameters studied was examined using the PearsonChi square test. Analyses were performed using SPSS 19.0 software(IBM, Armonk, NY, USA). A P-value of < 0.05 was considered to bestatistically significant.

3. Results

3.1. Assessment of tuberculosis status of the sacrificed goats

The combined results obtained from macroscopic, microscopicand/or isolation analysis showed that 101 of the sacrificed goatswere positive for tuberculosis (27 in herd A, 24 in herd B, 15 inherd C and 35 in herd D) while the other 61 goats were considerednegative for tuberculosis. An animal was considered positive fortuberculosis when M. caprae was isolated from the lesions and/orwas positive for Ziehl-Neelsen and immunohistochemistrydetection of Mycobacteria in the microscopic lesions. All M. capraeisolates obtained in the study (n = 58) belonged to the spoligotypeSB0157.

3.2. Sensitivity and Kappa index of the diagnostic techniques

The sensitivity (Table 1) of the skin test (44.5%, detecting 45 of101 goats with tuberculosis) was clearly lower than the other twotechniques, with the IFN-c release assay (65.3%, 66/101) and ELISA(66.3%, 67/101) showing similar results. When the IFN-c releaseassay or ELISA were considered as complementary tests performedin combination with the skin test, 90 of the 101 goats with tuber-culosis (89.1%) were positive in the skin test and/or ELISA, while 79of the 101 (78.2%) animals with tuberculosis were positive in theskin test and/or specific IFN-c release.

After analysis of the Kappa index (Table 1), the agreement be-tween skin test and the detection of tuberculosis in the animals(gold standard) was poor (j = 0.105) while the agreement between

Table 1Results of the diagnostic test in the 162 animal where the post-mortem analysis was performed.

Presence of tuberculosis Total Sensitivity (%) Kappa index

Negative Positive

Skin test Negative 41 56 97 44.5 0.105Positive 20 45 65

IFN-c release Negative 54 35 89 65.3 0.504Positive 7 66 73

ELISA Negative 54 34 88 66.3 0.493Positive 7 67 74

Total 61 101 162

1110 A.J. Buendía et al. / Research in Veterinary Science 95 (2013) 1107–1113

the IFN-release assay and ELISA with the presence of tuberculosiswas moderate (j = 0.393 and 0.504).

Considering the 498 animals on which diagnostic techniqueswere performed and not only the 162 goats where an accuratediagnosis of tuberculosis was obtained, the kappa index betweenthe techniques was the following: moderate (j = 0.425) betweenskin test and IFN-c release assay; poor ((j = 0.256) between skintest and ELISA; and moderate (j = 0.481) between IFN-c releaseassay and ELISA.

Only two goats showing tuberculosis were from the groupnegative for the three diagnostic techniques.

3.3. Relationship between macroscopic lesions and the results of thediagnostic techniques

After postmortem analysis and subsequent mycobacterial isola-tion, 57 goats with tuberculosis showed single lesions in the lungand/or regional lymph nodes, while 38 goats showed multipletuberculous lesions in the lung and other organs (mainly spleenand liver). Although they did not show macroscopic lesionsanother six goats were positive when isolation techniques wereused. The relationship between the presence of single or multiplemacroscopic lesions and the diagnostic techniques is shown inTable 2. Analysis of the results using the Pearson’s Chi-square testshowed a significant association (P < 0.001) between the type ofmacroscopic lesion and the results for ELISA in goats with tubercu-losis. In fact, the sensitivity of the technique decreased from 89.4%(34/38) in the case of goats with multiple lesions to 50.8% (29/57)in the goats with single lesions. Although there was no significantassociation, changes in sensitivity were also observed in the case ofskin test, the 49.1% (28/57) sensitivity in goats with single lesionsfalling to 34.2% (13/38) in animals with multiple lesions. In con-trast, no changes were observed in the results obtained with theIFN-c release assay, and sensitivity was very similar in both typesof lesions, with 64.9% (36/57) in the case of single lesions and 65.7%(25/38) in the case of multiple lesions.

Table 2Relationship between presence of macroscopic lesions in goats with tuberculosis(n = 95a) and the results of the diagnostic techniques using a Pearson’s Chi-squaretest.

Macroscopic lesions

Single Multiple Total P

Skin test Negative 29 25 54 0.191Positive 28 13 41

IFN-c release Negative 20 13 33 0.994Positive 37 25 62

ELISA Negative 28 4 32 <0.001Positive 29 34 63

Total 57 38 95

a Six animals with positive isolation showed no macroscopic lesions.

3.4. Relationship between microscopic lesions and the results of thediagnostic techniques

After histopathological analysis, microscopic lesions wereobserved in 96 animals (including the 95 goats with macroscopiclesions). Microscopic lesions could not be observed in the other fiveanimals which were positive for culture but lacked evident macro-scopic lesions. Among the animals with microscopic lesions, sevenof the 96 showed small granulomas (Fig. 1a), 34 goats showedlarge granulomas in single lesions (Fig. 1b), 18 goats showed largegranulomas in multiple lesions, 20 animals showed cavitarylesions (Fig. 1c) and finally, 17 showed fibrocalcified lesions.(Fig. 1d). The relationship between the type of lesion and the re-sults of the diagnostic techniques is shown in Table 3. Statisticalanalysis demonstrated that the ELISA results were highly associ-ated with the type of lesion (P = 0.002), sensitivity being low inthe case of large granulomas in single lesions (41.2%; 14/34) andincreasing in animals with cavitary lesions (75.0%; 15/20), fibrocal-cified lesions (82.4%; 14/17) and, especially, in goats with multiplelarge granulomas (94.4%, 17/18). The relationship between the skintest and type of lesions was not statistically significant (P = 0.068).The sensitivity of the technique fell dramatically in animals withcavitary lesions (20.0%; 4/20). The type of lesions did not signifi-cantly affect the results of the IFN-c release assay, although thesensitivity was higher in animals with large granulomas, whetherin single (73.5%; 25/34) or multiple (77.7%; 14/18) lesions, thanin animals with cavitary (60.0%; 12/20) or fibrocalcified (41.2%;7/17) lesions. The two animals with tuberculosis but negative tothe three diagnostic techniques showed cavitary lesions.

3.5. Relationship between the results of the diagnostic techniques andthe presence of pseudotuberculosis

Results from the detection of pseudotuberculosis showed 22cases with lesions that were culture positive for C. pseudotubercu-losis, indicating a frequency of 13.5% in this study. There werepositive cases in all four herds studied, with between 2 and 11cases among the sacrificed goats per herd. Eleven cases were mixedpseudotuberculosis-tuberculosis infections (10.8% of the goatswith tuberculosis) while the other 11 cases were among the 61goats testing negative for tuberculosis. The presence of pseudotu-berculosis did not seem to affect the sensitivity of the diagnostictechniques for tuberculosis.

3.6. Relationship between the results of the diagnostic techniques andthe serological detection of paratuberculosis

The results obtained with the Parachek�2 kit showed that 42goats were positive using this technique, indicating a frequencyof 25.9%. All four herds had positive animals, with between fourand 15 among the sacrificed goats per herd. Fourteen animals wereParachek positive and were tuberculosis negative, while the other28 animals were positive for tuberculosis (27.7% of the goats with

Fig. 1. Representative images of the type of microscopic lesions described in the study. (a) Small granuloma in the lung of a goat with tuberculosis. Note the minimal area ofnecrosis (asterisk), the lack of a connective tissue capsule and the presence of granulation tissue surrounding the lesion (arrows). (b) Large granuloma in the mediastinallymph node of a goat with tuberculosis. Note the large area of necrosis (asteriks) surrounded by epithelioid cells (arrows). (c) Cavitary lesion in the lung of a goat withtuberculosis. Note the presence of an area with numerous neutrophils (asterisk) surrounded by macrophages (arowheads) and granulation tissue (arrows). (d) Fibrocacifiedlesion in the tracheobronchial lymph node of a goat with tuberculosis. Note the central calcified core (asterisk) surrounded by collageonus tissue containing epithelioid andmultinucleated giant cells (arrows). Haematoxylin-eosin staining.

Table 3Relationship between presence of microscopic lesions in goats with tuberculosis(n = 101) and the results of the diagnostic techniques using a Pearson’s Chi-squaretest.

Microscopic lesions

NL SG LGs LGm CL FC Total P

Skin test Negative 1 2 18 11 16 8 56 0.068Positive 4 5 16 7 4 9 45

IFN-c release Negative 1 3 9 4 8 10 35 0.243Positive 4 4 25 14 12 7 66

ELISA Negative 2 3 20 1 5 3 34 0.002Positive 3 4 14 17 15 14 67

Total 5 7 34 18 20 17 101

NL, animals with positive isolation but not detectable microscopic lesions; SG,smallgranulomas; LGs, large granulomas in single lesions; LGm, large granulomas inmultiple lesions; CL, cavitary lesions; FC, fibro-calcified lesions.

Table 4Relationship between serological results for paratuberculosis in goats with tubercu-losis (n = 101) and the results of the diagnostic techniques using a Pearson’s Chi-square test.

Paracheck�2 results

Negative Positive Total P

Skin test Negative 38 18 56 0.268Positive 35 10 45

IFN-c release Negative 21 14 35 0.045Positive 52 14 66

ELISA Negative 26 8 34 0.502Positive 47 20 67

Total 73 28 101

A.J. Buendía et al. / Research in Veterinary Science 95 (2013) 1107–1113 1111

tuberculosis). The relationship between the serological results forparatuberculosis in animals with tuberculosis and the results ofthe diagnostic techniques is shown in Table 4. The results showed

a significant increase (P = 0.045) in the number of animals serolog-ically positive for paratuberculosis among the false negativesobserved for the IFN-c release assay, 14/35 (40.0%) in contrast withthe overall percentage of paratuberculosis in animals withtuberculosis 28/101 (27.7%).

Five goats of the 162 studied were positive for the three infec-tions, confirming the presence of a percentage (3.1% in our study)of tuberculosis-paratuberculosis-pseudotuberculosis mixed infec-tions. As regards the results obtained with the different diagnostictests, ELISA detected three of five animals while the skin test andIFN-c release assay only detected one of five goats; each triple-infected animal was positive in only one of the diagnostic tests.

4. Discussion

The data obtained in the present study confirm the poor resultsoffered by the comparative skin test for diagnosing tuberculosis incaprine herds with a long history of persistent infection with M.caprae. The other two techniques studied provided significantlybetter results, with a higher level of sensitivity than the skin test.In addition, the sensitivity of the comparative skin test was partic-ularly low in animals that had cavitary lesions. This is of someconcern, since these animals have the potential to disseminateinfection and disease. The specific IFN-c release assay and espe-cially the ELISA, clearly showed better results than the comparativeskin test in animals with this type of lesion. The only group wherethe comparative skin test showed better sensitivity that the othertechniques was the group with early lesions composed of smallgranulomas. However, this group was made up from a compara-tively small number of animals. The sensitivity of the comparativeskin test in the present study was lower than in other studies usingthe same technique in cattle, where 52–100% sensitivity wasobserved (reviewed in De la Rua-Domenech et al., 2006). In caprine

1112 A.J. Buendía et al. / Research in Veterinary Science 95 (2013) 1107–1113

herds, Gutiérrez et al. (1998) reported a sensitivity of 83.7% in astudy with 74 goats from herds located in an area without an erad-ication programme. However, lower sensitivity (below 50%) fordetection of tuberculosis in goats has been described in other stud-ies in caprine herds based on the comparative skin test (Liebanaet al., 1998; Álvarez et al., 2008; Bezos et al., 2011). This lowsensitivity, mainly in animals with severe lesions, observed inour four herd study, tested repeatedly by skin test over severalyears, could be associated with the desensitization described re-cently in cattle by Coad et al. (2010) as a result of repeated testingwith approximately 60-day intervals. These authors reported thatthe significant desensitizing effect was associated with decreasedIL-1 beta and an increased IL-10 response. This fact was observedin naturally infected animals after the third short-interval skin testand increased in the fourth and final skin test (Coad et al., 2010).Repeated skin tests have been demonstrated to have no effect onthe results of specific IFN-c production in cattle that were infectedboth experimentally and naturally (reviewed in Schiller et al.,2010). This might explain the differences observed in the sensitiv-ity between the skin test and IFN-c release assay in our study.Another aspect that could be related to the lack of sensitivity ofthe skin test is the presence of animals with cavitary lesions.Previous studies (Sanchez et al., 2011) showed a lower CD4+ T cellresponse associated with this type of lesion compared to granu-lomatous lesions. This low response could lead to a decrease inthe sensitivity of the test; in fact, only four of 20 goats with thistype of histopathological lesion in our study were detected usingthe skin test.

To analyze properly the results obtained with the skin test wehave to take into account the number of animals where it wasnot possible to establish a positive diagnosis for tuberculosis usingthe gold standard. In the present study have been found 61 animalswith negative diagnosis for tuberculosis after the histopathologyand isolation. Thirty-four of these animals belong to the group withnegative results for the three techniques, which implies that theseanimals probably were not infected. However other 27 goats werepositives for at least one of the diagnostic test, most of them (20 of27 were positive for the skin test), and taking into account thepresence of M. caprae in the herds and the relatively high sensitiv-ity of the skin test in cases with no lesions or small granulomas de-scribed here, it cannot be ruled out that a number of these animalswithout macroscopic lesions nor isolation were actually infectedwith M. caprae. In that case, we should recognize the limitationsto detect the infection of our gold standard and the possibility ofhaving underestimated the sensitivity of the skin test.

The sensitivity for the specific IFN-c release assay obtained inthe present study, although clearly better than those obtained withthe comparative skin test, were lower than in studies carried out innaturally infected cattle (Gormley et al., 2006), where sensitivityreached 88%. In the case of caprine herds, several studies (Gutiérrezet al., 1998; Liebana et al., 1998) have obtained high sensitivitiesabove 80% with this technique. Another more recent caprine study,with tuberculosis-paratuberculosis mixed infection, showed lowersensitivity similar to that displayed in our study (Álvarez et al.,2008). The results of the present study point to a significant asso-ciation between the presence of false negatives for the IFN-crelease assay and a serological diagnosis for M. avium subs.paratuberculosis infection. Interference in diagnosis caused by thepresence of paratuberculosis has been reported both in naturallyand experimentally infected goats (Bezos et al., 2012). Such inter-ference could be related to the production of anti-inflammatorycytokines induced by the infection, as described by Lybeck et al.(2013), in goats naturally infected with M. avium subs. paratubercu-losis. Due to the high specificity but low sensitivity of the serologicdetection of paratuberculosis (Collins et al., 2005), the number ofanimals infected with paratuberculosis has very probably been

underestimated, and the influence of mixed infection on tuberculo-sis diagnosis could be even greater.

In the field conditions of the herds studied (i.e. eradication pro-gramme, repeated skin tests and mixed infections) the ELISA basedon MPB-70 gave the higher sensitivity (66.3%) of the diagnostictechniques used. A boosting effect induced by the skin test, usingbovine PPD, on the antibody response to the protein MPB70 hasbeen described in cattle (Lightbody et al., 2000). This may havecontributed to the results observed in our study. In addition, theELISA identified a larger number of goats with more developedforms of the illness, for example, 89.4% sensitivity in animals withmultiple macroscopic lesions and 75.0% in goats with histopathol-ogical cavitary lesions. Animals with these forms of the disease aremore likely to excrete mycobacteria and spread them within theherd. The sensitivity of ELISA was clearly lower in goats with singleand granulomatous lesions, where the cell immunity-basedtechniques, especially IFN-c release assay, were more efficient.

In conclusion, our study demonstrated the relatively poor re-sults using the comparative skin test in herds with a long historyof tuberculosis. It also highlighted the need for an ancillary testto improve detection of infected animals. The relatively superiorresults of the ELISA for the detection of advanced forms of tubercu-losis and the complementarity observed when used with the com-parative skin test (with better results in early infections), make theMPB70 ELISA a suitable ancillary technique to improve eradicationprogrammes. After years in which the use of serological tests forthe diagnosis of tuberculosis in farm animals has been undervaluedcompared with tests to determine cellular immunity, the detectionof specific antibodies has been demonstrated by this and otherstudies, to be a useful diagnostic tool for tuberculosis in cattle,(Whelan et al., 2010, 2011; Clegg et al., 2011; Waters et al.,2011) and goats (Marassi et al., 2009; Shuralev et al., 2012).Furthermore, its simplicity and low price make it a suitable andpreferable option in the field of caprine health, where the costsare a fundamental aspect in the design of eradication programmes.

Conflict of interest

None of the authors of this paper have a financial or personalrelationship with other people or organization that could inappro-priately influence or bias the content of the paper.

Acknowledgments

We thank Javier Bezos from VISAVET for critical reading of thepaper and Pedro Gil and Juan Sanchez for execellent technicalassistance.

The study was supported by Fundación SENECA, Grant code11915/PI/09.

References

Álvarez, J., de Juan, L., Bezos, J., Romero, B., Sáez, J.L., Reviriego-Gordejo, F.J., Briones,V., Moreno, M.A., Mateos, A., Dominguez, L., Aranaz, A., 2008. Interference ofparatuberculosis with the diagnosis of tuberculosis in a goat flock with a naturalmixed infection. Veterinary Microbiology 128, 72–80.

Aranaz, A., Cousins, D., Mateos, A., Dominguez, L., 2003. Elevation of Mycobacteriumtuberculosis subs. caprae Aranaz et al. 1999 to species rank as mycobacteriumcomb. Nov., sp. Nov. International Journal of Systematic and EvolutionaryMicrobiology 53, 1785–1789.

Bezos, J., Alvarez, J., de Juan, L., Romero, B., Rodriguez, S., Fernandez-de-Mera, I.G.,Hewinson, R.G., Vordermeier, M., Mateos, A., Dominguez, L., Aranaz, A., 2011.Assesment of in vivo and in vitro tuberculosis diagnostic tests in Mycobacteriumcaprae naturally infected caprine flocks. Preventive Veterinary Medicine 100,187–192.

Bezos, J., Alvarez, J., Romero, B., Aranaz, A., de Juan, L., 2012. Tuberculosis in goats:assessment of current in vivo cell-mediated and antibody-based diagnosticassays. Veterinary Journal 191, 161–165.

Caro, M.R., Gallego, M.C., Buendia, A.J., Del Rio, L., Seva, J., Navarro, J.A., 2001.Differences in lymphocyte subpopulations from peripheral blood and lymphoid

A.J. Buendía et al. / Research in Veterinary Science 95 (2013) 1107–1113 1113

organs in natural caprine tuberculosis infection. Journal of Veterinary MedicineB Infectious Diseases and Veterinary Public Health 48, 81–88.

Casal, C., Bezos, J., Diez-Guerrier, A., Alvarez, J., Romero, B., de Juan, L., Rodriguez-Campos, S., Vordermeier, M., Whelan, A., Hewinson, R.G., Mateos, A.,Dominguez, L., Aranaz, A., 2012. Evaluation of two cocktails containing ESAT-6, CFP-10 and Rv-615c in the intradermal test and the interferon-c assay fordiagnosis of bovine tuberculosis. Preventive Veterinary Medicine 105, 149–154.

Clegg, T.A., Duignan, A., Whelan, C., Gormley, E., Good, M., Clarke, J., Toft, N., More,S.J., 2011. Using latent class analysis to estimate the test characteristics of the c-interferon test and a multiplex immunoassay under Irish conditions. VeterinaryMicrobiology 151, 68–76.

Coad, M., Clifford, D., Rhodes, S.G., Hewinson, R.G., Vordermeier, H.M., Whelan, A.O.,2010. Repeat tuberculin skin testing leads to desensitisation in naturallyinfected tuberculous cattle which is associated with elevated interleukin-10and decreased interleukin-1 beta response. Veterinary Research 41, 14.

Collins, M.T., Wells, S.J., Petrini, K.R., Collins, J.E., Schultz, R.D., Whitlock, R.H., 2005.Evaluation of five antibody detection tests for diagnosis of bovineparatuberculosis. Clinical and Diagnostic Laboratory Immunology 12, 685–692.

Daniel, R., Evans, H., Rolfe, S., De la Rua-Domenech, R., Crawshaw, T., Higginss, R.J.,Schock, A., Clifton-Hadley, R., 2009. Outbreak of tuberculosis caused byMycobacterium bovis in golden Guernsey goats in Great Britain. VeterinaryRecord 165, 335–342.

De la Rua-Domenech, R., Goodchild, A.T., Vordermeir, H.M., Hewinson, R.G.,Christiansen, K.H., Clifton-Hadley, R.S., 2006. Ante mortem diagnosis oftuberculosis in cattle: a review of the tuberculin tests, c-interferon assay andother ancillary diagnostic techniques. Research in Veterinary Science 81, 190–210.

Gormley, E., Doyle, M.B., Fitzimons, T., McGill, K., Collins, J.D., 2006. Diagnosis ofMycobacterium infection in cattle by use of the gamma-interferon (Bovigam�)assay. Veterinary Microbiology 112, 171–179.

Gutiérrez, M., Tellechea, J., García Marín, J.F., 1998. Evaluation of cellular andserological diagnostic tests for the detection of Mycobacterium bovis-infectedgoats. Veterinary Microbiology 62, 281–290.

Kamerbeek, J., Schouls, L., Kolk, A., Van Agterveld, M., van Soolingen, D., Kuijper, S.,Bunschoten, A., Molhuizen, H., Shaw, R., Goyal, M., van Embden, J., 1997.Simultaneous detection and strain differentiation of Mycobacterium tuberculosisfor diagnosis and epidemiology. Journal of Clinical Microbiology 35, 907–914.

Liebana, E., Aranaz, A., Urquia, J.J., Mateos, A., Dominguez, L., 1998. Evaluation of thegamma-interferon assay for eradication of tuberculosis in a goat herd.Australian Veterinary Journal 76, 50–53.

Lightbody, K.A., Skuce, R.A., Neill, S.D., Pollock, J.M., 1998. Mycobacterial antigen-specific antibody responses in bovine tuberculosis: an ELISA with potential toconfirm disease status. Veterinary Record 142, 295–300.

Lightbody, K.A., McNair, J., Neill, S.D., Pollock, J.M., 2000. IgG isotype antibodyresponses to epitopes of the Mycobacterium bovis protein MPB70 in immunisedand in tuberculin skin test-reactor cattle. Veterinary Microbiology 75, 177–188.

Lybeck, K.R., Lovoll, M., Johansen, T.B., Olse, L., Storset, A.K., Valheim, M., 2013.Intestinal strictures, fibrous adhesion and high local interleukin-10 levels ingoats infected naturally with Mycobacterium avium subsp. paratuberculosis.Journal of Comparative Pathology 148, 157–172.

Marassi, C., Almeida, C., Pinheiro, S., Vasconcellos, S., Lilenbaum, W., 2009. The useof MPB70-ELISA for the diagnosis of caprine tuberculosis in Brazil. VeterinaryResearch Communications 33, 937–943.

Pollock, J.M., McNair, J., Welsh, M.D., Girvin, R.M., Kennedy, H.E., Mackie, D.P., Neill,S.D., 2001. Immune responses in bovine tuberculosis. Tuberculosis 81, 103–107.

Salinas, J., Caro, M.R., Vicente, J., Cuello, F., Reyes-García, A.R., Buendia, A.J.,Rodolakis, A., Gortazar, C., 2009. High prevalence of antibodies againstChlamydiaceae and Chlamydophila abortus in wild ungulates using two ‘‘inhouse’’ blocking-ELISA tests. Veterinary Microbiology 135, 46–53.

Sanchez, J., Tomas, L., Ortega, N., Buendia, A.J., del Rio, L., Salinas, J., Bezos, J., Caro,M.R., Navarro, J.A., 2011. Microscopical and immunological features oftuberculoid granulomata and cavitary pulmonary tuberculosis in naturallyinfected goats. Journal of Comparative Pathology 145, 107–117.

Schiller, I., Vordermeier, H.M., Waters, W.R., Whelan, A.O., Coad, M., Gormley, E.,Buddle, B.M., Palmer, M., Thacker, T., McNair, J., Welsh, M., Hewinson, R.G.,Oesch, B., 2010. Bovine tuberculosis: Effect of the tuberculin skin test on in vitrointerferon gamma responses. Veterinary Immunology Immunopathology 136,1–11.

Seva, J., Menchen, V., Navarro, J.A., Pallares, F.J., Villar, D., Vasquez, F., Bernabe, A.,2002. Caprine tuberculosis eradication program: an immunohistochemicalstudy. Small Ruminant Research 46, 107–114.

Sharpe, A.E., Brady, C.P., Johnson, A., Byrne, W., Kenny, K., Costello, E., 2010.Concurrent outbreak of tuberculosis and caseous lymphadenitis in a goat herd.Veterinary Record 166, 591–592.

Shuralev, E., Quinn, P., Doyle, M., Duignan, A., Kwok, H.F., Bezos, J., Olwill, S.A.,Gormley, E., Aranaz, A., Good, M., Davis, W.C., Clarke, J., Whelan, C., 2012.Application of the Enfer chemiluminescent multiplex ELISA system for thedetection of Mycobacterium bovis infection in goats. Veterinary Microbiology154, 292–297.

Vordermeier, H.M., Chambers, M.A., Cokle, P.J., Whelan, A.O., Simmons, J., Hewinson,R.G., 2002. Correlation of ESAT-6-specific gamma interferon production withpathology in cattle following Mycobacterium bovis BCG vaccination againstexperimental bovine tuberculosis. Infection and Immunity 70, 3026–3032.

Waters, W.R., Buddle, B.M., Vordermeier, H.M., Gormley, E., Palmer, M.V., Thacker,T.C., Bannantine, J.P., Stabel, J.R., Linscott, R., Martel, E., Milian, F., Foshaug, W.,Lawrence, J.C., 2011. Development and evaluation of an enzyme-linkedimmunosorbent assay for use in the detection of bovine tuberculosis in cattle.Clinical and Vaccine Immunology 18, 1882–1888.

Whelan, C., Whelan, A.O., Shuralev, E., Kwok, H.F., Hewinson, G., Clarke, J.,Vordermeier, H.M., 2010. Performance of the enferplex TB assay with cattle inGreat Britain and assessment of its suitability as a test to distinguish infectedand vaccinated animals. Clinical and Vaccine Immunology 17, 813–817.

Whelan, C., Shuralev, E., Kwok, H.F., Kenny, K., Duignan, A., Good, M., Davis, W.C.,Clarke, J., 2011. Use of a multiplex enzyme-linked immunosorbent assay todetect a subpopulation of Mycobacterium bovis-infected animals deemednegative or inconclusive by the single intradermal comparative tuberculinskin test. Journal of Veterinary Diagnostic Investigation 23, 499–503.

Wilton, S., Cousins, D., 1992. Detection and identification of multiple mycobacterialpathogens by DNA amplification in a single tube. PCR Methods and Applications1, 269–273.