Validation of the BrockTB Stat-Pak Assay for Detection of Tuberculosis in Eurasian Badgers (Meles meles) and Influence of Disease Severity on Diagnostic Accuracy

Abstract

A lateral-flow immunoassay (BrockTB Stat-Pak) for detecting tuberculosis in Eurasian badgers was 49% sensitive and 93% specific against culture for M. bovis (n = 1,464) at necropsy. However, the sensitivity was significantly higher (66 to 78%) in animals with more severe tuberculosis, indicating that the BrockTB Stat-Pak may be useful for the detection of badgers with the greatest risk of transmitting disease.

Despite attempts to control bovine tuberculosis (TB), the average incidence of disease in cattle in Great Britain increased an estimated 225% between the years 1996 and 2006 (7), causing considerable economic loss to the government and the farming community. In parts of Great Britain and Ireland, Eurasian badger (Meles meles) populations constitute a reservoir of infection with Mycobacterium bovis and a potential source of infection to cattle (8, 14, 15). In an effort to simplify the serodetection of badger TB, we developed a lateral-flow immunoassay (13) that is now manufactured as the BrockTB Stat-Pak assay (Chembio Diagnostic Systems, Inc., Medford, NY). We report here the results of ongoing evaluation of the assay, in particular the influence of disease severity on diagnostic accuracy.

Badger sera were obtained in Great Britain from two sources and stored frozen at −20°C until testing (i) 1,464 animals killed as part of the defra/Independent Scientific Group Randomized Badger Culling Trial (RBCT) (8) and (ii) 68 animals captured, sampled, and then released as part of an ongoing ecological and epidemiological study in Woodchester Park (WP), southwest England (6). Each badger from the RBCT was subjected to routine necropsy examination and culture for the presence of M. bovis. Carcasses were stored at 4°C and examined by a standardized necropsy protocol within 3 days of submission. A range of tissues and lymph nodes were examined for evidence of TB. In badgers with no lesions the bronchial, mediastinal, and retropharyngeal lymph nodes were placed in 1% cetyl-pyridinium chloride prior to bacterial culture. In badgers with lesions the lesion tissue was always submitted for culture. Bite wounds were sampled separately. Tissues were processed within 72 h of sampling. The cetyl-pyridinium chloride was discarded, and the tissues were rinsed in saline and then emulsified before inoculation onto six slopes of modified Middlebrook 7H11 agar. The slopes were incubated at 37 ± 2°C. Any growth of organisms characteristic of mycobacteria was identified by spoligotyping (9). The infection status of WP badgers was determined by bacterial culture of clinical samples of feces, urine, sputum, pus from abscesses, and bite wound swabs as described previously (3).

Compared to the culture results from tissues taken at necropsy (n = 1,464), the findings were as follows: BrockTB Stat-Pak positive, culture negative, n = 74; BrockTB Stat-Pak negative, culture negative, n = 1,004; BrockTB Stat-Pak positive, culture positive, n = 190; and BrockTB Stat-Pak negative, culture positive, n = 196. Thus, the BrockTB Stat-Pak had a sensitivity for detection of TB of 49.2% (95% confidence interval [CI] = 44.1 to 54.4%) and a specificity of 93.1% (95% CI = 91.4 to 94.6%). These figures are within the ranges previously reported by us with a smaller sample size (13) and are equivalent to the performance of the existing serum enzyme-linked immunosorbent assay-based test for TB (Brock Test) (2, 5, 12). Although the specificity was relatively high, this is likely to be an underestimate. The badgers were sampled from parts of the country where TB is endemic, and it is likely that true infection was missed in some cases despite the use of an extensive postmortem culture protocol.

During the present study, samples that were hemolyzed and/or lipemic were identified prior to testing. Although previous work showed that the condition of these samples did not influence the Brock Test (unpublished data), the potential effect on the performance of the BrockTB Stat-Pak was unknown. A total of 216 RBCT serum samples from the 1,464 submitted to testing (14.8%) were either hemolyzed or lipemic. Compared to “normal” samples, those that were hemolyzed and/or lipemic exhibited a statistically significant reduction in sensitivity (P = 0.003, Fisher exact test), but no difference in specificity (P = 0.613, Fisher exact test) (Table 1).

TABLE 1.

Influence of blood quality of the performance of the BrockTB STAT-PAK based on 1,464 RBCT badgers

Serum	No. of isolatesa					Sensitivity or specificity (95% CI)
	TP	TN	FP	FN	Total	Sensitivity	Specificity
Normal	173	850	65	160	1,248	52.0 (46.5-57.4)	92.9 (91.0-94.5)
Hem/lipb	17	154	9	36	216	32.1 (19.9-46.3)	94.5 (89.8-97.5)

^aTP, true positive; TN, true negative. FP, “false positive,” i.e., the number of positive BrockTB Stat-Pak tests in culture-negative badgers. FN, “false” negative, i.e., the number of negative BrockTB Stat-Pak tests in culture-positive badgers.

^bHem/lip, hemolyzed and/or lipemic.

Subdivision of culture-positive badgers based on criteria associated with disease severity revealed an increased ability of the BrockTB Stat-Pak to detect infected badgers with more severe (progressed, disseminated) disease (Table 2). By inference, these individuals may pose a greater risk of infecting other badgers and cattle through repeated and/or greater excretion of M. bovis (4, 10). However, it must be borne in mind that the relationship between pathology in the badger and the likelihood of onward transmission of TB is unknown. Two criteria were used to identify such animals: either their “excretor status” (for WP badgers, which were live sampled and then released) or presentation of gross lesions at necropsy (for RBCT badgers). The BrockTB Stat-Pak detected 25 of 32 so-called “super-excretor” WP badgers, representing a sensitivity of 78.1% (95% CI = 60.0 to 90.7%). An animal was classified as a “super-excretor” after its second consecutive culture positive result or after it had provided two positive culture results from different samples (e.g., feces and urine) on its last sampling occasion (6). In addition, “super-excretors” have a higher mortality rate (17) and so probably represent animals with progressive TB that does not resolve. In contrast, “excretors” exhibit intermittent shedding or may apparently cease bacterial excretion altogether (6). The BrockTB Stat-Pak detected 15 of 36 “excretor” badgers, representing a sensitivity of 41.7% (95% CI = 25.5 to 59.3%). The difference in the sensitivity of TB detection between “super-excretor” and ‘excretor’ badgers was highly significant (P = 0.003, Fisher exact test), a finding consistent with our previous observation that seropositivity in badgers was associated with more severe disease (1). Similarly, the BrockTB Stat-Pak detected 66.1% of culture-positive badgers with visible lesions suspicious of TB at necropsy (VL) compared to only 34.4% of those with no suspect TB lesions at necropsy (NVL) (Table 2). The difference in sensitivity was again highly significant (P < 0.0001, Fisher exact test). We have recently demonstrated improved detection rates of TB in the badger using quantitative real-time PCR (16) and an enzyme immunoassay (5) to detect the antigen-specific production of gamma interferon (IFN-γ). For both the IFN-γ enzyme immunoassay and the Brock Test, the sensitivity was also significantly higher in badgers with VL than in those with NVL (5). However, the Brock Test and both IFN-γ tests require specialist laboratory facilities and equipment and take 3 and 48 h, respectively, to complete. In contrast, the BrockTB Stat-Pak can be performed anywhere and produces a result in no more than 20 min.

TABLE 2.

Sensitivity of the BrockTB Stat-Pak in tuberculous badgers with different levels of disease severity

BrockTB Stat-Pak parameter	Based on necropsy examination		Based on live sampling
	NVL	VL	“Excretor”	“Super-excretor”
No. of positive tests	72	117	15	25
No. of negative tests	137	60	21	7
Sensitivity (%)	34.4	66.1	41.7	78.1
95% CI	28.0-41.4	58.6-73.1	25.5-59.3	60.0-90.7

The BrockTB Stat-Pak had the lowest sensitivity in culture-positive NVL badgers, which would be consistent with the suggestion that this group is proportionally less likely to be excreting M. bovis (10). Moreover, not all VL badgers will excrete M. bovis, since this will depend on the anatomical location of the lesion, its structure, and the quantity of bacteria contained therein (10, 11). This is consistent with the lower sensitivity of the BrockTB Stat-Pak for VL badgers than that seen for “super-excretors.” Overall, these data are consistent with a positive BrockTB Stat-Pak test result being more frequently associated with advanced TB, especially that which gives rise to more extensive or persistent shedding of bacteria.

In summary, the BrockTB Stat-Pak is an easily executed and rapid test for the detection of TB in badgers. The BrockTB Stat-Pak could be performed directly in the field, although consideration should be given to the quality of the blood sample so that test sensitivity is not adversely affected. With a sensitivity of 49% compared to the “gold standard” of necropsy tissue culture, the BrockTB Stat-Pak is unlikely to be sufficiently sensitive for routine TB surveillance. However, the sensitivity was significantly higher in animals with more severe TB, classified by more frequent excretion of M. bovis or the presence of visible lesions at necropsy. This raises the possibility that the BrockTB Stat-Pak could be of use where a simple tool is required to detect badgers more likely to be at advanced stages of disease.