Meta-analysis of field studies on bovine tuberculosis skin tests in United States cattle herds
Introduction
Since the early 20th century, diagnostic capabilities for bovine tuberculosis (bTb) have challenged Tb eradication programs worldwide. In part this is due to inherent difficulties in detecting disease caused by Mycobacterium species (notably bTb and Johne's disease) with a high degree of accuracy stemming from less than ideal diagnostic test capabilities. Diagnostic tests are evaluated on the basis of sensitivity (Se), specificity (Sp), positive and negative predictive values, and when established these parameters provide test users an appropriate degree of confidence in the test result. Despite nearly a century of bTb testing in the United States, reference estimates for test Se and Sp of tuberculin skin tests used for ante-mortem diagnostics have not been firmly established. The current U.S. bTb eradication program utilizes individual animal tuberculin skin tests as a herd level indicator for disease, requiring a more comprehensive interpretation of variability in test parameters, test results and their relation to disease status. Overall the lack of rapid, sensitive, and specific bTb diagnostic tests for live cattle further compounds an already complex goal of bTb eradication.
To date, the tuberculin skin tests and gamma-interferon blood test for cell-mediated immunity remain the only live animal diagnostic tools for bTb approved for use in the United States. Skin tests currently used in bovines include; the caudal fold tuberculin (CFT) test, the cervical tuberculin (CT) test and the comparative cervical tuberculin (CCT) test, all utilizing delayed-type hypersensitivity (DTH) cell-mediated immune response to either Mycobacterium bovis, or a combination of M. bovis and Mycobacterium avium purified protein derivatives to indicate infection with bTb. The current U.S. bTb eradication strategy utilizes the CFT or CT test as an initial live animal screening test, with a follow-up CCT test (or in some cases gamma-interferon blood test) as a confirmatory test for animals with a positive response to the initial CFT test (USDA:APHIS, 2004).
For the CFT test in an individual, any visual or palpable inflammation at the injection site 64–78 h following tuberculin administration is considered a positive test response, and the animal is designated a “CFT reactor” (USDA:APHIS, 2004, Thoen et al., 2006).
The U.S. bTb eradication program utilizes the aggregated results from CFT testing of cattle within the herd as a “herd level” indicator of disease. Herd-level test sensitivity is defined as the probability that a positive herd yields a positive herd-test result, and correspondingly, herd-level test specificity as the probability that a negative herd yields a negative herd-test result (Martin et al., 1992). Herd-level test sensitivity and specificity (HSe and HSp, respectively) are different than test Se and Sp in the individual animal due to influence of within-herd disease prevalence, sample size and the herd reactor cutoff value (e.g. 1 or more test positive) used for the test (Christensen and Gardner, 2000, Jordan and McEwen, 1998, Norby et al., 2005).
To increase overall test Sp, reduce the proportion of false positives, and remove the burden of declaring infected herds from individual veterinarians, the USDA adopted the CCT test as a follow-up test for CFT responders in 1973. For the CCT test, skin fold thickness for both M. bovis and M. avium tuberculin injection sites is measured before and after PPD administration, and the difference in skin fold thickness for each injection site is then plotted on a scatter-gram (Roswurm and Konyha, 1973, Schiller et al., 2010, USDA:APHIS, 2004). Crudely put, animals which mount a stronger immune response to the M. bovis PPD versus M. avium are termed CCT reactors, and are considered positive for bTb.
Developing a more complete understanding on the accuracy of bTb tuberculin skin tests and establishing their most effective use in eradicating the disease comprise the primary rationale of this study. Our approach was to conduct a systematic review and meta-analysis of available data and to estimate herd-level skin test parameters.
Section snippets
Literature search
The first author (MF) searched the PubMed and CAB electronic indexes on 10 October 2008. Initial search terms for PubMed were MeSH terms “tuberculosis, bovine” OR “mycobacterium bovis” AND “diagnosis”. For the CAB index search terms were modified to comply with different syntax rules. PubMed index was searched again on 14 September 2009 and 6 August 2011 with additional MeSH term “sensitivity and specificity” (full search string “tuberculosis, bovine”[MeSH Terms] OR “mycobacterium bovis”[MeSH
Literature review and meta-analysis
Twenty-five point estimates of bTb tuberculin test Se and Sp in ten publications were identified (Table 1, Table 2). Two publications (eight point estimates) were excluded from further consideration because they did not meet inclusion criteria (Francis et al., 1978, USDA:APHIS, 1992), and five were excluded because they were calculated either from re-tests on animals already tested (Roswurm and Konyha, 1973), or from a sub-set of a larger sample population (Norby et al., 2004). In total, twelve
Discussion
For any given diagnostic test, test Se and Sp are best thought of as a distribution of values, which can be utilized to provide an accurate estimate of test performance in comparison to true disease status. The data reported here are intended to help establish an appropriate range of Se, Sp, HSe and HSp for both the CFT test and serial interpretation of CFT-CCT tests conducted in North American cattle, based on published literature.
The limited number of field studies used to validate bTb skin
Acknowledgements
The authors would like to acknowledge Dr. Barbara Knust, Epidemiology Intelligence Officer with the U.S. Centers for Disease Control and Prevention, Dr. Susan McClanahan, Senior Veterinarian with the Minnesota Board of Animal Health, and Dr. Katie Portacci, Senior Risk Analyst with USDA-APHIS-Veterinary Services, for technical assistance and critique of the manuscript.
The study was funded in part through the 2008 Minnesota Rapid Agriculture Response Fund.
References (21)
- et al.
A new probability formula for surveys to substantiate freedom from disease
Prev. Vet. Med.
(1998) - et al.
Herd-level interpretation of test results for epidemiologic studies of animal diseases
Prev. Vet. Med.
(2000) - et al.
Ante mortem diagnosis of tuberculosis in cattle: a review of the tuberculin tests, gamma-interferon assay and other ancillary diagnostic techniques
Res. Vet. Sci.
(2006) - et al.
Estimation of sensitivity and specificity of diagnostic tests and disease prevalence when the true disease state is unknown
Prev. Vet. Med.
(2000) - et al.
Herd-level test performance based on uncertain estimates of individual test performance, individual true prevalence and herd true prevalence
Prev. Vet. Med.
(1998) - et al.
Evaluating the health status of herds based on tests applied to individuals
Prev. Vet. Med.
(1992) Status of the state and federal cooperative bovine tuberculosis (TB) eradication program fiscal year 2006
Proc. Annu. Meet. U.S. Anim. Health Assoc.
(2006)- et al.
The sensitivity and specificity of various tuberculin tests using bovine PPD and other tuberculins
Vet. Rec.
(1978) - et al.
Evaluation of a five-antigen ELISA for diagnosis of tuberculosis in cattle and cervidae
J. Am. Vet. Med. Assoc.
(1996) Annual update for the state and federal cooperative bovine tuberculosis (TB) eradication program fiscal year 2010
Proc. Annu. Meet. U.S. Anim. Health Assoc.
(2010)