Challenges with the official serological tests for bovine brucellosis in South Africa and the way forward

J.G.K Kangumba

Laboratory Services and Epidemiology, Directorate Veterinary Services,

North West Province, South Africa

Summary

The Compliment Fixation test (CFT) was introduced in 1909 and was widely used based on the suggestions that it was highly sensitive and specific. In South Africa it is used with the Rose Bengal test (RBT) as official tests for the diagnosis of bovine brucellosis. The primary binding assays such as the FPA and ELISAs have been developed as more sensitive and more specific alternatives to conventional tests. Gally and Nielsen reviewed over 50 publications on the diagnostic performance of both the conventional and primary binding serological assays used for the diagnosis of bovine brucellosis and presented unambiguous evidence that the FPA and the ELISAs were of superior diagnostic performance (more sensitive and more specific) and less expensive than the conventional tests. The work done locally at the Potchefstroom provincial veterinary laboratory also suggestedthat the FPA and ELISAs had better performance indexes compared to RBT and CFT. The laborious and exacting nature of the CFT, inconsistency in the results obtained from time to time and the unreliable supply of reagents had prompted the South African Veterinary Laboratory Scientific Forum to look for evidence in order to convince theveterinary authority to discontinue the CFT in favour of the alternative assays.

Keywords

Bovine brucellosis – Conventional serological tests (RBT/CFT) – Alternative assays

Introduction

Bovine brucellosis is one of the most important infectious diseases controlled by the state in South Africa. As such, all tasks in the field as well as in the laboratory are regulated and carried out according to the state strategy on the eradication of the disease. Enormous amounts of resources are invested each year by the state in controlling bovine brucellosis since the inception of the brucellosis eradication scheme in 1968. The testing of serum samples using the official serological tests (RBT and CFT) are some of the routines that have not changed over the years. There is a perception that the brucellosis eradication scheme has failed. Role players in the public and private sectors have different opinions on what they think might be the gaps in the eradication scheme. Field workers have indicated several challenges justifying the current status quo of bovine brucellosis in the country. Some critics have pointed to the laboratory test results. Laboratory workers have, from time to time, also blamed field workers for submitting poor quality samples to the laboratories. Since the introduction of the quality management system (QMS) in the laboratories about ten years ago, thousands of poor quality samples (mostly haemolysed) have been rejected by the laboratories resulting in fruitless expenditures. Small changes in strategies have begun to surface and are being enforced by the regulatory body. The changes affecting the veterinary laboratories include the implementation of a QMS according to ISO/IEC 17025 standard and a system of approval for laboratory wishing to conducttests for controlled diseases introduced by the department of agriculture, forestry and fisheries in 2008/9. Despite all these efforts it appears as if bovine brucellosis is here to stay as there has been very little change in the number of outbreaks reported yearly (Fig. 1).

There are up to 20 government and parastatal veterinary laboratories (Easter Cape: 4 labs, Free State: 2 labs, KwaZuluNatal: 2 labs, Limpopo: 4 labs, Mpumalanga: 2 labs, Northern Cape: 1 lab, North West: 2 labs, Western Cape: 2 labs and the ARC-Onderstepoort veterinary institute) which, at some point in time, have all been involved in the testing for bovine brucellosis in the country. Smaller laboratories have no capacity for the complement fixation test (CFT). The total number of samples submitted to the provincial laboratories (excluding the ARC-OVI) is over 800,000 samples each year resulting in close to 900,000 RBTand CFT performed. The leading provinces in numbers of samples tested for bovine brucellosis are Free State and North West (over 180,000 samples submitted to each province between January and December 2013)followed by Mpumalanga (over 120,000 samples submitted to this province during the same period). The abovementioned figure exceeds 1,000,000 tests per year when the tests performed at ARC-OVI and other private veterinary laboratories are added.

TheRBT is used as the first-step test to screen all the samples. It is a very sensitive test but it can also give false positive results because of S19 vaccination or false-positive serological reactions. False-negative reactions occur rarely, mostly due to prozoning and can sometimes be detected by diluting the serum sample or retesting after 4–6 weeks. Nevertheless RBT appears to be adequate as a screening test for detecting infected herds or to guarantee the absence of infection in brucellosis-free herds.

All the samples reacting positive on RBT (and all of the samples intended for qualification of animals for import/export) are subjected to the CFT for confirmation. The CFT is an OIE prescribed test for international trade and is usually very specific. However, like all other serological tests, it could sometimes give a positive result due to S19 vaccination or due to cross-reacting organisms. The CFT, as such,is a relatively simple testwidely used and accepted as a confirmatory test for bovine brucellosis, but it is a very exacting procedure (because 5 variables are involved) requiring good laboratory facilities and adequately trained staff to accurately titrate and maintain the reagents.All reagents involved in the CFT must be used at optimal reactivity. Therefore it is imperative that all known reagents be carefully prepared and standardized to insure a completely balanced system. These standardizations, which involve titrations of sheep red blood cells, hemolysin, complement, and antigens before the test proper can be performed, render the CFT rather difficult and time consuming. An actual test requires 2 days to complete.

The South African Veterinary Laboratory Scientific Forum (SAVLSF) has taken account of all the facts given above, as well as the recurring unavailability or inconsistent supply of the RBT/CFT antigens and antiserawhen needed,to start looking for alternatives that will suit the test program of the veterinary laboratories, including the average laboratory.


Fig. 1	Fig. 2

South African map showing bovine brucellosis outbreaks South African map showing the Veterinary laboratories

Materials and methods

Despite the huge numbers of outbreaks and all the samples collected and tested every year for bovine brucellosis very little research has been undertaken on bovine brucellosis in South Africa. Stacks of raw data which provide no meaningful information are generated every year by the provincial laboratories and the ARC-OVI. However, common sense as well as the little available information led the veterinary laboratories, under the leadership of the SAVLSF,to a conclusion that, as a country,we have stuck with the conventional testsfor too long despite the availability of primary binding assays which were developed as more sensitive and specific alternatives to conventional tests. Talks and frustrations about the laborious nature of the CFT and the intent to replace it with a simpler but efficient test gained momentum about eight years ago, but no concrete steps were taken with the exception of a small scale validation work onELISA and FPA which took place in few laboratories. The work of Kangumba was published in the proceedings of South African Veterinary Association(SAVA) (Sun City, 2008) and International Society for Veterinary Epidemiology and Economics(ISVEE) XII (Durban, 2009) and created a platform for further thoughts.Thus, the only reliablesource of data to bring about evidence that it was time for South Africa to introduce alternative tests in the routine testing of samples for bovine brucellosis was the literature published in peer-reviewed journals.

Internet search

Using internet search engines we retrieved and studied the articles where comparisons between conventional and alternative tests were discussed. The publication of Gally& Nielsen was most useful and formed the basis for the conclusions and recommendations drawn in this article.

Experimental work for performance evaluation

Conventional tests

The RBT, Slow agglutination test (SAT) and CFT were assessed in the validation work conducted at the ARC-OVI and Potchefstroom Provincial Veterinary laboratory (PVL) between 2007 and 2008. All tests were conducted according to approved standard protocols.

Alternative tests

The performance characteristics of the Enzyme-linked immunosorbent assays (ELISA) and Fluorescent polarization assay (FPA) were compared in various experiments with RBT, SAT and CFT at the ARC-OVI and Potchefstroom PVL according to the prescribed protocols.

Proficiency test by interlaboratory comparison

In 2015 the first proficiency testing involving the conventional tests (RBT and CFT) and one primary binding assay (ELISA) was organised by JGK Lab AfriQa (www.jgklab.co.za).

Results

Table 1 represents the published sensitivity of different serological techniques used for antibody detection. The ELISA is clearly more sensitive than the CFT (Ross-Whybin, www.rcpa.edu.au/)

Table 1

Sensitivity of different serological techniques used for antibody detection (µg/mL)

Method	Sensitivity (µg/mL)
Gel-diffusion	30
Ring precipitation	18
Bacterial Agglutination	0.05
Complement Fixation	0.05
Passive Haemagglutination	0.01
Haemagglutination Inhibition	0.005
Immunofluorescence	0.005
ELISA	0.0005
Bacterial Neutralisation	0.00005

Table 2 shows the information extracted from the work of Gally& Nielsen. Over 50 publications in which the sensitivity and specificity values of assays used for the detection of exposure to Brucellaabortus were reviewed. The authors averaged the sum of the sensitivity and specificity values for each test to give a performance index (PI) and allow for a comparison between the different methodologies. A score of 200 was perfect. These findings were published in a peer-reviewed journal. We extracted and summarized the data affecting the two conventional tests (RBT and CFT) currently used in South Africa and the two potential alternative tests (ELISAs and FPA) which the SAVLSF is currently contemplating.The results show that the FPA and ELISAs had better mean PI and fewer variations (SD and %CV) compared to the conventional tests.

Table 2

Comparison between the mean of performance indexes, standard deviation and % coefficient of variation of conventional tests currently used in South Africa as official tests for bovine brucellosis and alternative tests (Gall & Nielsen, 2004)

Test	Number of data sets*	Total number of samples tested	Mean of sensitivities	Mean of specificities	Mean PI	SD	% CV
Rose Bengal test	11	12,146	81.2	86.3	167.6	24.8	14.8
Complement Fixation test	38	28,537	89.0	83.5	172.5	24.3	14.1
Polymerase chain reaction	5	539	82.0	98.6	180.6	21.4	11.9
Indirect Enzyme-linked immunosorbent assay	37	60,985	96	93.8	189.8	18.1	9.5
Competitive Enzyme-linked immunosorbent assay	14	15,865	97.7	90.5	188.2	16.6	8.8
Blood Fluorescent polarization assay	2	1,385	97.2	98.4	195.6	4.13	2.2
Bulk milk Fluorescent polarization assay	1	258	100	95.9	195.9	N/A	N/A
Fluorescent polarization assay	7	39,934	97.5	98.9	196.4	4.4	2.2

Table 3 summarises the results of the comparative study published by Kangumba in the proceedings of SAVA and ISVEE XII. Despite the limited number of samples tested, the work was well accepted as a platform for further work and earned its merit for the fact that B. abortus was first isolated from the herd before the samples were collected for serological investigation. Bacteriological isolation is still the only absolute method for establishing infection statusand is considered the Gold standard test, and therefore the preferred method, for defining sera to be used for estimates of sensitivity. This test was done in a dairy herd where milk sampleswere collected and tested for evidence of brucellosis. After the isolation of B. abortus100 animals were selected randomly and bled using redtop vacutainer tubes. Serum was decanted from each tube and enough aliquots were made to allow comparative testing with 7 tests: (1) RBT, (2) SAT, (3) CFT, (4) FPA, (5) iELISA (Pourquier), (6) iELISA (IDEXX) (short incubation mode) and (7) iELISA (IDEXX) (long incubation mode. The samples were of unknown values and the purpose for designing a blind testing was to determine the number of samples which each test would classify as positive.

Similarly, a confirmed negative herd without vaccination at the Potchefstroom College of Agriculture was used as the negative herd for the experiment. Forty cows were selected randomly and sampled. Aliquots of samples were subjected to the same testing regiment as above.

Table 3

Comparative study on performance indexes of RBT, SAT, CFT, FPA, iELISA (Pourquier), iELISA (IDEXX, short incubation mode) and iELISA (IDEXX, long incubation mode) on identical samples collected from confirmed bovine brucellosis positive and negative dairy herds (Kangumba, 2008)

Tests	Positive herd (n=100)	Negative herd (n=40)	Performance indexes (ps+ns/N)
Rose Bengal test	41	40	41.28
Slow Agglutination test	39	40	39.28
Complement Fixation test	39	40	39.28
Fluorescent Polarization test	43	39	43.27
Indirect Enzyme-linked immuonsorbent assay (Pourquier)	46	40	46.28
Indirect Enzyme-linked immuonsorbent assay (IDEXX) (short incubation mode)	49	39	49.27
Indirect Enzyme-linked immuonsorbent(IDEXX) (long incubation mode)	50	39	50.27

The sum of positive samples in the positive herd and negative samples in the negative herd was divided by the total number of animals tested (n=140) to give a performance index for each test. The ELISAs and FPA showed better performance indexes compared to the conventional tests.

Table 4 represents the results of the first proficiency testing (PT) by interlaboratory comparison organised by JGK Lab AfriQa during the first quarter of 2015. Twelve sets of data obtained from participating laboratories in South Africa and the SADC region were included in the analysis. Four borderline positive samples collected from a herd (previously negative without vaccination) that had just seroconverted, 2 unequivocal positive samples collected from a confirmed infected herd and 4 unequivocal negative samples collected from a known negative herd were include in the PT panel. From the results obtained, the iELISA (IDEXX) was clearly more sensitive than the CFT and did not miss any of the samples reported as positives by the screening test (RBT).

Table 4

Results of the PT scheme for bovine brucellosis organised by JGK Lab AfriQa during the first quarter of 2015 depicting a better sensitivity of the iELISA over the conventional confirmatory test (CFT) (Kangumba, 2015 – www.jgklab.co.za)

RBT	CFT	ELISA
Pos.	Neg.	Pos.
Pos.	784 IU/mL	Pos.
Neg.	Neg.	Neg.
Pos.	Neg.	Pos.
Neg.	Neg.	Neg.
Neg.	Neg.	Neg.
Pos.	Neg.	Pos.
Neg.	Neg.	Neg.
Pos.	Neg.	Pos.
Pos.	240 IU/mL	Pos.

Discussion

Ideal tests for screening should be highly sensitive, inexpensive and rapid, whereas confirmatory tests should be both highly sensitive and specific. The cELISA and FPA fit the criteria for both screening and confirmatory tests.

Since the inception of the brucellosis eradication scheme South Africa has used only the CFT as confirmatory test. The CFT test was introduced by Wasserman in 1909 and its wide use was based on the suggestions of the time that it was a highly sensitive and specific test. With recent advances in primary binding assays (McGiven at al.) such a consensus can no longer hold, but there is a common agreement that improvements in quality assurance (QA) had helped improved the performance of CFT as observed over the last few decades. Unfortunately, this is not necessarily the case in South Africa as very few veterinary laboratories have implemented a QA program to the level of readiness for accreditation. Cases of inconsistent RBT/CFT results are still being reported on routine testing as well as interlaboratory comparisons.

The exacting nature of the CFT procedure requires good laboratory facilities and adequately trained staff which are lacking in some provincial laboratories. In the past technical staff were able to produce critical reagents of CFT (guinea pig complement, amboceptor and veronal buffer) in house. But the new generation of laboratory workers does not have such skills; therefore all reagents must be sourced commercially. The Onderstepoort Biological Products (OBP) is currently supplying only the RBT/MRT/CFT antigens and the positive serum and lots of complaints are received by the SAVLSF from provincial laboratories due to inconsistent supply of these reagents. Thus, all of the other reagents must be imported from overseas and are being distributed in the country at a very high cost. Gally & Nielsen reported that the cELISA and FPA were generally less expensive than conventional tests with FPA having the least expensive cost per unit sample. Ironically the conventional tests were almost as costly, and generally less accurate, to perform as the primary binding tests.

Recommendations

Scientific evidence demonstrates that the FPA has greater diagnostic accuracy than RBT and CFT and can be used under field conditions to detect antibody to Brucella spp. in whole blood or milk. In addition, the FPA is less costly (i.e. one reagent), easier to perform (i.e. simple and rapid) and adaptable to automation and enhanced QA, making it suitable for eradication programmes, national serological surveys, and laboratories handling large volumes of brucellosis samples as is the case in South Africa.

Similar to FPA, the results presented above demonstrate that the cELISA has greater diagnostic accuracy than the RBT and CFT. The bigger provincial veterinary laboratories and the ARC-OVI are equipped with ELISA machines and have competent technicians who are already performing ELISA tests on routine basis. The large volumes of samples handled in these laboratories can be tested with ELISA in shorter period of time. Also, the cELISA was found less expensive than conventional tests.

The cELISA and the FPA are capable of distinguishing vaccinated animals or animals infected with cross-reacting organisms from naturally infected animals, thereby reducing the number of false positive reactions and subsequent trace backs or slaughter of animals in an otherwise negative or healthy population. These primary binding assays were specifically developed as alternatives to conventional tests. Therefore, there is no justification to be stuck with the CFT when all the laboratories testing samples for bovine brucellosis have unanimously expressed their intention to move on because of the laborious nature of the test, inconsistent results obtained from time to time and the unreliable supply of critical reagents.

A new test may replace an existing test if it is of equal or superior diagnostic performance and is cost-effective with regard to the amount of reagent(s) required to support national programmes (Wright et al). The CFT has served its purpose and the time has come to depose it and move to the primary binding assays.

The veterinary authority should approve the replacement of CFT by the FPA and cELISA without further delay. Laboratories wishing to continue screening the samples with RBT could do so but in order to reduce the number of false positives due to S19 vaccination one of the primary binding tests (either FPA or cELISA) must also be performed.

The SAVLSF should take charge in insuring smooth transition from CFT to primary binding tests which should begin with the verification of their diagnostic performance under local conditions and the harmonisation of the testing protocols.

To assist in harmonising the test results and improve diagnostic performance nationally and internationally all laboratories will have to participate in proficiency testing schemes.

References

Angus R.D. & Barton C.E. (1984). – The production and evaluation of a buffered plate antigen for use in a presumptive test for brucellosis. In Proc. 3rd International Symposium on brucellosis, Algiers, Algeria. Dev. biol. Standard, 56, 349-356.

Gall D., Nielsen K., Bermudez M.R., Moreno F. & Smith P. (2002). – Fluorescence polarization assay for detection of Brucellaabortus antibodies in bulk tank bovine milk samples. Clin.diagn. Lab. Immunol., 9 (6), 1356-1360.

Gall D. & Nielsen K. (2004). – Serological diagnosis of bovine brucellosis: a review of test performance and cost comparison. Rev. sci. tech. Off. int. Epiz., 23 (3), 989-1002

Kangumba J.G.K. (2008) – To maintain, supplement or discontinue Rose Bengal and Complement Fixation tests as official tests for the diagnosis of bovine brucellosis in South Africa? A comparative study with Enzyme-linked immunosorbent assays and molecular technique.Proceedings of SAVA congress, Sun City, South Africa.

Kangumba J.G.K. (2009) – Performance and limitation of CFT and the subsequent challenges on the diagnosis of bovine brucellosis in South Africa. Proceedings of ISVEE XII, Durban, South Africa.

MacMillan A. (1990). – Conventional serological test. In Animal brucellosis (K. Nielsen & J.R. Duncan, eds). CRC Press, Boca Raton, 153-197.

McGiven J.A., Tucker J.D., Perrett L.L., Stack J.A. Brew S.D. & MacMillan A.P. (2003). – Validation of FPA and cELISA for the detection of antibodies to Brucellaabortus in cattle sera and comparison to SAT, CFT and iELISA. J. immunol. Methods, 278, 171-178.

Mikolon A.B., Gardner I.A., Hietala S.K., Hernandez de Anda J., Pestana E.C., Hennager S.G. & Edmondson A.J. (1998). – Evaluation of North American antibody detection tests for diagnosis of brucellosis in goats. J. clin.Microbiol., 36 (6), 1716-1722.

Nielsen K., Wright P.F., Kelly W.A. &Cherwonogrodzky J.H. (1988). – A review of enzyme immunoassay for detection of antibody to Brucellaabortus in cattle. Vet. Immunol.Immunopathol., 18 (4), 331-347.

Nielsen K., Gall D., Kelly W., Henning M.D. & Garcia M.M. (1992). – Enzyme immunoassay. Application to diagnosis of bovine brucellosis. Agriculture Canada Monograph, Nepean, 203 pp.

Nielsen K., Kelly L., Gall D., Nicoletti P. & Kelly W. (1995). – Improved competitive enzyme immunoassay for the diagnosis of bovine brucellosis. Vet. Immunol.Immunopathol., 46 (3-4), 285-291.

Nielsen K., Gall D., Jolley M., Leishman G., Balsevicius S., Smith P., Nicoletti P. & Thomas F. (1996). – A homogeneous fluorescence polarization assay for detection of antibody to Brucellaabortus. J. immunol. Methods, 195 (1-2), 161-168

Nielsen K., Gall D., Kelly W., Vigliocco A. & Garcia M. (1996). – Immunoassay development: application to enzyme immunoassay for the diagnosis of brucellosis. Agriculture and Agri-Food Canada Monograph, Nepean, 216 pp.

Nielsen K., Gall D., Smith P., Kelly W., Yeo J., Kenny K., Heneghan T., McNamara S., Maher P., O’Connor J., Walsh B., Carroll J., Rojas X., Rojas F., Perez B., Wulff O., Buffoni L., Salustio E., Gregoret R., Samartino L., Dajer A. & LunaMartinez E. (2001). – Fluorescence polarization assay for the diagnosis of bovine brucellosis: adaptation to field use. Vet. Microbiol., 80 (2), 163-170.

Nielsen K., Smith P., Gall D., Perez B., Samartino L., Nicoletti P., Dajer A., Rojas X. & Kelly W. (2001). – Validation of the fluorescence polarization assay for detection of milk antibody to Brucellaabortus. J. Immunoassay Immunochem., 22 (3), 203-211.

Nielsen K., Gall D., Bermudez R., Renteria T., Moreno F., Corral A., Monro O., Monge F., Smith P., Widdison J., Mardrueno M., Calderon N., Guerrero R., Tinoco R., Osuna J. & Kelly W. (2002). – Field trial of the brucellosis fluorescence polarization assay. J. Immunoassay Immunochem., 23 (3), 307-316.

OIE Terrestrial Manual (2012). – Bovine brucellosis, chapter 2.4.3.

Public Health England.(2013). Complement Fixation Tests.UK Standards for Microbiology Investigations. V 18 Issue 3.2. http://www.hpa.org.uk/SMI/pdf.

Rylatt D.B., Wyatt D.M. &Bundesen P.G. (1985). – A competitive enzyme immunoassay for the detection of bovine antibodies to Brucellaabortus using monoclonal antibodies. Vet. Immunol.Immunopathol., 8 (3), 261-271.

Ross-Whybin – Fundamental of serology. www.rcpa.edu.au

Stack J.A., Perrett L.L., Brew S.D. & MacMillan A.P. (1999). – Competitive ELISA for bovine brucellosis suitable for testing poor quality samples. Vet. Rec., 145 (25), 735-736.

Wright P.F., Nilsson E., Van Rooji E.M.A., Lelenta M. &Jeggo M.H. (1993). – Standardisation and validation of enzyme-linked immunosorbent assay techniques for the detection of antibody in infectious disease diagnosis. In Biotechnology applied to the diagnosis of animal diseases. Rev. sci. tech. Off. int. Epiz., 12 (2), 435-450.