Journal of Microbiological Methods 97 (2014) 25–28

Contents lists available at ScienceDirect

Journal of Microbiological Methods

j ourna l homepage: www.e lsev ie r .com/ locate / jmicmeth

Note

Innovativemodifications to Rose Bengal plate test enhance its specificity,sensitivity and predictive value in the diagnosis of brucellosis

Shubhada K. Chothe, Hari Mohan Saxena ⁎Department of Veterinary Microbiology, College of Veterinary Science, Guru Angad Dev Veterinary and Animal Sciences University (GADVASU), Ludhiana, Punjab 141004, India

Abbreviations: RBPT, Rose Bengal plate test; STAT, StaiELISA, Indirect enzyme linked immunosorbent assay; CFT⁎ Corresponding author at: Flat No. 9, First Floor, Geeta

Nagar, Ludhiana, Punjab 141001, India. Tel.: +91 941714E-mail address: hmsaxena@yahoo.com (H.M. Saxena)

0167-7012/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.mimet.2013.12.005

a b s t r a c t

a r t i c l e i n f o

Article history:Received 19 October 2013Received in revised form 8 December 2013Accepted 8 December 2013Available online 15 December 2013

Keywords:Superagglutination testBrucellosisSensitivitySpecificityFalse negativeFalse positive

Current agglutination tests occasionally yield false results. Superagglutination test reduced false results, hadhigher sensitivity (95.88%) and negative predictive value (95.83%) than Rose Bengal plate test (RBPT), StandardTube Agglutination test (STAT), ELISA, and Complement Fixation test and specificity (89.32%) and positive pre-dictive value (89.42%) higher than RBPT and STAT.

© 2013 Elsevier B.V. All rights reserved.

Brucellosis is an important zoonotic disease caused by Brucella or-ganisms. It is of public health significance and causes huge economiclosses to the livestock sector due to reproductive losses in animals, abor-tions, placentitis, epididymitis and orchitis. Brucellosis is endemic inIndia (Aulakh et al., 2008) where it is estimated to cause a loss of US$58.8 million per year (Kollannur et al., 2007).

The Rose Bengal plate test (RBPT) is often used as a rapid screeningtest in the diagnosis of brucellosis (Ruiz-Mesa et al., 2005). Although thesensitivity of RBPT is reported to be very high, the specificity can be dis-appointingly low (Barroso et al., 2002). As a result, the positive predic-tive value of the test is low and a positive test result thus requiresconfirmation by a more specific test (Smits and Kadri, 2005).

The RBPT could sometimes give a false positive result. Suitable mod-ifications of the RBPT are, therefore, required to get accurate results. Wehave developed a novel Superagglutination test to enhance the sensitiv-ity and minimize false positive and false negative results of RBPT(Saxena andKaur, 2013). The present studywas undertaken to comparethe sensitivity and specificity of the novel Superagglutination test withthe available serodiagnostic tests RBPT, STAT, CFT, and ELISA to evaluateits efficacy on serum samples that may be either false positive or falsenegative by RBPT.

Guidelines of the Institutional Animal Ethics Committee werefollowed in the study.

ndard Tube Agglutination test;, Complement fixation test.njali Apartments, E Block, Rishi7813 (mobile)..

ghts reserved.

A total of 200 bovine (181 cattle and 19 buffalo) serum sampleswere derived from the animals in veterinary clinics, dairy farms andgaushalas (animal shelters) in and around Ludhiana. All the animalswere of age two years ormore. Brucellosis suspected herdswere select-ed for sampling primarily based on the history of abortions in the herdwhile normal healthy animals were sampled from the herds of the uni-versity dairy farmwithout the history of abortions and repeatedly RBPTnegative status. Common serological tests i.e. the RBPT, STAT, iELISA andCFT along with the Superagglutination test were applied on all theserum samples.

RBPT was done as per the standard method (Morgan et al., 1978).For performing Superagglutination test, equal volumes (2.5 μl each)

of RBPT colored antigen, test serum stained with 0.1% Coomassie Bluedye, biotinylated anti-bovine IgG (Sigma) and streptavidin (Sigma)weremixed thoroughly on a clean glass slide in the abovementioned se-quence. The slide was observed for 4 min for the formation of clumps.Ordinary hand lens was used occasionally for better resolution. Theslides were viewed under low power (10×) of an inverted microscopeto visualize the composition of clumps in case of doubt. Formation ofclear agglutination, within which the blue color (due to the CoomassieBlue dye staining the serum antibodies) and the pink color (due to theRose Bengal dye stained RBPT antigen) could be differentiated on mag-nification, were considered as positive, while absence of clear aggluti-nates was considered as negative.

The standard method recommended by the Office International desEpizootes (OIE, 2004) was followed for Standard Tube Agglutinationtest (STAT).

To perform Indirect ELISA (iELISA), a commercial kit was procuredfrom Immunobiological Laboratories IBL-America (Minneapolis, USA).

26 S.K. Chothe, H.M. Saxena / Journal of Microbiological Methods 97 (2014) 25–28

The testwas performed according to the instructions provided in the kitmanual.

Complement Fixation test (CFT) was performed as per the OIEManual.

The positive predictive value (PPV) and negative predictive value(NPV) for each diagnostic test were calculated using the followingformulae:

PPV ¼ Number of True Positive casesNumber of True Positive casesþNumber of False Positive cases

NPV ¼ Number of True Negative casesNumber of True Negative casesþ Number of False Negative cases

:

Observed proportion of agreement (OPA) and agreement beyondchance (kappa values) were determined usingWinepiscope-2 softwarepackage with 95% confidence level.

Out of the 200 samples, 97 were found to be positive by RBPT(Table 1). The percent prevalence of brucellosis varied with the test

Table 1Results of analysis of sera by various serological tests.

Sample no.

69(20/4/11), 80(20/4/11), 83(20/4/11), A, B, F, G, BB, BH, BR, DD, DL, K, L, M, N, S, V,AC, AF, AG, AP, AZ, BG, CB, CK, 68(20/4/11), 70(20/4/11), 71(20/4/11), 77(20/4/11)

78(20/4/11), 81(20/4/11), BKATT2(1/3/11), 13(1/3/11), 23(1/3/11), 76(T 5/10/11), 10(13/10/11), E, BA, BF, CY, DN, DR,AB,74(20/4/11), 3 (13/10/11)

AEAI, BC, BYAN, BW23(20/4/11), BP, 102(15/9/11), 7(20/4/11)73(20/4/11), 1(1/3/11), 20(1/3/11), 21(T5/10/11), 5(13/10/11), AD, BL, 66(20/4/11),3(1/3/11), T7, 6(13/10/11), T2061(25/11/11), J

T6CE, AVCN, AJCX, AO7(13/10/11)29(15/9/11), 40(15/9/11), 15(20/4/11), 17(20/4/11), 20(20/4/11), 34(20/4/11), 42(20/4/1151(20/4/11), 53(20/4/11), 54(20/4/11), 72(20/4/11), 76(20/4/11), T1, T3, 4(1/3/11),92(T 5/10/11), 11(13/10/11)

T8, 79(T 5/10/11), T 85 (25/11/11), ODQ, BE, AQ, 2413T 81 (25/11/11)T 2062 (25/11/11)H, W, X, AS, BTAL, AY, 2379CI2218, 2452, 2581, 101(15/9/11), I, AH, CR, DG, DJ, DO13(20/4/11), 79(20/4/11), 100(20/4/11)11(1/3/11), 19(1/3/11), 82(T 5/10/11)DU, BI12(13/10/11), CV, 16(20/4/11), BJ, BMP2308, 2417, 89(T 5/10/11), U, Y, AK, AW, CT, DB, DC103(15/9/11), 6(1/3/11), 9(1/3/11)T486(T 5/10/11)CM, T5, AAR, BDBX, DP, CZ2362, Q, AM, AR, AU, BS82(20/4/11)BO, BQ2490, T2, AX, BZ, 25(20/4/11)2574, 25822426, 2467, 2554, 2567, 104(15/9/11), 19(20/4/11), 32(20/4/11), 87(T 5/10/11), CD, CH, CS,88(T 5/10/11), 80(T 5/10/11), T 29 (25/11/11), T 77 (25/11/11), T 78 (25/11/11), T 84 (25

T 8 (25/11/11)T 86 (25/11/11)

and ranged from 43.50% to 48.50%. The test detected 6% less positivesamples than the Superagglutination test and showed a sensitivity of93.33%, a specificity of 88.18%, a PPV of 86.6% andNPV of 94.17%, respec-tively (Table 2).

In the case of the Superagglutination test, the clumps on the slidehad both blue and pink color. When the slide was viewed under thelowpower of a lightmicroscope, the agglutinate could be very easily dif-ferentiated into two parts, the antibodies were blue in color due to theCoomassie blue dye and the antigen was pink in color due to the RoseBengal dye (Figs. 1 & 2). A total of 104 out of the 200 serum sampleswere detected positive by Superagglutination test (Table 1). The test de-tected more positive samples than ELISA (16.5%), CFT (14.5%), RBPT(6%) and STAT (6%) and showed a sensitivity of 95.88% and a specificityof 89.32%. The positive predictive value (PPV) of this test was found tobe 89.42% and Negative Predictive Value (NPV) was 95.83% (Table 2).

STAT could detect 119 out of the 200 samples as positive. A titer of1:40 and above was considered as positive. A total of 81 samples hadtiters below 1:40, 36 samples had a titer of 1:40 and 83 samples hadtiters more than 1:40 (Table 1).

STAT titer RBPT Superagglutination test iELISA CFT

00 − − − −

00 − − − +00 − + − −10 − − − −

10 − − − +10 + + − −10 + + + +20 − − − +20 − − − −

20 + − − −20 − − + +20 + + − −20 + + + +20 − + − −

), 40 − − − −

40 − − − +40 + + + −40 − + − +40 − + + +40 + + + +40 + + − +40 + + − −80 + + + +80 − + − −80 − − − −80 + + + −80 + + − −80 + + − +

160 + + + +160 − + − −160 + − − −160 + − + +160 + + − +160 + + − −160 + + + −320 + + + +320 − + − +320 + + − +640 + + + +640 + + + −

2489,/11/11)

N1280 + + + +

N1280 + + − +N1280 + − + +

Table 2Sensitivity and specificity of the Superagglutination test and its agreement with other serodiagnostic tests.

Serological test Values (in percent) Agreement of Superagglutination with other serological tests

Sensitivity Specificity PPV NPV Prevalence of the disease OPA (percentage) Kappa value Degree of agreement

Superagglutination 95.88 89.32 89.42 95.83 48.50RBPT 93.33 88.18 86.60 94.17 45.00 92.5 0.850 Almost perfectSTAT 94.25 68.14 69.49 93.90 43.50 81.5 0.626 SubstantialiELISA 74.47 95.24 93.33 80.65 47.24 81.5 0.634 SubstantialCFT 82.80 93.46 91.67 86.21 46.5 81.0 0.623 Substantial

27S.K. Chothe, H.M. Saxena / Journal of Microbiological Methods 97 (2014) 25–28

STAT detected 6% less positive samples than the Superagglutinationtest and showed a sensitivity of 94.25% and a specificity of 68.14%. PPVof this test was found to be 69.49% and NPV was 93.90% (Table 2).

Out of the 200 samples, 75 were detected as positive by iELISA(Table 1). ELISA detected 16.5% less positive samples than theSuperagglutination test. Sensitivity of this test was calculated to be74.47% and the specificity was found to be 95.24%. PPV of this test was93.33% and NPV was found to be 80.65% (Table 2).

Out of the 200 samples, 86 were detected as positive by CFT(Table 1). In the present study, the CFT detected 14.5% less positive sam-ples than the Superagglutination test and showed a sensitivity of 82.8%and a specificity of 93.46% (Table 2).

Statistical agreement between the Superagglutination test and othertests was determined. The kappa values and the observed proportionsof agreements are presented in Table 2.

The agreement between the Superagglutination test and RBPT wasfound to be almost perfect, whereas the agreement between theSuperagglutination test and other tests was found to be substantial.

False positive reactions in RBPT have been attributed to residual an-tibody activity fromvaccination, cross reactionwith certain bacteria andlaboratory error. False negative reactionsmay arise during early incuba-tion of disease or immediately after incubation (Radostits et al., 2000).False negative results may be due to a small clump size in sera withlow titers of antibodies and false positive results due to the inability todifferentiate non-specific aggregates of antigen particles alone fromthe true agglutinates comprising both antigen and antibody. Crossreacting antibodies have been detected by RBPT and false negative reac-tions are believed to occur mostly due to prozoning (OIE, 2004).

Fig. 1. The gross view of the superagglutination test on a glass slide by naked eye.

In the present study, Superagglutination test, developed by us re-cently (Saxena and kaur, 2013) by introducing novel modifications intheRBPT, has been evaluated for its usefulness in the diagnosis of bovinebrucellosis by comparing its results with those of RBPT, STAT, iELISA,and CFT. The new test could detect more positive samples than RBPT,STAT, iELISA and CFT. In an earlier report (Akhtar et al., 2010) a lowspecificity of RBPT had been demonstrated.

The high sensitivity of STAT observed in the present study can be at-tributed to its ability to detect antibody titer as low as 1:40 which wasprobably missed by other tests. In serum agglutination tests, cross reac-tions with various bacteria for example Yersinia enterocoliticaO:9, E. coliO:157, Francisella tularensis, Salmonella urbana group N, Vibrio choleraeand Stenotrophomonas maltophilia have been reported (Corbel andBrinley-Morgan, 1984).

The Complement Fixation test is technically challenging because alarge number of reagents must be titrated daily and a large number ofcontrols of all the reagents are required. It is an expensive test and islabor intensive. Other problems include the subjectivity of the interpre-tation of results, occasional direct activation of complement by serum(anti-complementary activity) and the inability of the test for use withhemolysed serum samples. However, since only IgG1 isotype of anti-body fixes complement well, the test specificity is high (Poester et al.,2010). In an earlier study, Stemshorn and Forbes (1985) had demon-strated a sensitivity of 79% for CFT for the diagnosis of bovinebrucellosis.

Fig. 2. Microscopic view showing a two colored agglutinate formed in theSuperagglutination test with blue colored antibodies bound to the pink colored antigenparticles.

28 S.K. Chothe, H.M. Saxena / Journal of Microbiological Methods 97 (2014) 25–28

The highest sensitivity of all the tests observed withSuperagglutination test in our study can be attributed to the fact thatthe number of false negative results with Superagglutination test wasless when compared to the other tests. The anti-bovine IgG plays arole in increasing the sensitivity since if fewer antibodies against the in-fectious organism are present in the serum, the anti-globulin cross linksthese antibodies resulting in an increase in the clump size. Streptavidinbinds to the biotinylted anti-globulin increasing the clump size furtherby up to 4 fold due to the presence of four binding sites for biotin oneach molecule of avidin. The specificity of the Superagglutination testwas found to be higher than that of RBPT and STAT.

In the case of the Superagglutination test, the two colored trueagglutinates could be very easily differentiated from non-specific onecolored aggregates under the low power of a light microscope. The an-tibodies were blue in color due to the Coomassie blue dye and the anti-gen was pink in color due to the Rose Bengal dye. Each agglutinate hadboth the blue and the pink color, which aided in the differentiation ofthe true agglutinates from the non-specific aggregates of the antigenof pink color only. The antigen and antibodies which did not participatein agglutination reaction could be viewed under the microscope as ag-gregates of either blue or pink particles alone lying separately. Ordinaryhand lens was generally found to be effective in visualizing theagglutination.

The highest agreement with RBPT, combined with the higher speci-ficity and sensitivity of Superagglutination test ensures that it can serveas a more efficient screening test than RBPT.

The Superagglutination test had a higher sensitivity and a negativepredictive value than the other serodiagnostic tests like RBPT, STAT,ELISA and CFT. Its specificity and PPV were found to be better thanRBPT and STAT. The test can be used in the pen side diagnosis of bovinebrucellosiswith better results than the RBPTwhich is routinely used as apen side test for brucellosis. ELISA is not a cost effective test whenscreening has to be performed on herds with a large number of animals.In such situations, Superagglutination test can offer an advantage of in-creased sensitivity of screening compared to RBPT and STAT.

In our study, the agreement between Superagglutination testand RBPT was found to be higher than the agreement betweenSuperagglutination test and other tests. This agreement combinedwith the higher specificity and sensitivity of Superagglutination test en-sures that it can serve as a more efficient screening test than RBPT.

The Superagglutination test showed the highest sensitivity of all thetests (95.88%)which can be attributed to the lesser number of false neg-ative results obtained with Superagglutination test compared to theother tests. The anti-globulin played a role in increasing the sensitivitybecause if there are less number of antibodies against the infectious or-ganism present in the serum, anti-globulin binds to these antibodiesresulting in an increase in the clump size. Avidin binds to the biotinylat-ed anti-globulin increasing the clump size further by up to 4 fold due tothe four binding sites for biotin present on each molecule of avidin.

The specificity of the Superagglutination test was however, found tobe somewhat lower than that of iELISA and CFT. This may be due to thecross reactions of anti-globulin with non-specific or low affinity anti-bodies. Interestingly, the agglutinate in the case of Superagglutinationtest could be differentiated into two parts, the antibodies were blue incolor due to the Coomassie blue dye and the antigen having a pinkcolor due to the Rose Bengal dye. Each clump had both the blueand the pink colors, which aided in the identification of the true

agglutinates. This has not been reported by any other researcher asyet. Antigen and antibodies which did not participate in agglutinationreaction could be viewed under the microscope as blue and pink parti-cles lying separately or as single colored aggregates.

The results obtained in our study suggest that Superagglutinationtest has a higher sensitivity than other serodiagnostic tests commonlyused for brucellosis. The specificity of Superagglutination test washigher than RBPT and STAT but lesser than that of ELISA and CFT. Thenew test can be used as a screening test in the diagnosis of bovinebrucellosis. As ELISA is not a cost effective test, Superagglutination testcan offer the advantages of increased sensitivity, cost effectiveness, norequirement for skilled personnel and ease of performance.

Conflict of interest

The authors declare that there is no conflict of interest. The innova-tivemethods in the Superagglutination test are patented (Inventor: HMSaxena; Assignees: 1. GADVASU, Ludhiana and 2. The Department ofBiotechnology, Ministry of S&T, Govt. of India, New Delhi). The studywas funded by GADVASU.

SKC conducted the study as a Master's degree student under the su-pervision of HMS, a Professor of Immunology. HMS conceived the ideasunderlying the Superagglutination test. He designed, supervised and co-ordinated the study, analyzed and interpreted the data and drafted themanuscript. SKC carried out the experimental work and statisticalanalysis.

We are thankful to Dr. Mudit Chandra and Dr. Deepti Narang,Department of Veterinary Microbiology, GADVASU, Ludhiana for theirhelp.

References

Akhtar, R., Chaudhry, Z.I., Shakoori, A.R., Ahmad, M.D., Aslam, A., 2010. Comparative effi-cacy of conventional diagnostic methods and evaluation of polymerase chain reactionfor the diagnosis of bovine brucellosis. Vet. World 3, 53–56.

Aulakh, H.K., Patil, P.K., Sharma, S., Kumar, H., Mahajan, V., Sandhu, K.S., 2008. A study onthe epidemiology of bovine brucellosis in Punjab (India) using milk-ELISA. Acta Vet.77, 393–399.

Barroso, G.P., Pelayo, R.C., Extremera, G., 2002. Study of 1595 brucellosis cases in theAlmeria province (1972–1998) based on epidemiological data from diseasereporting. Rev. Clin. Esp. 202, 577–582.

Corbel, M.J., Brinley-Morgan, W.J., 1984. Brucella. In: Krieg, N.R. (Ed.), Bergey's Manual ofSystematic Bacteriology, Vol. I. Baltimore: Williams and Wilkins, pp. 377–388.

Kollannur, J.D., Rathore, R., Chauhan, R.S., 2007. Epidemiology and Economics of Brucello-sis in Animals and its Zoonotic Significance. ISAH Tartu, Estonia 466–468.

Morgan, W.J., Mackinnon, D.T., Gill, K.P.W., Gower, S.G.M., Norris, P.I.W., 1978. Brucellosisdiagnosis: standard laboratory techniques report. Series no. 1, Weybridge, England.

OIE, 2004. Bovine brucellosis, Section 2.3, OIE Manual of Standards for Diagnostic Testsand Vaccines, 5th edition. OIE, Paris.

Poester, F.P., Nielsen, K., Samartino, L.E., Yu, W.L., 2010. Diagnosis of brucellosis. Open Vet.Sci. J. 4, 46–60.

Radostits, O.M., Gay, C.C., Blood, D.C., Hinchcliff, K.W., 2000. Diseases Caused by Brucellaspp. A Textbook of the Disease of Cattle, Sheep, Pigs, Goats and Horses, Ninth edition.Harcourt Publishers Limited, London 881–890.

Ruiz-Mesa, J.D., Gonzalez, J.S., Reguera, J.M., Martin, L., Palmero, S.L., Colmenero, J.D., 2005.Rose Bengal test: diagnostic yield and use for the rapid diagnosis of human brucello-sis in emergency departments in endemic areas. Clin. Microbiol. Infect. 11, 221–225.

Saxena, H.M., Kaur, P., 2013. A new superagglutination test to minimize false negative andfalse positive results common with plate/slide agglutination tests for the diagnosis ofinfectious diseases. Int. J. Trop. Dis. Health 3, 199–209.

Smits, H.L., Kadri, S.M., 2005. Brucellosis in India: a deceptive infectious disease. IndianJ. Med. Res. 122, 375–384.

Stemshorn, B.W., Forbes, L.B., Eaglesome, M.D., Nielsen, K.H., Robertson, F.J., Samagh, B.S.,1985. A comparison of standard serological tests for the diagnosis of bovine brucello-sis in Canada. Can. J. Comp. Med. 49, 391–394.