Surveillance and testing strategies for the diagnosis of CBPP: Results of the
FAO/IAEA Co-ordinated Research Programme on the monitoring of CBPP in Africa

Roland Geiger

Animal Production and Health Section, Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture

Introduction and Background

Reliable and efficient diagnostic tests for the serological diagnosis of infectious diseases are the cornerstone of any disease control strategy. The requirements for diagnostic tests depend on the purpose of the diagnostic testing and the epidemiological needs. The monitoring of a region wide vaccination programme has different needs compared to the needs of export certification of individual animals. In the case of contagious bovine pleuropneumonia (CBPP) countries adjacent to infected areas may want to confirm the absence of the disease relying on serology and slaughterhouse inspections. Positive test results will lead to follow up investigations and, depending on the diagnostic tests used, may result in substantial expenses. Highly specific tests with a high positive predictive value are needed to limit the amount of follow up work.

Countries which are operating a disease control programme with no vaccination based on stamping out of positive herds and compensation of the farmers require tests which are highly sensitive and specific at the individual animal level and in the case of positive results several follow up investigations might be undertaken. In countries where the disease is present and control programmes based on movement control and vaccinations are operated the sampling efficiency must be high to enable constant monitoring and redirection of the ongoing programme.

Serological tests

For serological diagnosis the complement fixation test (CFT) is still the most widely used test. Some authors report it as highly sensitive in the acute phase, with lower sensitivity in later stages (OIE, 2002). Other authors report an overall sensitivity of only 63.6% Bellini et al., 1998). Specificity is reported to be high, but some authors report false positive results in up to 3.5% of all sera leading to the misclassification of up to one third of the herds investigated (Stark et al., 1995). An indirect ELISA which showed a high level of sensitivity was found to show many non-specific reactions and consequently a more specific competitive ELISA was developed and introduced into 11 African countries through an FAO/IAEA Co-ordinated Research Project (CRP) on the “Monitoring of contagious bovine pleuropneumonia in Africa using enzyme immunoassays” (Le Goff and Lefevre, 1989; LeGoff and Thiaucourt, 1998). The main objective of this CRP was to compare and validate the main serological tests for the diagnosis of CBPP in particular the CFT and the competitive ELISA. In CBPP free areas the competitive ELISA was reported to have a specificity of close to 100% (Amanfu et al., 1998). An indirect ELISA based on the specific lipoprotein LppQ of Mycoplasma mycoides subsp. mycoides small colony type (Mmm SC) showed a high level of sensitivity and specificity when used in one country but not enough data are available for reliable estimates under different epidemiological conditions (Bruderer et al., 2002).

A rapid latex agglutination test for the detection of circulating capsular polysaccharide (CPS) antigen of Mmm SC and for the detection of circulating antibodies to Mmm SC which was reported to have a high specificity was introduced in the last year into the CRP and compared to the results of the CFT and the cELISA (March et al., 2002).

Material and Methods

FAO/IAEA Co-ordinated Research Project

The CRP that was operational from 1997 until 2003 involved 11 African countries with various epidemiological situations. Three Research Agreement Holders (UK, France, Sweden) provided advice and scientific guidance to the CRP. Annual FAO/IAEA Research Contracts awarded to the participating institutions ensured the supply of reagents (primarily for the CFT and the cELISA) and equipment (IAEA 2002). Four meetings were held where the results of the CRP were presented. The results of the final CRP are being published.

Figure 1. Countries participating in the FAO/IAEA Co-ordinated Research Project “The monitoring of contagious bovine pleuropneumonia in Africa using immunoassay”

Research Contract Holders:

1. Nat. Vet. Lab., Botswana
2. Lab. Path. Anim., Côte d’Ivoire
3. NAHRC, Ethiopia
4. CVL, Ghana
5. CVL, Kenya
6. LCV, Mali
7. CVL, Namibia
8. Nat. Vet. Res. Inst., Nigeria
9. ADRI, Tanzania
10. LIRI, Uganda
11. CVRI, Zambia

Research Agreement Holders:

• CIRAD-EMVT, France

• Nat.Vet.Inst., Sweden

• Scient. Agric. Coll., UK

Sera and samples

Namibia and Botswana were the only countries where sera from confirmed CBPP free areas were available. In 9 countries (Ghana, Nigeria, Cote d’Ivoire, Mali, Ethiopia, Kenya, Uganda, Tanzania, Zambia no confirmed negative populations were available. In the endemic areas the Research Contract holders were advised to collect sera preferably from herds where CBPP was diagnosed clinically.

Diagnostic tests

cELISA

The reagents were provided in a kit form containing all the reagents (e.g. polystyrene microplates coated with Mmm SC lysate, a monoclonal Mmm SC mouse antibody, dilution and washing buffers, a monoclonal anti-mouse IgG peroxidase conjugate and TMB with stopping solution). The initially standard protocol of the competitive ELISA following the FAO/IAEA protocol (FAO/IAEA, 1998) that used freeze-dried antigen was replaced in 2002 through a simplified protocol using precoated plates (FAO/IAEA, 1998; Kit Insert, 2002).

Complement fixation test (CFT)

The CFT was carried out as described before (Campbell and Turner, 1953).

Latex agglutination test

This test was designed as a penside test and that could be carried out under field conditions. The test samples were mixed with latex beads on a glass slide and positive samples showed a visually detectable agglutination. The test was provided in three formulations and was carried out as described in the protocol (March, 2002).

• Blue test: Blue latex beads coated with whole Mmm SC. Designed to detect total Mmm SC antibodies in serum.

• White test: White latex beads coated with Mmm SC capsular polysaccharide (CPS). Designed to detect CPS antibodies in serum.

• Red test: Red latex beads coated with anti-Mmm SC CPS IgG antibodies. Designed to detect CPS (principle circulating antigen in serum and other fluids.

Results

cELISA: Sensitivity and specificity

In Namibia and Botswana the specificity of cELISA was between 98% (n=50) and close to 100% (n=1200). The latter value is correct because larger numbers of samples were tested.

Only the relative sensitivity of the cELISA compared to the CFT could be determined since not enough statistically significant numbers of samples from confirmed positive cases were available. Generally the CFT was able to detect more positive sera in newly infected herds whereas the ELISA was able to detect more animals in endemic situations. In an infected herd in Tanzania sequential bleedings of 29 cattle, the cELISA was positive for 25 cattle and the CFT for 21 cattle. In Botswana the testing of stored sera from cattle from the outbreak area with pathognomonic lesions detected in post mortems and which represented most probably true positive cases showed a sensitivity of 90% (n=82) for the cELISA and 93% in the CFT. In Cote d’Ivoire the relative sensitivity of the cELISA to the CFT was very variable and was dependant whether new outbreaks were investigated or whether herds with old infections were investigated (relative sensitivities were: 87% (n=170), 54% (n=176), 0% (n=211, but in the CFT only 4%, e.g. 8 positives) and 88% (n=480)). In Cote d’Ivoire the overall relative sensitivity of the cELISA compared to the CFT was 78% (n=1037).

In Nigeria the relative sensitivity of the cELISA was 100% (n=170). In Namibia the CFT detected 15 out of 44 animals from a herd with a new outbreak where the cELISA could not detect any animals. In Tanzania in recently vaccinated herds 96% (n=578) of animals reacted positive in the CFT, whereas all the animals were negative in the cELISA resulting in a specificity of the cELISA in vaccinated animals of 100%.

An important aspect in the testing of sera is that the results are repeatable and reproducible and how well a test can be quality controlled. In the ELISA the change to a format with precoated plates reduced the variability in 10 out of 11 countries to acceptable levels. Test results of the quality controls which were included in each kit and consequently on each plate (strong positive serum, weak positive serum, negative serum, conjugate control, monoclonal control) fell within predefined upper and lower control limits indicating that the results of unknown test sera would be equivalent in each of the testing laboratories.

An indirect ELISA based on the specific lipoprotein LppQ of Mmm SC (Bruderer et al., 2002) was also included and tested as part of the CRP in one country but the results were not conclusive.

Latex agglutination test

Blue test for the detection of total Mmm SC antibodies

Results from five countries (Tanzania, Nigeria, Cote d’Ivoire, Mali, Uganda) showed a high number of positives in infected herds (71% to 100%) that indicated a high sensitivity but probably also a low specificity because of non-specific agglutination. However, no results from infection free areas are yet available which is essential to exclude a possible low specificity of the test.

White test for the detection of anti CPS antibodies

Sensitivity in seven countries ranged from 37% to 100% in new outbreaks. However, all animals (Namibia, n = 50) tested from infection free areas were positive indicating a low specificity, although another report showed a higher specificity (UK, 94%, n = 32) (see Huebschele et al., this report). More testing is presently undertaken in Namibia and Botswana.

Red test for the detection of circulating CPS

No conclusive data on the performance of this test were available. Specificity on sera ranged between 98% in Botswana and 38% in Namibia. Sensitivity using sera was low between 0% (Uganda) and 26% in Namibia and it appears that the test is not suitable for the testing of sera.

Discussion

Studies on the distribution of the disease (surveillance)

For the establishment of successful control programmes it is essential to delineate infected areas from areas that are not infected. If CBPP positive animals are detected by clinical, post mortem or serological investigations in a new area, this area is becoming classified as infected. Abattoir surveillance was recommended as a useful tool for the detection of CBPP (Mariner, 2003). Although the lesions of CBPP in acute cases are pathognomonic the clinical or pathological diagnosis of new outbreaks should be confirmed by laboratory diagnosis through the isolation of the infectious agent.

If serological investigations are undertaken the detection of specific antibodies must be interpreted as a sign of infection until follow up investigations have explained the positive results. Serological positive animals can be the result of vaccinations, cross-reactive sera, or the results of infection with a field strain of Mmm SC. The expected serological prevalence in infected herds ranged in Cote d’Ivoire between 0% by the cELISA and 4% by the CFT, in another case 19.6% by the CFT and 20% by the cELISA up to 58% by the CFT and 64% by the cELISA. Modelling suggested that between 3% and 75% of herds if the herd size is between 75 and 500 heads will be infected but only 1.8% to 73% are continuously infected. Cross sectional serological surveys should be designed to detect this basic herd prevalence and within-herd prevalence. Although the established serological tests such as the CFT and the cELISA have low sensitivities, it is possible to compensate for a lower sensitivity by increasing the sample number; e.g. if the sensitivity of a test is only 50% it is possible to compensate for the reduced sensitivity by doubling the sample size. It is more difficult to compensate for a low specificity since the testing of primarily negative samples from supposedly disease free areas will result in high number of (false) positives; e.g. in a scenario with disease at a 5% level where a test with a sensitivity of 100% and a specificity of 100% is applied 59 samples are required to detect at least one positive animal with 95% confidence, if the sensitivity drops to 50% 119 samples are required. If in the same scenario a test with 95% specificity and 100% sensitivity is used 330 samples would be needed and the population would be considered as infected with 95% confidence if 23 or more positives are detected. Because of the reduced specificity of the test there are 5 false positive samples for every 100 true negative samples that have to be clarified through followed up investigations at the field level. This example illustrates that for the surveillance a less than perfect sensitive test is acceptable but a reduced specificity results in rapidly growing sample numbers and associated follow up work.

The specificity of the CFT was not assessed as part of the CRP but the reported results for the specificity of the CFT vary from 96.5% to 99.5%. The results of the FAO/IAEA CRP and other reports suggest a specificity of the cELISA close to 100% if large numbers of samples are tested. To assess the true sensitivity of the cELISA it would be necessary to determine what percentage of animals were detected by the CFT and not detected by the cELISA were vaccinated animals and which were true positive samples, considering that the specificity of the CFT is less than 100%. This was not possible in the present study. On the other hand it is known that the sensitivity of the CFT is less than 100% in particular in chronic infections. In a number of countries the cELISA detected more animals than the CFT. Considering that the specificity of the cELISA is close to 100% it must be assumed that these are animals that were missed by the CFT and that the true sensitivity of the cELISA is considerably higher than the relative sensitivity compared to the CFT. Estimates of the relative sensitivity of the cELISA to the CFT in chronic infection range from 54% (Cote d’Ivoire) to 100% (Nigeria). To calculate the necessary sample sizes for disease surveillance to delineate infected areas from non-infected areas it is therefore suggested to adopt a conservative estimate of 50% sensitivity for the cELISA. This is below the true sensitivity and will lead in the design of the sampling frame to an overestimation of the sample size. At the same time a conservative estimate on the minimum expected prevalence should be adopted which will avoid that control strategies are based on erroneous assumptions that areas are free of disease.

The required sample sizes (number of herds) to detect disease at certain prevalences is shown in Table 1. Table 2 shows the number of animals that have to be sampled to detect disease within a herd with a probability of 95%. Assuming that in an infected herd 20% of animals became infected this would result in a sample size of 15 animals to detect the disease with 95% confidence. If the sensitivity of the test used is only 50% twice the number of samples are required. Modelling suggests that the proportion of herds with active infection in an infected area is above 3%. The serological prevalence will be well above this figure so that the probability that antibody to Mmm SC will be detected in at least one herd would be above 95% if 105 herds were sampled.

During the surveillance activities to achieve an international recognition of freedom from rinderpest infection the member countries of PACE are implementing serosurveys which are designed to detect disease at a level of 1% infected herds. These surveys are based on randomized selection of sampling sites and since the stratification for a CBPP survey would be in many countries similar to the stratification of the rinderpest surveys it is possible to make relative precise estimates on the distribution of CBPP.

Table 1. Number of herds which have to be sampled to detect at least one positive herd with a probability of 95% if infection within a herd is also detected with 95% probability.

Table 2. Number of samples needed to detect at least one positive in large herds (>10.000). If a diagnostic test with lower sensitivity is used the numbers have to be increased proportionally.

Disease detection in buffer zones

Buffer zones or ‘cordon sanitaires’ are established to prevent the spread of infection from endemic areas into disease free areas. This is achieved through a combination of movement control, vaccination, stamping out and intensive surveillance. In this area serological surveillance is difficult since both the CFT and the cELISA will detect vaccinated animals at various levels. Results from Tanzania indicate that the specificity of the cELISA in vaccinated animals might be close to 100% in a situation where the CFT reacts with 96% of the animals. This would result in substantial numbers of false positive results if serological surveillance is used but considering the high specificity of the cELISA combined with abattoir surveillance, clinical surveillance and follow up of positive results through cluster analysis and correct epidemiological analysis serological surveillance can be a useful tool. None of the two tests was shown to be a reliable tool to monitor vaccination coverage.

Disease detection in surveillance zones

Early detection of the outbreak of disease in the surveillance zone is essential to avoid the further spread of the disease. No vaccinations are carried out in the surveillance zones, which is part of the buffer zones. Abattoir, clinical and serological surveillance should be carried out in this zone. The CFT and the cELISA are complementary and a parallel testing strategy should be adopted where a positive test result in any of the two tests will trigger a follow up action with outbreak investigations.

It is not possible to calculate the sensitivity of such a combined test system consisting of serological tests and clinical observations. Estimates of the combined sensitivity of two independent serological test systems is shown in Table 3). Parallel testing strategies increase the sensitivities considerably but the disadvantage is that the specificity of the test system is reduced proportionally.

Table 3: Combined sensitivities of two independent test systems in parallel testing, e.g. if a serum sample is tested in the CFT (assumed sensitivity 0.6) and in the cELISA (assumed sensitivity 0.7) in parallel and a positive result in either of the tests is used to classify the sample as positive, the combined sensitivity is 88%.

Prevalence studies in endemic areas

Prevalence studies are carried out in areas where the disease is endemic to assess the impact of the disease or to decide on a more appropriate control strategy (e.g. stamping out versus vaccination). Prevalence studies are best carried out through cross sectional serological studies. Reliable prevalence studies are only possible based on a random selection of the samples and abattoir surveillance is not suitable to estimate the prevalence since the study population will be biased. If the sensitivities and specificities of the tests are known it is possible to correct the prevalence estimates and to use tests with low sensitivities. Both serological tests (CFT and cELISA) might be recommended for mass screening but preference should be given to tests which can be quality controlled such as the cELISA.

Confirmation of outbreak

It is essential to confirm the clinical diagnosis or the post mortem diagnosis through laboratory investigations if outbreaks occur in areas previously not infected. This must include the isolation of Mmm SC, the CBPP agent. The serological tests most widely used, the CFT and the cELISA detect antibodies to Mmm SC at different stages of infection and no test can detect all infected animals. Serological confirmation should be carried out at the herd basis since both serological tests are not sensitive enough at the animal level to refute the diagnosis based on a single animal. A minimum of 15 blood samples should be collected from clinically affected animals and tested in both tests. The influence of antibiotic treatments on the performance of the serological tests has not been assessed and should be investigated.

Conclusions

None of the validated diagnostic test is sufficient on its own for all the needs of the diagnosis and surveillance of CBPP. However, if the CFT and the cELISA tests are used in an epidemiological and statistical meaningful way and if these tests are combined in correct testing strategies it is possible to detect new disease outbreaks and to make reliable estimates on the distribution of the disease which will enable the implementation of targeted and focused control programmes. The surveys which are needed to assess the distribution of the disease can be undertaken using the existing structures and resources which were established through PACE and through the FAO/IAEA support programmes to carry out rinderpest surveillance. The laboratory capacity to carry out the necessary large scale serological surveys are in place in most African countries, but the diagnostic capacity to isolate Mmm SC and to confirm outbreaks of CBPP exists only in few countries.

References

OIE (2000). Chapter 2.1.6., Manual of standards for diagnostic tests and vaccines, 4th edition.

Bellini, S. Giovannini, A., di Francesco, C., Tittarelli, M., Caporale, V., Sensitivity and specificity of serological and bacteriological tests for contagious bovine pleuropneumonia, Rev. sci. tech. Off. Int. Epiz., 1998, 17 (3), 654-659

Stark, K.D.C., Vicari, A., Tontis, A. und Nicolet, J. (1995) Untersuchungen zur Epidemiologie der Lungenseuche in der Schweiz. Schweiz. Arch. Tierheilk. 137, 92-100

Le Goff, C., Lefevre, P.C., 1989. Peripneumonie contagieuse bovine: test immunoenzymatique et cinetique d’apparition des anticorps au cours d’une infection experimentale: relation entre la fixation du complement, l’excretion et la recherché de l’antigene cirulant. Rev. Elev. Med. Vet. Pays Trop. 42, 365-369

Le Goff, C., Thiaucourt, F., (1998) A competitive ELISA for the specific diagnosis of contagious bovine pleuropneumonia (CBPP). Vet. Microbiol. 60, 179-191

Amanfu, W., Sediade, S., Masupu, K.V., Benkirane, A., Geiger, R., Thiaucourt, F., (1998), Field validation of a competitive enzyme-linked immunosorbent assay for the detection of contagious bovine pleuropneumonia in Botswana. Rev. Elev. Med. Pays Trop. 51: 189-193

Bruderer, U., Regalla, J., Abdo, E., Huebschle, O.J.B., Frey, J., (2002) Serodiagnosis and monitoring of contagious bovine pleuropneumonia (CBPP) with an indirect ELISA based on the specific lipoprotein LppQ of Mycoplamsma mycoides subsp. mycoides SC, Vet. Microbiol. 84, 195-205

March, J.B., Kerr, K., Lema, B., (2002), Rapid Detection of Contagious Bovine Pleuropneumonia by a Mycoplamsma mycoides subsp. mycoides SC Capsular Polysaccharide-specific Antigen Detection Latex Agglutination Test, Clin. Diagn. Lab. Immunol.,Vol 10, No. 2

IAEA (2002). Monitoring of contagious bovine pleuropneumonia in Africa using enzyme immunoassays, (in print), Proceedings of the Final Research Coordination Meeting of an FAO/IAEA Co-ordinated Research Project (CRP) on the “Monitoring of contagious bovine pleuropneumonia in Africa using enzyme immunoassays”, TECDOC.

FAO/IAEA, (1998). Manual for competitive CBPP ELISA, Bench Protocol Version CBPP 2.1, January 1998, IAEA, Vienna.

Kit Insert (2002). CBPP serum competition ELISA, Vers. P05410/01, Pourquier, Montpellier, France

Campbell, A.B., Turner, A.W. (1953) Studies on Contagious bovine pleuropneumonia of cattle: An improve complement fixation test, Brit. Vet. J. 29, 154-163

March, J., (2002) CBPP Latex agglutination test (LAT) data sheet, Moredun Research Institute, UK.

Mariner, J.C., 2003, The Dynamics of CBPP Endemism and the Development of Effective Control/Eradication Strategies for Pastoral Communities, Final Modelling Report.