SciELO - Scientific Electronic Library Online

Home Pagealphabetic serial listing  

Services on Demand




Related links


Ciência Rural

Print version ISSN 0103-8478

Cienc. Rural vol.23 no.2 Santa Maria May/Aug. 1993 







José Alfonso Barajas-Rojas1 Hans Riemann2 Charles Franti3




A herd of cattle (Holstein-Zebu crosses) was screened every two months by ELISA during a period of two years for IgG antibodies against 19 infectious disease agents. Two hundred and ninety five sera were collected from 157 young animals (0-4 months of age), 1037 sera from 292 developing animals (4-36 months of age) and 1468 sera from 259 producing animals (> 36 months of age). The results indicate that the difference in ELISA between positive and negative tests is associated with the overall prevalence of positive tests. When the prevalence of positive tests is low the difference between positive and negative tests is greater than when the prevalence is intermediate or high. This means that ELISA, presumably other serological tests for IgG antibodies, is more reliable at low disease (antibody) prevalence. This will tend to offset the declining predictive value of positive tests at low prevalence and may contribute to the successful use of serological tests in disease eradication.

Key words: serological screening, cattie diseases, ELISA.



Um rebanho bovino (cruza Holandês-Zebu) foi monitorado durante um período de dois anos pelo teste imuno-enzimático (ELISA) para detectar anticorpos contra 19 agentes infecciosos. Duzentos e noventa e cinco soros foram coletados de 157 animais jovens (0-4 meses de idade), 1037 soros de 292 animais em crescimento (4-36 meses de idade) e 1468 soros de 259 animais em produção (>36 meses de idade). Os resultados indicam que a diferença entre resultados positivos e negativos ao ELISA está associada com a prevalência geral de testes positivos. Quando esta prevalência for baixa a diferença entre resultados positivos e negativos é maior do que quando esta for intermediária ou alta. Isto significa que o ELISA é mais confiável quando em prevalências baixas. Isto tenderá a compensar diminuindo a probabilidade de valores positivos em prevalências baixas podendo contribuir no uso da sorologia em programas de erradicação.

Palavras-chave: triagem sorológica, doença de bovinos, ELISA.




Serological testing for circulating antibodies is convenient for detection of infectious diseases. The reliability of testing of is customarily expressed in terms of sensitivity and specifity of the test and as the derived predictive values, which depend on the prevalence of the infection. Specifity is a relative term since it depends on cross reactions with other infectious agents the prevalence of which may vary geographic location and age and species of the host. Sensitivity depends on the preparation of the antigen used and may vary according to the degree of seroresponse of the host.

The present study deals with observations which indicate that a low prevalence of an infection in a population is associated with a more reliable discrimination between positive and negative tests.



The study was conducted in a cattle herd at the center for Research, Teaching and Extension in Tropical Livestock (Centro de Investigación, Enseñanza y Extensión en Ganadería Tropical-CIEEGT) located in the North-central part of the state of Veracruz, México. Blood samples were collected bimonthly during 1988 and 1989 and the sera subjected to ELISA with 19 antigens prepared from infectious agents as described by BARAJAS-ROJAS et al (1993).

Two hundred and ninety five serum samples were collected from 157 young animals (< 4 months of age), 1037 samples were collected from 292 developing animals (4-36 months of age) and samples from 259 producing animals (> 36 months of age).



The mean percent ELISA and standard deviations for positive tests (tests above the cut off point) and negative tests (tests below the cut off point) shown in table 1. Plots of mean of percent ELISA an standard deviation for positive and negative tests against the prevalence of positive tests are shown in figure 1, 2 and 3. These plots suggest that as the prevalence decreases the difference between the mean percent ELISA of positive and negative tests increases; this applies to all 3 age groups. Plots of logarithms of percent prevalence against difference between positive and negative ELISA (Figure 4, 5 and 6) illustrates the association more clearly and indicates that the association is best for the producing animals and least clear for the most heterogeneous age group, the developing animals.















The results presented here may not be surprising since a high prevalence may mean a high endemic level of infection with seroconversion and a waning of antibodies going on at the same time with the result that a considerable number of animals have ELISA values dose to the cut-off point. The fluctuations in ELISA values related to season and state of pregnancy (BARAJAS-ROJAS et al, 1993) show that many infections (seroresponses) have such a dynamic behavior. A low prevalence on the other hand may represent a situation where diseases transmission (infections) is not very active. Most animals will then have reduced antibody levels and the ELISA positive animals may represent chronic or sporadic infection. It may be argued that the pattern found in this study is due to a higher sensitivity of the antigens associated with high prevalence with the result that many tests will cluster dose to the ELISA cut-off point, assuming that percent ELISA will rarely exceed 100. However this is not a likely explanation; furthermore there is no apparent association between the prevalence of positive tests and the recorded maximum percent ELISA (not shown).

It may also be argued that repeated testing of the same animal (maximum 5.7 tests per producing animals) may cause distortion. However the number of repeated tests is small compared to the total number of tests and the absence of complete "independence" of test will hardly cause any bias; furthermore all we deal with in serological testing is a population of test results and dynamics of many diseases suggest that result from an animal at one point in time may be quite independent of result obtained at a different time. This is indicated in the report by BARAJAS-ROJAS et al (1993) and will be discussed again in a subsequent article.

According to conventional wisdom, which assumes that test specifity is independent of the prevalence the condition tested for, the predictive value of positive tests wanes as the prevalence approaches zero. Taken to its logical consequence this means that it would be next to impossible to base eradication of a disease on serological testing. Practical experience show otherwise (e. g. eradication of brucellosis, bovine leukemia, infectious bovine rhinotracheitis and pseudorabies) and part of the explanation may be the widening gap between positive and negative tests as prevalence decreases.

The evidence presented here together with the known facts that test results may vary with geographic location as well as age, species and breed of animals make the conventional concepts of sensivity and specifity almost meaningless. At the very least estimates of sensivity and specifity should be based on specification of host species and age as well as on geographic locality and prevalence of the condition tested for.



We appreciate the assistance in sample collection and data collection of Dr. Eusebio García Neria and Dr. Rebeca Acosta and personnel of the Centro de Investigación Enseñanza y Extension en Ganadería Tropical-CIEEGT of the Faculty of Veterinary Medicine. National Autonomous University of México. Mexico City.



BARAJAS-ROJAS, J.A., RIEMANN, H.P., FRANTI, F.E. Application of enzyme linked immunosorbent assay (ELISA) for epidemiological studies of diseases of livestock in the tropics of México. Rev Sci Tech Off Int Epiz, v. 12, n. 1, 1993. In Press.         [ Links ]



1Doctor in Veterinary Medicine MSc, PhD. Professor Titular Dept of Bacteriology, Faculty of Veterinary Medicine, National Autonomous University of México, México City 04510, México

2Doctor in Veterinary Medicine, PhD Professor Dept. Epidemiology and Preventive Medicine, University of California, Davis, CA 95616, USA.

3Statistician, PhD, Professor of Epidemiology and Preventive Medicine, University of California. Davis, CA 95616, USA.

Recebido para publicação em 21.10.92. Aprovado para publicação em 06.01.93.

Creative Commons License All the contents of this journal, except where otherwise noted, is licensed under a Creative Commons Attribution License