EQUINE INFECTIOUS ANEMIA 



1 EQUINE INFECTIOUS ANEMIA IMMUNODIFFUSION TEST LABORA TOR Y SUR VEY. Pearson, James E. 



Diagnostic laboratories approved by APHIS, USDA, to perform the equine infectious anemia (EI A) 

 immunodiffusion (ID) test are required to test a set of unknown check test scrums at periodic intervals. The purpose 

 of this requirement is to insure that adequate quality control standards are maintained throughout the Nation and a 

 reasonable degree of uniformity in test results may be expected by the horse industry. The check list samples also 

 provided a form of training as the serums are more difficult than those routinely encountered. 



Check Samples 



Fifty samples were determined to be optimum for this evaluation. This number is considered statistically 

 significant in a test where the finding is cither positive or negative. A smaller number of samples would be required 

 in a test where antibody titers are determined. 



Forty -eight laboratories participated in the survey. 



Samples selected contained enough serums so that each laboratory received identical 0.5 ml. amounts. This 

 permitted each sample to be tested at least three times. Some of the samples selected were the most difficult that 

 were available. It was anticipated that few laboratories would get 100 percent. 



The samples were classified in five categories: negative— 13; very weak positive— 2; weak positive— 14; 

 positive— 16; and strong positive— 5. Several of the negative samples formed lines, of nonidentity with certain 

 antigens. Samples were obtained from a variety of sources. Several were from horses hyperimmunized with other 

 antigens that contained equine serum and, as a result, produced lines between the serum wells. Other samples were 

 from experimental horses that had been infected with EIA for various lengths of time. The remainder of the samples 

 were interesting samples we had received for routine testing. A few of the samples had been diluted or mixed to 

 obtain the desired reaction. 



Several of the samples were of very poor quality with a high lipid content. If these had been submitted for 

 routine diagnostic purposes another sample should have been requested. However to provide a standard comparison, 

 results for all samples were required. 



All samples were tested at least 10 times with at least five different antigens before being used as check 

 samples. Personnel that had received training at Veterinary Services Diagnostic Laboratory, Ames, Iowa, since June 

 1972 had tested at least half of the samples as part of their training. 



Evaluation of Results 



Each laboratory received their test score and a drawing of the reactions that were observed when the samples 

 were tested. The average number of serums scored incorrectly was 6.7. When the results of all the laboratories were 

 plotted the data was found to approximate a Poisson distribution. The laboratories that were Wrong on more than 

 10 serums fell below the 95 percent confidence limit and were considered unsatisfactory. 



Those laboratories that fell below the 95 percent confidence interval will receive special training, including 

 practice samples similar to those that they missed. Every one will receive another group of check samples which 

 must be tested with an improved score if the laboratory is to remain approved. 



Three of the 48 laboratories scored correctly on all of the samples. 



17 



