[Comparison of different methods for anti-HTLV-I assay]

In 130 patients, who were considered to be anti-HTLV-I positive or negative by the PA method, we compared the anti-HTLV-I detection rates and the specificity of the following three EIA methods: the Ei-test ATL and two new EIA methods using different recombinant antigens which recognize different sites. The results from the three EIA methods were consistent with the results from the PA method at a rate over 96.9%. The specificity and sensitivity of the three methods were excellent. In 8 (0.6%) of the 130 cases, however, the results from the four methods were not in agreement. All of these 8 cases had been classified (by the PA method) as weakly positive (low antibody titer). The use of the Ei-test ATL produced some false positive cases and some false negative cases (no false negative cases have been reported in tests for anti-HTLV-I antibody before). In 3 patients, the results of the two new EIA methods were not in agreement. Because all of these three patients had a low antibody titer, the discrepancy was difficult to explain based on the difference in the antigens used. Although the four methods had similar anti-HTLV-I detection rates, the results indicate a need to carefully evaluate the data in patients with low antibody titers. Therefore, it is recommended that a combination of multiple tests be used or that the results from one test be checked against those from another test.