Establishment of a novel diagnostic test algorithm for human T-cell leukemia virus type 1 infection with line immunoassay replacement of western blotting: a collaborative study for performance evaluation of diagnostic assays in Japan

Abstract

Abstract Background The reliable diagnosis of human T-cell leukemia virus type 1 (HTLV-1) infection is important, particularly as it can be vertically transmitted by breast feeding mothers to their infants. However, current diagnosis in Japan requires a confirmatory western blot (WB) test after screening/primary testing for HTLV-1 antibodies, but this test often gives indeterminate results. Thus, this collaborative study evaluated the reliability of diagnostic assays for HTLV-1 infection, including a WB-based one, along with line immunoassay (LIA) as an alternative to WB for confirmatory testing. Results Using peripheral blood samples from blood donors and pregnant women previously serologically screened and subjected to WB analysis, we analyzed the performances of 10 HTLV-1 antibody assay kits commercially available in Japan. No marked differences in the performances of eight of the screening kits were apparent. However, LIA determined most of the WB-indeterminate samples to be conclusively positive or negative (an 88.0% detection rate). When we also compared the sensitivity to HTLV-1 envelope gp21 with that of other antigens by LIA, the sensitivity to gp21 was the strongest. When we also compared the sensitivity to envelope gp46 by LIA with that of WB, LIA showed stronger sensitivity to gp46 than WB did. These findings indicate that LIA is an alternative confirmatory test to WB analysis without gp21. Therefore, we established a novel diagnostic test algorithm for HTLV-1 infection in Japan, including both the performance of a confirmatory test where LIA replaced WB on primary test-reactive samples and an additional decision based on a standardized nucleic acid detection step (polymerase chain reaction, PCR) on the confirmatory test-indeterminate samples. The final assessment of the clinical usefulness of this algorithm involved performing WB analysis, LIA, and/or PCR in parallel for confirmatory testing of known reactive samples serologically screened at clinical laboratories. Consequently, LIA followed by PCR (LIA/PCR), but neither WB/PCR nor PCR/LIA, was found to be the most reliable diagnostic algorithm. Conclusions Because the above results show that our novel algorithm is clinically useful, we propose that it is recommended for solving the aforementioned WB-associated reliability issues and for providing a more rapid and precise diagnosis of HTLV-1 infection.