Comparative performances of an HTLV-I/II EIA and other serologic and PCR assays on samples from persons at risk for HTLV-II infection

BACKGROUND: HTLV-I and HTLV-II are related exogenous pathogenic human retroviruses. Until recently, ELISAs based on HTLV-I antigens have been used to screen for the presence of HTLV-I or-II antibodies. The HTLV-I-based assays have not been as sensitive in detecting antibodies to HTLV-II as in detecting antibodies to HTLV-I. The Abbott HTLV-II HTLV-II ELISA uses a combination of HTLV-I and HTLV-II antigens to detect antibodies to the whole HTLV group. The performance of this ELISA was compared to that of several HTLV-I-based serologic assays and an HTLV-II PCR assay in cohorts of South American Indians and New York City IV drug users (IVDUs) in whom HTLV-II is endemic. STUDY DESIGN AND METHODS: Sera from 429 South American Indians and New York City IVDUs were evaluated for HTLV antibodies by the use of three ELISAs (rgp21-enhanced HTLV-I/II, Cambridge Biotech;Vironostika HTLV-I/II, Organon Teknika; and HTLV-I/HTLV-II, Abbott), and a Western blot (WB) assay. Peripheral blood leukocyte DNA from each person was analyzed for HTLV-I and HTLV-II pol DNA via PCR. The HTLV-II PCR-positive samples were further subtyped via cloning and sequencing and/or oligomer restriction. RESULTS: Two hundred four samples (48%) were positive for HTLV-II by serologic and/or PCR assays. All of the positive samples from the Indians and approximately one-third of the positive samples from the IVDUs were of the HTLV-IIB subtype. Comparative analyses indicate that the sensitivity and specificity of the various assays were: PCR, 98 and 100 percent; Abbott HTLV-I/HTLV-II, 78 and 95 percent; Cambridge Biotech HTLV-I/II, 76 and 96 percent;Vironostika HTLV-I/II, 71 and 98 percent; and WB, 73 and 100 percent, respectively. CONCLUSION: There were no significant differences among the sensitivities and specificities of the HTLV-I/II ELISAs (p values, 0.056-0.438). The WB and PCR assays were much more specific than the other serologic assays (p less than or equal to.03). However, the PCR assay is significantly (p<0.001) more sensitive than any of the serologic assays in the detection of HTLV-II infection. Thus, optimal detection of HTLV-II infection would seem to require both serologic and DNA PCR assays.