Sensitivity and specificity of a recombinant transmembrane glycoprotein (rgp21)-spiked western immunoblot for serological confirmation of human T-cell lymphotropic virus type I and type II infections.

Serum specimens (n = 2,712) obtained from individuals residing in diverse geographic regions and categorized as seropositive (n = 122), seroindeterminate (n = 523), or seronegative (n = 2,067) for human T-cell lymphotropic virus (HTLV) infection in accordance with U.S. Public Health Service guidelines were retested by recombinant transmembrane protein (rgp21)-spiked Western immunoblotting. Of the 122 HTLV-positive specimens, those from 85 of 85 (100%) U.S. blood donors, 2 of 2 (100%) Brazilians, 1 of 2 (50%) Indonesians, 14 of 14 (100%) Solomon Islanders, and 18 of 19 (95%) Papua New Guineans reacted with rgp21, yielding an overall sensitivity of 98% (120 of 122). Specimens from individuals whose infections were confirmed to be HTLV type I or HTLV type II by the polymerase chain reaction assay reacted equally well with rgp21. Of the 523 HTLV-indeterminate specimens, those from 21 of 379 (5.5%) U.S. blood donors, 3 of 6 (50%) Brazilians, 10 of 23 (44%) Ugandans, 8 of 49 (16%) Indonesians, 4 of 36 (11%) Solomon Islanders, and 5 of 30 (17%) Papua New Guineans reacted with rgp21. None of these 51 specimens reacted with native gp46 and/or gp61/68 in a radioimmunoprecipitation assay, suggesting a false-positive reaction (9.75%). Of the 2,067 HTLV-negative specimens, 12 reacted with rgp21, yielding a false-positivity rate of 0.6%. These data indicate that while detection of rgp21 is highly sensitive, it can yield false-positive results. Thus, specimens exhibiting reactivity with rgp21 in the absence of reactivity with native gp46 and/or gp61/68 by Western blot should be tested further by a radioimmunoprecipitation assay to verify HTLV type I or type II infection.