Immunoassays based on the highly immunogenic transmembrane protein of human T-cell lymphotropic virus type 1 (HTLV-1) (protein 21c) are capable of detecting antibodies in all individuals infected with HTLV-1 and HTLV-2. However, because of antigenic mimicry with other cellular and viral proteins, such assays also have a large proportion of false-positive reactions. We have recently identified an immunodominant epitope, designated GD21-I located within amino acids 361 to 404 of the transmembrane protein, that appears to eliminate such false positivity. This recombinant GD21-I protein was used in conjunction with additional recombinant HTLV type-specific proteins and a whole virus lysate to develop a modified Western blot (immunoblot) assay (HTLV WB 2.4). The sensitivity and specificity of this assay were evaluated with 352 specimens whose infection status was determined by PCR assay for the presence or absence of HTLV-1/2 proviral sequences. All HTLV-1-positive (n = 102) and HTLV-2-positive (n = 107) specimens reacted with GD21-1 in the HTLV WB 2.4 assay, yielding a test sensitivity of 100%. Furthermore, all specimens derived from individuals infected with different viral subtypes of HTLV-1 (Cosmopolitan, Japanese, and Melanesian) and HTLV-2 (IIa0, a3, a4, IIb1, b4, and b5) reacted with GD21-I in the HTLV WB 2.4 assay. More importantly, HTLV WB 2.4 analysis of 81 PCR-negative specimens, all of which reacted to recombinant protein 21e in the presence or absence of p24 and p19 reactivity in the standard WB assay, showed that only two specimens retained reactivity to GD21-I, yielding an improved test specificity for the transmembrane protein of 97.5%. None of 41 specimens with gag reactivity only or 21 HTLV-negative specimens demonstrated reactivity to GD21-I. In an analysis of additional specimens (n = 169) from different geographic areas for which PCR results were not available, a substantial increase in the specificity of GD21-I detection was demonstrated, with no effect on the sensitivity of GD21-I detection among specimens from seropositive donors. Thus, the highly sensitive, GD21-I-based HTLV WB 2.4 assay eliminates the majority of false-positive transmembrane results, thereby increasing the specificity for serologic confirmation of HTLV-1 and HTLV-2 infections.