False-Positive Western Blot Tests for Antibodies to HTLV-IIIN

To the Editor.—As part of the US Army quality assurance program for human T-cell lymphotropic virus type III (HTLV-III) testing, we recently sent identical proficiency panels to five large commercial firms that offered HTLV-III Western blot testing on a fee-for-service basis. Each panel consisted of 15 samples from healthy, low-risk adults whose serum was repeatedly negative for HTLV-III antibodies and five samples from HTLVIII-infected (culture-documented) patients. By Western blot analysis, all five positive samples produced definite, but not strong, bands at 41 kilodaltons (kd) and other molecular weights characteristic of HTLV-III proteins.1Scores ranged from 85% to 100% accuracy (Table). Four of the five laboratories reported at least one falsepositive test result. The six falsepositive results were all on different normal specimens, suggesting that the errors were due to technique and not to intrinsic biologic properties of the specimens. There was a substantial range of clarity of the