Susan L. Stramer PhD
Executive Scientific Officer, American Red Cross
Issues: Blood donation samples that are repeat reactive by FDA licensed HIV-1/HIV-2 screening assays currently undergo a confirmatory algorithm that involves HIV-1 western blot (WB) and HIV-2 EIA/WB for those samples that do not contain HIV-1 antibody. Issues with both sensitivity and specificity exist with the current blood donor confirmatory algorithm. The major issue with this algorithm includes high rates of samples that can neither be called positive or negative according to the FDA licensed inserts; these “indeterminate” results account for 40-50% of HIV repeat reactive donations, 30-40% of which do not even have bands to viral proteins but cannot be interpreted as WB negative. Fewer than 1:2,500 donors who test HIV-1 WB negative or indeterminate is HIV infected. Poor specificity is exacerbated when an individual falsely confirms positive on the WB (current rate approximately 1:700,000). Blood donors receiving these indeterminate and false-positive messages are confused and anxious, especially when their HIV uninfected status is corroborated by RNA negativity. Western blots and comparable confirmatory assays have not evolved since 1987 in contrast to screening assays which have made significant improvements in both sensitivity and specificity. As an example of improvements in sensitivity, one newly licensed HIV-1/HIV-2 (plus O) EIA is 74% more sensitive than the WB during seroconversion. This presentation provides options to the current use of only the HIV WB in blood donor confirmatory algorithms.
Description: Blood centers have implemented HIV-1 RNA testing by pooled NAT in 1999. Integration of NAT with serological test results improves the quality of the donor counseling message for repeat reactive donors by increasing the sensitivity early in infection when the WB typically would not be positive. It also helps stratify WB positive donors into those most likely infected versus those who likely are not (especially when coupled with S/CO ratios of the screening assay, RNA negativity of an individual sample, and donor follow up). However, greater than 95% of HIV-1/HIV-2 repeat reactive blood donors are negative for RNA in the pooled NAT format (and of those, the vast majority will test single unit NAT negative). Therefore, simply integrating NAT into confirmatory algorithms may address the sensitivity issue but does not address the specificity issue. Since the screening EIAs have ungone vast improvements in sensitivity and specificity, those samples testing RNA negative should be tested by a second licensed HIV-1/HIV-2 EIA and only those samples testing concordantly repeat reactive require further testing by WB. Feasibility of this “dual EIA” approach is based on the concept that if two assays with comparable sensitivity are composed of differing antigens, the false-positive populations should have limited cross over; the more unique the tests, the greater the separation of false-positive populations.
Lessons Learned: The dependence on the use of viral lysate-based confirmatory assays in use for HIV since 1987 should be addressed. The inclusion of new versions of HIV-1/HIV-2 screening assays that are not dependent on viral lysate antigens and licensed RNA testing should be included.
Recommendations: Modifications in the confirmatory algorithms for HIV-1/HIV-2 EIA repeat reactive blood donors are needed. Substitutions of a second licensed screening assay (to rule out the existence of HIV antibody) and RNA (to rule out the possibility of early HIV infection) for the WB should occur. The WB then may be used only to confirm antibody reactivity in those donors whose RNA is negative but who do have reactivity by two different licensed EIAs. Such an algorithm improves sensitivity and specificity (>98% of repeat reactive blood donors will no longer be counseled as “indeterminate”).

Last Update: April 8, 2005