HIV testing

Townsend Letter for Doctors and Patients, June, 2006 by Jule Klotter

How accurate are HIV tests? Rodney Richards, PhD, an organic chemist who worked on HIV testing for the biotech company Amgen (formerly known as Applied Molecular Genetics) says the Enzyme-Linked Immunosorbent Assay (ELISA), Western Blot, p24 antigen, and PCR and branch DNA (bDNA) viral load tests are hampered by the lack of a "gold standard." HIV has never been directly isolated in such a way that manufacturers can prove that a positive test result means that the virus is present in a sample.

ELISA and Western Blot look for antibodies, not a specific virus. In his interview for Zenger's Newsmagazine (2001), Richards acknowledges the work of science writer Christine Johnson, who identified over 60 factors that can cause false-positives in these tests. Antibodies to herpes, hepatitis, and malaria, and conditions such as pregnancy or having a flu vaccination can produce a positive result. Richards says that "every diagnostic test for any germ that's based on antibodies has the same problem. It's a well-recognized problem attributed to what's known as 'cross-reactivity.' In other words, antibodies to germ A may coincidentally react to germ B."

If the ELISA test is positive, US standard of care recommends testing with the Western Blot. The Western Blot consists of a series of proteins, believed to be unique to HIV, on a strip. If antibodies stick to the protein, that band can be seen. If the Western Blot gives a negative result, the ELISA results are discounted as a false positive. These proteins, however, may not be so unique. In an article for Alive and Well (May 2004), Matt Irwin, MD says that gp24, which is the protein most often used as definitive evidence of HIV, has been found in people with other conditions, including people with generalized warts, cutaneous T-cell lymphoma, and multiple sclerosis.

In the early days of HIV testing, at least 50% of healthy blood donors who tested positive on ELISA also had antibodies that triggered one or more bands on the Western Blot, according to Richards. Various health and government institutions selected banding patterns that "minimized the numbers of blood donors testing positive and maximized the numbers of people in risk groups testing positive." Criteria for a positive HIV result on the Western Blot--which is considered more reliable than the ELISA--varies from country to country and, sometimes, from agency to agency within the same country. Richards says manufacturer inserts for HIV tests make no claims that the results correlate with the actual presence of HIV. The package insert for Roche Diagnostic Systems' Amplicor HIV-1 Monitor Test Kit (June 1996) reads: "The Amplicor HIV-1 Monitor test is not intended to be used as a screening test for HIV or as a diagnostic test to confirm the presence of HIV infection." Genetics Systems ELISA package insert (2000) reads: "The risk of an asymptomatic person with a repeatedly reactive serum sample developing AIDS or an AIDS related condition is not known." Cambridge Biotech Western Blot package insert (1998) says, "The clinical implications of antibodies to HIV-1 in an asymptomatic person are not known." Yet, life and death decisions are based on the accuracy of these tests.

PCR (polymerase chain reaction) and related tests (QC-PCR and the branched NDA test) are used to measure viral load, a technique for monitoring AIDS progression. PCR replicates a tiny segment of DNA until a million copies exist. Researchers believe that the bits of DNA (not intact viral particles) that they are replicating belong to HIV. Irwin points to studies that show high viral load numbers often do not correlate with any sign of an active, infectious virus. Science writer Christine Johnson says, "As it stands right now, bDNA uses QC-PCR as a gold standard; QC-PCR uses regular PCR as a gold standard; regular PCR uses antibody tests as a gold standard; and antibody tests use each other. I have noticed time after time that studies which are 'verifying' an HIV antibody test will invariable state that they evaluated the performance of their test on samples which are known to be TRUE-POSITIVE or TRUE-NEGATIVE. How did they know this? It's simple: Without a gold standard, they didn't."

Conlan MG. Rodney R. Why the "HIV tests" can't tell you whether you have HIV. Zenger's Newsmagazine. 2001. Available at www.healtoronto.com/rodney_richards2.html. Accessed on March 6, 2006.

Irwin M. Questions on HIV-antibody tests. (reprinted from VirusMyth.net, edited and updated for Alive & Well, May 2004) Available at www.aliveand well.org/html/questioning/questions_hiv_antibodies.html. Accessed on April 5, 2006.

Johnson C. Viral load and the PCR. Continuum. November 2001. Available at www.virusmyth.net/aids/data/chjtests5.htm. Accessed on March 11, 2006.