IS ANYBODY REALLY POSITIVE?

By Christine Johnson

HEAL Magazine 1995

Well before 1984, when HIV was announced to be the cause of AIDS, researchers had found evidence that antigen-antibody reactions were nonspecific for all viruses, including retroviruses. It is often believed that an antigen, a foreign invader which causes disease, and the antibody it elicits are "soulmates" and only react with each other. In reality, antigens and antibodies are not so selective. Often antigens cross-react with antibodies that they don't belong with and antibodies cross-react with antigens in the same way. Since the principle of HIV tests is that HIV antigen present in the test kit will react only with any HIV antibody that might be present in a person's blood sample, this is a crucial point.

Thus retrovirus antigens will react nonspecificafly with a multitude of nonretroviral antibodies, retrovirus antibodies will cross-react with a variety of nonretrovirus antigens, and antigens and antibodies to one retrovirus will cross-react with those of another. Considering this situation, it would be expected that the initial development of HIV antibody tests would have been done very cautiously, with insistence on verification by an independent method called a gold standard. A gold standard means that the reaction (a positive or negative test result) must be correlated with the presence or absence of HIV in the body.

Robert Gallo, who developed the initial HIV antibody tests, never used a gold standard to confirm his ELISA test. His methodology entirely ignored the necessity of proving that his subjects were either infected or not infected by use of virus isolation, matching the virus isolation (VI) test results with antibody test results. If VI positivity matched antibody positivity, and if VI negativity matched antibody negativity in all cases, he would have had an ideal test, proven by a gold standard. Instead, Gallo used a second antibody test, the Western Blot, to confirm his ELISA. At that time, and even to this day, the Western Blot was an entirely unproven test of unknown accuracy in this context, yet it was accepted without question as adequate verification of Gallo's ELISA.

The lack of a gold standard wasn't the only problem with Gallo's study. He used a group of AIDS patients in his sensitivity determinations and used random blood donors for his specificity deternminations. (Sensitivity is the extent to which a test will be positive in patients with the disease in question, in this case HIV infection, and specificity is the extent to which a test will be negative in patients without the disease).

When Gallo began his study, he simply assumed that all his AIDS patients were infected with HIV. After all, his hypothesis stated that HIV caused AIDS, and therefore they must be infected. Conversely, he assumed that 0 the random blood donors were not infected. He thought that since the donors were healthy and didn't have AIDS, they must not be infected. He then assumed that any negative test results in AIDS patients must be false negatives, and any positive results in the random blood donors must be false positives. Using these figures, determinations of sensitivity and specificity were made.

In essence, Gallo's study insisted that something was true, even if it wasn't, perhaps in the hopes that his belief would make it so. The reality is that Gallo didn't know if his Western Blot was accurate, he didn't know if AIDS patients were really infected, he didn't know if the random blood donors were not infected, and more to the point, he didn't know if HIV caused AIDS. There was a tremendous amount of assuming going on in Gallo's study, and in subsequent years, no other researchers or test kit manufacturers made any different assumptions when they were attempting to verify the accuracy of their HIV antibody test kits. They all used essentially the same methods as Gallo did.

Gallo's methodology was utterly without scientific merit and it was invalid for him to conclude that his ELISA test was "both highly sensitive and specific and should be considered a reliable initial screening test for the presence of antibodies, and hence exposure, to [HIV]" (JAMA January 11, 1985). ELISAs eventually became notorious for their high rate of false positives (which is why they are always supposed to be "confirmed" with Western Blot). Despite the high rate of false positives, much of the developing ideology concerning AIDS was based on data provided by these early tests.

Let's imagine for a minute that Gallo had been a true scientist and was not willing to assume anything:- Therefore, not assuming that HIV caused AIDS, he might have asked himself, "Since we noticed that almost all AIDS patients were positive on this test, and almost all blood donors were negative, there must be a reason for this pattern. What do these reactions mean in each of these unrelated groups of people? In AIDS patients, what is our test reacting with? Is it reacting with antibodies to a specific infectious agent that might or might not cause AIDS, or is it reacting to something else?"

The obvious answer to this question has eluded the AIDS research establishment simply because they continue to assume that a positive HIV antibody test in a person with AIDS is without question a true positive. This reasoning has also been extended to include anyone in a risk group. Certainly no one spends research time and dollars investigating whether a positive result in a gay male, intravenous drug user (lVDU), hemophiliac, or a blood transfusion recipient is a false positive. In fact, all these groups have something important in common and therein lies the answer to the mystery. Every one of these risk groups has been exposed to a plethora of foreign antigens and infectious agents and thus they have numerous antibodies to many non-HIV antigens. Many of these antibodies have the potential to nonspecifically cross-react with the antigens in the HIV test kits.

For instance, hemophiliacs, who routinely administer clotting factor, pick up whatever is in the blood plasma of 20,000-30,000 people used to make one vial of Factor VIII. Needle-sharing and multiple sexually transmitted diseases among IVDUs lead to a similar state. According to the comparative charts in Robert Root-Bernstein's book "Rethinking AIDS" (pages 165-169), gay males have "the highest disease load of any North American or European risk group," even greater than that of hemophiliacs (and comparable to the microbe burden of Africans).

Gay males have been repeatedly exposed to, and thus have antibodies to, a range of diseases and microbes, including cytomegalovirus, Epstein-Barr, herpes simplex I and II, hepatitis A and B, HTLV I and II, Giardia, amebiasis, toxoplasmosis, Chlarnydia, syphilis, gonorrhea, mycobacteda, and other bacteria. In addition, semen absorbed through the rectal mucosa during anal intercourse can lead to production of anti-sperm antibodies, which also can cross-react with the HIV antibody test.

Over the years, a lengthy list of diseases and conditions have been well-documented to produce false positives on HIV antibody tests. In particular, false positives have been noted in persons who have been exposed to hepatitis B or even vaccinated for hepatitis B. This disease is common among AIDS risk groups in the United States and Europe, but not in the general population. Many gay men and almost all hemophiliacs have been infected. Interestingly, hepatitis B is present at endemic levels in the general population of Asian countries as well. According to Bryan Ellison, coauthor of "Why We Will Never Win the War on AIDS," 99% of all people in the Far East are born infected with the hepatitis B virus, where it is transmitted perinatally and exists as a latent infection.

Cross-reactions are such a problem in the risk groups that Eleopulos et al. have stated:. "It is to be expected that cross-reactivity with HIV antigens would be the rule, not the exception, in these groups." (Unpublished letter to Nature). Lee et al. (Lancet April 25, 1992) and Moore et al. (NEJM May 22, 1986) have stated findings of "a high rate of false positives" and "very high false-positive rates for HIV in the high-risk groups."

With this overwhelming evidence, there is no reason to assume that a positive HIV antibody test result in a member of a risk group is a true positive. It is most likely that 'the opposite is true.

The testing situation is even worse in Africa. Mainstream AIDS researchers have commented from time to time that HIV antibody tests may not be specific for HIV in African populations. This means that the false positive problem is so extensive that most or even all positive results in Africa are false. In 1985, Hunsmann et al. stated "...African sera [blood samples] yield a high prevalence of false-positive reactions with the ELISA for [HIV]." (Lancet October 26). Labius Mutanda, a Uganda scientist, commented in The Blue Sheet (October 23, 1991) that "ELISA and Western Blot assays may not always be able to reliably ascertain HIV infections in many African individuals." Mutanda told me that he had observed Africa blood samples to be positive when tested with the kits of one manufacturer, whereas the same samples would be negative with the test kits from another company. He noted this occurrence with both ELISA and Western Blot. Even Gallo conceded that for African populations that did not compare to the AIDS patients and blood donors he studied, the ELISA test might need to be supplemented with a confirmatory test.

A recent study by Max Essex, a well-known AIDS researcher, reported high levels of false positives in persons with leprosy, as well as persons with whom they had contact. It was proposed that these cross-reactions occurred not only to Mycobacterium leprae (the microbe associated with leprosy) but to other Mycobacterium species as well , and that "ELISA and Western Blot may not be sufficient for HIV diagnosis in AIDS endemic areas of Central Africa where the presence of mycobacterial diseases is quite high." (J. Infectious Diseases, February 1994).

The implications of this are clear. Tuberculosis (caused by Mycobacterium tuberculosis) is a fairly prevalent AIDS indicator disease all over the world, especially in IVDUs. In many areas of the world, especially those with "new, rapidly spreading AIDS epidemics," tuberculosis is a common, endemic disease. Half of the world's reported (as opposed to estimated) tuberculosis cases occur in South East Asia. Tuberculosis is particularly widespread in Africa and is estimated to kill over a million people there each year. Another member of the Mycobacterium genus is M. avium, involved in Mycobacterium avium complex (MAC), which is the most common systemic bacterial infection in American AIDS patients.

Another disease which has been implicated in false positive HIV antibody tests is malaria. This may help explain why HIV appears to be transmitted predominantly via heterosexual intercourse in Africa and Asia, but not in the United States and Europe, where AIDS and HIV still affect mostly males. Since mosquitos do not practice gender discrimination, anti-malaria antibodies are common in men and women alike in malaria-endernic countries. Similarly, tuberculosis affects each gender equally.

Traditionally, AIDS and HIV have been quite rare in Asia but now the World Health Organization (WHO) gives us the alarming news that HIV is spreading out of control in this region. According to the Los Angeles Times (July 1, 1994), "the global AIDS epidemic is now spreading in Asia faster than anywhere else in the world," possibly soon to eclipse even the alleged monumental "heterosexual spread" of HIV in Africa.

However, when looking at WHO's map of world-wide malaria prevalence zones, it is interesting to note that the malaria belt and the AIDS belt coincide almost exactly. The countries mentioned in the Times article (Thailand, India, and Vietnam), as well as sub-Saharan Africa, all lie neatly within the malaria belt. There are high correlations between the AIDS belt and regions where tuberculosis is endemic as well.

In 1991, over three million cases of malaria were reported in Southeast Asia and over 20 million in Africa. WHO estimates that 300-500 million clinical cases occur each year, with countries in tropical Africa accounting for more than 90% of these cases. Southeast Asia reported over two million cases of tuberculosis and Africa reported 373,000 (though WHO believes several hundred thousand additional cases go unreported in Africa). About half of the AIDS patients in Africa suffer from tuberculosis (Science 1993, 260:1266).

Mass immunization programs against hepatitis B have been carried out in Asian countries in recent years. In Thailand alone, over two million people have been vaccinated. This, coupled with the endemic levels of hepatitis B outside risk groups, creates a large potential for false positives in the general population of Asian countries which is not present in the United States or Europe. Here, very few people outside the risk groups are exposed to hepatitis B or feel the need to be vaccinated for it.

For the above reasons, we must ask ourselves: Is it possible that the "heterosexual spread of HIV" in these areas is mainly due to the spread of HIV antibody testing programs?

A different problem relates to the general population of America and Europe, people who are not exposed to many microbes and who have relatively low levels of antibodies. What does a positive test mean in this group? Not very many people from the general population test positive, and starting with Gallo, mainstream AIDS researchers have usually held the opinion that a positive test result in any member of a low prevalence population is almost certainly a false positive. Langedijk states that in both ELISA and Western Blot "[a]lmost all reactions, especially in low-risk populations, represent false-positive results." (AIDS. 1994. 6:1547). According to Lepine (J. Clin. Micro. June 1990), in low-risk groups, the majority of positive ELISAs will prove to be false by failing to be confirmed by Western Blot. Even in cases where the ELISA has subsequently been "confirmed" by a positive Western Blot, the person is not likely to be HIV infected if there is no clinical or epidemiological information to substantiate the possibility of infection.

A classical principle of diagnostic test interpretation is that when a person has no logical reason to have the disease they are being tested for, a positive test result is most likely a false positive. As Griner states (Ann. Int. Med. April 1981), "When the likelihood of disease is low, a normal result tends to exclude but an unexpectedly positive result is not particularly helpful in confirming the disease." It is obvious that this precept is being ignored on a regular basis, when we see one anomalous transmission case after the other crop up in the media. People in no risk group, practicing no risk behavior, who have absolutely no reason to be HIV infected, have blamed their HIV infections" on such diverse and absurd sources as dentists, oral polio vaccine, and surgeons somehow picking up HIV from one patient and transmitting it to another during sequential surgical procedures.

If most or all positive results in any group are false positives, then, as a practical matter, how should one respond to a positive HIV antibody test? Unfortunately, the situation isn't simple and the answer depends on whether you belong to a risk group or not.

In low- or no-risk groups, a positive test is of no concern unless there is a logical reason for the person to be infected, a reason which makes sense without having to stretch the imagination to incorporate bizarre theories such as dental transmission. Even on the rare occasion when a positive might possibly be a true positive, one still has to grapple with the issue of whether HIV infection has any relevance to the development of AIDS whatsoever.

On the other hand, in high-risk groups, a positive test can serve as a surrogate marker for AIDS risk -- not AIDS, not HIV infection, not certain death and doom but simply an indication that the person has been exposed to certain pathological conditions that might possibly lead to AIDS or other illnesses. For these people a positive HIV antibody test should be viewed as a signal that it is necessary to assess their life-style and medical history for possible health risks.

The cross-reactions which produce false positives can occur as a result of many conditions, both pathological and nonpathological, so it is wise not to jump to any conclusions but rather correlate test results with factors such as symptoms, medical history, life-style, drug use, risk behavior, etc. In any event, a positive test result certainly cannot be used to justify a diagnosis of HIV infection or the initiation of so-called anti-viral therapies such as AZT, ddI, ddC, d4T, or other prophylactic pharmaceutical drugs.

In spite of their limited applications, HIV antibody tests have probably caused much more human misery and tragedy than any possible benefit they might provide as nonspecific markers of AIDS risk factors. I most vehemently beg to differ with Gallo's statement that his antibody test "has by now saved countless lives" and recommend that its use be entirely discontinued. AIDS should be diagnosed on the basis of clinical symptoms alone.

Christine Johnson is on the Board of Directors of HEAL/LA and a member of Mensa. She has worked with medical records for over 20 years and has done extensive research into the scientific literature on HIV antibody testing. Ms. Johnson regularly corresponds with scientists all over the world, including the Eleopulos research team in Perth, Western Australia. Her articles on HIV antibody tests have been published in England and Spain, as well as the United States.