Playing Russian Roulette in the Laboratory: Can You Really Trust HIV Testing?

By Christine Johnson

"It was the most terrifying experience of my life. I just freaked out."
"When she found out, she went up to the roof of the hospital and jumped off."
"To me, the emotional impact was immeasurable; the fear was unmanageable."

Are these people talking about being diagnosed with AIDS? No, these are the reactions of healthy people without symptoms whose HIV test came back positive. The effects of getting a positive HIV test are so profound that no one should bear that kind of burden without knowing with absolute certainty that the test is a true and meaningful measure of the presence of HIV in the body. This statement introduces and states the purpose of a ground-breaking paper from Australia, "Is a Positive Western Blot Proof of HIV Infection?" published in the June issue of the Bio/Technology journal.

The authors, Drs. Papadopulos- Eleopulos, Turner, and Papadimitriou, demonstrate that even though we have been told that we can depend on the accuracy of these "AIDS tests," in fact we cannot. They discuss the tests and how they came to be accepted as virtually unquestioned proof that a person is infected with HIV. They also critique these tests on several different grounds: 1) the tests are not specific, 2) there is no standard way of interpreting them, and 3) the results are not reproducible. In order for an antibody test to be scientifically valid, it must meet these three criteria, and these three in turn must be validated by an independant means, namely the "gold standard".

Specificity and the Gold Standard

To begin, what does "specificity" mean? Specificity is the number of negative test results you will get in people who have definitely been determined not to have a given disease. A test that is 100% specific is always negative when the disease is absent. There are no false positive readings. How do you determine if the disease (in this case, HIV infection) is there or not? HIV antibody tests are based on the idea that if the antibody to a virus is there, then a protein of the virus, and thus the virus itself, is there. So in order to see if the test is doing its job, it is necessary to have an independent means of verifying the presence of the virus in an antibody-positive person and the absence of the virus in an antibody-negative person. This independent method is called the "gold standard."

The only suitable gold standard is isolation of the virus itself. Since viruses need the living cells of the host in order to reproduce, they are grown in tissue cultures. HIV isolation means that the virus is taken from a tissue culture and separated out from everything else in the culture so that only pure virus remains.

Let's say that 100 people have been determined to be HIV-free according to a virus isolation test. These people are all given an HIV antibody test and this results in 90 negative results and 10 positive results (false positives). This gives the test a 90% specificity (which is not very good). If the test were 100% specific, it would never be positive in anyone (whether that person has symptoms or not) who is not infected with HIV as determined by a viral isolation test.

Although many claims are made that the HIV antibody tests have a high specificity, Eleopulos and colleagues argue that this is not the case. A test that is not specific will give a positive reading in the presence of antibodies other than those it is supposed to detect. Clearly, if an HIV antibody test is not specific, a positive result is at best ambiguous.

Antibodies and Antigens

Let's briefly review the definition of antibodies and antigens, since we'll be talking about them a lot. Antibodies are the body's foot soldiers in the war against foreign invaders such as bacteria and viruses. The immune system will make a set of antibodies that have a unique biochemical attraction to the proteins of an invading virus. When the antibodies see a virus particle passing by, they attach to the virus and render it harmless. As the need arises, the immune system will make many thousands of different kinds of antibodies, each of which will attack one specific antigen.

Antigens are substances that are foreign to the body, things that might make us sick or be fatal to us if our immune system did not neutralize them with antibodies. Antigens include such things as viruses, bacteria, toxins, or tissues from other people's bodies (such as sperm which might find its way into the bloodstream during anal intercourse). Antibodies fit together with their antigens like a key in a lock and this interlocking or binding action results in neutralization of the antigen so that it can no longer harm us.

The ELISA and the Western Blot

These principles are used in the HIV antibody tests, which basically work like this: In the ELISA, a mixture of proteins, which are said to come only from HIV, is exposed to a blood sample and any antibodies in the blood that can bind to these proteins are allowed to do so. If all of the proteins in the mixture really do come from HIV, and if all of the antibodies recognize only HIV proteins, a positive reading in the test would mean that at some time in the past the person had been exposed to the virus. But, as Eleopulos and her colleagues demonstrate in their rigorously argued and exhaustively referenced article, these two essential conditions are not met in either of the antibody tests (ELISA or Western Blot) currently in use!

So we have two big problems already:

1. Not all the proteins in the mixture definitely come from HIV;

2. Not all the "HIV antibodies" definitely recognize HIV alone.

In fact, Eleopulos and associates point out that there is no proof that any of the supposed HIV proteins actually come from HIV. The reason is that any attempt at virus isolation has presented numerous insurmountable difficulties (which we will discuss later). Their position is that HIV has never been isolated and thus one cannot say for sure that any of the HIV proteins actually come from HIV. It follows that if you cannot know for sure that the proteins in the antibody test come from HIV, then you cannot know for sure whether the antibodies that react with these proteins are anti-HIV antibodies. There are many documented cases of antibody cross-reactivity where antibodies to another disease or condition altogether will cause a false positive test result. The only way to know for sure is to apply a gold standard.

In the Western Blot (WB), the presumed HIV proteins are present separately instead of in a mixture and, after being allowed to react with a blood sample, each protein is capable of giving a visible signal if it has bound an antibody. (See Figure 1 for a picture of what a WB read-out looks like.) Thus, while an ELISA can only determine that a blood sample contains some antibodies which appear to have been elicited by HIV, a Western Blot, at least in theory, can determine which particular protein of the virus elicited which antibody.

Because ELISA tests are well-known to be less than perfectly specific, Western Blots, which are thought to be highly specific, are generally used to "confirm" them, especially when it is important to give the appearance that great scientific care and utmost accuracy have been exercised. In fact, so much weight is placed on these Western Blots that a positive result is almost always taken to equal an active HIV infection. So, how accurate are these "definitive" tests?

First, figure 2 has a chart showing which proteins must be present in order for the Western Blot test to be positive, according to each of five different authorities. You can sort of mix-and-match; pick at least one from each required category and your test will be interpreted as positive. More on this later.

The picture in Figure 2 shows how the Western Blot test is read.

Each HIV protein/antibody reaction shows up on the strip in its own special area, called a band. There are several bands which are supposed to represent specific HIV proteins. The most important of these are p120/160, p41, p31-32, and p24. In reality, these bands may not represent HIV proteins at all!

The Proteins are not Unique to HIV

Let's take a look at each of these main HIV proteins and see whether their presence on a Western Blot strip is actually indicative of the presence of HIV:

P41: Even at the beginning, Dr. Montagnier, the discoverer of HIV, found that blood from AIDS patients reacted with a p41 protein that was found in HIV-infected cells and non-HIV-infected cells. It was concluded that the p41 band was the result of contamination of the virus by a protein called actin, which is a normal component of all cells, as well as other microbes besides HIV. In other words, many other things besides HIV can give you a positive reading on the p41 band.

P120 and p160: P120 is found on the knobs on the surface of the immature HIV particles and p160 is found only in infected cells, but neither is found in the mature, free virus particles. However, the blood from AIDS patients reacts with the "purified HIV" in the AIDS test, and bands corresponding to p120 and p160 develop. This is a contradiction. Since no p120 or p160 is found in the mature virus, there would not be any in the test kit antigen preparation either. So where do these bands come from?

If p160 is split apart, you get p120 and p41. It is thought, then, that the p120 and p160 bands do not actually represent p120 and p160, but rather are what chemists refer to as oligomers of p41. An oligomer is an integral number of subunits of something else. If you put four subunits of p41 together, you get p160; three subunits of p41 equals p120 (4 x 40 = 160; 3 x 40 = 120). Therefore p120 and p160 are not different proteins at all; they are just several "packages" of p41 which are hooked together by chemical bonds. And if, as discussed above, p41 is simply a contaminant of cellular material, then this disqualifies p120, p160, and p41.

P31-32: Proteins are composed of amino acids. In 1987, a study was done by L.E. Henderson (1) which compared the amino-acid sequences of "purified HIV" with those of a normal protein found in the human immune system called "Class II histocompatibility DR proteins" and found the DR proteins to be identical to the p30-32 proteins of HIV. Good-bye, p31-32!

P24: Detection of p24 is considered to be synonymous with the actual presence of HIV. However, Robert Gallo (co-discoverer of HIV), has stated repeatedly that p24 is not unique to HIV, but that another retrovirus (HTLV-1, which does not cause any disease) contains p24, and this will cross-react on the antibody test. The test may say you have antibodies to HIV in your blood, but it could really be HTLV-1 instead.

In fact, a p24 has been found in HTLV-1, HTLV-II, HIV-2, and in all endogenous retroviruses. The specific gene which contains the information to make p24 (called the gag gene) is found in all retroviruses. Hence it is not surprising that a p24 is commonly found in the absence of HIV.

If a p24 band appears by itself on the Western Blot test without any of the other HIV bands, the test will be called "indeterminate." This means the test does not show enough bands to be sure of anything, but generally an indeterminate test is not a cause for alarm. P24 is the predominant band that shows up to cause an indeterminate Western Blot test. The indeterminate pattern is very common and does not correlate to the presence of HIV. Reference is made to a group of patients who received blood which had been tested by Western Blot and found to be negative; however, within six months, 42% of these patients had developed indeterminate WB tests.

Eleopulos and associates give a list of conditions where patients have p24 antibodies without HIV: multiple sclerosis, generalized warts, cutaneous T-cell lymphoma, and even 1 out of 150 healthy individuals. In the studies cited, p24 antigen was found in only 25-50% of the HIV+ patients or AIDS patients investigated. It was concluded that the P24 antigen test was "erratic" and "should be interpreted cautiously". Looks like we can't count on p24 either.

In spite of all the evidence to the contrary, the p160, p120, and p41 bands are considered to represent distinct viral proteins.

In conclusion, Eleopulos states that "(in) view of the above, it is difficult to defend the view that the bands p41 (and thus p120 and p160), p32, or p24 represent specific HIV proteins." In addition, she thinks that even if these proteins were shown to be specific to HIV, you could not assume that the antibodies that react with them are diagnostic of HIV infection (we will discuss cross-reactions in a minute).

As you can see from the table in figure 2, the different agencies (FDA, American Red Cross, Centers for Disease Control (CDC), CRSS, and World Health Organization (WHO) consider the Western Blot test to be positive if at least two or three bands are present, and you can take your pick from the four HIV proteins we mentioned. However, since the p31-32 band represents a cellular protein, and p120 and p160 are just different forms of p41, all you really have left to choose from are p24 and p41, both of which may or may not represent HIV proteins.

In summary, the Western Blot is based on detection of five primary protein bands: p24, p32, p41, p120, and p160. Data has been given to cast doubt on the authenticity of any of these proteins as always (or even part of the time) representing HIV proteins.

The Test is not Standardized

If this isn't bad enough, the Australian paper expresses concern that there is no standard way to interpret the Western Blot test. There are various patterns of bands. What do they mean? For an antibody test to mean anything, it must be standardized. As Eleopulos states: "(The test results must give the same meaning) in all patients, all laboratories, and all countries." In reality, the blood sample from an AIDS patient:

-- may not react with all HIV proteins;

-- may react with non-HIV proteins;

-- may give different results on different blood samples obtained from the same patient at different times.

Because of this, the various agencies established the criteria as set forth on the table. Depending on whose criteria you use, the same group of AIDS patients yielded anywhere from a 50% to a 79% positivity rate. Eleopulos says that "in the scientific literature, no strips have been published of a standard positive Western Blot." In addition, the band pattern obtained will vary with the temperature and concentration of the chemicals used in the test.

As Zolla-Pazner et al. put it, there is "confusion over the identification of these bands" which has "resulted in incorrect conclusions...". (2) You might hope that it wasn't your test that gave one of these "incorrect conclusions."

The Test is not Reproducible

Finally, a test must be reproducible in order to be valid. Using a single blood sample, the band patterns should be the same or similar from one test to another, or on tests performed at different labs. The CRSS did a study where they took two known positive samples and two known negative samples and sent them around to 19 laboratories for Western Blot testing. The same serum was tested multiple times in each lab and the resulting band patterns showed quite extreme variations both from one lab to the next, and from one test to the next in the same lab.

It must also be noted that the labs involved were "Reference Labs," which is what they call the top quality labs. Eleopulos remarked that these top labs make up only a small number of the total number of labs performing this test, most other labs not being as top-notch, and presumably getting more erratic results, with more false positives.

The False Positive

The issue of false positives is a big can of worms with these antibody tests. Massive ELISA testing in Russia resulted in 280 false positives for every 1 "true positive". In spite of much evidence to the contrary, there is an uninformed "general consensus that proof of the specificity of the HIV antibody tests is firmly established". (3)

The belief that the antibody tests are "98-100%" specific is based on the work of both Gallo and Burke and their colleagues. As a gold standard, Gallo used the clinical syndrome. All AIDS indicator diseases have been around for centuries, so their presence does not prove there is HIV in the body. Yet Gallo was saying that as long as a patient had what looked like AIDS, then a positive test result must be correct. The fact that a person has one of the diseases in the AIDS syndrome does not have any bearing on the accuracy of the HIV antibody test.

Burke tested a low-risk group of military recruits and found 15 positives out of 135,187 total. Rechecking each of these positive results against four other different antibody tests, he found that 14 were still positive on all four tests and 1 was negative on all four tests. He calculated the false positive rate to be 1 out of 135,187, or 0.0007%. Eleopulos's group criticizes Burke's methodology for determining the false positive rate. By definition, a true positive is a positive test occurring in a person who is HIV infected as verified by an independent test (the gold standard); a false positive is a test in a person who by application of the gold standard, does not have an HIV infection. The four other antibody tests that Burke used to confirm his WB results do not provide any independent evidence of HIV infection or absence thereof. These four tests and the original WB test are all looking for the same antibodies -- they are all essentially the same test. A test cannot act as a gold standard for itself. Therefore, since Burke did not use a gold standard, it was impossible for him to measure the false positive rate for his WB tests. The authors do not estimate the actual false positive rate, but other sources have put it as high as 90% and it may be even higher. (11)

Multiple Testing

Another problem seems to be that if you keep repeating a person's initially positive test, a significant number of people will eventually end up testing negative. If you give two ELISAs and then two WBs in succession, many people will test negative on each subsequent test, so by the time you get to the fourth test, only a few will still be positive. If this last group of people were to be given only one, or even two, antibody tests, they would be considered to be HIV infected, when in fact they are not.

This is a serious flaw of the antibody tests. The official screening test is the ELISA, which has an astronomical false positive rate, and the CDC's current AIDS definition accepts a single ELISA test without confirmation. In other words, a single positive ELISA (plus an AIDS indicator disease) gets you an official AIDS diagnosis.

In practice, a positive ELISA may or may not be confirmed with a Western Blot, and in any event, as we have just seen, a single test, or even several more subsequent tests, may all be false positives and a person may go on to get a negative result on the third, fourth, or fifth test, and so on. So a simple screening test or even a screening test plus a confirmatory test (ELISA + WB), may often not give the true picture.

Cross Reactions

Neither Burke et al. (4) nor Gallo et. al. (5) tested people who had "other diseases where antibodies, some of which may interact with HIV antigens, may be produced for other reasons."

Eleopulos discusses the phenomenon of "biological false positives (BFPs)," false positives which will result in patients with any number of conditions unrelated to the condition being tested for. This is a common phenomenon in syphilis tests, for instance. I had a BFP reaction on a syphilis test once myself due to a previous case of cat scratch fever. BFPs to syphilis may occur in patients with lupus, auto-immune hemolytic anemia, idiopathic thrombocytopenic purpura, leprosy, or in drug addicts.

Some groups of people generally make a lot of antibodies because they are exposed to a much greater than normal amount of diseases and unhealthy conditions. This includes poverty-stricken Africans and IV drug users. According to Biggar et al., "reactivity in both ELISA and Western Blot analysis may be nonspecific in Africans..." (6) And, of course, all our ideas about a huge African AIDS epidemic are based on ELISA testing programs.

The scientific literature is full of references to data which show "the widespread presence of nonspecific interactions between retroviral antigens and unrelated antibodies." Eleopulos cites a long list of cross-reactions. A subject was injected repeatedly with HIV negative donor blood serum and his WB test was initially negative, but with each injection, it became more strongly positive. Mice injected with uninfected T-lymphocytes from another mouse strain developed antibodies to HIV! The test also cross-reacts in malaria patients, 25-41% of Venezuelan malaria patients being Western Blot positive, but with no AIDS.

Isolating Virus

Eleopulos and associates believe that all these difficulties with the specificity of the WB could be avoided by use of HIV isolation as a gold standard, which is the only acceptable method; however, "This has not yet been done, and may not even be feasible".

HIV isolation means to separate the pure virus particle from all other material in the cell culture. To use virus isolation as a gold standard, you need to be absolutely certain that the material you have isolated is actually the particular virus you are looking for and not anything else -- not another virus, not fragments of cells, not "virus-like" particles, and so on. The method used to isolate the virus involves separating out the liquid part of the cell culture (the supernatant) and spinning it in a centrifuge. This separates particles according to their different densities. It's sort of like telling the difference between two identical-appearing balls, one made of steel and the other of plastic, by throwing them into a swimming pool -- the steel ball sinks and the plastic one floats. See Figure 3 for an illustration of this principle.

It is believed that the material which settles at a density of 1.16 gm/ml is composed of virus particles and nothing else. The problem here is that some parts of cells have the same buoyant density as viruses. Some eminent virologists have pointed out that one cannot avoid contamination of the viral preparation with various other types of cellular material and cell fragments, since cells are going to rupture in the process. Since some of the cellular material has the same density as the virus, if you have a band showing up at 1.16 gm/ml, you still cannot be sure what it is. Because of this, researchers in the 1970s who were trying to isolate pure retroviruses visually confirmed their findings by using the electron microscope.

Using the above techniques, the best results could be obtained with supernatant fluids that contain a lot of virus and very few cellular contaminants. These conditions can best be satisfied by viruses which do not kill the cells they infect and in culture conditions where most cells remain alive (and intact) during infection. Here, retroviruses fit the bill. These properties of retroviruses make it much easier to separate them from everything else, including cellular contaminants. This would aptly describe most retroviral tissue cultures. However, just the opposite is the case with HIV.

Cellular "Lysates"

AIDS tissue cultures have very little virus, so little that it is hard to detect at all. The cells in these cultures are difficult to keep alive. Gallo, Montagnier, and other AIDS researchers have always found the concentration of virus-like particles in their tissue cultures to be very low. So here you have lots of cells and not much of these virus-like particles which they have called virus particles (it is pointed out that just because these particles looked like a virus does not constitute proof that they were in fact viruses). This situation will give you the least accurate results in terms of separating pure HIV particles from everything else that bands at 1.16 gm/ml.

In addition, the official party line on HIV is that it, unlike other retroviruses, does kill cells. There are many theories as to how it does this, including "apoptosis," a sort-of cellular suicide. If HIV is responsible for cell death one way or another, the result would be that the dead cells would have ruptured and the supernatant fluid would thus be full of cell fragments and cellular material from the ruptured cells.

Since some of this cellular material has the same density as retroviruses, then it is difficult to know what you really have showing up on the band that supposedly represents "pure HIV."

Eleopulos and colleagues contend that no one knows what particles, if any, band at 1.16 gm/ml and no one knows the density of what is called HIV. They state that "most if not all claims of 'HIV isolation' have been from cellular lysates." (3) (A lysate is the material formed by the fragmentation of cells.) In fact, "the available evidence... indicates that only about 20% of the proteins which band at 1.16 gm/ml are 'HIV proteins;' the rest are cellular..." (3) This situation is very likely to arise since "HIV-infected" cultures die and cultures are most often deliberately lysed.

To sort out this confusion, it would help to have a look at this material with an electron microscope, but, as Eleopulos and colleagues note, the AIDS literature contains no electron microscope pictures of the material that bands at 1.16 gm/ml, and there is no way for us to know whether this material "contains any such particles (pure HIV particles) whatsoever." They point out that the literature uses terms such as "HIV," "HIV isolation," "pure particles," "virus particles," and so on, and that these terms have a variety of meanings, but "most often without proof of the presence of a (virus) particle."

Claims of HIV Isolation

There are many accepted "proofs" of HIV isolation which turn out to be no proof at all. For instance, there is a substance called template-primer AndT15, which is copied if it is incubated with the supernatant or with the material that bands at 1.16 gm/ml. This activity is considered to be proof of reverse transcriptase activity and thus of HIV isolation. However, the same template-primer is copied when incubated with several other types of cells that are not infected with HIV, including normal uninfected spermatozoa. It is copied by reverse transcriptase (which is found in retroviruses) but it is also copied by cellular DNA polymerase (which is not found in retroviruses).

Certain particles are detected in AIDS cultures and these particles are considered by most AIDS researchers to be HIV. However, there are a lot of situations where HIV particles are found alongside non-HIV particles, or alongside "virus-like particles" that are "some what different" from what are generally accepted to be HIV particles, and so on. HIV is a Type-C retrovirus and type-C particles appear in non-HIV infected lymphoma cells which are metabolically impaired. "Retroviral particles" that have similar antigenic properties to HIV have been found in patients with Sjogren's syndrome. Virus particles indistinguishable from HIV are found in a variety of non-HIV associated lymphadenopathies. This goes on, with the conclusion that "the presence of such particles (does not, by itself), indicate infection with HIV." (7)

Again, the Gold Standard

It seems to be almost impossible to isolate HIV and know for sure that you have HIV and not any of the many above-mentioned entities. And even though researchers have accepted a wide variety of phenomena as representing HIV isolation, and have gone to extraordinary lengths to isolate the virus, "it is still not possible to isolate HIV from all antibody positive patients." (3) Conversely, with the same effort, HIV can be isolated from patients who do not have AIDS and who have a negative AIDS test! Using detection of p24 as a method of HIV isolation, positive results have been found in a large majority of "presumably uninfected" people with indeterminate Western Blot tests, and in all of a group of HIV negative blood donors.

There is absolutely no correlation between "HIV isolation" and a positive antibody test, and the CDC admits this. The CDC calls a positive antibody test a "documented infection," and yet "the virus cannot be detected in every person with a documented infection." (8) Therefore using HIV isolation as a gold standard to authenticate the validity of the antibody test is problematic at best.

Genetic Sequencing

Another way of trying to find the virus is by genomic investigations. The goal is to try and prove that AIDS patients have been infected with a unique, exogenous retrovirus. (Exogenous means originating outside the body, as opposed to endogenous, which means originating from within the body.) A genome is simply the total package of hereditary information found in the genes, which are made up of DNA or RNA. It seems that if you could determine that a certain retrovirus has a distinct genetic sequence, then when you find this genetic sequence, you could say you have found the virus. Well, it really isn't that simple.

Information about the genome is contained in RNA or DNA. A retrovirus does not have any DNA; it only has RNA. The way a retrovirus infects a cell is by using reverse transcriptase to turn its RNA into DNA, and by turning this DNA loose in the cell, where it then becomes integrated with the DNA of the cell. This DNA translation of the retrovirus's genetic information can then be used as a pattern, or template, to produce additional copies of the virus, which are then sent forth from the cell. The retroviral genome, when integrated into the cell's DNA, is called a provirus.

Various phenomena are discussed that provide a "stumbling block to any genetic analysis of (retroviruses,)" (9) but the main trouble spots are as follows:

No two HIV genomes are the same

"No two identical HIV have been isolated even from the same person," (3) at different times or at the same time. In the same patient, HIVs from different types of cells are different. Different types of cells used in the cell cultures produce different HIVs.

HIV sequences cannot be found in all AIDS patients

Try as they might, AIDS researchers just cannot find much virus, and in a lot of cases, they cannot find any! The polymerase chain reaction test (PCR) was introduced to improve detection of viruses. It is supposed to be the equivalent of being able to find a needle in a haystack, since it can find one gene or fragment of a gene and amplify it to the point where you can now detect it. This is the test you now see in all the ads, promising you accurate results within four weeks of infection.

Even with such a test as PCR, there "is the scarcity or apparent absence of viral DNA in a proportion of patients." (10) PCR cannot detect HIV in the majority of semen samples from AIDS patients. What does this do to our theory that HIV is primarily sexually transmitted? If PCR cannot find it, it is simply not there to be found.

However, Eleopulos and associates comment that even with PCR, there is some confusion as to the meaning of the test results, especially when one is trying to use PCR as a gold standard to confirm the actual presence of virus in the body of an HIV positive person. Blood samples were tested using PCR and the standard ELISA tests and the results were compared. From one lab to the next, correspondence between the two varied between 40% and 100%, which means that between 0% to 60% of the time, one or the other of these tests must have been wrong. False positives and false negatives on PCR were also observed.

The positive hybridization results may not be HIV specific

Hybridization is a lab technique which is useful in identifying cells replicating HIV. On Gallo's first hybridization studies, he found the bands were "faint" or "low signal." He thought this meant that there was not very much virus but conceded that it might mean the test was picking up a homologous virus (a different virus that is very similar in structure and origin). This paper argues that the latter is true. HIV-related sequences have been found in normal cells as well, so many HIV phenomena could well be cellular in origin.

Eleopulos and her co-authors conclude by stating that the use of antibody tests to predict or diagnose HIV infection, or for epidemiological surveillance, "needs," as she politely puts it, "to be carefully reappraised." I will be less than polite and tell you that these tests are not accurate; they are a treacherous deception and making any life-and-death decisions based on a positive antibody test is a very foolish thing to do. It can only lead to tragedy. *

REFERENCES