See the first addendum to this little missive for further developments since I wrote this article (written in 1992).
About 10 years have gone by since HIV was first discovered and isolated, and there has been a lot of debate recently about whether HIV really causes AIDS. The chief instigators of this idea have been Peter Duesberg, one of the world's leading experts on retroviruses, and Kary Mullis, winner of the Nobel prize in Chemistry for PCR, a technique that is used greatly in AIDS research to enhance copies of DNA infected with HIV (which is normally extremely hard to detect) for analysis. Claims and counter-claims, made by the various parties involved in this field, are abound and I'll attempt to address some of them in this missive, and examine the validity of each claim.
HIV alone cannot be responsible for the depletion of the T-cell count that we see in AIDS patients: This is the strongest claim that can be made toward the argument that HIV does not cause AIDS. Both sides of this issue agree to this; Gallo and Montagnier (the two founders of HIV), in the October 1988 Scientific American (an issue worth checking out simply because it was entirely devoted to AIDS and demonstrates how far we have come from that time), try very hard to convince us that HIV does indeed cause AIDS by providing tons of circumstantial evidence in this regard. But they do acknowledge, in the end, that even though HIV kills T4 cells, the direct killing is not adequate to explain the massive eradication of the immune system.
HIV is very hard to find in AIDS patients: This was once an extremely valid claim since Gallo and Montagnier had an extremely difficult time isolating it, and it is still valid, but not as much. A couple of papers published back-to-back in the May 1993 of Nature purports to explain this phenomenon. The papers essentially claim that HIV is present in great amounts in the lymph nodes (even though the papers disagree on the time during which this happens as HIV infection progresses), than in the blood, and these particles are carried to other HIV target cells. But it is also true that some people diagnosed with having AIDS have latent or inactive HIV. And even a single exception to any hypothesis should be adequate to question it, following a philosophy of science in the Popperian vein.
99% of people with AIDS have HIV: Here we run into slightly muddy waters. It is true that almost anyone who is diagnosed as having AIDS produce antibodies to HIV. The 1% discrepancy is explained away as errors due to testing techniques, and this is probably true. But the definition of when someone has AIDS is the one that is suspect. Many AIDS-like cases have been reported but are not diagnosed as AIDS simply because of an absence of an immune response to HIV. This is the medical profession failing to be self-critical---and this in itself says something about the whole nature of this controversy. It becomes an issue in semantics and thus when appropriately defined, all AIDS patients end up having been exposed to the virus. Yet, there is no doubt in my mind (contrary to claims by Duesberg and Mullis) that the correlation is adequate that any exposure to HIV is likely to cause AIDS---the only issue is the actual details of the mechanism by which it operates.
The sad fact is that the diagnosis for AIDS has become whether the patient has an HIV immune response or not (especially since there is demographic variance in the symptoms of the AIDS disease, which is not seen in any other disease known), and this should never be the case for any disease where the causative agent is not established and the mechanism of pathogenicity is unknown. Another problem is the fact that an antibody response indicates that the immune system has won against the invader, not given into it, but this argument has another side: HIV almost always enters into the lysogenic cycle at the beginning of each infection (this is when it remains in the DNA inactive while the cells keep reproducing). But I think most agree that the Koch's postulate which states that every person with the disease has the believed causative agent has been fulfilled adequately, at least as adequately as any other disease we know. But what of his next postulate?
A lot of people with HIV do not have AIDS: About a third of the HIV+ people, over a million Americans, to be precise. And this is a violation of Koch's second postulate. Of course, excuses in the form that Koch's postulates are outdated or too rigourous are abundant; a reason given for this fact is that everyone with HIV will eventually develop AIDS. This "eventually" has changed from 1-2 years (the number initially given) now to 5-10 years (most recently, as of 2000, as much as 20-30 years). Again, this is an example of non-rigourness that is at the heart of this argument. A million Americans alone with HIV with normal T-cell counts. Six million Africans supposedly. One should answer this discrepancy in a robust manner.
AIDS is not an infectious disease and caused by various toxins, including AZT itself: this is dependent on the definition of an infectious disease, wherein the distribution of AIDS is equal among all members of a population. If HIV were the causative agent, it should not care about whether you are male or female or heterosexual or homosexual, but statistics show it seems to. But this is nitpicking, because the mode of infection can be different and doesn't necessarily have to be the same and doesn't necessarily have to be distributed. This is like saying any other STD isn't an infectious disease (even though most STDs are distributed pretty equally among the population). But this is where I think Duesberg and others are wrong in their viewpoint. The problem is that he is forced to answer the question "so what causes it if isn't HIV?" Instead of following his own advice, he comes up with an hypothesis that even less justifiable than the HIV one. The fact that we have to question the HIV hypothesis is indicative of the shaky ground we're on, but Duesberg doesn't do his position any good by coming up with unsubstantiated theories.
Related viruses to HIV cause AIDS-like infection in other animals: this is true, especially in the case of the Visna (which is the Icelandic word for "wasting") virus that killed hundreds of sheep in Iceland. Similarly, the simian immunodeficiency virus (SIV) causes a full blown AIDS-like disease in Asian Macaque monkeys, but not in certain other species. Everything I've said so far has been about HIV-1, there is another less virulent strain, HIV-2, endemic to certain populations in West Africa. All this represents pieces of a fascinating jigsaw puzzle that we do not know how to put together yet.
HIV always causes AIDS: We do not know and we could be wrong about this. There is no gene in HIV that has been determined to be harmful, and the mechanism of pathogenicity is completely a mystery, yet people spend billions of dollars on basic research devising up various ways to combat HIV. Huge egos are at stake and Duesberg is the constant target of attack by Nature editor John Maddox (who interestingly doesn't let Duesberg respond to his barbs). The fact remains, however, that we are really no closer to finding a cure for AIDS than we were 10 years ago (my opinion is that the only cure is to live with HIV; if we can live for ten years with HIV without developing AIDS, then why not try to figure out how to extend that to live to a hundred).
The social issues do not concern me, however; more people dying means a reduction of the population and while this may sound callous, this is ultimately an evolutionary tenet (I believe HIV evolution is a feedback response human population growth--it doesn't make sense to have an evolutionary dead-end like HIV exist otherwise). What I am interested in is simply the basic truth as to what really happens with AIDS and whether effort should be spent on combating HIV when it is clear that there is at least some doubt. It simply means we should spend more of our time on the intellectual challenge that is AIDS pathogenesis. This also doesn't mean we should disregard all the caution that we use to make sure we don't become HIV+, since there is an enormous correlation between AIDS and HIV. Even if HIV is along for the free ride, it suggests that you will with extremely high probability end up with AIDS if you engage in behaviour that leads to HIV+. But every student of logic knows that correlation does not imply causality; unfortunately this is the path which we have chosen to take.
All my sources are from Nature, Scientific American, and an Interview with Duesberg in the East Bay Review. I was firmly convinced at one point that HIV causes AIDS and spent a lot time working on HIV and its evolutionary relationship to other retroviruses. But basic research on how HIV causes AIDS has not turned up anything. Perhaps it is time to admit that we could be wrong and look in another direction as to molecular causality.