From: sbharris@ix.netcom.com(Steven B. Harris) Subject: The AIDS Heresies ((all articles, combined)) Date: 22 Dec 1996 Newsgroups: misc.health.aids,uk.gay-lesbian-bi,misc.health.alternative,sci.med THE AIDS HERESIES Introduction: The following paper, _The AIDS Heresies_, was written in response to a large recent increase in unscientific and frankly paranoid theorizing about HIV and AIDS. The recent publication of a book by prominent retrovirus researcher Peter Duesberg, Ph.D., claiming that HIV does not cause AIDS, and the well- publicized statements of a Nobel Prize-winning chemist (Kary Mullis, Ph.D.) also publicly questioning the HIV-AIDS theory, have both contributed to a certain amount of skepticism within the HIV-infected community, if not the general public. It has become clear that some kind of detailed response is needed to Duesberg and his followers, to be used by professionals and laymen alike in the always difficult question of what to do about AIDS. This essay was begun at the end of 1994 with this end in mind. The issue of whether or not HIV causes AIDS is a life-and- death one. Until recently, the treatments for HIV infection were only moderately better than doing nothing, but there was still the question of prevention. It seems likely that the sexual behavior of someone who knows themselves to be infected with a virus which is deadly, versus the behavior of someone who is convinced his or her infection is less serious than a cold, will be quite different. To this end, it is likely that skepticism about the cause of AIDS has already contributed to many HIV infections, and many deaths, as people carrying HIV simply lie to themselves and their partners about carrying the virus. "Sex will not kill you," states Peter Duesberg, who recommends that people worry only about previously known, non-fatal sexually transmitted diseases. But such a low and inappropriate level of concern will be expected to be a disaster in a venereal plague which is, in fact, deadly. Progress against AIDS continues, and in late 1995 and early 1996 a new class of antiviral drugs (HIV protease inhibitors) was introduced. These antivirals have shown the remarkable ability to inhibit HIV replication almost completely, for periods of up to a year and a half, and perhaps far longer. Addition of these new drugs to the medical regimen of tens of thousands of people dying of AIDS has already caused many to regain weight and health, and, to their astonishment, the will to continue with their previous lives. Smaller AIDS wards across the country are literally being shut down for lack of sick patients. Yet in the face of such gains the critics continue to tell HIV-infected people that if they take no drugs of any kind they will surely live, but that if they agree to medical treatment, they will surely die. Never before has the question of belief in one of the more difficult theories of medicine been so important to so many people. This essay tells the story of the discovery of HIV and its lenitvirus relatives. The reader who perseveres will gain an understanding of not only HIV epidemiology and history, but also some understanding of the difficulties of using the inferential scientific method. When does a critic of science become a crank? What makes the establishment so sure about the role of HIV? How much room does the evidence give for alternate theories? All these questions will be addressed. This essay, written as it is for net distribution, is a work- in-progress. The author would be glad for any feedback from any reader. Steven B. Harris, M.D. Salt Lake City, Utah December, 1996 Email: 71450.1773@compuserve.com sbharris@ix.netcom.com N.B. The author is an internist, clinician, and experimental gerontologist who receives no grant money from the government AIDS research establishment, and does not feel himself constrained in his opinion on this subject in any way by his academic status or occupation. This draft information brief may be freely reproduced and distributed in this ASCII form by anyone without license, so long as the ENTIRE contents are reproduced, including author copyright information and notes, and so long as reproductions are not sold for money (this does not include normal connect time charges on computer nets for downloading, but would include any special charges for the article itself on a BBS). Those seeking other arrangements should contact the author. Illustrations and figures for some of the concepts are available, and can be obtained by contacting the author. A shorter edited and illustrated version of this paper appeared in SKEPTIC Magazine in vol. 3 no. 2, 1995, with follow-up letters and an author reply in SKEPTIC vol. 3 no. 3. For SKEPTIC information call (818) 794-3119. THE AIDS HERESIES A Case Study of Skepticism Taken Too Far (c) 1994-1996 Steven B. Harris, M.D. _Felix qui potuit rerum cognoscere causas._ [Fortunate is the man who understands the causes of things.] Virgil "It's the virus, stupid." Dr. David D. Ho AIDS Researcher A Dialogue In Inductive Frustration Let us suppose that you have a bright and iconoclastic friend who smokes three packs of cigarettes a day. You remark one day that you would like to see your friend quit the habit, since he is certainly increasing his chances of lung cancer. "Prove it," says your friend. "Well," you begin, "the Surgeon General and a lot of scientists and doctors say you should quit...." "Come now!" says your friend, "Since when did you become a fan of The Argument From Authority? I can find you scientists who do NOT believe I necessarily should quit, too, as well as a lot of intelligent business executives." "Sure, but all those scientists and all those executives are paid by tobacco companies or grants from the Tobacco Institute," you protest. "Well, what do you expect?" says your friend, lighting up and taking a satisfying drag. "Whenever any scientist takes an anti-establishment, anti-government position like that, all grant funding is cut off. Didn't you know that? The poor scientists then don't have anyone else to support their research _but_ the Tobacco Institute. Do you expect them to starve or drop out of research just because they hold unpopular opinions?" "Okay, let's look at the facts," you say. "What about the fact that 90% of lung cancer occurs in smokers?" "Yes," says your friend, "and that means that 10% of it occurs in non-smokers, doesn't it? Obviously the `cigarettes = lung cancer' hypothesis doesn't explain all lung cancer, and even for smokers there must be `co-factors.' Heck, my grandfather smoked 3 packs a day right up to the day he was hit by a drunk driver at the age of 92. A lot of people smoke for a whole lifetime and never get cancer." "Look, I didn't say the association was statistically perfect!" you protest. "But it is certainly there. Two-pack-a- day people have 13 times the lung cancer risk of non-smokers." "Oh, really?" your friend says, "Now, where do you get that number? I suppose somebody did an experiment where they got together a group of nonsmokers and randomized them to start smoking, or else stay smoke-free, and then made sure each and every person did as told for the next 40 years, so as not to bias the results. I must've missed that study." "You know there is no such study. That experiment would have been impossible, since you can't enforce a random protocol like that. People will start or stop on their own. And besides, any experiment where you try to keep people from quitting would be immoral, since smoking causes cancer." "Sort of circular argument there," sighs your friend, admiring a smoke ring, "since that is what remains to be proved. So you admit you don't have any study where the two groups of smokers and nonsmokers are exactly equivalent, and only differing by chance or random draw? In every study the smokers and the nonsmokers are self-selected for their behavior, aren't they? And bound to be different not only in smoking behavior, but also because of whatever made them smoke or not smoke to begin with! So your experimental `controls' are inadequate, even if I do say so. Not exactly great science, if you ask my layman's opinion." "But darnit, when smokers quit, we know their risk of dying drops!" "You mean with regard to the smokers who don't quit? So what? The people who quit smoking did so for a reason other than chance or the experimental flip of a coin, I'm sure, and again that means they will differ in some way OTHER than their not smoking. Again you don't have a proper "control group" of non- smokers who had that very same reason, but were _prevented_ from quitting. Look here: did you know that for the first year after quitting, the risk of death for a new quitter actually goes UP with regard to his fellow smokers who keep right on smoking?" "I knew you'd bring that up. The mortality goes up for the quitter group for a while after they quit only because those people who quit are quite often sick, and that's WHY they quit." "If so that makes my point about self-selection, doesn't it? These aren't true controlled experiments. You're saying that in that first year of quitting, the higher death rate of quitters is caused by another factor in our study other than smoking-- namely, sickness. Well, so long as we're talking about such third factors, I have a hunch that _stress_ causes cancer, and stressed-out people take up smoking to try to relieve the stress, and that's why there is more cancer in smokers, not because of smoking. Moreover, maybe the act of quitting stresses people out, and that's really why quitters die faster in that first year after quitting. Smoking is just a marker for stress, you see-- what you statistics people call a "proxy variable." "All this is ridiculous! You're just using your intellect to make you believe something you want to believe for other reasons. There is experimental evidence! Smoking causes lung cancer in lab animals! Are THEY stressed?" "Actually, yes-- have you seen what they do to them in a modern lab? Ever seen one of those rabbits with a leather muzzle over its nose, and a cigarette stuck in it which it can't take out? But anyway, I don't even believe you can find me a report of an experiment in which smoking causes lung cancer in animals." Back you go to the scientific literature. And you find---- nothing. There is no such paper.... ----------------------------------------------------- Medical Induction. Because there are many intellectual steps which are not perfectly secure in any generalization, even the most detailed inductive argument only goes so far toward proof, as our fictio- nal dialogue, which is based on facts, demonstrates. It is also unavoidable that the same evidence will mean different things to different people. In particular, it is more difficult to induce a person to follow a complicated inductive-reasoning trail when they dislike, or are threatened by, the conclusion at the end. In the medical sciences, assembling an irrefutable argument for causation is sometimes an impossible task for the same reason it is in astronomy or paleontology: the direct and definitive experiment cannot be done. Scientists cannot travel back in time to watch dinosaurs, nor can they influence the behavior of planets or stars. In medicine, a common difficulty is that the necessary human interventive experiments to perfectly assess "risk factors" for harm may be unethical, and so these risks cannot be studied directly by experiment either [1]. How, then, do we come to "know" (or confidently believe) what things cause lung cancer or AIDS? For that matter, how do we come to know with any confidence that tyrannosaurs ate meat, or what generates the sun's energy? In other words: how _do_ we ever infer causation from looking at events (or records of events) which we cannot influence? However we do it, it does seem that it can (to some extent) be done. Modern science depends on the fact that "correct" causal relationships can often be guessed entirely from logical and indirect observational tests of competing theories, even where direct experimentation is not possible. This is done using help from knowledge of simpler causal mechanisms which we have gained from similar systems in which experimentation _is_ possible. One of the most amazing things about the universe is that this kind of inference is possible at all, as Einstein once commented. Of course, the overall results of this kind of theorizing, like those of any inductive process, are never certain. Still, whenever inferential theories in science finally _do_ become directly testable by some new experimental technique, they often prove to be surprisingly sound (or "robust," as scientists say). Why this should be true remains the deep mystery that it was for Einstein. We are only little more than re-stating the mystery when we note that the universe seems to have a certain degree of uniformity and symmetry at many levels, allowing us to correctly guess at the structure of some levels of uniformity, on the basis of knowledge of others. Human intuition, trained by a core of more secure direct experimental knowledge, does seem often an adequate tool for doing some of this kind of guessing. On the basis of our past success, we expect (on the basis of the symmetry of past and future) that the present inferential methods of science will _continue_ to result in progress toward truth in many fields of knowledge. This progress in understanding will occur even before final experimental "proof beyond all reasonable doubt" arrives in each case. It is because of an inferential process, based on many lines of evidence, that we can be reasonably confident of the tobacco causation of much of lung cancer, even in the absence of a definitive experimental study. In the same way, an examination of a large body of related facts allows us reasonable confidence about the causation of other diseases-- even a disease far more complicated than lung cancer, and with even more money and passion involved on either side of the issues. The HIV-AIDS Hypothesis, and Its Challengers. Recently, several popular lay publications (Reason Magazine, Spin Magazine, the New York Native) have run articles calling into question the theory that the viral agent with the conclusi- on-asserting name, the "human immunodeficiency virus (HIV)," is the cause of the epidemic of human acquired immune deficiency disease, known as AIDS. What do we mean by talking of the "cause" of AIDS? We know that the common cold or the flu-- indeed all infectious diseases- - are in some sense "caused" not only by the pathogenic organism. Also important in the causal chain are host factors (such as immune response), and even simple host-overwhelming factors, such as the infectious dose of organism which enters the body (called the "inoculum"). These additional causal factors, which have nothing to do with the microbe itself, can be extremely important. They may in some cases outweigh everything else. Because (for example) the smallpox virus is necessary for smallpox, if not sufficient, medical science still regards it as "causal" in the sense that if there is no microbe, there is no illness. Eliminate the smallpox virus from the population and one eliminates the disease, as was in fact done in the 1970's. Yet even this kind of "causal" connection between a disease syndrome and infectious agent is what is under attack in recent articles about HIV and AIDS. Some skeptics have claimed not only that HIV is not the only external factor necessary for AIDS, but perhaps even that HIV is not always necessary for AIDS, so that if HIV were eliminated from the Earth, at least some AIDS would still be with us. Still others have gone further and claimed that HIV infection is totally harmless and never even _contribut- es_ to the development of AIDS. These people believe that if HIV were to disappear, AIDS would continue exactly as before. In what follows, we will examine the best evidence behind what most researchers believe is the role of HIV and other factors in AIDS. We will also examine leading skeptical views on the causation of AIDS. Because a great deal of published research is available on this issue, our examination of AIDS will also let us illustrate how science closes in on cause and effect, even when direct experimental "proof" is not available. We will thus be interested in not only AIDS, but also larger questions about science, and scientific debate. What makes a good scientific theory, and what makes a poor one? Are there reasons for hope in looking at the disease of AIDS in particular, and the workings of the biomedical scientific "establishment" in general? Or has science been completely corrupted in working on the cause and cure of AIDS, as many critics suggest, so that with this particular medical problem the scientific method has ceased functioning? Are we making any progress with AIDS, or just wasting billions each year chasing fantasies? This essay will argue that we are not wasting our money, and that when it comes to critics of the HIV/AIDS hypothesis, we have a practical case in which scientific skepticism has been taken too far. Science, we are happy to report, still works, and it is making progress with AIDS. That some critics have failed to recognize this only highlights the fact that science is only partly an empirical enterprise, and that it also has an intuitive and aesthetic side which is subject (to a certain extent) to arguments over taste. This is not a thing which is taught about science in school, but it is a concept key to understanding most scientific controversies. What Is This Thing Called AIDS? Scientific problem-solving begins with definitions, and in choosing a definition for AIDS we run immediately into the HIV/AIDS controversy. Some of the difficulty is that definitions, even in science, are chosen partly on aesthetic grounds. There are utilitarian principles, too, of course. In medical science we rarely know in detail at the molecular or even cellular level what causes most human illness, and so in our ignorance we are often forced to work with "disease syndromes," which are collections of symptoms and sometimes lab tests which seem to "go together." In order to usefully define a "disease syndrome," (which gets promoted to a "disease" after it stands the test of time and we get used to it) we need to pick our defining characteristics so as to include all of the sick people who we are interested in for good clinical reasons, and exclude everyone else. What are "good clinical reasons"? In medicine there isn't much point (other than pleasing some doctor's vanity) in defining a new "disease" which, when present, makes no difference in either prognosis or treatment. Nor is there much point in defining a disease so poorly that it fails to capture all the sick people who seem to have pretty much the same thing wrong with them from the prognosis or treatment viewpoint. If (as always happens) we lack information about what impact certain definitional characteristics _have_ upon treatment or prognosis, then we are forced to _guess_, as best we can, what disease definition will be most useful. It is at this point, in deciding whether two people have "pretty much the same thing wrong with them," that aesthetic and intuitive considerations unavoidably enter into medical science. Utility imposes other constraints, too. A disease definition which is to be used during a hunt for the disease's causation, should not assume any cause which is in question. To be specific, if we choose a definition for "AIDS" which in some cases requires evidence of infection with the HIV virus (the current way it is done in many countries, including the U.S.), then we will have chosen our terms so as be of little help in the question of whether HIV causes AIDS. Obviously, it would be nothing remarkable if we "discovered" that almost all people with AIDS were infected with HIV, if we had previously chosen our AIDS definition to make sure that it was so. In re-opening the question of the cause of AIDS, what we need then is a modified AIDS definition which does not involve HIV, so that the question of whether or not all AIDS cases are infected with HIV is an empiric one, not simply a semantic one. When we have a suitable HIV-free candidate definition for "AIDS," we can then ask two critical questions about it: 1) Have we captured with our definition all of the people with the new medical problem that we historically came up with the "AIDS" label, in order to describe and encompass in the first place? 2) If we test our defined group, are very nearly 100% of the people encompassed by our AIDS definition found to be infected with HIV, an otherwise rare virus in the population? If the answer to both these questions is yes, then HIV is promoted to a good candidate for the cause of the AIDS epidemic. If either answer is no, then our version of the HIV/AIDS hypothesis obviously has severe problems right from the start. Fortunately, it turns out that we can easily construct a workable definition of AIDS which does not include any reference to HIV, but which still describes the new epidemic we are interested in. Such a definition will not be the standard one, of course, but since the standard modern "HIV-containing" AIDS definition is unusable for this purpose, both we and the HIV/AIDS critics are required to construct special AIDS definitions, even to continue to talk about the problem of causation. It's unavoidable. Finally, we should note that in no infectious disease known to man is it possible to possible to define the disease syndrome clinically (i.e., by things observable at the bedside) or by any laboratory means which don't include direct tests for the causal organism, and YET still be able to achieve 100% specificity ("100% specificity" means that in no case before testing for the microbe is the physician ever fooled about whether or not the patient REALLY has the disease). We're not pefectly 100% sure we are seeing ANY infectious disease without a test for the disease causing organism. Since this is true for any infectious disease, we should expect this to be true of AIDS also. Simply put: no definition of AIDS which does not depend on HIV (as the critics insist it must not) can ever be expected to be perfect, and thus NEVER capture anyone who actually does not have HIV. But, as with all infectious diseases, a 99.9% score will be quite good enough for us. _AIDS Is An Acquired Immune FAILURE Syndrome_. What is the best way to define AIDS without reference to HIV? First, the term "AIDS" stands for Acquired Immune Deficiency Syndrome, but it has always been understood (except by certain critics, who we will come to presently) that the amount of immune deficiency we are interested in, is not a trivial one. AIDS is the name historically chosen for a new medical syndrome which is essentially 100% fatal, and thus in defining "AIDS" we are looking for people with an immune deficiency in the range which is life-threatening, and which will continue to grow relentlessly worse until life is impossible. One possible way to define immune deficiency would be to define it by what secondary health problems it causes-- for instance, one could pick people who have gotten so-called "opportunistic infections," which are strange infections which are seldom if ever seen in people whose immune systems are fully functional. In the early days of AIDS, before HIV was discovered, the syndrome was indeed defined using such opportuni- stic diseases (Fig. 1a), and people with these infections are still included in the latest 1993 Centers for Disease Control and Prevention (CDC) clinical surveillance definition of AIDS (but now, in most cases, such people are only included if they are also HIV infected). We will not be able to use this CDC definit- ion of AIDS (Fig. 1c). Not only does it require HIV infection in many (though not quite all) cases, but for historical, political, and technical reasons, it also is constructed in a way which does not assess current immune status in the best way for our purposes. Why is this? The basic problem is that only a limited amount of information about a person's immune system function flows from the bare fact that they have an "opportunistic" infection. Certainly there is a good correlation between immune function and what kind of opportunistic infections occur, but the correlation is far from perfect, since opportunistic infection risk is influenced by not only immune status, but also by the quality of what we may term the infectious "assault" to the system. The infectious assault in turn is influenced by a person's physical location, infection contacts, personal habits, and other exposure factors both known and unknown. In the end, infection assault differences insure that _some_ unlucky, highly-infection-exposed people manage to contract opportunistic infections when only mildly immune compromised (though these are rarely fatal). By contrast, the same assault differences insure that _other_ people who are badly immunologically impaired may escape opportunistic infections for an amazingly long time, simply by missing the microbes which can kill them. When it comes to immune function, then (what we are interested in with AIDS), it is better to have a direct test which is not subject to uncontrolled variables, such as which microbes happen to be in the air or drinking water, and how many. We would like a more direct test. Such a test exists. Quite early in the history of AIDS, it was found that the immune defect in this disease is peculiar, and that it most visibly involves a particular kind of cell in blood and lymphatic tissues (lymph "nodes"), called "T-lymphocytes." In the syndrome of AIDS, certain T-lymphocytes gradually disapp- ear from both blood and lymph tissues, and a simple T-lymphocyte count in the blood can tell how serious the reduction has been in both places (since blood lymphocytes come from the lymphatics). The arm of the immune system which is controlled most directly by T-lymphocytes (the body's defense against viruses and fungi) is what is most defective in AIDS, and viral and fungal infections are (not surprisingly) the main opportunistic infections which appear and cause death in AIDS. AIDS is so specific in its attack that scientists eventually found that only one _subset_ of T-lymphocytes was initially hardest-hit. This was the so-called CD4+ or "helper" T-lymphocy- te, which has the job of stimulating the immune system. The other major type of blood T-lymphocyte, the CD8+ or "suppressor" lymphocyte, is involved in shutting the immune system down; in AIDS, CD8+ lymphocyte blood numbers increase early in the disease, and are not decreased until near the very end of the disease process, when they may also disappear. [See appendix: "What is a CD4+ Lymphocyte"]. CD4+ lymphocyte blood counts tell much of the story in AIDS and other immunodeficiencies involving the T-lymphocyte immune system. A healthy adult might have a CD4+ lymphocyte count of 800 to 1000, with a CD8+ count half of this. These are normal values. Under physical stress, injury, or chronic infection, CD4+ lymphocyte count might drop to 500 (to even less than the CD8+ count), and mild, non-fatal opportunistic infections might be the result. A CD4+ count less than the CD8+ count was once used as a crude marker for AIDS, but today with progress we know that this immune state is non-specific. In AIDS, things eventua- lly become much worse than this, and the worse things get, the fewer possible other causes besides AIDS there are for the problem. In full-blown AIDS, as defined by opportunistic infections and other problems, the CD4+ count is usually below 200. It is at such count levels that Kaposi's sarcoma (a tumor perhaps caused by a virus) and life-threatening infections begin to appear, although approximately 95% of AIDS patients survive beyond this level of decline [2]. Another feature of AIDS, however, is that inevitably the count grows worse over time. Today, in the modern era of antibiotics and more knowledgeable care, 85% of AIDS patients live to see their CD4+ lymphocyte count drop below 50, which is also near the average point at which most AIDS patients in the modern era now suffer their first opportunistic infection [3]. Famous AIDS sufferer Kimberly Bergalis, for instance, had her CD4+ count drop to 41 before her disease was discovered [4]. Many AIDS patients today go all the way to CD4+ counts of zero before the inevitable final infection or other complication. It is because of the implacable and more or less irreversible loss of vital T-cells that AIDS remains a fatal condition, with an average time of less than two years between the first opportunis- tic infection and death. If we wish to define AIDS in terms of immune failure, the essential question is this: where do we draw the line, so as to include almost all people with the new immunodeficiency epidemic, who are going to die from it, but exclude everyone else? If we simply define "immune deficiency" as a CD4+ lymphocyte count of less than 200 (where death begins to become more likely), we will capture about 95% of people who die of what the CDC now defines as "AIDS" (Fig. 1b). Previous to the epidemic of AIDS, of course, people did die of immune failure with low T-lymphocyte counts (including low CD4+ counts) for other reasons, and they continue to do so now. Thus, we must also exclude from our AIDS definition all those people who have one of the classic reasons for a very low T- lymphocyte count-- reasons which were well-known before the AIDS era (cancer, malnutrition, tuberculosis, radiation, chemotherapy, etc). These people don't have AIDS, because the historical epidemic of AIDS consisted of people with no T-lymphocytes for no other known reason. These people had appeared (more or less) newly on the scene in the 1980's with evidence of a fatal kind of immune failure which was _acquired_, meaning that it was an epidemic problem of something "picked up" by previously healthy people. So let us simply collect all the people we can find with sustained CD4+ counts below 200 without known reason, and test them for HIV. When we do, we find that almost all are HIV infected, and any who aren't, don't look at all like typical AIDS patients (as we will see). This, despite the fact that only 0.3% of the general population of the U.S. carries the HIV virus, giving a huge population free of this virus which should still be at risk for any other causes of severe immune problems. Thus, at this point we have no evidence yet to directly contradict the simple theory that HIV causes almost 100% of our conservatively defined "AIDS." In fact, things are looking quite suspicious for HIV. AIDS Heresies, Part 2 _Enter the Critics_ Not everyone will agree to use the above definition of AIDS, of course. Before we go further, we will introduce two major critics of the HIV/AIDS hypothesis. Both are scientists who hold Ph.Ds. Neither one is a physician. [See Appendix: "Two Critics: Duesberg and Root-Bernstein"] The view that HIV plays no role in AIDS has been most notably put forth by Peter H. Duesberg, a molecular biologist spe- cializing in a group of viruses which are related to HIV. Duesberg's view, promulgated since 1987, is that HIV is harmless and has no causal role in AIDS, and that AIDS instead is caused by drug use and by immunosuppressive blood products (page references to Duesberg's views will be from his major 1992 paper on the subject, and to a recently published book by Ellison and Duesberg [5]). At the other end of the skeptic spectrum are hybrid arguments raised by Robert Root-Bernstein, an associate professor of physiology and author of _Rethinking AIDS_ [6], the most carefully-documented work to yet assail the prevailing medical views on HIV and AIDS (Root-Bernstein references will be given in the form of a page reference from this book). Root-Bernstein is less radical than Duesberg, arguing for a somewhat less central role for HIV in AIDS than is generally given it, but still allowing for the virus to have some part in the etiology of the disease. Since Duesberg's original challenge, which has been the cause of much formal debate in the literature [7], a number of scientists, physicians, and lay persons have taken up the cause for a "re-appraisal" of the idea that HIV is THE major causal factor, or even _one_ of the major causal factors, in AIDS. Most respectable is an organization called "The Group for the Scienti- fic Reappraisal of the HIV/AIDS Hypothesis," which has collected over 300 signatures of physicians and scientists, including those of Nobelists Walter Gilbert and Kary Mullis. This group has campaigned to remove the requirement for HIV infection from any medical definition of AIDS, feeling that using this criterion is at best premature, and prejudices any hunt for alternative explanations for the disease. Almost all critics of the AIDS/HIV hypothesis have one thing in common: they insist on using a much broader definition of AIDS than we have proposed, a definition which virtually guarantees that some people who fit the critics' "AIDS" definition will _not_ be HIV infected. To be fair, there is some historical precedent for using a definition of AIDS which relies solely on the patient developing certain of the most serious and specific opportunistic infectio- ns, since this was the way the disease was diagnosed before HIV testing became available in 1985 (Fig. 1d). Today we know that almost all such people with pre-1985 defined "AIDS" are infected with HIV---indeed this was known in late 1983, before the official announcement of viral cause was made the following year. But today we know this figure would not quite be 100% [13]. As we will see below, there is evidence that the few HIV-negatives in this group will be people with lesser degrees of immune suppression (higher CD4+ counts), who will _not_ progress to worse immune function, or quickly die. It seems reasonable, then, with what we know today, to simply exclude them-- since we know that this is not the characteristic picture of AIDS. Again, it is most reasonable for our purposes to diagnose AIDS on the basis of immune function (CD4+ levels) only, since it is immune function, not infection status, which correlates with prognosis in CD4+ immunosuppressed people. The critics, however, will have none of this, and in their definitions are seemingly less interested in clinical utility than they are in collecting ammunition for an argument. The more broadly AIDS is defined, the more "HIV-free AIDS" cases critics can assemble, and these can in turn be used as evidence to the lay public that HIV cannot be the cause of "AIDS." Duesberg, for instance, has insisted upon retaining the early 1980's observation that a CD4+/CD8+ lymphocyte count ratio of less than 1.0 is often seen in AIDS, and he has decided that such a ratio, even in the absence of opportunistic infection, is synonymous with AIDS (Duesberg, p. 260). Duesberg now calls this ratio an "AIDS-defining immunodeficiency," and counts people with this lab result as part of "HIV-negative AIDS," in his shocking and too-often repeated statistic that there are "3000 documented HIV-free AIDS cases" [8]. Here again, Duesberg's chosen definit- ion of AIDS is less than useful epidemiologically or clinically, because people with such mild immunosuppression as he uses to define "AIDS" are not the people who are dying, or are shortly destined to die (as will be made clear below). AIDS is nothing if not a fatal epidemic, and insisting that mildly compromised persons who may or may not eventually get any worse be labeled as having "AIDS," as Duesberg routinely does, only serves to confuse the issue (Fig 1e). Perhaps following Duesberg, there is a general trend for HIV/AIDS skeptics to overdramatize levels of immune deficiency which are not clinically very significant. For example, Root- Bernstein (p. 262), in characterizing a study of HIV-negative (HIV infection free) men newly infected with CMV virus, notes that for a time, some of the men had CD4+/CD8+ cell ratios of less than 0.4, a figure which he claims "represents extreme immune suppression." It would take a medically sophisticated reader to know that in AIDS this ratio would typically be far less than 0.3, and thus these men would not be mistaken for the current CDC immunological definition of AIDS, even if they were HIV-positive. The reader would also need to know that the level of immunosuppression associated with a ratio of only 0.4 is not associated with significant risk of death by opportunistic infection. Such a reader might wonder how we are justified in calling a ratio of 0.4 "extreme immune suppression," if people rarely die from it, as they are known to do in AIDS. Root- Bernstein does not say-- indeed, does not even raise the issue. An tendency toward overdrawn interpretation of the clinical significance of lab results, is one of the places in which the absence of medical training in the chief HIV/AIDS skeptics shows most clearly [9]. Duesberg's paper [5] and Root-Bernstein's book [6] each contain descriptions of groups of HIV-free people who are somewhat immunosuppressed due to low CD4+ counts, but not severely so, as defined by our straightforward criteria of having a significant risk of infectious death due to T-lymphocyte loss. These immune deficient patients in HIV/AIDS skeptic literature are presented along with the inference that perhaps somewhere there exist people with these immune suppressive factors, or combinations of them, who are _severely_ permanently T-lymphocyte immunosuppressed (as AIDS patients generally are), and yet still without having HIV. Duesberg and Root-Bernstein only have one difficulty in this argument-- neither has been able to actually _find_ more than a handful of such people in a country where AIDS sufferers infected with HIV have passed through this clinical phase by the hundreds of thousands [9]. _HIV-Free AIDS?_ Hypotheses may be disproved by the right data with relative ease, and cases of HIV-free AIDS would disprove the idea that HIV causes AIDS, in proportion to how often these are found (i.e., if 10% of AIDS cases were HIV-free, this would prove that HIV is not the cause of AT LEAST 10% of AIDS). Thus, Duesberg and Root-Bernstein aren't the only ones who have been looking for HIV-free people who are badly CD4+ lymphocyte immunosuppressed without any classic reason (i.e., good candidates for HIV-free AIDS). Very recently the CDC reported that after a massive search it had only been able to find less than 100 people without HIV infection across the country whose CD4+ counts were, at one measurement, less than 300 (not quite in the AIDS-class immunosuppression range of 200, but drawing close). This group of people, for ease of reference, was given a special syndrome name: "ICL" (idiopathic CD4+ lymphocytopenia), meaning "people with low CD4+ lymphocyte counts without a medically-defined disease, or other known immunosuppressive reason." Why wasn't ICL simply called "HIV-free AIDS"? Sometimes it is. The 1987 and 1993 CDC AIDS definitions do allow for AIDS to be diagnosed in the absense of HIV, and even with a negative HIV test. This requires certain infections and a CD4 count less than 400, and these things happen extremely rarely in the absense of HIV (fewer than 1 in 1000 cases of AIDS are tested for HIV and found repeatedly negative). Critics have darkly suggested that the reason "ICL" is not called AIDS, is simply a matter of politics, but in fact there were problems with considering these people as AIDS cases, which had nothing to do with AIDS politics or the HIV theory. One difficulty was that two thirds of people labeled as having "ICL" were found not to come from the AIDS risk groups. They did not use illicit drugs, had not been exposed to blood products, and had no evidence of sexual behavior which would have exposed them to a special infection risk. Thus, as we will see, the most popular alternative AIDS hypotheses did not explain the majority of these people _either_ --- a fact which did not keep them from being mentioned in nearly every skeptical treatment of the HIV/AIDS issue. What the skeptics had forgotten (or hoped their readers would not notice) was that the immune deficiency of people with ICL did not seem to be _acquired_ in any obvious way [10]. What justification was there, then, for calling it "AIDS"? To call a disease "acquired" it is not enough simply to show that it appears in adults without previous problems, for even some genetic diseases do this (Huntington's disease). A bona-fide acquired disease must have evidence of an etiology which is not genetic. The classic epidemic we know as "AIDS" does. Most ICL does not. The two syndromes are evidentally not the same thing, even looking at them from basic epidemiology. Moreover, people with ICL were not only epidemiologically, but often immunologically very much distinguishable from AIDS cases: their CD4+ lymphocyte counts swung widely in response to infections and were often much higher than 300 (in contrast to people with AIDS, whose CD4+ lymphocyte counts tend to stay low, and heading on an ever-downward trend). ICL people also often had low total lymphocytes or low CD8+ lymphocyte counts, again indicating that their type of immune failure did not make much distinction between CD4+ and CD8+ lymphocytes, as AIDS always does. Clearly, these people did not belong to the classic AIDS groups which began suffering with epidemic immune problems about 1980. They are not part of the new phenomenon of AIDS, something underscored also by the fact that, although suffering from opportunistic infections, ICL victims did not even seem to share the implacable death rate of AIDS [10]. Searches for HIV-negative people who have AIDS-type severe immune suppression have also been taken specifically within AIDS risk groups. Vermund reported in the United States Multicenter Cohort Study that of the 2713 persistently HIV-negative homosex- ual men in the study, who had had a total of 22,643 blood tests, only one significantly immunosuppressed man (CD4+ lymphocyte counts persistently less than 300) was found. This man was taking chemotherapy and radiation for cancer, and thus had a very good immunosuppressive reason other than his lifestyle to explain his lab results [11]. A similar review of another cohort of homosexual and bisexual men found no persistently lowered CD4+ T cell counts among 756 HIV-seronegative men who had no other cause of immunosuppression [11]. Finally, in the San Francisco Men's Health Study (a population-based cohort recruited in 1984) it was found that among 206 HIV-seronegative heterosexual and 526 HIV-seronegative homosexual or bisexual men, only one had consistently low CD4+ lymphocyte counts, and this man also had low CD8+ T cell counts, suggesting that he had general lymphopenia rather than the selective loss of CD4+ lymphcytes characteristic of AIDS [11]. If these three studies are indicative-- and there is no reason to think they are not-- then most, if not all, male homosexuals with AIDS-range immune failure are HIV-positive, since it has proved very difficult to find any who are HIV-negative. Much the same seems to be true in IV drug users: in a study of 1246 HIV-negative injecting drug users in New York City from 1984 to 1992, for example, only 4 were found with CD4+ lymphocyte counts less than 300 (if IV drug use per se was a major cause of AIDS, the number should have been far higher). In this small group of 4 people, even though infected with multiple other non- HIV viruses, and with a history of heavy drug use, immune function was stable and without the steady decline in CD4+ lymphocyte counts over a time span of years which is characteris- tic of all unselected HIV-positive cohorts [12]. A second study by others found much the same result [12]. Thus, in these studies also, the few HIV-negative people who could be found with even near-AIDS range immunodepression, were _still_ not behaving medically like people with AIDS. Studies of recipients of blood and blood products, and household contacts and sexual partners of transfusion recipients, also suggest that persistently lowered CD4+ cell counts are very rare or nonexistant in the absence of HIV infection [12a]. So far as we know, then, in the United States more than 99.9% of the people who are a part of this new phenomenon of permanently very low (and declining) CD4+s in high risk groups, have been infected with HIV. This does not prove that HIV causes AIDS, but it is surely an important clue. Why Not Merely Use the CDC Definition For AIDS, With HIV Taken Out? A persistent suggestion by critics is that it would be proper to use as an AIDS definition the current CDC definition (which includes all HIV-infected people with a much expanded list of infections and other problems), but with the HIV criteria removed. The problem with this suggestion is that definitions of diseases are chosen by the CDC for maximum clinical utility, and the HIV infection criteria in the CDC AIDS definition were not put there only to insure that there would be less HIV-free AIDS. Rather, HIV infection in a person with opportunistic infection is known to be (alone among all other viral infections) a very good predictor of whether immune status will continue to decay until the person eventually succumbs to opportunistic infections. In people with mildly compromised immune systems, the prognostic value of an HIV infection is large. Even critics admit that the prognostic value is large, without admitting causation. Thus, we cannot simply remove HIV status from the CDC definition and still have the definition do what it was designed to do, which is _predict_ impending death by immune failure. Critics know that if "AIDS" is defined only in terms of today's much broader list of "AIDS-defining" diseases and infections (which are meant to be used only in conjunction with HIV status), it is sure to be quite true that the definition will then become far too broad to be prognostic. Such opportunistic infections, as critics well know, sometimes happen occasionally even without the most severe CD4+ immunosuppression which is characteristic of people who die with AIDS. A study by Salvato illustrates this point [13]. In the study, medical records covering 6 years for 1500 HIV-positive patients were compared with records for 1000 HIV-negative patients who had Chronic Fatigue Immunodeficiency Syndrome (CFIDS) and evidence of immune suppression. It was found that the CFIDS patients had a syndrome much like that seen with early HIV infection -- fatigue, lymphadenopathy (swollen lymph "nodes") and low grade fevers-- but that over the course of 6 years their problems were not severe. Only one of them developed CD4+ lymphocyte counts less than 300 ("ICL"). Still, two had yeast esophagus infections, a severe opportunistic infection rarely seen other than in AIDS, and other people severely immunosuppressed. Three had active CMV virus disease of various tissues-- another disease very often seen in AIDS. A total of 486 patients had evidence of yeast infection of the mouth on exam, a condition suggestive of mild immune problems but one not limited to AIDS. The average CD4+ lymphocyte count in these patients (not including the single ICL patient) ranged from 500-1400, with an average of 650. This was significantly lower than normal, but much higher than typical for AIDS. In this study, 95% of the HIV-negative patients had previou- sly been infected with the EBV, CMV or HHV-6 viruses, and 48% had evidence for continued viral infection (critics such as Root- Bernstein have suggested that these viruses have roles in AIDS at least as important as that of HIV, but this study provides evidence against this idea). Most interestingly, these immuno- compromised HIV-negative patients were followed from 2 to 6 years, and none experienced progressive CD4+ lymphocyte decline (except for the one patient with ICL, who, with treatment of CMV infection, showed increased CD4+ lymphocyte counts again). Such CD4+ count stability is never seen in any random group of HIV- positive people, where an average CD4+ count decline over time would be inevitable. The study authors' conclusion: "Even after a methodical search in a practice that sees a large number of patients with immune problems, only 1 patient was found to have ICL. However, this study demonstrates that patients with normal CD4 counts can develop AIDS defining opportunistic infections [...] Upon long-term follow-up these patients do not appear to experience progressive CD4 depletion." Most importantly, no HIV-negative person died in the study, which illustrates the extent to which chronically virally infected, immune-suppressed people can _approach_ the clinical picture of AIDS, without crossing into the deadly long term immune failure which is characteristic only of people with HIV infection. The reader who is a bit confused at this point should keep in mind simply that the most important thing about the syndrome of AIDS (by any good definition) is that it inevitably and rapidly destroys the immune system, and then kills people by means of the infectious and other diseases they get as a result of having no working immune function. Thus, mild CD4+ cell depression and opportunistic infections are not always AIDS, for only some of these people (as it turns out, the HIV+ ones) will progress to immune failure, and death. It is immune failure (almost complete CD4+ lymphocyte loss) and death by opportunistic infection which is characteristic of AIDS; and it is only such people who are almost _always_ HIV infected. How and Where AIDS First Appeared The story of the detective hunt for the cause of AIDS is told with wit and clarity by Randy Shilts (an author who in 1994, at the age of 42, ultimately became a casualty of the disease himself), in the best-selling book _And the Band Played On_. Other good histories of the early AIDS epidemic are also availa- ble [14]. Historically, what happened in the U.S. in 1981 was that homosexual men began presenting to physicians in increasing numbers with very, very low CD4+ lymphocyte blood counts (but not lowered counts for other subtypes of lymphocytes), a destroyed immune system with lymphatic tissue destruction, opportunistic infections, and Kaposi's sarcoma. Almost no one who had treated diseases in the male homosexual community could remember having seen anything remotely like what had began happening on an increasingly large scale in the early 1980's. The year 1981 was not (in retrospect) exactly when the problem started, but rather when the problem first grew large enough in the U.S. to be brought to the attention of the federally-run Centers for Disease Control and Prevention (CDC) in Atlanta. It was in the Summer of 1981 that alert physicians brought to the attention of the CDC a mini-epidemic of immunod- eficiency and pneumonia caused by unusual organisms (A fungal organism called _Pneumocystis carinii_, and a virus called CMV) in homosexual men in Los Angeles. Before the epidemic of AIDS, there were only 50 to 60 cases of pneumocystis pneumonia reported to the CDC per year, in the United States. After ten years of the AIDS plague, this number had risen to tens of thousands of cases per year, and almost all of these in adults who had never previously been targets for this organism. Because many of the first people to contract AIDS had had sexual contact with each other, CDC researchers thought they might be looking at an unknown sexually-transmitted infectious disease. They also toyed for a time with the idea that sex- stimulant-chemical use or illicit narcotic use, both very common among the first cases of AIDS, might be somehow causing immunosuppression. Perhaps sexual contact was a red herring-- or merely a marker for a small and fairly tight-knit sub-community of people who shared common interests in non-sexual activities which might be damaging their immune systems. Those physicians treating infectious diseases in homosexual men thought not, however. Dr. Joel Weisman, one of the first doctors to put the AIDS puzzle together, noted that initially, within the male homosexual community, the disease seemed to follow lines of sexual contact more than it did drug or sex habits. Not all homosexual men were so promiscuous as to make contact-tracing impossible; Weisman observed that promiscuous men did not always contract the disease, but on the other hand, that even men with few sexual contacts were coming down with the disease if they had had sexual contact with the wrong person. In fact, men with severe immunodeficiency were eventually found to form sexual contact networks, of the kind that have always been seen by researchers using the classic epidemiologic tools for tracing sexually transmitted disease chains. The difference, however, was that for AIDS the contact networks stretched over years, indicating an infectious agent (if there was one) with a very long latency. Still, investigators found that of the first 19 cases of AIDS reported in Los Angeles, 9 had direct or indirect (one intermediate partner) sexual contact with a single French-Canadian airline steward, a man who was also sick with immunodeficiency. Then, starting in 1982, reports began to come into the CDC of the same CD4+ lymphocyte and lymphatic-tissue-destroying immune failure syndrome occurring this time in U.S. citizens who had received transfusions. Soon also came reports that an identical immune deficiency of a new severe variety was now being seen in men with hemophilia, a genetic disease in which sufferers must be injected with concentrates of protein clotting factors made from donated blood plasma. Reports of the first people with hemophi- lia and AIDS emphasized that, in these people, none of the same drug or male-homosexual behavioral factors were present that had been seen in the first group of AIDS sufferers [15]. Further, the same was true of those with "transfusion-related AIDS," who also did not fit into drug-using or male-homosexual lifestyles, and did not resemble them in sex or age either. Former tennis star Arthur Ashe is a well-known modern example. Ashe, like many of those with transfusion-related AIDS, had never had an intimate connection with anyone else with an immune problem, EXCEPT for a history of a blood transfusions years in the past, during the time in which transfusions were associated with AIDS. In late 1982 all this worried epidemiologists as the reports continued to come in. They knew that another viral disease called hepatitis B ("serum hepatitis") was also transmitted epidemically as a sexually transmitted disease in homosexual men, but much more rarely in homosexual women or heterosexuals in the U.S. Hepatitis B had historically also shown up early in people with hemophilia, who because of their large pooled blood-product exposure have historically seemed to be first to suffer from any new organism infecting the blood supply. Hepatitis B had also been known to be one of the worst disease-causing contaminants in donated blood for general transfusion. Thus, the same 3 groups of people who had historically been infected with hepatitis B in the 1970's, had now started coming down with AIDS. Hepatitis B was also a disease of IV drug users who shared needles, and it was not long before the first reports of IV drug users with AIDS came in. By 1983, the CDC was sure it had a new infectious disease on its hands, similar in epidemiology to hepatitis B, but with a longer latency period. Analysis of the habits of donors of the blood components that were transfused into people who had later developed AIDS, indicated one thing different about the donors: it was found that blood products AIDS patients had received had more often come from people who themselves were at "high-risk" for AIDS due to promiscuous male homosexual behavior. On the other hand, matched case-controls who had been transfused identically from the same blood bank but had _not_ developed AIDS after transfusion, were found to be not nearly as likely to have gotten blood components from anyone in a "high-risk group." This initial study concluded that there was only a 1% chance that the statistical association of transfusion-associated AIDS with the lifestyle of the blood-donor would be as close as it was found to be, if only chance had determined the lifestyles of the donors of blood to people who later became sick. Such a chance association would have been expected if there was no contamin- ation, and instead there was something about normal transfusion blood itself, or perhaps some other factor unrelated to trans- fusion, that was causing AIDS in transfusion recipients [16]. The remarkable fact-- from which there was no escape-- was that AIDS in a transfusion recipient PREDICTED the lifestyle of a blood-donor he or she had never met (a donor which generally turned out to be a promiscuous homosexual man who had thought himself to be perfectly healthy). Nothing but an infectious agent (or more than one) could explain a statistical connection between a blood donor's sexual habits, and risk to the person receiving the blood. As for drugs or immune toxins, it was impossible to believe that any chemical toxin could be present in a relatively small amount of blood component coming from a single nominally healthy person, in sufficient quantities to cause total immune failure in the recipient, and do it years after the transfusion. Eventually, with many cases like Arthur Ashe's on record (but showing up in the early 1980s, earlier than Ashe's did), AIDS looked epidemiologically _very_ much like hepatitis B. The hunt was on for the microbe, or microbes, which caused the new syndrome. When the virus now known as HIV finally hit the world news in the Spring of 1984, there was a great deal of skepticism in the scientific and lay communities alike. With the ability to test for antibodies to HIV in 1985, however, there came a way of powerfully sifting through putative causal factors for AIDS, and comparing them with the factor of past HIV infection. HIV infection has emerged from these tests as the clear champion of competing AIDS-causation theories, convincing at present all but the most die-hard skeptics [14]. [See Appendix: "What is an Antibody? and What Does HIV-positive Mean?"] AIDS Heresies, Part 3 The Skeptics and the History of HIV and AIDS But what if AIDS and immune failure are not really new-- perhaps (as some critics suggest) we just look harder for them now that we recognize them? Could our new theories be warping our views so completely that by now that we have made a new "plague" out of something that was here all the time? Epidemiologically, what can we fairly say about the period before 1980, keeping this possible bias in mind? With the new ability to test old preserved tissue specimens for HIV, the first thing that becomes apparent is that AIDS is indeed older than 1981-- perhaps far older. Deaths from what has since been recognized as HIV infection with immune failure have been seen clinically, without being understood, for at least 35 years, and probably much longer. An HIV-infected British sailor, who had traveled widely, was once thought to have died with severe immune deficiency and HIV infection in 1959, the earliest proven case of modern AIDS. The diagnosis was made by means of preserved autopsy tissue specimen HIV testing, 30 years after the fact, although this case is still not without controversy. [Note added in Jan. 1996-- this case has finally been disproven. Apparently reports of this man's tissues containing HIV were in error, see ref 17]. This man's death alone would have provided good evidence that HIV is not a product of deliberate genetic engineering, for in 1959 biologic science was simply too unsophisticated to work with lymphotropic (lymphocyte-infecting) retroviruses like HIV, let alone engineer them [103]. HIV appears still to be an accidental infection of humans with an African primate virus, if it is anything at all. The genetic material of the most common HIV-1 strain is most similar to that of a virus known to natura- lly infect chimpanzees, and it may be that HIV's ancestors have been present in Africa, perhaps even in humans, for a very long time-- perhaps thousands of years [18, 121]. In West Africa, a close cousin of the U.S. HIV-1 strain, called HIV-2, is almost identical to several indigenous African monkey viruses, and almost certainly has been derived from them quite recently in virus evolutionary time (less than several centuries). In the U.S., AIDS-like deaths have been reported as far back as 1934, as Root-Bernstein notes. The first AIDS or AIDS-like death that we believe could have been associated with HIV infect- ion was that of a 17 year-old possibly homosexual male, who died of strange opportunistic infections in St. Louis in 1968 [19]. This early AIDS-sufferer had never been out of the country, showing that the virus was already active in the Western Hemisph- ere in 1968. Unfortunately this case is the clearest on record, and even it has not resulted in recovery of HIV DNA sequences (which would provide absolute proof). Some reports of early HIV are frankly errors: for instance a 4% fraction of preserved serum samples from IV drug users in this era (1971-2) in the U.S., found to be "HIV-positive," were apparently false-positives as shown by later followup [20]. It is possible that HIV viral infection has been present in small contingents of both drug users and homosexual men for some time in the United States, but the case has yet to be absolutely proven. Let us suppose that preserved specimens finally show that HIV was present in America long before the late 1970's. Why, then, was the U.S. first hit with an AIDS epidemic only in the 1980's, with HIV infection quickly rising to 50% in some risk-groups? The answer may be that if it was not the simple presence of HIV virus in the United States that changed, perhaps it was the social milieu. In the late 1960's drug use became far more widespread in the U.S., and the invention of the disposable plastic injection syringe about 1970 made IV drug abuse possible for the first time on a large scale. Also beginning around 1969 (the date of the New York City "Stonewall riots"), homosexuals in the U.S. began to take open political power, and concomitantly one faction of male homosexuals began to engage in the high-infection risk "bathhouse lifestyle" chronicled by Shilts. In addition, the American homosexual-male community was apparently many times re- infected by many world-traveling disease "vectors" from other countries in the 1970's, including the previously mentioned airline steward (described in Shilts as the CDC "patient zero") who traveled widely in Europe, Canada, and the U.S., died of AIDS, and is known to have had sex with no less than 40 of the first 248 Americans to be diagnosed with AIDS by April, 1982 [14]. What happened in the late 1970's in the U.S. is that when a large enough fraction of the American homosexual-male population became infected with HIV, the U.S. blood supply, maintained with volunteer donations only, finally became contaminated with the virus. (This started in 1978, as we know from later testing of archived serum samples taken from homosexual men originally for hepatitis B studies). Similar archived samples tell us that in 1978 the U.S. plasma supply used to make clotting factor for hemophilia treatment became HIV contaminated, no doubt primarily by IV drug users selling plasma to support a drug habit. The dates are not coincidental-- a crossover between initial HIV infected groups occurred as some homosexual men experimented with IV drugs in the late 1970's, and male IV drug users in large cities turned to homosexual prostitution in order to obtain drugs. The resulting new epidemic of transfusion and hemophilia- associated AIDS, beginning in 1982 and rising sharply in 1984, helped to bring the acquired nature of AIDS into focus. The small incidence of AIDS in the American homosexual-male and IV drug-user communities before the late 1970's in no way subtracts from the reality of the dramatic increase in AIDS which took place in the early 1980's on the heels of exploding HIV infection rates in these groups. Although relatively mild immune suppression has apparently always been widespread in many AIDS risk groups, the more complete and deadly immune failure characteristic of AIDS itself has been sporadic and very rare in young cancer-free people in any of these groups, until the 1980s. It is, to be sure, difficult to retrospectively evaluate the health of male homosexuals before the first prospective studies of gay men's health were done in the 1980's AIDS era, but we can be reasonably sure that an epidemic of deadly immune failure among young American men before 1980 would have been duly noted by epidemiologists. The HIV/AIDS skeptic Root-Bernstein docume- nts a few cases of unexplained opportunistic infection deaths from the medical literature before 1980, but clearly an epidemic of immunosuppressive deaths cannot be seen in the historical record before 1980 by any act of imagination. By contrast, at present "AIDS" (a new epidemic of immunodeficiency deaths) shows a high and rapidly rising incide- nce among young men and women in the U.S., and these deaths cannot be simply a new label for an old problem. The reason is that _total_ mortality and cumulated years of life lost to premature death in young persons are observed to be rising rapidly, with all of the change due to "AIDS" deaths, _at_ _the_ _same_ _time_ other leading categories of mortality remain stable. If mere re-labeling of deaths into different categories was a problem, these "newly recognized" AIDS deaths would come out of other previously defined mortality categories, and this clearly isn't happening [21]. AIDS, the disease, may be old, but AIDS, the epidemic, is indeed something new. [See Appendix: "Hemophilia and Life Expectancy in the 80's"] People with hemophilia, unlike homosexual men, represent a well-defined group with long-term documentable changes in morbidity and mortality, since they had been well-studied as a group before the era of AIDS. This research shows that people with hemophilia began to die of dramatically different things, starting about 1982 [22] See Fig 2. A recent check shows little evidence of a special incidence of opportunistic diseases in people with hemophilia in the U.S from the turn of the century up to 1979, although a low incidence of AIDS could not be ruled out in this study, mostly because some cases of fatal pneumonia had no identified infecting organism [23], and because people with hemophilia as a group are immunosuppressed enough to be somewhat more susceptible than normal to bacterial infections. Signifi- cantly, however, in the years before AIDS, people with hemophilia had never been noted to be particularly susceptible to the more obvious _fungal_ infections, such as candida esophagitis, common to AIDS patients and others with low-lymphocyte type immune deficiency. After 1984, however, this type of AIDS-associated opportunistic infection and immune failure rapidly became the single most common cause of death in people with hemophilia in the U.S. [24]. The rise in total mortality risk in people with hemophilia was sudden: total mortality in this population, which had been stable in 1982 and 1983, suddenly increased by a factor of approximately 900% in the first quarter of 1984 [25]. Such an increase in raw numbers of deaths was consistent with an epidem- ic, or some new very toxic contamination of the clotting factor supply. It is not consistent with slower social changes, slower toxin or immune suppression models, multifactorial causation models, or the idea that people with hemophilia were actually at no greater risk than before (i.e., that again perhaps there had been some kind of "cause of death" re-labeling in response to AIDS hysteria). Mortality figures in hemophilia patients also showed something else important, which was that the new deaths of the late 1980s, by virtue of all being judged "AIDS," demonstra- ted that most or all of them occurred in people with hemophilia who were HIV-positive. Since these deaths accounted more or less for the entire new increase in mortality, it could be inferred that the mortality rate for HIV-negative people with hemophilia did not increase much in the 1980's, if at all. How significant was the increase in death rate for HIV- positives in this group? In one _Journal of the American Medical Association_ study it was found that in a cohort of 111 people with hemophilia infected with HIV in the early 1980's, one third had died by 1992 [26]. The reader is asked to imagine any group of this age (a high school class, perhaps) and imagine an overall 33% mortality rate in less than 10 years. Of the estimated 10,000 people with hemophilia to have been infected with HIV in the early 1980s in the United states, a quarter had been reported to the CDC to have died of AIDS by July of 1993. Such death rates were especially shocking in view of strides in hemophilia treatment which had been made in the years before. Total life expectancy in people with hemophilia had risen as clotting factor treatment became available through the 1970's, until by 1980 it was nearly normal [23]. After 1984, however, life expectancy in this group began a steep decline, and by the early 1990's was at a lower level than at any time since before World War II [24]. In the 1980's, total mortality for hemophilia increased in all age groups above 9 years of age, and age at death shifted markedly to lower ages, decreasing from 57 years of age in 1979-1981 to 40 years of age in 1987-1989 [27]. About 50% of people with hemophilia in the U.S. had been HIV infected by early 1986, when screening and treatment of the clotting factor concentrate stopped HIV spread [28]. Still, the long latency of the virus (as long as 15 years for 50% progress- ion to AIDS in this group) caused death rates to rise for long after the window of new HIV infection closed (they are still rising as of this writing, although clinical AIDS or severe lymphocyte loss (ICL) has yet to be reported in children with hemophilia born after 1986). The fact that there was a massive and silent HIV infection of half of the people with hemophilia in the early 1980's is beyond question. The HIV/AIDS skeptics' quest to divorce this event from the epidemic of deaths by AIDS in this same group over the next decade has resulted in some remarkable and curious stateme- nts about hemophilia mortality. Duesberg, for instance (p. 216) quotes only older statistics for hemophilia patients from the pre-1986 period, before AIDS deaths became very large. His practice of using randomly reported AIDS and mortality data for people with this disease (which is often notoriously unreliable in the best of circumstances [29]) instead of the much more reliable cohort study data, also results in figures which minimize the impact of AIDS. Cohort data shows mortality in hemophilia patients to be far higher than Duesberg acknowledges [30]. Duesberg has not been alone in ignoring major trends in hemophilia mortality in the last decade. The very misleading statement that people with hemophilia are living "longer than ever" today is one of the standards among the HIV/AIDS skeptic community. Root-Bernstein does no better than Duesberg at providing updated information in this area, offering one paper's 1979 pre-AIDS statistics [23], without update and without qualif- ication, to represent _contemporary_ life expectancy in people with hemophilia in 1993 (p. 247). This represents very sloppy scholarship (something which stands out particularly in Root- Bernstein), but the oversight does allow the author to skip discussion of the pronounced and otherwise awkward peak in life expectancy for hemophilia in the middle 1980's. Duesberg, though he seems to believe that people with hemophilia have suffered no mortality increases in the age of AIDS, does suggest that people with hemophilia live longer than ever due to recent factor concentrate development, and thus live long enough to die of immunosuppression caused by longer treatme- nts with clotting factor concentrate, instead of from hemophilia (Duesberg, p. 220). Although clotting factor does indeed appear to be mildly immunosuppressive (albeit in a different way than AIDS-- see appendix hemophilia section), the main problem with the hypothesis that clotting factor itself causes AIDS is that two studies of HIV-positive people with hemophilia have found that HIV infection and not clotting factor use is the critical risk for AIDS. These studies found that once a person is HIV- positive, risk of AIDS is _not_ related to amount of clotting factor used or severity or type of hemophilia---effects that would have been expected if clotting factor carried a significant immune risk independent of its HIV content [31]. Available statistics thus strongly suggest that the known association of clotting factor use and AIDS risk is merely due to the increased risk of being infected with HIV the more clotting factor has been consumed; once HIV infection has occurred, it doesn't matter how much clotting factor is used. Can a Viral Epidemic Explain the Historical Timing and Epidemiology of AIDS?: Attacks on Straw Men. It is an unfortunate fact that a great deal of the debate over AIDS and HIV has been over what rhetoricians call "straw- men." A straw man is an argument or viewpoint set up in a debate only for the purpose of being knocked down, and one which the opposite side never really defended or held; or one which is not very important to the central issue of the debate, even if it _has_ been held. Straw man arguments often result from debaters talking "past each other," without understanding the opposing side's position (straw man arguments may also, of course, be used deliberately solely for rhetorical purposes-- a practice genera- lly considered beneath respect in scientific debates). In the HIV/AIDS debate, straw men set up by heretics have most often been medical hypotheses which have previously been put forth in the context of the HIV theory and which have turned out to be wrong, but which were never important corollaries necessarily deduced from the idea that HIV causes AIDS, or were in other any other way central to it. Other straw men are ideas that the orthodox scientific "establishment" never put forth seriously at all, though they may be attacked vigorously by heretics as though they are current medical dogma. We will presently see samples of both. An example of an epidemiologic straw man is the timing of HIV arrival in the Western hemisphere. Root-Bernstein discusses cases of possible AIDS as far back as 1932, notes documented HIV infection with AIDS as far back as 1968 in the US, and argues that these data are anomalous (p. 2) if the virus was transferred for the first time to the Western hemisphere around 1978, as was originally thought (and which is the "just-so story" told by Shilts). And so they are. But if the HIV virus was transferred much earlier than 1978 to the new world, and remained at low levels in male homosexuals and injecting drug users in America until changing social factors in the 1970's encouraged its spread (exactly as Root-Bernstein himself indirectly suggests), no real damage would be done to a suitably modified HIV/AIDS theory. An example of a bad prediction made by the orthodox medical establishment which is not necessarily derivative of the HIV theory, was (or is) the official idea that AIDS is due to be a heterosexual pandemic in America _any_ _time_ _now_. It is argued by Duesberg (p. 203), that the "viral hypothesis" has failed to predict the course of the AIDS epidemic-- namely that AIDS has (at least so far) shown no clear inclination to spread rapidly by a complete heterosexual-sexual-transmission mechanism in the U.S., even though it apparently does so in Africa. It is also asserted in a related argument by Root-Bernstein that the HIV/AIDS hypothesis does not explain the generally-low measured levels of HIV virus in semen, the low (but not zero) rate of HIV infection in mates of HIV-positive men with hemophilia, or the nearly zero rate of infection in U.S. heterosexual prostitutes (unless they are drug users). If AIDS is an infectious disease, ask the critics, then why doesn't HIV infect very well? All these arguments are against straw men, so far as the cause of AIDS goes. There is nothing in the HIV/AIDS theory which demands that any particular transmission mechanism be the chief cause of the spread of HIV infection in any given place, or which demands that the HIV virus be as infectious in one locality as another. For example, it now seems likely from many studies that efficient sexual transmission of HIV requires mucosal tissue trauma, which is much more likely with anal intercourse, and/or a concomitant inflammation or ulcer from a second sexually transmi- tted disease. Because transmission may be inefficient even so, promiscuity also greatly enhances the chance of HIV spread. These requirement(s) for efficient HIV sexual transfer easily explain the difference between spread of HIV in tropical Africa vs. the developed counties. They also adequately explain why a disease which spreads well sexually only in populations with an extreme level of both promiscuity and rectal mucosal trauma (i.e., one sub-segment of American homosexual men) has not yet become a generally spreading epidemic in the U.S. It isn't that the HIV/AIDS heretics haven't come across such explanations. Root-Bernstein, in a good discussion of the epidemiology of AIDS, admits that there is nothing especially strange about a sexually transmitted disease which spreads easily in homosexual males but not heterosexuals. Both syphilis and hepatitis B in the 1970's have been examples of such a phenomen- on, and the "odd" differential epidemiology of both diseases with regard to sexual-preference groups is easily explained by differential _behavior_ in the homosexual and heterosexual populations in those years. Thus, Duesberg argues that a disease which restricts itself to classes of people in America, but not in Africa, cannot be explained by a micro-organism. But while he is doing so, fellow heretic Root-Bernstein (pp. 281-303) is noting that infectious epidemiology in one group of American homosexual males (who may have ulcerative sexually transmitted diseases and in addition be sexually infected with giardia, parasites, amoebas, hepatitis A, and B, shigella, salmonella, etc.), may resemble far more the disease epidemiology of some African countries than that of heterosexuals living next door (p. 290). In this, an AIDS caused by an infectious agent such as HIV may behave just as AIDS statistics suggest it does, and yet merely follow a pattern already amply demonstrated before AIDS, with many another infectious disease. Root-Bernstein is sometimes too competent a scholar for his own good: his Chapters Eight and Nine-- which address the epidemiologic differences and commonalities of U.S. homosexual men and African heterosexuals due to sexual practices and social changes which appeared newly in the 1970's and 1980s-- not only believably explains and refutes most of Duesberg's epidemiologic problems with AIDS (Duesberg, p. 209), but also does the same with many of Root-Bernstein's own epidemiological problems, raised in Chapter One. Unfortunately, Root-Bernstein is willing to let lifestyle and habit differences explain epidemiologic differences when it suits his argument's needs, but much less willing to consider them when they don't. An illustrative example occurs as Root-Bernstein discusses the rectal traumas and infections which occur during certain male homosexual practices, writing of these (p. 283-4): "It is now accepted that such injuries and infections greatly increase the risk of con-current infections (HIV or otherwise) and of semen gaining access to the immune system following anal intercourse." Yet when Root-Bernstein discusses the statistical association of AIDS with receptive anal intercourse (p. 225) he shows an odd difficulty with the same concept: "One possibility is that it is much easier to transmit HIV to a receptive partner than from a receptive partner. No other sexually transmitted disease behaves this way, however. [...] HIV would be the first disease agent to be able to make the discrimination, unless some other factor is involved." Here, unfortunately, Root-Bernstein is wrong, and wrong for the very reasons that he himself discusses in the quote preceding the last. Much like HIV, hepatitis B infection in homosexual men _also_ correlates with rectal trauma and receptive anal intercourse [32], and there is little reason to believe that the "other factor" is anything other than the fairly straight- forward mechanical injury that Root-Bernstein has already helpfully identified for us (see [33] for statistical development of a "rectal trauma index" which partly predicts risk of HIV infection). It is a characteristic of Root-Bernstein's style of argument that it makes causal mechanisms as mysteriously complic- ated as possible-- very often far more complicated than required to explain the facts. Here is another example of Root-Bernstein's difficulties with simple explanations: The known fact that HIV is difficult for an asymptomatic HIV carrier to transmit to another person by needle stick, or by heterosexual contact between married couples, does _not_ necessarily argue for the need for additional unknown co- factors or immune suppression in the more common cases of HIV transmission. Low infection rates for needle sticks may just as easily be explained by the known fact that virus blood levels in "healthy" HIV-infected people are low. Similarly, the known fact of low HIV infection risks during some kinds of heterosexual intercourse similarly admits to several interpretations, but one which requires no additional hypotheses is surely (and simply) that certain kinds of _homosexual_ behavior are more dangerous from an infection transmission standpoint (as discussed above) by their very _mechanical_ nature, without need to resort to the dubious and insupportable idea that the individuals who practice them must also be grossly immunologically compromised for HIV infection to occur. Root-Bernstein, eager to draw attention to any factor other than HIV in the causation of AIDS, does not take into account the most obvious physical factors. "...what is clear from existing studies," he asserts (p. 45), "is that HIV is extremely difficult to transfer to a healthy individual." In fact, existing studies establish no such thing. Studies quoted by Root-Bernstein never demonstrate that only "unhealthy" people in known risk groups contract HIV, only that certain traumatized risk groups (promiscuous gay men, hemophiliacs, transfusion recipients) are _on average_ somewhat unhealthy to begin with. This, of course, is not the same thing. Indeed, there is evidence that within risk groups, even the healthiest of individuals (immunologically) are capable of contracting HIV. Although men with hemophilia and homosexual men are on average mildly immunosuppressed even in the absence of HIV, it is by no means true that all are. A study of army recruits (surely a carefully screened group for health) shows that those who seroconvert to HIV (demonstrating HIV infection) may initially (by the criterion of CD4+ count) have immunity which is in the normal range. This is true in other groups as well [34]. Perhaps the most bloated straw man assailed by Root-Bernstein (and the one that provides the major theme of his book) is the idea that the causal agent of an infectious disease such as AIDS must be both _necessary_ and _sufficient_ to cause the disease in every sense of the terms; and moreover that since Dept. of Health and Human Services Secretary Margaret Heckler's dramatic announc- ement in 1984, most scientists have considered HIV to play this very role for AIDS. Root-Bernstein spends much time attacking what he calls the "HIV-only" theory of AIDS, an idea which actually has never flown, except possibly in the popular press or (at worst) occasionally when some scientist expresses a rash opinion without considering his formal training. (Dr. Robert Gallo, official co-discoverer of HIV, must by now badly regret his remark about HIV being able to cause AIDS in Clark Kent [35]; but this is hardly the N.I.H. official position). The subtitle warning of Root-Bernstein's book is _The Tragic Cost of Premature Consensus_, and it appears from the book that it is upon the "HIV-only" theory of AIDS that the "premature consensus" of the establishment is in dire danger of settling, if it hasn't already. Fortunately, it can safely be said that no such thing is occurring in the biomedical consensus, or about to. This does not prevent Root-Bernstein (p. 331) from logically blasting the somewhat cartoonish view he attributes to medical science: "Two of the most important implications of the HIV-only theory of AIDS are that all the risk groups should develop AIDS at approximately the same rate following HIV infection and that the symptoms they manifest should, on the whole, be the same." Alas for Root-Bernstein, however, since AIDS has from the beginning involved opportunistic infection organisms which vary in prevalence among populations, and since there has been reason to believe from the first that AIDS risk varies greatly with the biological _age_ of the HIV-infected person, scientists have never, even at the beginning, seriously considered such a theory as Root-Bernstein here lays out. "One logical implication [continues Root-Bernstein] is that the immunological status of an infected person should be irrelevant to susceptibility to contagion or to the progression from infection to disease. Acquisition of the retrovirus should be the sole factor determining whether an individual develops AIDS. Everyone should be at equal risk for AIDS, just as everyone is at equal risk for hepatitis B virus, syphilis, or measles." The most troubling thing about such writing is that an unwary lay reader may leave Root-Bernstein's book with the impression that the author has single-handedly discovered that infectious disease risks depend partly on host immune defenses and host behaviors and environments. The reader might well decide further that the biomedical community today does not in general think in terms of individuals having differing resistances to various diseases, and is accepting such advanced ideas only under duress, due to political pressures resulting from the penetrating logic of popular writers such as Root-Bernstein who are "re-thinking AIDS." The facts are more mundane. Obviously, since no microbe infects 100% of people exposed to it, or even causes disease in 100% of the people it infects (not even HIV has been shown to do this), there must be other factors to explain why some exposed people become ill with ANY infectious agent (viral, bacterial or parasitic), and some do not. Medical science certainly recogni- zes such factors, but does not use them to argue that there is in general something badly wrong with the germ theory of disease. Instead, as discussed earlier, medical scientists regard "caus- ality" in infectious disease in merely the sense of "necessity" (i.e., the "causal" microbe is necessary, but not sufficient). Medicine has not regarded the pathogenesis of any natural infection in terms of a "germ only" theory such as Root-Bernstein describes, since Pasteur, referring to disease, said that "The seed is nothing, and the soil is everything." Thus, Root- Bernstein spends many chapters assailing an idea that physicians have not held since the late 19th century, and certainly have never generally held in the case of AIDS. No infectious agent is usually "sufficient" to cause disease in a natural host, although in a laboratory (or perhaps very occasionally in nature) it may be sometimes true that inoculum (microbe "dose") may be so high as to make host resistance almost irrelevant. Naturally-occurring infectious disease organisms at reasonable doses, however, always rely on a chink of some kind in host immunity, with regard to that particular microbe (this is not to say that we must consider any host that is successfully infected to be "immuno-compromised"-- that would cheapen and overly broaden this useful term). The idea that deficiencies in host defense in some sense "permit" all or most infections is indeed a standard medical teaching [36], although a lay reader of Root-Bernstein might be surprised to learn of it after Root- Bernstein finishes misrepresenting the standard views of modern medicine. AIDS Heresies, Part 4 "Why is there such a huge and medically unprecedented variation [in time between HIV infection and death from AIDS]?" asks Root-Bernstein (p. 89). The answer to this rhetorical question is that such variation is _not_ medically unprecedented. Other infectious diseases, from malaria to syphilis to tuberculo- sis to viral hepatitis, may kill years after initial infection-- or within a much shorter time. In a cohort of newly-infected people, any study of a chronic infectious disease cannot help but produce steady increases in the "average" time between infection and death, as deaths accumulate slowly while the study follows the infected cohort prospectively onward in time. Once again, medical science has long assumed that there must exist host factors and other factors which explain why some people die of (say) tuberculosis or serum hepatitis 3 months after infection, and others not until after 3 decades-- but again these factors have nothing to do with our standard way of speaking of infectious disease "causation." The question of whether eventual liver failure from chronic infectious hepatitis is "caused" by the viral infection, is medically simply another way of asking whether or not it would occur _without_ the virus. It does not mean that other causal elements in the chain are being denied by medicine-- only that for very practical and necessary reasons they are being ignored at present (we will return to these reasons later in discussing polio). "No theory based solely on HIV can explain the phenomenon [of variable times of death]," writes [[ part lost ]] was explained as being merely an early warning of impending immune failure in these groups. (In HIV/AIDS skeptic lore, HIV is a harmless bellwether virus which can tell when certain members of modestly immunosuppressed groups are headed for future immune failure, and rapidly infects them ahead of time). The process of multiplying causal theories in order to minimize HIV responsibility for AIDS has finally culminated in the work of Root-Bernstein, which contains an eclectic "multifac- torial" view of AIDS which is so formless and complicated as to be epidemiologically unfalsifiable, even in Root-Bernstein's view (see his p. 92, quote below-- we will turn to this theory later). At the time of the early drug/toxin theories of AIDS, the leading toxin candidates were the inhaled amyl and butyl nitrite street drugs ("poppers") used heavily and almost universally as sexual-experience enhancers in the 1970's and early 1980's by the same fraction of homosexual men who indulged in high risk, promiscuous sexual practices causing injuries to mucosal tissue, and who also historically were the first U.S. group to develop AIDS as an "epidemic" [14]. Since this group was the one that suffered the first major impact of AIDS, a number of early studies found high statistical correlations between AIDS risk and nearly everything to do with this group's lifestyle. Later, after the HIV virus was identified, the CDC found that HIV was universally present and active in such men who developed AIDS. Almost as prevalent were a number of other chronic viruses, such as CMV (cytomegalovirus), HZV (Herpes Zoster virus), EBV (Epstein-Barr virus), and HSV-1, 2, and 6 (Herpes Simplex viruses 1,2, and 6). Many of these chronic viruses were found to be replicating actively in homos- exuals with AIDS. This state of "viral re-activation" (a product of immune suppression) was less common in AIDS sufferers from other risk groups, mainly because other groups had not been infected with as many chronic viruses in the first place. Sorting through the drug and infection variables among promiscu- ous homosexual men with AIDS was a statistical nightmare, although it became easier to separate out important AIDS risks when AIDS in other groups with different lifestyles was consider- ed. Epidemiologists fought it out in scholarly journals. After the main battle was over, they even tried to decide who hadn't guessed from the beginning on epidemiologic grounds that the problem might be infectious, even before a specific causal virus was proposed--- occasionally lambasting each other's past methods in print with words like "Neanderthal" [37,38]. Before HIV was identified, however, the basic problem for epidemiologists was that statistical methods could not by themselves suggest which lifestyles or practices (if any) were causal for AIDS, and which were merely an associative marker for some other causal factor which (perhaps) had not been measured. _After_ HIV was identif- ied, however, a second statistical appraisal could be taken using HIV status as a statistical factor, in an attempt to see if HIV had a closer associational (and therefore presumably more likely causal) relationship with AIDS, than other previously identified factors [39]. It did [40]. Much the same thing happened with other viruses, especially when statistics were extended across different risk groups. Infectious HIV was finally found to be present in essentially 100% of AIDS cases in ALL risk groups-- a higher proportion than was seen with any other virus [41]. Furthermore, although the prevalence of HIV was high (as much as 50%) in certain risk groups, such as homosexual males in some cities, or in people with hemophilia, the difference in HIV status for a risk group, and those who actually contracted AIDS within the risk group, was by far the largest _change_ associated with any virus, increasing from 50% to 100%. Furthermore, as noted, most of the change in HIV infection status had occurred before 1984 in people with hemophilia [28], long before the worst incidence of immunosuppression and in- creased death rate in this community, proving at least that HIV positivity was not derivative of _severe_ immunosuppression, since it preceded it. As judged by CD4+ lymphocyte counts in people who were followed over time, most of the loss of immune function in _individual_ HIV-positive people with hemophilia, came after HIV infection, as well [42]. Finally, it was found in several studies that while HIV- negative homosexual males might be mildly immunosuppressed, their immune function was never seen to drop as low as AIDS-class- immunosuppression (immune failure), as defined by CD4+ lymphocyte counts below 200. Moreover, when followed over time, HIV- negative homosexual men did not become MORE immunosuppressed, but HIV-positive ones did, and in prospective studies when men were followed by blood tests as they actually contracted HIV, this same slow and steady decay in immune status happened to the newly infected group after contracting HIV infection, starting immedia- tely after infection. HIV, when contracted by men being followed in studies, was generally contracted during a time when immune status (CD4+ lymphocyte count) was reasonably good [43]. Men who were severely immunosuppressed (CD4+ lymphocyte counts below 200; essentially immune failure) with no other explanation (such as cancer), invariably had become HIV positive already, or in other words, had become HIV-positive _first_ [44]. Such tight corre- lations between timing of immune failure and time of infection do not hold for any other known viral infection in immunosuppressed people. Lifestyle factors such as non-injected drug use and exposure to blood products (as in transfusions or hemophilia treatments) did correlate with risk of developing AIDS, but this association could be completely explained in the statistics by the fact that these behaviors (including even perhaps nitrite ("popper") use [45]) also increased risk of contracting HIV. To discover which was most important to risk, HIV or drug use, epidemiologists statistically "controlled" for HIV status (i.e., compared people with each other only within HIV status groups), attempting to discover if drug use or blood product exposure was important to AIDS risk AFTER the HIV virus was contracted, or independently of it. The answer, it turned out, was generally no. By and large, drug use and promiscuity were not independent variables after HIV infection was taken into account (with two exceptions, to be discussed). Important studies finding this included the followi- ng: * With regard to risk of AIDS in homosexual men, Susan Buchbinder followed 588 men in the San Francisco Men's Health Study who were infected with HIV at well documented times in the early 1980's (this is known because their sera was being stored for use in a hepatitis study). Of 588 men, 51% had developed AIDS by 10 years, and 69% had developed AIDS within 14 years of becoming HIV-positive. The San Francisco study also found that HIV- negative controls (both homosexual and heterosexual) had stable immune status, and did not develop AIDS, nor AIDS defining diseases. Of 538 men who became HIV-positive before 1983 in this study, only 8% were "healthy" HIV-positives as of Jan 1, 1993. This group of 1 person in 12, who for reasons still not underst- ood possibly will be long term survivors of HIV infection, was found to be quite time-stable in immune function, though mildly abnormal in immune status with respect to HIV-negative controls. [Italics]: _Significantly, the long term healthy HIV-positive men were as likely to have had sexually transmitted diseases or been users of illegal drugs as the other HIV-positive men in the study who became ill_ [46]. Those who progressed to AIDS and those who remained healthy, reported equal and significant use of marijuana, nitrite inhalants, and amphetamines. Both groups had similar herpes and hepatitis B infection immunity on lab testing- - objective markers for risky sexual behavior [46]. Regarding people already infected with HIV, there is no support from this study for the idea that either of the two possible outcomes (eventual progression to AIDS, or the long term health which may be possible for a small subset of HIV-infected people) has anything to do with non-injected drug use or sexual habits _after_ HIV infection, within a fairly wide range of behavior. A specific re-analysis of non-injected illicit drug use in this study [40] showed that 3 stratifications of drug use groups into low, moderate and heavy use, had NO statistical effect on risk of progression to AIDS in groups (even if done imperfectly, as Duesberg contends, some effect should still have been seen). [Fig. 2]. All 3 HIV-positive drug-use groups over several years- time lost CD4+ lymphocytes with equal rapidity and developed AIDS with equal probability; and all 3 HIV-negative drug use groups did not lose CD4+ lymphocytes at all, and did not develop AIDS defining diseases. Crude death rates in groups agreed with CD4+ loss rates and AIDS diagnosis rates, suggesting no bias problems with any of these markers, as has been suggested by critics. Statistical association of nitrite use with AIDS disappeared when only homosexual men, or only HIV-positive men, were examined, and thus nitrite use was seen to be merely a marker for homosexual behavior or HIV risk in San Francisco men, not an independent AIDS risk. Previous investigators in this study had also examined the statistical effect on AIDS risk of reported recreational non- injected drug use, and certain sexually transmitted infections, while controlling for other variables, and also found no associa- tion [47]. * In the Multicenter AIDS Cohort Study following a group of 1835 HIV-positive homosexual men, it was found that the 59 men who developed AIDS over 15 months during the course of the study were no more likely to have used inhaled nitrite "poppers" in the previous two years than were 295 matched controls picked from the same group, who did not develop AIDS [48]. In another paper from this study entitled "No evidence for a role of alcohol or other psychoactive drugs in accelerating immunodeficiency in HIV-1-pos- itive individuals..." Kaslow and coworkers found that the proportion of HIV-positive men at enrollment who developed AIDS during the following 18 months ranged from 5% to 8%, and that among HIV-positive men with low CD4+ lymphocyte counts, those who continued to use drugs showed no significantly higher 18-month risk of AIDS than non-users. No other manifestations of immunodeficiency were positively associated with substance use prior to enrollment in this study. Prior use of non-injected drugs was not associated with low mean CD4+ lymphocyte counts at enrollment, and continued drug or alcohol use after enrollment was not associated with greater subsequent decline in CD4+ lymphocyte counts. The authors conclude that as used in a large cohort of homosexual men, psychoactive substances did _not_ enhance the progression of HIV infection to lower CD4+ counts or AIDS [49]. * Likewise, the Vancouver Lymphadenopathy-AIDS Study (VLAS) comparing HIV-negative to HIV-positive men found no sexual behavior variables which correlated with CD4+ lymphocyte counts independent of HIV status, and for the 25 men who developed AIDS during the study found no significant differences in sexual behavior or illicit non-injected drug use as compared with 80 controls who were HIV-positive but remained AIDS free [50]. These studies included 78 HIV-negative men who were heavy users of many drugs who did not lose CD4+ lymphocytes or develop AIDS defining diseases, and also 19 HIV-positive men who reported no recreational drug use of any kind (and who were followed, incidentally, before AZT became available, yet lost CD4+ lymphoc- ytes steadily [51]). Again, AIDS risk in all studies correlated highly with the use of both intravenous and also certain non-injected drugs in homosexual men, but just as highly with promiscuity and high-risk sex in such men. If non-injected drug abuse (particularly nitrite use) was simply part of the lifestyle which caused one to contract the HIV virus, we would expect the correlation between non-injected drugs and AIDS to be high, without necessarily being causal. To separate out any independent effects of orally administered, inhaled, or intravenous illicit drug use, it is necessary to look at only men who had been infected with HIV while still showing reasonably good immune function (CD4+ lymphocyte count), and ask the question of whether drug use THEN, after infection, had any effect on the risk of later AIDS (which in most cases, appeared years later). The statistical answer was that after HIV infection, contin- ued injection of narcotics did seem to increase risk of AIDS, but that use of no other drug had any effect at all. Specifically, several early (pre-HIV) studies found connections between nitrite use by male homosexuals and Kaposi's sarcoma; but (in addition to the many studies cited above) later studies found that this association between nitrite use and AIDS disappeared after controlling for HIV status [52]. If recreational nitrite inhalants or oral drugs caused AIDS or Kaposi's sarcoma, and HIV was harmless, there should be some continued statistical associa- tion between AIDS risk and nitrite or other drug use, even when looking only at HIV-positive men, or only at HIV-negative men. No such association has been found in more than half a dozen studies [37-40, 46-52]. Illicit drugs did not cause AIDS, with the qualification that injection of drugs was implicated as a cofactor (see below). In the end, the "drug-only" hypothesis of AIDS pathogenesis failed all careful epidemiologic scrutiny. Even among IV drug users, although short term overdose deaths tend to swamp any necessarily long-term consequences of HIV infection [53], studies have generally shown that HIV infection is an additional risk factor for IV drug users [54-56]. Injected or IV drug use, of course, proved an excellent way to _contract_ HIV, if needles were shared. There was no evidence, however, that injected drugs themselves ever led to severe AIDS- type immunosuppression (CD4+ lymphocyte counts less than 200), in the absence of HIV. There was some evidence that IV narcotic use could be quite immunosuppressive (leading at least in part to fatal infections), and a co-factor for rapid AIDS development in people HIV-positive. Some studies found that continued IV injection of heroin, but not use of other drugs, hastened progression to AIDS in HIV-positive people [57], but other studies have suggested that heroin does not decrease CD4+ counts as AIDS does, so the immunosuppression of heroin users is apparently not due to heroin itself, but possibly drug impurities, or to other factors in drug-injectors [22]. One study [58] suggested that HIV-positive IV drug users who switch to methadone (an oral heroin substitute) may have slower progression to AIDS, but there was no mortality difference between using methadone and quitting narcotics completely, so IV injection per se, rather than narcotic use, was possibly the offending practice. Studies of IV drug users who continued IV drug might be implicating not drugs themselves in the rapid production of AIDS, but rather simply continued needle-sharing leading to acquisition of more virulent strains of HIV. The same was possibly true of extreme promiscuity, which also continued to be a risk factor after HIV-infection, in one study [59]. Acquisition of CMV was also a possible consequence of risky behavior, although the role of this virus as cofactor was limited at best [60]. In summary, the search for controllable cofactors in develop- ing AIDS, once HIV infection is present, was disappointing. Needle-sharing and extreme promiscuity very possibly hastened AIDS in HIV-positive people, but if they did so, it wasn't by much. Nor was there any evidence that avoiding these behaviors led to significant delays in AIDS, as compared with risk groups such as HIV-positive hemophiliacs, who still had a 50% chance of developing AIDS over 15 years. The Critics Don't Give In The above results have not convinced those who champion the drug hypothesis as the cause of much of AIDS. Duesberg, for example, accepts a causal role for drugs in AIDS (and for that matter, also for smoking in the causation of lung cancer: Duesberg, p. 213) on much the same grounds which he rejects for assigning a causal role to HIV-- namely, epidemiol- ogical correlations and suggestive lab experiments. The irony of this position is that the correlations are not nearly as good statistically for drug use and AIDS as they are for HIV infection and AIDS, and lab animal experiments with retroviruses mimic human AIDS immunodeficiency far more precisely than lab animal experiments with drugs have. Nitrite recreatio- nal drugs, for instance, do not cause lymphocyte abnormalities, lymphadenopathy, or the almost complete disappearance of CD4+ lymphocytes at the same time as CD8+ lymphocytes counts rise, as all happen in AIDS. Yet chronic retrovirus infection of experimental animals routinely causes these odd and specific phenomena. Nor have opportunistic infections or lymphomas ever been seen in a nitrite treated animal, and yet these are hallma- rks of lentivirus infections. Apparently, Duesberg's standards of evidence change greatly with the hypothesis he likes. Duesberg actually accused a group of scientists of data fabrication [61] after a paper in _Nature_ reported findings not in line with Duesberg's drug hypothesis [40]. Duesberg's letter to _Nature_ calling some results of the paper from the San Francisco Men's Cohort study less than honest, was refused print by _Nature's_ editor, with an accompanying editorial [62]. An independent institutional review board cleared the researchers of Duesberg's charges, which have been answered in print by the authors [63]. In the study printed in _Nature_, the authors had found in the San Francisco cohort men no connection between the four most commonly reported kinds of illicit drug use and later progression to AIDS, after results were controlled for HIV status (i.e., heavy drug users had the same likelihood to progress to AIDS as light users, if HIV-positive, but HIV-negative men did not progress toward immune failure, no matter what their drug use). [See Fig 3]. Moreover, these results held also for the 1985-1986 period before the drug AZT (the use of which Duesberg has suggested may cause AIDS to develop in HIV-positive people) was available. Duesberg's objections were that the study had not controlled carefully enough for drug use between HIV-positive and HIV- negative groups in the study, but Duesberg did not address the obvious question of why such considerable controlling for drug use as _was_ done, had absolutely no effect on differential AIDS risk seen. Duesberg also complained after seeing the raw data that supposedly "AIDS-defining" diseases in the HIV-negative group had not been counted as "AIDS," despite the author's denial that this had happened. Here apparently much depends on a disagreement between Duesberg and others as to what constitutes clinical AIDS. A recent article in _Science_ suggests that one difficulty is over the question of whether mild opportunistic conditions such as oral candida (thrush) constitute clinical "AIDS." Duesberg, ever ready to define AIDS broadly, argues they should [22]. In any case, the specifics of Duesberg's reanalysis of the Nature paper have never been printed, and death rates in this study again underscore the fact that Duesberg's broadly defined "AIDS," which strikes HIV-negative people, somehow does not kill nearly as well as the standard variety. As for worsening immune failure in groups over time (seen as declining CD4+ counts in the HIV-positive men, independent of drug use, but not in HIV-negative men, no matter how much drug use), Ellison and Duesberg have noted that this phenomenon isn't so clear before the data is reduced to averages. This, however, seems a odd complaint (making group trends clear is _why_ scientists calculate group averages). The bottom line is that, for now, the drug hypothesis of AIDS has no epidemiologic associational evidence behind it which is independent of HIV infection. HIV infection, by contrast, is heavily associated with AIDS risk, independent of drug use. [See AZT Section in Appendix on the question of whether AZT contributes to AIDS] Previously Known Viruses Because Duesberg does not regard any virus as being capable of causing a fatal disease long after the body has generated an antibody immune response to the microbe, he rejects a causal role for any virus in AIDS. The evidence for multiple viral infecti- ons in many of the early victims of AIDS, however, has caused many such "non-HIV virus" theories to be generated and tested. For example, though Root-Bernstein does not regard HIV as always a bystander virus in AIDS, he does regard other viruses in AIDS to be just as important as HIV. Are the other viruses (CMV, EBV, Herpes, etc.), or at least their antibodies, present as often in AIDS as those of HIV? Here statistics can help. According to Root-Bernstein, evidences of replication of the viruses CMV and EBV "are just as frequent concomitants of AIDS as is HIV replication" (p. 260). Unfortunately Root-Bernstein fails to note that this is true only in homosexual men with AIDS (where co-infection with EBV and CMV along with HIV is nearly universal). In science, situations in which several possible causes are all nearly 100% associated with a particular effect do not help us to differentiate causality, a point that Root-Bernstein makes (p. 279-280) without taking the next logical step. What are needed with viral studies and AIDS, obviously, are AIDS cases where some of the putative viral causes are present less frequently than 100% of the time. Such cases are available. In both hemophilia and transfusion-associated AIDS, HIV infection is universal, whereas infection and re- activation with other viruses, such as CMV and EBV, is variable [64]. In short, some people with AIDS in these groups have never been infected with CMV or EBV viruses at all in the past-- but all have been infected with HIV. Much the same is seen to be true when people in AIDS _risk_ groups are examined for antibodies to various viruses. EBV or CMV antibodies alone, in themselves, are not predictive for development of AIDS in risk groups where exposure prevalence of CMV and EBV is much less than 100% [60,65], a fact that eludes Root-Bernstein, who writes with too much generality that CMV and EBV antibodies are "synonymous with AIDS" (p. 103). In fact, as we have noted, CMV and EBV antibody are not unusually common in either HIV-positive people, or AIDS patients, from groups such as transfusion recipients, and people with hemophilia, in some countries [66]. Nor are antibodies to these viruses a clear extra risk factor for development of AIDS, even in HIV positive people [60]. Cases of people multiply infected with many chronic viruses, but not HIV, are also well known, and such people are easily differentiable from AIDS on clinical and laboratory grounds (see the study of Salvato [13] above). Despite Root- Bernstein's suggestions, statistical associations clearly differentiate HIV from other viruses in AIDS. It is necessary for heretics to come to grips with the crucial point (hard to explain if HIV has no causal role in AIDS) that the utility of HIV antibody screening is _exactly_ that a positive HIV screen, found in only 0.3% of the population, is predictive of risk for development of severe immunodeficiency, i.e., 50% risk of developing severe, life-threatening immunodef- iciency within less than 15 years. By contrast, EBV and CMV viral immunity and antibodies are acquired by most (well over 50%) of any normal, healthy population of humans during a lifetime, and thus are not predictive of future severe immunod- eficiency and death, except to the extent that they are markers for membership in particular risk-groups (such as gay men). Like many other factors, the association of viral antibodies with AIDS across risk groups disappeared when people were compared within groups-- except for HIV, where the association persists. The fact that HIV antibody is even 50% predictive of disease or death within 15 years is exactly the burden that those who argue that HIV is not causal must labor under, if they are to face the evidence. Other viral hypotheses fail this most simple test of prognostic value. As any life or medical insurance company knows, HIV infection status (HIV-positivity status) is more surely predictive of future death due to future severe immune failure than any other known piece of medical information related to viral infection. In science, if one has a list of factors which are independently statistically associated with an effect (such as AIDS), this does not prove that any of them are causal. If one of these factors (such as HIV infection) is far more highly associated than any of the others, however, it becomes difficult to argue convincingly that it is no more important than the others are. We will return to this point later. AIDS Heresies, Part 5 Why Did Medical Science Historically Implicate HIV? The standard method of trying to identify a new virus in a new disease suspected of being caused by a new virus, is to attempt to culture a new virus from an infected person, then show that antibodies to this virus are present in all people with the disease, but less often in people who are not ill. It is also helpful to show that persons develop antibodies to the virus during the acute illness. Sometimes viruses can be very difficult to culture in lab glassware. This is especially true of viruses which grow in human T-cells-- cells which could not be grown well without certain growth factors only discovered in the 1970's. In 1980, Dr. Robert Gallo of the NIH formally reported isolating a virus he named "Human T-cell Leukemia Virus," (HTLV) which infected T- cells and which was thought to cause some cases of T-cell leukemia in humans. This virus was a retrovirus (a virus type associated with cancer causation), and it was a distant relative of the "Feline Leukemia Virus" (FeLV) which caused leukemia in housecats. Because the transfusion results had shown that AIDS could be infectious, and because AIDS patients had abnormal-looking T- cells which looked something like those from retrovirus-infected animals, or T-cells in cultures infected with retroviruses, early AIDS researchers began hunting a T-cell retrovirus. In early 1983, a team of French scientists led by Dr. Luc Montagnier isolated a new retrovirus which they reported in May of that year, calling it eventually Lymphadenopathy-Associated Virus (LAV), because it had been isolated from tissues of a French patient with enlarged lymph tissues, or "lymphadenopathy" (this man died of AIDS in 1988). The French had been alerted to the possibility of a retrovirus in AIDS patients by an American team, led by Gallo, which was convinced that the AIDS virus was another variety of HTLV. It wasn't. [See appendix: "Montagnier, Gallo, Slip-Ups and Wrong Paths"] The new virus discovered by the French was a tiny, spherical, membrane-coated, protein-studded virus 1/100th the diameter of a lymphocyte, with an inner protein viral core shaped like a truncated cone, with a dense base. Under the electron microscope it didn't look like the feline FeLV or the human HTLV leukemia viruses (Fig 4), which had no distinct cores. Eventually, the LAV virus was correctly understood that Summer by the French team not to be a leukemia virus as they had thought, but rather to be the first human "lentivirus." This hypothesis was first formula- ted when Montagnier, at a suggestion from a colleague, began reading about "lentiviruses" or "slow viruses" -- a class of animal retroviruses he'd never previously heard of. In one book was an electron micrograph of the "equine infectious anemia virus," a virus which sometimes produced a familiar-sounding immunodeficiency and lymphadenopathy disease, after long latenci- es, in horses. Montagnier found himself looking at a tiny membrane-coated virus shaped like a sphere, containing a protein viral core in the shape of a narrow cone [67]. Antibodies against the horse virus cross-reacted with Montagnier's new virus, but not with Gallo's HTLV. Most importantly, coded AIDS patient serum provided by the CDC contained antibodies to LAV, but not HTLV-III, and Montagnier proved his lab could easily pick out AIDS samples from normal samples in the CDC material, without knowing the codes. Thus, Montagnier had his answer (his "LAV" was our modern HIV), but nobody would believe him for almost a year. Lentiviruses and Latent Diseases Most readers will remember that viruses in some sense are not complete living organisms. Animal viruses when outside cells don't metabolize, and cannot reproduce or grow by themselves. Instead, most viruses are little more than tiny floating packages of genetic material, sometimes without much other equipment. Viruses can reproduce themselves only by entering a living cell and commandeering the cell's synthetic machinery to subvert it into making more virus particles, which are then, in turn, released to infect more cells. A metaphor for a virus would be a truck-load of blueprints which rolls into a completely automated factory, and once there, is somehow able to use the blueprints to control the factory's machinery to cause it to make more sets of blueprints and more trucks to carry them, all of which are then assembled and sent out to take over more factories. [See appendix: "How Do Viruses Hide From the Immune System?" and "What is a Retrovirus?"] Members of one class of viruses use RNA as their genetic material, and are called "retroviruses," because their synthesis of DNA from RNA proceeds retrograde, in the opposite direction to what is "normal" in the rest of biology. Retroviruses avoid the body's immune system by inserting themselves into the DNA of the host cell. Most retroviruses cause no major disease, but not all are harmless. A sub-class of retroviruses, called "lentiviruses" is capable of slow infections resulting in death. Lentiviruses typically spend lengthy waiting periods in hiding in the cell nucleus (music lovers will recognize the Latin root _lento_, meaning "slow"), and even lentiviruses may never cause overt disease in their natural hosts. Sometimes, however, the lentivirus disease produced after a latency period can be devastating, though sometimes difficult to detect epidemiologically, due to the delay between initial infection and death. Lentiviruses were named in 1954 in honor of several very slow- acting brain infections of farm animals. The classic example was a sheep disease with a latency as long as a decade, called "visna" (an Icelandic word for "shivering"). Visna wiped out most of the sheep population of Iceland in 1939 because it had not been realized in 1933 that apparently-well sheep brought to the island from Germany had actually been carrying a latent disease. The ability of visna to cause disease many years after infection has since been demonstrated in a series of controlled experiments with sheep [68]. The visna agent proved eventually to be a retrovirus and a lentivirus, and the ominous pattern of the visna epidemic will become familiar to the reader during the course of this essay, as we discuss other better-known viruses in this unique class. HIV, our main subject, is a lentivirus. It is clearly related to the visna sheep lentivirus in structure and genetics, and more closely to "equine infectious anemia virus" a lentivirus which infects macrophages in horses. It is yet more closely related to FIV and SIV, two other immunosuppressive lymphocyte-infecting (lymphotropic) animal lentiviruses which the reader is shortly to meet. How To Dismiss All Laboratory Animal/Virus Experiments Before we meet these viruses, we should note that molecular biologist Peter H. Duesberg's scientific heresies do not only extend to his views on HIV. Duesberg, for instance, is unable to believe that any disease which confines itself relatively rigidly to classes of susceptible people, can qualify as a genuine communicable disease caused (in any sense) by a treatable microbe. In these diseases, Duesberg tellingly prefers to focus entirely on the host, refusing to consider the role of the microbe at all (as we will see, the issue of "causation" is a difficult one for HIV/AIDS skeptics). Thus, Duesberg and Ellison dismiss (for example) Legionnaire's disease and hepatitis C as fantasy diseases made up by the biomedical establishment for purposes of profit and self-aggrandizement (pp. 35-38, 57-59). Duesberg also does not believe in slow infections in general. The modern physician perusing Ellison and Duesberg's AIDS book will be astounded to read that syphilis microbes do not cause neurosyphilis, and that leprosy is not an infectious disease nor is it caused by the organism the medical textbooks say causes it (Ellison and Duesberg, p. 35). Nor does Duesberg believe in slow viruses at all, in the sense that he does not believe that any virus is capable of causing general infection or fatality years later in an organism which had been healthy prior to the initial viral infection (Duesberg, pp 209, 233). "There are no slow viruses," says Duesberg, "only slow virologists" (Ellison and Duesberg make it clear that they consider one of these to be D. Gajdusek, who won the Nobel Prize in 1976 for slow virus work, see pp 50-54). Nor does Duesberg believe that viruses cause cancer in animals, with a single kind of exception, the partial mechanism for which was co-discovered, as it happens, by Duesberg himself. Rous Sarcoma Virus (RSV), a retrovirus, causes muscle tumors in chickens (this fact won F. P. Rous the Nobel Prize in 1966). In 1970, Duesberg and P. Vogt showed that RSV caused tumors by carrying a cancer gene into the cell. (Thus, Duesberg's conten- tion that cancer is never caused by chronic viruses which have no cancer genes, such as HTLV and HIV, perhaps coincidentally has the effect of magnifying Duesberg's own discovery). A review of the evidence that Legionnaire's disease, leprosy, and hepatitis C are real infectious diseases, or that viruses other than RSV cause cancer, is beyond the scope of this article. With regard to other latent diseases it is worth noting that Duesberg's rejection of all current theories that viruses are causal in any latent cancer or disease of normal animals or humans requires that he reject not only all epidemiological evidence, but also all experimental laboratory animal evidence for viral cancer causation and fatal viral immunosuppression. Thus, cancer and immune failure in virally infected lab animals are uniformly dismissed as being responses of weak, inbred lab animals under "odd conditions" (Ellison and Duesberg, p. 91). The reader should note that, if valid, Duesberg's view about lab animals would have the effect of making almost any claim for a virus-caused cancer or immunodeficiency nearly impossible to prove. Epidemiology, after all, can always be dismissed in the absence of lab data. Moreover, it is difficult to impossible to collect experimental lab infection data on animals which have not been bred in captivity to be free of infections other than the ones being studied. In short, if infection-free lab animals were useless for the very infection experiments in which they are needed and now used, this would leave scientists with no good way to study infectious processes at all. Fortunately, however, enough experimental data is available to reject Duesberg's claims in the matter. With the exception of rodents, most animals used in laboratories are in fact no more highly inbred than are pets or domestic animals (or many societies of humans, for that matter). Modern science works with many tumor-virus models, and there is no reason to regard (for instance) the "specific pathogen free" cats used in feline viral leukemia experiments [69], or the woodchucks used in hepadnavirus liver tumor experiments [70], as being inbred, weak, or in any way intrinsically immunologically abnormal. Moreover, although Ellison and Duesberg (p. 84) say that feline leukemia virus and all other retroviruses only cause disease in young inbred lab animals, experiments have proven them wrong. In the next section we will discuss a retrovirus which causes fatal immunodeficiency and cancer, even when injected into so-called "random-source" domestic cats, such as one might obtain from a pet store or animal shelter [71]. Animal Lentiviruses Suitable For Laboratory Study In this section we will examine two very similar animal lentiviruses called FIV and SIV, and will note something of their effects in different animal hosts. The details about these two virus/host systems are given because all are crucial to a pattern which will be apparent by the end of our survey. The two viruses we are about to describe were actually discovered several years _after_ HIV, but it is more illuminating to tell about them first, for nature seldom provides her good clues in proper order. The reader should again bear in mind that our modern idea of the cause of AIDS is based on induction and inference, and inference depends on recognition of common _patterns_. Some of the crucial information for one of these patterns follows. By the end of the discussion, the reader should have some idea for why the "odd" effects of HIV in humans are no longer a surprise or shock to most scientists who study the matter. They've seen it all by now in animals, under controlled conditions, and realize that the way HIV works is much the way other viruses in its family work. _They Call It "Cat AIDS"_. The feline immunodeficiency virus (FIV) is a lentivirus discovered in Petaluma, California in 1986, where it was first obtained from the blood of two domestic cats living in a household in which there had been a number of deaths of cats from a strange immune deficiency disorder. The virus was isolated from the housecats by experimentally infecting special disease-free laboratory cats with their blood, then isolating the resulting infectious virus in tissue culture. The FIV microbe, once characterized, has since been identified in many species of cats around the world. The FIV virus is a tiny ball of lipid and protein, measuring four millionths of an inch in diameter, which is less than the span of 1000 hydrogen atoms laid next to each other in a line. An FIV virus particle is thus 1/10th the length of a small bacterium, or 1/100th the diameter of a white blood cell. The electron microscope shows the virus to be composed of a roughly spherical coat of phospholipids stolen from the host cell membrane when the newly made virus particle leaves or "buds" from the cell. This outer membrane is studded with protein molecule spikes, a bit like cloves stuck in a Christmas orange. The membrane closely covers a protein capsid shell which is a sphere or icosahedron (20-sided spheroid object made of triangular protein faces). The shell, in turn, contains within it a protein lozenge (called the viral core) in the shape of a blunt or "truncated" cone-- a cone which has a base which shows up notice- ably darker (denser) in the electron microscope views. The final result is a complicated structure not seen in any other virus class but the lentiviruses (Fig. 4). Inside the hollow protein cone which makes up the core of FIV are packaged two identical (or close to identical) coiled strands of viral genomic RNA, each containing 9 genes (a gene is a region of RNA which codes for and controls production of a main functio- nal virus protein, or set of proteins). Also present along with the RNA in the FIV viral core are molecules of a particular variety of reverse transcriptase enzyme peculiar to several previously known lentiviruses, including HIV. This reverse transcriptase functions best with a particular concentration of magnesium ion, and does not work (as this enzyme from other retroviruses sometimes does) with ions of the chemically somewhat similar mineral manganese. The RNA molecule genome of FIV is organized very closely along the lines of several other lentiviruses and is clearly related to them, with most genes and proteins having identifiable analogues from one strain of lentivirus to another. FIV, however, infects only the cat family. It causes disease in domestic cats (pets, feral cats, and lab cats), but apparently infects lions without causing disease. Not surprisingly, once FIV infects a cat, a DNA virus copy finds its way into the cat DNA, and stays with the animal for life [72]. It is what happens to FIV-infected domestic cats, however, that makes the FIV virus most interesting: in late life, ap- parently many years after being first infected, some FIV-infected cats come down with an immune deficiency syndrome. Older cats infected with the virus are worst affected. The presence of antibody in the cats' blood is not protective against damage, since the presence of FIV antibody makes no difference in later development of severe symptoms and death in older cats. Indeed, even with antibody present, the immune system of infected cats was found to be declining in a peculiar way-- CD4+ lymphocyte were disappearing while CD8+ lymphocytes were relatively unaffec- ted [72]. The lymphocytes of cats also happened to be the cells preferentially infected by the FIV virus in cell culture (a incubated glass dish containing cells and a nutrient solution). FIV virus infection sometimes killed cells in culture and sometimes not, depending on the strain of the virus and the type of cat cell. In FIV-infected cats, immune systems began to malfunction (sometimes years after infection) as CD4+ lymphocytes numbers declined. As scientists followed naturally-infected older cats, they saw some cats begin to waste away and develop lymphadenopathy (swollen, enlarged lymphatic organs or "nodes"-- the filtering structures which hold immune cells and trap invading microorganisms). Infected cats also developed fevers, diarrhea, and chronic neurological dysfunction-- the last a result of direct brain infection by the FIV virus. Finally, after a year or two of being ill, the unfortunate felines usually died of opportunistic feline infections [73]. Naturally, scientists wished to prove once and for all that FIV was the entire cause of the disease syndrome which they were seeing in FIV-infected cats, now informally dubbed "feline AIDS" (i.e., a life threatening immunodeficiency syndrome caused by a retrovirus, in cats without cancer). With experimental cats in the laboratory setting, experimenters were able to do what could not be done in humans, which was to cause deliberate infections in an animal with a lentivirus. FIV experiments with cats have taught us much, and allowed us to infer still more. When young cats were infected with FIV by injection in the laboratory, they suffered lymphadenopathy, coat and skin problems, and lethargy. They also developed very slow and progressive and specific loss of CD4+ lymphocytes and a decline in immune response over a period of years. The FIV virus was also active in the brains of deliberately infected cats, and they suffered various abnormalities in neurological function [74]. Older cats were worst affected, just as with HIV in humans [75]. Cats experimentally infected with FIV were observed to sometimes enter a carrier phase after infection, but this might progress to lymphadenopathy, wasting, and death as long as 3 years after infection. It didn't happen in uninfected control cats. Experimentally infected specific-pathogen-free lab cats (domestic breed cats raised in the lab without many microbes or parasites that affect pet housecats) often developed lymphomas eventually. "Random-source" domestic cats, exposed to and carrying many more normally-present "cat germs" from the outside world, not surpris- ingly were more likely to develop immunodeficiency and AIDS-like opportunistic cat infections when brought to the lab and given FIV [71,76]. Domestic cats infected with FIV are known to develop candidiasis and cryptococcosis, rare cat microbial infections also seen in human AIDS. Toxoplasmosis, a common cat parasite usually benign in both cats and people with good immune systems, was another organism deadly in both human and feline retrovirally acquired immune deficiency. In all FIV-infected cats, destruction of lymphatic tissues in a manner similar to that seen in AIDS patients (but no other human infectious disease) was found, with FIV virus growing actively in the lymphatic tissues [76]. What was really needed for further study, however, was an animal even more closely related to man, which could be infected with a similar immunosuppressive lentivirus. _A Tale of Two Colonies: Simian AIDS._ The needed non-human animal model of AIDS arrived by chance, a decade before it could be recognized or put to use. In 1976, a colony of stump-tailed macaque monkeys at the California Regional Primate Research Center (University of California at Davis) developed a disease epidemic-- an infectious syndrome of immunos- uppression, lymphocyte loss, lymphadenopathy, wasting, and death. Scientists at that time had no idea what kind of infection was killing the monkeys, but they did preserve some of the obviously diseased lymphatic tissue of one dead monkey in 1977. Many years later, in the mid 1980's, they isolated from this frozen tissue a virus which was by then familiar in both structure and function. In 1977, however, nobody recognized simian AIDS [77]. Macaque monkeys infected with the new agent developed a peculiar and severe immunodeficiency, and sometimes they develo- ped it after unprecedented delays, causing mistakes to be made in assuming that monkeys were recovered and well, when they were in fact chronically infected and doomed. At the primate colony at the Yerkes Primate Center (Emory University, Atlanta, GA), which had previously been "closed" to contact with other lab primate colonies since 1964, four apparently healthy stump-tailed macaques were introduced from the apparently "clean" California Regional Primate Research Center in 1981 [78]. In retrospect this was to prove as much a disaster as the visna-infected sheep which passed the quarantine in Iceland. The California Primate Center colony was not discovered to be still infested with its old 1970's immunodeficiency plague until 1986, when the plague broke out in California again. At that time, newly available antibody survey tests showed that two of the animals previously transferred to the Yerkes colony in 1981 were indeed infected with the latent retrovirus, but (in 1986) were still without symptoms after 5 years in their new home. It was not until 1988 that much of the Yerkes colony, still "closed" to other contact with other lab primates, began to die of an immunodeficiency syndrome, with loss of half of the animals in the colony over the next year. Many of the Yerkes macaques, including the two infected transferred animals, began to first develop immunodeficiency more than 7 years after first being exposed to the infectious agent from the other colony, and their disease syndrome consisted of opportunistic infections, lymphadenopathy, and selective loss of CD4+ lymphocytes. Scientists found that all the dead adult animals (including the transferred monkeys) had antibody in their blood to the newly identified viral agent, so again, antibodies had not conferred immunity to the immunodeficiency disease. The agent responsible for the Yerkes disaster was by this time known to be a lentivirus, a virus that has since become known as Simian Immunodeficiency Virus, or SIV. Eventually, a number of different SIV strains were isolated from both captive and wild African monkeys, each with its own story of detection, and sometimes also disaster in an expensive commercial colony [79]. The SIV virus strains, like FIV, turned out to be interesting in a number of ways. SIV was, by the mid-1980's era of AIDS, also a horribly familiar-looking virus. It was a tiny sphere, 1/100th the diameter of a lymphocyte, and it had a lipid membrane studded with proteins, and an inner protein core that was a truncated protein cone with a dense base. Inside were two identical or near-identical RNA molecules, associated with a peculiarly magnesium-dependent reverse transcriptase enzyme. The organization of its 9-gene genome was similar to other lentiviruses, and much like that of FIV. Like many other viruses in many classes, SIV turned out to be very host-specific in its ability to cause disease. SIV infected its natural hosts (apparently a number of species of _African_, not Asian, monkeys) without causing any obvious sickness or immunosuppression in them. The hosts developed antibodies to the virus, but in the case of SIV, like FIV in cats, these antibodies did not signal immunity, since the virus went on quietly hiding in the nuclei of cells until ready to reproduce and infect other animals. Only when transferred to genetically different _Asian_ species of primates (several species of macaque monkeys) did SIV cause illness. It killed animals quickly when antibodies did not develop, but of more interest was that it was quite capable of killing monkeys after more than a year, long after a strong antibodies response did appear. SIV had originally been trans- ferred to Asian macaques perhaps by housing primates from Africa and Asia in the same cage (in retrospect, not a good idea). Apparently, the African SIV virus did not know quite how to behave in the Asian macaques (which do not carry it naturally), and in these monkeys SIV caused both rapid and slow syndromes of immunosuppression and death, depending on the host's immune response. If the virus from a dying macaque was transferred back into African monkeys, however, it went right back to being benign-- infecting and reproducing without causing sickness. Again, all this was an old story to virologists, who had long known that severe viral disease and death syndromes usually result from mismatched duets between a virus and a "new" species of host for which the virus is not yet adapted. Viruses, after all, do not evolutionarily "want" to make their hosts deathly ill-- they want their hosts to be up-and-about, and spreading virus. Generally, a virus does not kill its normal host. For SIV, the consequences of mal-adaption to Asian monkeys caused the scientific interest--- for SIV infection in macaques caused what looked more like human AIDS than anything seen yet. Again, in SIV infection, T-lymphocytes were infected, and CD4+ lymphocytes were preferentially killed, leaving CD8+ lymphocytes. Again, simian lymphatic tissues became swollen, and were even- tually destroyed during SIV infection. Again, the disease progressed inexorably in some infected animals, even in the face of antibody "immunity," and without showing much active virus in the blood (progression was in the lymphatic tissues, which became damaged in a way sometimes indistinguishable from the stereotyp- ically-destroyed lymphatic tissues found in human AIDS patients). Again the SIV virus paradoxically caused little damage to lone CD4+ cells in culture, but obviously also did something in the host which totally destroyed the CD4+ system in the living animals. At the end of the SIV disease course, the CD4+ lymphocyte-depleted primates developed not only a few monkey- specific diseases, but also an eerie number of the same opportun- istic fungal and parasitic infections that were familiar from human AIDS: oral candidiasis (yeast infection of the mouth), _Pneumocystis carinii_ pneumonia (a fungal pneumonia), cryptos- poridium diarrhea (a protozoal intestinal infection), cytomegalo- virus (CMV) retinitis (a herpes-type viral eye disease), and _Mycobacterium avium_ (a tuberculosis bacterium relative). SIV also caused a "retroviral encephalitis" -- a brain infection caused directly by the SIV virus which was microscopically similar to the disease changes seen in human AIDS brains. Macaques with SIV also developed lymph-cell-associated tumors (lymphomas), a frequent AIDS complication in humans [80]. Many monkeys, like cats, have been sacrificed in scientific studies to learn about infections that could not be studied in humans, but in the history of science few monkeys have been sacrificed so usefully as the laboratory monkeys deliberately infected with the various strains of SIV. Again, scientists wondered if the SIV virus was itself capable of all the immune pathologies seen in SIV infected macaques. Was the SIV virus necessary to simian AIDS? Was it sufficient for simian AIDS, or were other infectious agent co-factors (perhaps even other viruses) being accidentally transferred with SIV from animal to animal, in experiments? At Harvard and elsewhere, the question was answered using the techniques of molecular biology. Teams using harmless viruses and bacteria as "carriers" were able to clone, and thus purify, a provirus, or DNA-copy of the SIV genome. They were even able to clone half of the virus genome in each of two cultures of bacter- ia, and then reconstruct and piece together a working SIV virus pro-genome (set of DNA-based instructions), out of the two cloned parts of the virus DNA. No other infectious agent but SIV could have survived this specific manufacture of SIV. The resulting cloned SIV genome, after being used as "instructions" to manuf- acture fresh whole SIV virus in cell culture, was found to cause immune destruction with lag periods of up to 11 months in rhesus macaques [81]. Whatever viral information was being cloned in the SIV genome, that genetic information was enough to cause deadly immune destruction in Asian primates. SIV "molecular clone" experiments proved that SIV was sufficient to cause the AIDS-like syndrome it produced in macaques, with no co-factors or co-infections being required (other than those that all healthy macaques normally possessed). Uninfected primates remained healthy; primates infected with SIV genomes manufactured by cloning often died deaths of immunosuppression or cancer months later [82]. Thus, SIV genes alone somehow held the key to, or perhaps--metaphorically-- the fuse to, the total immune destruction seen in the lab animals. Science was able to prove this much, simply because deliberate and pure infections could be carried out. When it came to simian AIDS in the laboratory, the SIV virus was capable of being the sole inciting culprit which made the difference between health and death. Although SIV and FIV were discovered after HIV, there remains a lesson for us from each of them: Lentiviruses alone in some circumstances is quite capable of causing slow immunos- uppressive death. AIDS Heresies, Part 6 The Critics Again Or, at least they are for those who are willing to look. Dr. Peter Duesberg spends very little time on the monkey SIV experim- ents in his writings, and what time he spends is spent on select experiments. Concerning SIV, Duesberg notes that in one experim- ent, lack of good antibody response to a cloned strain of SIV in monkeys predicts simian death, suggesting that the virus is powerless against a good immune response. Unfortunately, Duesberg fails to note that this finding applies only in the short term in a paper where animals were not followed for the long term [81]. Thus, while it is true that experimentally infected monkeys die quickly of SIV in the short term if they cannot mount an antibody response, it is also true that they can also often be expected to die of SIV-caused immune suppression after 1 to 3 years of infection, despite good antibody response [83]. This has also been seen in naturally transmitted infectio- ns: a long SIV infection progressing to AIDS even with good antibody response has been observed after experimentally-monito- red sexual transmission of SIV from animal to animal [84]. Also, as noted previously, the natural latency period for SIV before it causes immune disease and death has been observed to be as long as 7 years in animal colonies, and such long-latent disease, when found, also occurs in the presence of good antibody response [78,79]. Duesberg believes that the presence of antibodies to a retrovirus confers immunity on the host (p. 233), and thus that it is unlikely that those people infected by HIV who develop antibodies (making them "HIV-positive"), should develop active, damaging HIV infection later on. But it has long been known in general that antibody response does not protect against long delayed immune failure in the slow but fatal immune deficiency disease caused not only by SIV and FIV, but also by many retro- viruses, such as the simian leukemia virus and feline leukemia virus [85]. Nor does antibody response protect against the persistent and sometimes later damaging infection in lentiviral diseases in sheep and goats [86]. In writing of SIV infection, Duesberg is also at great pains to emphasize differences between the infection and human AIDS, pointing out that SIV infected monkeys do not get dementia (how would we know if they did?) and Kaposi's sarcoma, but failing to note that monkeys do get most of the other human AIDS infections, and have much the same basic pathology evident in their destroyed lymph tissues and immune systems as do human AIDS patients. Indeed, Ellison and Duesberg (p. 135) go so far as to character- ize simian infections with SIV as not resembling AIDS at all, instead merely being "...in all respects, very traditional viral flu-like diseases..." This is a pathological absurdity which we have already addressed and refuted at length. If Duesberg fails to discuss the relevant animal work with any degree of care or detail, Root-Bernstein does not even try. Any argument based on induction must turn on standards of evidence, and bias in standards of evidence abounds in arguments about the cause of AIDS, just as it does in all scientific arguments (the arguers being human). Thus, Root-Bernstein (p. 330) asserts: "The entire case for HIV as the cause of AIDS rests upon epidemiologic correlations..." He knows better. Root- Bernstein himself is willing to use a wide range of animal experimental evidence in indirect support for his own multifacto- rial AIDS hypotheses, yet he steadfastly maintains a peculiar blindness for the same kind of evidence when it supports the HIV/AIDS hypothesis. For instance: Root-Bernstein is quite capable of noting for purposes of inference many instances where a possible non-HIV causal factor of AIDS causes relatively mild unselective lymphoc- yte suppression, or even relatively mild selective CD4+/CD8+ ratio suppression, in laboratory animals (p. 117, 121, 124, 127, 130, 137, 138). Yet there is total silence from Root-Bernstein about the immunological effects of infection by animal lentiviru- ses like SIV and FIV. Indeed, the gap in Root-Bernstein's discussion of causes of immune suppression is so glaring that one is left wondering how it is that a literature search as careful as that represented by his book entirely failed to note the effects of certain viruses on immune parameters. Root-Bernstein discusses the effects of many micro-organisms, including viruses, on immune function, and even gives us the effects of CMV, EBV, and hepatitis B viral infection on CD4+/CD8+ ratios (p. 151,152,- 153). The results: the ratios may change as in AIDS, but not nearly as drastically. Yet Root-Bernstein writes nothing about the effects of any _retrovirus_ of any kind on these immune parameters, whatsoever. This, even though lymphotropic retrovir- uses are the only organisms known which in the laboratory mimic the especially severe CD4+ lymphocyte loses, severe CD4+/CD8+ lymphocyte ratio suppressions, lymph node destruction and opportunistic infection susceptibility found in AIDS. Nor does Duesberg mention such effects, except in one instance in which he makes an observation that SIV retrovirus "barely reduces the T-cell levels of ill monkeys" (p. 232), a statement which is false as a generalization, since there are in fact a number of studies which have found significant T-cell and CD4+ lymphocyte depletion in some SIV-infected monkeys [80]. Note that it is not that the effects of retroviruses on immunity are discussed and dismissed in heretical literature; rather, they are generally not discussed at all. In reading the AIDS here- tics-- even those heretics doing detailed surveys of the bio- medical literature about AIDS-- it is as though information on the effects of lymphotropic retroviruses on animal immune function did not exist. The Retrovirus From the Human Epidemic SIV clearly caused immunosuppression and death in monkeys, but no such simple answer was forthcoming with humans when it came to the question of what was causing AIDS, although certain lab and blood donation incidents came close to being controlled experiments with HIV. At the NIH, three workers have so far been accidentally infected with pure molecularly-cloned HIV, with the result that two are running abnormally low CD4+ counts some years later, and the third has lost almost all CD4+ cells, and develo- ped opportunistic infections. (Though many of Duesberg's "HIV- free" AIDS cases are based on nothing more than slightly abnormal CD4+ counts, in a display of double standards Duesberg reportedly refuses to acknowledge two of the lab HIV infections as AIDS by the same criteria) [22]. In another example, an African study found that children who had been HIV infected by blood transfusion given for malaria and other reasons, had a far higher death rate than matched children who had received the same amount of blood for the same illnesses but did not contract HIV [87]. In this study, 6% of patients who received HIV-positive blood and had been infected, but no HIV- negative transfused controls, had developed clinical AIDS after one year. Such studies underscore the close relationship between HIV-positivity of blood and risk to the recipient of dying. Duesberg has claimed [88] that no study shows a higher death rate in HIV infected people than in matched HIV-uninfected people-- yet here is such a study. With the new infectious immunodeficiency syndrome in humans first identified before FIV and SIV were known, science at first had only epidemiologic correlations to work with. Later, some results from viral cultures and matched clinical studies arrived. Nevertheless, the scientific community decided early that it was working with a viral disease of unusually long latency. After HIV was identified, statistical studies quickly picked it as the most likely answer yet proposed for the cause of AIDS. But there was more than statistics-- also present in HIV and AIDS was an overall pattern of viral structure and immune failure which proved far from unique as time wore on. HIV had the typical lentivirus structure, identical to that of FIV and SIV. Inside its cone-shaped core were two more or less identical RNA molecules, each containing 9 genes, and associated with a peculiarly magnesium-dependent reverse transcriptase enzyme. People infected with HIV were found to develop antibodies to it in the blood usually a month or two after infection, and keep them for life (although levels of antibody "targeted" to stick to the core proteins of the virus often did decline toward the end, in full-blown AIDS, as numbers of virus particles in the blood rose). The presence of antibodies in the blood for years did not seem to keep people with HIV from eventually becoming very ill, however. Since HIV was discovered before the many careful studies with FIV and SIV, the progression of HIV infection to serious illness in the face of serum antibody "immunity" to it, puzzled some scientists quite a lot. Today, we know from much closer experimental study that lentiviruses are routinely able to mutate to escape host antibodies, so that antibodies found in the same blood with these viruses often do not neutralize them, especially late in the course of disease [86]. The HIV virus proved to have a special affinity for CD4+ lymphocytes, the very cells which disappeared in AIDS patients (and, as we have seen, disappeared also in many lentivirus-infected animals). The HIV virus grew readily in CD4+ lymphocytes in culture (not usually causing them harm, but sometimes doing so [89]), and with the ability to grow the HIV virus in cultures of cells and to detect antibodies to it, came the ability of scientists to detect infections caused by it. Because HIV infection almost always resulted in HIV antibody, but since the antibody did not signal cure or the end of virus infection, the HIV antibody test was later used to track cases of transfusion-associated AIDS. In nearly all cases where archived samples of transfused blood could be tested later, people with transfusion AIDS were found to have gotten units of "HIV-positive" blood. In the U.S., 28,000 people received HIV- positive blood products before testing halted such transfusions in 1985, and of these, 5,879 have developed AIDS as of July, 1993. These figures, allowing for less than 100% transmission of HIV by this route, are quite similar to the AIDS rate in HIV- infected people with hemophilia. HIV-positive blood was later shown to hold not just antibodies to HIV, but also the infectious HIV virus itself. The Slow Development of HIV Infection in the Body. Although people could not be deliberately infected with HIV for ethical reasons, antibody testing to see who had been infected with HIV suggested that HIV was indeed transmitted sexually (though with very low efficiency), and also through contaminated blood products. In time, the virus was "caught in the act," as a number of people were identified by antibody testing and viral culture as they were first undergoing infection. This "primary" infection turned out to sometimes be rather like an attack of infectious mononucleosis ("mono") -- a long and severe sore throat, swollen lymphatic tissues, fever, and tiredness for weeks. Sometimes it was less severe, and was not even noticed (a good sign which was found to predict a longer time to development of AIDS [90]). After the initial onslaught, however, the infected person developed HIV-antibodies (i.e., become "HIV-positive," or as epidemiologists say: "sero-converted"), and then began recovering. Occasionally a person newly infected with HIV would rapidly seroconvert, lose most CD4+ lymphocytes in the blood, and go on to develop full-blown AIDS in as little as 8 weeks [90], so scientists knew this could happen (just as some monkeys developed AIDS in a few months after SIV infection). Most people, however, apparently recovered completely after initial HIV infection, and felt well. But only apparently. The virus was not gone, but had only gone underground, into the DNA of many of the host's cells. As with other lentiviruses, the long latency period of HIV before secondary disease was striking. Throughout the largely symptom- less latency period the antibody against HIV always continued to be detectable in the blood, making the infected person HIV- positive. But there was always more than antibody. Even years after infection, viral DNA could still be detected essentially 100% of the time in an HIV-positive person's CD4+ lymphocytes, and other cells as well. Especially easily infected were immune system cells called monocytes and macrophages. In a person with AIDS, up to 13% of lymphocytes and monocytes in the blood were found infected [91]. From 93% to 100% of the time (depending on the study), infectious virus and viral DNA could be recovered from the blood of asymptomatic people who were HIV-positive, and 100% of the time from people with AIDS [41]. Higher levels of virus could be cultured from the lymphatic tissues of such people, where it was multiplying actively, even in people who appeared healthy [92]. Like SIV-infected monkeys and FIV-infected cats doomed to future immune failure, HIV-positive people were still infected, and most were still slowly losing immune cells in lymphatic tissues. In some sense they were still "sick," even though they might feel and appear healthy. By following large groups of HIV- positive volunteers with blood tests and exams over more than a decade, scientists began to piece together what was happening. Over time, the average numbers of CD4+ lymphocytes in the blood of groups of such people (but not in HIV-negative people) were slowly falling at a steady rate, although not at the same rate in everyone. At the same time, until late in the disease, levels of CD8+ lymphocytes actually rose. The lymphatic tissues of some infected people were found to be under viral attack, and after some years began to show this by enlargement, as was also often seen in lentivirus-infected animals. This phase had early been called the "AIDS related complex" or ARC syndrome. In the lymphatic tissues, a particular stringy virus-trapping cell called the "follicular dendritic cell" disappeared, and with this loss came dysfunction of the other immune cells which depended on the follicular cells to catch and digest viruses for evaluation. (Cells in the immune system evaluate infectious organisms by partly digesting them into smaller pieces and then "tasting" and sharing the fragment- ary results among themselves, for later recognition). Even- tually, in an HIV-positive person with lymphadenopathy, most of the CD4+ lymphocytes in the lymphatic organs (where 98% of the CD4+ lymphocytes in the body were normally to be found) were gone. Exactly what had happened to them was not clear, but they had disappeared from the blood as well, and had obviously been destroyed. At about the time the lymphatics were reaching the end stage of destruction, the HIV virus, escaping filtration by the now-destroyed immune system in the lymphatics, began to enter the blood in larger numbers once again. The largely asymptomatic and reasonably healthy latent period was now at an end [93]. Levels of CD4+ lymphocytes had been falling at an average rate of 60 per year, and when levels of CD4+ lymphocytes in the blood reached the critical count of 250 to 200 (about a quarter of normal levels), fevers and other symptoms began. Opportunistic infections often also appeared, such as candida yeast infection in the throat (often the first infection), or the unusual fungal pneumonia caused by _Pneumocystis carinii_ (this organism is particularly suppressed by CD4+ lymphocytes, and so is one of the first infections to arrive when their numbers fall). In homosex- ual men, but much more rarely in other AIDS patients, a peculiar vascular tumor called Kaposi's sarcoma might appear on the skin. (This disease has long been thought to be caused by a second infectious agent, which only recently has been detected by DNA sequence and antibody detection [93a], and suggested to be a new virus in the herpes class, called HSV-8). With the appearance of these markers of secondary disease, now the infected person was said to have "full-blown AIDS," or simply, AIDS. Estimates of the time between initial HIV infection and later AIDS were found to vary strongly with the age of the person, and to a lesser extent on the group infected (homosexuals, with many other concurrent infections, developed opportunistic infections sooner, and also Kaposi's sarcoma). The latency time did not depend heavily (if at all) on the sex of the infected person [93b]. Generally, the period taken for half of a given infected group to become sick with full-blown AIDS was about 10 years-- with perhaps a few more years for younger people, and a few less for older people. The development of AIDS signaled a permanent loss of much of the body's ability to effectively fight viruses and fungal diseases, although the ability to fight bacteria was only mildly impaired. This gave the immune deficiency in AIDS a very different character from the kind of immune deficiency caused by many cancers, drugs, and other diseases. In AIDS, as, first, CD4+ lymphocyte count dropped toward zero, then finally at the end of life, even CD8+ lymphocyte count dropped sharply, finally even more devastating infections arrived, such as viral diseases of the retina and brain. Chemotherapy-resistant lymphomas became common. By the time AIDS suffers died, usually a year or two after first developing full-blown AIDS, it was not uncommon for them to have no CD4+ lymphocytes in the blood at all, and for them to have lymphatic tissues which had been essentially destroyed. Before scientists became acquainted with the lymphotropic (lymphocyte-infecting) lentiviruses, they had never before seen _any_ disease, toxin, or condition specifically destroy nearly every CD4+ lymphocyte in an animal, while leaving CD8+ cells relatively untouched until near the end. The specificity of lentiviruses in causing destruction of one part of the immune system, while leaving another part relatively intact, was awesome and very strange. It remains so today, not least because scientists still have not discovered exactly why it happens. One good guess is that the body's immune system somehow destroys infected CD4+ lymphocytes (as it does other virally infected cells) before they have time to make it into the blood, until eventually they are all gone. Many other possible mechanisms for destroying these cells have been proposed, and limited evidence for each proposed mechanism exists. That infection with a lentivirus can destroy an immune system is not the issue, since lentiviruses are known for certain to have this capability in animals, and thus a mechanism for this certainly exists. The only issue left is what this mechanism is. Telling Nature How Her Viruses Must Behave It is unfortunate that HIV/AIDS skeptics spend so little time on the behavior of lentiviruses in animals, for HIV is a typical lentivirus, and many of the things critics are saying HIV cannot do in humans are things which lentiviruses are known quite well to do in animals. The inference is not difficult to draw. It is one thing to assume that HIV cannot cause disease long after there has been a good antibody response to its initial infection, and another thing to assume this even though many examples are known in animals of the viruses' cousins operating in exactly this way. Animal models quite often provide "existence proofs" of mechanisms which overconfident scientists, on the basis of limited understanding of biology, might otherwise dismiss as being unlikely or impossible. That is one important function of animal models: to provide humility to biologists. Usually models are successful in this, but not always. The status of Dr. Peter Duesberg as a leading retrovirologist has made it easier for him to be taken seriously when he expres- ses disbelief at the way in which HIV is hypothesized to damage the immune system. Duesberg's objections of disbelief are made possible by the fact that we do not presently understand every step in the process by which lentiviruses (or indeed most viruses) cause disease. This does not prevent Duesberg from dictating how it should, or must, happen. "If HIV were the cause of AIDS," writes Duesberg, "T-cells would drop and AIDS would appear during the primary infection, when HIV titers [amounts of HIV virus in the blood] are high and there is no viral immunity" [8]. Unfortunately, based on the way both FIV and SIV are documented to kill animals after long intervals with good antibody production responses, it would appear that Duesberg has no good grounds to believe that HIV should not, or could not, do the same as a pathogenic agent in humans. Similarly, a great number of other properties of HIV which Duesberg believes argue against HIV's pathogenicity, also happen (to the detriment of his argument) to be properties of the demonstrably pathogenic "simian-AIDS" virus SIV, which (as we have seen) closely resembles HIV. There is, for instance, no question that SIV infection alone is sufficient to cause simian CD4+ cell-loss immunodeficiency and death, and that this may happen with delays of up to 3 years in experimental infection [81, 83]. In short, much of Duesberg's inference that HIV is probably harmless is built upon the very shaky proposition that HIV-1 is not likely to be able to do in humans what FIV, SIV, and (as we will see) HIV-2-- all very similar retroviruses-- are already known both observational and experimentally to do, without assistance, in animals. Properties of SIV infection, in common with HIV, include: similar low level viral titers in animals which are doomed to die later of retrovirus infection, lack of pure CD4+ lymphocyte cell killing by SIV in culture [89], with paradoxical profound loss of CD4+ lymphocytes and lymph organ destruction in SIV-infected animals [94], and finally, absence of any obviously special SIV genes which might confer virulent properties with regard to other retroviruses (instead, the infection latency and eventual death with SIV infection in some monkeys is a property of the host- virus system as a whole, and can be seen with one strain of HIV-2 or SIV, and not another). All these viral properties or infection characteristics have been claimed by skeptics to indicate that HIV is probably harmless, and yet we know that their presence with both SIV and FIV certainly does not indicate harmlessness for these organisms. By analogy it would seem, if anything, that HIV is not necessar- ily harmless either. The Koch Postulates of Disease The parallels between SIV and HIV turn out to be useful also in an odd debate involving the "Koch postulates" of disease causality, which has lately been resurrected by HIV/AIDS skepti- cs. Duesberg, almost alone among virologists, believes that viruses must fulfill these postulates to be proven to cause disease. Koch's postulates, named after a 19th century physician Robert Koch, date from a time before viruses or antibody testing were known, and have been discarded by most scientists in the modern era, since they can lead to bad inferen- ces. To repeat a theme of this essay: statistical correlation causes many difficulties in trying to infer causation, yet it is still useful. For instance, there are many people with sore throats who are not infected with the micro-organism popularly known as "strep," and many people are infected with strep for decades who don't have sore throats. However, we could not legitimately conclude from these data that strep does not cause _any_ sore throats. Rather, the issue is what fraction, if any, of sore throats are caused by strep. In order to even begin to answer this question, we need to know at minimum how "tight" are mathematical correlations between the presence of the putative infectious agent (strep), and the disease syndrome (sore throat) we think it may cause. Dr. Robert Koch postulated some simple rules which he felt should be fulfilled for an organism to be proven to cause a disease. The first was that the organism be findable in essenti- ally all cases of the disease. Strep would fail the first Koch test as the "cause" of the syndrome of sore throat, but today we would not say that this was proof that strep did not cause sore throats. Rather, we would measure the prevalence of strep in various populations of asymptomatic and sore-throated people, and then employ statistical methods in order to arrive at a probabil- istic estimate of what fraction of sore throats strep was likely to cause. Koch, too, would use such statistical methods later in life when he was forced to abandon his own rules in order to inductively guess the cause of cholera. In short, the world has moved beyond Koch's postulates, and even Koch in his own life time did so. But not Duesberg. We have examined Duesberg's attempts to define AIDS in such a way that many AIDS cases would be HIV-free, rather than define it in any clinically useful way. These are misguided attempts to disprove the HIV/AIDS hypothesis using Koch's simple un-statistical laws. Another of Duesberg's arguments relating to the first Koch postulate holds that HIV is not present in _sufficient quantities_ to cause disease in humans who have AIDS. But how do we know what "sufficient quantity" is, when it comes to lymphotropic lentiviruses? This ought to be a case where animal models should help. HIV is in fact present in the blood of AIDS patients in quantities very similar to the amount of SIV in the blood of SIV-infected monkeys, and we know that without doubt SIV is sufficient to cause severe immune deficiency disease leading to death in monkeys [82]. (During much of both SIV and HIV infections, most of the body's retrovi- rus is present in the lymph tissues, not the blood). Thus, animal experiments tell us that there seems to be something wrong with Duesberg's intuition regarding how much retrovirus "should" be observed in the blood, during a retrovirus infection which will eventually be fatal. We also know that the replicating HIV virus is indeed present in high quantities in the lymph tissues, where the destruction in AIDS is primarily seen [122]. Another of the Koch postulates requires that an organism be isolated in pure culture (not even close to possible for a virus when Koch was alive, but possible now with modern viral culture and molecular cloning), and that this culture then be used to transmit the disease to a susceptible host. Duesberg has pointed out that HIV does not cause disease in non-human primates, and has not been demonstrated by deliberate and controlled direct transfer to cause disease in man. However, we should note that failure to cause disease by experimental transmission is true only for HIV-1, since recent experiments have now succeeded in producing animal disease with another strain of HIV, called HIV- 2. HIV comes in two strains, HIV-1 and HIV-2. HIV-1 is the "familiar" strain isolated from almost all American patients; HIV-1 does infect and replicate in chimpanzees, but except for lymphadenopathy does not cause noticeable illness in them (this is not strange behavior for a virus-- deadly yellow fever and hepatitis B viruses behave in the same benign way in chimpanzee- s). Quite recently, however, when researchers were able to isolate a different lentivirus from AIDS patients in West Africa (with a few cases in other parts of the world), they found that the new virus, eventually dubbed HIV-2, was only 50% identical in genome to the more familiar HIV-1, but was genetically almost indistinguishable from the SIV family of monkey viruses (The figure of 50% may seem a significant amount, but it is for instance similar to the difference between the two common varieties of herpes simplex, either of which can infect the mouth or the genitals, and both of which cause very similar diseases). If HIV-1 and HIV-2 are as close genetically as Herpes-1 and Herpes-2, then the question of whether or not HIV-2 causes an AIDS-like syndrome in monkeys would become somewhat relevant to a causal debate involving HIV-1. Since HIV-2 is the only retrovirus isolated from some West African AIDS cases [95] it was natural for researchers to see if the HIV-2 virus, so similar to SIV, killed Asian monkeys (but not African monkeys) in the manner of SIV. They found, by injecting HIV-2 into Asian monkeys, that this virus does indeed behave like SIV [96]. HIV-2 destroys the immune system in Asian monkeys, somehow causing lymphadenopathy, depletion of CD4+ lymphocytes over time periods of up to a year, and fatal opportunistic infections. In other words, HIV-2 infection gives the monkeys the simian version of AIDS [97]. Further, the longer the chain of infection in monkeys, the more deadly the human virus becomes for them. HIV-2 also causes an AIDS-like condition when injected into baboons [97a]. In all this work, clues have emerged about where the HIV-2 virus came from originally. Humans appear not to be the natural hosts of this virus, just as they are not the natural or earliest reservoir for HIV-1. The result of this work is a chain of inferences, but with help of the new HIV-2 results, not a difficult chain to follow. The retrovirus known as HIV-2 has been found by direct experiment to cause low CD4+ lymphocyte counts in animals, destruction of animal immune systems, opportunistic infections, and death. This HIV-2 virus had been isolated from humans dying in Africa from an epidemic of something that destroyed CD4+ lymphocytes, destroyed immune systems, and as a result allowed opportunistic diseases resulting in death (in other words, AIDS). If HIV-2, which causes AIDS-like CD4+ immune destruction in monkeys, is not causally involved in the severe human immune problems in West Africa where it is isolated from AIDS cases, it would rank as one of the greatest scientific coincidences of all time. The magnitude of the coincidence would be similar if HIV-2 does indeed cause AIDS, but the 50% genetically homologous HIV-1, already implicated for many other reasons in AIDS, does NOT cause disease in the human AIDS cases from which _it_ is isolated. The art of scientific inference lies in deciding how many coincidences it is wise to believe in. AIDS Heresies, Part 7 The Ultimate Proof The HIV/AIDS hypothesis has not been proved by experimentally infecting completely healthy humans deliberately with HIV under controlled conditions. Thus, many of the standards of proof required by those who are skeptical of the HIV hypotheses would seem to be impossible in an ethical age. Apparently no organism but man becomes significantly ill with HIV-1, just as is the case with a number of other viruses, such as hepatitis B and yellow fever virus (both of which historically were proved to be human disease organisms in direct human experiments which would never pass a human subjects protection review board today). Root- Bernstein notes that human-specific viral diseases satisfy Koch's postulates only by being transferred deliberately and experiment- ally from a sick person to a healthy one (p. 95), and sometimes such a trial appears to be what skeptics demand for HIV. Root- Bernstein states: "..correlation, no matter how good, is never grounds for asserting causation. One must have experimental control over the disease." (p. 101). It is not clear what Root-Bernstein means by "experimental control over the disease." Should enough of the extreme HIV skeptics (which, to be fair, Root-Bernstein is not) take a page from the notebooks of self-experimenting disease-hunting doctors like John Hunter (who died after inoculating himself with syphilis) or Walter Reed's brave colleagues James Carroll and Jesse Lazear (who allowed mosquitos carrying yellow fever to bite them, resulting in the death of Dr. Lazear) and deliberately infect themselves with HIV-1, we might have something. Many physicians have historically risked and even given their lives for science and their beliefs, when there was no other way to continue an experimental program on human subjects [1]. Peter Duesberg has indicated unwillingness to be inoculated with HIV, saying it would prove nothing if he survived in good health, as he expects to. Duesberg is probably right in this, but since HIV is thought to cause steady and implacable immune decline in more than 90% of people infected with it, highly statistically significant results would be generated if even two healthy skeptics infected themselves with HIV, and both survived without any evidence of increasing immune deficit, for 5 years. Thus, it may be that putting together the necessary experiment lacks only some organization. One Florida physician named Robert E. Willner, author of an amazingly misinformed AIDS heresy book called _Deadly Deception_ [98] has already performed at least two TV-televised needle inoculations on himself with the blood of an HIV-positive man with hemophilia (no word yet on Willner's HIV antibody status, but from what we know of viral blood burdens, he is inoculating with too little blood to realistically be of much risk from a non-ill HIV-positive person). Such cases will probably remain too scarce for conclusions, especially if the adequate pre-testing necessary for good science is not done. Still, if any self-inoculating heretic does contract HIV, and then AIDS, this way, it will certainly make news [98]. Meanwhile, how far can we go with the AIDS statistical results that we have? How does associative statistics, coupled with knowledge of related mechanisms, actually work in guiding formation of a good scientific theory? Induction: Telling Causation From Mere Correlation Since the writings of the 18th century philosopher David Hume it has been known that certainty about physical causes and effects is not to be had from mere association, even if the temporal sequence is correct. (In fact, according to Hume, absolute certainty about positive causation is not to be had even when we can do direct experiments. Most scientists ignore this most extremely pessimistic view of causal knowledge, or else they view the high degree of inductive certainty gainable with experiment to be "good enough.") It seems that Hume is at least correct that to some extent the causal conclusions of science are always uncertain, because they involve mechanisms and rules we can never be completely sure of understanding, even if we had some way to guarantee that future events will continue to be bound by any "rules" we _could_ be sure we understand now. A favorite fable, which philosophers tell to illustrate the problems of induction, features a turkey on a family farm who observes that every time the farmer comes to his turkey-run, he comes to feed the turkey. The turkey thus hypothesizes a causal connection, and predicts feeding each time the man comes-- and as time goes on, sees this theory "verified" many times. But on the fourth Thursday of November the turkey's well-verified theory suddenly becomes mal-predictive, because the universe is a much more complicated place than turkeys can comprehend. Humans are sometimes in the same position with human science: unfortunately, the universe is such a complicated place as to sometimes make turkeys of us all. Nevertheless, to the extent that today's future, which we cannot know, continues to be like "yesterday's future," which we can check, progress can be made. A turkey's existence is made easier for a time by having even an incomplete model of reality, and we may ourselves say that even partial understanding of causes brings partial "knowledge." Inductive knowledge is uncertain the way a weather forecast is uncertain, but this does not mean that science does not uncover a certain kind of truth. A weather forecast may not guarantee the future, but the point is that it is usually better than the results of flipping a coin or guessing. Probabilistic knowledge, therefore, is still genuine knowledge of a special kind. Indeed, it is the only kind of knowledge we can have about what events will occur in the future. Because the association of events may be always controlled by a third factor we do not suspect, the problems with induction become most acute in trying to tease out causal relationships in systems where we cannot experimentally influence events in order to rule out possible causal relationships. In such systems, statistics help us tell which factors are important and independ- ent of each other in _predicting_ future events, but such associations only _suggest_ causation when present. This is because any factor "predictive" of an event may still only be merely a good "proxy," or marker, for an even more predictive, mechanistically causal factor. Thus, the strength of statistical associations between events and possible causes are far more helpful in ruling _out_ possible causes than ruling them _in_. Obviously it is very difficult for a factor to be causal of an event if its presense is not _at least_ independently predicative of the event. Here an example may be helpful: If we find from a multivari- ate analysis of (say) food sales and crime, that weekly ice cream sales during a calender year predict juvenile delinquency arrests from week to week, but spaghetti sales do not, this is reasonably good evidence on the face of it that spaghetti sales are NOT causally connected to juvenile delinquency. On the other hand, for ice cream and juvenile delinquency, the associative connect- ion found by the statistics is merely suggestive-- here we may be looking at cause and effect, _or_ in the case of ice cream may be looking at a variable which is a proxy for something else which is more directly causal of delinquent behavior. But if we doubt ice-cream causation of delinquent behavior for other reasons, we must guess what the real causal variable is, before statistics can help us further. If we find that, say, the daily average _temperature_ predicts juvenile behavior even better than ice cream sales predict it, and that knowledge of ice cream sales gives no predictive power independent of knowledge of temperature, then we now have evidence that ice cream is not causal, but is _merely_ a proxy, or marker, for _temperature_ during the year. Even with this better association, we are not still not assured of the causal role of temperature, although we might guess for "mechanistic" and even indirect experimental reasons that we are closer to the right track for temperature than we were with the ice cream sales. So we've made headway. This story, though silly, illustrates a major path by which science in general makes progress. Good and Bad Theory-Making: Occam's Razor and the Multiplication of Hypothetical Causes. As in the example above, to make progress in science requires that we continue to propose new and better causal factors for effects which interest us, and then test these putative causal factors statistically and experimentally. Independent prediction serves as the best statistical test for factors which we cannot vary experimentally. The power of prediction is thus all- important in evaluating candidate causal factors for the cause of effects which we cannot directly manipulate, such as AIDS. As noted, statistics cannot do independent thinking for us, and we must rely on something other than statistics to come up with our list of possible causal factors (such as the average temperature in the example) to test with our statistical methods. This is the business of the human imagination, and very possibly also the human sense of order and beauty. It has been observed by the late Karl Popper, noted philosopher of science, that almost no theory is ever ruled out by experiment, because with enough imagination, nearly any theory can be tinkered with after the fact, so that it continues to "explain" all data. Thus, if one causal factor does not explain results statistically in a given situation, it is not necessary to drop it-- one may instead postulate an additional factor which explains results in the case where the first one fails. In fact, if one persists in hypo- thesizing new factors each time an old factor fails, one need never drop any old hypotheses at all. At some point, however, doing this makes any theory simply too ugly and ungainly to be believed, and at that point (if a better alternative is in view) many scientists may decide to discard the old theory [99]. (Or, at least, most scientists will-- for as physicist Max Planck pointed out, only death removes the last die-hard believers in _some_ theories). When it comes to AIDS, it is of course possible to construct a theory in which AIDS is NOT caused by an infectious agent, but a theory with so many causal agents that it still covers all the available evidence. To do so, however, requires that a newly emerging simultaneous and significant cause of mortality in many diverse groups of people be explained rather "unnaturally" (at least to this author's sense of such things) in a great number of different ways. For example, Duesberg suggests that AIDS in Africa is caused by malnutrition and tuberculosis and new misclassification (p. 220), in U.S. male homosexuals by a new habit of recreational nitrite use (p. 248), in female IV drug abusers and their newborn children, by newly popular drugs other than nitrite (p. 215), and in many of those HIV-positive people who do not use illegal drugs, by use of prescription AZT since 1987 (p. 241), etc. [100]. Deaths in non-drug-using groups like people with hemophilia in the era before 1987 are explained as being due to immune suppression and extra life extension by clotting factor concentrate, new since the early 1970's (p. 219) [101]. For none of these suggested causes of AIDS is there any mechanistic or experimental evidence of good quality, since none of the agents proposed, or anything related to them, causes in animals the severe and specific suppression of CD4+ lymphocyte numbers characteristic of AIDS. In Duesberg's theory, AIDS deaths from transfusions are argued to be non-existent (p. 214), rejecting the orthodox idea that simple transfusion or trauma- associated mortality usually covers them up (but see [87]). AIDS deaths in contaminated hospital workers are rejected as anecdotal and perhaps due to something else, such as personal drug abuse (p. 211). AIDS deaths in wives of men with hemophilia (and their children) are dismissed as being due to normal aging and misclas- sification (p. 219), etc. One of the chief arguments against this kind of theorizing may be aesthetic, but a less subjective matter has to do with utility. Because there are no statistical data to estimate a quantitative role for many non-orthodox postulated risk factors for AIDS (such as immunosuppression from undiscovered viruses or from certain behaviors) the problem with very complicated _post hoc_ explanations is that they are retrodictive, but not predictive. If HIV is only a proxy factor, or marker, for a number of habits or practices which are supposed to "explain" immune failure better than HIV infection does, then it follows that the quantitative and even the qualitative presence of such problems should be usable to predict future development of AIDS in asymptomatic people, to even better accuracy than HIV status does. Nothing of the kind has been shown, however. On the contrary, cohorts of HIV-positive people have been shown to develop AIDS at a predicable rate, fairly _independently_ of most of the factors that have been suggested by heretics to be the real causal factors for which HIV is merely a marker. (The reader may see [100] and internal references, and also the review in [102] for studies of the development of AIDS over time in prospectively tracked cohorts of HIV-positive male homosexuals, people with hemophilia, and transfusion-infected people). Duesberg's hypothesis, as compared with Root-Bernstein's, actually has the charm of a certain simplicity, since for Duesberg immune failure is said to be due to only two categories of toxins: foreign blood proteins or drugs. In Duesberg's view, these two causal factors are supposed to have independently (and, apparently, coincidentally) begun producing an epidemic of immune deficiency in 1) people with hemophilia and 2) everyone else-- both starting in 1982 or so. "Multifactorial" hypotheses such as Root-Bernstein's, however, posit so many different contributing causes of severe immunosuppression (again, failing to differentiate between significant ones and non-significant ones) that one or more of the putative additional risk-factors identified by Root-Bernstein is likely to be present in addition to HIV, in almost every AIDS case. _Rethinking AIDS_ argues the case that besides HIV infection, AIDS might be caused by blood product infusion (both whole blood infusion and clotting factor concentrate injection), by surgery and/or anesthetics, by accidental trauma, by age (both extreme youth and old age) by use of any and all common illicit drugs and most pharmaceuticals (especially antimicrobials used in AIDS), by concurrent infection with any of dozens of microorganisms (known and unknown), by rectal insemination, by malnutrition, and even by sunlight exposure. Considering all these hypothetical cofactors, Root-Bernstein writes (p. 92): "If such agents exist... it is not enough to demonstrate that HIV is present and highly correlated with AIDS. It is also necessary to demonstrate that these other agents are not present in AIDS patients. That is impossible." Here Root-Bernstein somewhat overstates the case, since the epidemiologic statistics contains methods for testing putative risk-factor contributions in ways far more subtle than looking for people with simple presence or total absence of all factors being assessed. Far more important to the issue of causation, for those putative risk-factors which can be quantitated, is what the mathematical effect of a larger "dose" of one factor is upon the total risk, independently of the others. Example: looking at statistics, is cancer risk smoothly increased by smoking more cigarettes per day, other factors being equal? If so, that tells us something very suggestive. On the other hand, are smokers who spend various amounts of time in the sun, from a little to a lot, at larger risk with each larger increment of sun exposure, all other factors (such as cigarettes smoked) held equal? If not, this tells us something, too. In general, it is not necessary statistically to find cases where each factor which is a possible cause of a disease is totally absent, in order to amass evidence that the factor in question is not causal. It is merely enough to show that the increasing or decreasing _quantitative_ presence of the factor in any "experiment of nature" has no independent statistical effect on probability of the disease. Such a demonstration never rules out causation completely (due to annoying problems with induction which always allow for the insertion of additional hypotheses to "explain" failures in prediction), but it does make causation much less likely-- at least for those who like their causal hypotheses as simple as possible. Such statistical tools (i.e., multivariate analysis) are among the most powerful which scientists can use for exploring initial possibilities for causation, in complicated and difficult-to-manipulate systems. Without such statistical tools, for instance, it would never be possible to exonerate any one of any common set of behaviors (such as drinking coffee, walking, or eating certain foods) from a causative role in ANY disease. The reason is that we live in a complicated world, and participation to some extent in one or more of a certain set of behaviors describes ALL of us, and certainly all of us who develop diseases. But we are fairly certain that walking and driving a car (for instance) do not contribute to (say) lung cancer, because of multivariate analysis of life-style, NOT because somebody has described a group of lung cancer sufferers who never left their houses. The following example may illustrate how, if too many un-quantitated factors are postulated to explain an effect, statistical science cannot help us sort through them, since only quantitation can separate effects or lack of effects of possible risk factors which are widely prevalent. For our example, consider a new hypothetical "multifactor" theory that (say) lung cancer is not caused by smoking alone, but also by "hidden immunosuppression," caused by certain behaviors. This immunosuppression might be generated (say) by: certain kinds of sexual contact, exposure to unshielded sunlight, exposure to recreational and pharmaceutical drugs, chronic stress from depression and urban life, previous surgeries, and concurrent infections (including lung infections and infections by herpes and other chronic viruses). If we generate a long enough list of hypothetical co-factors for lung cancer, we will have found, for almost everyone who smokes and develops lung cancer, some other hypothetical "risk-factor," other than smoking, which has never been measured quantitatively in relevant groups. If, in additi- on, the very few people with cancer who smoke but deny all of our new hypothetical risk factors, can be dismissed as being of questionable veracity, we will have constructed a brand new "multi-factorial" hypothesis of lung cancer. This theory will be complicated, difficult to test, and not possible to evaluate statistically immediately, because the data do not exist to do it-- nobody having previously bothered to measure them. But just what good have we done by constructing such a new theory? Just how useful would our new model be? If, for none of our new hypothesized risk factors, can the extent of exposure to these risks be used now to statistically predict lung cancer rates independently of smoking (in the way that, say, fruit intake, age, or radon exposure can)-- then none of our newly hypothesized risk-factors are useful at all. Perhaps they may be useful in the future when measurements are done. Or perhaps not. But would we be justified _now_ then in using their _possible_ existence to berate those scientists who would like to focus study on known statistically independent explanatory risk-factors for lung cancer-- such as smoking, age, diet, and radon-radiation exposure? Would we be justified, having done nothing more than having constructed a new hypothesis of lung cancer involving new un-quantitated variables, in writing books on our new causal proposals with chapter subtitles like: "How could so many scientists be so wrong?" (Root-Bernstein, page 350). The answer is no. Nor is Root-Bernstein justified in this kind of thing, based on what evidence he has. Formless hypotheses involving large numbers of unknown etiologic factors which will be prohibitively expensive to search for until we have some better indication of where it will be most profitable to look, do not help the cause of human knowledge. When considering complicated multi-factorial hypotheses, where most of the multiple factors are unquantitated, missing, or un-named, one is reminded of the great physicist Wolfgang Pauli being asked if he thought the theory in a new scientific paper was wrong, and his remark that that would be too kind. Said Pauli: "It isn't even wrong." [99]. In the case of the AIDS heretics, actually, the situation is even more extreme than in our lung cancer example, since in the case of AIDS, statistical data DO exist to evaluate some of the heretically proposed "co-factors" for AIDS. The statistical influence of many of these (illicit non-injected drug use, AZT use, total clotting factor concentrate use) have been found to be statistically _not_ independent of HIV exposure, and their AIDS- risk predictive power disappears completely in study after study, when HIV status is controlled as a variable. All these proposed co-factors, then, are proxy variables. There is thus very little justification in continuing to advance hypotheses which involve them, since none have passed even the most basic initial statist- ical standards used by epidemiologists. AIDS Heresies, Part 8 Impact of Belief on Action: Do More Complicated Theories Gain Us Anything? As we have seen, epidemiologic criteria implicate HIV as a necessary causal agent in all AIDS cases, so long as AIDS is defined most narrowly (and most usefully) as the newly epidemic syndrome of life-threatening immune _failure_ (_not_ mild immune suppression) which occurs in certain groups of body-fluid exposed people known to be at risk for this new acquired disorder. Moreover, an examination of the pathogenesis of virally-caused CD4+ immune failure and slow death in animals infected with viruses related to HIV, demonstrates that it is quite possible biologically that HIV operates in humans to cause the same sequence of pathologic events as we see in AIDS. As regards the last point, our retrovirus "proof of mechanism" information is useful, if only in a negative way. As we have seen, because of such information it does no good for heretics to express disbelief that there could be any mechanism for retroviruses to do the immunologic damage which orthodox medicine thinks they do (perhaps the common theme of every heretical HIV/AIDS argument). Since we know from direct experiment that retroviruses _do_ in fact do this kind of immune damage in animals, even in the presence of serum immunity and with low blood viral levels, we _know_ a such a mechanism must exist in animals. Thus, such a mechanism is _not_ unlikely to exist in humans as well. The fact that we have not found yet what we logically know must be there in animals, should not bother us too much, and by extension, _the same is true of our lack of success so far in humans_. Let us frankly admit it: we know very little of the actual molecular mechanisms behind _most_ of the causal sequences which we have identified by experiment in biology. Thus, at this point the fact that we don't yet know exactly how HIV and the other retroviruses "work," is no more and no less odd than the fact that we don't yet know exactly how cancer, or fetal development, or memory, or the aging process works. Complaining in 1995 that we've spent a decade and lots of money on AIDS-- or for that matter the aging process, or embryo development, or cancer, or the basis of biological memory-- and only gotten more bogged down, is childish. This list of biological problems includes some of the most complex problems science will ever face-- so complex, in fact, that in the 20th century we're simply out of our league as regards most of them [103]. Again, (pace Root-Bernstein) _all_ infectious diseases are known to be influenced by inoculum and host defense, and there is certainly no reason to believe that HIV infection is an exception. _Intrinsic_ qualities of the host which are important in all known diseases, and so too presumably for an HIV-caused AIDS, would include ordinary variations in genetically-determined immune response, and also things like host age. These qualities we may term "intrinsic host factors." Root-Bernstein refers to host age as a "cofactor," but for purposes of our discussion it may be more useful to differentiate intrinsic host factors which cannot be changed, from those more conjectural environmental cofactors which might be subject to manipulation--- such as diet, dangerous habits, or infection by other organisms which are susceptible to antibiotics. The reason, of course, is that the last are where the treatment payoffs lie. It is, I believe, generally assumed by orthodoxy that risk of HIV infection is influenced by such standard infectious disease variables as virulence of the HIV strain, amount of the HIV virus inoculated, and route of inoculation into the body (which will not be independent of amount inoculated). The role of none of these things is fully understood for HIV infection, but the idea that there is a large role to be found for each of them, when we finally understand AIDS more completely, is not really controversial. Possible cofactors in the acquisition of HIV and development of AIDS have been discussed in the mainstream scientific literature for years, and are not in any danger of being dismissed any time soon. Real controversy in the clinical treatment of AIDS is almost entirely centered about the role of conjectural extrinsic cofactors which are manipulatable _after_ the time of HIV infection, and thus not related to sexual mechanics, the genetics or age of the host, or the inoculum and virulence of the HIV strain in question. Again, Root-Bernstein differentiates between AIDS theories in which HIV 1) plays a necessary role _along with_ such additional _and also necessary_ conjectural extrinsic cofactors; and 2) those theories in which HIV may play a role in AIDS, but is not necessarily required. The latter theories in which HIV may contribute to AIDS, but is not required, all suffer, like Duesberg's hypothesis that HIV is totally harmless, from this problem: full-blown AIDS, as defined by very low CD4+ lymphocyte counts seen epidemically in groups at risk for acquired immune failure, is essentially never seen without HIV being present. Again, this is true for no other infectious agent. It would be odd indeed if an infectious agent which is not really necessary for AIDS, is the only agent found always to be present in those who die of it. And if HIV is so good at finding out and infecting _all_ those who are immunosuppressed before they die, why is HIV not also found in most or all people who are somewhat CD4+ immune-suppressed for classical reasons, such as cancer or tuberculosis? By contrast, theories which posit that HIV is necessary for AIDS-- but that other immune-suppressive "extrinsic" or change- able cofactors are necessary ALSO, are more difficult to evalu- ate. We have examined a number of the arguments against specific proposals for hypothetical necessary extrinsic co-factors-- such as licit and illicit drugs, specific micro-organisms, and blood- product-effects on the immune system--and noted that studies do not find these ideas borne out by independent statistical regression, controlled for HIV status. This is not to say that we will not one day identify an important extrinsic risk factor; only that the evidence is against the ones so far proposed. We have also argued that intrinsic host factors, viral strain, viral dose, plus extrinsic cofactors like sexual mechan- ics factors and the presence of other mucosal lesions (as from sexually transmitted diseases) at the time of HIV infection, are sufficient to explain most broad variations in AIDS epidemiology. Of course, there is no guarantee that here again nature will not prove to be more complicated than our simplest theories. If extra complexity proves necessary to describe the risk of contracting HIV, however, we shall have to wait for clues that it does, in order to know how best to proceed toward a more compl- icated hypothesis. Such clues, as argued above, have not arrived so far. How important is it that we understand causation in AIDS as fully as we can, and as soon as we can? Obviously, having a more realistic view of the pathogenesis of AIDS matters, in that it partly determines where we put research dollars into prevention and treatment studies, to say nothing of insurance dollars into standard clinical practices. If HIV is merely a bystander virus, it won't be very helpful to study it, and it may even be counterproductive to treat it, if the treatment carries some risk on its own (as AZT and similar drugs certainly do). Similarly, even if HIV is only partly causal in AIDS, if another cofactor modifiable after HIV infection is necessary to development of AIDS as well, then (as Root-Bernstein points out) AIDS might be prevented or cured by treating something which we have some control over, even after a person shows up HIV-positive. This would give us something to do to improve or even guarantee health in HIV-positive people, and that something might be more effective, and not as dangerous, as taking long term antiretroviral medications. The correct view of the pathogenesis of AIDS should also important in prevention as well. If HIV is merely a bystander virus, there would be no more reason for an HIV-positive person to tell a prospective sexual partner about a previous infection with HIV than there would be to tell about a childhood case of chicken pox. And there would be no more (and no less) reason to use a condom if one was HIV-positive than if one was not. Obviously, an incorrect view of the pathogenesis of AIDS might be dangerous and expensive to many people with the wrong idea (and their sexual partners), no matter which view ultimately turns out to be correct. Prevention, of course, does not only apply to sexual transmi- ssion: if viral theories of HIV causation are correct, we should shortly see the positive effects on crude mortality rates, as a result of preventing people with hemophilia from becoming HIV- infected from their clotting factor concentrate, an intervention begun in 1985-1986. If this does not happen (and because of the long latency of this disease it is too early to tell now), we will know that we missed something very important. Duesberg has already proposed that we save the money that the HIV test for every pint of donated blood costs us, and stop doing it. If we cannot with time prove that halting blood-borne HIV infection has stopped transfusion-associated AIDS, to say nothing of almost all AIDS in younger people with hemophilia, we will have little excuse for continuing to test blood or clotting factor concentrate. Alternative views of the cause of AIDS are probably becoming more popular today, simply because they offer more hope than conventional views of the cause of this disease. A virus is implacable-- especially one hiding in the DNA so intimately that it is part of the genetic fiber of one's being. It would be much preferable to some to believe that AIDS is due to a combination of more tractable problems. This possibility has led to some unusual biomedical research suggestions: Root-Bernstein suggests (p. 369): "If the multifactorial, synergistic theory of AIDS is valid, then treating mice, rats or rabbits with some combination of, say, multiple blood transfusions, cytomegalovirus infection, anesthetics and surgery, and opiate painkillers could well result in the immunological suppression typical of AIDS." And maybe not. The idea that a lot of small immuno- suppressions might add up to a large one, is an interesting idea, but thin fare to base anything but a grant proposal on, until the experiment is actually done. In this author's opinion, it would certainly be inexcusable to base a book which advocates changes in the standard management of an often fatal disease on such theorizing (Root-Bernstein stops just short of this, but many of his readers will not). By contrast to experimental models of a multifactorial immune suppression "hypothesis" (if such a vague entity is worth of the term) of AIDS, we know experimentally that infecting Asian primates with SIV or HIV-2 virus DOES IN FACT sometimes result in exactly the type and severity of immunological suppression typical of AIDS. We have the added bonus in these experiments of producing this interesting result with nothing more than a single virus, quite similar to the one we always find in AIDS patients with immune failure (and the very same virus, in fact, as we find in West African AIDS patients). One would think that this kind of animal model would excite skeptics who wish understanding of the nature of this disease to come to science-- but for some reason it does not. Skeptics are silent about HIV-2. Perhaps the reasons have to do with other than pure logic. In _Rethinking AIDS_ Root-Bernstein ironically quotes Einstein's version of William of Occam's philosophic "razor": "Keep hypotheses as simple as possible, but no simpler." Failure to follow this maxim can be very costly in the real world. Resources for research are not infinite, and in practice, diversion of resources toward experiments to test _a priori_ epidemiologically unlikely hypotheses, can result in much wasted time and (of course) lives lost. One can imagine for instance, how long-delayed might have been the program to control polio (a disease which was also, for a long time, refractory to the "one agent--one disease" approach criticized (p. 108) by Root-Bernstein) had multifactorial theories of disease been pursued on all fronts for polio paralys- is. To this day we still cannot explain why the polio virus causes paralysis in some few, but not most, infected people--- yet science succeeded in finding an effective preventive strategy for polio paralysis, nevertheless. We long ago stopped polio by simply stopping the polio virus, which we had found was a necessary element in the disease. We _ignored_ the other cofactors (whatever they were) which caused some people to be paralyzed, and not others. There is perhaps a simple lesson here. Root-Bernstein, in the concluding paragraphs of his book, brings us to the crux of this practical issue (p. 372): "Assura- nce in science, he writes, "comes only through elaborating as many possible explanations as can be imagined for a phenomenon and eliminating all that can be possibly eliminated." A historian of science ought to have inserted major qualifications in such a statement. There are actually an infinite number of possible explanations which can be imagined for any phenomenon, and trying to eliminate them all would leave both science and government paralyzed and bankrupt. Real scientists, and the real government agencies that fund them, need a better plan. In practice, only a certain relatively small segment of possible kinds of theories are accepted or acceptable in science, and these are based mostly on the success of previous small classes of theories. As both evolutionary biologists and successful businessmen know, no efficient search of any large set of possibilities proceeds by random search of all possibilities, but rather must succeed (if it is to succeed) by exploration of a much more limited set of combinations (randomly chosen or not) of previously successful possibilities. Major hypotheses worthy of the expense and time of experimental testing in science are not generated by elaborating every possible explanation. Rather they come through a still mysterious process of mentally narrowing down certain classes of possible explanations to a few "good ones," and then testing those. (If there were no mystery about this process of narrowing down millions of theories into a few good candidates, it would not be necessary to speak of scientific "genius"). Again: "assurance" in science comes not from "eliminating all possibilities that can be eliminated," as Root-Bernstein seems to unwisely suggest, for (again) the resources for this do not exist in the real world. Assurance in science comes from publicly testing for "prediction power" a relatively *few* leading candidate explanatory theories, in ways that will usefully differentiate them from each other, and THEN finding that one particular theory (or small class of theories) repeatedly survives such tests, while the other likely theories do not. At present, there is no getting around the fact that the HIV- infection theory of AIDS is the leading and most predictively successful causal hypothesis for AIDS. Progress is thus most likely to be made by continued testing of this theory, and likely variations of it suggested to us by various kinds of statistical evidence. Testing of a causal theory of disease, of course, is most profitably done by testing its prediction of the occurrence of disease, or by its usefulness in preventing disease. A theory is not "tested" by requiring that it provide the route to a successful treatment for disease as well. Our present failure to cure either AIDS or lung cancer does not speak to the truth or usefulness of theories about their causation. Allocation of resources for the testing of more radical causal theories for any disease (of which there are, of course, an infinite number) must be made on the basis of some prior likelihood of their being more useful or more explanatory (our only current tests for "truth"), since otherwise much money would be spent uselessly chasing moonbeams and bad guesses [104]. When a new causal theory "predicts" (retrodicts) a phenomenon only after the fact, and predicts it no better than our current best theory, which is much simpler and already available for no more money-- then our current theory is to be preferred. On the public level of funding, there would seem little reason to waste money on the infinite number of alternate theories of disease which may be more comforting than the standard one, but which have (given current knowledge) too little chance of being true to bet much money on. It is to this point that we have come with AIDS. In AIDS research, as in most research, science is now heading generally (though never perfectly) toward the area of investigation in which the largest number of medical scientists and epidemiologists think maximum payoff is most likely to lie. This is the research direction which seems likeliest to most experts to save the most lives, with some caveats (see appendix on biomedical funding problems). We all wish there were modifia- ble necessary extrinsic cofactors for AIDS, but we also know that (in keeping with a saying as old as the metaphor) if wishes were horses, then beggars would ride. It is a long way from postulat- ing such modifiable co-factors for AIDS and thinking how useful they would be if there, to actually finding them. The bottom line is that, since there are an infinite number of such possible co-factors, those who wish for change in research direction policy in AIDS had best come up with better arguments for why we should divert a great deal of our limited attention and resources to look for things that we presently have too little reason to believe may exist. Nobody cares, of course, if a small fraction of research money is diverted to wild ideas and bold guesses (it's impossible to keep this from happening, actually); but survey funding is one thing, and major policy changes are another. In the meanwhile, is there any legitimate and useful and honorable role for the scientific heretic in matters of life-or-death public policy? That depends on whether said heretic publishes his contrary views in scientific journals, or _Spin Magazine_. There is a great difference between debate within the scientific and medical community, which may be (and o often must be) quite freewheeling, and the recommendations which this community should formally offer the lay public, which ethically must emerge from a better consensus of the best minds, and be somewhat more conservative in nature. From the workings of Darwinian evolution to the workings of business; from speaki- ng, to writing, to work in any field of science, there is no creative act in which the selection/editing process is not critical. Attempts to circumvent the particular expert-review editing process which produces progress in science certainly do not help matters. In particular, initiating biomedical scienti- fic debates at the lay public level, where most chemical, immunological, and mathematical or statistical evidence for one view versus another is bound to be lost, is to invite public policy disaster. [Footnote] ---------------------------------- Footnote: It is not best that medical treatment debates be kept secret, but perhaps it is best that they sometimes be left confined to professional journals where those especially-well motivated lay people who work hard enough to root them out and decode them will probably in the process also gain enough knowledge to begin to see the data in perspective. Certainly, the idea of partly shielding debate may sound patronizing, especially as it involves the field of medicine, which has come under intense criticism of late for this very flaw. And certainly, questioning authority is reasonable when one's interests are at stake. (One is reminded of Napoleon's remark to a famous beauty that he did not approve of women involving themselves in political discussion, and her reply that in a country where women were having their heads cut off, that it was natural for some of them to want to know why). Although the struggles of medicine in this area have been especially controve- rsial, the problem of the "expert" does not only arise in medicine, for we are all experts at something, and we are all ignorant of most things. Any expert asked for advice holds a special obligation (knowing that the asker may not have the time or the power to check conclusions), to edit knowledge, and to pass along only conclusions which are relatively sure, and which have been independently checked by other good minds. The asker _trusts_ the expert to do this, and indeed the essence of relationship between the knowledge-seeker and the expert is one of this kind of trust. When the expertise and the advice involves a life-or-death matter, the level of obligation and the level of trust are amplified, and this can result in behavior in those trained in the medical profession which may appear more paternalistic than it is. More often, such conservative behavior is better underst- ood as a manifestation of the wisdom, understood by responsible adults, to refrain from saying everything one thinks, in situa- tions where what one says really matters. The person who gives medical advice about life-and-death matters to the lay public in popular publications is under an ethical obligation to refrain from pushing a private theory with which most experts disagree, no matter what it is (no rational person, after all, can be so certain of his own conclusions as to bet other people's lives on them before his theory has been well-checked by others). Linus Pauling, to his credit, probably did no harm with his popular common cold suggestions, or probably even with this suggestion to give mega-dose vitamin C to terminal cancer patients. The AIDS heretics, however, are in a different ball game. -------------------------------------- End Footnote The Future and AIDS in the U.S.: Many Reasons for Hope The HIV viral plague is the great plague of the 20th century, infecting as many as a million people in the U.S. alone, with 1/3 of a million people already having developed AIDS. Almost all of the people with AIDS now alive can be expected to die within a few years, barring some immediate major technical advance or discovery. What about the rest of people with HIV? If there is no evidence that converting to a drug-free, medication-free lifestyle assures the harmlessness of subclinical HIV infection (and there most assuredly is NOT), then what can be done? And what about those people who are still HIV-free? Are they destined to see the disease sweep through their communities, eventually catching everyone who is not perfectly chaste, careful, and lucky? Let us, for the moment, follow William of Occam and assume that there are no important post-HIV-infection modifiable cofact- ors for AIDS left to identify, since at present we really don't have much reliable AIDS epidemiologic data that cannot be explained in broad terms under the headings of host and behavio- ral cofactors that we assume exist already. Suppose HIV is indeed the major culprit for development of AIDS, exactly as appears, and that risk of AIDS, once HIV infected, is more or less predetermined already by the virus and host, with only minor modifications in AIDS risk possible though modifications of the most extremely risky behavior, such as discontinuance of IV drugs and the more extreme forms of promiscuous behavior. Studies show that very long-term HIV survivors, as we have discussed, are NOT saints [46], and thus there is no evidence that saintly behavior, once a person is HIV-positive, does any significant good for survival (however psychologically appealing for some this thought may be). Is there much hope for the population of the world, even on these grounds? Perhaps. I believe there is reason for optimism, even if the simplest AIDS hypothesis which fits the facts (which is the one now commonly accepted) turns out to be substantially correct. The first bit of optimism is on behalf of people who do not have HIV yet. It appears that the HIV virus is very hard to catch, with favorable odds seen so far, even for U.S. call-girls who report that even they do not use condoms with some trusted regular customers. The risk for transmission from a woman to a man is particularly low in the absence of other sexually transmi- tted diseases-- so low, in fact, that when American women contract HIV from men, the men they contract the virus _from_ almost never contracted the virus themselves from other women, but rather from needle-sharing during injected drug use, or from male homosexual contact. Since it is usually difficult to hide the scars and the lifestyle which are corollaries of the kind of illicit drug use which involves drug injection equipment-sharing, it follows that a discriminating American woman's major risk of AIDS is from heterosexual contact with a secretly bisexual man. Even this risk can be greatly reduced by avoidance of anal intercourse, and sexual contact with men who do not appear per- fectly healthy. (As we have seen, HIV is far more transmissible in the acute early "mononucleosis-like" stage of the disease, and also much later in the final ARC and AIDS stages of the disease). As for men's risks, exclusively heterosexual men in America who have sexual partners at low risk for other concomitant sexually transmitted diseases, have a very low risk of acquisition of HIV. These facts alone are probably enough to insure that in the present sexual climate, condom use or not, HIV will not spread to any great further extent by exclusively heterosexual transmission in U.S. The heterosexual AIDS pandemic we have been warned of for a decade, in other words, is not clearly coming at all [105]. Indeed, there is every reason to believe that such an epidemic would be here already if it was destined to arrive soon. Constant numbers of HIV infected persons in the U.S. for the last several years point strongly to the probability that at an infection prevalence of only 0.3% of the total population, HIV has largely run its course in most populations susceptible to it in this country [101]. (Because of latency delays between infection and disease, new yearly AIDS cases will continue to rise rapidly for years, even so). In the farther future, sexual mores may yet change greatly, of course-- and it may be that even now, among select groups of heterosexuals (inner city teens, for instance) promiscuity and lack of treatment for other sexually transmitted diseases will allow for a future small heterosexual- ly-transmitted AIDS mini-epidemic. Much remains to be seen. What does seem likely, however, is that for the near future, most of the U.S. population will escape AIDS--- even should no vaccine or treatment be found. [Note from 1996: recent results suggest that strains of HIV transmitted heterosexually in the Eastern hemisphere may be different from the HIV "B" strain transmitted primarily intravenous drug use and homosexual contact in the U.S. If some strains of HIV really do prove to be far more easily transmitted by heterosexual sexual contact, and these strains have not yet reached the U.S., the assumptions made above may yet prove optimistic. We do not yet have enough data to be sure.] AIDS Heresies, Part 9 For any individual, use of a condom during heterosexual contact further greatly decreases the odds of HIV transmission, so much so that the odds of HIV transmission after condom failure with a carefully-chosen heterosexual partner, are probably on the order of the risk of dying on a routine freeway excursion, or perhaps even as low as fatality from a commercial passenger jet airline flight [footnote]. In short, conservative protests that use of condoms gives heterosexuals a "false" sense of security, are incorrect. For the American heterosexual, use of a condom and common sense probably results in at least as much security from accidental fatality as is the norm for participation in many other activities such as driving, swimming, skiing or boating. Certain homosexual behaviors, by contrast, are indeed considera- bly more dangerous, unless extraordinary precautions are taken. The estimated risk of transmission of HIV (at least for the familiar U.S. HIV "B" strain) from male to female during vaginal intercourse in the absence of other sexually transmitted disease is estimated to be about 1 in 250 to 500 per act of intercourse, and around half this for transmission from female to male (this data comes from married couples discordant for the virus). If a properly used condom fails 1% of the time, this figure would result in an infection risk of 1 in 50,000, even with an HIV-i- nfected partner. (These figures are consistent with studies of HIV-discordant married couples which show that consistent condom use even over several years time reduces HIV infection rate to essentially zero). Since far less than 1% of non-IV drug using women will be infected with HIV (prevalence is less than this, even for many classes of prostitute), total HIV-infection risk of heterosexual contact for an American man should be less than 1 in 5 million, which is the range for mortality-risk on an average commercial jet airline flight. Again, the "new sexual partner" risk is very much greater for women, who must exclude contact with bisexual men in order to bring risks into the same range of risk as for men. Concomitant sexually transmitted disease other than HIV also contributes to HIV transmission risk in a significant way. Studies of healthy Thai soldiers who acquire HIV heterosexually from prostitutes suggest that transmission risk for HIV in the presence of other sexually transmitted disease(s), may approach 1 in 10. In light of such facts the Surgeon General would of course remind us that limiting one's number of lifetime sexual partners is a time-honored (if not always followed) way of decreasing one's chance of acquiring sexually transmitted diseases, including HIV. Statistically, however, this strategy works only if one's partner(s) are doing the same. Requiring formal HIV testing before the beginning of any monogamous sexual relationship helps these statistics. For those demanding life by a shorter timetable, a rapid saliva HIV test is under development in Thailand, and may be on the U.S. market in a few years. For those people already infected with HIV, there is much hope as well. Very long-term HIV survivors do tend to be people who have avoided long-term early treatment with AZT alone [46], but it is hard to know if this is a cause or effect of good health. In any case, although it seems likely that AZT monotherapy for long periods is worth avoiding, the same does not necessarily follow for combinations of anti-retrovirals. New antiviral drugs are being developed at a rapid pace, and even a mutating virus may have difficulty keeping up with many different antiviral drugs administered at a time, or in combinations in a rotating fashion. For example, preliminary reports indicate that the new antiviral drug 3TC/lamivudine is particularly effective in combination with AZT at eliminating replicating HIV. The newly approved protease inhibitor drugs also appear to add to the effectiveness of nucleoside analogues such as AZT and 3TC, and have recently been shown to decrease viral loads drastically in AIDS patients, and improve survival over simple combinations of nucleosides. Other interesting combinations of antivirals, and even rotating combinations of combinations, are sure to come. Protease inhibitors are examples of the new drugs which are are continuously being tested against viral replication steps not affected by the AZT-like drugs. Other preliminary data suggest that an old anticancer drug called hydroxyurea might even be able to do the "impossible" and remove HIV viral genes inserted into the DNA of infected cells. For all these treatments and others yet to come, rapid assays of HIV viral reproduction, such as the RNA PCR test for HIV, can be used to rapidly pre-screen the effectiveness of new candidate treatments, without requiring the large and long trials which employ clinical endpoints such as morbidity (illness), or death. Other ways of controlling viral infection are also being tried by medicine, including methods of "booby-trapping" body cells by inserting a gene for a poison which is controlled by a virus gene, so that cells self-destruct before making any more new virus (the author has suggested inserting a poison gene in cells which is constructed so that it can be properly translated only after being copied by the viral reverse transcriptase enzyme). One or all of these techniques may buy much time, since the body can well afford to lose a few cells, if doing so stops viral replication. (Indeed, the word "cytotoxic," so vilified by skeptics in describing AZT, is an excellent description of the body's own chemical defenses against viruses, such as those provided by the sometimes cell-killing interferons). As long-term animal and human survival with retroviruses demonstrates, immune failure as a consequence of retroviral infection is a host-specific response to infection, and has little to do with any property of the virus itself. Almost certainly, AIDS is an evolutionary "mistake," caused by a particular (and perhaps unnecessary interaction of HIV virus and host [106]. We know that retroviruses survive well in many of their natural hosts without causing illness. We also know that in SIV infection of macaques, deletion of a single viral regulatory gene (called _nef_) can give rise to SIV viruses which replicate well enough to produce antibody, but produce no disease (just as they do not in African monkeys). It is thus very probable that HIV will be successfully dealt with eventually by finding a way to turn it into the kind of harmless infection it is in chimpanz- ees, long before a way is found of removing the virus wholesale from the body. In Australia, scientists have followed for some years a group of 5 people who are HIV-positive but healthy (save for one who died of AIDS after being treated with immunosuppress- ant drugs for lupus erythematosus), after all being infected by a blood donation from a single healthy HIV-positive donor who may harbor a more or less harmless strain of HIV [107]. There seems little doubt for many reasons that such a thing is possible, though perhaps rare. The particular HIV virus strain carried by these long-infected-but-well people no doubt has much to teach us. There is also much to learn from the 8% of HIV positive people harboring conventional strains of HIV, who are also still healthy after 14 years of infection, with no sign of becoming more immunosuppressed [46]. Do these people have less pathogenic strains of HIV, or are their own immune systems simply better? We know their blood load of virus is smaller than HIV-positive people who are progressively deteriorating in immune status. Recently identified is an interferon-like factor made by CD8+ cells in HIV "non-progressors" and other healthy people, which suppresses HIV infection in CD4+ cells without killing the cells. The factor is being intensively studied, and should it prove useful, may one day be commercial produced by genetic engineering techniques [108]. Of all examples of evolution we know of, viruses are the most spectacular, since they evolve so quickly that new species are seen on Earth regularly on the scale of a human life-time, and some of these newly "emerging" viruses emerge as plagues. Indeed, with HIV, there is good reason to think that the average human HIV host is infected by not one, but rather many closely related strains or subspecies of the HIV virus, all evolving furiously in parallel. During the asymptomatic phase of HIV infection, virus cultured from the host reproduces slowly in culture, but at the end of the disease when the virus is more active, cultured viruses have often turned into strains that are able to reproduce rapidly in culture. It is even possible to track the immune system as it attempts to produce antibody to a changing virus, and watch the virus as it attempts to escape by transforming itself. Usually, after a long time, the virus gains the upper hand in this battle with the immune system, but does so only because it infects the host slowly enough to give the virus time to evolve over a decade into something the host's defenses finally cannot handle. Even so, the virus can win only by slowly destroying the body's ability to deal with any invader, including itself. Possibly the attack on the immune system and the viral evolution are both needed to cause HIV's exceptionally high mortality rate after a long delay (at least 50% in 15 years, depending on host age). The fact that the HIV virus only prevails after it has destroyed the host immune system is itself a hopeful sign that the virus can be kept at bay for years, as long as an immune response can be maintained. Most interesting is the fact that in AIDS, cell-mediated immunity is not the only host immune defense against HIV which is destroyed by the virus at the end of the disease course: levels of antibodies against the core of the virus also decline during AIDS (perhaps due to loss of immune cells which stimulate antibody-making cells to produce them). Antibodies against the antibody envelope remain, but are often not neutralizing (do not stop the virus from infecting cells). If antibodies are not being made in AIDS, or the wrong antibodies are being made because the virus has mutated, it seems a natural treatment to try infusing anti-HIV antibodies from other AIDS patients who are in an earlier phase of the disease. Indeed, such experiments have recently been reported and confirmed: transfer of sterilized plasma, containing immune factors and anti-HIV antibodies, from early HIV-infected patients into AIDS patients, considerably cuts their risk of opportunistic infection, and raises their numbers of lymphocytes. Such primitive adoptive immune system therapy, using first antibodies targeted against HIV, and later soluble cell-growth factors and immune modulators (cytokines), seem likely to play an increas- ingly large role in the treatment of AIDS in the future. _The Future_. What will the future bring in the way of new diseases like AIDS? To the question of whether or not we will see more viruses newly evolving to be able to infect humans with new plagues, the answer is that no doubt we will. We know from the examples of viruses like Ebola Zaire that a virus can be as deadly as HIV without needing a latency period. But a short incubation and less than maximum infectivity was the undoing of Ebola, or otherwise modern world civilization might indeed have been destroyed by it [121]. The real question is: what kinds of viruses can manage greater spreading capability by having a long infectious latency period, yet be able to cause high mortalities as well? It may be that all viruses with as long a latency period as HIV must, in order to escape the active and tracking immune system over that time, either replicate in a place hidden from the immune system (such as the brain), or must be able to attack the immune system itself, in order be able to eventually emerge in enough force to cause host death. Perhaps only retroviruses can manage these feats. The only variable which may be possible to change is communicability, and one can imagine a "super-flu" which is as communicable as influenza or measles in the initial stages, but has a latency period followed by the long-term consequences of HIV, years down the line. At present we know of no retroviruses which are that infectious, but we don't know enough about biology yet to say whether or not this fact is due to necessity or luck. Certainly the fact that all known retrovi- ruses are membrane-coated and fragile-- easily destroyed by the environment outside the host-- has been a blessing. Let us hope that lymphotropic retroviruses are all forever somehow necessarily constrained in communicability, and that biotechnology progresses rapidly enough to find a way of suppressing our current crop of these microbes. Finally, let us also hope biotechnology does not progress in such a way as to let molecular bioengineering of such a virus become a possible basement hobby project (as it surely will one day) before we understand how to protect against the possibility of such a deliberately-caused, or deliberately-modified plague. AIDS Heresies, Part 10 APPENDIX: SPECIAL TOPICS 1. "What Are CD4+ and CD8+ Lymphocytes?" T-lymphocytes are lymphocytes which mature in the thymus (T is for "thymus") and which coordinate the immune response. They make up most (about 80%) of the lymphocytes in the blood in humans. The particular T-lymphocytes now known as "CD4+ lymphocytes" are known also as T-helper cells, because they help the immune system respond to invaders. CD4 (Cluster Designation number 4) is one of various "CD" designations referring to protein marker molecules found on the cell surfaces. CD4 is the cell surface molecule upon which the HIV virus "docks" and binds, before being taken inside by the cell (HIV can get into cells in other ways, though). The positive (+) sign for a given cell type (such as "CD4+") indicates that CD4 is present. T-Lymphocytes may be CD4+ or not. CD4+ lymphocytes make up about 2/3 of T- lymphocytes in the blood, with most of the remainder being T- lymphocytes classed as "CD8+ lymphocytes." The normal numerical ratio of CD4+ to CD8+ lymphocytes is about 2, but in stress can fall below 1, and in AIDS almost always falls below 0.3. Two HIV/AIDS Critics: Duesberg and Root-Bernstein Dr. Peter H. Duesberg is professor of molecular and cell biology at the University of California at Berkeley. A German emigree, he was originally trained in chemistry. On arriving in the U.S. in 1964 he began work in the field of viral molecular biology at Berkeley, where in 1970 he co-discovered the genetic basis for the carcinogenic action of the Rous sarcoma retrovirus. In 1987 he began publicly questioning the role of HIV in AIDS, a stand which has made him the center of the present HIV/AIDS controversy. Duesberg's most recent book is called _Why We Will Never Win the War on AIDS_ (1994), co-authored by a Berkeley graduate student and one-time protege Bryan J. Ellison. The book has been plagued by trouble. According to a message issued October 13, 1994 by the Group for the Scientific Reappraisal of the HIV/AIDS Hypothesis, this manuscript was published unilaterally by Ellison without Duesberg's consent, following failed editorial negotiati- ons with the original contracting publisher (St. Martin's Press). According to Duesberg, the editor had asked for additional documentation, clarification, and elimination of material which might be considered unfair to individuals. Duesberg was willing to cooperate but Ellison was not. Following Ellison's publicat- ion of the manuscript at Inside Story Communications (a political newsletter edited by Ellison), Duesberg severed relations with Ellison and is seeking an injunction against further publication of the book. In fact, Ellison and Duesberg are at this writing suing each other, with Ellison claiming injury from being forced to leave his Ph.D. track. The cooperation of James Tabulse, publisher of the Group's newsletter _Rethinking AIDS_, with Ellison, has meant that the Duesberg's Group has decided to severe relations with this publication as well. They now publish a new newsletter called _Reappraising AIDS_. Since Duesberg has questioned only publication and editorial rights for the new book and has not repudiated any of its contents, the book is also used in this review as an additional source of Duesberg's views. Robert Root-Bernstein is associate professor of physiology at Michigan State University (East Lansing). He held a MacArthur Prize Fellowship (the famed MacArthur "genius award") from 1981 to 1986, as the result of work in which he analyzed public policy and scientific creativity. He is the author of the recent book _Rethinking AIDS_ (MacMillan, 1993), and also author of numerous scientific articles dealing with the immunology of AIDS and other diseases. 2. What is an "Antibody" and What Does "HIV-Positive" Mean? Antibodies are blood proteins made by immune cells, which stick very specifically to microbial invaders, targeting them for destruction by the immune system. For many years after an infection by a microbe, antibodies specific to that microbe can be detected in the blood. A person who tests positive for antibodies to HIV by two different kinds of lab tests, is said to be "HIV-positive." In the case of infection with the average microbe, a person may test antibody positive for years or even a lifetime after the microbe is completely gone from the body. For the chronic viruses which hide in cell nuclei, however (retroviruses like HIV; and also CMV, EBV, and other herpes-class viruses), the presence of antibody is generally a clue to the continued presence of the virus, active or inactive, somewhere in the body. In some cases modern sensitive tests for viral DNA can actually detect these hidden viruses directly. Recently, development of sensitive tests for viral mRNA are able to test for continued viral activity, or "reactivation." The presence of viral mRNA indicates protein manufacture, which indicates virus reproduction. It is now known that chronic viruses often reactivate when the immune system is suppressed, but only sometimes cause illness as a results (well known examples of illness are cold sores and shingles). Tests for HIV RNA show HIV to be active in essentially all HIV-positive people, in blood cells and especially lymph nodes. This is true even if the person is asymptomatic. Continued reproduction of the HIV virus at all times after infection is thought to underlie the gradual immune decline seen over many years' time in 11 out of 12 HIV-positive people. The remaining 1 in 12 people with HIV show antibodies but no immune decline, indicating an immune system which is somehow able to keep replication to very low levels. This small subsegment of HIV-infected persons may be long-term survivors of HIV; at present with less than 15 years of tracking, it is still too early to tell. Over the years since the discovery of HIV, critics of the HIV/AIDS hypothesis have had to struggle to keep up with sensiti- vity increases in HIV testing. Initially, critics complained that HIV virus was not present in most HIV-positive people. When it became clear that infectious virus could be found in almost 100% of such people, if cultures were done correctly, critics claimed that most HIV was dormant until reactivated in culture. With new sensitive tests for HIV RNA showing that HIV virus is active in the body's lymph nodes, critics have fallen back to the position that it may be active, but not active _enough_. This is a question which can only be answered indirectly, by other studies. In the case of HIV, however, Ellison and Duesberg assert (p. 124), that HIV is rarely to be found budding from cells in patients, and that "...in most individuals with AIDS, no virus particles can be found anywhere in the body," implying that this absolves the virus from any disease role. Actually, however, even actively reproducing HIV may spend very little of its total life-cycle actually budding through a cell membrane or floating free as a particle in the blood before being picked up by another cell. Studies of viral RNA in the body show that there may be anywhere from roughly 10 million, to as much as 1 billion particles or actively replicating HIV genomes in a gram of lymph tissue-- a significant amount by the standards of most other kinds of virus [122]. Furthermore, a number of studies now show that amount of free HIV virus in the blood plasma, shown by more sensitive tests to be much larger than previously thought, correlate with rate of progress of HIV infection to AIDS [123]. 3. Hemophilia and Life Expectancy in the 80's. Hemophilia is a collection of genetic diseases which have in common an inability to make sufficient amounts of active blood clotting proteins. Many people with hemophilia require transfusions of active clotting factor proteins in order to avoid bleeding episodes. Just before the era of AIDS, such treatment had resulted in nearly normal life expectancies for people with hemophilia in the U.S., though the average age of the population was still low and had not yet recovered to U.S. normal. Approximately half of all people with hemophilia in the U.S. were infected with HIV in a silent plague beginning in 1979, peaking in 1982, and ending in 1985 when factor sterilization and testing ended new HIV infection. AIDS was first identified in people with hemophilia in 1982, with a sharp rise in AIDS deaths after 1984. Since the beginning of the AIDS plague, about 1/3 of HIV infected people with hemophilia have died of AIDS. This has resulted in the life expectancy of people with hemophilia falling sharply since 1984, during the same period that the average age for this population has been rising. These seemingly disparate trends represent the effects of the AIDS plague and advancing medical technology acting in opposing directions. The _average age_ of a population is an integrated effect of mortality risks over many previous decades: this is still increasing rapidly for hemophilia patients because of hemophilia treatment gains over much of this century, particularly in the 1970s. The _life expectancy_ for a population reflects the sum of current mortal- ity risks for people of all ages at a given time: this number has been decreasing continuously for hemophilia patients over the last decade because of increasing losses to AIDS. Duesberg and other HIV/AIDS critics have chosen to emphasize average age in their discussions, a very bad index of how risks are changing in a population on a year to year basis (as they are in the AIDS plague). Worse still, Ellison and Duesberg confuse life expectancy with average age (pp. 123, 188), further making things even more difficult for their unwary readers. _Hemophilia and Immune Suppression._ HIV/AIDS critics have recently emphasized the fact that hemophilia itself is immunosup- pressive, without emphasizing that this immunosuppression does not resemble that of AIDS, since it is a product mostly of higher CD8+ (suppressor T-cell) counts, rather than lower CD4+ (helper T-cell) counts [22]. Critics have also suggested that the much higher death rates in HIV-positive people with hemophilia are due to AZT use, despite direct survival statistics showing the reverse, and also the fact that this trend was well-established long before 1987, when AZT became available (see section on AZT). Although most hemophilia is found in men, approximately 3% of those with the disease are women, and female patients have a quite different spectrum of clotting factor use. Female hemophilia sufferers thus provide another piece of evidence that it is HIV, not clotting factor use, which causes AIDS, since male and female hemophilia patients have different factor usages and different AIDS risks, but the differences are entirely explained by HIV. This is easily seen in the fact that in people with hemophilia, average clotting factor use by type is different for HIV-positive males and females, yet AIDS risk over time is the same [109]. 4. AZT: Panacea-- Pandora Plague--- or Perfectly Pedestrian Pharmaceutical? [110] _The Drug and How it Works_ It is a greater task to find antimicrobials which interfere with virus reproduction than bacterial reproduction, simply because viruses use much of the cell's machinery to do their work, and thus it is difficult to find inhibitors of the viral reproductive process which are not poisons to cells as well. Viruses which use their own special enzymes, different from cell enzymes, may have Achilles heels. Retroviruses (including their subset, the lentiviruses) contain the unique enzyme "reverse transcriptase," which transcr- ibes the RNA instructions of these viruses into newly-made DNA. Since normal cells do not contain this enzyme (or at least not much of it), reverse transcriptase has been a natural target for antiviral drug developers. Eventually it was found that this particular enzyme is both the strength and weakness of lentiviruses when it came to AZT-like drugs. Advantage came in the fact that lentiviruses had a transcriptase enzyme which has no proof-reading capability, which a cell's reproductive enzyme systems have. The virus enzyme was thus prone to making mistakes in copying virus genes, giving retroviruses a high mutation rate which helped them evolve rapidly to adapt to new drugs, or to new hosts. The disadvantage to the virus was related: the viral enzyme would pick up and put into viral DNA altered or wrong building-block chemicals which the cell's enzymes were too fastidious to use for cellular DNA. Some of these synthetic chemicals, once inserted into a growing virus DNA chain, made the chain useless for further building, and so reproduction of viral genes was selectively stopped while cellular DNA could continue to be replicated. The first such chemical found to act this way was 5'-azidothymidine, or AZT. AZT had first been synthesized in the 1960's as a possible chemotherapeutic agent, but it had proved too toxic to normal cells at doses that interfered with cancer (at high doses, even the cell's enzymes could be fooled, particularly the DNA synthesizing enzymes in mitochondria). However, AZT proved less toxic at smaller doses which still inhibited HIV. The initial 1986 placebo-controlled trial of AZT in AIDS patients was so dramatic that the independent ethical monitors of the study (who had the codes for the otherwise double-blinded experiment) stopped the trial early so that AZT could be given to all the participants. The study was halted when 19 people had died in the group of AIDS patients receiving placebo, but only one patient had died in the equal-sized group receiving AZT [111]. A second study of AZT vs. placebo in Europe also showed that AZT delayed progression of HIV-positives to AIDS [112]. AZT was not benign, however. In the initial trial the 1500 milligram daily dose, 3 times what later became standard, was enough to cause some patients to require transfusions due to AZT's toxicity to their blood-forming bone marrow. The drug also caused side effects in the GI system and muscles. A number of placebo-controlled trials at lower doses showed that AZT did indeed delay the onset of AIDS and lengthen survival when given to immunosuppressed HIV infected men (both male homosexuals and people with hemophilia [113]), although the earlier the drug was given in the course of the HIV disease, the more the toxicity of the drug counterbalanced any symptoms of immunodeficiency that were delayed [114]. In the end, AZT alone turned out to represent a few extra months of life to the average AIDS patient, nothing more. Nor did (in smaller studies) early use of the drug before AIDS appeared seem to offer clear survival benefits over waiting until later stages of HIV disease [115]. Eventually it became clear that the reason for this was that the fast-mutating HIV virus was capable of evolving enough under the pressure of toxicity to escape the effect of AZT after approximately one to two years of treatment [116]. Because of this, many physicians began to suspect that this relatively short time of effective viral suppression was probably best left until late in the course of the disease, when the virus was causing the most damage, since AZT itself could damage quality of life in HIV-infected people who still had good immune function and were feeling well [117]. This suspicion was confirmed by the Concorde Trial, the largest trial of AZT ever done and one which had more patients and statistical power than all the other trials combined. When Concorde examined the benefits of AZT started quite early in the course of HIV infection, as opposed to waiting for AIDS or severe immune problems before starting AZT, there was a disturbing trend toward greater mortality on early AZT than deferred AZT. Some of the excess AZT group deaths, however, came from auto accidents, and it wasn't clear if these should be counted; AZT is accused of modulating immune problems, after all, not causing traffic accidents. However, with the study showing an 85-87% chance that the drug had done more harm than good when started earlier vs. later in HIV disease, it was finally apparent to most physicians that AZT was at best a short term drug, and a drug to be used later rather than earlier in HIV disease, at least when used alone [118]. Critics had thus scored a point in suggesting long ago that immense political and economic pressures had introduced AZT too early into clinically well people, a group for whom overall benefit was never well documented, and who in fact (as now appears) were being very expensively treated without being helped. The Concorde trial, however, proved also a blow to the severest critics of AZT, since it made clear once and for all that AZT at the proper doses was not a very toxic drug, and certainly was not a major AIDS-producing drug. In particular, Concorde proved that AZT could NOT be used to explain the large numbers of extra AIDS cases in HIV positive groups which are routinely seen in studies. This work of improving AZT-like drugs continues. Along with better and more specific treatment of the fungal and viral opportunistic diseases which AIDS patients develop, especially the previously often fatal pneumocystis pneumonia, the new anti-viral drugs have apparently slowly increased the time from the diagnosis of AIDS to death. Starting in 1987, the year AZT was introduced and began wide distribution, the average survival time of AIDS patients began to steadily increase. It continues to increase today: the average survival time after first diagno- sis of full-blown AIDS (the last stage of HIV infection) is now 18 to 24 months. Unfortunately, AIDS is still, by and large, a fatal condition. _The Controversy_. Despite the results of the Concorde trial which showed no significant evidence of any AIDS-producing effect of AZT, the effects of AZT are routinely discussed in heretical literature as though they might somehow be mistaken for the immunologic losses normally seen in the AIDS syndrome. But AZT does not cause selective loss of CD4+ lymphocytes with severely skewed CD4+/CD8+ ratios, nor does AZT at standard doses cause the kind of immunosuppression which can in any way be compared with that in AIDS [119]. This does not slow critics, who simply lump all kinds of immune deficiency together. Duesberg writes of one study: "Although the AZT-specific diseases were not diagnosed as AIDS, neutropenia generates immunodeficiency, and thus AIDS." (p. 242). Here Duesberg generates yet one more definition of AIDS of his own. Physicians, however, know that neutropenia, a deficiency of different kind of blood cell than the lymphocyte, predisposes to very different kinds of infectious problems than classic AIDS does, so again Duesberg's proposed definition is hardly clinically useful. Despite the complete lack of any findings showing that AZT causes selective CD4+ lymphocyte loss, or that AZT treated groups have significantly more deaths or immunological deaths than placebo or non-AZT groups (something that one would think might be an important point), skeptics led by Peter Duesberg have continued to characterize AZT as "AIDS by prescription" (p. 252) [120]. Thus, in heretical AIDS literature, AZT has become the scapegoat which "explains" modern studies which find a great deal of AIDS in HIV-positive people who do not use drugs. Duesberg blames AZT, for instance, for much of the increase in AIDS in HIV positive people with hemophilia, despite the fact that statistics show that HIV-positive hemophilia patients who take AZT are less likely to contract AIDS, not more likely [113]. Duesberg has also publicly blamed AZT and other AIDS drugs for AIDS-diseases in non-illicit-drug-using, non-hemophiliacs who have died of AIDS after contracting (supposedly harmless) HIV. He has, for instance, blamed AIDS medications for the death of clean-living transfusion-HIV-infected tennis star Arthur Ashe, despite the fact that Ashe's rare brain toxoplasmosis (thought initially to be a stroke) and his low T-cell count, proved the destruction of his immune system had taken place long before he was diagnosed with HIV, or took any drugs to treat it. Most famously, Duesberg has blamed AZT for the AIDS problems and death of Kimberly Bergalis, the non-drug using Florida college student who contracted HIV either sexually from a boyfriend, or else from her HIV-infected dentist during a dental procedure in December, 1987. According to Duesberg (in a favorite story which he has repeated in many interviews, and has included in his book), Bergalis began only with a non-pathologic oral yeast infection (one of Duesberg's medical eccentricities is his apparent belief that oral thrush is "AIDS-defining" in men but non-pathologic in women). In Bergalis, thrush was followed by what Duesberg describes only as "a brief pneumonia," after which Bergalis was inexplicably tested for HIV by her doctors "out of the blue," and when found positive, finally poisoned to death by her physicians who treated her with AZT and other AIDS drugs (Ellison and Duesberg, p. 217). According to a recent book-length account of the incident by Runnells, however [4], the complete story is a quite different than Duesberg's macabre account of a basically healthy woman being killed by her doctors. Bergalis had months of severe fatigue, wasting, chronic yeast infections, and a CD4+ lymphocyte count of only 41 just 5 days before entering the hospital, in December, 1989, where she was diagnosed by broncho-alveolar lavage with bilateral _Pneumocystis_ pneumonia, an AIDS-defining disease (the reader will remember that a CD4+ count of 41 signals nearly complete immune failure, and defines AIDS also by the more strict criteria proposed in this review). Only at this point (and hardly out of the blue) was Bergalis tested for HIV, on suspicion of AIDS, and only after testing HIV-positive, was she treated with AZT. But her immune system was nearly destroyed already. Bergalis ultimately survived another 24 months after her _Pneumocystis_, comfortably longer than similar patients in the era before AZT (average survival less than a year). Arthur Ashe lived more than 3 years from his first opportunistic infection and diagnosis of AIDS. There is no reason whatsoever to believe that AZT drugs shortened the life of either Bergalis or Ashe. Nor is the experience of these two AIDS celebrities unique, for there are quite enough studies of both average incubation time and survival time for AIDS in the era before AZT (eg, [113], [102] to be sure that (whatever else AZT does) it does not hasten either the onset of AIDS, or death from AIDS. What about studies in which people taking AZT have fewer AIDS diseases than people on placebo? For Duesberg this is only reason to suspect that placebo patients are using more illicit drugs-- even when the AZT and placebo groups have been matched by collecting them from a randomizing process. The chance of getting more illicit drug-users into one randomized group versus another, when the groups are picked by the equivalent of coin toss, is obviously low. This does not prevent Duesberg (p. 242) from grasping at a statistical straw and suggesting that this might somehow have happened in each of two separate studies, even though the chance of _this_ is "low" squared. 5. Montagnier, Gallo, Slip-Ups and Wrong Paths In the first weeks after the first AIDS cases were reported in 1981, the brilliant Don Francis of the CDC (played by actor Matthew Modine in the TV-movie version of Shilts' _And the Band Played On_), a scientist who had spent years working with the FeLV virus, hypothesized that AIDS was "feline leukemia in people." [ref. 14, p. 73] This was an inspired guess, but as it turned out, not quite correct (FIV, the actual cat analogue of HIV, would be discovered only in 1986). Sometime later Bob Gallo of the NIH (Alan Alda) would at Francis' suggestion make essentially the same assumption, thinking that AIDS was caused by a human leukemia retrovirus in the virus family that he himself had discovered. The French researcher Luc Montagnier isolated the correct virus in early 1983 from a French patient who later died of AIDS. However it wasn't until Gallo's lab cultures were contaminated with one of Montagnier's generously provided virus strains, which was then mistakenly identified as coming from an American AIDS patient, that the Americans were ready to also announce in 1984 that this virus (Montagnier's virus) was the AIDS virus [67]. Incredibly, the French-provided virus was also "re-isolated" after a similar lab contamination in England, as well. These fortuitous lab incidents, which tended to rob the French of sole credit and sole patent rights, were later proven by genetic fingerprinting of viruses; but government hearings failed to convict Gallo or anyone else of any misconduct (helping the American and English cases for accident was the fact that even Montagnier's own lab had been accidentally contaminated by the same fast-growing strain, which was not his original first isolate). Meanwhile, however, the American team led by Gallo had discovered a way to grow the virus in quantities which permitted an economical screening test to be produced from it, and had shared formal credit in 1987 with the French for the discovery of what became known as HIV, and by international agreement shared royalties for the HIV antibody test as well. The Americans and Gallo have since made far more money at this than the French, and thus, particularly in the eyes of HIV/AIDS skeptics who see a capitalist conspiracy in AIDS, have emerged as the clear villains of the tale. The personable Montagnier, despite his prime role in discovering HIV, has been granted partial absolution even by HIV/AIDS skeptics, due to his more recent speculations that HIV might not be able to cause AIDS without help from other microbes. 6. How Some Viruses Hide From Your Immune System The actual blueprints used by both viruses and cells to reproduce themselves, are long tape-like molecules called DNA and RNA. Either DNA or RNA is capable of carrying genetic information, and this information is coded by the linear sentence-like structure of the molecule. There exist viruses which use DNA, and others that use RNA, to carry their own instructions for making themselves. DNA and RNA as chemicals are a bit like recording media: biologic information can be copied from DNA to RNA in all living systems, and sometimes (in retrovi- ruses) even from RNA back to DNA, as we shall see. It is the genetic information, not the medium that holds it (DNA or RNA), which is important. Animal and plant cells use only DNA to archive their information. Some DNA-containing viruses are able to place their own DNA into close association with the cellular DNA located in the nucleus of the host cell, so that viral and cell DNA are regulated together. This is a survival tactic: virus DNA sometimes associates with cell DNA before halting new virus production and going into "hiding." The normal immune system of an animal can seek out and destroy actively virus-infected cells, and infected cells themselves also make antiviral substances called interferons, which destroy active viruses within other cells. The very effective strategy of associating non-active viral DNA with cellular DNA, however, sidesteps both these defenses, and results in infected cells which harbor virus DNA, but which are apparently very difficult for either the outside immune system, or the cell's internal defense systems, to detect. The virus which produces "cold sores" on the mouth and lips (a type of herpes simplex virus) is a familiar example of a virus which uses this trick, and which the immune system, despite all efforts, sometimes can never get rid of. Many people who are afflicted with a cold sore propensity are afflicted for life because there is no known way to remove herpes virus DNA chromosomes from association with the DNA of nerve cells, with which it is sequestered. Each outbreak of cold sores represents an event in which "sleeping" herpes DNA, hiding in nerve cells near the base of the brain, re-activates and directs the cell's biochemical machinery to now produce more virus DNA. The newly- made virus DNA then uses the nerve cell's inner transportation machinery like a subway train to move down through a long branch of the cell to the mouth, where it breaks free of the cell to infect and destroy skin cells to form a blister. Such outbreaks may occur sporadically for life. Each time the normal immune system stops an outbreak, the virus simply goes back into hiding inside the nerve cells. Certain viruses called retroviruses hide in cells in an even more intimate fashion, by making DNA genes and then inserting them directly into the cell's DNA. By directly controlling and influencing cell DNA activity in this manner, retroviruses may cause cancer or cell death after long latency periods. |
|