By Robert Root-Bernstein

Rethinking AIDS

The first definition of AIDS appeared in the September 24, 1982, issue of Morbidity and Mortality Weekly Report published by the Centers for Disease Control:

CDC defines a case of AIDS as a disease, at least moderately predictive of a defect in cell­mediated immunity, occurring in a person with no known cause for diminished resistance to that disease. Such diseases include KS [Kaposi's sarcoma], PCP [Pneumocystis carinii pneumonia], and serious OOI [other opportunistic infections]. These infections include pneumonia, meningitis, or encephalitis due to one or more of the following: aspergillosis, candidiasis, cryptococcosis, cytomegalovirus, norcardiosis, strongyloidosis, toxoplasmosis, zygomycosis, or atypical mycobacteriosis (species other than tuberculosis or lepra); esophagitis due to candidiasis, cytomegalovirus, or herpes simplex virus; progressive multifocal leukoencephalopathy, chronic enterocolitis (more than 4 weeks) due to cryptosporidiosis; or unusually extensive mucocutaneous herpes simplex of more than 5 weeks duration. Diagnoses are considered to fit the case definition only if based on sufficiently reliable methods (generally histology or culture). However, this case definition may not include the full spectrum of AIDS manifestations, which may range from absence of symptoms (despite laboratory evidence of immune deficiency) to non­specific symptoms (e.g. , fever, weight loss, generalized, persistent lymphadenopathy) to specific diseases that are insufficiently predictive of cellular immunodeficiency to be included in incidence monitoring (e.g., tuberculosis, oral candidiasis, herpes zoster) to malignant neoplasms that cause, as well as result from, immunodeficiency.(1)

It is evident from this definition that the CDC was not sure what AIDS was in 1982, other than that it appeared to be due to immune deficiencies of unknown cause that could be manifested by any of fourteen different opportunistic diseases. Crucial to the definition, however, was the statement that diagnosis for AIDS could be made only in people with these opportunistic diseases if they had "no known cause for diminished resistance to that disease." The reason for this caveat was that a number of groups of people had previously been identified as having a significant risk for each of these opportunistic diseases. Patients undergoing various cancer chemotherapies, transplant patients, people treated with high or chronic doses of corticosteroids to control inflammatory and autoimmune diseases, and people born with defective immune systems are prone to opportunistic infections of all kinds. They were excluded as AIDS patients by definition, as were men over the age of sixty who developed Kaposi's sarcoma, since such men were already known to be at risk for this cancer. A diagnosis of AIDS required no identified cause of immune suppression.

The definition of AIDS has evolved along with the disease itself. Just how much it has evolved can be seen from the following example. In May 1991 a new and unexpected AIDS risk was broadcast to the world. "Organ recipients test positive for AIDS virus!" screamed headlines. LifeNet Transplant Services of Virginia Beach, Florida, announced that three­people who had received organ transplants-one the heart and two others a kidney apiece- from a man who had died of gunshot wounds in 1985 had developed AIDS and died. Three other recipients of the man's tissues also tested positive for HIV. The frightening aspect of the cases was that the gunshot victim had been tested twice for HIV prior to the transplants and had been found to be HIV free. Subsequent reanalysis suggested that the tests used during 1985 did not have the sensitivity necessary to identify the man's very low level of infection. On the other hand, it is equally possible that the patients had latent HIV infections that were activated by the transplant procedure. In either case, the cases raised the spectre, validated by similar instances, (2) of hidden HIV infections unwittingly being transmitted to or reactivated in a significant number of transplant and blood transfusion recipients. The story, coming as it did at the same time that a number of states were considering banning HIV­infected surgeons and dentists from performing surgery, added fuel to the hysteria that perhaps there is not, and never can be, any real protection against AIDS. Even the most scrupulous and clean­living individuals might, by chance and through no fault of their own, still contract this modern scourge through an improperly screened blood transfusion or an unwanted visit to the hospital.

No one seems to have realized that just seven years earlier, the same three organ transplant recipients could have died of exactly the same opportunistic infections without raising an eyebrow and without being diagnosed as having AIDS. They would have been in a group specifically excluded from being considered for a diagnosis of AIDS: transplant recipients. Their causes of immune suppression were known: the drugs they were treated with in order to prevent their immune systems from rejecting their new organs. These drugs, along with the rigors of surgery itself and the possibility of an immune system disorder called graft­versus­host disease in which the lymphocytes in the donated organ attempt to kill the recipient's body, result in very high rates of morbidity and mortality in organ recipients compared with the general populace or even with other surgery patients. Morbidity is the physician's term for sickness; mortality for death. Two of the transplant patients who died of "AIDS" received kidneys. Their probability of dying within three years of their operation was 20 percent if they developed no complications and 40 percent if they did. This figure rises to nearly 60 percent at five years for patients with complications.(3) Since the two patients who died clearly developed complications manifested as opportunistic infections, they were in the high­risk group. Thus, from a purely statistical point of view, each of these people was more likely to have died than to have been alive in 1991, no matter what their HIV status. The same approximate statistics apply to the unfortunate individual who received a heart transplant.

Chances are also good that the three would have died of the same symptoms and the same opportunistic infections whether they had contracted an HIV infection or not. HIV­negative transplant patients are prone to the same sets of opportunistic infections that characterize AIDS patients, including Pneumocystis pneumonia (originally known as "transplant lung"), cytomegalovirus, varicella­zoster virus, disseminated herpes simplex, and toxoplasmosis infections. (4) The only difference between the transplant patients who died of AIDS and those who die of the same symptoms but are not given a diagnosis of AIDS is the presence of antibody to HIV in the former group.

What, then, is AIDS? Why do we call a patient who dies of Pneumocystis pneumonia following a transplant operation unfortunate but one who dies of Pneumocystis pneumonia and HIV an AIDS tragedy? Ironically, this definitional problem has existed since the very beginning of the "epidemic." In the first report of GRID published by Michael S. Gottlieb and his colleagues at UCLA, one of the five patients was a twenty­nine­year­old male homosexual who had a known cause of immune suppression. He had been successfully treated with radiation therapy for Hodgkin's disease (a cancer of the white blood cells) three years earlier. (5) Radiation therapy is a well­recognized cause of immune impairment. Nonetheless' this case stands as one of the benchmark cases heralding the discovery of AIDS.

Beginning in 1984, the definition of AIDS was changed to make the Hodgkin's case less anomalous and eventually to include transplant patients and other immunosuppressed individuals under certain circumstances. The CDC revised its definition by adding to the list of diseases diagnostic for AIDS any lymphoma (cancer of the lymph system) limited to the brain. (6) The discovery of HIV and its identification as "the cause of AIDS" during 1984 caused a second revision in June 1985. (7) To the previous set of fourteen diseases predictive of cellular immune suppression, the CDC added seven more diseases. If a person was found to be HIV seropositive by any test and had histoplasmosis (a fungus) disseminated beyond the lungs or lymph nodes; isosporiasis (a protozoal infection) causing chronic diarrhea for more than a month; bronchial or pulmonary candidiasis; many types of non­Hodgkin's lymphomas; Kaposi's sarcoma over the age of sixty; chronic lymphoid interstitial pneumonitis if a child; or any cancer of the lymph system diagnosed three or more months after a diagnosis of any opportunistic infection, then he or she was an AIDS patient. Thus, a number of groups that had previously been excluded from diagnoses of AIDS, such as certain cancer patients and elderly men with Kaposi's sarcoma, were suddenly potential AIDS patients despite previously demonstrated risks for opportunistic diseases. The crucial question was whether they had become infected with HIV as well.

Even more important in the light of recent questions concerning the necessity of HIV for causing AIDS, the 1985 revision of the AIDS definition also stated that some opportunistic diseases previously diagnostic for AIDS would be diagnostic in the future only if HIV was present: "To increase the specificity of the case definition, patients will be excluded as AIDS cases if they have a negative result on testing for serum antibody to [HIV], have no other type of [HIV] test with a positive result, and do not have a low number of T­helper lymohocytes or a low ratio of T­helper to T­suppressor lymphocytes." (8) In other words, people with the same clinical symptoms as an HIV­infected person (for example, disseminated tuberculosis) but without evidence of HIV or obvious immune impairment were not AIDS patients. This alteration causes problems. Twelve of fourteen cases of Kaposi's sarcoma diagnosed in individuals without identified risk factors for AIDS during 1981 and 1982 had normal immunologic results and were not tested for HIV (since HIV had not yet been discovered). (9) According to the 1985 definition, they might not have been diagnosed as AIDS patients. Even more interesting are the more than twenty HIV­negative cases of Kaposi's sarcoma among homosexual men with normal immunologic results that have been reported in the medical literature during the last two years. Do these people have AIDS? If not, is there a second epidemic of Karposi's sarcoma (and perhaps other opportunistic diseases) superimposed upon the so­called AIDS epidemic and appearing in the same risk group? How are these two diseases, if they are two, to be distinguished? What do they tell us about the necessity of HIV in AIDS?

These issues become more confused in the light of the next set of alterations announced by the CDC in August 1987. According to this set of revisions, the list of opportunistic infections indicative of AIDS grew to twenty­four, again enlarging the pool of potential AIDS patients. One set of twelve opportunistic diseases, including Pneumocystis pneumonia, Kaposi's sarcoma, disseminated cytomegalovirus infection, and esophageal candidiasis, were diagnostic for AIDS regardless of whether there was any evidence of HIV infection. Twelve other diseases were diagnostic for AIDS only in conjunction with a positive HIV antibody test. This meant that a large number of AIDS patients (45 percent of all cases diagnosed in the United States during the past decade and 1 percent of patients specifically tested for HIV seropositivity continued to be diagnosed as having AIDS in the absence of evidence of HIV infection. (10) By far the most important of the changes made in 1987 was the statement that "regardless of the presence of other causes of immunodeficiency, in the presence of laboratory evidence for HIV, any disease listed . . . indicates a diagnosis of AIDS." (11) In other words, acquired immune deficiency syndrome attributed to HIV infection is now diagnosed even among people who were born with congenital immune deficiencies; who have demonstrable, preexisting, or coexisting causes of immune suppression due to chemotherapy, radiation treatment, or corticosteroid use; among transplant patients who are on regimens of immunosuppressive drugs for life; and so forth.

AIDS, in short, has become a schizophrenic disease. Some people with diseases identical to those classically used to define the syndrome, such as disseminated tuberculosis, are not AIDS patients in the absence of HIV. Some people are AIDS patients if they develop opportunistic infections even in the absence of evidence of HIV. And in the presence of HIV, almost any rare disease is diagnostic for AIDS regardless of whether the person has other, more fundamental causes of immune suppression.

The definition changes are apparently not over. In 1992, the CDC proposed altering the definition of AIDS to include any person who had developed a significant loss of a particular type of white blood cell called T­helper lymphocytes. (12) Normally, a healthy person has a T­helper lymphocyte count of around 1,000 cells per cubic millimeter of blood. AIDS may now be diagnosed when the number of these T­helper cells falls below 200 per cubic millimeter of blood if the individual is HIV seropositive and even if he or she has no opportunistic infections. In other words, the primary criterion that allowed the identification of AIDS in the first place-that a person have an opportunistic disease in the absence of an identified cause of immune suppression-may be abandoned completely. People may be diagnosed as having AIDS even if they have no infections typical of AIDS, as long as they have a significantly low number of T­helper cells and antibody to HIV.

This latest proposed definition change has little, if any, scientific merit. Indeed, the CDC itself has been fighting against the definition change, and Dr. James O. Mason, assistant secretary for health in the Department of Health and Human Services, says forthrightly that changing the definition "messes up the baseline for comparison from past to future" and that it "will make interpretation of trends in incidence and characteristics of cases more difficult." (13) Then why alter the definition?

The reason for this latest definitional alteration is social and economic, not scientific. AIDS activists are now dictating how AIDS is to be diagnosed and who is to be included in the count. (14) For them, the issue is not one of correct diagnosis or elucidating the cause of AIDS; it is the understandable desire to increase access to health care. As Erik Eckholm has written in the New York Times, "The definition [of AIDS] has become a political as well as a medical question as people infected with the human immune deficiency virus, HIV, compete for treatment. For years, people weren't considered to have AIDS until they showed symptoms of certain infections and cancers that invade the body once the immune system breaks down. But after complaints that many ailing people were being excluded from the count, the Federal Centers for Disease Control has begun revising its definition. . . . It has been estimated that the broader definition . . . will add 160,000 people to the current caseload of 200,000 classified as having AIDS." (15) In other words, the number of AIDS cases may double with one fell swoop, not because AIDS has suddenly spread to new risk groups or even because it has spread within acknowledged risk groups but by definitional fiat.

It is worth putting these developments in historical perspective. Mirko Grmek, a French physician and historian of medicine, notes in his History of AIDS that AIDS "is not a disease in the old sense of the word, inasmuch as the virus is immunopathogenic, that it affects the immune system and produces symptoms only through the expedient of opportunistic infection or malignancy... Its pathological manifestations could not even have been understood as a disease before the advent of new concepts resulting from recent developments in the life sciences. In the past, a disease was defined either by clinical symptoms or by pathological lesions, which are morphological changes in organs, tissues, or cells. Nothing of the sort, neither clinical symptoms nor lesions, observable by the old means, characterizes AIDS. It is not a disease in the sense given to the term before the mid­twentieth century. Persons affected by HIV virus suffer and die with the signs and lesions that are typical of other diseases. As recently as twenty years ago, these opportunistic disorders were the only reality that physicians could observe and conceptualize." (16) In other words, AIDS is new not only in the sense that it was only recently recognized; AIDS is also new in the way that biomedical researchers have defined it. These are important points to remember when we try to determine what AIDS is, what causes it, and whether its causes are in fact new. After all, if the biomedical tools and concepts did not, as Grmek asserts, exist twenty years ago for recognizing AIDS, how could it have been observed even if it had existed?

The schizophrenic and metamorphic nature of the definition of AIDS are of considerable importance in evaluating the possible cause or causes of the syndrome. Consider an analogy. A man drowns. The pathologist finds that he has much too much carbon dioxide in his blood. From a purely factual standpoint, we know that too great a percentage of carbon dioxide in the air one breathes can be fatal. This is the point of the rebreathers that divers sometimes use; they absorb the carbon dioxide from the air supply, allowing prolonged reuse of the air. We also know that when people drown, the level of carbon dioxide in their blood increases dramatically since their cells continue to respire even when their lungs cease to exhale. Yet it does gross injustice to logic to maintain that the level of carbon dioxide in a drowned man's blood is his cause of death. One must take a step back and ask why the man's carbon dioxide level became so high; that reason, quite clearly, is that he could not breath; he could neither exhale nor inhale. Thus, the high level of carbon dioxide in his blood is what is known to pathologists and philosophers of science alike as an epiphenomenon-a secondary or additional symptom or complication arising during the course of a malady, treatment, or experiment. Clearly the drowned man had many problems besides this buildup of carbon dioxide. For instance, he also ran out of available oxygen, a problem at least as severe as the increase in carbon dioxide levels that he experienced. Yet neither the buildup of carbon dioxide nor the lack of oxygen is, in a purely logical sense, the primary "cause" of death. Indeed, there is no single cause of drowning, no matter how similar the outcome. At the most fundamental level, the man drowned because he could not swim, because he got a cramp that incapacitated him, because he had a heart attack, because he struck his head on something and passed out, because someone held his head under the water until he was unconscious, or any number of other reasons. In short, the existence of high levels of carbon dioxide in the man's blood is factually correct, it is a finding invariably present in drowning victims extremely rare in other people, but it is most definitely not the primary cause of death.

The drowned­man analogy is highly relevant to understanding AIDS. We must be absolutely certain that HIV is not an epiphenomenon of AIDS before we assert that it is the primary cause. The fact that it is an extremely frequent finding in AIDS patients is not logically compelling. It is only suggestive. Other active infections, such as cytomegalovirus, are also nearly universal among AIDS patients. If both are correlated with AIDS, which is the cause? Or are both viruses reactivated by previous and perhaps more diverse causes of immune suppression? How do we know what is cause and what is effect?

The existence of the full range of AIDS symptoms and opportunistic infections in both HIV­free and HIV­infected transplant and cancer patients warns us that this logical caveat is one that must be acknowledged in AIDS. HIV infection may be an epiphenomenon of immune suppression rather than a necessary cause. Immune suppression may predispose people to HIV infection (just as it predisposes them to other opportunistic infections) rather than resulting from such an infection. I argue in my book Rethinking AIDS, in fact, that HIV may be just such an epiphenomenon. Every AIDS patient has multiple causes of immune suppression other than HIV, many of which precede HIV infection and some of which occur in the total absence of HIV. The existence of these largely unrecognized immunosuppressive agents in AIDS not only requires a rethinking of the definition of the syndrome as occurring mainly in people without previously identified causes of immune suppression but also necessitates a critical look at the role of HIV as a causative agent in AIDS.

Before turning to the adequacy of the arguments supporting HIV as the sole, necessary cause of AIDS, two final comments are necessary concerning the definition of AIDS. The effects of the definition changes go far beyond mere questions of who has AIDS or how it is to be diagnosed. Much of our public health policy rests upon calculations of how fast AIDS is growing and into what groups it seems to be spreading. Each time the definition of AIDS changes, all of these calculations change as well. Previously exe eluded people suddenly qualify as AIDS patients. Diagnoses skyrocket. The 1985 definition change resulted in about a 4 percent increase in the number of diagnoses, a small enough fraction that translates into 2,000 additional cases a year in the U.S. The 1987 revision resulted in about a 30 percent increase in diagnoses, or some 10,000 cases in 1988 and some 15,000 additional cases during 1991. The proposed 1992 definition may double the the number of diagnoses overnight. In consequence, a significant proportion of the continued explosive growth of AIDS throughout the past decade has been fueled not by the transmission of AIDS to new groups of people but rather by the inclusion of previously excluded groups of people into the category of AIDS. People fitting these revised definitions of AIDS had always existed, but they were not counted as AIDS cases. Indeed, prior to 1981, they were not even recognized. Thus, despite claims that AIDS is the worst plague since the Black Death of the Middle Ages, despite the fact that AIDS is now the tenth most common cause of death in the United States, and despite the fact that there are no new miracle cures for the most common causes of death-heart disease, cancers, diabetes, stroke, and accidents-life expectancy for people in the U.S. has increased every year since 1980 at an almost constant rate. (17) One could justifiably argue that the AIDS epidemic is due at least partially to the grouping of two dozen causes of death under one rubric rather than to a new disease.

Finally, it is imperative that one gaping lacuna in the AIDS definition be pointed out: There are no criteria listed in any definition of AIDS that allow for a person to fight off AIDS or to be cured of it. Once a person is diagnosed, he or she will have AIDS forever after, regardless of any improvement in state of health and regardless of whether death results from a non­AIDS associated disease (for example, heart disease or diabetes). This is another way in which the definition of AIDS is a medical novelty. A person has pneumonia as long as he or she is symptomatic and the germ causing the disease is present. Destroy the germ and eradicate the clinical symptoms, and the person is cured, regardless of the fact that both antibody to the germ and scarring of the lungs may persist for their lifetime. Even in slowly progressing diseases such as cancer or heart disease, five­year survival is often taken as tantamount to a cure if disease symptoms are essentially absent. No such criteria exist for AIDS, despite the fact that some AIDS patients are still alive a dozen years after diagnosis with Kaposi's sarcoma, Pneumocystis pneumonia, and other opportunistic diseases. As AIDS survivor Michael Callen writes in his inspirational book, Surviving AIDS, (18) long­term AIDS survival does occur, but no one, once diagnosed definitively with AIDS, has ever been taken off the lists kept by the CDC except at death. This makes AIDS the first disease that no one can survive, by definition. Not only is this description of AIDS logically bankrupt, it sends the demoralizing and inaccurate message to people with HIV or AIDS that they have a disease that is not worth fighting. A more legitimate, and more hopeful, definition must be devised. *

Robert S. Root-Bernstein, an associate professor of physiology at Michigan State University, East Lansing, is the author of Rethinking AIDS: The Tragic Cost of Premature Consensus (New York, Free Press, 1993) and Diversity (Cambridge, Mass., Harvard University Press, 1989). He is a former MacArthur Fellow (1981-1986).


1. Centers for Disease Control. 1982. Update on acquired immune deficiency syndrome (AIDS)-United States. MMWR 31 (37):507­508 .

2. Associated Press. 1987. Patients infected with AIDS after kidney transplants. Lansing State J. 29 Sep. 3A; DummerJS, Erb S. Breinig MK, et al. 1989. Infection with human immunodeficiency virus in the Pittsburgh transplant population. A study of 583 donors and 1043 recipients, 1981­1986. Transplantation 47:134­140.

3. Kjellstrand CM, Hylander B. Collins AC. 1990. Mortality on dialysis-on the influence of early start, patient characteristic, and transplantation and acceptance rates. Am J Kidney Dis 15:483490; Held PJ, Brunner F. Odaka M, Garcia JR, Port FK, Gaylin DS. 1990. Five­year survival for end­stage renal disease patients in the United States, Europe, and Japan, 1982­1987. Am J Kidney Dis 15:451­457.

4. Salt A, Sutehall G. Sargaison M, et al. 1990. Viral and toxoplasma gondii infections in children after liver transplantation. J Clin Patholol 43:63­67; Singh N. Dummer JS, Kusne S. et al. 1988. Infections with cytomegalovirus and other herpes viruses in 121 liver transplant recipients: Transmission by donated organ and the effeet of OKT3 antibodies. J Infect Dis 158:124­131.

5. Centers for Disease Control. 1981. Pneumocystis pneumonia- Los Angeles. MMWR 30 (21):250­252.

6. Selik RM, Haverkos HW, Curran JW. 1984. Acquired immune deficiency syndrome (AIDS) trends in the United States, 19781982. Am J Med76:493­500.

7. Centers for Disease Control. 1985. Revision of the case definition of acquired immunodeficiency syndrome for national reporting- United States. MMWR 34 (25):373­375.

8. Centers for Disease Control. 1985. MMWR 34(25):373­375.

9.Selik et al. 1984. Am J Med 76:493­500.

10. Centers for Disease Control. 1989. Update: Acquired immunodeficiency syndrome-United States, 1981­1988. MMWR 38:229236.

11. Centers for Disease Control. 1987. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. MMWR 36 (Suppl lS):3S­lSS; Imrey HH. 1988. AIDS case clefinition. Science 240:1263.

12. Centers for Disease Control. 1991. Extension of public comment period for revision of HIV infection classification system and exe pension of AIDS surveillance case definition. MMWR 40:891.

13. Cimons M. 1991. Federal government to expand definition of AIDS. LA Times, 9 Aug. A37.

14. Cimons. 1991. LA Times, 9 Aug. A37; Associated Press. 1992. Center holds off on AIDS definition. Lansing State J. 5 Jan.

15. Eckholm E. 1991. Facts of life. More than inspiration is needed to fight AIDS. NY Times, 1 August, sec. 4, pl.

16. Grmek M. 1990. RC Maulitz, J Duffin, trans. History of AIDS. Princeton: Princeton University Press, 109.

17. Centers for Disease Control. 1992. Mortality patterns-United States, 1989. MMWR 41:121­125.

18. Callen M. 1990. Surviving AIDS. New York: Harper Collins.