Preliminary results from a Veterans Administration (VA) study of AZT therapy were presented on 14 February 1991 in Washington, DC, at a special meeting of the Antiviral Drugs Advisory Committee of the Food and Drug Administration (FDA). The findings indicated that early AZT therapy conferred no benefits in terms of overall survival, and might well be harmful to black and hispanic patients.

Widespread panic and consternation ensued-in the AIDS Establishment, among "AIDS Activists", and in the stock market. The very next day, shares of Wellcome Plc (the parent company of Burroughs-Wellcome, the manufacturer of AZT) fell more than 10 percent, based "on market worries about FDA approval stemming from an article in the New York Times." (Reuter, 15 February 1991) A well-orchestrated campaign of flak began. Immediately the National Gay & Lesbian Task Force, the National Association of Black and White Men Together, and the National Minority AIDS Council issued press releases attacking the validity of the VA study.

Following is a report on the two-day meeting of the Antiviral Drugs Advisory Committee, which I attended:

Surrogate Markers

The first day of the meeting, Wednesday 13 February, was devoted to the question of surrogate markers-various blood tests which might be used to evaluate the efficacy of experimental drugs, in lieu of more traditional measures, such as survival or prevention or amelioration of illness. This concern reflects pressure being put on the FDA to provide a shortcut for new drug approval-to forego the usual testing procedures for efficacy and safety. To date, the only federally approved "AIDS" drug is AZT, which many "AIDS" patients are unable to take, owing to the drug's toxicities. Other nucleoside analogues, like DDI and DDC, are waiting in the wings. AIDS organizations and self-proclaimed "AIDS activists" are now demanding that the FDA approve these drugs on the basis of what is already known about them, which is not much. If a magic surrogate marker could be found, then a drug could be evaluated without performing placebo-controlled trials, which necessarily take time. In effect, the "activists" are calling for an end to drug regulation, which would mean a return to the marketplace anarchy of the 19^th century, when deadly patent medicines and poisonous food additives were sold without constraints as to advertising, labelling, or anything else.

In brief, various presentations indicated that the CD4 or T4 test is the only marker to have any predictive value for a patient's prognosis, and that it is not all that good. The CD4 test was described in such terms as "a partial marker". A certain amount of nonsense was spoken about how AZT's benefits were reflected in temporary rises in CD4 counts, but that AZT's benefits might somehow went beyond CD4 counts. All this was pretty nebulous.

The most absurd moment of the CD4 talks occurred when a slide was shown depicting a patient's CD4 counts. The line was roughly horizontal until he was given AZT, at which point- ZINGO! -- the line shot sharply upwards. And then-ZONGO! -- the line immediately shot right back down again to where it had been before. It looked like a horizontal line interrupted by a highly-elongated upside-down V. And this, presumably, was meant to show the benefits of AZT therapy.

In fact, temporary rises in CD4 counts following the initiation of AZT therapy may represent a well-known rebound phenomenon, and may not be good for the patient at all. The molecular biologists Peter Duesberg and Bryan Ellison explain it in the following way:

The other reason for an apparent benefit of AZT lies in the observation that many patients on this drug experience short-term increases in their immune system cells. This, however, is a temporary pseudo-benefit; when the body is initially exposed to any toxin that depletes its blood cells, a compensatory reaction begins to produce large quantities of new blood cells to replace the poisoned ones. The temporary increase in all blood cells, including immune cells, is likely to be the result of the body's reaction to AZT, which later proves futile in the continued presence of the drug.(1)

One thing Wednesday's meeting did was to lay to rest, once and for all, the P-24 antigen test, which for several years has been used to claim benefits for AZT. The P-24 antigen test is now regarded as useless, and not a word was said in its defence. It means nothing at all.

The Veterans Administration Study

The main focus of the meeting on Thursday 14 February was Veterans Administration Cooperative Study 298, whose preliminary findings were presented by John Hamilton, MD. This study involved 338 HIV-positive individuals whose T4 cells were in the range between 200 and 500 and who showed some "symptoms or signs of HIV infection", including thrush, oral hairy leukoplakia, zoster, unintentional weight loss of 10% or more, unexplained persistent diarrhea, fever (100.5 degrees Fahrenheit), night sweats, fatigue, dermatitis, or lymphadenopathy. According to "risk group" category, 65% were gay men, 15% were intravenous drug users, and 8% were both gay men and intravenous drug users. Ethnically, 65% were white and 35% were black or hispanic. Patients were enrolled into the study as early as January 1987 or as late as January 1990.

The patients were randomized into two treatment arms: the first (early treatment) received 1500 mg. of AZT per day; the second (later treatment) received placebo followed by AZT at the point where their CD4 count fell below 200 on two successive occasions.

Hamilton stated his overall conclusions twice-at the beginning of his talk, and then again at the very end. His first statement of conclusions is as follows (verbatim):

Early zidovudine delayed progression to AIDS, as compared to later treatment. But no benefit for either treatment arm was detected for survival or the combined clinical endpoints of AIDS and death. Early zidovudine resulted in transitory benefits in whites and neutral or harmful effects in black and hispanic patients.

At the end of his talk, Hamilton expanded his conclusions somewhat, as follows (verbatim): We found that early zidovudine therapy delayed the progression of AIDS. We also found that survival was comparable in the two treatment groups. That is, no benefit-no detectable benefit. We found that early zidovudine resulted in transitory benefits in whites and neutral or harmful effects in black and hispanic patients. And we conclude that further studies are mandatory in minority populations.

I made the two brief transcripts above from listening dozens of times to the tape I made of Hamilton's talk. They are, word for word, what he said. I emphasize this because reports on Hamilton's talk have strangely distorted, or neglected to mention, the conclusions that he himself presented. For the record, there they are.

Hamilton described the toxicities of AZT treatment, all of which were found more often in the early treatment group.Comparing early vs. later, more patients on early zidovudine: were anemic, were neutropenic, had nausea and vomiting, had diarrhea, had central nervous system abnormalities, and had headaches. In response to questioning, Hamilton said that transfusions were given when necessary. (See Table 1 below)

Not only did AZT not confer any benefit in terms of survival, a slightly higher proportion of patients died in the early treatment group (14%) than in the later treatment group (11%). One astounding finding was that in the early AZT group, 10 patients (6%) died without ever progressing to CDC-defined "AIDS", whereas none of the patients in the later AZT group did so. One must ask, then, what these patients died from, if not from "AIDS"; and the answer is that they probably died, at least in part, from AZT poisoning. It would seem a dubious benefit to take a drug that will prevent you from progressing to "AIDS" by killing you first. (See Table 2 below)

By far the most controversial finding was the possibility that AZT treatment might be more harmful to black and hispanic patients than to whites. Among the minority patients, the death rate was significantly much higher in the early treatment group (14%) than in the later treatment group (2%). One can only speculate as to why there should be ethnic differences in the response to AZT treatment. But, at least for the patients in this study, the differences appear to be real.

Non-responses From Burroughs-Wellcome

After Hamilton's presentation a number of talks were given by people who are, officially or unofficially, in the Burroughs-Wellcome camp: Margaret Fischl, Paul Volberding, Stephen Lagakos, Sandra Nusinoff Lehrman, Terry Creagh-Kirk, and others. Nothing any of them said was particularly memorable or relevant. Paul Volberding discussed Protocol 019 in a desultory way, striving to downplay the toxicity of AZT. Margaret Fischl dismissed AZT's toxicities by showing slides analyzing various toxicities in terms of "rate per 100 person years", whatever exactly that meant-no one asked her to explain. Stephen Lagakos' talk consisted of meaningless generalities, and he showed slides with handwritten words on them; one slide, which attempted to explain why the pro-AZT studies (016 and 019) differed from the VA study, gave as a possibility, "016/019 data spurious?", which at least caught my attention. Terri Creagh-Kirk, who was supposed to talk for half an hour, spoke for only a couple of minutes. Other Burroughs-Wellcome presenters whizzed through a series of slides that apparently had something to do with survival.

Clearly Burroughs-Wellcome was faced with the problem of damage control. It was incumbent on them to say something in response to the findings of the Veterans Administration. But without any idea what to say, all they could do was talk. Which they did. More than one person fell asleep.

Discussion

An open public hearing followed the presentations. The first speaker was Wayne Greaves, MD, an infectious disease specialist at Howard University and a consultant to the panel. He said he was disturbed by the findings of the Veterans Administration study, which indicated that early AZT treatment might have no benefits, or might be harmful for minorities.

I was the second speaker. Since I had only five minutes to make my case, I decided to pull no punches. The audience of several hundred people listened intently to what I had to say, though I found out later from a young man in the FDA that some of his superiors were red in the face from rage at what I was saying. (See text of speech.)

The third speaker was Mark Harrington of ActUp New York, who said that for black and hispanic people, the cry should now be: "Ten years, one billion dollars, no drug."

Next followed a closed session of discussion among the panel members. For the physicians, the question was posed, whether they would change their practices regarding AZT for asymptomatic HIV positives, based on the findings of the VA study. Wayne Greaves said that they "cannot sweep the issue under the rug." Greaves said, "I will tell them the data contradict earlier studies, that early AZT therapy may not be useful and may even be harmful to minority patients." Greaves questioned the propriety of the current "Living With HIV" campaign sponsored by Burroughs- Wellcome.

Neil Schram, a gay physician from California, said he would not routinely prescribe AZT for asymptomatic HIV positives, that he would have long discussions with them first, pointing out the possibilities that AZT might have no benefits or even be harmful. He said it was his "gut feeling that AZT may be poison for black people."

Anne Gershon, a pediatrician at Columbia University College of Physicians and Surgeons, expressed a key sentiment of the panel members: "These are very disquieting results. We don't know where we stand."

Deborah Cotton of the Harvard Medical School said she considered Dr. Greaves concerns to be important, and she commented on the inappropriateness of the Burroughs-Wellcome campaign. She discussed the difficulties of conveying the uncertainties of AZT's benefits to practising physicians, who typically receive most of their information from "detailmen" (salespeople for the pharmaceutical companies).

Donald Abrams, a San Francisco physician, pointed out that the VA study had been carried through to completion-it had not been prematurely terminated, as had all of the studies used for the approval of AZT. It was questionable to prescribe AZT for asymptomatic patients, he maintained, in the absence of any survival benefits. The findings from Protocol 019, discussed last year, merely showed the annual progression rate to AIDS being cut from 4% (without treatment) to 2% (with AZT); either way, the vast majority of patients did not develop AIDS. He said patients have a right to know that different results are coming from studies done by different institutions.

John Hamilton said he would not change his policy with regard to AZT, because he has never prescribed AZT based only on HIV status and CD4 counts.

David Ho said, "AZT is not a great drug." In light of the budget crisis, he advocated developing more and better drugs, rather than doing any more research on AZT.

Perspectives

My overall impression of the VA study is that, while it is not definitive, it is good, honest, competent research, which deserves to be taken seriously. The tables that John Hamilton showed in his presentation were clear and meaningful, and were acceptable by the standards of professional research. Hamilton scrupulously indicated the confidence intervals for his data, and let the audience know when caution in interpretation was called for.

In contrast to the VA study, the research used to claim benefits for AZT has consistently been bad. In my book, Poison By Prescription: The AZT Story, I go into a detailed analysis of the Phase II trials(2), on the basis of which the FDA approved AZT for marketing, and the major AZT survival study(3). Both studies can fairly be described as not just sloppy, but fraudulent.

The report by Paul Volberding and Stephen Lagakos on Protocol 016(4) (AZT treatment for asymptomatic HIV positives), which was the basis for the FDA decision to approve AZT for HIV positives with T4 counts below 500, is unmitigated garbage. The description of methodology is inadequate and incomprehensible. None of the tables in their report are acceptable; none of them make sense. None of their tables even show bases-looking at one, it is impossible to tell either what people or how many people the table is based on. Nor is this information to be found elsewhere in their report. In fact, there are really no hard data at all in the report, just a lot of unjustified generalizations and meaningless numbers and verbiage.

The press release issued by the National Gay & Lesbian Task Force (NGLTF), "Government Study on AZT and People of Color Questioned by NGLTF", is a disgrace. It attacks the VA study from a standpoint of total ignorance-of the study itself, of statistics, and of research techniques. I called the two contacts listed on the press release: Belinda Rochelle and Robert Bray. Rochelle refused to answer "difficult questions", and said I should talk to the press officer, Bray. When I said that someone listed as a "contact" ought to be willing to answer questions, she responded that she was a "lesbian of color", and how dare anyone question her qualifications, at which point she hung up. Bray, on the other hand, was more than willing to talk, but incapable of giving straightforward answers to questions. After giving him several opportunities to describe the qualifications of himself or anyone else at NGLTF to evaluate statistical research, and instead being subjected to an endless stream of irrelevant rhetoric, I gave up. Clearly Bray himself knows absolutely nothing about statistics. Although Bray denied that the press release was pro-AZT, there is no other way to interpret it. Why has NGLTF chosen to attack the VA study, when it has never questioned the really bad research that has been used to claim "benefits" for AZT-the Phase II AZT trials or Protocol 019, for example? Why should NGLTF strain at a gnat, and swallow a camel? NGLTF ought to be doing everything it can to stop the pharmacogenocide of gay men; instead, it is collaborating in that genocide.

Aftermath: Burroughs-Wellcome Strikes Back

Burroughs-Wellcome lost no time in trying to counteract unfavorable publicity arising from the results of the VA study. A "Dear Doctor" letter was sent out to physicians on 15 February 1991, signed by Sandra Nusinoff Lehrman, Head of the Department of Infectious Diseases at Burroughs-Wellcome. Following is the text:

The FDA Antiviral Drug Products Advisory committee met February 14, 1991, to review new data on the efficacy and safety of RETROVIR brand zidovudine. This data [sic] reaffirmed the usefulness of RETROVIR in treating individuals who are at less advanced states of HIV infection. These data are described in the attached "Talk Paper" issued by the FDA following this meeting.

It is possible that some lay press coverage may cause patients to have questions about therapy with RETROVIR. For this reason, we have sent you a copy of the FDA "Talk Paper," which may be of value in responding to patient inquiries. A copy of the RETROVIR prescribing information is enclosed for your reference.

The FDA "Talk Paper", dated 14 February 1991, reads as though it had been drafted by Burroughs-Wellcome. A more false and distorted version of the meetings could hardly be imagined. The first sentence claims that the committee "reaffirmed the drug's usefulness in treating individuals who are at less advanced stages of infection with the AIDS virus." Then the "Talk Paper" goes on to summarize the VA study in such a way as to obliterate the most important finding: that AZT treatment conferred no benefit in terms of survival. To fully appreciate the dishonesty of the FDA "Talk Paper", compare Box A, in which John Hamilton summarizes the results of the VA study, with Box B, in which the FDA falsifies the same findings by eliminating most of what is unfavorable to AZT.

It should be clear that Burroughs-Wellcome is a thoroughly unscrupulous company, and that collusion between the FDA and Burroughs-Wellcome is as strong as ever. *

TABLE 1

*Difference (6% vs. 0%) is significant at the 99.99% confidence level.

Veterans Administration Cooperative Study 298

TABLE 2

*Difference (28% vs. 13%) is significant at the 99% confidence level.

**Difference (14% vs. 2%) is significant at the 98% confidence level.

Veterans Administration Cooperative Study 298

BOX A:

OVERALL CONCLUSIONS OF VETERANS ADMINISTRATION COOPERATIVE STUDY 298 AS PRESENTED BY JOHN HAMILTON WASHINGTON, DC -- 14 FEBRUARY 1991

(Verbatim)

1. Early zidovudine therapy delayed the progression of AIDS.

2. Survival was comparable in the two treatment groups. That is, no benefit-no detectable benefit [to early zidovudine treatment].

3. Early zidovudine resulted in transitory benefits in whites and neutral or harmful effects in black and hispanic patients.

4. Further studies are mandatory in minority populations.

BOX B:

DISTORTED VERSION OF VA STUDY 298 FDA "TALK PAPER" 14 FEBRUARY 1991

One study presented to the committee was conducted by the Veteran's Administration. Preliminary results from this study, in general, confirmed that earlier use of zidovudine was beneficial for delaying the onset of AIDS. However, the study when analyzed by various demographic factors also indicated that zidovudine's effects might vary significantly among different patient groups. For unknown reasons, among the African-American and Hispanic patients in the study, those who received zidovudine at a later stage of their infection may have fared better than those who received earlier treatment with the drug. The results regarding the outcome of African-American and Hispanic patients were not conclusive, however, and thus no definitive changes in practice were deemed appropriate by the committee.

References

1. Peter H. Duesberg and Bryan J. Ellison, "Is the AIDS Virus a Science Fiction?: Immunosuppressive Behavior, Not HIV, May Be the Cause of AIDS", Policy Review, Summer 1990.

2. Margaret A. Fischl et al., "The Efficacy of Azidothymidine (AZT) in the Treatment of Patients with AIDS and AIDS-Related Complex", New England Journal of Medicine, 23 July 1987.

3. Terri Creagh-Kirk et al., "Survival Experience Among Patients With AIDS Receiving Zidovudine [AZT]: Follow-up of Patients in a Compassionate Plea Program", Journal of the American Medical Association, 25 November 1988.

4. Paul A. Volberding and Stephen W. Lagakos, et al., "Zidovudine in Asymptomatic Human Immunodeficiency Virus Infection: A Controlled Trial in Persons with Fewer than 500 CD4-Positive Cells per Cubic Millimeter", New England Journal of Medicine, 5 April 1990.