The Science of "Concordology"
As treatment activists focused on such drugs as the non-nucleosides, the tat and protease inhibitors, and still more experimental approaches, controversy continued to swirl around AZT. This made sense: AZT was still, by far, the most commonly prescribed AIDS antiviral, and it was considered the best drug with which to begin antiretroviral therapy. The drug had a high profile, in part because Burroughs Wellcome's marketing of the drug was sophisticated and relentless: the company's efforts included promotional videos mailed to doctors, a toll-free "AIDS information" hot line designed to peddle the company's product, and mass media and billboard ad campaigns preaching the benefits of knowing one's antibody status and initiating antiretroviral treatment early. Yet questions about the drug's use had never fully been answered to the satisfaction of many.
Longtime opponents of AZT, including HIV dissenters such as John Lauritsen and Chuck Ortleb, attacked those who had climbed aboard the AZT bandwagon. Lauritsen labeled Project Inform and the Gay Men's Health Crisis as "gay quisling groups" for their endorsement of nucleoside analogues, while the Native referred to John James as an "AZT pimp." Such attacks became more vehement after some of the treatment activist groups began accepting donations from Burroughs Wellcome. In 1992 Project Inform received a $150,000 grant from the company to upgrade its computer equipment. TAG and ACT UP/Golden Gate received small donations, but TAG also brokered a $1 million grant from the company to AmFAR in support of community-based research. "If I have to take money from the devil to save my life and the lives of my friends, I'll do it," said Peter Staley of TAG. Yet even if the donations came without strings attached, the acceptance of such funds left treatment groups vulnerable to attack by opponents of AZT. One leaflet that made the rounds in San Francisco in 1992 and purported to come from a group called "ACT UP/Underground" described Burroughs Wellcome's "blood payment" to Project
Inform. In a memorable phrase, the leaflet accused the organization of having "promised to deliver its own community to nucleoside slavery, bound in the chains of corporate greed!"
At a different volume level, mainstream authorities continued to debate the merits of early intervention with AZT in people who were asymptomatic. "After 5 Years of Use, Doubt Still Clouds Leading AIDS Drug" was the New York Times headline of a 1992 article by Gina Kolata that charted the undercurrents of uncertainty. No study had yet demonstrated that AZT actually prolonged life in people who started taking the drug when they had fewer than 500 T cells, although the Veterans Administration study, once completed, indicated that AZT did prolong the length of the disease-free state. (Concerns about racial differences in response to AZT based on a preliminary report from the study were generally allayed by retrospective analyses of earlier studies that showed no treatment differences based on race.) This study was considered too small to provide a definitive answer to questions of survival; instead, researchers now looked to the completion of the Concorde study in Europe.
Meanwhile, doctors and patients grappled with uncertainties: Was it better to start treatment early, before the immune system had deteriorated and while the patient could more easily tolerate a serious drug, in the hope of delaying the onset of symptoms, thus keeping the patient healthy until more effective drugs arrived on the scene? Or was it better to wait, sparing outwardly healthy patients a toxic drug that could cause anemia or liver damage and avoiding the development of resistance, so that the drug would be effective later in the course of illness at the point when there was clearer evidence that it actually provided a benefit? As a 1992 survey of 448 physicians reveals, the vast majority simply followed the NIH guidelines and the FDA labeling, prescribing AZT for anyone with fewer than five hundred T cells per cubic millimeter. But doctors and patients who tracked the science of AIDS more closely were often less certain about the best course of action.
Still, nobody seemed quite ready for the news when the Concorde researchers released their preliminary report on April 1, 1993, in a brief letter to the editor in Lancet . Eight hundred seventy-seven patients, asymptomatic upon entry into the study, had been randomized into the "immediate treatment" arm of the study and had received one thousand milligrams of AZT daily; 872 had been placed in the "deferred treatment" arm, receiving a placebo unless they became symptomatic,
at which point they were put on AZT. In total, the study provided 5,328 person-years of data, making it the largest and longest study of its kind. The researchers reported that the patients in the "immediate treatment" arm experienced more of a CD4-count boost, on average, than those in the "deferred treatment" arm. However, "by contrast with the differences in CD4 count, there was no significant difference in clinical outcome between the two therapeutic strategies." The three-year survival rate was 92 percent for those who had begun taking AZT early and 93 percent for those who were put on the drug later. And similarly, there was no significant difference between the two groups in terms of disease progression and the development of opportunistic infections.
In the eyes of the Concorde investigators, Ian Weller in Britain and Jean-Pierre Aboulker in France, the study did not so much contradict 019 and the other studies of AZT use in asymptomatic patients as subsume them by proceeding for a longer period of time. The investigators had found an initial trend suggesting benefit in the early use of AZT, similar to that found by Paul Volberding in ACTG 019. But since the trend wasn't statistically significant, the trial was kept going, and as time passed the benefit simply disapeared. The bottom line, according to the Lancet report, was this: starting AZT early, as opposed to starting it later, did not extend survival or even the disease-free state. And CD4, the "surrogate marker" that had been employed in the licensing of ddI and ddC in the United States, could not be considered predictive of long-term treatment differences.
It was a painful moment that left many people dazed—or defensive, or angry, or jubilant, depending on their previous commitments in the AZT controversies. "Rarely have a mere eight paragraphs sparked such fury, hysteria and hyperbole," said an article in a San Francisco lesbian and gay newspaper. Not long afterward, one researcher with the World Health Organization would suggest that "a whole tropical rain forest has disappeared as a result of this study," which ushered in a brand new domain of scientific inquiry he dubbed "Concordology." And in the space of a few days following the report in Lancet , the stock value of Wellcome PLC dropped by more than $500 million.
Billed as the "definitive" study of AZT use in asymptomatic HIV infection, Concorde initially appeared to settle nothing; it seemed almost infinitely malleable in the interpretations it could generate. "You can interpret [Concorde] as your bias dictates," said the director of
inpatient services at San Francisco General Hospital. Weller, the British principal investigator, who described himself as having already been "very conservative in the use of zidovudine [AZT] monotherapy in asymptomatic patients," now said: "Those physicians like myself who tended to wait until symptoms appeared are going to be more sure they were doing the right thing." Douglas Richman, one of the principal investigators in the AZT studies in the United States, commented: "This is a chronic active persistent infection. It's incomceivable to me that not treating it is the way to go." One "knowledgeable and compassionate AIDS doctor told [GMHC's publication, Treatment Issues ] privately that he would continue to prescribe AZT for his asymptomatic patients even if it did not work. He said there was nothing else for him to do."
Those who wanted to score political points had plenty of opportunity. Larry Kramer, for instance, used the study to highlight the overall failure of the research effort in an overwrought editorial he penned for the Advocate , entitled "AZT Is Shit." Other responses followed predictable patterns. "Advocates of the AZT-is-poison school have had a field day," wrote Tim Kingston, a reporter for the San Francisco Bay Times . "Not only do they assert that early AZT intervention is ineffective, but also that all AZT treatment and all HIV antiretroviral therapies are ineffective." One example was the comment of HIV dissenter and AIDS activist Michael Callen, quoted in the Los Angeles Times: "Taking AZT is like aiming a thermonuclear warheard at a mosquito." A San Francisco gay newspaper whose AIDS writer had been opposed to AZT ran a cover graphic of a hand tipping a gigantic bottle of Retrovir (the brand name of AZT), with capsules spilling onto tombstones in a graveyard.
A Los Angeles-based group of AZT dissidents, called Project AIDS, International, sought to spread the word to AZT "victims": "You need to know your rights. Ask yourself the following questions: Were you misled by the U.S. studies or information that was given you through theFDA or CDC? Were you pressured into taking AZT … by your doctors or an AIDS organization? … If you have answered yes …, you have a right to compensation.…" In London, newspapers reported that the widow of a person with AIDS was suing Wellcome PLC, claiming that her husband had died from the effects of AZT. And a London group calling itself Gays Against Genocide picketed an AIDS organization and a hospital where a clinical trial of AZT in children was in progress, accusing their targets of "murder"
and "baby-killing," respectively. Meanwhile, opponents of these views argued that if Concorde did anything it disproved , once and for all, the notion that AZT is poison: after all, a large number of people had taken a high dose of AZT for four years and only a small percentage experienced adverse reactions to the drug.
European AIDS authorities, by and large, tended to accept the study conclusions as given. But U.S. authorities rushed to explain that nothing really had changed. Until more data were available, "no physician or patient should change the approach they're using based on this study," Daniel Hoth of NIAID told USA Today . David Kessler, the FDA commissioner, told the New York Times that the results of Concorde were not a surprise, since experts "have known that AZT has real but limited benefit"; this of course sidestepped the issue of the FDA's labeling of AZT for use in any HIV-infected person with fewer than five hundred T cells.
East Coast AIDS groups, less well-disposed overall to the nucleoside analogues, criticized the official "spin control." GMCH's publication, Treatment Issues , commented: "The blizzard of press releases responding to Concorde reveals the extent of the personal and professional investment many U.S. researchers have in AZT's efficacy. While one should expect a profit-driven corporation like Burroughs Wellcome to interpret the data favorably, U.S. government agencies should be held to a different standard." But West Coast groups like Project Inform and publications like AIDS Treatment News that were more committed to early intervention reacted quickly to the study, and they rushed out their methodological heavy artilerry. Fighting against the "black-boxing" of Concorde, they strove to highlight as much contingency, messiness, and uncertainty as they possibly could.
First, patients in the study had received an unusually high dose of AZT—twice the dose that had become standard by the time of the study's terminatinn. From the "pragmatic" perspective on clinical trials, this meant that the study had questionable relevance to the real-world treatment decisions of 1993. A larger concern was the modification of the study protocol in midcourse, after the results of ACTG 019 had come out. To satisfy ethical requirements, the Concorde investigators had allowed patients with fewer than five hundred T cells to choose at any point to begin open use of AZT rather than remain in the double-blinded trial. In all, 282 of the 872 patients in the deferred treatment arm had switched from placebo to AZT. The accepted statistical practice, however (called the "intent-to-treat" rule), was to
analyze patients as if they remained in the original groups to which they had been randomly assigned . That is, the 282 patients who began taking AZT prior to progression to illness were analyzed as if they had been on placebo all along.
The purpose of the intent-to-treat rule in biostatistics is to put the burden of proof on the drug whose efficacy is under investigation. For example, if some subjects in the placebo arm of a trial accidentally receive the study drug, researchers continue to count them in the placebo group. Then, if a difference in outcome is still found between the two arms of the study despite the blurring of regimens that occurred, the researcher can be especially confident about the results. The problem, however, is that if no difference is found, the researcher might reasonably wonder: is the drug really useless or did the contamination of the original study design prevent a true difference from manifesting itself? No one representing the Concorde study was suggesting that the change in the study protocol might account for the negative outcome—that if no placebo patients had switched to open-label AZT, the "deferred treatment" arm would have done worse relative to the "immediate treatment" arm. But to many U.S. researchers (and certainly to Burroughs Wellcome) this seemed eminently plausible. In effect, they argued, two roughly similar groups were being compared, so it was no surprise when the outcomes were similar.
John James also raised questions about the surrogate markers issue, which, as he noted, was so vital to the program of accelerated approval of new drugs. The Lancet letter had reported only that the median CD4 counts over time failed to predict the outcomes of the two arms of the study. Most U.S. researchers, wrote James, thought it was insufficient to look just at the median counts: "This question must be addressed by case-by-case analysis of whether individual patients whose T-helper counts rose after starting the drug seemed to show improved prognosis as a result." Meanwhile, treatment activists around the country worried that the initial reports from Concorde would fuel a backlash against the changes that had been instituted at the FDA to approved drugs more rapidly. "The spin, I predict, will be 'This is all the activists' fault," wrote Larry Kramer. "New York Times chameleon Dr. Lawrence Altman said as much in his ineptly reported article.…"
Altman had written a follow-up to his initial news article on Concorde, a commentary entitled "AIDS Study Casts Doubt on Value of Hastened Drug Approval in U.S." The article explained how Concorde
researchers had "persisted" against the prevailing construction of belief about AZT, continuing their study even after ACTG 019 had ended. That persistence had paid off in a finding that overturned the conventional wisdom. Furthermore, "in challenging the reliability of the CD-4 count in evaluating AZT, the Concorde study rekindled a long simmering dispute between many European and American researchers over the validity of surrogate markers in H.I.V. and AIDS," wrote Altman, noting that the British government had refused to license ddI on the basis of surrogate markers. Altman quoted Ian Weller, the British principal investigator, about the "lesson" in the Concorde study: "Don't stop trials too early…. Whatever the pressures are, keep going as long as possible."
Some reacted defensively to such charges, while others were despondent. Yet all the responses seemed to reflect the same underlying disquiet. It wasn't just the study; it was the whole state of the science, and Concorde was merely the last straw. An editorial in GMHC's Treatment Issues did a good job of capturing the sentiments that seemed to feed many of the reactions to the study: "Concorde underscores the uncertainty many AIDS researchers and clinicians feel. A few years ago, many physicians believed that the nucleoside analogs … would transform AIDS into a so-called 'chronic manageable' condition. There is now a deep and growing sense among many that some of the basic assumptions underlying AIDS drug development need to be reexamined."