The Genie in the Bottle
It was in the wake of these decisions that the FDA's Antiviral Advisory Committee sat down in April 1992 to consider the approval of ddC. First the committee turned back to ddI, looking at the data that had since been accumulated from the first of the Phase II trials to be completed. With a collective sigh of relief, the panel concluded that faith in ddI had apparently been borne out: its impact on CD4 counts had indeed predicted its clinical benefit. Nor had the approval of ddI impeded the successful completion of the trial, as committee members like Cotton had feared. "We took a risk in approving DDI and today I think that on balance we did it right," said Kessler, expressing the philosophy of what some were calling the "new FDA." "We cannot wait for all the evidence to come in when people are suffering and dying from these devastating diseases."
With renewed confidence in the use of CD4 counts as a surrogate marker, the committee then turned to ddC. A week before the meeting, Kessler had made official the new option of what would now be called not "conditional" but "accelerated" approval, along the lines originally proposed by activists. According to David Feigal, who would replace Ellen Cooper as head of the Antiviral Drug Division, Kessler emphasized that such approvals should indeed be thought of as "conditional."
But Kessler couldn't call the policy by that name for fear that third-party payers might decline to reimburse patients for their purchases of such drugs. Here again, the FDA was forced to be acutely sensitive to the vast financial and human consequences of the signals the agency sent out—despite the fact that, officially, "our approvals are not intended to be the basis for reimbursement decisions."
Hoffman-LaRoche asked that ddC be conditionally approved as a monotherapy (that is, for use by itself) for patients who were intolerant to AZT; but there was no evidence to support such a labeling, and the committee quickly nixed the idea. Instead, the committee turned to the evidence from ACTG 106 and other small studies that showed positive effects of ddC/AZT combination therapy on CD4 counts. Ironically, as Michael Botkin, an activist and writer on AIDS treatment issues for a San Francisco gay newspaper, described it, Burroughs Wellcome "[pulled] their rival's chestnuts out of the fire." In past years, Burroughs Wellcome had been decidedly uncooperative about research involving the use of AZT and the products of other companies. But with the growing interest in combination therapy, and with the widespread recognition that AZT alone was no solution to AIDS, the company stood to gain from an official endorsement of the ddC/AZT combination. They presented data from a study the company had commissioned that corroborated the results of ACTG 106 by showing the drug combination's effect on T-cell counts. By an eight to three margin, the committee voted in favor of conditional approval of ddC when used in conjunction with AZT. (Final approval would depend on the outcome of other studies, particularly ACTG 155.) As Botkin commented, that put Burroughs Wellcome in the "delightful position of proving that their rival's treatment is effective—but only when taken with their own product!"
The FDA ratified the advisory committee's recommendation on June 22, in what Health and Human Services Secretary Louis W. Sullivan described as "another step forward for patients with AIDS." Many treatment activists, like Delaney and James, were pleased by the outcome. Others were decidedly cynical. "For once, politics played in our favor," said G'dali Braverman of ACT UP/Golden Gate, adding that the Antiviral Advisory Committee meeting had a "scripted" feel, with dissenters being "strong-armed" by Kessler. In Botkin's analysis, Kessler had been seeking support from all sides, playing to the AIDS activists and the deregulation advocates by adopting accelerated approval, while placating consumer protectionists with simultaneous, well-publicized campaigns to beef up vitamin and food labeling.
Meanwhile, the underground ddC was a threat and a growing embarrassment. "The only answer, from Kessler's perspective, was to co-opt the buyers clubs by quickly approving ddC, thus demonstrating that the government is moving fast enough to introduce new AIDS treatments and that independent efforts aren't needed."
Comments by some committee members supported these activists' perceptions that the results had been foreordained. Deborah Cotton, who voted against approving ddC just as she had voted against ddI earlier, told Science that in her view, the committee had been asked to "pound [the data] into a scientific conclusion." FDA Commissioner Kessler "clearly wanted it," Cotton afterwards recalled. "David spent the entire two days at the table, which is usually how you tell that he's invested in an issue. …" From Cotton's standpoint, the ready reliance on surrogate markers was a mistake, because it threatened to delay the process of obtaining solid data about treatments. "We really have to ask whether relying on surrogate markers will hasten a cure or hinder it," Cotton told the reporter from Science: "We're getting into a situation of such complexity that we may have a large number of agents being used and no way of distinguishing among them." She added: "It's sad that we may have nothing to offer people in 1992. … It's sadder that in 2000 we may have nothing, too. In 2000 we'll look back and say, 'If only we'd done this in a more rational way.'"
Kessler assured the worriers that accelerated approval demanded careful postmarketing studies. If the surrogate marker evidence didn't hold up, the FDA would then "[put] the genie back in the bottle." Yet as Kessler's own phrasing perhaps inadvertently suggested, removing an approved drug from circulation might be easier said than done. Short of a truly catastrophic result in the postmarketing study, was it really likely that the political environment would permit the withdrawal of one of only three antiviral drugs approved against HIV? Or had the construction of belief passed the point of no return?