Today, some voices question whether accelerated approval is necessary any more, since there are already 17 HIV antiviral drugs on the market. Where is the urgency, they ask, that warrants the safety risks inherent in accelerated approval? Indeed, there is a greater concern over drug toxicity today. Some of the side effects people are dealing with, such as lipodystrophy and related cholesterol problems, were not clearly evident until well after certain drugs were given their accelerated approval. Is it time now to roll back the clock and make safety a much higher priority before allowing new HIV drugs on the market?
The mood began to change after initial studies of AZT (once known as "compound S") showed promise for people with advanced disease. Though far from a cure, it looked like the drug could extend life by perhaps 6 months to a year. Today, this seems inconsequential, but at the time, it was a breakthrough. This led to the first proposal, put forward by the FDA itself, for speedier access to experimental drugs. It was called the "Treatment IND" and, in theory, seemed promising. It offered mechanisms for making a new drug available to people in need at the first sign that it was effective and seemed reasonably safe. Rhetorically, it sounded much like today's accelerated approval, but HIV public policy groups were split on the proposal, with some prominent organizations arguing against it as a "sellout to the pharmaceutical industry." The opposing view prevailed and the proposal was approved in 1987, just in time to deal with the impending demand for AZT.
Using the new approach, AZT was distributed to more than 5,000 people after completion of a single controlled clinical trial that showed efficacy. Whatever the limitations of AZT, it was a satisfying program since it opened wide the door for purportedly better compounds, such as ddI and ddC, which were coming up behind it. However, it took less than a year to see that the FDA's heart really was not in the new approach. Both ddI and ddC stalled in the development process. Three more years went by with only AZT available. People began demanding faster drug approval through media campaigns, lobbying, and demonstrations. Groups like ACT UP made it their primary mission, while others worked intensively with the FDA leadership to get them to understand how the new Treatment IND had failed. At best, the Treatment IND was being implemented in a way that offered to cut only 6 months off the drug approval process. Whatever the regulations said, the FDA was still a bureaucracy, and the bureaucracy ruled.
During his campaign for the White House in late 1987, George Bush Sr. ordered his people to "help" the FDA draw up new rules to speed things up further. These rules, the second attempt at solving the problem, were known as the "Sub-Chapter E" provisions (also known as Subpart E; see diagram in "FDA Overview" in this issue). Though not well known or widely discussed, these rules spelled out new ways for the FDA to interact with drug companies from the earliest stages of drug development. Their goal was to create the conditions that would make earlier approval feasible, as it had been with AZT. They laid the groundwork, which ultimately made accelerated approval possible, though they had little immediate effect at the time.
Still searching for short-term solutions and waiting for the "second drug" to fight HIV, AIDS activists collaborated with Anthony Fauci, head of the National Institute of Allergy and Infectious Diseases (NIAID) to come up with a plan called "parallel track." Fauci angered the FDA by announcing his support for the plan at a public meeting in San Francisco, without prior consultation with the FDA. He forced their hand through public disclosure and the Department of Health and Human Services followed through by creating a national task force to develop and implement the plan. In short, the plan called for making drugs available under special programs, not through pharmacies, while they were still undergoing clinical trials. Access was provided on a track that ran "parallel" to the clinical trials of the drug. While the rules were being spelled out, the first parallel track program (later called "expanded access") was being established to distribute ddI to people who had failed on AZT. More than 16,000 people eventually received the drug through this program, which distributed it for free.
Two problems became apparent in the ddI program. First and foremost was a safety problem. Though only a few cases of pancreatitis were seen in clinical trials of ddI, many hundreds more occurred in the parallel track/expanded access program. When a New York Times article sensationalized a number of deaths associated with the problem, it nearly killed the program. But scientists pointed out that the people in the parallel track/expanded access program were generally much sicker and weaker, at the most advanced stages of AIDS, and the death rates could not be compared to those in clinical trials of people with mid-stage disease. Still, it presented a sobering reality that early drug access came at a price. No such program since then has experienced as severe a safety problem and ddI is still widely used and considered to be of only moderate toxicity.
The second problem with the ddI program was one of fairness. Since the drug was being distributed directly from the company to physicians and clinics, it tended to get only to a certain class of patients, namely people with private insurance and doctors who had the time to fill out the necessary paperwork. People who got their health care through public clinics and emergency rooms, in contrast, were much less likely to be able to take advantage of the program. Physicians in such settings did not have the time or support to handle the paperwork and follow-up. Nothing short of simple availability in pharmacies would solve this problem. It was this realization that led the FDA and some activists to pursue the final step of FDA reform, which was to be called "Accelerated Approval."
Accelerated approval combined the pharmaceutical company and FDA planning processes spelled out in the Sub-Chapter E rules with the concept of parallel track. The difference was that through careful planning with the FDA, a company could be in the position to get licensing approval, rather than just the right to give the drug away for free in parallel track/expanded access. The model for parallel track had been ddI, which then also became the model for accelerated approval. Though not all clinical trials were complete, careful planning with the FDA made it possible for the agency to grant accelerated approval to the drug well ahead of schedule. The key provision was that the company had to agree to the completion of certain clinical trials after the drug had been granted accelerated approval.
This solved two problems. First, putting the drug on pharmacy shelves made it routinely possible for public health clinics and emergency rooms to provide the drug as needed, without cumbersome participation in a special program. Second, it largely forced third-party payers (insurance companies, Medicaid, etc.) to pay for the drug since it could no longer be considered "experimental." The concept of accelerated approval was initially going to be called "conditional" approval (based on conditions of completing certain clinical trials), but third-party payers made it clear they would consider the word "conditional" to be a signal that the drug was still in experimental stages, and thus they would not pay for it. Daniel Kessler, M.D., then FDA Commissioner, changed the name just a few days before the accelerated approval hearing for ddI.
Accelerated approval, often preceded by a period of parallel track/expanded access, thereafter became the standard for AIDS drug development. The process achieved 2 critical goals. It put the drug in the hands of people in need much earlier than normal, and it acted as an incentive to pharmaceutical companies to work in AIDS, as it was initially the only place they could get accelerated approval from the FDA. From their point of view, the sooner they could start charging for a drug, the more attractive it was to them.
What happens in most such situations is that the companies set their goals based on the expected date of accelerated approval. Production is geared to reach maximum around that time, with little more than the amount needed for clinical trials produced before that. This has created considerable tension with AIDS activists and people in need.
It is true that the number of people today who desperately need a new drug is much smaller than it once was. With so many options and possible combinations, most people can find solutions that do not require something new. But for those who do need new drugs, those who cannot make an effective and tolerable combination with approved drugs, the need for expanded access and accelerated approval is as great as ever. To the highly treatment-experienced patient, it doesn't matter how many drugs are on the market if none of them work.
Some have argued that we only need expanded access, not accelerated approval today. In theory, this seems to make sense, in that such a system could still provide the new drugs needed for people in desperate straits. Yet this view overlooks the other issues that led to accelerated approval in the first place. The first of these is providing access in public health clinics and hospital emergency rooms. Expanded access has never been very successful in this regard, and it is often even less so today. Pharmaceutical companies today frequently offer expanded access only through a modest list of specific sites. Yes, people from other locales can be served through these sites, but only if all those involved are willing to do the paperwork. Again, people in public health clinics and emergency rooms are often left behind. Accelerated approval solves this problem today as much as it did in the past.
A second issue or situation that weighs heavily in favor of continuing accelerated approval is its value as a motivational tool encouraging companies to work in the field of AIDS. Today as never before, pharmaceutical companies face a long list of disincentives to work in AIDS. For one, the market size is not growing much (at least not in countries that are able to pay for the drugs). Another is the public pressure AIDS puts on companies, pressure to provide drugs free or at greatly reduced costs in developing nations. Yet another is the sheer number of drugs already available for HIV and AIDS. From the drug company point of view, each new drug shrinks their potential market share. All told, HIV is no longer the hot market it once was for drug companies. Thus, at least maintaining the potential advantage of accelerated approval may be critical to keeping companies involved.
The third objection people have raised against continuing accelerated approval is the fear that offering a continuous stream of new drugs without first fully testing them for safety is a time bomb waiting to go off. Sooner or later, they argue, some really bad side effect is going to show up that did not appear in the early studies of the drug. Some point to lipodystrophy, lactic acidosis, and mitochondrial toxicity as examples of side effects that were unleashed by hasty drug approval. This argument, however, doesn't stand up to the facts. The difference between regular and accelerated approval is today at most 1 or 2 years. Had we withheld the use of protease inhibitors for an extra year or two, we would have seen 2 outcomes. First, the death rate would have continued unabated throughout 1996 and 1997, with tens of thousands going to their graves unnecessarily. Second, we would almost certainly still have approved the drugs, lipodystrophy and all.
Surely, no one today is suggesting that we should withdraw approval of protease inhibitors because of side effects and toxicities. For example, lactic acidosis and mitochondrial toxicity are both believed to be most associated with the use of nucleoside analog drugs (our earliest class of antiretrovirals). We did not learn of these side effects until roughly 1998 or later, even though the contributing drugs were approved many years earlier. Could or should we have withheld the availability of nucleosides for upwards of 10 years, until we saw this side effect? Of course not.
In short, it seems that unless we are willing to accept greatly elongated drug approval times, we cannot rule out the potential for late-appearing side effects. Long-term side effects are simply a risk that must be accepted with any model of drug approval. The only alternative is to not approve anything.
The future of people with HIV is best served by maintaining the current regulations for accelerated approval, combined with a careful reading of the rules themselves. Many of the concerns that people voice today about accelerated approval were addressed and thought through years ago when the rules were written. A good regulation stands the test of time and does not require annual tinkering or revision. For now and the immediate future, the accelerated approval rules continue to meet a real need, even though their implementation may look quite different than a decade ago. But it is the situation that has changed, not the rules.
Martin Delaney is the founding director of Project Inform, a national HIV/AIDS treatment information organization (www.projectinform.org).
Back to the RITA! Summer 2002 contents page.