DDI: COMPASSIONATE ACCESS ANNOUNCED

On July 13, Bristol-Meyers Company announced that it would make the experimental anti-HIV drug DDI available to persons with AIDS who did not meet the criteria for clinical trials and had a "critical" need for the drug. This program, described by the company as "compassionate use", follows the "parallel track" ideas of NIAID director Dr. Anthony Fauci (see "Fauci Proposes 'Parallel Track' Treatment Access," AIDS TREATMENT NEWS # 82, June 30, 1989). However, many important details are not yet available.

DDI OVERVIEW AND IMPORTANCE

DDI emerged from the June Montreal AIDS conference with widespread professional consensus that it is the most important new AIDS antiviral at this time. The drug appears to be much less toxic than AZT, and the toxicities it does have are different- -opening doors to more effective doses, as well as combination therapies.

Like AZT, DDI is not a cure, and will have to be used as a maintenance treatment. It can be taken less frequently than AZT, probably twice a day.

Although the drug is in the same general class as AZT, there does not appear to be cross resistance -- meaning that strains of the virus which have become resistant to AZT are not automatically resistant to DDI. Therefore, DDI may be effective for people for whom AZT no longer works well. DDI may also be synergistic with AZT, meaning that the combination may work even better than would be expected by adding the efficacies of the two separate drugs together. But no one knows for sure, because as far as we know there have been no human tests of the combination.

Eventually AIDS virus strains will probably develop resistance to DDI, as with AZT. But the new drug should at least work for some time for people who cannot use AZT, or for whom AZT is no longer effective. And laboratory tests have suggested that when different drugs attack the virus in different ways (as suggested here by lack of cross resistance), it may take much longer for the virus to develop resistance to the combination than to any of the drugs separately.

Some scientists also suspect that bone-marrow toxicity from prolonged use of AZT might make it more difficult for the immune system to recover, even if the virus causing the immune deficiency can be stopped. DDI may provide evidence of whether or not this theory is true, by allowing AZT to be compared with an antiviral which has no bone-marrow toxicity. It will be important to see whether T-helper cells (for example) recover fastest with DDI alone, AZT alone, or a combination of the two.

In short, DDI will be most important for those who cannot use AZT. But also it may open doors to a whole range of new treatment possibilities, making possible creative research which can advance HIV management for the benefit of everyone.

These potential benefits, however, may be slowed or blocked by the ineffectual system of clinical research now in power. For example, no matter how clearly DDI works, the drug will have to go through a two-year ritual in which a statistically significant number of deaths and serious infections must accumulate in those in a control group not receiving the treatment. Trials to look directly at which patients do or do not improve while using the drug could be conducted much more rapidly, and would provide exactly the information patients and physicians want to know -- but such trials would not be accepted for drug approval, because it is hard to measure patient improvement scientifically.

Since important antivirals will take years to go through the approval process, and patients cannot wait for reform of the current unproductive research system and the entrenched interests behind it, immediate discussion has focused on programs to make drugs available before full approval to those who need them most -- after the drugs have passed safety tests and shown some evidence that they work. The basic fact shaping this discussion is the conflict between the interests of patients, who want to have more and better treatment options available, and the interest of institutions which, for differing reasons, want to restrict access.

Today there are increasing efforts to establish dialog among the different parties involved in treatment-access issues. Much of this dialog is focusing on the immediate and obvious point for negotiation -- the specific rules on which groups of patients will or will not be allowed certain treatment options.

DDI NEGOTIATIONS

What made the recent Bristol-Meyers announcement possible was the near-unanimous professional consensus coming from the Montreal conference that DDI looked good. (The only doubt we have heard so far was from a medical expert who questions whether HIV -- which is inhibited by DDI -- causes AIDS.) It would be hard for a company or for the FDA (U. S. Food and Drug Administration) to flatly deny access to a treatment when the medical profession is convinced that it could save lives. However, practical early access is not guaranteed, because there is still no consensus that the system should provide early access, even if the drugs clearly seem to work. And the all- important details of who will be able to get DDI have not been determined.

Bristol-Meyers made its announcement after intensive negotiations involving the company, the FDA, Dr. Anthony Fauci of NIAID (the U. S. National Institute of Allergy and Infectious Diseases), and the AIDS community, primarily represented by the Treatment + Data Committee of ACT UP New York (the city where Bristol-Meyers is based). ACT UP's Treatment + Data Committee has done an outstanding job; without its work the current opening for DDI might not exist.

The Treatment + Data Committee defined four categories of patients who should have access to DDI:

* The AZT-intolerant, meaning those who cannot take AZT due to drug toxicity;

* The AZT-resistant, those for whom AZT is no longer working well, probably because the virus has developed resistance to it;

* The protocol-intolerant, including those too ill to qualify for the formal trials, those who need to stay on another medicine which disqualifies them, and those with symptomatic HIV infection but T-cells too high for the trial; and

* The protocol-inaccessible, including those living too far from a trial site, those who could not enroll because the trial was full, and those whose physicians were unable to enroll them and who could not change physicians (for example, those whose primary care was at a public-hospital emergency room where physicians could not take the time to get them enrolled in trials).

This list reflects the fact that those who most urgently need a new treatment are those who cannot effectively use any standard one. It also reflects the requirement of Fauci's "parallel track" access proposal, that use of experimental treatments must not be allowed to interfere with ongoing trials (see AIDS TREATMENT NEWS # 82, cited above).

It is unclear at this time who will be allowed to use DDI. We have heard (but not confirmed) the following:

* At this time the only group fairly well assured of getting access is the AZT-intolerant -- because the FDA is likely to insist that if a standard therapy is available it must be used in preference to an experimental one, unless, due to toxicity, the standard therapy cannot be administered.

* Informal discussions with FDA officials have raised some hope that four other groups might also be included: those too sick to qualify for the trials, those who must stay on another medicine which disqualifies them, those who live too far from a trial site, and those who cannot enter a trial because the trials are already full.

* The problem with the "AZT-resistant" group may be the difficulty of defining it. For example, one trial design being considered for AZT-resistant patients would accept anyone who has been on AZT for a year or more, since it is believed that the AIDS virus may develop resistance after that time. Even for a formal scientific study it is not feasible to do viral cultures for every potential subject to prove that viral resistance to AZT has in fact developed.

The potential difficulty in getting access to DDI for patients who can tolerate AZT but do not benefit from it recalls the example of trimetrexate, a pneumocystis treatment which at first was allowed only to those who could not tolerate the standard treatments, but not to those who could take them but did not respond and had no other alternative except death (see "Trimetrexate With Leucovorin: Decisions That Save Lives, Decisions That Kill, AIDS TREATMENT NEWS #52, March 11, 1988). Following public outrage, these patients were also allowed access to trimetrexate. But what is not well known is that they were included through special procedures intended to avoid setting a precedent for the future. (We do not know why a special effort was made to preserve so clearly inhumane a procedure. One possible motive is that each early-access exception highlights a failure of the overall approval process, meaning that those who operate and support that process have reason to ration the exceptions as tightly as possible.)

THE POLITICS OF EARLY TREATMENT ACCESS

Ideally, persons facing a life-threatening illness should, with their physicians, be able to choose treatments based on medical merit, taking their whole medical situation into account. How can abstract, general rules make better decisions than those who know the specific, often unusual or even unique facts of a particular case? But the political reality is that institutions have more power to pursue their interests than patients do.

How do the different groups involved view the issue and perceive their interests? It would be difficult to answer this question fully. The insights and viewpoints below are some which have helped us in understanding what is happening. They refer to the general issue of access to treatments before full marketing approval, not specifically to DDI.

* Pharmaceutical companies. One clinical-trials expert described the position of pharmaceutical companies (toward the FDA) as, "Tell us what we must do, and that if we do it we will get the NDA." The NDA, or new-drug application approval, gives the company permission to market the drug -- permission often worth hundreds of millions of dollars. Everything the companies do is seen in the light of whether it will help them get the NDA, or not.

Providing access to experimental drugs before the NDA (through "compassionate use," "treatment IND," or "parallel track,") is an expense and bother to companies. They must pay not only for the drug, but also for associated research and administrative expenses, and they can seldom be reimbursed, let alone profit from this activity. There may also be manufacturing, quality assurance, and liability concerns.

But the overriding issue is whether providing early access will help the company get the NDA. Early access primarily for treatment use could also provide data to help prove the drugs effective, and therefore support the NDA. But historically, the "compassionate use" system has provided poor data, probably because it relied on busy physicians to fill out forms which they would rather not deal with.

Starting almost two years ago, on October 26, 1987, compassionate-use access to drugs (at least for AIDS) became much more difficult to obtain. On that date an FDA advisory committee recommended against approval of ganciclovir (DHPG), and the pharmaceutical industry interpreted this rejection as punishment of Syntex for making its drug available to thousands of patients through compassionate use. Only recently, when NIAID's Dr. Anthony Fauci proposed his "parallel track" for access to certain treatments during trials, has the misinterpretation been corrected. When company officials said that they could not use the parallel track because they feared that what happened to Syntex would happen to them, Fauci pointed out that Syntex got in trouble not for making its drug available, but for failing to do scientific trials early.

A key difference between the old "compassionate use" system and the newer "treatment IND" or "parallel track" is that the newer systems provide access under a protocol, so that better data can be collected. Community-based research organizations may be able to monitor patients, collecting data according to the protocol and relieving primary-care physicians of unwanted paper-work.

(Why then did Bristol-Meyers use the older term "compassionate use" in its announcement on DDI? Apparently the company did not want to be seen as taking sides in a fight between NIAID, with its "parallel track" proposal, and the FDA, which wants to revive its "treatment IND", which already exists but has not been used much because it has been interpreted so conservatively.)

For pharmaceutical companies, the bottom line is getting their NDA, which is granted by the FDA. Therefore, the most important factor determining whether or not these companies will be willing to provide their drug before marketing approval is the rules, often unwritten, set by the FDA. The public and the AIDS community have often failed to recognize this fact, in part because pharmaceutical companies and the FDA have an inner relationship between them which is hard for outsiders to penetrate. For example, which party takes the heat for an unpopular decision can be decided as part of a larger negotiated arrangement. For this reason, when treatments are not available, it is often hard to know whether the real problem is with the company which holds the exclusive rights to the drug, or with the FDA.

* The FDA. In discussions with FDA officials, the word "thalidomide" is likely to come up. Thalidomide was a drug disaster that led to the birth of thousands of deformed children in Europe in the early 1960s. No animal or other tests gave any advance warning of the problem. Fortunately, an FDA official noticed an obscure report suggesting other toxicity, and withheld U. S. approval long enough that the danger became known before the drug was distributed here.

After thalidomide, Congress amended Federal law to require proof of efficacy as well as safety before a drug could be marketed. This new law would not have stopped thalidomide, but in practice it has made new-drug approvals enormously more difficult. The current cost of drug development in this country is about $120 million for each new drug. And even aside from AIDS, analysts have questioned whether the current system saves as many lives as it destroys (see "FDA Reform: Major New Position Paper," AIDS TREATMENT NEWS #58, June 3, 1988).

The FDA has taken its mission from thalidomide and the resulting Congressional mandate -- to protect the public against dangerous or worthless drugs. Neither the FDA nor any other institution evaluates the risks or costs of not approving a drug which should be approved. People forget that thalidomide was a sleeping pill, and casually apply the standards appropriate for a new cold, cough, or baldness remedy for the development of life-saving treatments for diseases like cancer and AIDS. Thousands if not millions of people with diseases which are or should be treatable are sent away to die on their own, and there is no institutional responsibility. But if anyone is hurt by the other kind of mistake -- approving a bad drug -- then the FDA, the company, the researchers, and everyone else involved can expect to be blamed. The result is more than a distortion; it is a situation where half of the decision-making process does not take place at all.

The bottom line is not that we should weaken drug regulation, but rather strengthen it by balancing the costs of both kinds of errors. The rules of drug approval have immense and often hidden effects, not only on individual patients denied existing treatments, but also on the speed and creativity of the entire enterprise of medical research and development. Federal regulation largely controls the ability of medicine to respond quickly to new emergencies like AIDS, and to develop new treat- ments for old diseases like cancer.

* We do not have a clear picture of the other major institution in AIDS treatment development, namely NIAID (the National Institute of Allergy and Infectious Diseases), a branch of the National Institutes of Health. At this time the AIDS community is grateful that NIAID director Dr. Anthony Fauci proposed a "parallel track" system of allowing access to important new drugs while formal efficacy trials proceed. However, we are hearing of resistance to early access by some of the AIDS principal investigators working through NIAID contracts at sites around the country. Apparently these researchers fear that early access will deprive them of subjects for their trials, and as a result they are pressing to restrict such access. These reports have not yet been confirmed, but they have been greeted with anger in the AIDS community, which sees the NIAID research effort as unproductive in view of the time and money it has had, and which sees the notorious recruiting problems at academic research centers as being caused by poorly designed trials (that people cannot volunteer for even when they want to), not by the existence of other options for patients. This situation must be watched, because it could threaten access to lifesaving treatments.

WHAT SHOULD BE DONE?

Pharmaceutical companies and government agencies alike have been unenthusiastic if not hostile toward early treatment release. The old "compassionate use" system, for example, was supposed to apply to only a few patients. The "treatment IND" applied to groups, but the FDA has used this system after the full burden of proof has already been met, when only paperwork remains before approval of the NDA. And pharmaceutical companies have had little incentive to use either system.

What is needed instead is a flexible burden of proof that takes account of the uniqueness and potential value of a drug and the urgency of the need for it. For example, consider DDI:

* It will take probably two years or more to get statistical proof that DDI (or any antiviral) increases AIDS survival or reduces the frequency of major infections.

* During this time, 50,000 people will die in the United States alone, unless better treatment becomes available.

* Of the new treatment possibilities, DDI looks best at this time.

In this emergency the FDA should waive the requirement to prove reduced death or opportunistic infections. Instead, it could work with Bristol-Meyers to design much faster trials using p24 antigen, T-helper count, and clinical measures such as weight gain to show drug efficacy, in a program of clinical trials involving perhaps several hundred patients, testing different doses and testing the drug in different patient populations. Placebos could ethically be used in some cases, because patients would not be on the trial for long -- and after the trial would have the option of using the drug. If these rapid trials showed that the drug was useful, and the long-term experience available did not show serious problems, then the developer should get its NDA and be allowed to market the drug -- provided that post-marketing studies continued. (The approval might be called an "emergency NDA", to alert physicians that the drug was approved with less testing than usual because of the urgent need, and therefore should be used cautiously.) No legislation would be needed to implement this approach; the FDA has the power to start tomorrow.

This system would be clean to administer, and would provide enough incentive to insure that pharmaceutical companies conducted their trials rapidly. Then the medical community could evaluate the evidence available and make recommendations to guide practicing physicians. Patients and physicians could then decide whether to use the drug or to choose other options instead.

Why hasn't such an approach been used already? The reason is that the outcome measures available -- p24 antigen, T-helper cell count, and overall health of patients -- all have flaws and therefore are not technically attractive in the academic world which sets the tone for these decisions. Fantastic scenarios can be concocted in which drugs could look good after the trials suggested above, but really not give any benefit to patients. But the small chance of mistakenly approving such a drug must be balanced against the certainty of tens of thousands of deaths caused by the built-in, two-year delay of the kinds of trials currently required.