D4T (Stavudine): Approval Recommended

The experimental anti-HIV drug d4T (generic name stavudine, brand name Zerit), now widely available in a "parallel track" program for patients who cannot successfully use either of the approved drugs AZT or ddI, was cautiously recommended for early approval by the Antiviral Drugs Advisory Committee, at an all-day public hearing on May 20. The FDA [U.S. Food and Drug Administration] must make the final decision on whether to grant approval on the basis of the data so far available; it does not have to follow the recommendation of its advisory committee, but usually does.

There has been considerable confusion about this hearing and what it means. We believe that most of the confusion does not involve d4T itself. Rather, it arises from the difficulty of fitting the particular situation of this drug into the existing body of approval regulations, in such a way that a reasonable result is achieved.

[Note: This writer did not attend the hearing. This report is based on discussions with people who were there.]

Background

A major clinical trial of d4T (known as the '019' study) is still underway in the U.S.; it is expected to finish at the end of 1994. After completion of this trial, it will take months to analyze the data and schedule another hearing of the Antiviral Drugs Advisory Committee; as a result, approval will be delayed for about a year if it is necessary to wait for the final data and its analysis. To avoid this delay, the developer of d4T, Bristol-Myers Squibb Company, asked for an earlier approval under the new "accelerated approval" regulations. These regulations allow the FDA to approve a drug for serious or life-threatening illnesses, when no other satisfactory treatment is available, on the basis of "surrogate markers" (for example, blood tests such as the T- helper count, which seems to provide a crude measure of a drug's effectiveness). But the pharmaceutical company must finish its clinical trials, in order to eventually obtain conclusive proof that the drug is clinically beneficial to patients.

D4T Results

The May 20 hearing focused on partial data from two human studies of d4T. One is the Bristol-Myers '019' study -- the ongoing clinical trial mentioned above. The other is the parallel track program, run by Bristol-Myers with FDA permission primarily for compassionate purposes, to make the drug available before approval for persons with no other therapeutic option.

The '019' study includes 822 patients. All of them had been on AZT for at least six months (but often considerably longer). When they entered the study, each patient was randomly assigned to either continue AZT, or switch to d4T. The study is "blinded"; neither the patients nor their physicians were told who is getting which drug.

At the May 20 hearing, data was available for 359 patients. Data from the other 463 patients could not be used, because their treatment is still blinded. (Researchers are very reluctant to break the blind prematurely, since doing so could bias the study.)

About 14 percent of the patients in this study entered with a T-helper count under 100. About 50 percent were in the range of 100 to 300; and the other 36 percent were over 300. Only 10 percent had an AIDS diagnosis at entry.

The data showed that:

* Those assigned to d4T had an average rise to a maximum of about 20 to 25 in their T-helper count, compared to baseline (just before their d4T treatment began). Counts then declined, however, and by week 16, the average increase from baseline was only about 10. For those assigned to continue on AZT, T-helper counts decreased. According to Bristol-Myers, a difference of 50 in the T-helper count between groups was maintained for up to 60 weeks.

* HIV titer (of PBMC-associated virus) dropped 53 percent from baseline, for those on d4T. For those on AZT, the titer increased 11 percent.

* The d4T group also showed improvement in weight gain over the AZT group. The weight gain on d4T was small, however, and there was no significant weight gain for those with T-helper counts under 100. But the average difference between the d4T and AZT groups exceeded 2 kilograms (4.4 pounds).

* In an interview with AIDS Treatment News, Lisa Dunkle, M.D., who presented some of the data for Bristol-Myers at the May 20 hearing, said that the d4T group showed hematologic improvements over the AZT group. She did not present the hematologic data at the hearing, however.

* According to Dr. Dunkle, there was also a marked difference in favor of d4T in reduced side effects -- with the only adverse event being peripheral neuropathy in 12 percent of the patients after one year, and in only half of those was it serious enough to require discontinuation of the drug. (See other views of side effects, below.)

The data now available from the parallel-track program was less clear. This program is only for patients who had already tried and failed both AZT and ddI. As would be expected, they tended to be at a more advanced stage of HIV infection; their median T-helper count at entry to the parallel track was 41, and 75 percent of these patients had a T-helper count under 100 (compared to only 14 percent for the '019' study, as reported above).

In order to collect scientific data from this program, patients were randomly assigned to two different doses of d4T -- either the standard dose of 40 mg twice a day (the dose used in the '019' study), or half that dose, 20 mg twice a day. After a median time of 40 weeks, 79 percent of the patients were still alive. As a check to make sure that patients were not left on a dose known to be inferior, a group called the Data Safety Monitoring Board (DSMB) secretly unblinded the study and compared the two doses. It found that there was no survival difference between the doses, but that those taking the larger dose had a somewhat higher risk of peripheral neuropathy -- 21 percent of those receiving 40 mg twice a day, vs. 15 percent of those receiving 20 mg twice a day. As a result, the DSMB asked that the 40 mg dose be discontinued in the parallel track, and those who had been receiving it were switched to 20 mg.

Since the survival rate was the same with the 40 mg and 20 mg doses, does this mean that the two doses both worked equally well -- or does it mean that the drug did not improve survival for those patients at all? We cannot tell from the parallel-track data. There was no placebo or no-treatment group (which would have been unethical in this situation); and it was also impossible to compare d4T with AZT or with ddI, since the patients had to fail both of those drugs as a condition for getting into the parallel track. (The '019' study is designed to look for a survival difference -- between switching from AZT to d4T, vs. continuing on AZT -- so we should have information after this study is completed.)

Patients in the parallel-track program were not required to obtain regular T-helper counts. Somewhat fewer than half of them did choose to do so; and for them, T-helper counts tended to be stable for over a year.

Regulatory Issues

There has been some confusion about just what the May 20 hearing did decide, and what it means. To understand the current status of d4T, certain background should be kept in
mind:

* The studies are not comparing d4T against AZT (or against ddI or ddC, the other approved anti-HIV drugs) as a starting treatment. (Early, smaller studies evaluated d4T as a starting treatment in some patients, and found that the T- helper count rise appeared to last for a prolonged period -- but the large studies now ongoing are not designed to confirm this result, or to make sure that it translates into clinical benefit.) Also, there is no study comparing switching to d4T vs. switching to ddI, as a second treatment after AZT. Nor is there any comparison of d4T in combination with other anti- HIV drugs. These obvious question, that doctors would like to have answered, will not be addressed in the approval of d4T.

* Despite the limited information, and the disappointment due to the small T-helper count improvement seen in the '019' study, there is a widespread feeling that, on balance, d4T should be approved -- without waiting an extra year for the final data from the '019' study. Everyone knows that existing HIV treatments are inadequate; and all the evidence suggests that some people do benefit from d4T. At the same time, the existing parallel track program restricts access to the drug, and limits the normal development of clinical experience by physicians. Bristol-Myers has provided this program as a public service, at a cost of millions of dollars which it may not be able to recover, instead of denying the treatment to thousands of people. (But there is increasing concern that as the overall process of drug development becomes more expensive, other companies will avoid developing potential AIDS drugs, and may even conceal important treatment leads.)

* The regulatory complication about approving d4T now is the following. The FDA's new "accelerated approval" regulations would need to be used, since the data so far shows improvement in blood tests and in weight gain, but not in "clinical endpoints"; there is not enough data yet to provide statistical proof that, in the long term, certain patients treated with d4T will do better than those treated otherwise. The accelerated approval system was designed for just this situation.

But the accelerated approval regulations require that no other treatment be available; the point is to allow the increased risk of approving a drug early, without clinical- endpoint data, in situations where the need for the drug is very urgent. This is the case with those in the parallel track program, as they have already failed standard treatments.

The complication is that while accelerated approval does apply to one group of patients, the best data available are from another group, those with less advanced disease (who are in the '019' study). An additional complication is that the dose which generally appears best and will probably be approved, 40 mg twice a day, is the dose which the DSMB discontinued for those in the parallel-track program -- the patients that an accelerated approval will need to address.

In other words, there seems to be a general consensus that d4T should be approved; but the factual situation of this drug does not exactly match what the regulations were designed for. The FDA staff seems to have sidestepped this potential red-tape gridlock by asking the Advisory Committee to answer particular questions which have the effect of leaving the situation open. This gives the FDA staff more freedom to make the right decision on the basis of all available data, as it comes in, without needing to wait a year for the '019' study to be completed.

The FDA asked the Antiviral Drugs Advisory Committee to vote on the following questions:

"1. Does there exist a population for whom stavudine [d4T] will provide meaningful therapeutic benefit over existing therapies? If the answer is yes please comment on the characteristics of that population.

"2. Has the applicant provided evidence from adequate and well-controlled clinical trials establishing that stavudine has an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit?"

The Advisory Committee answered yes to both questions, by majority vote. This is considered a recommendation in favor of approval.

Side Effects Controversy

Over a year ago ACT UP/Golden Gate, a San Francisco-based activist group, started a series of public forums for people in the d4T parallel-track program, independently of Bristol- Myers or anyone else involved. ACT UP also developed a survey, and its own computer database to record peoples' reports, and was in phone contact with patients around the country. In July 1993 ACT UP/Golden Gate reported that more than half of the patients in its database had some degree of peripheral neuropathy or other side effect that they thought might be drug related. The organization presented an updated version of its report at the May 20, 1994 hearing on d4T.
[Note: This writer usually attends the weekly meetings of the ACT UP/Golden Gate Treatment Issues Committee. We were not involved in the d4T study, however.]

A practical result from the ACT UP study is the suggestion that, if people have sleep disturbances while taking d4T, the problem can often be corrected by not taking the second dose at bedtime, but taking it earlier in the day.

In addition, the Conant Medical Group in San Francisco did a retrospective study of the case reports of about 100 of its patients who participated in the d4T parallel track program; it found that more than half of them reported adverse effects which might have been drug related. While this study has not yet been published, the Conant Medical Group sent a summary report to the FDA before the May 20 meeting.

These numbers are considerably higher than those reported by Bristol-Myers, and also by the DSMB. No one knows why there is this difference. Self-selection may have had a large effect on the ACT UP numbers, since those who have problems during a treatment may be more likely to come to a meeting and be recorded in the database. But self selection would not be expected to have affected the Conant study.

The differences might also have occurred if Bristol-Myers and the DSMB made more effort to screen out minor side effects or those which appeared not to be drug related, while the other reports counted everything. The major side effect of d4T is peripheral neuropathy, which can also be caused by HIV itself. This makes it hard to interpret reports of side effects -- especially since the patients in the parallel- track program usually had advanced AIDS or HIV infection, which would have caused a number of cases of peripheral neuropathy that were not due to the drug.

The issue of side effects should be kept in perspective. ACT UP/Golden Gate, which first reported seeing a larger number of side effects, still supported approval of d4T at the May 20 hearing -- but asked for neuropsychiatric studies focusing on possible adverse effects of the drug. And the Conant Medical Group is continuing to use d4T through the parallel- track program. The issue of side effects will affect medical management of d4T use, but it has not changed the belief that the drug can be useful and should be available.

Comment

At the May 20 meeting, the FDA also asked the Advisory Committee to meet again, on the subject of "surrogate markers" in AIDS drug approval (for any drug, not focusing only on d4T). This concerns the ongoing controversy of whether we can trust blood tests such as T-helper counts, or clinical indications such as weight gain, as proof that an AIDS drug is really a benefit to patients.

On this issue, we believe it is time for AIDS community as a whole -- as well as the FDA, and its advisory committee -- to step back and take a look at the big picture. And the big picture, we believe, is that far too much effort is being spent agonizing over the details of marginal drugs which will never work very well. Instead, we should quickly approve the drugs when there is good safety data and some indication of benefit. Then we can re-focus research attention and resources on the discovery and development of fundamentally better treatments.