written by Jules Levin

Yesterday, January 11, the FDA Antiviral Drugs Advisory Committee was convened to discuss "clinical trial design issues for patients with HIV who have treatment experience and limited therapeutic options". This report is a brief summary of highlights, but a full report is being prepared now and will be communicated within 2-3 days. The Committee members discussed approving new drugs more quickly for patients with limited treatment options, how to improve collection of drug toxicities, new study designs for patients with limited options, and new parameters by which to judge new drugs for this patient population. At times the hearing was filled with taught confrontational discussions over some of the questions. The morning session was more like this, but in the afternoon members settled into a more productive discussion.

Several issues received particular attention:

This meeting was designed to invite participants to suggest new novel study designs for this group of patients. Participants were instructed to present novel ideas. Actually, designing registrational (for approval) studies for this group of patients is difficult for many reasons. For one, the collective of these patients is very diverse in their treatment experience and health status (CD 4, viral load, clinical). And traditional parameters such as percent undetectable may not be applicable, because for patients with no options achieving undetectable may not be realistic in all cases.

The use of "modified factorial trial design" received mixed support, but I think the FDA went away with support for this type of study. I think the industry will have to consider conducting more than 1 type of study for the diverse group of treatment experienced patients. This could include a modified trial design and a more traditional type study.

Genotypic-Phenotypic Resistance Profile

And there are a number of considerations that need to be addressed. I think for appropriate drugs we need to well characterize the resistance profile of the drug and the response of study patients with particular resistance profiles. Abbott conducted these types of studies for their new drug Kaletra. They reported that people with certain levels of gentoptypic & phenotypic resistance would respond in certain ways to Kaletra. For example, people with only 2-3 genotypic PI mutations would respond much better than people with 7-8 PI mutations. They also characterized how people with less phenotypic PI resistance would respond better to Kaletra than people with more phenotypic PI resistance. A full report on this is available on the NATAP web site in the Resistance Workshop and ICAAC Conference Reports.

As well, this received an extensive write-up in the September NATAP Reports post-Durban newsletter. The point is that patient's resistance profiles must be well characterized before entering these studies so we can well characterize how they respond to the new therapies. Then, the company can report information on how patients with particular levels of resistance can respond to the new drug. This will help interpret study results and help guide doctors & patients on how to use new drugs after approval.

What is a modified factorial trial?

Next year if and when we have 3, 4, or 5 new drugs in pre-approval development, this study design would utilize 3 new drugs in one study rather than the standard trial drug companies use which is standard of care plus a new drug versus standard of care. The FDA was looking for Advisory Committee support in attempting to tell the drug companies that they would like them to conduct this type of study. I think the FDA got that support. This will mean drug companies will have to collaborate on studies which in the past has not occurred much.

Here is an example of a modified factorial design:

For the sake of discussion lets include these 3 new HIV drugs: DAPD, T-20, and PMPA. Each person in each study arm would receive the best regimen from available drugs including what might be available from expanded access. We'll call this optimized best regimen (OBR).

OBR + DAPD + T-20

In the past this type of study would be OBR+new drug vs OBR.

Some questions did emerge: would we be able to sort out toxicities & drug interactions; in other words, with so many multiple combinations, could we assign an emerging toxicity to a particular drug. The FDA felt this could be accomplished but some concerns about this were expressed.

16 Week Data and Drug Toxicities-Side Effects

Another key question receiving attention is whether or not 16 weeks data is enough to receive accelerated approval. The regulation now is a 24 week study is required for safety and efficacy for accelerated approval and 48 week data for full approval. Because new drugs are needed quickly for people with varying degrees of treatment experience and because people with extensive experience are deeply in need of new treatment options, the committee considered shortening the requirement to 16 weeks for people with few treatment options. The European authorities appear to be prepared to use 16 week data. The committee in general supported using 16 week data. Everyone agrees we need therapies more quickly and 16 week antiviral activity data ought to be adequate for accelerated approval. However, the community panelists, which included myself, raised the concern that we need better collection and reporting of toxicity & side effect data from new drugs. I suggested that drug companies with new drugs coalesce to collect a database of this data from all new drug studies. 

For the collection of toxicity data16 and 48 week data is inadequate. Hepatotoxicity can take years to develop. Sometimes new toxicities don't emerge until after years after approval following a drug's use by thousands of patients. It took 2 years to appreciate the emergence of lipodystrophy. So a long term organized collection & reporting of data would be helpful. FDA regulations do not appear to require such reporting for accelerated approval. Drug company representatives were in the audience and heard all the community panelists agree on this issue. Community panelists will follow-up with the drug companies after this hearing. Hepatotoxicity is probably the most crucial unanswered question today. Do HIV meds cause harm to the liver for people with HCV or HBV, and does liver disease progress more quickly as a result of ongoing exposure to HIV meds. We don't really know the answer to this question.


For the very treatment-experienced patient the expectation of undetectable viral load may not always be achievable, so what is a satisfactory parameter by which to judge a new drug: CD4 increase, viral load reduction of 0.5 log, toxicities, a composite endpoint including some or all of these endpoints. These issues were also discussed yesterday. As you may know, data has beenemerging that when viral load increases oftentimes CD4s remain elevated for a period of time. This CD4 count elevation may be adequate to sustain health. Committee members felt that a VL reduction of a 0.5 log may be adequate for drug approval in an experienced patient. It appears that a number of studies have shown that a 0.5 log reduction does result in longer term clinical benefit.

NATAP is preparing a full report of this hearing which is expected to be forthcoming in the next few days.