Some edited excerpts from discussions by the FDA Antiviral Advisory Committee on the new drug application for Aptivus (tipranavir) 250mg capsules indicated for treatment of patients with HIV. 

 

May 19, 2005

 

Janet Englund (Committee Chair) – Welcome and turn off your Blackberries.

 

During the first half of the daylong meeting, safety and efficacy data and analysis concerning tipranavir/ritonavir (TPV/r) were presented, first by the sponsor (Boehringer Ingelheim) and then by the FDA. After clarifying questions from the Committee to the sponsor and FDA, the members of the Committee discussed several questions posed to them by the FDA. The Committee also voted on the key question of whether TPV/r was safe and effective in the proposed population.

 

Question 1:

Do the data demonstrate that tipranavir/ritonavir (TPV/r) is safe and effective for the multi-drug resistant HIV-1 infected population?

            If no, what additional data are needed to provide evidence of safety and efficacy?

            If yes, please address the appropriate population for TPV/r use considering the following:

                        - Limited inclusion criteria of the RESIST trials

                        - drug-drug interactions

                        - resistance information and patterns associated with optimal use

                        - safety considerations

 

 

Englund – What is your risk/benefit assessment of TPV/r given the data provided for safety and efficacy in the treatment of “heavily pretreated” HIV-infected individuals?

 

Richard Haubrich – The evaluation of safety is in the context of need. The biggest need in the clinic today is for patients with multidrug-resistant HIV. This is who we clearly need new drugs for. The risk/benefit ratio is different for these patients. The bar would be much higher for treatment-naïve patients. We are willing to tolerate more toxicity in these patients because we need new drugs. The promising drugs on the horizon won’t be available in the clinic on Monday morning. Most of the safety issues with tipranavir are reversible. While they are serious, they are manageable. But this is the reality of HIV care in 2005.

 

Maribel Rodriguez-Torres – I treat the complications. There is a need for a treatment for patients with resistance. Why not approve tipranavir in combination with Fuzeon? If not, then that should be a strong recommendation to use them together.

The prevalence of hepatitis C virus (HCV) is much higher in the clinic than in these studies. It is as high as 60 percent in our Baltimore patients. Simply measuring ALT does not tell enough about the severity of the disease. We need more biopsies. Did patients with high LFTs also have lipid elevations?

I’m also concerned about so many drug interactions. How will garden-variety treaters sort this out? I’m concerned about interactions with methadone – we need to know about this. Also, it is usually a different doctor who is treating with methadone. How are we going to deal with women who may be treated with hormones by another doctor? We need to give some paradigms that primary treaters can follow.

 

Kenneth Sherman – I appreciate what Dr. Rodriguez-Torres said. From the viewpoint of the hepatologist, it is the patients with end-stage liver disease that are the problem. We see lots of patients dying of end-stage liver disease with undetectable viral load.  This study has short-term HIV endpoints. We have seen nothing about prevention of opportunistic infections or long-term survival.

There was also a failure to address issues with early hepatic signals in these trials. A high proportion of patients had high liver enzymes but there was little attempt to get more biopsies to distinguish between those with different stages of liver disease.

Short term use of these drugs does not occur. This is the last drug for many people and it will be used for an extended period of time. The concern is not short term flares, but patients who cruise along with elevated LFTs for years. There is much more disease in coinfected patients. Sulkowski recommends more monitoring in coinfected patients. There is a poor understanding of liver injury in community and there is poor monitoring. But once you have symptoms, the game is over. You have to worry about liver injury well before that point. While we are transplanting more people with HIV these days, these patients with advanced HIV disease are the least likely to be transplanted.

Finally, drug-drug interactions that may lead to increased liver toxicity have not been well characterized to this point.

 

Lynda Dee – We have to look at the risks vs. the benefits. Tipranavir may have a limited role in the clinic. We’re talking about heavily pretreated patients. I was a guest here on the indinavir hearings. I think there has to be more education on what drugs you can take this with. I don’t want to harp on the safety considerations. I think the liver studies are very important. And I don’t why I never heard about rash before.

 

Robert Munk – I think we can characterize the patient population, but what about the prescribing population?  I don’t think this can be turned loose on the market without education. I’m concerned the package insert won’t cover all the issues.  If I don’t see a drug studied for interactions, I presume it is benign. But there are many potentially serious interactions with this drug that haven’t been studied yet. I’m concerned about whether tipranavir is ready for prime time.

 

Lauren Wood – The patients who need this will also likely have high baseline LFTs.  Clinicians will face situations with patients who need this who have LFTs above the upper limit of normal. What do they do? The company should generate data that tells us the magnitude of elevation over time. When you’re talking about grade 3 and 4 LFT elevations sustained over months and years -- that has much different implications for long term toxicity.  We didn’t discuss other parameters of liver function – coagulation studies. 

 

Scott McCallister (Boehringer Ingelheim) – Patients had AST or ALT elevations. Coags were not elevated.

 

Victor DeGruttola – There appears to be a patient population that has a favorable risk/benefit ratio for this dug. But how well can we predict who will have liver toxicity and who will have virological response benefit? It would be useful to put analyses together to classify how well you can predict toxicity. Regression analysis doesn’t do that. Who will respond best? Doug Mayers made a crucial distinction between mutations that have a direct causal impact and those that are merely associated with mutations that have causal effect. If the mix of mutations does not change over time, that’s fine, but that mix could change and the L90 could cease being associated with the causal mutations. Try to determine what the best classification is for patients who are likely to respond best. What is the chance of the individual patient to respond well?

 

Englund – I’m anxious to get drugs that will benefit my patients and I think I can follow and manage my patients.  We should stress this drug’s use by experienced HIV treaters, but that is hard to do

 

Gene Morse – One of the concerns I have is that these patients may be on 8, 12, 15 drugs at a time. And many have coinfections. You can identify who is most likely to be safe, but beyond that how will those patients be managed? How will long-term toxicity be followed?  We need studies.

 

Robert Grant – We’ve seen data that established that a sub-group that will benefit.  I’d like to hear more about the patients in the boosted PI study. We didn’t see efficacy data from the .51 study. The proposed language of the indication mentions salvage settings, but the people with the most resistance were excluded from the data presented.

 

Douglas Mayers (Boehringer Ingelheim) – A small percentage got a durable drop, but most began to fail afterward. We’ve seen good anecdotal information on the Fuzeon-naïve patient that combines them.

 

? – Why don’t we see short-term survival data or fewer OIs? Maybe we haven’t followed them long enough.

 

Englund – Let’s go around the table and vote. Safe and Eff for MDR population ? 

Wood – Yes

DeGruttola –yes

Rodriguez-Torres – No. Drug interaction studies are needed, as well as studies of histology and on the outcome of liver disease.

Munk – yes

Sherman – No as a rapid approval pending long-term data and clinical outcomes.

John Gerber yes

Ronald Washburn – No. There is a need for long-term efficacy follow-up.

Grant – yes

Veronica Miller yes

Frank Maldarelli – yes

Morse – yes

Edmund Capparelli – yes

Stephen Hall yes

Englund – yes

All yes votes came with concerns and reservations.  11 Yes; 3 No votes

 

 

Englund – Please discuss your highest degree concerns.

 

Hall – My concerns are with long-term outcomes and clinical management issues.

 

Capparelli – There needs to be a greater focus on use with other drugs. Few thymidine NRTIs were used and there was little variety in the background regimens.

 

Morse – My problem is that I can’t tell my patient how well this will work.

 

Maldarelli – Efficacy is evident because new resistance mutations emerge, but the durability of the viral response remains uncertain.

 

Miller – I am concerned with long-term liver toxicity. It is too bad that 48-week data was not available. We need more information on rash, including its clinical management and the role of CD4 count in predicting rash.  We should stress the need for physician expertise in using this drug.

 

Grant – I agree we should emphasize that the use of this drug in these patients should be restricted to experts.  I’m concerned that the evidence is not yet sufficient to justify the risk/benefit ratio in women, particularly women on birth control pills.

 

Washburn – I remain unconvinced that the risk/benefit ratio is acceptable based on a summary of data from a short-term, unblinded trial.

 

Gerber – A drug proposed in this advanced population--especially a drug with these interactions--should have some clinical endpoint. We also need to see information on interactions with fibrates and lipid lowering drugs.

 

Sherman – If pretreatment status shows elevated LFTs, then patients should be thoroughly evaluated for stage of liver damage.

 

Munk – More drug-drug interaction studies are needed as are more studies of treatment in co-infected patients and in women. Better characterization of resistance is also needed.

 

Rodriguez-Torres – I agree with Dr. Sherman to fully workup patients for liver disease.

 

DeGruttola – It is very hard to interpret the surrogate endpoints here – we need clinical endpoints. It is also important to make the best use of mutations at the start of treatment. A person with a 1.0 log drop, but still not undetectable, was counted as a success in this study. We need to know more about durability. We need to know how to classify those patients who are at most risk and those at low risk of toxicity.

 

Wood – It is also important to tell clinicians that there is no indication for this drug if a patient has susceptibility to other protease inhibitors and that this would be the only PI allowed in a regimen. Make that clear. It also needs to be reinforced that tipranavir needs to be accompanied by another active drug to gain the maximum benefit. Another priority focus for interaction studies are diabetes drugs. Also, look at additional oral contraceptives in women. That has to be at the front gate.

 

Englund- We are not trusting oral contraceptives at all.

 

Dee – What authority does the agency have to recommend education to doctors?  

 

Debra Birnkrant (FDA) – We think there should be adequate educational materials. We never seem to learn whether these educational program work or not. We never get a testing of the materials in a large group. But we don’t regulate this.  

 

Douglas Fish – The short follow-up of 24 weeks is a casualty of accelerated approval.

 

Princy Kumar – I have a different take on this compared to my hepatologist colleagues. As a clinician looking at who was treated in these trials, I see an extremely treatment-experienced group of patients. Of course we worry about safety in these patients. But all the concerns should not deter from the fact that at the present time this is one of the few agents we have that can work in this group. As Dr. DeGruttola asked, which patients can we give this to safely? No single agent is going to be durable if it is the only active agent. What can be done to allow other, more potent drugs down the line to be added to this drug?

 

Haubrich – My comments are closely aligned with Dr. Kumar’s.  There is no way this study could show clinical benefit since it was designed to let people drop out if they were failing. Doing a study like that would be infeasible because you would have to keep people on a control arm. I hope people will stay away from that idea.

 

Englund – My summary: We as a committee feel that the need for this drug is high, but the risks are high. We want to know about long-term durability. We need to know clinical incidence rates over the long term. We have questions as a group concerning management of toxicities; drug-drug interactions; about the lack of women in the trials. To hear that the ongoing studies are not preferentially set up to enroll women is troubling.  We need much more information on hepatic function follow-up and on the prediction of toxicity and rash incidence.

 

Jeff Murray (FDA) – I’ve been here 13 years and wanted to say something about clinical endpoint studies for registration. Over the years both investigators and participants have told us that such studies can not be enrolled and they don’t want to participate in them. You recall the ritonavir study that required patients to stay on a failing regimen until they developed an OI. No one wanted to wait to get sick. If you want clinical endpoint studies you have to be willing to wait until a patient has an OI. I thought this issue for registration had been laid to rest. I would like everyone to think hard about what it means to propose a clinical endpoint study in 2005.

 

DeGruttola. When we use surrogate endpoints it is because we believe they predict a clinical benefit. We have to come up with a model that says the effect on this surrogate means it will have benefit or it won’t. We can’t force patients and docs to do something they think is unacceptable, but we could collect clinical events so that we could compare those who got tipranavir from the start from those who delayed or didn’t start. Then we could maybe go back and put together information from all salvage studies and try to identify some important predictors so we can classify patients from the start. It is harder to do, but will we ever get the information we need any other way?

 

Birnkrant – I would like 10 minutes discussion on the indications

 

Englund – Let’s discuss the indication. Who should this be used for? It has been clearly stated that it needs to be used in those patients who have failed several PIs.

 

Miller – They many not necessarily have advanced disease because they may have acquired highly resistant HIV.

 

Grant – I would recommend PI resistance not PI experience as the indication.

 

Munk – It should specify multiple PI resistant virus.

 

Gerber – You need genotypic and phenotypic testing. You want to be sure the virus is susceptible to tipranavir.

 

Maldarelli – Getting both tests may be overkill – get the genotype first.

 

Grant – I agree that either would be adequate and I wouldn’t specify which.

 

Miller – If a resistance test is in the label, to what extent will that limit use because of reimbursement issues?

 

Englund – If it is not required then I will have trouble getting reimbursement.

 

Munk – The fact that it is required doesn’t mean that it will be reimbursed.

 

Mayers – Thirteen of the 50 states will not support genotypic or phenotypic testing.

 

 

 

 

Question 2

Given the data on transaminase elevations, please provide your recommendations for:

            TPV/r use in patients with underlying liver disease

            Monitoring and management of hepatotoxicity during clinical use

            Future studies

 

 

Sherman – You should biopsy patients who have abnormal liver enzymes.

 

Rodriguez-Torres – A study in HCV coinfected patients is needed.

 

Mark Sulkowski (BI Consultant) – In a patient with few options we are in the pre-HAART era. ARV toxicity is not unique to this patient. In HIV patients, 12-15% have grade 3 LFTs; if they are coinfected with HCV, then that percentage goes up. I’m concerned about the biopsy recommendation. It is not accessible to many patients.

 

Munk – I think it is unreasonable to put the whole question of HIV and the liver on this one drug and this sponsor. How do we deal with coinfected patients in general?

 

Englund – There is a concern that LFTs do not indicate liver true liver function.

 

Sherman – This study had a comparator arm and tipranavir had more liver toxicity. This indicates the tipranavir was more toxic than the comparator arm.

 

FDA – We also saw more liver toxicity in the healthy patients.  Please comment on using tipranavir in patients who are coinfected at baseline.

 

Englund – Is there a patient in whom you would not use this drug?

 

Sherman – As Dr. Sulkowski indicated, those with HBV/HCV have greater risk of having liver toxicity. We don’t exclude those patients from being treated. They plan to study interaction with IFN/ribaviran.  All of these are risk/benefit assessments in individuals. You look at each patient. But you’d really like to know what risk you are putting certain patients in.

 

? – There seemed to be a concentration relationship with liver toxicity.

 

Kumar – From the practical standpoint it will be impossible to get biopsies on people with mild LFTs. If we are required to get biopsies before using this drug, we will not be able to use it. We manage patients clinically and with lab measurements. I don’t see why we can’t do that with this drug. 

 

Dee – What authority does the agency have to make the sponsor educate about the problems with this drug?

 

Birnkrant – We could negotiate with the company about what would be in the letter they receive. That would be binding.

 

Daniel Kuritzkes (BI consultant) – There is no doubt about the hepatic toxicity, but I don’t think the RESIST study can serve as a study of this. It selected against preexisting or known toxicity of the existing PIs. It was not a randomized comparison of liver toxicity of this drug vs. the comparators.

 

Rodriguez-Torres – Patients with moderate to severe fibrosis should not receive this drug. It is different if a patient has 400 CD4 or has 5 CD4.

 

DeGruttola – I agree with Dr. Kuritzkes that because you are selecting patient who are doing well on their current PI that that bias exists, but …

 

Birnkrant – We still need input as to monitoring and management.

 

Fish – I envision myself doing this: I would not start anyone with LFT over grade 2, I would monitor every 2 weeks initially. I would not use in a patient with suspected cirrhosis – normal LFTs, but high platelets and albumin, for example. I would await data in patients with compensated cirrhosis.

 

Wood – Vigilant monitoring would have to remain in effect.

 

 

 

 

Question 3

The limited amount of data on females with HIV infection in the TPV program shows an increased incidence of rash in females. Please provide your recommendations for:

            Investigation of this safety signal in future studies with TPV.

 

 

Miller – I’d suggest that older women not on birth control also be studied. Women on hormone replacement therapy.

 

Wood – It should recommend in the label how to manage individuals with rash. It is encouraging that there is no evidence of SJS.  But do you treat through or discontinue and rechallenge?  It would be a disservice to deny access of the drug to women.

 

Englund – After so many years on this panel, I’m getting ready to recommend that a drug not be licensed if there is so little data on women.

 

Dee – How many women are likely to enter future trials?

 

FDA – The naïve study is fully enrolled and has about 20% women.

 

 

 

Question 5

Given the high inter-patient variability in TPV exposures following fixed doses and exposure (blood levels)-virologic response relationships, could a biomarker such as Cmin/IC50 be used for the individualization of TPV/r therapy? Please discuss the studies that would supplement the data presented today. 

 

 

Grant – The concept is promising but premature to recommend for monitoring. We need a validated protocol for collecting samples. We need assays for measuring the drug level and interpreting protein binding. The phenotypic assays have not been validated. This is not ready for management for individual patients at this time. It will require more study of timing, drug resistance phenotype, drug level assays and adjustments for protein binding.

 

Hall?  – I’m not sure protein binding is a necessary component of that. ACTG is conducting a study with 30 sites. There is a general lack of enthusiasm for TDM. But if I were a patient with a cut off of 3, then I’d want to know what my concentration was in relation to susceptibility. But there are other ways to get numbers for that denominator. It should not be required but some patients could benefit.

 

Dee – The companies won’t do TDM unless there is a message from on high that it is needed. They use it in Europe and people in the community here are wondering if it could help them.

 

Jonathan Shapiro (BI consultant) – TDM is an interesting concept  and  we see correlations between concentration and resistance but it is challenging to turn this into a useful tool without damaging the patient. How do you determine within two weeks if a patient is tolerating the drug at that dose so you can increase it? Two weeks may be too soon to increase a dose. Also reducing a dose too soon is also risky. TDM needs much more study before it can be mandated.

 

Birnkrant – We recognize that the assays available today should not impinge on the approval of this product. We want to know what to ask companies to do on the next trial.

 

Gerber – I see potential for studying this as a way of reducing toxicity. But what happens if your concentration is changing? I’ve been against TDM for these reasons. Also the labs don’t have quality controls.  Efficacy might be more difficult to do because you need phenotype.

 

Maldarelli – Difficult to predict usefulness. Making a recommendation based on this is difficult.

 

Capparelli – There is also intra-patient variability. So anything used must measure it well. I don’t think we’re going to implement a trial that could show this conclusively.

 

Englund – If there were a focus it would be in more difficult patients.

 

? – Isn’t it Cmax that should be looked at for toxicity?

 

Gerber – Cmax is impossible to measure.

 

FDA – There should not be an issue between Cmax and Cmin.

 

 

 

Question 7

Please discuss and recommend future study designs /data acquisition for the heavily pretreated population.

 

 

Fish – A phenotype would be helpful in addition to genotype.

 

Dee – I think BI is to be commended for studying these patients and for having that 8 week escape clause. But we are now stuck with that 24-week trial depending on 8 weeks of data. The community tortured the agency into having that salvage meeting. The agency is not opposed to using two investigational agents. Unless other sponsors want to sit on this hot seat, they need to get together with other companies to work together.

 

Wood – It doesn’t need to be proven that only adding one active drug is not effective. So I’d stress that the companies work together. I’d plead for the aggressive inclusion of women.

 

DeGruttola – It would be useful to have some clinical data in these studies; to have full suppression below 400 copies; to have analyses of baseline mutations that also predict durability.

 

Gerber – The only potential concern I have in trials without a real comparator arm, is if you have a drug that is more toxic than the usual drug. How do you evaluate how it performs in the overall scheme of things? We want to be able to understand how the toxicity affects overall survival.

 

Grant – We need more Africans, more women and more Asians.

 

Capparelli – Enrich the population in terms of ethnicity, race and pediatrics. I appreciate that we lose information after they switch over, but can we then investigate alternative dosing or other information that helps clinicians manage patients?

 

Hall – I think it would be helpful to focus on the outliers. Do the people with low concentration have something in common?  P-gp genotyping should be a part of future studies.

 

Mayers – We learned a lot from this trial. If I can ever do a placebo-controlled trial in his population, I will.   It becomes an 8-week immediate vs. deferred tipranavir study.  So how can you distinguish bad things when they happen when they’re both on your drug?

 

Also we would ask dropouts to stay for monitoring to see what happens to them in the long term. In time the comparisons get worse. How do you keep the options attractive to the patient and not blow their last chance? It was a good trial for efficacy, but not the best for safety.  

 

We are doing everything possible to make tipranavir available to other companies for studies. But their drugs all use the same liver enzymes and we need to do the drug interaction studies.

 

Birnkrant – Thanks everyone.