The complete panel transcript is now available!
PANEL HIGHLIGHTS
Opening Remarks:
COLMAN: (00:07 – 00:16)
Thank you. I’d like to welcome back all the folks that were here yesterday, and also welcome those of you who are joining us for the first time.
As you’ve heard, we’re going to be talking about a new drug called lorcaserin, which is a 5HT2C receptor agonist, which is being developed for weight loss. I want to mention one particular question that we’ll be asking, and that has to do with potential for valvular heart disease. Given that this class of medication, and data that suggest that activation of the 5HT2B receptor is responsible for the valvulopathy that was observed with fenfluramine and dexfenfluramine, we’re obviously very concerned with this compound since it is somewhat of a chemical cousin to these drugs in terms of its interaction with the 5HT2B receptor.
At the beginning of the development plan, the Division spoke with the Company. We all agreed that echocardiograms would have to be done to best evaluate for the potential for valvulopathy. The question was, what was a reasonable sample size. I believe the Company initially proposed a non-inferiority margin of 2; and we were uncomfortable with that degree of uncertainty, we felt 2 was a little too high. So we had continued discussions, we ultimately settled on 1.5. I think that necessitated that their sample size go from 3-4000, to 7000. But we felt that that was a reasonable place to start the process. We were in the dark in some areas, for example we had no idea what the incidence of valvular heart disease would be if you took a group of people at baseline who did not have it, followed them for a year on placebo, and then observed them to see how many actually developed FDA-defined valvulopathy. So there were a lot of unknowns that we weren’t aware of, and the only way to get some data was to start and just see what you get.
So I want to make it clear that the 1.5 was an initial goal-post just to allow us to proceed and get some data. From our standpoint it was never: the non-inferiority margin is 1.5, so if you come in at 1.4, that’s a game-winner. We always anticipated that it would be, we would look at the degree of uncertainty with the valvulopathy in context of the drug’s efficacy, potentially other benefits of the drug, and other potential risks of the drug. So I don’t want to convey the message that 1.5 was the win-all number. Because, again, we were working in the dark in many areas here.
Um, do you want to go over the questions? (to meeting chair)
DR. GARY WILLIAMS PRESENSTATION OF CARCINOGENICITY
00:22 – 00: 41
DR. WILLIAM SHANAHAN PRESENTATIONON CLINICAL EFFICACY
00:41 – 01:00
DR. CHRISTINE ANDERSON PRESENTATION ON LORCASERIN SAFETY
01:01 – 01:29
DR. STEVEN R. SMITH PRESENTATION ON RISK/BENEFIT
01:30 – 01:25
PANEL QUESTIONS FOR SPONSOR
DR. WEIDE: Question about LDL cholesterol worsening in both treatment & placebo arms, despite weight loss – though less in treatment arm
DR SHANAHAN: No explanation; it happens commonly.
DR WEIDE: Cancer risk: Slide C82 shows 2 lung cancers & 2 multiple myelomas in drug arm but none in placebol
DR ANDERSON: That was just partial listing of neoplasms – full listing of all neoplasm terms shows no imbalance – no imbalance in tumor types associated with rat data (breast cancer, brain cancer) – no neurological system tumors
DR. DOUGLAS (01:39): Question regarding 4 MI’s and adjudication
DR ANDERSON: Explained adjudication process utilized by Arena.
DR. DOUGLAS: Question regarding statistical power calculation in Slide 94.
DR. KAUL (01:46) “While I don’t have any overt concerns about valvulopathy, I don’t necessarily agree” with Arena data analysis. Makes detailed argument about correct way to make these calculations.
DR SHANAHAN BRINGS ARENA STATISTIAN INTO DISCUSSION.
DR. COLMAN (01:52): ( Interrupts discussion) “ I was afraid this was gonna happen. That’s why I mentioned something earlier. When the Company … discussed what was appropriate … given what you originally proposed 2, we thought that was too high, we thought 1.5, that requires 7000 patients. It seemed like a reasonably good number of patients to start with. [Emphatically] However, it doesn’t mean, if you come back, your upper confidence limit number is 1.49, that, Whoo, we win! That’s the end of the story!
“It’s a place to start, and I want people to look at the boundary, and take into consideration everything else we know about the drug. Because the issue is how much uncertainty with the echocardiogram data do you feel comfortable with, when you take into account the drug’s efficacy, the other safety issues. So I don’t think we should be getting hung up on what the upper bound is numerically. We started off at 1.5, because it was a reasonable place to begin in the process. But I think it’s misplaced to focus on that number. The question is, how much uncertainty from the echo data are people comfortable with, when they take all other available data into account.” (end 01:54)
DR. PROSCHAN (01:54): “Yeah, I don’t want to belabor it because there’s not evidence of terrible harm in valvulopathy. I still, I’m sorry, but I absolutely disagree” (with Arena’s statistical analysis. Detailed explanation of what Arena did wrong in making calculations.) “So I absolutely disagree that that’s an appropriate analysis. But, given that, there’s still other analyses that show, at least, it’s in the right range. It’s just that I don’t think they actually win on the 1.5 – I do think it’s a little bit more than 1.5.”
(end 01:55)
DR. VELTRI: Question about LDL, other apolipoproteins; since 33% had dyslipodemia, what was breakdown in statin use; what was effect on LDL on those who were not on statin.
DR SHANAHAN: Don’t have statin breakdowns; shows slide showing changes in hypertension & dyslipidemia meds during trials – favor lorcaserin, but only a trend.
DR GARDNER: Would like to ask Dr Williams about breast cancer in mice and potential risk to in utero children of mothers taking lorcaserin; potential development of breast cancers in the children of mothers who took lorcaserin.
DR. WILLIAMS: Question should go to Dr. Shanahan for developmental toxicology data.
DR. SHANAHAN: No effects seen on reproduction or birth abnormalities in rats and rabbits. Didn’t follow “pups” later on. Dr. Williams’ presentation implicated a vulnerability to tumor development during adolescence: can’t extrapolate that to developing fetus.
MELANIE COFFIN: Question about Adverse Effects (AEs), especially amnesia: what constitute mild and moderate amnesia.
DR. ANDERSON: “Amnesia” term did not refer to “medical definition of amnesia.” (Amnesia term used from ?some other list of terms?) Referred to minor forgetfulness episodes, like forgetting where keys were. All events resolved, and typically were not considered to be related to lorcaserin by the study investigators.
DR GOLDFINE: Question about relative receptor potency; discrepancy between 2004 and 2009 data.
DR. SHANAHAN: Produced slide relevant to question (Dr. Goldfine wanted look at more closely over break).
-------------------------
BREAK
02:22 – 03:42 FDA PRESENTATION
PANEL QUESTIONS FOR FDA
DR. THOMAS (03:42)
We’ll now take clarifying questions from the committee for the FDA. I’d like to remind the committee that during this session of the meeting that the questions should be focused specifically to the FDA or their presentation. If time remains, we will go to questions for the sponsor, but please restrict your questions to the FDA for the current time.
The Felner-Colman Exchange (03:46)
DR FELNER (03:46)
A question for Dr. Colman. If – assuming you could throw out all the adverse events data, assume no adverse events—based on the 2 criteria that you have for benefits for the drug, meaning the 5% weight loss in the first, which they didn’t meet, they only met the second criteria-- I assume this wouldn’t be coming to the meeting here, but how would the FDA have looked at it just by that very minimal weight loss, compared to the other drugs that are available? How would you have looked at that?
DR. COLMAN:
I’m not sure I can answer that question. Are you asking how would we view the efficacy without any consideration of potential safety issues?
DR FELNER:
Well, let’s assume there are no safety issues. Let’s assume that. And I know you have the two criteria there, either meeting the 5% or more above placebo, or I think more than 35% weight lost – would you have been able to say that’s efficacious just based on that, the minimal weight loss? What I’m trying to figure out is, we know there is some safety concern obviously, but I want to know what’s felt on the FDA’s end of the efficacy, because if you’re not completely satisfied with the efficacy based on the 2 criteria that you have, I mean I think it makes things a lot easier to think about. Right?
DR COLMAN: (Colman gulps, laughs uncomfortably; other laughter)
There’s a little bit of pressure on me there, huh?
Well, the fact of the matter is they did satisfy one of the criterion in our guidance. You know (small laugh), it had to be approximate in the one study, and I believe I was quoted as saying it met it by a slim margin, so…(trails off)
DR. FELNER:
I mean, that’s what I was getting at. I mean, you even said “by a slim margin” on the efficacy issue, so I wanted to …(Colman nodding, Felner’s comment trails off)
(end 03:48)
LUNCH BREAK
PANEL DISCUSSIONS OF OFFICIAL QUESTIONS
Question One – Efficacy
Question Two – Valvulopathy
Minute 6:33 – Michael Proschan, Ph.D (Mathematical Statistician, NIH)
“I think they did about as much as they could do [to rule out valvulopathy.] The relative risk is close to 1… I’m not so sure you can ask them to do more than they’ve done so far.”
Question Three – Neoplasm Risk
Minute 6:38 – Jessica Henderson, Ph. D. (Acting Consumer Representative)
“I didn’t have a problem with the efficacy of weight loss but this neoplasm risk is my #1 concern for risk. Problem within this area was that there was no safety margin in the female rats. I’d like to see if we can have more animal studies to see if that can be determined, safety margin, and we’d definitely need to keep a registry if this drug was approved. “
Minute 6:39 – Edward Gregg, Ph.D. (Chief, Epidemiology and Statistics Branch, CDC)
“I’d have to admit; I was kinda ‘spooked’ by the same thing and particularly the assessment that there were mammary tumor risk all the way across spectrum of dosage and the statement from the FDA that there was no safety margin identified. I do feel a bit unqualified on how we make the judgment on the risk from rat studies to humans. And so I’d actually love it if there were a bit more discussion around this and somebody could either make me feel better or not.”
Minute 6:40 – Jacqueline Gardner, Ph.D., M.P.H. (Professor, Dept of Pharmacy, University of WA)
“I concur completely and don’t know how to make that leap and also don’t know how to suggest studies that would help us make that leap and so someone who could help us do that. I do know that even if we decided to proceed with adequate or what we perceive to be adequate warnings it is going to be very difficult for women to hear there might be an increase in breast cancer with a weight loss drug. I don’t know how we could do it. So how do we learn more about this and what we would recommend to the sponsor and the FDA?”
Minute 6:41 - Michael Proschan, Ph.D (Mathematical Statistician, NIH)
“I’d also like to echo that sentiment. I really feel lost on this. When the company made their presentation on this point I thought, Gee that makes sense. And then when the FDA made their presentation, I thought, Gee that makes sense too. And you know I absolutely have no idea how to translate from animals to people. So I feel particularly lost on this one.”
Minute 6:41 - Jodi Segal, M.D., M.P.H. (Assoc Professor of Medicine, John Hopkins Univ)
“Estrogen is still on the market. Some Drugs cause a little elevated risk of breast cancer but there are still a use for them. Can I ask the FDA a question? Is it sometimes better to approve drugs even if you have to pull them later, is there a certain number of drugs that you are just going to have to get rid of later? [(laughter) Dr. Kaul responds with, “especially with weight loss drugs.” (making reference to the Meridia panel he participated in the day before)]
Curtis Rosebraugh, M.D. (FDA – Director of Office of Drug Evaluation CDER) responds with, “Yeah, well, I guess if you voted to approve it, that your name would be on the vote thing and then if we yanked it, how would you feel about that? So, ya, I can’t answer that question.”
Segal: “But with some drugs, you won’t be collecting the data until post approval possibly. That’s why there are post-approval studies.”
Minute 6:43 – Pamela Douglas, M.D., M.A.C.C., F.A.S.E., F.A.H.A. (Ursula Geller Professor of Research in Cardiovascular Diseases, Director, DCRI Imaging Program)
Asks a question to the FDA about identifying the neoplasm risk in more than one species of animal.
FDA response --Todd Bourcier:. “These are 2 year studies. They take a long time. I think the hazard has been identified in the rats and I don’t think if let’s say we do hamsters for 2 years and that comes out clean, that’s gonna really make us feel very warm and fuzzy about it. I think, regarding the mammary tumors, our fundamental disagreement is in how they arose. And we understand the argument about prolactin but we just don’t feel like the data supports that sufficiently. If we were able to come up with a set of experiments that is able to clarify how this happens, is this a tumor promoter, is this working through direct or indirect mechanisms, and were to define that a little better, then I think we are in a much better place to say how this is relevant to human biology. If it turns out in fact that prolactin is involved, well we have a lot of anti-psychotics out there that cause mammary tumors in rodents…. Then we have something to work from. It is a different indication but we have experience with that. Right now, we don’t feel like we know how it works. And that is scarier than knowing how it works. So I think there is a path forward and this pertains not only to the mammary tumors but also to the astrocytoma, although there is a safety margin there which reduces our concern.”
ABSTENTION ISSUE:
THOMAS (6:58):
As I mentioned yesterday, please try not to vote Abstain, because the FDA not only values your comments after the vote, but also the actual vote itself. And there’s probably a lot of peer pressure from the panel not to vote Abstain.
DR. GOLDFINE (6:59)--Question regarding whether they could vote “provisionally” based on current information & wait for diabetic study (Bloom DM) results to see if confirmatory.
DR ROSEBRAUGH, FDA DIRECTOR OF
Just so you kinda understand how this works, we’re on a time clock, and we’re not gonna have time to look at that data... And I just kinda want to go along with the Abstain- much like we can’t wait because we’re on a time clock, we can’t abstain down on this end either, so it is nice if everyone make their vote based on what you think should be done.
PANEL VOTE REMARKS
SEGAL: YES (07:03)
Met efficacy; met valvulopathy risk; cancer a little unsettled. Depression and dissociative problems not a issue. Need active surveillance, large post-approval studies, registry.
GOLDFINE: YES (07:04)
“I found this a very very hard decision to make.” Very small magnitude of weight loss relatively, “but I think that that can be clinically quite important.” Surrogate endpoints moving in the right direction “important to me.” Very concerned with limited population; should be used in “multidisciplinary teams that counsel on behavioral, lifestyle, and other sorts of factors” including breast cancer risk. Breast cancer risk is “very concerning.” Need post-approval trials; more detailed analysis on cancer mechanisms in animals.
GREGG: NO (07:05)
“This was a very difficult vote. I think this was a promising drug. I think the weight loss, though not great, was acceptable. I thought the profile of risk factor benefits was encouraging, actually, and there wasn’t any clear evidence, actually, that this was a dangerous drug. That said, there was enough doubt raised about some of the risks, lack of understanding of some of the mechanisms underlying those risks, so that it forced me to take a harder look at what I saw as the crux of the benefit, which would be on the magnitude of the risk factors, not necessarily the spectrum but the magnitude of the risk factor benefits. After you think about that, and take into account loss to follow-up& so on, and when I did it that way, the risk to benefit just didn’t quite meet that bar. On balance, I think the drug is promising, but it’s not quite there yet, in terms of showing the evidence.”
PROSCHAN: YES (07:06)
“I voted yes; this was also a very close vote for me. I thought carefully about whether yes or no was the right answer, and just by a small margin came out on yes. I agree with the comment about part of it was because of benefit in some surrogate endpoints. It’s not clear if those were because of the weight loss, but whether it’s because of the weight loss or something else, that’s a good thing.
I definitely agree though, that there are doubts stemming from the fact that the population was pretty restricted; and from the troubling results in some animals. But again, I don’t know how to translate that to humans, and I thought the evidence in people did not really raise a huge safety flag in my mind.”
THOMAS: YES (07:08)
“I voted yes. I have to say this was one of the harder decisions to make, I was vacillating between yes and no for quite a while in preparation for this meeting and even throughout this day. And the reason for this is just that the actual efficacy is very small, and I think that’s not a problem with the sponsor – that’s actually what they were required to do… If we want more effective drugs for weight loss, we have to set higher limits.
The concern was this efficacy was in a very selected population that was generally healthy for an obese population. And when this is translated into clinical practice, the weight loss is going to be much smaller, and then you wonder is it worth taking compared to placebo.
However they did meet the criteria that the FDA has put forth. I think they did their best in trying to assess many of the other risks that were of concern with this type of agent, including valvular heart disease.
I am concerned about the tumor issues, and think there has to be some type of registry, very significant follow-up if approved, to look at these agents. And if there are any animal studies that could be done, they probably should be started now
One reassuring thing is that, unlike some other agents, the benefits of weight loss, even though modest, did seem to correlate with benefits in what we think are important factor- glycemic control, blood pressure, heart rate. So I think the safety signal, even though this was a selected population, was better.
I do have to say one thing, I think you can say, ‘Damned if you do, damned if you don’t.’ If you studied a riskier CV population, we’d probably would talk about, oh, we don’t have enough data and what’s the benefit and we may not see as much efficacy. I think the point Dr.Goldfine made earlier that there are patients that you want to treat before they get these diseases, if this drug is effective in that population, maybe it should be limited to people without some of these diseases, and that could be a good use for this agent.
FLEGAL: NO (07:09)
“I also vote no, and I also vacillated until the last minute. I think this is a promising drug, and it fills a niche. The weight loss is small—I found it encouraging, however, that actually the change in biomarkers pretty much tracked with the degree of weight loss in both the treatment and the placebo arm, which suggests that’s really is how this is operating.”
Unanswered concerns about the risk that really need to be dealt with. And also we need to how this would benefit the population that stands to benefit more from it. “ So, on balance, I voted no, but again, it was a close call.”
WEIDE: NO (07:10)
When I read over all the stuff before the meeting, I was most concerned about the tumors and, although I’ve been told that the cause of cancer has been identified and it’s rat abuse [a joke—laughter], I don’t know how that correlates to the human. I was encouraged by what we saw that there wasn’t really a difference between the placebo group and the active group; but again, the population was limited, the time-span is limited, some tumors take a long time to grow. Having said that, what bothered me the most was that there was only about 3% -- a little more than 3% difference between the placebo and the active group; and the limited population that was studied. That was just overwhelming for me, I just couldn’t get over that we didn’t have a broader population that was representative of the people that are most likely to get this. So I can’t see how we can do that. I’m also a little disturbed that the study with diabetes that is coming only has 600 patients, and if it’s one to one, that means 300 diabetics; and even if you have a good retention rate, we’re not going to have an N of 300 and that’s just too small, period. So I’m not sure what we’re going to get out of the additional study. So I’d encourage you to do a broader study, and hopefully show us that we get more bang for what’s going on, and I would like to know the mechanisms of the breast cancers at least, as well.
FELNER: NO (07:12)
A lot of my comments are very similar to what Dr. Gregg had said. I really have a lot of difficulty, just like a lot of other people on this committee. I think the sponsor did a great job. They performed the study, they met the criteria that were asked of them; but I think when it comes down to it, I didn’t really have a lot of issues with the risk. But obviously a lot of other people here did, and it made me look at it a little more closely, and it made made me look back at the real benefit, at least today, of this drug; and I just didn’t see it as being that efficacious for even the group that was studied.
And what bothered me I think, a little bit more about looking at it was the second year of the data, when they didn’t maintain weight as we’ve seen in some other drugs. And if they would have continued past two years, it may have even been right back to baseline, even on the drug. And so these things concerned me, but I think there’s still a possibility of getting this through, from the sense of doing some more pre-market studies. I agree with Dr. Weide that we’re probably gonna need more than the 600 diabetics. That was the real key to me was that I had a lot of trouble seeing the real benefit …I hope they can get the work done and have it come up again.
HENDERSON: NO (07:14)
I agree with all the comments that this is a promising drug, and I would encourage the sponsor to re-apply. I loved the quality of life data, that’s one of my complaints when there’s not the patient perspective there, and that was there. But I voted no because of too much uncertainty – too much uncertaintly about the brain and breast tumor development, and too much uncertainty of the real life population this would be used in. I do think this is promising, but just too much uncertainty at this time.
DOUGLAS: NO (07:14)
I voted no because of concerns about marginal effects in a selected population, with continuing concerns for side effects. And that aren’t completely put to rest. The argument that there isn’t much else in this space and there’s an urgent patient and public need – I don’t think really mitigates the concern of putting out a drug that doesn’t do enough, and may do harm on the market.
KAUL: NO (07:15)
I voted no. I really didn’t think the current portfolio of evidence was enough to persuade me that a desirable benefit/risk has been demonstrated. Highly selective population that is not representative of the real world was chosen. I mean there were 42 exclusions in Bloom and 35 in Blosson. And disease conditions that cluster with obesity were either excluded or minimally included, and as a result, in my opinion, the benefit was likely overestimated and the risk was likely underestimated. I’d like to echo what Dr. Douglas said, that in our desire to offer anti-obesity therapies to our patients, we should not lower the bar or our expectations. I think it’s important for us to get it right the first time around.
COFFIN: YES (07:16)
I voted yes, and I do want to applaud the sponsor for including some diversity in the ethnicity of the patients, and I hate to say it, but I wish they were sicker. I wish they were sicker obese people that were part of the study. That said, it did meet the requirements of the FDA. I’m not pleased about the weight loss being only 3%, and it is only 3%.
GARDNER: NO (07:17)
I voted no for some of the reasons that have been discussed here. I do applaud the sponsor as well, for diversity and for looking at other endpoints besides just weight loss. I think my concern about the tumors- I was encouraged by Dr. Bourcier’s suggestion that maybe there’s a way of shedding some light on this going forward. We’re also waiting for some data relating to diabetes, and I think that the line in the question that stopped me, in particular, was when used long-term in a population of overweight people. My role in these committees has been to try to think about what can be done in a post-market to try to improve the risk management or risk/benefit profile. I guess we have failed to do that often enough that I’m not at all confident about the ability to reverse problems in the post-marketing environment; and I guess I agree that I’d like to get it as right as we can the first time.
CONNOLLY: NO (07:18)
I voted no. I echo the comments of many of the individuals here. The selected population studied demonstrated modest efficacy with this medication and in my opinion the potential risks of the medication outweighted the potential benefits. I also was very concerned about the “long-term” phrase in the question, and for that reason I felt that the potential risks outweigh the benefits. Therefore I vote no.
DR THOMAS: Does the FDA have any more comments? Dr. Colman, any comments?
COLMAN: No. I feel as though I should apologize for putting you through this, um, unpleasantness. We certainly do appreciate your input and your time. And we do go back to the office and discuss the comments you’ve had and the suggestions you’ve made. And you should just be grateful you don’t have to do this everyday for your living. And I do want to thank Arena, I know they worked extremely hard, I’m sure this wasn’t the vote they had hoped for. So again thanks to everyone.
PRECIS OF PANEL REMARKS ON VOTING DECISION
I went through my own transcript of the reasons the AC members gave in support of their decision, and pulled out these remarks, the number of panelists making them, and a breakdown of their vote.
Reference to cancer risk or general risk (which seemed to basically refer to cancer): 10
Yes: Goldfine, Proschan, Thomas
No: Gregg, Flegal, Weide, Henderson, Douglas, Gardener, Connolly
Study population too limited/too selective: 8
Yes: Goldfine, Proschan, Thomas
No: Flegal, Weide, Henderson, Douglas, Kaul
Negative allusion to small or modest efficacy: 8
Yes: Goldfine, Thomas, Coffin
No: Flegal, Weide, Felner, Douglas, Connolly
Efficacy basically acceptable/ good enough: 6
Yes: Goldfine, Proschan, Thomas, Coffin
No: Gregg, Felner
“Very hard” or “very close” decision: 5
Yes: Goldfine, Proschan, Thomas
No: Gregg, Flegal
Called “promising” or made reference to future approvability: 5
No: Gregg, Flegal, Felner, Henderson, Gardener
Offered compliment to Arena: 5
Yes: Thomas, Coffin
No: Felner, Henderson, Gardner
Mentioned “long-term” use indication requested as problematic: 2
No: Gardner, Connolly
Negative allusion to second-year data: 1
No: Felner
Reference to valvulopathy (positive): 1
Yes: Segal