Hiatt did nothing at Adcomm but immediately and persistently attack Vascepa's safety. He even accused Colman of glossing over it. He used phony statistics and had to receive an explanation of how ridiculous it was. The only advocacy was to commend Amarin for Reduce-It. He asked for an explanation of the FDA's agreement SPA with Amarin and how it could be disregarded:
10 DR. HIATT: Thanks very much. I have a
11 question about your safety data from ANCHOR. The
12 designs of both ANCHOR and MARINE were short-term,
13 small number of patients, and therefore the number
14 of safety events is rather sparse.
15 If you look at your slide number 58, the
16 total bleeding events, there's a numeric excess on
17 the 4 grams versus placebo, which by definition,
18 none of these become statistically significant. So
19 I think the idea of even introducing a
20 statistically significant argument doesn't really
21 help us much. But really, I think 95 percent
22 confidence intervals around the difference, the 1 rate difference between the 4 grams and placebo,
2 would be helpful.
3 Can you provide the 95 percent confidence
4 intervals around that difference?
9 DR. HIATT: I think the approach to the
10 conclusion that this is a safe therapy I think
11 needs to be challenged a bit, and the absence of
12 evidence doesn't really support that. So knowing
13 that the differences in event rates, the
14 2.6 percent versus 1.7 percent, and the upper bound
15 of the 95 percent confidence interval is around
16 that difference, would give us some sense today of
17 what the potential upper bound of the excess risk
18 for bleeds might be. And that might better inform
19 us than to compare 2.6 percent to 1.7 percent.
20 If you don't have that now, could you
21 provide that later in the day?3 DR. HIATT: Okay. Then a similar question
4 on slide 61 around the A1c values. So here the
5 nominal p is .089. Once again, on the right,
6 looking at a numeric imbalance, if you will,
7 suggesting that glycemic control could be worsened
8 by 4 grams per day -- again, the magnitude of that
9 excess was supported by an increase in fasting
10 blood sugar concentration. An estimate of that in
11 a larger population would be helpful, too.
9 DR. HIATT: So just to be sure I understand
10 your comment, I'm not talking about the point
11 estimate. I'm talking about the upper boundary of
12 around .2 to .3 percent increase in A1c. Do you
13 think that's clinically significant?
8 DR. HIATT: I'd like to ask the FDA whether
9 you think the ANCHOR safety database is adequate.
10 And the reason I ask that is that the proposed
11 label is to treat a surrogate endpoint, as you say,
12 with the presumption that there's clinical benefit.
13 But the issue on the table today is whether we
14 should approve something as a surrogate endpoint.
15 In my mind, if that's the question, then the
16 safety database has to be really good, and it
17 should rule out things, as were mentioned earlier
18 today, such as worsening glycemic control,
19 transitioning people from prediabetes to diabetes,
20 or causing major or minor bleeds.
21 The background in both the sponsor and your
22 talk don't really say much about it, and I just
1 want to know if you all think that it's adequate
2 for approval for this indication.
13 DR. HIATT: I get that. That's not my
14 question. My question is, at the end of the day,
15 we're going to judge something on risk and benefit.
16 Right? So clearly the drug significantly lowers
17 triglyceride levels, and you focused on efficacy,
18 and that is fine.
19 My question is, from a regulatory
20 perspective, do you think the safety database is
21 adequate? You didn't even present it. And I'm
22 asking you to judge collectively why we're here
1 today, if we're trying to prove something on a
2 surrogate endpoint -- not a clinically relevant
3 endpoint, a surrogate endpoint -- with a really
4 sparse safety database. And we can discuss what's
5 concerning about it later in the day, and I think
6 we probably will.
14 DR. HIATT: There were several questions
15 from this morning that were pending potential
16 sponsor review, such as providing the relative risk
17 and confidence intervals around bleeds in the
18 ANCHOR study. Could we maybe find out if that
19 information is available before we discuss this
20 question?
7 But I want to ask you, again, do you think
8 the safety has been adequately described to approve
9 this label indication?
15 DR. HIATT: That's what I figured. Just to
16 point out, though, it's a short exposure. It's not
17 many patients. There aren't many safety events.
18 And therefore, you're assuming the absence of
19 evidence is evidence of safety. I just am asking
20 that question. I guess I can table it for now, but
21 that's my concern.
16 DR. HIATT: I disagree. I think both are
17 important. But the perspective here is not to get
18 lost in what's up here. We're looking at an event
19 rate of 1.7 percent on placebo, and 2.6 percent on
20 4 grams per day. That's roughly a 50 percent
21 increased risk.
22 DR. HARRELL: I don't know how you say that.
1 DR. HIATT: What?
2 DR. HARRELL: I mean, the confidence
3 interval for that goes from essentially zero to
4 infinity.
22 DR. HIATT: Okay. So consistent with many
1 previous meetings of many advisory committees, the
2 idea of understanding safety when you are looking
3 at an endpoint that is, for example, in this case a
4 surrogate endpoint, and you therefore expose
5 patients to a short interval, 12 weeks, to find out
6 if the lipid parameters got better or not -- and we
7 know that -- is the same approach that you take in
8 understanding a diabetes drug; does that improve
9 glycemic control? A little different in a weight
10 loss because that takes a bit longer.
11 But typically, these are populations that
12 are at risk. We agree with that. These people are
13 at risk for cardiovascular events. So by
14 definition, there's a huge data deficiency here.
15 There's sparse data around the actual risk of a
16 bleeding event, which I'm saying is really
17 uncertain.
18 I'm not -- hold on -- I'm not landing on any
19 particular thing. But the upper bound of seven-
20 fold increased risk is only the high upper bound
21 because there are so few events. So it is a
22 numbers game.
1 But the point of the question is, do we have
2 any confidence that this drug increases bleeding
3 risk or not? And I would posit that you can't say
4 no. You can say, I don't know, and it certainly is
5 a numeric trend for an adverse effect here.
6 I'm just trying to understand what the
7 boundaries around that adverse effect are. And if
8 they're really, really broad, then I just get more
9 concerned. And if they're really, really tight, I
10 at least have some confidence to say to a patient,
11 if you take this drug, your triglycerides will drop
12 significantly and oh, by the way, your bleeding
13 risk might go up by 50 percent. And I guess I
14 can't have that conversation today because of the
15 gross uncertainty around those numbers.
4 DR. HIATT: I completely agree with that.
5 So in other words, can you tell me definitively
6 today that this drug does not cause bleeding?
7 That's my only point.
13 DR. HIATT: It doesn't appear to be an
14 interaction. It's the kind of expected, maybe a
15 little low, use of antiplatelet therapy in a high
16 risk population. But again, the data to support
17 aspirin for primary prevention aren't very good, so
18 I don't see gross imbalances here. And it doesn't
19 suggest that there is a greater proportion of
20 patients taking 4 grams a day that are also taking
21 a platelet inhibitor that might increase their risk
22 of bleeding. And given the size of the study, I
1 wouldn't have expected it to be gross imbalanced.
2 I don't see an interaction here.
3 The only point was an earlier point to
4 Dr. Roberts and the FDA, is that you seem to gloss
5 over the safety of this drug. And I just wasn't
6 sure why. And I don't know if there's a bleeding
7 risk or not.
8 I don't know if there's a risk of worsening
9 diabetes, really, or not, or transitioning people
10 from prediabetes to diabetes. I just don't know.
11 And when you're trying to judge the balance of
12 safety and efficacy, I would say that we don't
13 really have much certainty around the safety of
14 this drug.
15 It wasn't really presented, and I see
16 imbalances that I'm obviously trying to point to
17 simply to illustrate the point, not to say I know.
18 It's just to say I don't know. And the uncertainty
19 makes it difficult, I think, to judge the balance
20 of efficacy and safety. That's the only point I
21 was trying to make.
3 DR. COLMAN: Yes. Maybe there was a little
4 confusion. But we have been looking at various
5 Omega-3 formulations for the past decade in our
6 division, and after review of the original NDA for
7 Vascepa and then looking at this supplement, we did
8 not see any obvious, glaring safety issues that
9 hadn't been touched upon before.
10 We saw what the company was going to present
11 in terms of potential safety issues. And it's not
12 unusual for us to say, if we agree with the company
13 on one issue, that we're just going to acknowledge
14 that we agree with them, which Dr. Roberts did, and
15 let the company give their side of the story.
16 So I don't think we glossed over it. We
17 didn't see anything that was certainly a
18 significant -- that appeared to be a significant
19 problem. And we in general agreed with the
20 company's assessment of the safety profile from
21 these two studies.
22 DR. HIATT: I just don't know why you say
1 that when the exposure is so low. How many
2 patient-years of exposure is there? Would you say
3 the same thing if this was a new nonsteroidal to
4 treat arthritis pain? I mean --
22 DR. HIATT: I appreciate that. I'm still
1 just wrestling a little bit with -- the only reason
2 I could think of to lower triglycerides or raise
3 HDL or change LDL composition, Apo B levels,
4 particle size -- the only reason to want to do that
5 is to prevent cardiovascular events. I can't think
6 of any other reason to do that.
12 I think that that trial will perhaps inform
13 whether that's a biologically accurate placebo or
14 inert placebo. Even if you subtract that out, the
15 changes from baseline are pretty dramatic on
16 triglyceride as the primary, so I don't think the
17 drug is inert. It's biologically active. It's
18 changing the primary endpoint significantly.
News 
Market Data 
Discover 
AI
