 Well, I guess right now the floor is open for general discussion So if you have any questions for any of the speakers in this session, this is a time To ask I know there are some people want to ask questions I have a question for Mike about the you have to show the slide in the About the OCCUP, OCCUP and ZEMPIN Since because the structure is similar to COZEMPIN So people are always thinking about where the genetic link is to the BFINAL2 Have a similar reaction similar link and my question is Because by clinical observation the instance even in Asian countries instant quite low Although the drop structure is similar and also the severity is also low So I consider the genetic test that the COZE effect is how I would like to know how the FDA consider that That is whether Could you I'm sorry, how FDA considers what? To label if I know the BFINAL2 have been labeled for OCCUP and PIN in Europe You've been coming also in Taiwan, but I don't know whether how the FDA to Think it how it's possible to label BFINAL2 for OCCUP and PIN is like like I Will ask Sally to comment, but I would just say that we can't really speak about a sort of specific labeling considerations necessarily, but you know the principle would apply that if you have sufficient suspicion that You know one drug in class causes an event and you can predict it and identify patients who are likely to have toxic response that You would use any available information to help inform other similarly Compounds that have similar viabilities that we can't necessarily label or restrict use based on theoretical concerns So, you know, there does have to be some obvious risk that's that's been detected You mentioned at the conclusion of your talk about the FDA having some desire to potentially you utilize clinical decisions support to help with The avoidance of these severe adverse events This may be an unfair question because I suspect it's in another portion of this large agency that you don't necessarily touch But FDA has expressed the Opinion that they have the ability to regulate clinical decision support as a device Because a concern about safety and so my question is is Do you see that as inherently in conflict or is this a way that we could potentially move forward to try and Get through this potential Mindfield of problems relating to the device Perception of the FDA Yeah, I think the so that is managed by the Center for Devices and I will I'm from the Center for Drugs so So, you know, I the spirit of that that comment was more that You know FDA labeling as a communication tool to inform prescribing is not the most agile And and I think that there probably is some value in having dynamic clinical decision support systems that would able be able to Inform prescribing in a real-time basis based on accumulating evidence even if imperfect Because they're the other systems aren't necessarily going to to be able to react so quickly That was the nature of that comment I we work with devices a lot If I may actually just follow up to that point I mean there had there has been interest with all the next generation sequencing talk as as Evidence in previous workshops that there is an intense amount of interest in trying to figure out other ways to Communicate clinical validity and have that be established that's outside of FDA walls There's been very little said about African Americans or Africans and the risk for these complications But my as far as in liver disease African Americans are over represented when it comes to liver injury from allopurinol and the aromatic anti-convulsants do we have any idea of the rates of Stevens-Johnson syndrome and in blacks Yeah, this is an issue we looked at many years ago One of the problems in the literature was that cases kept being re-reported and accumulated So when somebody said, oh, there were so many thousand cases. It was you know, the same cases kept being reported over and over again so we looked at it in Toronto with the people that we were studying and We looked at it and compared it to the census data Which was about a year old at the time And there did not appear To be a strong Signal it might be for the anti-epileptics. It looked like it might be twice as high as In the white population But then there was a paper around the same time to say that said that epilepsy was probably twice as prevalent in a black population So I don't know what that means. However, I would say that when blacks have had Stevens-Johnson TN They would have perhaps more severe disease and worse outcomes and part of that might well be the recognition of the rash early on People have maybe tender skin, itchy skin, the rash isn't picked up That's my concern and that extra delay of even three days or four days Whether it's access to health care or good diagnosis could result in worse outcomes From higher mortality rates and higher morbidity For sure, so I think it's a it's a very important but potentially a loaded question and trying to figure out exactly all that's going on But there's others are going to comment on it as well So it goes to the issue of FDA labeling if you say for Asians that It's worthwhile screening for this The problem is this this whole problem. We don't have data. It's not like the data is negative. We don't know How do we deal with them? I mean we we try to communicate that uncertainty the best of our ability. I mean it was a little Unsettling to hear that there's a genetic predictor in Italians and not the rest of Europe And I shuddered to think how you could write something around that you know, so There's uncertainty and I think for a lot of adverse reactions and For micro genetic biomarkers that you know if there is a absence of an effect that's often stated if There's uncertainty about what the implications of it are that will also be communicated Well, we certainly know the rates of these HLA types and African-Americans are they overrepresented Well, I mean some of it some of them are 58 oh one is is pretty much represented across races, although it's most prevalent in Southeast Asians it is represented in blacks and And Caucasians and the same with 1502 but 5701 obviously is much lower in blacks And you know when you look at African-Americans, obviously you're looking at admixture So you have to take into account admixtures. So the Prevalence the carriage rate is going to be somewhat intermediate and we saw that obviously was 5701 where it's less than 1% in Africans continental Africans, but two to three percent in African-Americans and You know, unfortunately the actual true Prevalence rates haven't been worked out in terms of actually there's continuing admixture. So it's it's a bit of a moving target I mean clearly there's admixture continuing in the United States. So some of these things I think need to be re-evaluated over time It's the same with what we call Hispanics, which is a diverse Racial mixed, you know the Mexican-Americans are different than like Cuban-Americans and like more represent Native peoples Gene the genes of Native people. So these types of things would be very I think very important to look at the rates of these these alleles in different populations and can you say for sure that the risk of these severe adverse reactions is different in Different races independent of the The clue that we should be looking if you see actually I mean the initial clues That actually led to looking for HLA associations with drugs to begin with actually were that there was an over Representation in specific racial groups So it goes both ways if there's actual epidemiological evidence that there's a clear clue that there's a racial preponderance and that is actually an important clue to actually look first for an HLA Association and the other issue that Manier was talking about yesterday with never a pin and C401 That is an allele that's well represented across races. Although there still may be other metabolic factors In certain racial groups like blacks Where there is actually a heightened risk because of that for instance, so there's complex effects and the phenitone Sip to C9 star 3 issue In Southeast Asians is another example where you're not just looking at an HLA association There's complex effects, but when you see a strong independent Epidemiological effect, I would say that's that's actually a strong indication in an immunologically mediated drug reaction That it is almost going to be a pure HLA effect So obviously as niceties are very important issue in this. So can we have Maya to comment on this? Actually my point was what what Neil said what they observed in the black population and what you asked when we did a study or We part of our epidemiological study in Europe was to look for certain exposures And when we compared the people with SJS TN exposed to never a pain It turned out that most of them were of African ancestry in in different countries, but mainly in France and So when we compared this group of of never a pain induced SJS TN patients to all other Patients the severity of the disease was about the same the outcome was about the same So this is not black and white comparison But it is among the eight these never a pain HIV infected patients the majority were Africans so what I was saying is we couldn't see that the severity of the disease was higher and we didn't see a higher death rate But the HIV infected patients are usually younger than the other ones We have and we know that age is one of the factors for Higher death rate. So just to put this a little bit into perspective I just want to ask a question on the drug induced liver injury We also collecting samples for the entity be drug induced liver injury and Because it's a cocktail and when I asked the doctors, you know, what do you do when you actually found a liver injury? The answer is never the same because the causative agent can be any of the you know form Main items in the cocktail all this while you thought it was I said nice it and then I think Thailand reported that is more or is it was it Rika in Indonesia that said that is probably more it than but all So, you know, I was wondering with all these uncertainties That would actually prolong the treatment that is required and for for For the entity be part. So perhaps I like a comment from you Because I think that what we are trying to do this Southeast Asian group is actually get all our Cases together and do it as a project at the moment. It's actually individual I I'm sorry, I didn't understand the question Perhaps you could talk about it during lunch because we do have to move on. We have a couple of more questions Yeah, sorry Did you switch me off? Comment first of all I was involved in the working party that Changed the label with the European medicine agency for carb may spin in 1502 And the longest time of discussion that we had was well, how do you define an Asian? How do you find a Southeast Asian and and obviously we're coming to how do you define occasion? Etc. And I think that is a big challenge as we go deeper into genomics That is facing all of us and the regulators in particular I think that needs to be some harmonization Across the different regulatory agency in terms of how that is done And how that's defined because you'll see that the European label in terms of how it defines with 1502 is completely different from the From the FDA label. So I think it is important That maybe the international conference harmonization might be an avenue where people need to sit down and talk about Well, how they're going to I should start defining this within drug labels That's a comment question for Mike. So Let me just put in disclaimer first as a senior author on the new general medicine 31 or one paper and and you compared and contrasted the 1502 and 30 one or one And 30 one one does lead to is I completely agree that the effect size is less than with 1502 and it does lead to a wide variety of Adversive X some which are mild but some which are also very serious including dress, you know Where it's been replicated in many different ethnic populations in in the I'm gonna use the terms again Caucasians Asians Chinese hand Chinese etc so With so much replication What evidence do you need to be able to make them label more sort of mandatory rather than Information only do you need a prospective study? Do you need a prospective trial? What would you need to be able to change the label from what it is to be consistent with what's seen with the 1502 Yeah, I mean it's it's complicated to answer. I mean there's probably no I mean the evidence is there the it's the effect size is smaller and the the efficiency of screening in terms of True positives that you're going to get in terms of positive predictive value negative predictive value We're likely going to shift quite a bit You know whether there's some odds ratio threshold that you cross that you make some More affirmative recommendation That's not been defined So it's hard to say I don't if you have suggestions and I agree it is difficult to say but but you know and Matt may want to comment on this and Matt did some power calculation But you would need a huge prospective study to be able to and get the same precision of estimates that can be seen with the case control studies that are already available and And my worry is that sometimes we have a different level of evidence needs for genetic markers compared to other markers So if you had let's say serum rhubarb, which said that they predisposes to carbamase been hypersensitivity because serum rhubarb might be easily measured in a laboratory and Everybody knows about that test. It might be easy. It might be easier to get that into the label than to get a genetic test marker and That is a worry and then that you can see that genetic exceptionalism Everywhere within within medicine. Yeah, I mean there is that exceptionalism and I'm not sure why it exists I think if you can make some assessment risk assessment for an individual patient based on age and race You should do just the same for genetics, but the reality is right now We don't necessarily have all that information and there's a societal burden of imposing testing on costs and all the other things and We also don't necessarily dictate medical practice absent contra indications and indications so The professional societies may have a say in this as well. I see pick I'm not sure So any questions not for mark Leave leave for Mike alone Yeah, I'll just answer I do not know question for Mary Harrington because What's the best technology now and cheapest to just do the HLA type for um, you know an individual because It we if we can just do it once and then the patient knows it for alipurinol and and carbamase apine and You know at the back of here and and then all the other risks that you mentioned for disease Maybe that's the most efficient use of our health care dollars The big problem is it it's actually in our lab we can do ABC DRD QDP for $80 The problem is is that when we do that we're doing nearly 400 people at a time so the real issue here is trying to Come up with an inexpensive way of typing one individual very quickly and that's what you need and I would say You see so what I'm saying is it's much more cost effective to do a whole lot of people at once Then it is one person every day for five months or something and So there are kits that can be purchased and I and you know, they are meant for typing And I think they're from Dino They can be purchased and they are meant for typing a single individual at a time And I think a lot of transplant labs use those They're not they're not hugely inexpensive, but I don't think they're outrageous I actually don't know how much they they cost But um, but they're quick and they'll do one person at a time They might not be huge resolution, but it doesn't sound to me like that's really what you need Maybe I could I could follow up if you don't mind with with Mary on that on that question I think Alan raised a very insightful question about how about looking at all of the the variation I mean it's variable of region as that is it's hard to imagine that all of the types that are known Possibly encompass all of the variation that that could be there So so is there working on to identify more serotypes or is it that that humans are a young population? And so really we don't have all of the of the variants there You know there somebody mentioned this I Maybe it was Elizabeth that at the be low because there are over five thousand alleles Okay, so and there there will be more leels Most of those the vast majority of those are hugely rare The problem is is that all of the techniques that we've used historically They keep the rare ones rare because they make the assumption. Oh This I've got this sequence I don't have sequence from Exxon 4 and the most common one is you know 1502 so it must be a 1502 and that keeps the rare ones really rare so but now Deep sequencing is used routinely for HLA typing You know there is like the junior the ABI junior could be used to sequence a single individual and I don't really know the cost I could look into what those kind of costs are maybe you know Elizabeth but And that gives really high resolution data and But but I think right now what I would do is look at what the transplant labs are doing They're getting high enough resolution that you know, they can go ahead and do a transplant and they're using You know techniques that are the least expensive And you know, there are a lot of techniques, you know, we used to use Sequence specific primers and it was a really inexpensive way to do it You need a good bit of DNA, but I don't think that would be a problem here It's inexpensive. You do one person at a time. There are things so and that's basically what that dinoc it is as well so If it was useful, I could try to generate Identify information on that in the cost for doing different types, which you know a lot of people I don't know if there's somebody here from a transplant lab that could comment Is it is one of our clinical center colleagues here who does the transplant typing or maybe I Might just make a comment that research labs that have high throughput and our lab for instance has both research and clinical accreditation But as Mary said, it's the high throughput nature of it It gets it down to reagents costs and with reagent costs You can get things down to 50 to 70 dollars a sample for sure But but at the end of the day even if you have all the alleles available with you to you You're still left with the decision support aspect of things and getting a report The other caveat is that it really is important to have quality labs doing this and labs that actually have quality assurance programs I mean we saw this early in the back of your story that once the clinical a story and The association had been worked out that all the safety issues moved to the lab And there were episodes of false positives and false negatives for 5701 There were blacks in the UK being called as 5701 Positive when they were negative and there's there were smattering of cases That were safety issues where they were false negative for 5701 in the early days So it really is important to get the lab technology right and that the final sort of point I'll make is really There's other sort of innovative techniques flow cytometry should not be overlooked There's actually a lab in San Francisco. I think Alan Wu is doing HLA B 57 by B 17 Using a B 17. I think it's the one lambda antibody But because B 58 is part of that same serotype He could actually effectively if he had Q8 it for this screen for both alipur and a baccadir with that B 17 Antibody so there are other innovative ways of doing this But I would still at the end of the day say if you ended up with information on 5,000, you know, you know Across the population on different HLA Leels, how would what information would you provide to patients? That's the problem So to a final comments It's actually done in my laboratory at the NIH clinical center Department of Transfusion Medicine and We looked into that very carefully and with the format genomics for for HLA in the clinical center program It's not that cheap and one has to do it at a Clinical approved Level to make a clinical decision that's different from Simply doing research and doing screening that can be done relatively inexpensively But once you actually want to make a clinical decision, it does cost some money in particularly in the United States And one has to consider that big issue and perhaps one could contribute something to that is Forwarding the information, you know, you have to do it only once in a lifetime if the documentation is properly Maintained and that can be used Decades from now and I we would be perceived a lack of Infrastructure to have this documentation in place and perhaps a cost-efficient Aspect would be to get a good documentation of what is done on can be done today typically One would probably look for sequencing results at this time, which is expensive. That's not $80 But it's worthwhile along a term to do it at the sequence level and also identify rare Leals, which otherwise would be missed if you do it inexpensively Thank you last comment give you an actual cost For immune the immunology lab at the hospital. I work to do MHC one MHC two is four hundred dollars I can also tell you this you need to be very careful about discussing cost because often people are discussing your agent costs So it's costing him $40 to for the reagents But then it may be costing another two hundred dollars for the manpower or person power So to say it the test only cost $40 isn't a real reflection of the economics of the test Okay, I think we just got here and This is a great session I would like to send all the speakers and all the comments and questions and next time We should have more time for question. I think The way we get more time for questions is the speakers stay to their times So so that was a challenge, but yeah, so thank you and thank you Mike So now we have a brief break for lunch We we only have 20 minutes to be able to run down get our lunch and bring it back So we're gonna be eating lunch while Lauren is speaking and and Lauren is going to start at 12 45 because we want to Have time for discussion, so we'll see you in 20 minutes