 And this little one page is something of a guide to it. That's all I have right now. Before we get to the point where we're rushing to the airport, I thought it'd be important now to talk about scheduling the next meeting so we get some sense of people's calendars while everyone is here. I think, I've heard from several that June is out. How about July? Is that out of the question for anyone? Paul? Okay. Chris? Holger? Okay. So Mike, do you want to have someone from your office try to schedule a specific time? If we're able to do it now, probably three days. Well, yeah. But also we have a lot of stuff on the disks that I think will be really generate some questions. Is it the July meeting when we will have people coming in and talking to us? Yeah, that's the plan. And, you know, I think we do have a lot of data available. It's just we haven't had a chance to absorb it all. I think that certainly try to have something by, we'll have something by then. That was the third week better than the fourth week? Okay. Third week? Fourth is better. 26. Last week is not as good as the third week. If we had it toward the end of that third week, would that help? We have the 21st, 22nd, and 23rd. We'll have to fly in on the 20th. Some of us. I think it's important that you be here. So we drop down to the 26th. What are the issues, Holger, for you? 26 to 28 would be okay. Okay. Let's schedule it for the 26th, 27th, and 28th. And now, I guess the question is if you have the testimony on the 26th and then meet the, have a regular meeting the next two days? Yes, I think that would be the strategy. Okay. Okay. So moving on, we were going to discuss today a point that Bern brought up about what is acceptable risk and how are we going to deal with it specifically in terms of endpoints. But before we do that, Andreas was going to make a presentation that hopefully will enlighten us in this discussion. Do you have anything to add? Well, it's fit here. Maybe I might have to take them out. Just to let you all know, this is being recorded. So when you speak, please turn on your microphone and speak into the microphone. Thank you. We'll work the slides or I can do it. The red button is the pointer. Arrow. And I speak into this microphone. So I thought I'd start off by talking a little about the general background of our work, some mixture data, and then our little attempt at cumulative risk assessment, which as you will see will probably, will most definitely show up the data gaps and problems which we have, but maybe also useful for illustrating a way forward. This is just a reminder. Our work is motivated by investigating the role of chemical exposures in what is called the Testicular Dysgenesis Syndrome. This is a concept developed by Sharpe and Skackebeck. There is at the moment in Europe, and I gather here in the U.S. as well, the great concern about disease trends for non-descending testes, cryptocardisms. The trend is going up in Europe as well. Poor semen quality in parts of Europe. The semen quality in young men approaches crisis proportions. They have, according to a study which was carried out as part of an EU project, which I coordinated, it was found that 20% of young Germans, 18 to 19-year-olds, have semen numbers below 20 million per milliliter, which is the official limit by WHO, where you can require assisted fertilization. So this is not good, and later in life it's ain't going to get better. So there's great concern about that. Then we have hypospadias. This is a malformation of the penis, as you know, where the urethral opening is on the underside of the glans penis and not at the top as it should be. And finally, testis cancer. That's also going up in Europe in parts or levelling off at a fairly high level. And Skakebek and Charp have proposed that all these disorders are indicative of a malfunctioning of androgen programming in fetal life. So they all go back to a problem during development related to androgen insufficiency. Let me clarify, they say that not all cases of hypospadia have to do with what they term the testicular dysgenesis syndrome. Not all of them. There are, in particular with hypospadias, some well-defined genetic disorders which lead to these problems. That's not necessarily to do with the TDS. But the concept they propose has been very fruitful in directing research. So it is really, that is what the summary of a lot of research, both here in the US and in Europe, is that the action of fetal androgens makes a man a man. So if there are problems with fetal androgen action, the affected individuals show signs of feminization. So we know now that certain chemicals can disrupt hormone action in fetal life by either blocking the receptor, the androgen receptor, or by suppressing sex hormone synthesis. And this leads to demascularization. And we have in the rat a very good model that is able to recapitulate all three of the disorders that make up the testicular dysgenesis syndrome. Only it is impossible to study testicular cancer in the rat, which has to do with when the spermatocytes mature in a rat's life. But all the other elements of the testicular dysgenesis syndrome can be recapitulated in the rat. And the male offspring of females treated with certain chemicals then show signs of demascularization. The problem is, or the question, definitely this model has relevance to human because the programming in fetal life is basically the same in these mammals and in humans. The question that is not resolved is, do chemicals really play a role in these adverse disease trends in the human? Now here, a little jump to how toxicology would assess chemicals normally. Here's the example of vinclosoline that's a pesticide often found in red wine. It is a fungicide, so people who love red wines from Italy you can show the traces of that pesticide in their urine. The breakdown product, you will find. Vinclosoline is one well-studied chemical in this rat model and typically demasculinizing effects are seen at doses around 5 milligram per kilogram body weight per day. That is the dose to the pregnant female. The human exposures are in Europe at least in this range which you can, if you divide the human exposure by the dose that leads to first effects in the rat or just not effect, you can see there's a margin of safety as it is called of between 200 to 1000 and at this point most regulatory toxicologists would say, fine, case closed, no case to answer, move over. But there is the issues of toxic cocktails as it is called or mixtures or exposure to combined chemicals or many chemicals simultaneously. And the question that has inspired a lot of our work during the last 10 years is can chemicals that are safe on their own, if you like, gang up in our bodies to produce effects that cannot be anticipated by studying those chemicals in isolation? And secondly, are these effects predictable if we know the potency of the individual players? And here's a quick summary I'll take you through in stages through some key findings relevant to our work here. First of all, there's this landmark paper by Earl Grace Group, how to sell a tile, it came out in 2008 which shows that phthalates, combination of phthalates act additively in producing mixture effects. The endpoint they looked at was testosterone production. So here you see these gray lines are individual, dose response curves, the regression models for the phthalates how to sell a tile studied. So you see here as we escalate the dose, the testosterone synthesis goes down in fetal life. They prepared a mixture of these five phthalates at a certain mixture ratio and then using those gray curves, it is possible to construct what the predicted combined effect of these five phthalates should look like. It is for a mixture of specific mixture ratio. And then that's a very strong heuristic tool because you can design the mixture experiment on the basis of this prediction and that's what how to sell a tile did and these are the observations and they match fairly well what can be predicted on the basis of the individual dose response curves. The conclusion from this was that these five phthalates act together in a dose additive fashion to produce suppression of testosterone synthesis. Maybe a little word about what those addition is. There are two competing models one by which such additivity predictions can be calculated or constructed. One is called dose addition. It works on the assumption that one equity effective dose of one component in the mixture can be replaced by the equity effective fraction of another one. The concept finds wide use in the application of TCDD equivalence factors in evaluating the oxygen mixtures. The competing concept would be called independent action or some people call it response addition and there are other considerations and other approaches going into this. Put simply this model works on stochastic principles and assumes that each chemical does what it does in a fashion independent from all the others. Mathematically these two concepts are different but often they produce exactly the same prediction for certain mixtures. For example in this case here you could calculate an additivity prediction for independent action and it would sit very closely near the dose addition prediction. Under certain circumstances these two predictions would diverge. The question then is which of those two competing models can approximate the observed mixture effects best? So far as you will see in this area of study the answer is that dose addition is fairly useful. So that's the first landmark paper. We together with Ola Hasse's group in Copenhagen, the Danish Technical University have spent a lot of time studying these chemicals in a rat developmental toxicity model. This slide just shows you briefly how this experiment is designed. This here is the lifeline of a female pregnant rat so they get pregnant here. Dosing begins on gestational day 7 and normally throughout the entire pregnancy through birth up until postnatal day 16. Then here they give birth to offspring typically six males and six females and here's the lifeline of the young male rat. Very early after birth the anal genital distance is measured. That can be done in a non-invasive fashion so these little rats live on. Then a little later on postnatal day 16 samples are taken, some are sacrificed and the organ weights are measured off organs that are under the control of andrin action so for example the prostate and various other sex accessory organs and also malformations of the penis are evaluated. Then later that's the postnatal day 47 where malformations of the penis are looked at again so we would look at things that recapitulate hypospatias in humans. Here's the first result of a long study which we conducted. The intention of this experiment was not to model environmentally relevant experts who just wanted to refine the tools. It has never been done before at the time we needed to know whether andrin receptor antagonists produce mixture effects in a predictable fashion so they were selected these players here is Vincloselin, flutomide and prosimidone these two are pesticides, flutomide is a drug they were selected in terms of their effectiveness and their mechanism so this is if you like a lab mixture of environmental experts but this was terribly important to us to verify certain concepts. So what you see here is a dose response curve for anal genital index here you see how with dose escalation the offspring gets more and more feminized. Maybe I... no that's fine. Now this is the outcome of the mixture experiment the green curve is what was anticipated on the basis of these individual dose response curves for a mixture with specific mixture ratio and what you see here these dots are the observations the black dots the outcome from individual litters the red dots the overall litter means and we were quite stunned how well the observations matched what was predicted again this shows that these three androgen receptor antagonists work together in a dose additive fashion. So step three we then looked and this paper came out very recently last year in EHP just before Christmas we became interested in looking at combinations of antioxidants that work in a totally... that work by mechanisms that are very different. So finasteride for example blocks the conversion of testosterone to dihydrotestosterone vinclosaline blocks the androgen receptor as we've seen. DEHP interferes with testosterone synthesis and prochloraz is a mixture it is able to antagonize the androgen receptor but also interferes with some important enzymes in steroidogenesis. What we did was studied the individual dose response curves of all these players again and then mixed at a certain ratio and this is what happened with respect to a multitude of endpoints anal-general distance this is musculos-leviator ani that's a funny muscle that's also under the control of androgens in development nipple retention and prostate weight and what you see here in green again are the predicted curves according to dose addition and the blobs here what we observed and again with these players here that work in a dissimilar fashion there was good agreement with dose addition in this particular case let me say that the prictions based on infant action were actually congruent with with those for dose addition now we were very pleased but now comes the one result which throws a spanner in the works which I personally hate but there it is penile malformations you remember at this very last time point postnatal day 47 here we have we constructed a prediction window if you like this is what we expect on the basis of independent action this is one very conservative prediction based on dose addition and these here in blue is what was really observed so you see a shift towards lower doses with this very end point a very important end point penile malformation this shift here if you project it down onto the dose axis is probably not very large it's a factor of four or five with respect to here this is larger is nearly an order of magnitude but what is a little disturbing is that in a dose range where basically all the models predict that nothing much should happen the observations show us that we're near maximal effect that is because these dose response codes are so steep we concluded that this is a synergism but we can often know mechanistic explanation as to why this should be when all the other end points are essentially dose additive maybe there's something about penile development which we don't quite understand also it seems to us that in this case here the presence of phthalates and other players sensitize the tissue to the effects of Vinclosolin Vinclosolin is by if I haven't shown the individual dose response curves but Vinclosolin is the only player in this mixture that will give a complete dose response curve with penile malformation all the others give curves that are sort of stumps a little stump so as I say we don't like this very much ourselves because it throws a spanner into this predictive approach but it has to be reported yes because it's you Chris so the dot there is on the high level is the mean the sample mean and the bar is what? these are confidence intervals how many observations in each dose group do you remember? quite a few I don't remember off the top of my head but we normally looked at between 6 and 10 liters and each liter has 5 male offspring but some of them were sacrificed before that so I really have to look it up how many? but I'm just wondering because in the previous slides you had dots for actual observations the actual observations here were there some animals that were lower or did they all sort of cluster up with the confidence interval well they cluster up pretty much here otherwise you wouldn't get at this point you didn't have an animal that was down here okay so just a resume of what we found so far that we've also I'm not showing you this here we've done a lot of work with mixtures of Easter egg chemicals in vitro also with fish so we have joint effects with mixed mode of action anti-androgen and that's quite important because there's this myth out there in the regulatory world and otherwise that says people think that combination effects are only to be expected with chemicals that show a similar mode of action this is evidently not born out by our empirical observations combined effects are always greater than the effects of the most toxic chemical in the mixture and this leads us to think that the usual chemical by chemical approach to risk assessment may lead to underestimations of risk I say may not necessarily in all cases but we can define parameters where underestimations of risks are likely that's particularly when we're dealing with mixtures of many components more than 10 right let's these are what follows now this is the empirical results I thought it might be useful so you have a feel of what's possible now in the area of mixture toxicology now the question how can we approach reality and how can we make use of that in risk assessment when you want to do this the first question we have to face is how are we grouping chemicals which chemicals do you want to consider together in risk assessment there are mechanistic criteria so that's out there that's very much discussed by EPA by other regulatory agencies in Europe you can say I grouped chemicals that show a similar mechanism or mode of action but the question then is how narrow do you want to be and what precisely do you mean by mode of action there's a discussion about it to a large degree pretty fruitless in my opinion if it would be to adapt a phenomenological criteria by saying pretty regardless of mechanisms or the details of mechanisms we should group according to similar adverse outcomes so this is for example the approach that the Salad Report of the National Research Council has taken so the question then becomes which chemicals can cause de-masculinizing effects in male offspring and let's group those together regardless of the details of mechanism and mode of action the question then is that concerns a lot of people are we not casting the net too wide here so this is the dilemma either you're probably too narrow or too wide let's just highlight this if you do it if you overdo it with mechanistic criteria you will always find arguments where not even combinations of phthalates are regarded as sufficiently similar to warrant joint assessment and I think then it becomes absurd there's always slight differences in the molecular mechanisms of how phthalates act in this kind of system so a grouping like that would be if it's too narrow it would lack credibility on the other hand how wide do we want to cast the net this is in my opinion one of the most difficult questions in mixed toxicology today to define on a sound scientific basis good grouping criteria for cumulative risk assessment this grouping has to be plausible and credible I'll give you one example I mention this every now and then for example US EPA groups, organophosphates carbamates, styrocarbamates and dicylcarbamates in separate assessment groups for cumulative risk assessment when essentially all these chemicals work by the same mechanism contributing acetylcholine esterase so in my opinion that's an example of a grouping strategy that lacks both credibility and scientific plausibility so how can we make headway with these anti-androgens and phthalates and here's an insight by someone called Earl Gray who whenever asked says the developing tissue does not care if the phthalate is compromised whether by suppression of androgen synthesis or antagonism at the receptor the outcome is the same and that opens the door this is a biological or we call it physiologically based grouping criteria and instead of getting so hung up about mechanisms so here's a chart which appears in the other anti-androgen compounds other risk factors through this funnel if they all act on androgen action the outcome is altered male reproductive outcomes and this has to be taken into consideration so here's a proposal for grouping of course our favourite chemicals the phthalates have to show up but here's a list of pesticides that many of which act at the androgen receptor level we have environmental chemicals that show evidence of anti-androgenicity PPDDE is if you like the classic one certain PCBs show these effects certain dioxins and furans as well and then there are other certain parabens for example there in vitro androgen receptor antagonists UV filter substances and some others so I now I'm moving towards my example of how this could be played out but before I should lay my cards on the table in terms of risk assessment approaches and this will be touched on yesterday there are two fundamentally different approaches to risk assessment they've been discussed by Glenn Suter at length he's written many useful papers about this there's one first of all risk assessment to protect human populations in the sense you know we want this deterministic risk assessment we want to know does the red light begin to flicker or not and this kind of risk assessment out of necessity has to have a bias towards conservatism in case of uncertainty and that is necessary to protect human populations the aim is mostly to evaluate is there a case for action this has to be contrasted with and may not be confused with a quantitative risk assessment where the aim is really to assess the magnitude of a health impact at current exposure levels that's some it's of course the two approaches are related somewhat but it is different and both approaches use different methods to contrast this the quantitative risk assessment approach would very much rely on probabilistic methods and approaches the question is in our case are actually sufficient data at hand to move towards quantitative risk assessment and to cut a long story short I don't think there are so what follows is very much in the camp of this deterministic risk assessment we started doing this myself and my collaborator and friend Michael Faust just in order to see let's use the information that's out there and will this kind of cumulative risk assessment is there a case to answer or not with the TDS and the role of chemical exposures so how can this be done we're all familiar with the classic single chemical approach you divide intake estimates by what you think is a tolerable daily intake and if this quotient is smaller than one everyone is happy there is an approach which is called the hazard index approach to cumulative risk assessment which builds on this and uses the same quotients if you like that's a hazard quotient but does this for every chemical in the mixture the demand is then that the sum of each of these quotients here should also be smaller than one that's an approach the hazard index approach that has been widely used and is still widely used by US EPA for example for the assessment of superfund sites normally what they do with superfund sites the demand is that these tolerable daily intakes does not necessarily have to be related to the same endpoint or health outcome they sum up across totally different endpoints but we thought we can't do this in this case let's utilise this approach and sum up over endpoints relevant to anti-andronicity so let's not mix up apples with pears really for the Chris I put this in for you right so in selecting in making a list of chemicals we want to subject to the hazard index approach we had to reflect on an effect criterion and what we chose were only chemicals with in vivo evidence of anti-andronicity we disregarded all in vitro anti-androgens exposure information had to be available as well and let me tell you these two criteria which probably seem very straightforward and trivial to you already proved to be very restrictive a lot of chemicals had to be thrown out for a stunning lack of information if possible our goal was to distinguish between median exposures and high exposures so here's the first list these are the chemicals which we selected I hope you can see it it's a little small but I'll talk you through these these are the chemicals we have five phthalates we have vinclothalimplochroasprosimidone linorone, fenetrothione here these are well documented anti-androgens in vivo and finally PPDDE down here the effect looked at or taken into account here was suppression of fetal testosterone synthesis retained nipples for the pesticides mostly so this is the chemical distance for fenetrothione and retained nipples in male offspring for PPDDE now here is what I would like to define as point of departure but for anti-androgen endpoints not for anything this is derived from experimental studies for example how to sell just a point of departure it could be either a benchmark dose or a no observed effect concentration or dose here's an uncertainty factor and here's already the second limitation it turns out that many of these studies in fact were scientific exploratory studies not designed to estimate very low effects so but the information was out there we took account of this by combining then by using an uncertainty factor larger than 100 normally 100 you use to define a reference dose but in 100 we've used here for these pesticides so by combining those point of departures from animal experiments with uncertainty factors we arrived at what we call reference dose for anti-andronicity microgram per kilo in a day let me say this was important to us as well in order to avoid comparing apples with pear let's look at Phinitrathion. Phinitrathion if you like has an official ADI which in fact is lower than the reference dose for anti-andronicity which we took into account here because Phinitrathion is also a neurotoxin and occurs at doses lower than anti-andronic effects so we have not taken the ADI for Phinitrathion in order we feel that would have muddied the waters okay these here's the second selection it becomes a little hazier here and here you see already the problem here's a flame retardant BDE-99 there is one paper that describes suppression of testosterone levels and reductions in anal genital distance in male offspring only one study Bisphenol A there are some studies in Japan that describe decreased testosterone levels in male offspring but not very well documented and then propyl and butylparaben again here suppression of testosterone levels decreased adidimus weights and decreases in spam production but in general these chemicals the evidence for anti-andronicity I would regard as fairly limited the data could be better nevertheless just to try out how far we get with this we include those chemicals as well then intake values and yeah okay that is the first attempt to cut a long story short if we carry out this exercise cumulative risk assessment using the hazard index approach for median intakes on the basis of 15 chemicals the answer is there's probably not a case to answer because the hazard index the sum of hazard quotients here is 0.38 smaller than 1 let me just talk you through this table so here you see those chemicals here's the median intake the next column is the reference dose for anti-andronicity the hazard quotient is just intake divided by reference dose here you make the sum and here we constructed the ratio of the hazard quotient to the overall hazard index and because it's 15 chemicals if they all contributed equally you would expect this to be around 5% highlighted in gray are those that contribute more so this is a wonderful tool for prioritization all of a sudden you see which chemicals are important players and which ones do not contribute a lot to the hazard index for example these here dbp, dhep those pesticide and butylparaben I have to say something about this later contribute disproportionately to the overall hazard index but for example not ppdde not brominated flame retardants lots of people are hung up about brominated flame retardants and anti-andronicity interesting the human mind without aid cannot easily aggregate prevalence and potency but this is a wonderful aid to do and a tool for prioritization okay let's look at high exposures and here the story changes and in the interest of deterministic risk assessment we should take this account we're constructing a worst case again here similar construction, high intake reference doses, hazard quotient and here the relative contribution what we see here is that dhp is important probably dbp as well all those pesticides bisphenol A butylparaben however what worries me is that with bisphenol A butyl and propylparaben both the intake estimates and the potency estimates are very shaky however the overall sum is 2.01 that's probably over accurate so let's say 2 and you see here butylparaben alone contributes 50% so we could say let's take butylparaben out for the sake of the argument because the quality of data is not very good we would still be left with something quite close to 1 which I would conclude to mean maybe we should be a bit careful we've then carried out well that's a summary here for high intakes we have 7 chemicals 90% of an expected combination effect there's poor data quality for bisphenol A the same for paraben and the surprise to us if we compare intake values was that flame retardants and ppdde contribute very little this goes against the grain of the hobby chemicals of many people in this area but now we have to have an uncertainty analysis we asked ourselves and you see here the Empire State Building gets a little hazier what are factors that decrease our risk estimates and definitely the first one is the question of combinatorial exposures we have assumed in this worst case that everyone is or no that certain people are exposed to the high end it's really that someone high in psalates is also high in pesticides we don't know for a lack of data what would be needed to address this would be that we have samples from one and the same person analysed for multiple chemicals, these data are not available but surely I think it can be argued that combinatorial exposures will definitely decrease our risk estimates against this we have to balance factors that will definitely increase our risk estimates and these are first of all my worries are that we are dealing with an incomplete set of chemicals if you remember the selection criteria which we set ourselves meant that very definitely we had to leave out those chemicals where there are suspicions of anti-androgenicity so if we had better data and our data gaps and if we included those as well the risk estimate would go up again, the hazard index would increase secondly I have concerns about underestimations of potency you remember when I explained how we arrived at those reference doses for anti-androgenicity many of these studies were really not designed to estimate no observed effect levels or benchmarks and my concern is that really those risk estimates in future studies, better conducted studies might actually show that we are lower that would increase our risk estimates as well so you can see this just you probably guessed it with all this we've made the assumption that the exposure of women in the reproductive age is very similar to that of the general population I think that's a fair assumption really we want to protect the male fetus so here if we go back to the formula for hazard indices you can see mathematically so we have the exposure levels here if they go up then the hazard index will increase the acceptable levels for single chemicals if they go down hazard index will increase and finally that worries me a lot the total number of chemicals that go into this formula here n we restricted ourselves to 15 if we have better data for more potential anti-androgen which will enable us to put these into this formula as well well for help we can turn to this man and reflect a little unknown the Invitro-Androgen receptor antagonists our lab we are pursuing this with vigor in our lab we are testing quite a few chemicals and a lot of surprises happen there is at least 35 pesticides in use which we know of are Androgen receptor antagonists in vitro there is UV filter substances they are too and certain dioxins and furans the question really is you know as well as I how expensive it is to do vitro developmental toxicity study do we really want to test all of these in these models it would be very expensive on the other hand there is this information out there about a lot of Invitro-Androgen receptor antagonists what do we do the most serious stumbling block in my opinion is this poor and fragmentary information about exposure levels for example about the parabens there are very crude estimates about exposures done by European scientific committees but we are not sure better data are needed so to conclude I believe that the hazard index approach is a good tool and probably the best we can use at the moment to make headway with at least prioritising chemicals in this area and for making headway with getting a feel for whether there actually might be a case to answer or not it's a powerful tool also to identify knowledge gaps and I couldn't anticipate that before we did this and I have to say I was a little surprised about some knowledge gaps there are loads of pesticides out there quite frankly they are in use and in vivo data about anti-androgenicity industry has got away with not providing these data we can also use this to direct epidemiological research so epidemiologists might be interested in this and say right we can tell them rather than looking at your favourite chemical for example there the environmental epidemiology in this area is littered with studies where only the effects of PPDDE on hypospadias and similar things were studied this consideration shows us that epidemiology has probably been barking up the wrong tree for a long time and instead should focus on some other chemicals it may also be used to focus regulatory action so to conclude I would say we can try out cumulative risk assessment it is better than anything that happens today simply by pursuing single chemicals anything considering mixtures is better than the status quo and with this I'd like to thank you and thank you any questions, comments I have a couple of questions are there experiments where you are looking at the mixtures you make a mixture or Earl Grey or somebody maybe half a dozen chemicals they are generally roughly at equipotent doses is that correct these are what we call the lab designer mixtures where the aim is to refine prediction tools to see whether we can really anticipate the effect of absolutely this is necessary to do so because you have to safeguard against the danger that really you don't do a mixture experiment although you put some chemicals together and that's the case when maybe one or only a few players dominate absolutely and also you're doing experiments at doses you have very measurable effects and my question is after reading that chapter in the NRC report what happens as you move away from that sort of middle ground when the exposures get lower where you know humans exposures are not at the 50% effect level they're somewhere lower and where the mixtures maybe a little more skewed do you think that what's likely to happen will additivity do you think dose additivity will still apply or do you think it may be less than dose additive I haven't shown this but we've also nested in some of these studies we've done all our something from nothing experiments where for example with the mixture of 300 receptor antagonists in vivo it was possible to define an observed adverse effect level and we then deliberately directly combined the three chemicals at precisely that dose to see is there a mixture effect or not and there was and it was predictable well by dose addition just think with all the chemicals we're exposed to and we know that some of them are anti-antigenic probably a lot we don't know and it just seems to be to me if you keep adding chemicals to your list it's going to find that the risk is so high it would be unbelievably implausibly high and I'm just wondering as you get to lower and lower exposures to more and more chemicals does that you know do you deviate from additivity conceptually no although if someone gives us the money we do the experiments if you want proof positive but it's going to be damn expensive we think I personally don't like doing those experiments they are necessary to prove I would say if you then move on into theory it is to be expected that if you combine many chemicals at low doses if they then the toxic units the hazard quotients end up to suitably high level then you will get a mixture effect of course if you look at the mathematics of this you can anticipate situations where for example if you have very few relevant players combined with relatively low exposures nothing is to be expected but the answer to that question and I think this is quite important to separate out is not an issue of mixture toxicology it's one of incomplete information about exposures and also the relevant chemicals which we need to take into consideration I agree totally that the issue is on exposure because to have all these things at the high levels in any one place in any one time is quite remote and unfortunately when people have done risk assessments using Monte Carlo techniques to take the 95th percentile of everything and they throw it in and say well we got a risk sure of course you do because you've thrown everything into the hopper and you've only looked at the real high end and I think we have to be very careful that if you look at the realistic probabilities of coming up with some situations where there may be one, two or three chemicals that occur simultaneously because either high production or they're in products that everyone uses every day and they're high levels in those projects but for the most part I just don't see where that 90th percentile is going to come into play in any reality but there's nothing going for us because the phthalates and endhanes they're measuring a set of phthalates and so if you could it wouldn't be easy to do but I think it could be done if there were a bioequivalent a bio-monitoring equivalent whatever you want to call it a metabolite level and if you could apply some adjustment factor that would adjust for not only the pharmacokinetics and then also the potency you could sample and we can get the download the data from NHANES for individuals? if you can download for individuals and you can have the range of levels from the body you can at least get an idea of how many are present and how many are not but that means you have to download the whole data set because there's so many of these samples in which there's non-detects I didn't think there were that many that were non-detects it depends upon the individual and the compounds because we're talking about a whole variety of compounds here and I tend to people reporting things like the 97th and the 95th and the 50th percentile they don't report the 5 and the 2 and the 1 percentile which in fact may in fact be there might be a lot of data if we knew how to do that I think it's conceptually possible it would be very valuable but we thought about this as well and the problem is one of those metrics if you have information about tissue levels in humans it would be really nice to have the equivalent from animal experiments where we know that the outcomes are observable but unfortunately there's another data gap we don't I'm currently involved I'm currently coordinating an EU project where we're trying to fill this gap but that's going to be a lot of work but there's only one hint of what might come that's a study by Earl Gray where they looked at where they gave phthalates to the female rats and measured in the amniotic fluid the phthalate levels and compared to what's measurable in amniotic fluid in humans and when observed when they found effects in the rat the margin of exposure on that dose metric was frighteningly small but that was done only for two single chemicals but this is the kind of stuff we would need and then human tissue levels would come to life and we could use them at the moment I find this difficult because we haven't got the information from the animal experiments to use a read across or a parallelogram approach you mean like even comparing a rat urinary level to a human urinary level that's not possible I don't think that's possible at this time in fact I think there would be kind of dangerous in terms of coming up with any kind of reasonable assessments oh no I wouldn't think it's dangerous someone needs to do it and see what happens using the data again biomarkers in their own right there are single points in time and I don't think unless we have time series that we have any way of using that effectively for a risk assessment which we're talking about a risk assessment not for just one day in a life but over a year or maybe even 70 years so we really have to be really careful how we extrapolate these numbers yes absolutely well I mean I look at it like you said it's a snapshot but it's of a large population so is it any different from looking at one person from day to day absolutely absolutely it's different from day to day well for an individual it is but for the population think about use of pesticides I mean in the United States if you have people who have you know farms gardens people who have pests in the home and they spray their own houses they will have periodic exposures and those exposures will lead to levels in the body that will be decreased over time and so you sample a person at time X you find nothing, you sample a time Y you find levels that are through the roof you will have characterized that individual's exposure and you're using a general population study sure it gives you statistical distribution of the mean but not the people who read the high end or the low end so you really got to be careful how you extrapolate that data you really do do you think the goodness of FET that you've seen here is unique to anti-antigenic activity or might you find it with other hormonal activities or even more distantly hormonal mechanisms of action I'm not sure I understand your question goodness of FET with the dose response curves or and also the mixed mechanism but they still involve androgens yes, yes in the one mixture the three players antagonize the Andron receptor you saw the quality of dose response analysis which we conducted the regression models with confidence belts etc and I'm not sure I'm getting your point but the diagnostic criteria on which we use to assess agreement with the prediction is an overlap between the prediction belt around the predicted curve and the observed data if these two overlap we conclude it's in line with dose addition and that was the case with all the examples I've shown you accept this synergism I'd like to come back to the biomonitoring data I think we shouldn't over complicate the interpretation of the biomonitoring data the interpretation of the external those estimations for example by the chewing experiments and so on because you can chew and you can't chew you can take out the dose you can't it's the same with the biomonitoring however picking up the lead from Paul I think with the delates we are in a position that we very well know how consistent the exposure is within the general population you have for at least six delates always positive biomonitoring values you have positive detects of at least six delates in the urine and to make it clear detecting has nothing to do with risk or a hazard but we can extrapolate the metabolite levels pretty reliable, pretty reliable to the dose taken up and there are other studies I think also can talk about it that we pretty much know that those who have been in the upper quartiles are pretty consistently in the upper quartiles for most of the delates and for others they are not and yes we can individually look at the delayed exposure of each individual we've done this in Germany for the German environmental survey we can up the delayed levels on an individual basis so we can really sum up the dose of each individual regarding the delates and looking at the Haines data plus other anti-antigens as long as you have time series you can do that yeah but Paul isn't it the same with the external dose extrapolations with chewing or not chewing taking the teeth in the mouth or not measuring air concentrations and dust concentrations I can be in the room and I can't be in the room I think the issue of the external dose the external exposures if you're talking about where the exposure is occurring if it's occurring in the home yes you have consistency because the levels aren't going to change from day to day because you have all these products that have phthalates in them so that's fine under that circumstance but if you're looking at the high end exposures who knows it can be for anything but for things where there's a consistent day to day reasonable consistency in levels which you probably have in most homes I see no problem with it's just those high end peaks you're going to find them that that is not easy to predict no but I think we have to have a reality check here I think we also found that the exposure information for phthalates is as good as it gets it's sufficiently good to do this exercise and that's fine we can, you know, there's good data from the US, the NANES data and also good data from Germany very extensive data and you can construct 95, the 95th percentile as a criteria for high exposure that's sufficient we have to be realistic you can dot the ice and cross the T's if you like but I don't think it'll alter much with a general picture I mean there have been smaller studies where they looked at people for what, a week, two weeks or something like they looked at day to day and even within a day I mean there are variations I was just going to comment on that there are those studies where there's repeated measures where you just have to take a leap if we're limiting our discussion to phthalates there's no, I don't have a problem with it when you start burning, throwing in pesticides you can't it's a totally different set of circumstances in applications and source patterns when you say you can't what can you not do you cannot assume that it's going to be a constant level because you have applications periodic and it's not every day of the week and you don't know whether the person is getting a very high exposure on day X versus day Y it's a difference in terms of source distribution and source release the emissions from a power plant are consistent from day to day alright, that's because the source is constant the emissions of phthalates in the home will be fairly constant because you have similar sources every day the difference between that and other kinds of sources will in fact depending upon the periodicity the intensity and the contact that you have there's a whole different set of variables yes, but all this is true but again we need a reality check the exposures or intakes of a large number of pesticides are regularly estimated on a mandatory basis in Europe the information is there well, I guess it depends if you're exposed from residues in food it may be fairly consistent if it's the one you spray at home that's a little bit different which is basically where the highest exposure is Ketokonazole I think is used for human, you know athlete's foot or something so that may be a little more sporadic but I think that the phthalate exposure is also sporadic because apart from home exposure there's a use of personal care products you use it in the morning then you may reapply something in the afternoon, the evening and then I think a lot also depends on the timing of the urine sample you could have someone that uses a product at 6 a.m. take a urine sample at 8 and at 10 a.m. you're going to get very different levels because of the short half-life so I think for phthalates too there's a lot of intermittent exposures phthalates that are used in personal care products semi-volatile or they mostly non-volatile they're the same that's the butyl and the ethyl so they're like semi-volatile but you do absorb them through the skin so there will be apart from maybe your home or office environment where there's an indoor air level as you apply these products as you eat different foods throughout the day you're going to be dosed at different levels but again for those ethyl and butyl the exposures may be more episodic whereas for DEHP, DINP maybe it is a little more constant I mean it's I think DEHP and Holger has some data on this presented at a meeting a few years ago was primarily through foods right so that would also be episodic and then a lot depends too on the spot urine samples when they're taken relative to the so that's probably why you're seeing the variability isn't as much as what someone's exposed to as much also as when that urine samples taken relative to when that exposure occurred and you're detecting these in whatever 90-95% of the population I mean my feeling isn't that few percent that you may not detecting it it may be that they're using these products or eating something but their exposure is far enough remote from the urine sample that you're not detecting it meaning that there's widespread we're going to need some exposure scenarios or exposure characteristics to deal with what we're going to be dealing with the 50th percent of the 90th percent of the key will be how one constructs these profiles to get some level of understanding of how it relates at some point to biomonitoring because without that then all this stuff could be stochastic there is some data there's data that actually people have taken urine samples multiple times throughout a day relative to a point exposure and then there's data on people repeated measures you may have 10 or 15 on the same person over a month or several months so those can be used as our benchmarks yes well it is to answer your question it is very important if you pursue the hazard index approach to get the dose metrics right it weren't tissue levels would be of no use for that at this present state because there's only one demand with the hazard index approach the dose metric of the numerator has to be the same as the one of the denominator which is intake the numerator is intake however if in the future there's more information available about how human tissue levels compare with tissue levels for example in the rat that are associated with the effects then you could do a better read across but at the moment this kind of data is not available but this is not only relevant for interpreting biomonitoring data what you said that's relevant for interpreting all data oh yeah I'm happy to concede that yeah I have a quick question which is kind of in a different direction what you've presented I think you've presented it clearly but it's complex and so you're looking at compounds that may act through common adverse outcome mechanisms but then we know that there are chemicals that may alter the metabolism of chemicals so it could become even more complex so for instance if there's another chemical that may increase or decrease the rate of metabolism and therefore clearance of phthalates is that something that can be considered in these I mean it would be I think another step forward because you're looking at common adverse outcome or mechanisms this is a very important question and this is where the limit of this predictive approach to mixotoxicology that's where the limits are of this approach which we pursue because what you're talking about is a situation where in isolation chemical A doesn't do anything but say combined with an anti-androgen and phthalate would exacerbate the effects of phthalate or the other anti-androgen with mixture effect prediction and assessment this only works for combinations that all produce the effect you're just interested in and then you can model so what you've described is a very important point but at the moment essentially unpredictable quantitatively you may arrive at qualitative predictions and say right if we combine a phthalate with this and that chemical qualitatively we would expect the effect to get worse or better but you cannot easily predict at this stage to what degree the effect would get worse or better this is done very frequently with drug interactions you know someone's on drug A and you say well don't take drug B because its metabolism is impaired this you're actually taking is not 10 milligrams that your body will see but you know 100 milligrams because drug A is impairing its metabolism so our goal is to come up with some numbers to determine whether or not there's going to be a ban on these products in toys and in other kids products so how do we approach this going toward what kind of degree of uncertainty or certainty we're going to allow on both sides of the questions to at least come up with some reasonable estimates because if we don't set some reasonable targets we can spend our time arguing about how much uncertainty we're going to allow on any one measurement and I'm fearful that that's going to cause us a lot of problems well I think that leads us right into the next thing we wanted to discuss and I think at this point maybe we should take a break okay come back at 10-15 and pursue you know what is an acceptable risk the point that Byrne brought up and acceptable risk with what endpoints those are critical issues we need to really discuss and come to some kind of consensus I like the idea of reasonableness with respect to doing it I'm going to figure out what reasonableness is don't have everyone here at the moment but I think we should get started again and now I'd like to start a discussion about some of the issues that we're going to have to come to consensus on and I think the first one we deal with is and it's going to be one of our charges is what is acceptable risk and in order to really talk about that we're going to have to talk about with respect to which endpoints so I'm going to have Byrne lead this discussion Byrne would you take take the mic yes I got your cue well I'm not sure that it will take us a long time to initially reach agreement on what are the endpoints of toxicologic response that we want to consider as the prime endpoints and we've seen them outlined we've been talking about them in that context already but I think we need to have a discussion of are we going to cast the net fairly widely and move by brute force through these or are we going to cast the net fairly narrowly at least initially so that we would pick ones that we would most likely be able to reach agreement on for acceptable risk because we have the data to be able to consider that and then if there are endpoints beyond that where it's more difficult we can always come back and pick those up later my suggestion would be that we not try the brute force approach and cast the net widely and look at all kinds of endpoints I would say that we should get our teeth a better defined and a better agreed upon one or two or three and see if we can work out an understanding of what acceptable risk would be for one or two where we have a lot of information and it's a repeatable observations we're spending our time on what everybody would agree to is the important endpoint for considering risk so before we talk about how we would reach agreement on what that risk or that acceptable risk is perhaps we can have a discussion of what are the endpoints that we should really consider and Mike you had I don't know if it's on the slide I don't think it is but if we're talking about the developmental effects in the males I mean it's the range is from it's not here but the range is from you know testosterone starts with testosterone production and then from there on up to testosterone production analgenital distance nipple retention cryptorchidism and hypospadias I think is the range of effects that you see you could add in a prostate weight or something there's a couple other things you could include but those are the highlights and I guess the question is do you want to I don't know that you can have the whole range of effects on all the relevant phthalates and I don't know do you it's a matter of I think in the NRC report was it did you was it testosterone production or lethality I can't remember what the end point that you focused on testosterone yeah yeah I guess that's the other extreme but testosterone levels and that's fine usually for that you would pick a 10% drop you know you pick some number that you consider to be adverse I mean that's a toxicologist kind of decision but you know do you want to pick one end point I mean you can do that apply that to all the phthalates I guess the other option is depending on the data that are available you might use some combination of the end points anti-endrogenic end points but that's I think that those are the options on the table for the male developmental effects you know which of these end points or combination of do you do you want to use or are they at this point maybe still all on the table maybe before we try to take the male developmental effects to a final decision let's look a little more broadly what about the female developmental effects I I don't know off hand if anybody wants to jump jump in I mean there are a variety of reproductive and developmental effects and for example if you expose earlier during gestation there are effects in both males and females skeletal variations visceral the organ malformations and so on then later in gestation is where you see the anti-endrogenic effects and I'm not sure at birth what you see in the females you know there's at least one report where there are behavioral effects later on but that you know it's it's sort of a a mix of things but you know the male pups are the ones that are affected the most and you know that's probably I think a good starting point is to say for any kind of developmental reproductive effects the anti-androgenicity is going to be the most potent endpoint for the active phthalates now if you want to expand that to all the phthalates you would have to look at something else like any developmental effect or maybe perhaps the developmental effects that occur earlier in just earlier in gestation well we're fortunate in this if you want to consider this a good fortune that the most sensitive effect is also a serious effect so we're not looking at a trivial effect it's a severe effect yes I don't know the literature all that well in terms of embryonic exposures to phthalates but I suspect there's not a whole lot there do you know Bern? I think that's right and certainly most of the attention has been what Mike has just talked about there's a fair amount of studies I mean that was the problem early on the studies I can't remember exactly the details but they occurred during organogenesis and so you didn't see these anti-androgenic effects you did see malformations things in the skeletal things heart defects and even kidney defects and a lot of those are covered in the CERHR reports more recently that people started exposing later in gestation during the critical period for sexual differentiation so I think there's there's a reasonable amount of data on both maybe yeah yeah I mean dose response is always I'm not sure how good that is but even in the older studies there was usually an effort to try to find a no effect level so that's one reference point that we would be working from a lot of doses but yeah I mean there were generally no wells so we've talked about the male developmental effects and to some extent female developmental effects those are both categories of responses that we might want to consider for an acceptable risk discussion what about exposure to the young animals or adolescents as opposed to adults and the conceptus is there a group of effects that are observed primarily when adolescents are exposed for example not that I know I think they're just less severe you know similar anti-antigenicity there's similar kinds of effects just that the juveniles are not as sensitive as the fetus and of course even in the adults at high enough doses it'll damage the testes but they're just less sensitive if we're going to be looking at products that children are using after they're born and while they're growing up at a young age what is going to be the end point that we're going to target these estimates of risk I mean you say well you know sort of glibly you know it goes down but you know that's where the target of our investigation is supposed to go even though we are to include the pregnant mother I don't understand we have to be a little bit more firm than say well you know I think you have to consider that they're young children and they're more maybe less susceptible than the fetus but they're more susceptible than adults how do you make that point with some clarity can we make that quantitative that's a real important point that I don't know because I think the data are limited maybe it's something you apply safety factor to and there is there's a recent paper by Richard Charp's group they looked at the effect of suppression of androgen synthesis in the what they call male programming window but also then were able to elucidate that for proper growth and development of external genitalia in these rats androgen action after birth so if you compromise androgen action after birth it definitely doesn't get better there's a nice paper published in international general of andrology the same volume yeah I noticed there's actually I guess there are they I don't know if they're reviews but there's a lot of stuff in that particular issue so maybe that would help us interpret this issue is my sense correct Mike that we're more data rich when we're talking about developmental reproductive endpoints definitely then we're talking about postnatal right so our strategy might be then to use the more data rich endpoints and then hopefully at some point be able to do some extrapolations to children and adolescents and adults where the sensitivity is you know it's decreased well I think this is the time for us to be complete and not have to revisit these things later on when we think we've done the job only to find out that we have to now go back and look at something else so if we can eliminate some things now Paul I think your question is a good reminder yeah because if you look at the next category now in this progression from developmental effects to the adolescent effects now to the adult effects I think the question I would have of what we've seen from the adult exposure data would be the question for example is there a risk of cancer when neonates are exposed and you get liver cancer as adults they're not going to get liver cancer as as young children or infants are there changes in the liver that predispose them to liver cancer later on yeah I don't think anyone knows no I don't think there's any information on that but again it's questionable whether the motive action is relevant to humans that's correct so there's a lot of uncertainty there that keeps this as a low priority concern yeah I think so but the question that I'll come back to is if I'm going to consider exposure as an issue or young child growing up with these toys and personal care products we're going to need something to link to I mean obviously what you said some information that helps us with the mail programming that would be important because that alone is going to have to be the one single thing that's going to make the difference as to how much uncertainty we start with because otherwise it's going to be rather difficult to make quantitative assessments one of the things that Phil and Mike and I were talking about at the break was our need to be very articulate in identifying the assumptions that we make because that I think will be as important as anything else that we would have in a final document is what assumptions did we make as we tried to determine what the risks were and if we're going to come down with recommendations on restricting uses it certainly has to be there has to be a clear identification of the assumptions that we made in arriving at the numbers that we arrive at to make that conclusion are there any other considerations from the adult exposures that direct us toward a potential risk consideration for children well from the I mean chronic studies target the liver and kidney among other things and it's not just tumors but other chronic toxicity there are also we should consider I think reproductive effects in adults and just as an example the slide that's up here right now it sums up what the CERHR has said about this set of chemicals or I guess all the chemicals that they reviewed I mean it covers developmental and reproductive but there are reproductive effects too in you know two gen studies it's not all developmental so for example well let's see I could find a high concern I guess the most serious ones are developmental and the reproductive effects in adults are through the same mechanism for males yes females I'm not so sure yeah I think some of the questions you're answering I think Paul Foster would really be you know the one that would be able to answer some of these because I know we're thinking too about you know who to bring here in July specific questions about how the mechanism may differ in your role versus adult I know some of the data you know I'm not as familiar with the laboratory animal data so I know there are differences and I think Paul Foster would really be able to help us through that definitely yeah he'd be the best person to answer a lot of these questions okay that's a good thing to put on the side and we can pick that up when we have the discussion yet today about the questions and the people thanks for us I think he would be helpful to confirm our suspicion that the critical effects are reproductive and the critical life stage in utero fetal exposures he will also like Russ said be able to tell us something about other concerns and I'm wondering bringing in someone we need probably someone who will dispel our concerns or the niggling doubts we have about carcinogenicity we I think all feel that this is of limited relevance but maybe we should have this confirmed by an expert confirmed or otherwise I'm thinking of who would be suitable Jeffrey Peters he would be fantastic if we can coax him down here I think he'd be willing that for a day Jeffrey Peters is at Penn State he was on the last CHAP he's an expert in the PPAR receptors that's another good thing to add to that list just so many questions that come up you know the developmental effects he Jeffrey Peters had a paper where in the PPAR alpha in Elmice they looked at treated it with DEHP and there were developmental effects even without the PPAR alpha but these were the garden variety developmental they weren't the late gestation the antiandrogenic ones so you know you always wonder all these questions keep popping up and he's the he would be able to answer all of those is that because they didn't look well at the time exposures right at the time they didn't look at exposures they didn't look although they didn't do the experiment but the structure activity relationships are such that they don't correspond to the induction of that pathway so they're all I mean there are a lot of nuances but even if they're it's not uncommon in these developmental talk studies in the standard design to see decreased fetal body weight and even if that wasn't mediated by an antiandrogenic mechanism it's still important absolutely and there are plenty of those effects all developmental effects and the adolescence and the adults and we have questions but perhaps we have looked at the horizon enough in this regard to know that we have an order of priority in terms of the nature of the response and therefore the target organ and it happens to correlate also with sensitivity one question to come back to the male developmental effects since the effects of concern are mediated through this hormonal effect is it enough for us to look at the hormonal effect at the hormone changes and say that everything else biologically is downstream from this so we don't have to look at each one of these endpoints under the testosterone effect individually is it sufficient for us just to look at you mentioned a 10% decrease that way are there could be other ways to say here's the threshold for an effect on antiandrogenic effects I would welcome your thoughts on that because that really gets at how we're going to do acceptable risk what are we going to look at testosterone production or AG it's a little harder the problem with testosterone is what's the level that is adverse 10% as a hypothetical I have no idea what should be but that's a judgment that's probably very difficult to judge but what is clear from the data published is start for thalates the suppression of testosterone fetal testosterone synthesis is the most sensitive endpoint so if you base everything on that you protect against a lot of the other effects you can ask Paul Foster for confirmation there but I think as far as I know the literature and to judge it that's the case with thalates for all other chemicals the most sensitive one is nipple retention and with thalates you're getting pretty close there too so if you regard nipple retention in the RAT model is indicative of disturbance of androgen action in fetal life now there's an argument among some toxicologists who helpfully point out that this effect is not of relevance to the human but that's an argument to be had I mean that's the question just where do you draw the line yeah if you accept that we can't I mean RATs are different from humans thankfully in many ways so we are different from them if you now take the stance that we are dealing with a model anyway and this is a model for disturbance of androgen action in fetal life I think one can get round this problem and the RAT that is indicative of disturbance let's use that and extrapolate to the human although this plays out in a different way in the human it would help me to see the dose responses together like this level of testosterone reduction causes this level of nipple retention and so on that might help me sort it out but just to conclude I think you are also driving at this other issue you have with phthalates effect spectrum typical of disruption of androgen action and that differs in some respects with the effect spectrum which you get from androgen receptor antagonists they are distinct in certain for example flutamide or inclosaline will not lead to reductions in testosterone synthesis but there is a commonality there is an overlap of the effect seen with phthalates in the RAT which is seen with androgen receptor antagonists so one could say we could go for the most sensitive common effect of all these chemicals which I guess would be nipple retention the alternative would be to say we have played this through in the paper which I showed you this morning we have said right we go for the most sensitive end point regardless we weren't looking for common effects or an overlap but that's open for debate really my feeling is that considering all the other uncertainties we have to deal with that's the least of my worries I think if we go there and really begin to wax lyrical about this we are going to guilt lilies dot i's and cross t's rather than address the real problems I don't know whether that was helpful and Mike I think what you said is helpful as well we need to know the dose response here we need to know whether the anti-antrogenic effect the dose response for that predicts these other downstream markers and if they don't then we have to look at those we have to look at hypospedias we have to look at nipple retention we have to look at others in addition to the effect to the anti-antrogenic effect if in fact you can't look at a dose response curve for the anti-antrogenic response in the animal and predict that yes this is where we would see nipple retention this is where we would see hypospedias I mean I think that's true that they follow but we need to have that actually Earl Gray might be the best person to ask about that and we may want to get into this discussion a little bit more when we talk about acceptable risk because that's where it becomes extremely critical at this point we're just trying to identify what are the end points that we're going to focus on so maybe we've for now maybe we've had enough of that discussion and we could proceed to thinking about you know we have a list now of end points and responses that we would use as the model and if we use these somebody else isn't going to come along and say well they've failed to include a very important toxicologic response this one I mean the problem is the problem is it's in the eye of the beholder and you know if you look through the scour of the literature you'll find an occasional tumor site parathyroid or you know the couple of studies about trying to link phthalate exposure to asthma but it's only I mean it's not a strong link they're not strong links and they're sporadic I mean you see it with one or two phthalates I'm just trying to be the devil's advocate and you know there's something for everybody in there but they're just not consistent and I don't see how you know that would be important to include in something where you're looking at total phthalates and I think to cover ourselves if we need to do that in the introduction or some place you list all of the known adverse outcomes and then proceed to say for effects of phthalates on parathyroid tumors there's you know not enough information for us to really go through an acceptable risk analysis for that endpoint and do it that way then we've covered our bases and made it transparent what is recovering and why I think we can do that I think it might also be reasonable for us to now summarize in writing this list of endpoints that we think are most important for determining acceptable risk and we can ask that of the Jeffrey Peters and Paul Paul Foster and others who might come in and ask if they disagree with that I think that's a fair evaluation a fair opinion to seek from them and if they say oh you really missed this one well then we need to consider that can I suggest to maybe also include Grae in that because he's equally expert with these things as Paul I would in plus I would include him sure plus knowledge of mixture effects as maybe we should invite the whole combo it's going to be fun yeah there's always a risk that we want to invite somebody who will be insulted because they weren't invited but nonetheless there's you as a group have a good feel for who are the greatest contributors I mean if I had to pick you know three people to invite in it would be those three so yeah however still taking into the consideration the task of this committee here and screening the last let's say two years of literature in regard of epidemiological evidence I see publications on delayed exposure with less male typical behavior in young boys attention deficit or hyper activity disorders obesity insulin resistance and so on I'm not saying anything about the quality and reliability of the papers and the data but still I would like to have it checked and at least discussed I think it would be a shortcoming not to look into it and that's what I was trying to get at in terms of I think somewhere we have to have a list a table here all the adverse effects that have been reported and we're not making a judgment about quality of data or conclusions but you're right I think that's absolutely necessary but also from the standpoint that we have we don't want to be criticized by the epidemiology community for looking only at mouse data or rat data so I think in order for us to not be criticized in terms of whom we bring in in July to advise us on their thoughts about this we should include an epidemiologist or two if there are such people who can advise us broadly on what the epidemiology literature say beyond in addition to what we can get. I agree I can do that I'm familiar with those studies I'm actually giving a talk next week that includes the Stephanie Engel study on the aggressive behavior and then Shana Swan's paper on the play behavior there's probably four or five so I am very familiar with them I could put something together for July potentially bringing in some of the authors if we want but it's an individual for each of those four or five papers yeah I mean that's that was a big consideration in assembling the CHAP and you know we wanted to cover the reproductive or developmental toxicity and make sure we covered epidemiology and if we could have had more epidemiologists but that's definitely something we can't overlook then I think we should make a decision here do we want to invite an epidemiologist or do we want to rely on Russ giving a summary of I know that we will get good information and help from Russ what I'm addressing is what the image is I was going to say that two issues you know in terms of the interpretation which I feel comfortable doing but also in terms of the balance in the image of also hearing from epidemiologists we could invite Shana Swan or is there someone else you might suggest there's Stephanie Engel from Mount Sinai that's probably the paper you're referring to Holger she works with Mary Wolfe who's also on the NRC report but I think Stephanie would be good was there other papers Holger that you saw? it's at least reference to 6 or 7 papers right now which came up in the last couple of months so I wouldn't restrict it to Stephanie Engel I'm not sure maybe we should discuss it in more detail later on she would really just be familiar with that single paper I have concerns too with the way they group the phthalates by molecular weight but that's beyond today's discussion anyone who would have more of an overview of the epi field with respect to what we're interested in I'd ask like on a swan would probably have more of an overview okay that would be similar to my overview but I think for balancing the presentations and you know the other I'm just trying to think if we invite these people in and invite testimony from the public there's going to be so we may need well if it ends up being a day and a half of testimony then that's okay I think Bernie you were going there too that we should formulate questions for the people coming in so they have time to prepare a question like what percent change in testosterone would be associated with x-fold increased risk of hypospadias in that study or drugidism just as an example and I would hope also that by giving them a lead time that they would come prepared to express the opinions of their colleagues as well and that we don't hear just somebody reiterating what they've already said in their publications you have to be tactful to get that across well without insulting them they're just presenting their research we have a specific question a set of questions we want their opinion but we also want what their colleagues would say because they're representing their colleagues as well want the state of the art is what we want them to give us so Phil maybe we've taken this as far as we need to at this moment if you're ready to step off into the other discussion well I'm wondering in terms of questions for these individuals should we leave it at all of us will submit questions after we review whatever literature we wish to Mike who will then convey them to these individuals that we're inviting is that the strategy? I think we would benefit from some more discussion today because hearing the opinions of others is very helpful at least to me you mean in terms of formulating the questions well not wordsmithing but at least what kinds of questions do we want to have them be prepared for and then we can wordsmith it and I'll help you with that Phil we'll also just add to it too because you want to proceed developing questions ideas for questions for these individuals that we're going to invite I think today we need to have a discussion of what kinds of information do we want to have brought to us and then who can bring that kind of information and then a third approach would be what are the questions that we specifically want to ask of Paul Foster and Jeffrey and Earl others and also as we in terms of the public testimony we're inviting testimony from the public we could spell out the questions we have like about product uses or unpublished data whatever it is we could pose a list of questions and say this is what we want to hear we might not get it but that's usually what we do let's proceed how do you want to go back to the discussion of what do we mean by acceptable risk and develop questions in that arena wherever you want to go next that would probably be beneficial to do that first because that may then generate questions that we want to provide to these individuals who are going to come and also to the public to do which to go to the discussion of what do we mean by acceptable risk do you want me to follow that or are you going to no go ahead where wants to jump in on that I think it would be real easy for us to reinvent the wheel here and for some of you who have been involved in these discussions before at the NRC or other places where you have already had this discussion will you please start this discussion for us and then we can modulate it as we go and add to it as we go but for those of you who have already done this would you please start us I think if I remember right from our discussions we started off I think talking about the multiple potential endpoints we had a little bit of discussion on cancer, hepatic et cetera and I know Paul and Kevin contributed a lot there and so I think we started off very broad but then we I think relatively quickly narrowed down to the anti-androgenic testosterone effects and if I remember right I don't think we spent a lot of time on other endpoints and I think part of it was maybe the data but also the sensitivity of the response to low dose so my question would be if you already narrowed it down when you went to the NRC what is it that we're going to get that's new from these three highly qualified people the question I have is what kind of questions we're going to repeat to say the question of what are the most important effects of concern or what are the new data that can help us reduce the uncertainty about this particular suite of effects and have we learned something that puts something else on the table that would be my question the second part of your state that it would be really trying to new data it hasn't been that long but there probably is some and then I think there are thoughts about some of the questions related really to developmental reproductive effects rather than I think we can relatively quickly narrow it down but then I also think bringing in Peter Jeffries in terms of making sure that the hepatic and the liver cancer is something that we should not pursue before we confirm that decision Could I add to that I think the aim to invite those is not really to generate new information we can read, we can all of us read but I think it is to remove doubts that might linger about a special take which we have already of the situation we all think that developmentally reproductive effects are the critical one and we think that cancer is probably not so urgent but that's my interpretation of the literature and others probably do the same but I would like to have this view confirmed by the experts so only that it's rather than emphasis on new data we can all read them it's sort of matters relating to judging identities etc etc so if you're focusing on reproductive my question would be to them is there enough information to link or at least suggest the linkage between the reproductive effects and the effects that occur in a young child after birth that's very difficult that question we will not be able to answer no one can with certainty at the moment but can at least shed because that's going to be a critical question that we're going to have to grapple with and I would love to hear a discussion to me because it's going to be the toughest question we have a critical discussion as to whether or not that those links can be made, what degree of uncertainty and what pitfalls will there be in trying to make those links because I think if you're going to have three highly qualified people coming in they're going to help us with this task to me those are the kind of questions I'd like to see answered to whatever degree possible you know I agree I think there are good questions that we should strive for answers to but I don't think but I think within their respective fields they can give us perspectives of the animal and the human data and then you know discussing links but I also think you know as you were saying the new data we could read but there's also there was talk about NIEHS putting 30 million dollars towards new Thale research so it would be interesting to know where they're going with that I mean is so potentially the new data if it's published we can read it but there are kind of new directions or thoughts that Earl and Paul can provide in terms of you know what next, what are the next steps the data may not be ready for our conclusions but I think it will help inform us if they let us know where they're going it lets us know where some of the gaps are because acceptable risk question is something that I've been involved in those discussions many times but I must have said I don't listen real hard because it isn't something that I've had to do from day to day so help me understand if we say that we've arrived at a decision that a risk is acceptable or not acceptable for a chemical for certain age children for certain endpoints is this a bright line kind of a decision or is it like we make for drugs and foods and other things where there's a margin of safety that you consider or is it based what numbers do you feed into that last vote when you make a decision of yes it's acceptable or not can I well I'm aware of those discussions in the cancer arena where what is an acceptable risk one in a million or something like that that question doesn't shape up like in this way in this particular arena the discussion there is rather what is an acceptable or unacceptable exposure and that's resolved by making those comparisons between points of departure say from relevant animal models and human exposures and then judging margins of safety or etc etc but it's as far as I'm aware not answerable in the sense that's done in the cancer arena one in one million etc so we would have to rely on dose or intake comparisons okay for these endpoints that we've talked about the developmental endpoints we would know no effect levels or we would that's one thing that we would nail down from the animal studies so then to create a ratio and a number what is the exposure data that we need to have because eventually our recommendation is going to be the level at which these phthalates are found in products that children are exposed to is either acceptable is either above or not above a threshold that we decide yeah so we're going to recommend either to lower it or leave it alone then what numbers do you put in that ratio for us to say okay it's okay or it's not okay if the no effect level in an animal model is one of those two numbers what's the other one the other one one would be intake human intake kids, adults pregnant women whatever well it would be the if you know what the no effect level is and you have an idea what the intake is from the products we're concerned with the level in are they within reach of one another are they three ores of magnitude difference that will allow us to determine whether or not there's some margin of safety we may have to apply to some number to say that this is acceptable and does the way in which we use this product today provide enough margin of safety for that intake to ensure that we're well below a no effect level I mean that's the logic I would use can I just sorry Paul remind all of us that we also according to the charge have to consider intake from all other sources so not only from those toys I'm talking about in terms of banning a particular toy we have to look at it relative to everything else and if it's minuscule compared to everything else which is what the only thing you can do then you may or may not want to change the number right now or desire banning it yeah I mean the question is when we talk about everything else how details does it have to be is biomonitoring good enough because that gives total exposure like Paul said you estimate exposure from the children's products compare that to the total exposure or combine it cumulative risk assessment and so on I don't know that we necessarily have to articulate all the what all the other exposures are like so much from food so much from cosmetics I mean we can but it would be nice we don't necessarily have to do that I think well it says very clearly in the charge that we do well we said we have to consider all sources of exposure that's right I think Andreas is right we have to do that we have to to the best again going at number six we have to quantify to the best of our ability so that when we make this determination that either these products are or not significant proportion of the intake that we have to consider that has to be a reality that we have to address and make decisions about we just can't leave that well you know we know that there's a high intake from food well we have to really be able to provide some degree of number generation plus uncertainty around those numbers and can do a comparative analysis of comparative risk with respect to what we suspect would be the intake from a toy that includes the material in a child's world there are several more specific things that I think we need to talk about one at a time and Paul one of the things that you raised was the proximity of the no effect levels to the exposure usually we look at that from the standpoint of a margin of safety that's acceptable so is a thousand fold difference from the well the threshold of concern or is it 500 fold or 10 fold or unity there's so many different experiences out there that people have had with these things that you can have a whole range of numbers and you have I think we have to make our best judgment based upon a number of factors including the source the frequency of use the likelihood that the levels may increase or decrease they're the number of other alternative ways of getting it's really expert judgment at that point and we're going to have to wait a while to come to that but it's good putting it on the table the thing that I'm concerned about though is we're not in this discussion we're not pointing out the fact that for one chemical it depends on what the other chemicals are it's not just the no effect level for a single chemical you have to have that in context to other things I thought we were talking about all the phthalates we are going to be considering as a group not just one phthalate I wasn't thinking of one phthalate I think your point is that whether is this discussion related to one phthalate at a time or is it the composite it's the integral I thought but except the way to get to that is to do them one at a time and combine them that's one way to do it or did you mean that when you discuss margins of exposure which normally don't take combined exposures into account do you mean that we have to increase the margin or make allowance for that possibly we do or maybe even play a game of what if these chemicals are at this level then what is that margin some kind of a sensitivity yeah there will be a sensitivity analysis and I think that is a progression as we move forward but I think your point is right we can't lose sight of the fact that this is a suite of phthalates and not one and what I felt comfortable with Andreas's presentation this morning is that we're not dealing with a complex degree of synergism it looks like linear additivity has some degree of confidence in terms of demonstrating effect in the animal so I don't feel uncomfortable right now I guess kind of what I'm looking for is whether there are presidents that have been set that say that a 10 fold margin of exposure is the standard or 100 fold well for most people if it's based on animal data would say 100 fold if it's human data if it's children if you're worried about sensitive populations some people might want like a little more yeah up it a little bit 110 and 20 are the ones that I hear floated around a lot there's the sort of defaults you know you could use you can set them at whatever you want but those are the defaults I guess what I'm trying to anticipate Mike is whether we're going to go 99% of the way to making a decision and the last percent we're going to be arguing no margin of 10 isn't enough it should be 15 or it should be 30 we might end up doing that but then it becomes a unique judgment tool of this committee as opposed to a state of the art kind of approach not no pressure but one of the points one of the Roman numerals is to use all appropriate safety factors so you know to make sure you I guess that means to make sure we're protecting sensitive populations and that's one approach if you for example if you don't have data for juvenile males you take the adult data and add another safety factor I mean that's one way to do that yes well or three or two or whatever but if you have a specific mechanism to say that a factor of 1.3 is scientifically justified you can do that but it doesn't happen very often but the only thing I would add to that is to use these these considerations the appropriate approaches appropriately and to not just use them because they're there and we keep adding on factors we totally agree to be comfortable in this committee with making those decisions or do we need other outside input well there's a huge list of people who would like to add their names to those who get to vote on this and we can't bring them all in and we can't entertain all of their biases in how to do this so I think it comes down to the comfort that CPSC has in having selected this particular team I mean we're put this way I'm comfortable with whatever the panel decides okay so that that helps me having had this discussion to know that we're in a range of a margin of safety that 10 100 and maybe us more than 100 if it is need be and we can tailor that to the situation as we get there in the other question now that we've discovered frequency distributions and don't just look at point estimates now I mean no one can answer this do you apply it to the average exposure do you apply it to the upper bound 95th percentile you know I think it's the answer might be to the median exposure but sometimes you use upper bound yeah 95th percentile that's where I think we should get some guidance from approaches that have been used well let me put it this way you know because you brought this up in the data you have a 95th percentile exposure is that person at the 95th percentile every day or we just hit them on a bad day well that's the point if we're going to have somebody do some calculations of risk we have to make sure that exposure profiles we're going to have to make sure we understand we define the scenario form and define in such a way that we can come up with a reasonable number yeah I think that's an important point chronic exposure, point exposures and I think in this context we might also have to discuss the concept of the TDI chronic exposure if it is relevant for the delights or if there is a single exposure enough at a certain point in time during sexual differentiation so I think we might discuss this whether a single exposure might be enough to set the damage and I think after the slides Andreas has shown I think the red data shows there are certain windows of susceptibility that we might have to take account of but I think or Paul might say something also a discussion every under the curve or concentration and just adding is kind of pharmacokinetic differences between rats and humans because I know in those studies with the critical window it's a single dose but is the half life similar or not and then how is it metabolized etc any responses to that comment I don't know I suspect that in humans the window is broader no it's probably broader a critical window for exposure to the fetus yeah it's probably I guess it depends what you mean by broader it probably ranges in rats a few days I'm not that familiar two days in humans it could be first trimester longer in weeks but potentially a single point within that window maybe just as relevant in a eight week window as in a two day window that's a good point a couple of other points that I would like to have a little more discussion on and so we've talked about the animal data and no effect levels you raised the point Andreas I think earlier about uncertainty and what this could include the quality and quantity of data available and I think we have talked about endpoints where we're not looking at a single study and somebody's interpretation of a single study and the possibility that it's not even repeatable we're looking at pretty solid endpoints here I don't think there's a large uncertainty there may be uncertainty about whether nipple retention has an application to another species but from the mechanistic standpoint if it's an antiandrogenic effect it should predict other antiandrogenic effects in another species where that hormonal activity is developmentally important so I'm not too worried about being able to defend that if we get into that corner but from the exposure standpoint what is the quality of data and I'm not being critical I'm expressing my lack of knowledge and the quantity of data for us to be able to know the exposure sufficient enough to match the quality of data for the biological response it's going to be hard we're going to have a lot of uncertainties we're going to and we're going to have I think you and I are going to have to sit down and think about what are the scenarios that are the most realistic and consider what kind of distribution functions one can ascribe to them it's going to be a challenge and for little kids it's going to be tougher because the data is sparse on activities for kids you have the simulations you have the adult surveys I hate to say this but memory of parents on what kids do is rather very poor we've done studies where we've had the parents fell questionnaires and we've done videotaping sometimes we wonder where the parents were during the entire day because they don't match up at all and it makes things a little bit cumbersome when you're trying to deal with it but those problems aside we still can we can still come up with some reasonably logical scenarios and it's a matter of in the end where you all think when it always comes back to toxicology is it the cumulative dose of all these low level exposures is it the periodic frequency of the periodicity of frequent infrequent or frequent high doses these are all considerations that we have to have to be able to come up with a meaningful exposure and those are questions that I think we need to have answered because exposure is just not a number it's a number with a biological response and a biological half time associated with somewhere well I think we're going to have to have a fairly important section of a report on exposure taking into and not just generalities but there are some very specific things about this we're talking about sensitive periods of pregnancy and we're talking about chemicals that that are rapidly absorbed and rapidly eliminated but they're stored in fat those two things drive some critical considerations for risk and one that was made before is very important and often isn't taken into given isn't given much consideration and that is that a single exposure at the critical time can be more important than a prolonged exposure that gives you the same area under the curve but may never reach a critical concentration to cause the effect absolutely so that is the kind of thing that we need to put in there as statements of our understanding of how to reach a judgment about risk assessment and that's for one endpoint it may be well be the peak exposures but for another endpoint it would be the average ones and also I think we need to be able to address the question are there any phthalates in this composite mixture that by themselves are of considerable concern for risk and whether there are or not how much greater is the risk when you consider the do they have co-exposures or are they different with them as a total exposure or are we dealing with them because they are unique in terms of the source that they can go on and be on their own as a separate entity in this risk assessment but there are critical questions if we come to a conclusion that the composite exposure to phthalates is what is important that doesn't help the manufacturers because their product doesn't have that they have one or two phthalates and they will back away from us and say well you're talking about a situation that isn't ours fortunately as a result we don't have to do anything when in fact that's why we need to be able to address the individual phthalates and say yes or no on them as well as the composite that's a good point not to lose because again it goes back to we're looking at products products will have specific phthalates we're not looking at one product that has every phthalate and the differentiation of cumulative exposure versus product and single phthalate exposures complicates the equation rather dramatically but it is the real world so we have to focus on that as well so for us to ignore that would be a terrible mistake it would be a disaster Mike a question of how we capture ideas there are a lot of things that we'll talk about there occasionally is something that we would like to have put in the parking lot and save it to be able to identify the list of things that we put in the parking lot whenever we want to to make sure that we don't forget them how can we do that I mean I'm trying to take good notes and what the best thing to do might be for us to compare notes by email after the meeting of what action items or as things as you said parking lot items I'll go through my notes and try to pick those things out okay Andreas you had a comment I'm trying to go back to your issue with uncertainty which I think is a very important one the way I see this going is as follows roughly I think we have to go through a process of quantification maybe quantifying the unquantifiable by starting with individual salates and then putting it together maybe the hazard index approach provides a suitable framework if we accept that that's what we would have to do but then if we must be careful in my opinion not to get too hung up about numbers and taking them as absolutes in view of all this uncertainty that's I think the main issue one of the main issues we have to face or deal with at some point is when we know there is exposure to chemicals that have potential in vivo antiandrogens when we know already they are active in in vitro assays we can never be sure absolutely until they are tested but within the lifetime of our deliberations and this committee I don't see that this situation changes at all so we can before we guilt the lilies with quantifying this for salates I think we have to be aware of these major sources of uncertainty and take this into account by choosing better margins of exposure etc. or maybe we should funnel this into a qualitative question then at the end taking into account all available evidence and on balance are we of the opinion that exposure to salates should go down or not and if the answer to that question is yes then we then consider the charge in relation to proposed bands or whatever but I don't think we should get too hung up about numbers and I don't think that the outcome of our deliberations will be a risk number or something like this you see what I mean you're saying that we should do weight of evidence in terms of how we're going to make our judgments I think that the data situation will force us into something like that we'll see it may be the right way to go but I I think we're going to have to come up with a number and when we part of the consideration I don't know where you put this in a report is that as we talk about possibly the need to restrict the inclusion of a phthalate in a product we need to be cognizant of the fact that that will drive a replacement of often unknown we'll know less about the replacement chemical than we will about what's being taken out that's a risk consideration as well we need and it doesn't mean that we should do nothing it just means that we need to be aware that that will be a consequence of our decision absolutely and have that in there some place I just wanted to when you said that to put something in the parking lot and that's about the alternatives because some of the or all the experts were thinking of bringing at least for the July meeting are those that are familiar with phthalates of the alternatives because that's something we're going to have to look into so I don't know now is the time to discuss it but just to put it a bookmark there so we know to come back to that from my experience in the chemical industry people from the major companies know what's going on in all the companies and could express thinking about how do we select alternatives and based on what criteria based on what kind of a database would a criteria, would an alternative be accepted, be considered and if perhaps we would want to ask somebody to address how are alternatives selected one possibility would be to ask to solicit testimony from the companies on that certainly but they obviously have internal thinking they don't want to share with the public right but they can explain I'm sure there's a lot they can say yes in general and I think it would be a very good discussion to have it would be good for us to hear that I'm not suggesting that we're coming close to a closure on this but one thing we want to find out if anybody is quite uncomfortable with the philosophy that we have been talking about as an approach to making this decision because if somebody is quite uncomfortable with this general approach it would be good to have some of that discussion now rather than on the day that we're trying to make a decision I'm not asking to put you on the spot but Paul I just want us to be cautious about not coming up with a number because that's going to leave us more confusion than anything else I mean we have to come up with a bright line or moderately bright line somewhere or else industry is going to have a problem with dealing with it in terms of what's going to be in their product the other communities which are concerned will have a problem dealing with what do we mean by the certainty that we're ascribing to this it's we have to just sit down and think hard and I'm not asking for an answer now I'm just saying that as we go along these deliberations we have to really consider what it is we're going to recommend and how what form we're going to put the recommendation in so that it's understandable to all those on each side of the issue I second that even if you know maybe that there are parts of this where we can have numbers in other parts where we can Andrea can I just urge you to look at this and AS report signs and decisions this is very helpful to us they and I don't think we can afford falling falling back behind that standard and level of discussion I sympathize if we can I'm the last to be against coming up with numbers but my prediction and my feeling is that we are not going to be able to do that and we have to deal with this in some other ways that's not to say we're going to be unable to come up with useful recommendations I don't think that but they're probably not we have to brace ourselves for the possibility that these are not going to be numbers necessarily or maybe at the moment I would like a plea here to say let's keep this open and let's keep suitably flexible before closing doors on certain thoughts yeah I think that's good I think all of us probably have been in these committee situations where we're trying to avoid a bright line because some people don't like bright lines but one alternative that is high, medium and low levels of concern where there are just as many people who find that useless to have some committee that understood what a low level of concern meant the day that they talked about it but there is no way to explain to anybody else how this translates to any other decisions so are there any alternates are there any other ways to do this we have a bright line and we have a high, medium and low level of concern is there any better way to do this for those of you who have done more of this than I have then we're not going to lock us into one or the other but we're mindful of this just leave our options open and do the best we can I mean we know it's not perfect and part of the problem is we know the limitations better than the audience does but I just don't know of a better way to do it in terms of justifying a recommendation if the recommendation will be to take some regulatory action some kind of quantification a hazard index or something even with the uncertainties makes it a much stronger case so for your purposes you don't want us to necessarily be satisfied with a high, medium and low level of concern you need more than that I would think so yes because we're not going to make a decision we're only going to make a recommendation true well you're going to make a decision to make a recommendation in fact I will make a recommendation and the five commissioners will make a decision and that decision obviously has to do with parts per million in a product ultimately yes what other things would be good for us to discuss today on this topic I think after the meeting one of the things I could draft something up for a federal an announcement of the next meeting and say we're looking for testimony on these topics and share it with the panel so you could mark it up or whatever I think that's something we would want to do as soon as we can and then as far as questions for the speakers I've already jotted a few things down and we can continue to expand that I'm assuming that that's another part of this discussion that's going to follow so if I'm sounding like I'm coming to a closure on that I didn't mean to yeah one thing we might want to ask EPA and they give a presentation is what level of analysis they may be able to provide on these other sources of falleys and includes food, dust and other food pathway, inhalation pathway and dermal that since they are doing something it's important I think to get an idea where they are in the process what they can be able to provide us and at one point in our process we may be able to see something that we can use effectively in terms of dealing with some of these issues and I'm not sure exactly I am going to meet with them in a couple in a week or two I'll see what I can get about where they are in their process but they're you know like us they're in well the ones who would be looking at exposure OPPT is in a regulatory mode so they might not be able to share a lot of things I don't know but who's going to be doing the analysis the exposure assessment in Washington or in RTP well they've got four at least four programs looking at falleys right now iris which is just the risk part not exposure OPPT is looking at potential regulations so they'll look at everything they're going to look at the exposure assessment I think clearly we need to have some information from them and what they're going to be able to what they're doing, what the level uncertainty of their data is and how that can be useful to us and the sooner we get that the better well I have a feeling they're saying the same thing right now that they need to get all the information they can from us and from the chap but we'll see we'll find out about that actually it was pretty much the same thing Paul had said I guess we were having telepathy but I was just thinking in terms of a bookmark again for EPA as we're talking about information that we may want to get and potentially someone to visit us visit with us in July just to make sure that that was on our on our list well we can ask someone from iris and or OPPT to come and talk to they're two different processes one which is dealing with the questions that the burnt was raising and the other one is the questions that Holger and I and Russ have just raised absolutely and we'll invite both just to complete the regulatory survey is there anything that we need from does the FDA regulate the alites in a way that would be helpful to us well they have authority over you know food contact articles cosmetics medical devices and there we can certainly invite them to come as well who am I talking to I've been talking to well CIFSAN which includes cosmetics in food packaging I guess are the main the main groups I mean we can invite medical devices but that's kind of it to me such a narrow a narrow group that may be a low unless there are pediatric devices well I mean there are in little kids who have to have major surgery or even preemies they can have a very intensive exposure but it's a whole different balance it's a risk benefit consideration the other reality is if you want some information from several different centers in the FDA you have to invite several different people it's hard to find somebody who can speak for the whole FDA that's true but also we can we'll talk I'll talk to them and see what they how much they have because I don't know how how much they they really have that can be helpful Andreas can I suggest we suggest to invite someone from the committee on improving risk analysis approaches from the National Academy of Sciences that report Thomas Burkey chair this he's from Johns Hopkins around the corner I think that might be helpful information what was his name? Tom Burke oh Tom Burke yeah Tom Burke is more than one on the cover and yeah that's that's good but give him they have really made specific recommendations to EPA so I think we could utilize that really which questions would you want to ask them in dreams because this would be more meta or concerning a framework or approach to to this issue which we could use they have a wonderful decision tree or matrix kind of stuff developed I think it would be helpful it would not help us to address directly any questions related to satellite health risks but it would help us how to frame the problem how to go about it etc etc that part of the science and decisions that's right yeah he chaired that committee one of the disks Mike it's it's not but I could look certainly look into getting it I think that I have it on my computer you can have a copy straight away in fact I'll play it onto my brain but with with that report I think the critical questions that I would like to see is they have the decision tree but how do they deal with the decision tree in approaching the mixtures question that we're really seriously gonna have to deal with and I think that that's the kind of framing of the discussion I think we'd like to have with with Tom so that he at least knows where we're coming from and what our critical concerns are you know the endpoints how you going extrapolation endpoints from reproductive to early childhood the questions of the mixtures how do you put it into something that can be quantifiable I think is a good point for departure with for him because I think it's something that you've been you just articulated a little while ago you're gonna struggle with well I've got the report your emails are in there there is a page somewhere with everybody's emails let me find out where we are I think we're done with the discussion about acceptable risks and we've now morphed into this issue of what data do we want for the July meeting and I don't know if we're done with that discussion or not yet you had a list of questions that you're going to distribute to each of us for further input yes I think here it is I just jotted these down while we were talking and we'll I'll share this and we'll add to it or whatever refine it a little bit these are just some bullets we'll try to frame specific questions for like Jeff Peters and Paul Foster yes so some of these will have specific individuals attached to them and we're hoping that people who aren't experts behind these questions won't attempt to answer them yeah we just don't have time but in addition you are going to be contacting a broader list of people on a standing list well you're going to announce we'll invite those particular specific people that you requested we'll also just publish a federal register notice and something on our website asking inviting people to present testimony probably some of the people sitting here today will be presenting testimony we can expect people representing the fallate manufacturers and so on substitutes as well as others so we will do that but we can frame some specific questions for them like if you we can ask about substitutes for example what the process is that they go through what are the considerations and you know I'll draft something up or I could unless you wanted to take the first stab at it but I could draft something up like you know we've done for other meetings start with that and share that with you Mike I think that's a helpful way to proceed so if you want me to take the first crack at it yep we give our input okay I mean I as long as that's okay I just offering is there something more that you need from us today to be able to do that today you know I don't think so I think I'm in pretty good shape as far as that goes I'll you know draft stuff up and circulate it and I'm sure you know the panel and so on but we'll get started on that planning that meeting soon okay Ross I just had a I guess a question in terms of staging of this I mean because it seems like there's a lot of people and information we've listed would it be you know information overload for July and then we're gonna come up with some new questions and needs for our next meeting which will be in the fall I mean is should we because it sounds like there's like five to eight or more presentations that we may have you know I'm gonna be too much I was wondering about that and would you want to break it up have the public testimony in July and then the scientists the invited speakers at the subsequent meeting or in person it's something to think about I think we want we want input from the experts as possible and if we have to have but I do think we want the public input too as soon as possible also so this just could be a data dump July but I'm just wondering if you know if it's two days full days of presentations doesn't leave us a lot of time and then we're probably gonna come back with you know other needs I don't know I'm just thinking in terms of we may want to prioritize well I think we can try it yeah I think last time we had you know a day of public testimony and so if we have end up with a day and a half of listening I don't think that's too intolerable to digesting the information so not that they'll data dump all this information on us and then we leave a few hours later but we I think we need time as a panel to digest it the talks the actual presentations we can keep them limit the time and they can submit lots of data and give a short summary Mike how about if you would put together an agenda for these public testimonies both scientists, EPA, FDA and the general public and with some times associated with that then we can start to we can do that I mean the only thing is we don't know how many people are going to want to give, present until we ask but we can put together a straw man or something like that say that we like Earl Grey to just talk with us for an hour not necessarily present for an entire hour but have a dialogue going on maybe that's not enough I don't know but maybe have the three experts for a dialogue for an hour that would be more effective at the same time because then we won't have to ask the same questions over and over again we have these three very competent people in the room we want to just get a information exchange then we will need to take a more active role in the moderation of that discussion so that we're not here at six o'clock in the evening from nine o'clock in the morning listening to those three people yet we could devote an entire morning to those folks and maybe the afternoon to from EPA, FDA and again have them but I would also just caution to make sure that there's time for us so if they present all morning block out two hours so we can talk about what they presented rather than two days straight and then we all come together on the third day and we don't remember what if we can't internalize it won't be of much help to have them come and just talk to us have a block of time for us after each so the type of three coming for two hours and then we block another hour where we sit and talk about what they told us that would be good and have EPA and FDA and NIH just talk have a block of time at the end of that for us to talk that would be fine now because that will be an open public meeting we may have two hours of them answering our questions but then they'll sit in the audience and hear what we say about it and that's the way it is that's a public meeting but I mean there the chairman I mean you're inviting them you know they could sit at the table if that's what you want and participate in the discussion for as long as you want so are we talking about like having the invited speakers on the first day how do we want to do this why we'll be here on the first day remember so it's up to you guys I'll put them on the second morning okay I'm here probably comment on that first morning okay got our agencies in the afternoon and the experts on Monday morning or the second day morning I've just emailed everyone their report and maybe we're done with this piece I'll turn it back to you I think it's time for lunch that was an easy one