 So we'll now start with the presentations from panel three and on my list Amy Bachman is First okay My name is Amy Bachman. I Do have a PhD in toxicology. Although it doesn't have my credentials in here, but We do have a good panel this afternoon. I think Representing ExxonMobil and ECPI I wanted to start out with just a couple comments before I got into my presentation just to point out a few things for everyone ExxonMobil has provided extensive written comments I believe it's on a CD that will be provided to all of you if you don't have it already There's quite a number of documents in there The main document does attempt to address all 12 of the questions that are posed to the chap And there's plenty of supplemental data that are provided in there I'm not going to go through the full list of comments But if you do have any questions on all the attachments that are included, please do let us know So I'm going to start out with kind of a general overview of dinp and didp Points of interest that may not be covered or may have been covered Already or presentations to come And then I'll be followed by Nina Hallmark who's representing ECPI Today on a presentation of endocrine disruption. She will be followed by Alan Godwin who's also from ExxonMobil talking about uses and various phthalates So my outline then includes just a brief Point on phthalate chemistry the hazard characterization comparison that we've set up and Just a little bit on cumulative risk I just want to make a brief point about that although you'll be hearing more about that later this afternoon through ACC presentation presentations The phthalates do constitute a broad class of chemicals with a range of physical chemical and toxicological properties The properties are structure-dependent and what's very important and to be very clear is that phthalates are not all the same And they are not all toxicologically equivalent. So in previous presentations today, you hear the broad use of phthalate being thrown around I think that term needs to be very clearly defined because all these phthalates are not created equal Phthalates are widely used because they have technical properties suitable to many applications But suitability is related to specific technical properties Phthalates are sometimes interchangeable. You can hear more about that from Alan I'm not going to go into that but that interchange is not always appropriate for all uses Non phthalate alternatives do exist which you've heard about already But some are not always as technically suitable or well characterized toxicologically or readily available or cost competitive So to put this in kind of an overall table Calling this the overall hazard characterization. So we're looking at all the major end points for R2 phthalates dinp and didp in comparison with some of the alternatives that you've already heard today I'm not going to talk about the alternatives the data that's presented as data that we were able to obtain from Publicly available sources at this point my focus is on dinp and didp and the comparison table up here Is just to show that dinp and didp have low toxicity which is on par with most like the most likely alternatives So when people are talking about how safe these alternatives are I think safe is not the perfect the appropriate scientific term to be thinking about safety has connotations of risk and all of that When you're looking at toxicity to compare the toxicity across our phthalates and alternatives You'll see that they're pretty much on par I'm not going to talk about repeated dose which is highlighted in yellow I think that you'll get comments on the liver and kidney effects for dinp and didp in other presentations And probably tomorrow in your discussions with experts I'm going to focus on developmental just to point out what's been seen for dinp and didp So for dinp minimal developmental effects are thought to be due to maternal toxicity observed in the standard assays Currently they're not classified under the EU classification labeling packaging regulation and UN globally harmonized system It's GHS So for dinp the three main studies developmental studies that you see are listed at the top the references for each of those And I want to make a note that for any study that you see that is ours Exxon mobile is willing to provide the full study report if need be which I believe in the past We have done so you can have the full thousand page document if you choose The critical effect that are seen specifically in the Waterman and the Hellwick study skeletal variations Are the main? Effect that's noted This is seen to really be due to the maternal maternal the dams are being tox I'll start over The dams are considered to be toxic at those doses and so therefore any developmental effects are related to those Effects which you see is reduced weight gain and food consumption The fetal no al is seen to be 1,500 and 200 across those studies But you see the effect level is at 1,000 for each of the two studies that are showing effects And that's just based on the dosing that you see the no al to be different For didp there are two studies again. You see the same types of developmental effects for didp and again The same concept applies here the effect level you're seeing it is at 1,000 your no al based on dose spacing is 500 and 200 physical and chemical properties and exposure Didp and dinp are liquids with very low vapor pressure water solubility and dermal absorption And again, I think Alan's going to speak to the first two of those a Points to be made exposure sufficiently high enough to induce adverse effects and humans are not plausible given the inherent physicam properties of Didp and dinp, and I think that Alan again will speak to that point Just a note on exposure then just as a visual to take a look at this if you plot The 95th percentile dinp exposure estimated from all sources You see how low they are in comparison to what's in blue, which is the adi, which is 120 micrograms per kilogram That's the CPSC adi In comparison to that the doses which induce effects and rats are up, you know four to five orders above where you're seeing Exposures and in that red box you see the liver effects the no al for the liver effects a low al for developmental effects Middle dot and the no al for reproductive effects So I think it it speaks to how low the exposures are specifically for dinp in relation to the adi and where you're seeing effects in animals In comparison to low molecular weight phthalates This speaks to the fact that again as I started out with not all phthalates are toxicologically equivalent And I'd like to point out a couple of differences between the two groups So if you group the low molecular weight phthalates into bbp Dbp and dhep and high molecular weight being dinp and didp You can see as you look down the list and it is a one-to-one comparison across there So we've listed them out Together as you go down the list you do see a number of effects with the low molecular weight phthalates And in the green box you'll see for dinp and didp that those differences exist So you don't see any reduction in anal genital distance with the high molecular weight phthalates No nipple retention. No hypospatias. No cryptorchidism. No sex organ weight decrease No adverse testis histopathology and no decreases in fertility We do have a couple of footnotes to draw your attention to for nipple retention and adverse testis histopath as there have been reports Well at least one single report gray at all for those two end points for your consideration But importantly they're not classified for reproductive and developmental effects as the low molecular weights are in the EU So the NTP CER HR review has also concluded for dinp that there's minimal concern for adverse reproductive or developmental effects And for di dp negligible concern for adverse reproductive Effects and minimal concern for adverse developmental effects. So these phthalates have been extensively reviewed and the conclusions have been Similar to those that are posed up there for CER HR Dinp and didp are not selective reproductive or developmental toxins Dinp and didp are not endocrine disruptors and a needle hallmark will be speaking on the endocrine disruption endpoint So you'll get a full 15 minutes on that dinp and didp do not pose a risk to male reproductive develop track development Okay, and a note on cumulative risk so the main point of this is not to Extensively look at what other people have done for cumulative risk or even to analyze, you know, how they've conducted that cumulative risk These are two that have been published Dr. Courtney camp. I'm sure you're very familiar with your own data And I don't want to misrepresent that data in that we did pull out those six phthalates that were in that report But there were a number of other chemicals that were included in that chemical risk assessment But if you just compare those six and you create a pie chart for kind of what's the Contribution of each of the phthalates in that group to the overall Toxicity of that mixture you see for dinp, which is included in both of these instances that it's very minimal to negligible So the three main low molecular weight phthalates are really driving the toxicity of the mixture So for dinp and this is regardless of endpoints now endpoints were chosen in these two studies for their own You know choices, but if you look across all the endpoints, I think It doesn't matter what end point you're gonna choose Dinp and didp have low toxicity and they have very low exposure and those two factors put into a cumulative risk assessment Are gonna cause them to kind of fall out of that mixture in terms of adding to significantly to the toxicity of mixture Because that toxicity is going to be swamped out by the low molecular weight so in conclusion The high molecular weight phthalates are dinp and didp are widely used in commerce The uses relate to structures of specific molecules high molecular weight phthalates again are not toxicologically equivalent to other phthalates The imp and didp have an extensively tested and widely assessed and I've listed six Points on that so ill see RSI and the CPSC have assessed the carcinogenic potential The NTP assess the reproductive and developmental effects. The EU is conducted comprehensive risk assessments Their specific uses have been considered by the CPSC FDA and CIR There's very low exposure that's been estimated via biomonitoring data and dinp and didp are currently reached registered I do want to make one comment on carcinogenic potential of dinp and didp We did provide extensive comments on the kind of recent hypothesis on Alternative mechanisms to carcinogenesis in our comments. They're part of our prop 65 comments that were provided to California I believe last year so we are we address in those comments a number of the studies and it's not just a select few There's quite a body of literature that we looked at beyond ETO and Geiten So there's quite a number of other mechanisms that are being considered and papers to consider in that in that discussion We've covered them all So the final point then on cumulative risk contribution to an overall cumulative risk of phthalates by dinp and didp is minimal Due to low exposure estimates and low toxicity That's all I have Thank you questions on the committee Like you look like you have a question Yeah, which phthalates does Exxon mobile manufacturer dinp and didp Thank you for this presentation I Think there can be no doubt that dinp is not as potent as some other of the Low molecular white phthalates But I find it a little tendentious that you Seem to ignore Two pieces of evidence which you quoted in the footnote. I just spotted it That's the study by our gray 2000 gray et al and Bosch et al 2004 Where they showed that the INP? Can reduce fetal testosterone levels Now it is well known that you know all the other low molecular weight phthalates do the same and they're not very potent at say inducing changes in AGD or nipple retention, but would you not? agree that the fact that we see Reduction of fetal testosterone synthesis points to Activity in terms of the phthalate syndrome It's very hard to say based on that one single study because there was no dose response included in that study It was a single dose at 750 milligrams per kilogram to study For decreased testosterone there was the Adamson study, which saw no decrease in Matsutomi study one of the other ones saw no decrease and one saw a decrease from my recollection But I can certainly look in that to clarify the effects But regardless I haven't seen a dose response done for testosterone endpoints for dinp So it's very hard to make a determination as to what How potent dinp is in terms of decreasing testosterone and what effect if any That continues on to be because in the two generation studies, which we have for dinp and didp for that matter You don't see any of those effects That you would expect if it was in fact causing a major decrease in testosterone leading to further effects due to that So taking the two gen as a definitive study I would say that the decreases in testosterone are not contributing in the way that they do For some of the low molecular weight phthalates Questions one one more Along the same lines Yes, you're only talking about the phthalate syndrome types of effects, but are there not other developmental effects such as in the hell they get all and Waterman Studies right so those are the effects that I showed in the table Which we do not believe that has that has anything to do with it being an anti-androgen the dams are They're getting doses that are toxic at that point and so therefore the developmental effects the Skeletal variations that you see are due to that and not excuse me not due to any sort of anti-androgenic potential of the chemical It's a clarification or question on slide eight where you Estimate dinp exposure from all sources Did those studies measure was it the monowest or minp or because that's a very small Percent of the metabolites or did they use the oxidative metabolites? I believe the woodus like paper Was oxidative metabolites and Haynes and cone was minp Which is only a few percent right of Well, yeah, I mean that that is brought up I mean I think the point there is that when they've compared the exposures that are back calculated from the urinary metabolites of the Mono ester versus the oxidative metabolites. They're really not that far off And I think the the paper the most recent paper from Wittesack 2010 Demonstrates that that the exposures are fairly similar and comparable. I mean you can see that in this chart here They're not that far off if one's oxidative and two of them are minp But I mean the overall conclusion there is the exposures are way below the ADI regardless of whatever Mono ester or you know oxidative metabolite you're using to back calculate exposure So the small differences that you may get I don't think are going to cause That much of a difference in the numbers that you wind up with nor the concern that you're gonna have from the exposures I have one question on and slides five and six For the developmental talks the dosing was done during embryonic periods basically six to fifteen Are there any studies where the dosing goes? more into the fetal period Not studies that we have done there are studies public studies. Yes, I believe so In what the gray study took it out today. I guess gd-19 and then they looked out past that postnatal day So you can certainly and You haven't referenced those here. No, we were just using these as kind of the standard one gen two gen developmental talk studies Is what we wanted to talk about in our studies particularly, but certainly there are other studies that have looked at dinp with substantially similar results or different Different endpoints were examined in other studies. So there aren't any other Developmental, you know standard developmental studies out there to specifically compare to these are the three But if you're gonna look at other endpoints then you can start looking into the published literature for other endpoints So no, there isn't another study out there that has looked for skeletal variations per se but in terms of the Effects that you indicate on slide On page 10 where there was no reduction in the anal-general distance Nipple retention hyperspace, etc. When you dose more into the fetal period Do you see those things or not? No, and that's we're claiming no reduction. No nipple retention. That's based on the published literature I'm told absolutely not just on those. No, not just on those developmental studies. That's in total. Thank you Just like to carry on where well and our chairman left off first of all This table that's not surprising that you don't see very much effects in terms of nipple retention hyperspace, etc with phthalates Because the pattern of effects they induce typically In these studies is different. It's not like you just wouldn't expect these effects very much very weak if anything Let me come back to the three studies which you Convoted I think it was slide five or six Yeah, wonderful so they Where these studies carried out a protocol That examined all the offspring all the males and where these studies designed specifically To look at changes. We would expect for phthalates. IE fetal androgen synthesis reduction, maybe weak Effects on AGD, etc No, these Most of those endpoints were not included in these developmental studies That's that wasn't a standard. I mean the standard OECD 414 protocol was used for this and those endpoints are not included in that Would you therefore not conclude that these studies are not so relevant in terms of the concerns and endpoints we're discussing here Um, if you're gonna use them to look for effects of decreased testosterone and in a general distance Obviously, they're not going to be relevant because they didn't look at them, but I think the overall developmental toxicity. Yes, these are relevant Thank you. Thank you very much Next speaker dr. Hallmark Hey, good afternoon My name is Nina Hallmark and I'm standing in front of you today to try and keep your wake off to lunch I hope you've all got enough coffee. This is Often the way after lunch, but um to try and entertain you and I'm going to be wearing two hats today One of my direct employer, which is ExxonMobil. I'm based in Brussels, Belgium But also I'm here officially representing an industry trade group today the European Council of Plasticizers and Intermediate We're part of a broader European chemicals trade group known as Cephic CEFIC and the ECPI just to keep those acronyms in there is has a broad membership of chemicals manufacturers and Just to show that this is not an ExxonMobil presentation per se I'd just like to point out one of my One of the other members of this group and that's dr. Reiner Otter from BSF So what I'd like to Do for the next 10 minutes or so is pick up on one of the endpoints that I know that this panel has been tasked with Including in the spectrum of endpoints that they're considering for this this commission Endocrine disruption is something that's very close to my heart. It's something I have a deep research interest in And I understand it's something that the public are very sensitive to as well So I think it's really worth spending a few minutes to consider this specific issue Just to outline the talk for just now I was going to spend a moment to introduce two of the chemicals for to focus on I think Amy's done some of that and Alan will pick it up So I'm going to go through that very quickly And come quickly on to the point of this talk, which is to discuss what is an endocrine disruptor How do you identify one and walk through some case studies? And the framework that I'm going to use to evaluate the hazard potential hazard Poised by the chemicals in the case study is the OECD conceptual framework for endocrine disruption I've chosen that specifically because there are many ways that people can Determine the hazard that chemicals pose But I believe the OECD framework is the most robust, widely accepted And it's something that changes with time. So I think it evolves to reflect the data that are generated And then finally share my conclusions with you on two of the chemicals that are the purpose of the case study So very quickly, what are DINP and DIDP? Captured in this slide is a I think you'd like to bring your attention straight away to the bottom of the slide To websites which expand a lot more on this information And I've just taken a synopsis of the details from those home page of those websites and presented those here So DINP is diisononothalate A substance composed of esters of phallic acid and isononol Also known as a C9 C9 thalate because there are typically nine carbons in the in the chains and then it's Similar chemical, DIDP, diisodesylthalate. Also known as a C10thalate because it has up to 10 carbons in the side chains. So these are also known as high molecular weight thalates. So to the endpoint of interest for this presentation is endocrine disruption. And I think it's worth noting that endocrine disruption is not considered a toxicological endpoint per se, not like a developmental toxicity where you have a skeletal variation. It's really indicative of a functional change that leads to an adverse effect. And actually that description is not my own. It's taken from the IPCS global framework on endocrine disruption. That framework came out in 2002 but it builds on and includes definitions of what an endocrine disruptor is. And that's something the panel has been deliberating with already this morning. How you define some of the terms that we that you need to handle during this evaluation. So I offer you two definitions here for your consideration. The perhaps one of the first definitions of endocrine disruption was developed at a workshop in Weybridge UK back in 1996. So this is not a new issue. This is something that's been deliberated for quite some time. And according to Weybridge an endocrine disruptor is an exogenous substance that causes adverse health effects in an intact organism or its progeny secondary to changes in endocrine function. To my interpretation that means you may see a reduction in for example fetal testosterone but unless that is paired with actually a phenotypic change and adverse health effect that would not meet the definition of Weybridge of an endocrine disruptor. And that's mirrored in the IPCS definition which says that an endocrine disruptor is an exogenous substance or mixture that alters functions of the endocrine system and consequently causes adverse health effects in an intact organism or its progeny or subpopulations. And I'd just like to emphasize subpopulations because even though we're here to discuss human health these definitions also imply to environmental endpoints as well. So to come to the the framework to evaluate an endocrine disruptor what I have here is the OECD framework and it may not... let me see if I can make it a little bigger. How do you think we might do that? Oh there's just this guy here okay thank you very much. I'll leave it even with that help and I thank you. It still may not be easy to read from the rest of the room but obviously this has been made available in advance and a link to the OECD website where you can find this has been provided. But really the purpose of this framework is is a tool it's not an answer it's a mechanism to help structure how you organize the data that might be available and how you evaluate it to contribute to a weight of evidence approach to help draw conclusions on the biological complexity and detect the potential hazards for endocrine disruption. This is broken into five levels and I'm going to walk you through each of those levels and give you some examples and some proposed conclusions. As I do that you'll find that I've tried to make this very simple for today and use a color coding system but what I have not done in the interests of time today is to provide an exhaustive reference list but I'd like to offer that as a follow-up to today's meeting so that you can double check the references that I used and repeat that investigation if you so wished. So the first level in this concept is to start with level one it's called the sorting and prioritization based on existing information and I think that reflects where one of the tasks that the task force was given this morning starting this de novo evaluation starting with gathering the information that already exist and in terms of this framework that comes under three headings the physical and chemical properties the human and environmental exposure and the hazard and frankly the rest of this framework focuses on the hazard. Now for the chemicals that I'm going to look at today I've chosen four phthalates, two what we call low molecular weight phthalates the dibutal phthalate and the diethylhexyl and two high molecular phthalates the diosanonal and diocidesal and what I'd like to bring your attention to immediately is it under the EU system. Two of these phthalates are already classified and this is in the bottom row of the table both for reproductive effects and actually DBP for aquatic effects as well. So I think of a reasonable conclusion from this slide is that comprehensive hazard assessment data sets are already available sufficient to demonstrate that not all phthalates are the same and already to enable risk assessments to the point where classification or non-classification can be decided. That brings us on to level two and level two provides a framework to evaluate in vitro assays and there's a list of the typical assays that may exist OECD is not prescriptive they don't restrict you to that information so what we are able to do is to evaluate the literature and interpret the study designs a little more fluidly these don't have to be OECD guideline studies this can include a broader spectrum of study of available information and when you do that it's reasonable to conclude that in vitro data already exist for these chemicals and they are sufficient to demonstrate that DINP and DIDP are not hormone receptor antagonists but to be fair and hence the yellow coding in vitro data for other mechanisms are inconsistent and inconclusive but that leads us naturally on to the next level in the framework. So level three looks at what we might also consider as small in vivo studies that are focused on specific endocrine mechanisms and effects for example estrogenic androgenic and I believe that when you have a look at these data and some of those include eutotrophic assays going back to the 90s we've got Hirschberger assays that were done 2007 and when you look at all these screening tests I think you can already start to see a differentiation between the different phthalates and to pick up your comment Andreas about beg your pardon Professor Corbett about the borsche and the grey data I think it's reasonable to conclude that they should be considered hence the yellow the inconclusive coloring at this point but what that also leads us to is the conclusion that in vivo assay data already exists for these chemicals but the weight of evidence is sufficient to demonstrate that DINP and DIDP are not estrogen or androgen mediators and this is built on further in level four. Now to pick up on a point from earlier we do not have the precise OECD 407 studies for any of these particular chemicals those guidelines came out quite recently but when you take a weight of evidence approach and you look at the for example the chronic toxicity data done in a range of species we've got rats we've got primates even dogs there are no data to indicate any endocrine excuse me endpoints of concern for DINP and DIDP so in these level four data looking at whole animals providing data about multiple end crime mechanisms and effects we're really confirming the differentiation story that we saw hinted at with the lower hierarchy data so in conclusion for the level four again in vivo assay data already exist sufficient to demonstrate that DINP and DIDP are likely not reproductive system mediators and I say this and confirm it with the final highest highest hierarchy tests the apical studies looking at for example the OECD 416 the 414 to a lesser extent but I mentioned the 416 the multi-generation study because that not only considers the susceptible window of concern but also considers postnatal exposure so really extrapolates and gives you an opportunity to explore that worst case scenario again when we weigh all these data in by all these data up we can see that we confirm this differentiation story these in vivo data exist these studies have been published certainly for DINP and DIDP and to reiterate what Amy said earlier though the summaries are published in the peer reviewed literature if you had an interest in full study reports I believe they could be made available to you but suffice to say that not only for the rodent studies and these are typically in rats I'd also like to take this opportunity to mention that environmental studies have also been conducted and there's a multi-generation study in fish in the Japanese Madaka that looked at DINP and they also saw no indication of any endocrine endocrine effects so in conclusion for the top level of evaluation I believe it's reasonable to conclude that in vivo data from apical studies already exist for these chemicals and they are sufficient to support the position that DINP and DIDP are not endocrine disruptors and on this final slide of summarize the conclusions from the previous steps and I'd just like to add a few further comments one of them is that in line with the definition of the endocrine disruptors with waybridge and IPCS I believe that in order to be defined as an endocrine disruptor whole animal effect should be observed not just inferred and finally I just ask you to consider the relevance of humans to the effects seen and I mention that because one of the emerging data sets looks at the mouse and for example Dibutothalate and there they've been trying to identify whether the mouse is as sensitive as the rat for endocrine disruption effects and surprisingly I believe this is work published by Kevin Guido in his group in 2007 the mouse did not share the rat sensitivity to these endocrine mediated effects and to pick up on the comments from the previous panel I think human relevance is critical to this this is why we do this this is all about protecting human health so I would support the idea that further research into the human would be ideal but also I'm aware of the sensitivity that doing experiments using pregnant women is morally unacceptable so I think they have to have to be some alternative approaches and I believe that some of that research is underway in the US and France and in the UK and my understanding is that preliminary conclusions suggest that the human is not sensitive to those anti-antigenic effects seen in the rat the human is more like a mouse so I think I will leave you with that thank you very much thank you questions dress thank you for this presentation I can say I can agree with most of what you've said but one problem one issue for clarification if we turn briefly to your slide number seven I think the level two in the OCD conceptual framework it is no surprise that none of these chemicals produce any convincing effects in these essays are you aware of the new upcoming essay looking at steroidal synthesis in vitro H295R cells yeah are you aware of any data concerning phthalates in that in vitro essay not off the top of my head I have to tell you okay so the INP and the IDP have not been tested in that essay not enough I know it's very new it's currently I don't know where they are with OCD I think it's undergoing validation or it's not yet validated I believe yeah okay the issue I can agree with most of what you said because the problem with these essays in the OCD conceptual framework with most of them is that they're not exactly tailored to look for these effect were interested in many of them cannot recapitulate the phthalate syndrome for example so that is quite widely acknowledged even by OCD and they're working on that so all in all it's not very surprising what you present to us today it is correct but how can I say it sort of somehow misses the important issues we aim to discuss here with endocrine disruption and phthalates in that there's no essay that would look at say fetal reduction of fetal antigen synthesis etc etc they are not validated at OCD level in vivo so this information is really missing and the other issue with a lot of these OCD framework essays is there are serious question marks as to their sensitivity in capturing some of these endocrine disrupting effects I think captures some good points where to begin I think you're right OCD they have developed this framework and really I think the mechanism they had it the biological mechanism they had in mind for this is something that is hormone receptor mediated and to the best of our knowledge we believe that that is not how that low molecular phthalates induce the effects that they that we observe having said that I believe that's where the apical studies come in that's where your multi-generation study come in comes in because it doesn't take doesn't make any assumptions about the mechanism that's at play that's a whole a whole animal assay and I think if you still see no phenotypic changes so adverse health effects at that level to me that that's the answer you're looking for I was intrigued by your last comment how did you put it the human the human is more like the mouse in response to phthalates what is the scientific evidence you base this on is this a published peer reviewed study or yeah the mouse work was done by Kevin Guide I think that was published 2007 and he basically repeated a rat protocol with debutothalate he repeated in the mouse where he was looking at exposure of the vulnerable window and he did the fetal testosterone assays and to his surprise he did not find equivalent results compared to the rat he did not see that a testosterone change he saw other changes but he he did not did not see that androgen change and the other paper I'd bring to your attention was by a French group and I forget the lead author but it's the group of Renee Haber and they were able to get hold of legally human fetal testicular tissue from legally terminated pregnancies in Paris and culture those in vitro and expose those to phthalates and bigger pardon I think it was debutothalate to see if they could mimic the the reduction in testosterone as well and so that's a direct test albeit an in vitro assay you might consider a screening assay it would probably come in level two in this framework but that initial screen saw saw no so no effect on testosterone again it had some indications of other other changes and the other paper I'd like to bring to your attention is one that I authored where also a few years ago but that did a quick comparison of the rat the marmoset and the human using a similar protocol to her bear so try to do a multi-species comparison looking again for those testosterone effects and it was seen in the rat model not in the human and we saw different changes in the marmoset and they're all published is this that's I guess hallmark et al 2007 in EHP leave so yeah and the other one you're referring to as long brood two thousand yeah we are aware of that thanks the results that you've talked about at these various levels of testing are all in the individual chemicals yes but one of our concerns is mixtures if you will for look for example at the test results on level four where you have clearly distinguished DBP and DHP from the other two yes if you tested in this same animal model DBP and DHP together with one of those that are non-toxic in these assays what do you expect the impact would be if you use the same dose level of DBP for example but loaded that with one of the non-toxic phthalates what would the model pick up rather than guess I would refer you to a paper that professor corporate and count mentioned earlier it's by Borsche tell from 2004 and there they took debutal phthalate a big upon death or hexal phthalate DHP and dinp and did a if you like a simple mixtures experiment and that from what I recall of the paper I believe that DHP produced a distinct significantly significant decrease in fetal testosterone I'm gonna say production because she didn't they didn't take the testes and measured the testosterone content alone they also took the whole testes incubated them ex vivo and measured the testosterone they produced for about three hours ex vivo so they actually try to do quite a robust investigation there and I think what what I recall of the paper is that when you add dinp and DHP you don't see much difference compared to DHP by itself so the major contributor of that toxicity would be DHP actually while I think about it I believe that the the dose of dinp was about triple what was provided for debutal phthalate because of this possible difference in there and the strength of their effect any other questions thank you and our next speaker whose name I don't have my business for you so that I have to find that good afternoon my name is Alan Godwin I'm a senior research associate with excellent mobile chemical and I've been working plasticizers for about almost 30 years in various capacities from technology manufacturing as well as the marketing in US Asia Pacific and in Europe so a little bit experience in some of the questions we heard this morning about telling more about what phthalates are being used and so forth I hope to address that now recognize that Exxon's position in the market is really based on the products we sell and enter into the business and those end uses we cover so we really don't have a complete picture where everything is used but we have a pretty good picture where our types of plasticizers are used so if you think about plasticizers and the whole scope I mean there's a study I guess the comments before there are 29 plasticizers other than the six that have been identified on reality it's probably a lot less than in terms of phthalates that have any really significant or in importance even the list that someone else had of several plasticizers they saw were issues some of those are really no longer commercial products even the data changes as we go through with it but we talk about plasticizers and general purpose and we think general purpose those are the plasticizers that are used in a wide variety applications a wide variety processes and just cover numerous end-usans segments there's only a handful of those and that's the DHP the DINP and the DIDP and the DP HP so it's really those four that really take up the majority plasticizer use around the and they just really they just offer the best best compromise or price performance and quality there's a whole host of specialty plasticizers they've been with us for a long time you know this whole technology of plastic as a plasticization using chemical plasticizers is over a hundred and sixty years old phthalates themselves were first discussed in literature in the early 1900s with DOP took off and in the 1920s as inventions to discover very quickly phthalates became the number one plasticizers because of price and performance if you look at a study that was published in 1943 at the time the author concluded that 20,000 different chemicals had been proposed at plasticizers since that time based on information we would say probably the number is close to 30,000 different chemicals in the literature at some time another have been discussed as plasticizers of those you know various numbers went into test marketing and so forth we probably had three or four hundred actually went into commercial production many of the ones we see today or were developed a long time ago I think you heard Eastman Mark Holt talk about DOTP have been around for 40 years but recognize of those categories of those 30,000 majority those were non phthalates which ones moved in the top four or five well it was those DOP the DHP the INP the IDP just because they offer the best compromise of product and performance one of the chart we have over here at the side is looking at global capacity of plasticizers and that number kind of reflects an estimate of about 10 billion pounds annual production of plasticizer and capacity and where we are today with non phthalates is just very small at this point but to get to that area if you had to replace that it's going to take a substantial investment and substantial in that time in order to displace the phthalates and we don't have a copy of this presentation with you but we'll make sure that you have it on your system so again not all plasticizers are used this or phthalates are used as plasticizers and there's some categories and solvents such as you see with the cosmetics and issues some are used as fragrances and issues but we get into where the plasticizers are and those are their added polymers to make them more flexible once they're fused this is fairly stable but plasticizers are not chemically bound they can migrate under heat or extreme stress or with some sort of extracting media plasticizers are generally poor solvents for the for the polymer at room temperature they have a low volatility they have to have the ability to make it more flexible there's a recent study that was just published this year researchers in Spain where they actually took DHP and chemically bound it to the PVC molecule and when you look at the data it sounds nice from the topic of the paper but when you look at it the material is no longer flexible so that so these are all things that take into place for phthalate esters you can have just a wide variety of plasticizers different from all the other molecules the range of alcohols can go from C4 to C13 its combinations is various out there you can throw some aromatic rings in there and you get a whole variety problem or plasticizers you know we look at different plasticizers that around we look at general purpose plasticizers the attributes you like to find or low-cost UV stability adhesion heat stable stain resistance these are all categories that someone making a product is going to look relative to their plasticizer a lot of alternatives around we hear things talking about citrates and oxidized so we've been all or natural products or DOTP or in dents and they all have applications specific that are growing we see those things they have some some they're not as good in many ways they don't process as well they may not have the long-term compatibility they may not have the long-term stability you know DH or DOTP has been around for for 40 years and it's only really in the last couple years it has come across as a phthalate replacement because people are looking at alternatives and this seems to be you know one of the high ones in the list along with dents and a few other ones you know if we look at the US plasticizer market and so forth this is taking a 2008 phthalate ester demand here's your major ones that you can see on the chart where the deal at dinp's and didp's and the dphp's those c9 and c10 phthalates take up about 50 percent of that market you'll see a little bit of DOTP also on this chart but this 2008 scale other phthalates the question was asked before what plasticizers is the exon produce well we produce a dinp and didp those are two main products but we also sell DTDP which is a C13 plasticizer an L11 L911 and a 911 products so and they all have specific end uses that they're tailored to when the phthalates themselves they represent about 85 84 to 85 percent of the US market but that other 15 to 16 percent being the variety of non phthalates applications they go to the pie chart at the lower level you probably can't read that from your seats but you know flooring wiring cable consumer products so those are all the big ones medical is smaller automotive you know coatings and in a variety of smaller end uses but the big ones are the wiring cable building construction consumer flooring those those are the big end uses for phthalate plasticizers this chart again if we just take kind of a quick summary of our just a real quick snapshot of the market that we would see not to meant to be completely accurate in every scale of possibility but you can see where the top ones at the top of the chart we have the solvents creams cosmetics and fragrances those are your very low molecular weight plasticizers or phthalate esters that the dimethyl diethyl dipropyl dibutyl but as you start getting more towards specific end uses you can see how the center part of the table has a lot of checks that's really where the dinp's didp's and dphp's are have potential use and as we go through these table looking at these you can find that if people look for cosmetics and creams are going to choose things like C1 to C4 phthalates if they wouldn't use some of the higher molecular weight ones that you would would find that we'd be talking about the dinp's and di dp's you move towards wiring cable classic point would be the building wire you have been inside your home and you put that you buy a new home you want to make sure that wire is safe and it lasts for 30 or 40 50 years because a replacement is cost and the safety risk is very high so people tend to make choices about plasticizer based on those products that will give them a long safe life and not have to worry about it so inside that wall you're going to find the insulation material may use trimelatates and linear phthalates as well as on the jacketing material would have di dp or dphp looking that you wouldn't find low molecular weight plasticizers or citrates or or anything else because those things just won't be seen as lasting long and other applications again the specifications of the performance of the wires always tend to dictate what kind of products a wide variety of products can be used and things but in general you're not going to find the low molecular weight products you may find some with dhp or dotp used there in addition to the other products you move on to construction products non building wire non flooring you know again the varieties can change you may find the acrylic caulks and sealants such as you know your painter's caulk that you would use in rooms they would tend to use more the low molecular weight plasticizers that the butyl benzyls or possibility in that where you get into a couple of times as mentioned vinyl siding using plasticizers actually that's a very very small in use most of your vinyl siding is is rigid PVC which may have some type of acrylic cap stock on it but there are some products out there that will have a plasticized PVC cap stock just that thin layer on the outside surface for the pigment is in the antioxidants so that they can provide you know a long lifetime history of good color stability and good stability in that product for automotive interiors one of the key sections on key criteria is the inside the car is ability to maintain low levels of fogging if you remember back in the 70s and so forth you would always tend to have on the inside the cars when the sun would hit it you'd find these oil films develop on the inside your winch windshield well those are coming by volatile components used inside the car and the automotive industry over the past really beginning in the in the early 1980s has been working to reduce and prove those emissions and they changed the plasticizer so while it used to be a DEHP was used to a lot inside a car it's almost not used there today they moved to heavy and like it way products as the IDP or dphp or even some of the heavier linears 11s and trimelatates and so forth again you wouldn't find the low volatility products being used there under body seating sealants you know a lot of the foreign cars some Japanese imports made in the US would have a undercoding applied underneath the chassis to protect it for long service life and again you would choose plasticizers based on those products which help improve that efficiency of application to keep it worth you've got to heat these cars up to the fusion temperature to seal the products so they basically coat the car throwing an oven heated up to a fusion temperature and so your higher molecular weight products just have a more difficult to process there if you look for vinyl flooring or resilient flooring or PVC carpeting industry is changing a little bit you're finding a tremendous growth of the DOTP products growing in these areas but also the Dibenzo weights as Mark Holt mentioned earlier butyl benzo phthalates used there issues they would like to have in choices or the stain resistance long color good stability non yellowing things they want to have a product that lasts for a long period of time for PVC back carpeting people but choose plasticizers that when they process it it doesn't damage the fibers of the carpet you know there's a temperature limit that the nylon would have and it says it higher temperature that they're going to make choices based on that so they would tend to use plasticizers like dinp the lower molecular weight of the series although they may get by with the IDP or DOTP and again in those applications you wouldn't find the C1 to C4 phthalates being used for medical products we're not in the medical business we don't we don't not a DHP producer so our view of this is probably less accurate than some of my colleagues and other companies may provide but again long history safe use can it resist changes in color change the sterilization low-color high purity those are all the things people look for for plasticizers they want products that they know are safe and be used for a long time if you look for shoes or for clothing products shoes inks and handbags they all have different choices people would use your you could find dinp and didp dphp DOP DOTP all used in those applications at various levels for products in the last slide here just a series of miscellaneous applications from furniture to garden hose to tablecloth to shower curtains floor mats and even toys and you're going to just to find a variety of plasticizers used there but again for most of those the exception of toys you would find dinp and didp and dphp and DOTP having potential use in all those areas but you wouldn't find the C1 to C6 phthalates unlikely to find citrates unlikely to find the heavier phthalates in those areas for toys you know the world is changing there there was a 2008 study published at a analytical journal and in Switzerland where they looked at toys I think they analyzed over 300 parts of 270 some toys and and again dents as was mentioned earlier before represented about 50% of those products but you also find on that list had about another 20 types of plasticizers there being used in addition to citrates and DOTP a whole host of some you know mesomol products as well as some polyolesters from algae chemical and and variety things epoxodisobionol and so forth so it's a wide variety of things that are being used so that was my quick overview and happy to answer any questions thank you questions and you will be providing us copies of your presentation and one of the issues that we have to deal with is what are the sources of phthalates and we're happy to provide more time more information on that but questions down we'll spend a lot of time working that for you great thank you so thank you for that overview so if I'm sitting in my car and I've got some coming in heating things up is it inhalation exposure that I'd be worried about not from phthalates most of the studies done now where they come back and look inside cars and so forth they're finding it's it's coming from the the fabrics themselves as maybe polyurethane chemistries and so forth so you know it's but your discussion then you went into carpets and things in relation exposure then in all of those products that you're thinking about for human exposure or is it I mean I recognize lotions I'm gonna worry about dermal what it what kind of exposures would you know I'm not in the expert to get into that but you're I'm certainly it's probably it's not food but it's either gonna be inhalation or skin contact and automobile itself for most of your your vinyl products that you have your body's not resting on so it's you know in a leather car a leather seat in your automobile it's it's the the sides it's the back you know it's underneath which is which is PVC your arm rest where you sit in your back or see it as all leather so and those mechanisms I would say yes the only potential route you would have would be inhalation that would be but again the automobile industry has looked at that through setting their fogging specs to to really minimize that exposure potential and and there's certainly tests for people show that when they do this that they don't they don't see that being a problem at this point it's now other types of chemicals they're there carpeting in a in a building where you got vinyl back carpeting yes that the vinyl is is on the floor there's no way that you could come and contact that with you know I guess unless you wanted to crawl underneath the carpet but the only way you could get that exposure would be inhalation but again these think products at very very low vapor pressures and you know and so the exposure risk is a non toxicologist expert would be would be very minimal follow-up on that wouldn't I mean you're just talking about the carpet in the vinyl backing I mean wouldn't degrade with time and form small particles and dust and adhere to a way in which children you know very young children are exposed on the floor and I don't know what the mechanism of that is be honest I've heard you know those kind of questions coming before and we've exchanged ideas when the company but you know how it how it falls apart and so forth and breaks up I really don't know as you would expect though that one the level to be very very long in time intervals I mean you know PVC flooring can last in a home 25 years and people replace it not because it's you know brittle and falling apart but usually because the housewife wants a new look you know so it's still still most of the plasticizers in there but certainly if you did have extreme conditions where heat was hitting something and and you had a higher vapor pressure than you would like to you know seen with the plasticizer choice you could get some embrittlement and friction on top of that could create some some small particles again I'm just trying to think of where all the stuff is coming so people have been talking about shower curtains hot water on a shower curtain is that gonna what's the problem is it that is it an inhalation problem potentially there is it a what again the vapor most of your shower curtains that are made you're going to be with you know DHP or DINP or DIDP and the vapor pressure on those molecules is fairly low I would think that if anything where you're seeing plasticizer loss you know could come from just repeated water hitting it which would go down the drain I think there's always confusion that people would say the new vinyl odor is plasticizer you know or if you get by that shower curtain and you open that bag up you know there's an odor with that well that's not the plasticizer that some of the stabilizers that are added into the system so but as far as you know anybody measuring the vapor pressure released you know of a plasticizer in a hot shower you know I've never seen any studies on it. A lot of my questions come from I'm thinking about you know studying in Haynes data and where it's clear in biomonitoring studies that humans are exposed to phthalates so and then we have this list of all these products that these phthalates are coming from what in your best judgment where do you think the problems are? It's hard for me to understand you know well phthalate you know if you look around a room like here you know you have these rubber mats here you have this strip across the floor as PVC some of the wiring may be plasticized PVC and it in a television set probably nothing is there they they worry about migrations there so I mean it's a round a lot this floor mat may have the back end the backside of that could in fact be plasticized PVC so you know it's just my point is that biomonitoring studies show that that we have these chemicals you know I heard people say that food was the number one you know contribution from it but food contact for things you know and it's not my my area of expertise I mean potentially how how far down the supply chain do you know or have information on you know obviously you don't make carpets and cable and all these other things but how much can information do you have because that's the kind of thing we're going to need I think from most of these end uses we can probably come up with pretty good I mean if we're selling to people making carpets and we have a pretty good idea what they're using how they're using what types of products they go into the one big gap is in the industry is that it's a compounding industry where we would sell plasticizers to a company who would take and convert those to pellets and those pellets would go to a wide variety of systems and in uses that we would you know not be familiar with and so you can get pretty close you can probably cover you know probably 80% of the market between companies like Exxon and Eastman and Faro and BSF working where those products go but you know getting a hundred percent maybe everything may be really tough and I want to head right on to panel for a good afternoon thank you very much for rearranging the schedule for my convenience I appreciate that my name is Jane Tata I have over 30 years experience as an epidemiologist designing conducting reviewing studies I had my training at Yale University where I have my masters in biostatistics and my doctorate in epidemiology and if you look at this first slide and you see DRPH that is not a phthalite that is Dr. Public Health her a lot of D's and P's and H's today and I'm here at the request of the American Chemistry Council phthalate esters panel I want to recognize the contribution of my colleagues that exponent on the fuller report that I hope you've all had a chance to look at and that is Dr. Jean Manson who's a reproductive toxicologist and a clinical epidemiologist and Megan Wagner another epidemiologist and their colleagues of mine at Exponent which is a consulting company based in California over the past decade an increasing number of epidemiology studies have appeared on the scientific literature investigating in utero biomarkers of phthalate exposures and subsequent birth outcomes and developmental effects in children and infants several studies have investigated associations with gestational age while one study each considered an association with cryptorchidism or undecided testicles and antigenetal distance a number of studies have focused on neurodevelopmental effects such as reduced cognitive function and adverse behavioral outcomes using biomarker phthalate metabolite exposure measurements at older ages the neuro studies a few other studies have explored possible hormonal disruption using phthalate metabolite measurements in maternal breast milk or from children's urine samples in general these studies have a number of common characteristics they all start out with a proposed hypothesis based on animal results they spend a fair amount of time enumerating most of them anyway that the general limitations of their work many of which they have in common and I will discuss those and then they talk about possible mechanisms for any of the positive or statistically significant associations that were observed and routinely they call for more research doesn't surprise you virtually all of the studies report one or more statistically significant associations causal interpretation of these findings is problem problematic however due to several limiting factors often well described by the investigators I will discuss the strength of the epidemiologic evidence in light of the challenges that researchers have faced that include these exposure misclassification lack of consistency in findings lack of replication residual confounding weak associations and multiple comparisons let's start with exposure misclassification most authors raise the issue of whether their study captured the appropriate window of exposure given the short half lives less than 24 hours of phthalate biomarkers DEHP metabolites were shown not to be particularly stable or highly reproducible I should say during the last six weeks of the third trimester by a DB at all other studies however have suggested relative stability of phthalate biomarkers for weeks or months that is among men and non-pregnant women the metabolism of pregnant women changes rapidly and they have a 30% increase in circulating blood volume of particular concern our cross-sectional neuro behavioral studies that attempt to link spot urine of elementary school children and link it to parental or teacher neuro behavioral performance ratings these are typically cross-sectional studies so they're taking the urine sample and the phthalate level at the same time they're doing the testing and their older children wolf at all appropriately call for a more comprehensive integrated exposure assessment prenatally and before puberty let's talk about consistency where there are multiple studies of some endpoint the results have not demonstrated reasonable consistency this particularly is the case for studies that have examined length of gestation at gestational age of delivery and there are a number of them why it at all report a statistically significant inverse relationship of DEHP metabolites and gestational age of delivery this finding is consistent with that of Latina Lantini at all who measured MEHP in cord blood it is inconsistent however with wolf at all who reported statistically significant positive associations of low molecular weight phthalate metabolites and MEHP with gestational age a DB at all also reported increasing gestational age with increasing DEHP metabolites it is noteworthy that although the results were divergent both Wyatt at all and wolf at all their studies were conducted among African-American and Hispanic New York City mothers so they had different results but they were using very similar populations with three hundred and eleven mothers in the former study and four hundred and four in the latter metabolite concentrations were also similar in these two studies me curate all reported a positive odds ratio borderline statistical significance for preterm birth for the metabolite MEHP comparing cases and controls all of whom had greater than median levels of exposure while the DB at all reported an inverse association for preterm birth and DEHP metabolites in a nested case control study of cryptorchidism main at all reported no differences in the metabolites measured in mothers milk or differences in gestational age between boys with undescended testicles and controls still sticking with the consistency issue angle at all did two neuro behavioral studies one in newborns and one when these children reach four to nine years of age both reported sex specific effects newborn girls showed a significant linear decline in orientation and quality of alertness with increasing concentrations of high molecular weight phthalate metabolites boys showed a different pattern with some indication of better motor performance with higher concentrations of low molecular weight phthalates metabolites in the study of older children however boys exhibited poor scores with increasing concentrations of low molecular weight phthalate exposures lack of replication first time reports of associations particularly studies employing novel methodology require replication before a causal association can be considered seriously for example there have been no other studies of anti-genital distance and exposure to phthalates in humans other than the one by swan at all and her colleagues and a subsequent study by swan alone studies that examined similar outcomes are not true replications if the source of exposure is substantially different such as studies were exposures were sampled from urine versus cord blood or breast milk this is a little more of an epidemiologic concept but I think it's very important in these studies and it's called residual confounding I think most folks know what confounding is it's another explanation for a result in order to have confounding the factor has to be a cause of the outcome but it also has to be associated with exposure those two things to be a confounder when we say residual confounder we mean whatever the measure was for the confounding may have not captured at all so if we use for example education to reflect socio-economic status for example we may capture some of that confounding but not all of it because we don't have the true confounder measured with many studies when when folks talk about error they say well the results could have been even stronger but there was misclassification or differential or non-differential misclassification when you're talking about confounding when you have errors in confounding it can result in either under or over estimation of risk it could go either way when you don't have the confounder right errors in measurement or inadequate surrogates for confounders can result in residual confounding and bias risk estimates in either direction the existing studies have collected information on numerous variables we heard about that this morning that may be confounders most from parent questionnaires there's a long list of covariates that are known or suspect risk factors for developmental effects in children for example age sex ethnicity race mother's pre-pregnancy height her weight body mass index smoking history education IQ marital status and history of asthma hypertension and diabetes to confound an association of risk factor also would have to be associated with exposure to phthalates main at all note the possibility of unknown factors related to phthalate exposure while knowledge is incomplete of the distribution of phthalate exposures a few important characteristics have emerged minorities those of lower socioeconomic status and obese persons typically have higher levels of phthalate metabolites in their bodies many of the study authors tested for confounding with these variables by examining the effect with and without the factor in the statistical model which is standard technique the effectiveness of this approach relies on the accuracy of parental recall though you're asking women what they weighed and you have to trust that you're getting the right answer for example wolf it all reported a positive but small statistically significant correlation between low molecular weight phthalate metabolites and body mass index but raise the issue of crude estimates of maternal anthropometric features based on maternal self reports a key problem it's statistical but I think it's very important is that most of the researchers had some cut off for putting a confounder in the model because you don't want if you put too many things in the model it's not parsimonious you have to estimate too many variables and you lose power so people try to cut it down to the important one so they say well let's let's only put it in if it creates a change of 10% or more problem you have is then you knock some off because they didn't come to the 10% so you don't put them in the model now you get a positive finding of 4%, 5%, 6% could have easily have been the confounder but you didn't put in the model for good reason so it still leaves that issue outstanding it has been estimated that 90% and this is an important statistics of DEHP intake except for infants is from food 90% this is according to Kavlock in a paper in EHP in 2006 90% from food the most influential food products are fats oils and dairy products so the real question now is does body mass index capture that if that's a confounder it's related to exposure we expect people who eat heavy fats oils have more adverse reproductive effects or developmental effects in their children so have we really captured that if these food products are independent predictors of risk for developmental effects or hormonal alterations they could also be confounders in an association between DEHP and developmental effects while most studies that have examined this association control for body mass index or body weight there is uncertainty as to whether these variables adequately control for potential dietary confounding no study has attempted to control specifically for dietary intake weak associations when epidemiology studies have weak associations that is the magnitude of the increases small or the dose response is borderline significant those kind of weak associations then any kind of error in your study whether it be confounding or poor information could easily have been an explanation for what you see we always feel more comfortable with a big risk because then we think well it's less likely than any mistakes in the conduct of the study could have explained it though many of the associations observed in these studies were not statistically significant and therefore assuming no error chance findings or they're insensitive due to small sample sizes when when they're not statistically significant but they show an association where you do get statistical significance they are generally of low magnitude which could easily be due to study limitations another challenge the interpretation with weak associations is whether they are clinically relevant the endpoints examined in the developmental studies are of unknown significance either because of subtle changes observed or because the relationship to a clinical diagnosis is absent seven years ago latinia at all discussed the need for a clearer understanding of the clinical relevance of their measured associations between phthalate metabolites and shorter gestational age in 2008 wolf it all noted the small effect sizes related to longer length of gestation db at all who also reported longer length of gestation associated with higher levels of phthalate metabolites commented and I quote the clinical pop the clinical or population significance of two to three days in gestational length is difficult to evaluate in the wolf it all studied the weak associations between low molecular weight phthalate biomarkers with the timing of puberty in girls was in contrast to the inverse relationship with high molecular weight phthalate biomarkers none of the adjusted prevalence ratios was statistically significant and there were no notable differences from the null angle at all reported poor parent rated behavior and executive functioning in boys not girls associated with low molecular weight phthalates not high although few children and this is what's important for the clinical significance few children met the standard at risk or clinic clinically significant criteria main at all on the other hand did a study where they actually took a clinical end point crept or criticism and reported no association with any of the six phthalate metabolites measured in mother's milk finally multiple comparisons since the studies are exploratory in nature many of them have analyzed multiple phthalate metabolites and multiple outcomes given a large number of comparisons some associations will be statistically significant just due to chance meeker at all referred to the quote large number of statistical comparisons wolf at all noted that their statistically significant associations may be due to multiple comparisons multiple comparisons are particularly problematic for studies of neurodevelopmental effects angle at all for example included behavioral rating inventories of eighty six and a hundred and thirty item questionnaires that were examined separately for boys and girls so you have eighty six and plus a hundred and thirty and then they did them for boys didn't for girls didn't for high metabolites didn't for low molecular weight metabolites resulting in over eight hundred comparisons wanted all include over five hundred comparisons while examining reduced masculine play in boys biomarkers studies of development developmental effects pose numerous challenges the criteria for causation such as consistency biologic plausibility magnitude of risk and even temporality as several these were cross sectional in design are far from satisfied across the whole gamut of the studies of phthalite metabolite exposures and birth outcomes and developmental effects a few studies suggesting a dose response are uncertain due to questions of residual confounding and exposure misclassification because of these challenges authors are appropriately cautious in their interpretation of the observed weak and inconsistent associations thank you any questions thank you questions I guess quite quite a few but just I think what you're pointing to and in terms of epidemiologic studies is is correct I mean not just the epidemiology of phthalates and reproductive or developmental health but air pollution studies lead etc. I mean all of these limitations that you point out are possibilities in epidemiologic studies so I wouldn't disagree I mean this is inherent in the field it's not specific to this topic area well they are known epidemiologic limitations that one has to consider in drawing conclusions about causation but environmental epidemiology is particularly subject to difficulties I have done occupational epidemiology studies for 30 years and do not suffer our studies don't suffer from as many of these limitations this is a particularly difficult when you're doing biomarkers a single sample that's different from other studies it's more challenging well possibly but in terms of these limitations when you you just mentioned a biomarker a single urine sample most of the time or all the time you really need to think about the magnitude of the bias and the potential direction of the bias and I agree with you that with confounding it can be in either direction with mis-measurement of exposure if you only have a single urine sample during pregnancy or at a single time and point in time that will most likely lead to what we call measurement error and misclassification and that's where you mentioned you know the traditional bias towards the null if it's a binary exposure variable once you get more than two categories three or more categories in your exposure then that general rule of thumb fails on occasion and it's been published about that in other words if you have more than one that's true I mean it can fail on occasion it's rare but I mean that gets too much finer point but I guess a bigger point one of the things I wanted to caution though is when you're talking about clinical relevance is to be cautious in terms of clinical relevance to an individual versus a population because a small change you presented on some of the gestational age studies I mean I would have thought maybe some of the hormonal studies or others but a small change in a continuous measure may not be adverse for an individual and the best example which data that was collected you know several decades ago looking at lead and IQ the lead literature showed about a two point change in IQ and for an individual that probably or would not be clinically relevant I mean you would probably have the same job that you're doing now you'd probably be you know sitting in this room or not regardless of whether your IQ was two points above above what it is now or lower but clearly on a population level if you shift the population central tendency down to IQ points you can affect the tail of the distribution I'd like to mention however this was not original on my part that the authors themselves raised this and what I found I'm not a biologist as you know my background is from statistics into epidemiology not biology but what I found extremely odd is that the studies that had a longer gestation one or two days would have a neat little biologic mechanism to explain it and the ones that had shorter also had another explanation for that and they both treated it as if it was a bad thing I mean one or two days either way I couldn't even figure out myself what's a good result and what isn't no one's going to be right on gestation age especially with the difficulty of even measuring it and many of these studies just use mothers last menstrual period and not many of them use clinic so that's what I mean how relevant if for sure you knew for sure you were extending gestation by one or two days and that was truth maybe that means for population that might mean something I wasn't so much referring to the gestational age but just a careful distinguishing between what is a clinically relevant effect for an individual which is more in the clinical medical realm versus a population level I would tie it more closely to things like shifts and in a general distance in populations which may be indicative of let's say hormonal changes or changes in serum hormone levels in individuals rather than you know I know that gestational age literature is murky I was actually an author on the Wyatt and the adibi paper which you know found opposing results but I think for some of the effects that the panels interested in in terms of in a general distance hormonal changes that be cautious in terms of what's clinically relevant for an individual and on a population level. But when you say hormonal changes that's one thing I think we heard this morning the association with adverse effects for an endocrine disruptor as opposed to modulating someone's hormonal levels again I'm not a biologist but I know a lot of things modulate our hormonal levels so it's got to tie itself somehow to a clinical entity that's why I pointed out the one study that actually did look at crypt orchidism and didn't see anything but you may have seen some changes in hormones so just pointing out that's just another thing to be cautious in interpretation but yes. You've highlighted multiple comparisons you highlight the problems with multiple comparisons which I think that point is well appreciated is your specific point with the study by Engel et al and Swan et al both 2010 that they didn't make adjustments for multiple comparisons or what is your point in my training at Yale we were made very well aware of multiple comparisons but we were not trained to make adjustments for it I think things can get very conservative if you start doing Bonferroni adjustments what you do is you take it into account in your interpretation it makes the case for replication or consistency so much stronger but you can make adjustment for multiple comparison Bonferroni correction is one there are many other methods as you feel well know what is your point specifically that Swan et al and Engel et al did not make these adjustments? No that's not my criticism my point is when you evaluate these studies and you go to interpret them in terms of do you think this is really cause and effect you have to consider how many comparisons they did how many significant findings they had and did anyone else see the same thing was it an a priori hypothesis in other words if an endpoint is stated up front this is what I'm looking for this particular endpoint not just neuro behavioral effects but you know more aggression or more and they set it up front and they got it be less likely to say it's multiple comparisons but when you just start let's look at all neuro cognitive neuro behavioral we got 150 of these and you get something you have to say well let's look there's got to be more studies to look at this I can't make much of it is your point and that multiple comparisons would not be made? No there are ways to make it less of a problem which are can you give an example I'll give you an example although some of the grouping trying to put the phthalates into categories now you can sit around and debate what categories a lot of people say the low and the high was not a great one because some of the animal studies some of the low showed a problem and some of them didn't and you put them both together may not be the best thing to do but suppose you group those that show effects and rodents and make that one group so you're not looking at eight different phthalates and then set up a front your outcome I'm going to look at length of gestation for example now that doesn't mean later on somebody does a study and they see another outcome that you didn't focus on you may go back to your data and look at it but try to in your protocol design it to reduce the number of comparisons that you make well yes your points are well taken multiple comparisons is a problem but through adjustment one can still distill out as you full well know statistical significance this is possible which leads me on to my next question you made a lot of points very well taken and well acknowledged in epidemiology especially environmental epidemiology and we all agree with you I think that this is a very difficult and challenging area however the field has developed in the last 20 30 years and there are quite sophisticated methods available to distill out and correct for possible buyers confounding etc etc can you highlight any of the studies which you quote that in your opinion fall behind the scientific state of the art in that area apart from the general problems every study of this kind grapples with well I'd certainly consider any of the cross sectional studies weaker and put them in a whole different category weaker I would agree with but would they fall below the scientific state of the art you know they get published cross sectional studies get published they have to be considered we consider all the data but I wouldn't give them the weight I would give to the studies that have a longitudinal pattern to them I think everyone in this room would agree with you there but can you highlight a study then you made many criticisms and implied criticism can you highlight any study or a series of studies which you quoted which in your opinion are beyond the payer I think they're all very similar I tried to make that point in the beginning when I said they have the common characteristics other than the cross sectional they all seem very similar they all took one spot year in random they all did that they focus pretty much on high and low phthalates filled with confounders all of them I saw a lot of similarities in the limitations that you don't usually see across the board in one area so I'm not able to say this one is you know better than all the others they all struggle and I'm not it's a difficult area I again agree with you there but the points you describe are not really specific to an ethylite epidemiology what everyone struggles in environmental epidemiology with confounding with misclassification with multiple comparisons etc etc and there are ways of dealing with it and that's the scientific state of the art so would you say that the phthalate epidemiological studies as a group fall behind what other researchers do in this area of environmental epidemiology I think many of these limitations are characteristic of this particular area of research and I think it's the one time measurement which may not be a limitation some other kind of environmental study but I think in general this is characteristic of environmental studies but I do know there are new methods coming out new approaches to studying phthalates at the right period of time I'm not that's not my area but I've talked to Dr. Manson about it and there are new techniques to improve I do think something has to be done about diet they have to really get at and that's not characteristic of every other environmental study a lot of them have socioeconomic status as an issue and this may too but I don't know that we're really getting the residual confounding I haven't seen that as a huge problem a lot of other studies but I think here it is because we don't know enough about the characteristics of exposure to phthalates in people you know why is it higher in African-American women why is it the weight thing is it the diet fatty I mean that's different I agree with you this is that is the case in the area of phthalate epidemiology but it is also the case in environmental epidemiology concerning other pollutants so I'm sorry to press you a little on this point it is a very important one for our work as you will understand and we will have to rely on your expert judgment there so again the question is or is not the phthalate epidemiology in your opinion below scientific standard in other areas of environmental epidemiology or not no I think all environmental epidemiology has more difficulties than say occupation or cardiovascular in general environmental epi is more difficult doesn't mean because it's on a standard a par with all the others doesn't mean you can jump ahead to draw conclusions just because this is the best there is if it's not good enough to draw a causal inference thank you very much so I am familiar with one of the swan papers and I don't remember if it's one of the three you listed here but where she actually accounted for the fact that these were chemicals together so she accounted for the mixture effect in some way do you are you aware of other studies that have done some approach of instead of just having tables of ten different phthalates but to try to account for the fact that many of them like I said did try to group them into low and high molecular weight and they had like a global score she did that and some of the others not my area but Dr. Manson disagreed with that strongly because she felt that you shouldn't group them by whether they're lower high molecular weight but you should group them differently by which ones are most plausibly related because of animal studies but no I don't think anyone did that you're saying a high molecular weight so it would be out of the class of high molecular weight some way of finding a summary score of explosion and the problem you have is if one or two of them is not related and the others are you're going to have dilution and you won't see something that might be there so she suggested doing it based on animal data putting them together that way I would agree with that and I think that in the swan paper I mean they defined the low and high molecular weight phthalates differently than I think in just recent presentations where you know the she was using combining diethyl phthalate or mono ethyl phthalate with mono butyl phthalate and I think as you pointed out which I agree with is if you combine the two together whereas in toxicological studies the ethyl is not biologically active but the butyl is you'll be less likely to find an association in the human study and then I also agree with you in terms of a single spot measure as being not giving you an exact precise estimate of exposure over time because these phthalates have short half lives but I think you would agree that if that's one of the main limitations of the studies it would make it more difficult on average or in general to find an association right so if you did a pregnancy study and you had ten urine samples over different time points and you could combine them in some way to get what the woman's exposure was during different critical windows or over pregnancy you'd probably be more likely to find an association if one did exist than if you only use one single measure I mean would you agree with that? In order to throw it off it would have to be like a confounder so what you're talking about misclassification so is it differential or non-differential misclassification is what you're saying? Yes. I think you'd have to try it out and see if it's differential with respect to the outcome I mean I don't know enough of the biology of all this to think it would go like that but that's the way it would have to go but in general if two measures is better than one and three better than two you would think that only having one single urine sample would make it more difficult because of your misclassifying exposure I would think you may be right I'm not sure I think you'd have to do a validation and say what would the results be if we had exposure assessment would we have more power maybe if there's something there to detect it that may be and a better measure of exposure but also with small samples you can get errors of increase to just not enough data so I'm not sure I think that's a good area for research what difference would it make if we had a full exposure assessment versus a single spot I'm not going to think about it some more I don't think you may be right thank you and Chris are you going see if I can do two things at once in the interest of saving some time here I'm going to try to navigate this and get out of this presentation I'm not sure where it is there we go didn't do so well at multitasking but I've got it up now my name is Chris Borgert I'm with a small consulting firm in Gainesville Florida called Applied Pharmacology and Toxicology Inc I've been working with the phthalate estrus panel of the American Chemistry Council on cumulative risk approaches for anti-androgens and phthalates I have to make the disclaimer that while they have heard my views I'm not sure that they agree or disagree with them but they are my views and we will publish them as they are effective agreement or disagreement from the funder today I want to maybe change pace a little bit at least with this presentation and talk about things from a much more theoretical standpoint I've titled my talk uncertainty versus certainty and cumulative risk assessment of anti-androgens and so I'm going to brush through some generalities rather quickly and you can take exception with them but I'm just trying to orient my comments here so to summarize my points those are the conclusions that I'm driving toward the dose additive model which is really the basis for a hazard index approach dose additive cumulative effects have not been demonstrated except at doses near the observable effect range now that is admittedly a limitation of toxicology studies it's difficult to show effects at miniscule doses below the observable effect range but nonetheless we have the data that we have and we don't have other data I don't think that dose addition is a generalized phenomenon and by generalized I mean extending from the first point I don't think we have the basis for extrapolating dose addition to much lower levels we simply don't have the data another point I want to make that may not seem absolutely related but I hope to develop it is that it's not possible to be protective without accuracy and I think that there is a way to apply a cumulative approach rationally and in a protective fashion with an adjustment and the adjustment would be what I call a human relevant potency threshold so I'll try to explain those things during my talk so the background Dr. Gray and Dr. Kortenkamp's group have independently reported cumulative effects of anti-androgenic chemicals which would include drugs, pesticides, phthalate esters on male reproductive tract malformations, testosterone levels and other secondary sex characteristics in the rat and that based on these results I know that there is discussion of a cumulative risk approach that uses a hazard index calculation for anti-androgenic chemicals that affect the male reproductive tract and I would note I think that the approach is based on a precautionary preference for the concentration addition or dose addition model of combined action so that sets the general framework but I think that we have to admit and I want to be absolutely clear these are not criticisms of the studies in any way they're simply looking at the realities of conducting these kinds of studies in whole animals and it leads to some uncertainties and when I talk about uncertainties I'm not talking about the results obtained by the authors or the conclusions made by the authors I'm interested in this business of extrapolation because I'll get to another slide or two I think that's really what we're talking about here with the cumulative approaches is extrapolating from the observed region to regions where we have a much more difficult time doing studies so both Gray and Corten camp evaluated a single mixture ratio of the components we know that the type of cumulative effect that's observed can change with different dosing ratios again not a criticism of the study simply a reality of trying to do these very labor intensive types of experiments in whole animals is very challenging to do say a full factorial design both Gray and Corten camp evaluated scored endpoints sometimes that makes analysis of the variance a little difficult and again I'm talking about the variance that we're talking that would apply to extrapolating the effects that's what's important here I'm not going to dwell on these because I'm not sure that these are the most important issue I want to get to some larger issues there was some discrepancy in the types of combined action that Gray found versus the Corten camp group Gray reported that the response addition model did not predict any mixture effects whereas Corten camp generally found that both concentration addition and response addition predicted the mixture effects but they also found some evidence for synergism in some instances scored endpoints were combined I'm interested in the variance not only for the study at hand but that's less important than my interest in variance for extrapolating conclusions or results that conform with one model or another to very different regions of the dose response curve so I'm working with the statistician George Cassella another toxicologist Bob Golden to understand uncertainties and certainties and that's what I want to get to a little bit later is the certainties and exploring how we might modify a cumulative risk approach with a human relevant potency threshold so just a background slide to orient everyone to what we're talking about this is a fairly simplistic graphic representation of the difference between these two models and I'm stressed that these are simply mathematical models of how one might predict the effects that a mixture would produce they are not intricately tied to an understanding of biology or pharmacokinetic pharmacodynamic models they're simply two ways of adding let's consider a hypothetical dose response curve for any toxic effect and doses of chemical A chemical B and chemical C present in a mixture at about half their threshold concentration the dose addition model would predict the response of the doses adjusted for their relative potency added together whereas the response additive model would add simply the responses for each chemical at each dose and so if these chemicals are present at doses below their threshold they would give zero effect and response addition would predict zero effect dose addition would actually predict a super threshold effect if these are present at half their threshold concentrations because 0.5 plus 0.5 plus 0.5 would give you a measurable effect so this is a bit of the groundwork for these various mixture models that we consider but I want to make the point that again irrespective of whether we're talking about a decreasing dose response curve or an increasing dose response curve the critical issue is what happens down here in the very low dose region and distinguishing an observable response from one that's not observable and that's where our point of departure really is and where I think these issues of certainty and uncertainty come into play because we are extrapolating then these models not across a small dose range but across orders of magnitude many times in some precautionary assumptions that we've long made in cumulative risk assessment now I want to switch to a different point here for just a minute and ask with that background in mind and our discussion of extrapolating these additive models whether we're really making progress overall in our research and regulatory methods for assessing cumulative risk it's often been criticized that our risk assessment models do not account for mixtures but in fact for decades a hazard index approach has been used at least in superfund style risk assessments etc and that hazard index approach is based on the target organ of toxicity so that single chemicals are assessed for their no observable adverse effect in an animal study that no observable adverse effect level is then used to derive a reference dose which usually means reducing that dose by 100, 1000 or more to be protected for human health and then the hazard index approach say at a hazardous waste site would compare the expected exposures with the reference dose add them all up for all the chemicals at a site and determine whether that the entire hazard quotient exceeded one and if it does and one goes back and breaks it out on the basis of the target organ well from that though there is an additional so anyway first of all we have assessed mixtures for quite a while we've done it in this somewhat general and more crude way but there's an interest in being able again to be more predictive of what might occur at lower doses and what might occur across species and so from that interest was generated a desire to look at the mechanism by which chemicals worked so that perhaps we could better add a larger group of chemicals based on what we know about their mechanism not just about what target organ they might in a whole animal study and so we came up with the concept that similar modes of action in other words chemicals that have what are determined to be certain mechanistic features in common would be assessed by a dose additive approach whereas dissimilarly acting chemicals would actually and I've got dose addition here I'll get to that in a minute but dissimilarly active chemicals were assumed to behave by response addition and that led to relative potency approaches such as the toxic equivalency factor approach for dioxins and dibenzofurans when that was eventually tested in the last few years the dose addition approach works for some cancers doesn't work for others for some cancers even found synergy and now there's a new research on the differences between AH receptors the aryl hydrocarbon receptor which is the basis for the TEQ-TEF approach of dioxins that there are differences between the rat and human receptors so that some of those mechanistic inferences may be in question now that was also this type of approach was also applied to organophosphorous pesticides and the EPA had to do quite a bit of work to demonstrate actual dose addition there some then criticized that this was getting too restrictive and too difficult to demonstrate mechanistic similarity and so now we are where we are today with the phthalite esters such that dissimilarly acting chemicals may also be dose additive and we see some demonstrations in the literature from Dr. Courtenkamp and Dr. Gray that chemicals with dissimilar modes of action can also be dose additive but then I ask you where is our ability to predict because if we are saying mechanism doesn't matter but only the target organ system we've sort of taken out of the equation our basis of on mechanism which is what we hoped would allow us to extrapolate to lower doses and to other species so it's a conundrum that's difficult to solve now why is that important well we one of the things we know and now I'm moving to the level of certainty we know that single chemicals at least are dose dependent in their mechanisms and this was a publication by Slicker et al it was actually an ILC workshop that identified all of the various bases that we know about for dose dependent transitions in mechanisms I simply ask you then if we know that the mechanism by which single chemicals operate changes through the dose response range would not also the interactions or non interactions or types of additivity also be expected to change across different doses some of this has been investigated for instance a study by Charles et al published in toxicology and applied pharmacology ask a cumulative risk question with estrogens they ask at what dose would synthetic chemical estrogens or putative estrogenic active chemicals be additive in combination with phytoestrogens so they gave active doses of the phytoestrogen and determined the dose of the synthetic chemical mixture that was necessary to produce an estrogenic response greater than the phytoestrogen alone essentially the take home point was those synthetic chemical mixtures increase the estrogenic response over the phytoestrogen background only when each chemical in the mixture was present at about greater than half its individual no L in the estrogenic assay so this is one reason that I think that I'm driving toward this human relevant threshold potency and I think probably the most compelling example for hormonally active agents is when we look at that example rather well known example of diethylstilbestrol diethylstilbestrol was widely prescribed to four to five million pregnant women until 1972 in the mistaken belief that it would prevent miscarriage I say the mistaken belief because there was no properly conducted clinical trial that ever ascertained the efficacy and ascertained the effective doses as a result we had large numbers of males and females exposed in utero to widely different dosing protocols of DES as we all know it was discontinued in 1972 on the discovery of vaginal adenocarcinoma in the female offspring of the mothers who were administered DES and male reproductive tract malformations in some of the male offspring so there were hundreds of clinical studies on DES exposed men and women continuing forward to today and what we find is in those widely differing dosing protocols we find the evidence for a true threshold in the lower dosing protocols we actually see no effects in the DES males whereas those effects are clear in the higher dose protocols and the same can be said for the vaginal adenocarcinomas so there's clear indication of a threshold here from human data with a potent estrogen but I would make the point that those women in the 50's and 60's were also exposed to phytoestrogens in their diets to bisphenol A to a host of other putative environmental estrogens and so if all of these things really do add together regardless of their potency and regardless of the amount we couldn't possibly see this kind of an effect so what I'm getting at is that we have evidence for cumulative effects and that could be justified for high potency chemicals or at doses near the observable effect range so I'm not taking exception with the published data but I'm simply pointing out that our ability to extrapolate to lower doses in my view is unsupportable I think it's at least questionable and somewhat uncertain and here's where I want to point out that I think we need to be careful and maybe we need to modulate this type of cumulative approach that we take because overstating theoretical risks can cause misprioritization and can mask real risks now there are a number of examples of this one of the most famous was the water disinfection problem in South America when chlorine was removed as a disinfectant and cholera cases resulted you can think of numerous clinical examples protective doses of postmenopausal estrogens given to women turned out not to be so protective but that was the assumption that we can think of a day when dietary fats were all considered to be bad and now many of us are taking fish oil which is a fat so it wouldn't have been protective and probably wasn't protective to try to eliminate all dietary fat so we've got to have some accuracy in our assessments in order to truly be protective and precautionary so we're proposing that a human relevant potency threshold can be used to differentiate the point at which the cumulative approach starts to be unsupportable and one would limit the application of the dose addition model and we believe that this is pretty well demonstrated for DES and I think that the same kind of an approach can be derived from data on human pharmaceuticals we're trying to forge some bridges with folks who might be able to help us with that data from the pharmaceutical industry so with that I'll conclude and ask you if there are questions. Thank you. Questions. Thank you very much for these interesting thoughts. Again I cannot disagree with a lot of the things you have said. Just one point for clarification, you seem to suggest that application of dose addition will always lead to the prediction that we will have effects at sub-threshold levels of the individual component. I think we can very easily agree that this is not the case, it simply depends on the number of chemicals you combine and really on their level so only if the sum of the toxic units reaches a certain level is there an effect to be expected so if there are any fears that blanket application of dose addition will always lead to prediction of effects I can allay these by saying that's not the case. What do you agree with that? Well I would agree with that obviously but I don't want to lose my point is that if dose addition applies to all potencies and to all levels then we could not possibly see that clinical threshold and this was not a few hundred or a few thousand women in their offspring. It's where millions of women who were treated with DES think about it they were exposed to many of the environmental estrogens and dietary estrogens that were exposed to today so if dose addition is applied without any cut off then we can't possibly have the result where we have a DES threshold because all of those estrogens would also be adding to the response and so that's where I'm proposing the human relevant potency threshold. There's got to be a point where we decide that a chemical is simply not potent enough to add to the effect or is not present at concentrations near a no observed effect level and so that it would protect us from being almost absurd in our prediction that everything will add together because if you think about it and I published a paper on this in 2003 where we went through all the data on relative estrogenic potency and as we initiate the endocrine screening program we're going to identify hundreds, maybe thousands more potential estrogens and the same may be true for compounds that can interrupt steroidogenesis, compounds that could have many of these anti-androgenic effects and that kind of butts up with then the idea that they're all adding together so I certainly agree with you that the dose addition model when applied to any discrete group of chemicals will not always predict an observable effect but if one were to apply that to all the chemicals that it ostensibly should be applied to then we come up with a result that doesn't match reality so that's I think the distinction I am really interested in your HRP team model and could you detail for us a little more how you wish to apply this to overcome the usual problems in toxicology which is demonstrating effects at very low doses etc etc I haven't quite got this point I'm sorry I didn't catch my what approach you're human relevant thank you I have to forgive my hearing it's probably a cumulative effect of too much noise over the years I haven't got the entire approach worked out yet certainly for anti-androgens I would dearly love some data from the pharmaceutical industry for DES for estrogens I think what we do is we look at the potency of DES relative to 17-beta estradiol since that's the control used in most studies I think from those data I shouldn't have closed out the slides but I think you can see there and incidentally this approach was vetted a couple of weeks ago at a conference I attended on validation of alternative toxicological methods and Dan Dietrich from the University of Constance in Germany seemed to agree with this approach which we're trying to discuss this further with him as well he presented some of the same conclusions on diethylstylbestrol as Bob Golden has so we think that probably the human relevant potency threshold for DES and estrogens can be derived from that human data on those adverse effects and it's probably within the range some might say two to five others might say ten we can argue about exactly where that potency threshold is but I think we have the human data to do it there with respect to antiandrogens I think we need more human pharmaceutical data and I'm in the process right now trying to call through the literature for all of that and that's going to take I believe more of a multi-step process because thankfully we don't have that kind of mistaken experiment that we had with DES but it'll involve looking for human thresholds with the antiandrogens and then trying to relate those to the potencies of those same antiandrogens in animal studies. Any other questions? I would add one thing I've requested data from Dr. Quartenkamp and from Dr. Gray so we're hoping to work with them they've been very gracious and contacting their collaborators so we're hoping to come to a broad approach that we can all that's our research group's goal and we're reaching out to others to do that and we're thankful that they've been responsive and we hope that that will work out. Thank you. In deference to those who need a caffeine break or a bladder break we'll take you have to go. You may be now all unduly stressed waiting for me here but sorry about that. Good afternoon my name is Kathy Clark and I'm here on behalf of the American Chemistry Council to present some information describing exposure estimates to humans to Thalade Esters. I'll first begin by providing a brief description of an extensive database the American Chemistry Council has compiled which includes measured concentrations in the environment and in human biological samples. I'll describe two main approaches for estimating human exposure and identify primary sources of exposure for some selected Thalade Esters. I'll then present a comparison of human exposure estimates for those Thalade Esters. This first slide is just designed to remind everyone and this point has been made already that there's a wide variation in the physical chemical properties of Thalade Esters. I've shown on this plot eight of the more common ones and it represents the number of carbons in the straight chain on the y-axis versus the log octanol water partition coefficient and the octanol water partition coefficient being a representative of solubility in water and partitioning to organic matter and so fate in the environment. And the range for dimethylthalate to diisodesylthalate from one carbon to ten and the log octanol water partition coefficient varies over a factor of ten to the eight. So you have a large difference in how these chemicals will behave in the environment and that affects also their selected use, but how they behave in the environment then affects what are the sources of exposure. The ACC has compiled a database which now includes about 800 references and it includes 21 Thalade Diasters and 25 Monoesters and before any references to the input to the database it's reviewed and categorized according to data quality and it includes a review of the analytical and sampling methodologies where contamination precautions implemented, quality assurance and quality control measures were blanks used, where the data corrected for blanks and a very small portion of the data, less than 5%, are categorized as not reliable and those data are included but are separated from the rest of the data in the database. The database includes measured concentrations in a wide variety of media surface water, sediment, soil, indoor air, outdoor air waste water, various foods and as well as human biological samples. There are two main approaches to estimating human exposure, indirect approaches and biomarker approaches. The indirect approaches use the concentration of a thalade ester in a medium and the intake rate of that medium so if you're looking at the exposure due to ingestion of drinking water it's the concentration in the drinking water times the ingestion rate of the drinking water times an absorption factor divided by the body weight and this approach is useful in helping to identify the relative importance of different sources of exposure. The other way to estimate exposure is through back calculation with biomarker data and this uses measurements of thalade ester metabolites in urine to back calculate exposure to the parent diester. This is a good method for looking at total exposure. It doesn't identify what the sources are but another advantage to it is that it has less concerns about contamination issues which are a prevalent problem when you're looking at the diesters with the indirect approaches. In the literature there are many studies reporting indirect estimates of exposure and these studies vary often in the pathways that are considered. Some include only the diet, some include as well environmental media like air or soil and then others may include exposure to consumer products. Exposure estimates have been made using the concentrations in the ACC database and they have been made examining the pathways of ingestion of food and drinking water inhalation of indoor and outdoor air and incidental ingestion of dust and soil. The estimates have been conducted for five age groups adults, teens, children, toddlers and infants. In these next couple of slides I'm going to present the sources of exposure for a toddler for four selected thalade esters. On the left is diambutylthalate and on the right butylbenzylthalate and for both of these thalade esters diet represents the primary source of exposure, 78% and 68% respectively. Diabutylthalate being more volatile and having other sources in indoor air from consumer products and personal care products indoor air represents 14% of exposure for the toddler and ingested dust a lesser amount 8%. Soil and ambient air are negligible. For butylbenzylthalate again diet is the most important followed then by ingested dust at 31% and indoor air is small only at 1%. This presents now the sources of exposure for diethylhexylthalate and diisononothalate, again diet being the primary source of exposure and dust being the second largest source. The next few charts present a comparison of indirect and biomarker estimates of exposure and these estimates are for adults because we have a lot more data, there are many more studies reporting this but I also have results for the other age groups as well just less data so I'll try to present some toddler numbers as I go along. So the y-axis is median intake of diambutylthalate ranging from 0 to 6 micrograms per kilogram per day on that scale and on the left are a variety of indirect studies and they're from different locations, some have different pathways so you have fair variation there from less than a half a microgram per kilogram per day up to a little under 5 and then biomarker studies are shown on the right hand side and again you see variation of about a half a microgram per kilogram per day up to a little over 4. In general though the indirect and biomarker studies they compare within an order of magnitude and as I said there are specific explanations why there are differences in those studies which I could provide an explanation of and in all cases they're less than the reference dose from the US EPA of 100 micrograms per kilogram per day and the numbers for toddler children and toddlers for indirect studies range from about 1.5 to 3.5 and for biomarker studies they're higher about 2.5 to about 7.5. This shows the same sort of comparison for butylbenzylthalate again for adults and the intake ranging on the y-axis there from 0 to 0.6. Indirect studies varying from less than 0.1 to 0.5 and biomarker studies from about 0.1 to 0.4 and again are all less than the reference dose of 200. This shows a comparison for diethylhexylthalate, the indirect studies ranging from about 2 micrograms per kilogram per day up to 11 and for the biomarker studies about 2 a little less than 4. And the final one is diethylenolthalate with far less data for this thalate ester and it's part of the reason why you see there's more variation here. The indirect studies the range of the estimates are less than 0.05 up to about 0.65 and for the biomarker studies from about 0.2 to 0.7 and again they're less than the acceptable daily intake of 120. In summary then human exposure can be estimated using either indirect or biomarker methods and the two methods agree with each other within an order of magnitude and there are many reasons for the differences. They can be attributed to how use of consumer products are accounted for in the exposure estimates. There's a general lack of data on absorption factors and variation in assumptions on absorption create some of those differences. There are also regional and temporal differences. Biomarker methods are useful when it's difficult to quantify all the sources of exposure. For example all the possible use scenarios of different consumer products makes it difficult to account for that with indirect methods. But the indirect methods are useful for looking at various sources of exposure and comparing the relative importance. Using either the biomarker or indirect method the exposure estimates are less than the US EPA's reference doses for diambutyl phthalate, butylbenzyl phthalate and diethylhexyl phthalate and for diisononyl phthalate they're below the ADI. Thank you. Does anybody have any questions? Thank you. Questions? Yes. You will have a copy of that. Sorry. In terms of the sources for the dietary exposure did you know is that from processing, packaging the inks? Or all of the above or yeah. Diets a hard one where it's been difficult to figure out what is the source and a lot of the dietary study in the database is very old back from when the phthalates were perhaps used in greater quantity diethylhexyl phthalate in materials that came in contact with food or they're confounded by contamination problems too. That's the other problem with the analysis, the proper lack of use of blanks. So some of those may be biased high as well. Definitely a nice data set you presented here. I would also be interested in the other 50% of the general population. So I would be more interested in the distribution of exposures not only the median exposures because I think both from the indirect studies and the biomarkers studies we can learn more about the exposures of the upper percentiles of the general population. I agree with you and we have a draft paper which I believe has been submitted to your committee and in there I include both estimates of median exposure and 95th percentiles so I showed that as an upper limit. I was just trying to simplify here for presentation. We have that paper. I have a question for the panel or the phthalate esters panel. I'm not sure any one person can answer this. We've heard different numbers 20, 30 phthalates that are commercial products but we see biomonitoring data and other studies on a much smaller number of them and what I'm wondering is how do we, how does the CHAP prioritize these different phthalates in terms of their importance and why their importance for human exposure and why don't we see more of them in biomonitoring studies? I'm probably not the one to ask. Ray, can you help me out? I'm not sure that it's a question of not seeing them in the biomonitoring studies. I think it's a question of making sure that they are evaluated so if nobody's really looking for some of those let's call them dogs and cats then there's simply no way to go back and evaluate what the exposures might be other than through the indirect methods. What I'm getting at is and I don't expect you to be able to answer it but something for you as a group to think about is we have the HPV program in the report that the ACC I think submitted on behalf of the panel has a long list of compounds, a long list of phthalates yet we hear a lot about DINP, DE, IDP, DEHP but there's a whole long list linear and branched and so on of any every size and shape and how do we get a handle on which ones of those are most important and how might people be exposed to them so I think this is for next baby for next time. The data you presented to what year would you fix it that it would represent the intake or the exposure would you say would be the most up to date exposure estimate in terms of indirect approach and direct approach? I tried to make the cutoff to not use data older than 2000 if possible as far as currency the most current data were from about 2007 I think so they're within that window they're not all the same time they're fairly close. Does that answer your question? Looking at the last let's say for the last set of data it would be 2007 or 2006 how would you take into account the shift or substitution effects within the delayed market? If I can help I think if you look at the CDC data over the course of almost 10 years the last report came out what last year I think was the last reporting of phthalates and there is a minor shift in what appears to be a minor shift in the amount of lower molecular weight phthalates versus say some of the higher ones or two materials that are not on the CDC's panel What would be the year the data of the CDC is based on? I'm sorry? So what would be the time limit of the biomarker data? Well I think the data were collected as part of the 2007 or 2008 NHANES and they didn't report it it took them a while to do all the analysis so it doesn't get reported until early 2009 I think in blocks of time so I would say 2007 would be the last real collection. Any other questions? Thank you if not we'll go on with our last speaker Well I think at this point it's you probably need to just go on and finish. For those of you who don't know me my name is Ray David I'm a toxicologist not just with BASF Corporation but I am here on behalf of the Phthalate Esters panel and the last speaker of the day and I'm mindful of the time so I'll try not to deli I wanted to talk about some of the research that's been done not only by the panel but by others to look at species responses and how different species respond to treatment the reason I'm doing that is because I think as you do your assessments this is the kind of information that you will hopefully take into account when you do your assessments. The panel has long been involved in looking at how different species respond. At first it was focused on peroxazone proliferation. The reason we did that is because peroxazone proliferation was the link to liver cancer which was the primary concern at the time and we have since moved to or we have also looked at responses on the reproductive tract particularly the testes. Our biggest challenge is looking at how to evaluate developmental effects in other species and I'll give you some examples we've looked at cells in culture so hepatocytes, human hepatocytes, we've also looked at non-human primate species to use them as potential surrogates for humans and what we've learned really is that human hepatocytes don't respond the same way as say rat hepatocytes either in culture or in the entire in the intact animal and these are just a couple of examples of the literature. The same is true for adult marmosets or cinemagous monkeys either of those monkeys don't respond the same way at least in terms of the liver effects as rats do. We've tried to for a long time we've struggled with that phenomenology I'll call it and try to relate that to what the mode of action is and back in the 90s when PPAR was identified it gave us a way of really associating the species response with some mode of action, some mechanism obviously with some of the more recent information that debate is rekindled and you'll hear from people tomorrow about whether or not that is still relevant for the wild type animal or for humans and by the way if I can I've been sitting here for a long time listening to the questions and answers and you'll forgive me I'm kind of a long time student of phthalate toxicology so if I can respond to questions as they came up I'd like to do that. There was a question about reversibility, recovery of some of the effects. There are actually studies that have looked at the recovery of hepatic effects and cancer and non-cancer endpoints in rodents that were treated either with DEHP or DINP so if that's of information I mean it's certainly one of the reports you have already from the last CHAP and the others have been published so if you are interested in looking at those we can provide you information. With respect to reproductive effects of course the male rat is very sensitive to the effects of some of the phthalates like DEHP and dibutal testicular effects are very obvious as per metagenesis for example but when we look at other species we don't see these kinds of effects particularly we've used the marmoset monkey and the synomologous monkey here are two or three studies in which those adults were treated for a period of time without any evidence of the kinds of testicular effects that we would see in rats. A few years ago we did a study with our colleagues in Japan looking at juvenile marmosets so these were weanlings that were treated with a very high dose of DEHP for a considerable length of time and again we didn't see the kinds of testicular effects that we would have seen in a rat. Now couple this with, and by the way I have my scribbles here, there's another recovery paper if you're interested. Many years ago and I'm sorry I've forgotten the first author's name there was a study in which rats were treated with a high dose of DEHP, they were neonates they were treated for about one week and then allowed to recover and they then followed how well the testicular let's say the testis function returned to normal. The interesting part about the non-human primate studies is that we have some evidence that the same kinds of, or let's say the lack of response occurs in humans. Two studies in particular where we know humans were given a particular dose the first one was a group from Denmark that smeared volunteers with cream that had dibutal phthalate and then they looked at hormone levels and they looked at testicular well, they looked at sperm counts and they really couldn't find any kind of an effect on those volunteers. There's another group that looked at a very small population of neonates that were exposed to DEHP as part of a procedure called ECMO which uses a tremendous amount of PVC tubing and therefore the dose of DEHP to those neonates was probably the highest that's been estimated for humans. In this small population they couldn't see any evidence of the kinds of genital effects that we might have seen say in rodents. Now again we lost a dose and recovery kind of an experiment in a human. Can this phenomenology be associated with any kind of mode of action? The answer is no because for the past 30 or 40 years we have been trying to understand what the mode of action is for phthalates on the testes and we really don't have a good answer to that. So what we have is just the phenomenology, the fact that we can test another species and we don't see these responses. For developmental effects I think Nina already referred to the study by Kevin Guido in which he treated pregnant mice and then looked at the male offspring to see if they had the same kinds of effects that were seen in rats. And the answer was no, not at the same dose level. He could not reproduce the kind of testicular agenesis or dysgenesis that is observed in rats. He couldn't reproduce it in mice. Recently there was a study in which Richard Sharpe's group also treated pregnant marmoset monkeys and again they could not reproduce the same kind of testicular effects that one might see in rats. And yet we have the study from Shayna Swan that suggests that there is some kind of an association. How is it that we can reconcile that lack of effect where we know there's a treatment and some of the human data that is being reported? It's very difficult to do that I think. I have a scribble here. One more recovery study for you. The first author's name is Barlow, B-A-R-L-O-W. He worked with Paul Foster and as a postdoc they looked at the recovery of animals that had been treated in utero and then allowed to recover. They were looking primarily at nibble retention for example and some of the other sensitive endpoints. So that's another one that you may want to look at. We as a group have been trying to delve deeper into the issues of developmental toxicity. How is it that we can come up with a system by which we can either screen or we can investigate the mode of action? We went to some investigators at Michigan State and we asked them to conduct some studies for us. There was a pointer here. They got some cells, mouse lating cells that do express testosterone. Thank you. They treated the cells with either MEHP or DEHP for up to 48 hours. I've just listed some of the effects here. What I wanted to show you was this graph. Someone mentioned to me why am I showing this graph? The reason I'm showing this graph is because the testosterone level goes up with increasing dose of MEHP. That's not at all what we wanted. It was supposed to go down. This as a model system obviously wasn't going to work. We've also been working with a group at Brown University that has working on a technique in which they transplant tissues from a fetal rat into an adult mouse and from fetal mouse into an adult rat. Looking at whether or not the fetal tissue responds the same way in a host species, a different species. I can't tell you the entire results. I think there's been one paper published from the group so far. It does appear that the native tissue responds the same way as it would in the intact animal. So the mouse does not show the kind of testicular related effects that would be seen in an intact mouse, even though it's implanted in a rat and vice versa. This group also has a grant, not from us, but they have a grant to look at a human tissue from abortuses. The data are very preliminary but it would suggest that maybe humans are along the same kind of responders as mice. In other words, they can't seem to find the same kind of testicular related effects. There are a lot of questions here. The kinetics, for example, how robust is the transplant. Hopefully this will lead to a mode of action. At least that's what we're looking for. So what I want to ask you is a question about just how appropriate is the rat as the model. If we know that we have these other species that don't seem to respond the same way and with a rat we got all kinds of effects, is it appropriate to use only the rat? I mean you are not bound by statute to use the most sensitive animal species. You have the scientific freedom to really adjust your conclusions based on the most appropriate species. I hope that you can do that. The other question I have for you is a bit self-serving. I work for industry. We in industry are faced with a situation where we don't have a mode of action for many of these things. What level of evidence could we provide to you as an example of independent evaluation group? What level of evidence could we provide to you to demonstrate the lack of an effect in a human? I'm not sure I know what that is. Short of determining what the mechanism is in a rodent and demonstrating that that doesn't occur in a human, I don't know what else we could do. Perhaps that's something that could be discussed over the course of time. Thank you. Thank you. If I could add something to your last statement that makes it a neutral statement as opposed to a biased one. When you declare that you want to find that something isn't relevant, you're indicating a bias when in fact we are supposed to be independent and I think it's important to us to find out what are the data that shows either no relevance to humans or relevance to humans. There's an equal chance that this would be relevant. So I don't think we should be looking just under those rocks that might show that it's not relevant. Another point, you said that humans and mice seem not to react like the rat and I think that's a very important word, seem. Absolutely. There are kinds of issues that could make if we were to look at them, marmosets and other monkeys and mice react similarly to rats. At a simple level, it would seem to me that a comparison just starting off with in vitro looking at rat cells and showing that they respond at any point testosterone synthesis or whatever. And then try to find a mouse line that you could compare it to. That apparently hasn't been done. That's correct, it has not been done. And the reason it hasn't, well that it hasn't been done with mice because my colleague Nina could probably answer that better than I was. Those tissues don't seem to retain the same capability for the synthesis once you take them out of the intact. Maybe it's a matter of the combination of cells, the environment that we put them in. As Nina mentioned, there was a study by a French group, Lambrose, the first author's name. And they looked at cultured rat cells and they looked at cultured human cells. And they got a response sort of with the rat cells and didn't get any with the human cells at least with respect to testosterone effects. Whether that's convincing enough, that's a question you're going to have to unfortunately decide. Another issue I think relates to when I was reading the study on marmosets they didn't see in effect when they look at the endpoint. But assuming there is a mode of action that phthalates are inducing. There must be, right? Then really the critical question is does the phthalate get to that whatever that's going to cause the effect. It obviously doesn't get to the rats because we're getting in effect. Does it do so in the marmoset monkey? And we don't know that. As far as I can tell from this one study, and I'm not an expert in this area, but just reading that one study, they make no mention of what the target dose is. That's true, they do not. And until you have that, seem has to be the word that you use. You are correct that how much pharmacokinetics, or toxicokinetics plays into the lack of response isn't entirely known. Certainly there are reports in the literature that talk about the blood levels in marmosets relative to rats. Pregnant marmosets versus pregnant rats. That was a kinetic study that again, we sponsored because to try and answer exactly that same question. And there is very much a difference in the dose that reaches the target in a rat versus a marmoset. At least based on the blood levels. In a way that would explain the results. I'm sorry? You said there's a difference. There is a difference. And is that difference as a go in a way consistent with the results? That is effect in the rats and not in the marmosets. So is it lower? It is lower by a factor of, I've forgotten, five or eight, is it? I don't know if that is enough of a difference in the dose to account for the lack of a response. I bring these up because these are the issues that we're going to have to grapple with and in the absence of data it's a challenge. I do respect the fact that you will have a lot of information to review. If you find there are questions like that I would encourage you to funnel them through CPSC to the ACC. There may well be information that you simply don't have at your fingertips. Thank you. The mouse may not react in the same way as the rat. Can you suggest a mechanism for these differences? I wish I could. Can you suggest a mechanism as to why the marmoset doesn't react in the same way as the rat? Are you suggesting that mouse and marmoset are more like humans than the rat? I'm suggesting that a marmoset could be more like a human. Is there any evidence you would base that statement on which would help us make headway with these comparisons? What kind of evidence would you feel comfortable with? For example molecular mechanisms but you said already that's not available. Any other evidence you want to suggest really? I'm at a loss but you need to help us. I would be very grateful if you could. Can I give that some thought? Or is at this point the only fact that might suggest more relevance of the marmoset to the human that both are primates? Is that all? Certainly that is one factor. At this present point in time you're not aware of any other additional material evidence. Maybe Nina is. The lead author of Scott in 2009 and that's a really state of the science summary of what's known about fetal steroidogenesis across several different species. That highlights research areas, strengths and weaknesses. Just bring that to your attention. Just to be transparent about it that report was commissioned by CEPHIC so that was industry funded. To repeat Chris's comments earlier CEPHIC had no input on the conclusions of that study. The idea was to try and generate research ideas for further programs. This Scott et al report is not published in the peer reviewed literature but it is. Where is it? I think it's in endocrinology reviews. It's like a 20-30 page paper. It's very thorough. I think one of the questions Andreas was asking was one of the slides you were alluding to. I think there were two studies in humans, the Janu study where they put the cream on men and then also the ECMO study, the neonate study, as showing no associations with human health endpoints. Those studies, especially the NICU study where they looked at these children when they were younger adults or in puberty was an extremely small study. I think there were 19 children followed. It was published but I wouldn't really put any weight on it just because of the small sample size detecting anything would be nearly impossible. From a statistical perspective, absolutely. I agree with you. There have been discussions within FDA about doing a follow up to a larger population but as I understand that is not a simple undertaking. So I guess the question of priority has come up and whether or not that's something that is worthwhile funding. You were hinting to species differences and hinting to the fact that the blood levels of the delates might be different. That was the mama set in the red study by Kessler and maybe you know that you also have some blood data in the meantime available also for humans. We had a single data set from our study a couple of years ago but in the meantime I think also Michael had the honour to be at the Berlin conference and some new data has been presented and to my recollection in essence the outcome was that the human data was more comparable to the red data than the mama set data. I think Reiner it was an eCPI study. It was the eCPI study dealing with the conversion factors and it was also confirmed by Johannes Filzer who independently together with Fromme did the same sort of study but we did the study of 10 individuals and Johannes had a lower amount of people. Again to make the point my recollection was that human data was very much comparable with the red data and not the mama set data. Is this data available for the committee to see? I think Mike Babich was at this closed BFR environmental protection workshop and he has the data. We have probably to clarify whether he can get the whole data collection but I think from our study that will be available once upon a time. To be clear the study is still in the reporting phase at the moment. We haven't got a final report available. We debated extensively amongst ourselves as to whether it would be useful to provide a summary at this point but frankly until the data analysis and reporting is complete I don't think that's reasonable to do just yet. But given the duration of your review I think it's reasonable to conclude that we should have that available before your deliberations are concluded. We certainly appreciate having that. Any other questions, comments? If not we will take a short break and then the committee will reassemble. Thank you to all the speakers. All right can we just briefly come back to the meeting. One two before we quickly adjourn. Just to thank everyone again for their presentations and to encourage you to submit to us any information that you feel will help us in our task. I think that's really critical and we really value that. Since as you all I think can tell this is a very difficult charge that we've taken upon ourselves and we want to walk away from the report being confident that we've done the best job we can. Having said that if there are no comments from the committee I'm going to adjourn this meeting and we'll re-adjourn tomorrow morning at I believe it's 8.30. Again thank you all.