1 MEETING
2 OF THE
3 SCIENTIFIC REVIEW PANEL ON TOXIC AIR CONTAMINANTS
4 CALIFORNIA AIR RESOURCES BOARD
5
6
7
8
9
10 SOUTH SAN FRANCISCO CONFERENCE CENTER
11 255 SOUTH AIRPORT BOULEVARD
12 SOUTH SAN FRANCISCO, CALIFORNIA
13
14
15
16
17
TUESDAY, APRIL 13, 1999
18
9:00 A.M.
19
20
21
22
23
24 Janet H. Nicol
Certified Shorthand Reporter
25 License Number 9764
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
ii
1 APPEARANCES
2 MEMBERS PRESENT:
3 Dr. John Froines, Chairman
Dr. Roger Atkinson
4 Dr. Paul D. Blanc
Dr. Craig Byus
5 Dr. Gary Friedman
Dr. Anthony Fucaloro
6 Dr. Stanton Glantz
Dr. Peter S. Kennedy
7 Dr. Hanspeter Witschi
8
REPRESENTING THE CALIFORNIA AIR RESOURCES BOARD:
9
Mr. Lynton Baker, Staff Air Pollution Specialist
10 Mr. Bill Lockett, Deputy Ombudsman, Northern California
Mr. Peter Mathews, Office of the Ombudsman
11
12 REPRESENTING THE OFFICE OF ENVIRONMENTAL HEALTH HAZARD
ASSESSMENT:
13
Dr. George Alexeeff, Deputy Director for Scientific Affairs
14 Dr. Melanie Marty, Senior Toxicologist
15
REPRESENTING THE DEPARTMENT OF PESTICIDE REGULATION:
16
Mr. Paul Gosselin, Assistant Director (Via Telephone)
17 Dr. Lori Lim, Staff Toxicologist
Dr. Jay Schreider, Primary State Toxicologist
18
19 PUBLIC SPEAKERS:
20 Dr. Dale Hattis, Clark University
Dr. Lorenz Rhomberg, Harvard University
21
22
23
24
25
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
iii
1 INDEX
2 PAGE
3 AGENDA ITEMS:
4 1 Discussion of margin of exposure (MOE) and 99
reference exposure level (REL) approaches.
5 DPR and OEHHA
6 2 Scientific presentations regarding non-cancer
risk assessment.
7
Old problems and Some New Approaches in 3
8 Non-Cancer Risk Assessment.
Lorenz Rhomberg
9
Inter-individual Variability and 63
10 Quantitative Low Dose Risk Projections.
Dale Hattis
11
3 Review of findings for DEF as a toxic air 140
12 contaminant.
SRP
13
4 Update on the schedule of pesticide documents 161
14 to be submitted to the SRP.
DPR staff
15
Adjournment 175
16
Certificate of Reporter 176
17
18
19
20
21
22
23
24
25
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
1
1 P R O C E E D I N G S
2 CHAIRMAN FROINES: Let's begin. Welcome,
3 everybody.
4 The first thing I'd like to do is to introduce two
5 people, Dr. Elinor Fanning and Lesley Dobalian, who are
6 sitting back there. Elinor and Lesley are working for the
7 panel at this point.
8 The load on this panel has increased dramatically,
9 especially insofar as we are trying to plan workshops, we're
10 trying to address issues outside of having specific
11 documents, and we are having, as everybody knows, 51 acute
12 documents that we had to review the individual compounds.
13 And George Alexeeff tells me that we are about to
14 receive sometime in the next few months the chronic document
15 which will have how many chemicals?
16 DR. ALEXEEFF: 120.
17 CHAIRMAN FROINES: 120.
18 So we decided that we could use some help, and
19 Lesley and Elinor were kind enough to agree to work with us.
20 And so that, I think, is going to provide us with some
21 technical backup that will be useful.
22 We'll make sure the panel has information about
23 how to reach them and vice versa, so that questions that the
24 panel members have can be raised with them.
25 The second thing that I wanted to do today was to
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
2
1 change the order a bit, and because we have been dealing
2 with noncancer risk assessment, first with the 51 acute
3 documents, soon with the 120, or what have you, chronic
4 documents, and with the anticipated pesticides that we're
5 going to be addressing, we have gotten more and more
6 involved in issues of noncancer risk assessment, and we have
7 had to be concerned with the methodologic implications of
8 how one does that risk assessment.
9 So what I did was to invite Lorenz Rhomberg from
10 Harvard University and Dale Hattis from Clark University to
11 talk about their research on trying to take some fresh looks
12 at noncancer risk assessment.
13 And I think that it would be better to start out
14 with Lorenz and Dale, because I think what they have to say
15 may have relevance for the discussion of the margin exposure
16 and reference exposure level approaches that DPR and OEHHA
17 use, and so I thought we would switch that.
18 So we didn't talk about who goes on first.
19 Dale, why don't you -- oh, Lorenz is going to go
20 first. Okay.
21 Why don't we start with Dr. Rhomberg and
22 Dr. Hattis and then move to the margin of exposure
23 discussion.
24 And I had one question, housekeeping question. Is
25 Paul Gosselin with us?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
3
1 DR. SCHREIDER: I'm Paul Gosselin for the day.
2 Paul has to appear in the Legislature. For this week he's
3 acting director and chief deputy and assistant. So he will
4 try to make it by telephone call.
5 CHAIRMAN FROINES: Well, that's what I was
6 wondering. Is there going to be a conference hookup?
7 MR. MATHEWS: Yes, there will be.
8 CHAIRMAN FROINES: So he'll pop in at some point
9 and you'll let us know.
10 MR. MATHEWS: I will.
11 CHAIRMAN FROINES: I don't know if the panel has
12 any other issues to raise before we get started at this
13 point.
14 Hearing none, why don't, Lorenz, why don't you
15 begin.
16 And this is intended to assist the panel in
17 looking at noncancer risk assessment in the hopes that over
18 time we can develop new methodologic approaches that advance
19 the state of the art, so that we start to look at these risk
20 assessments with more confidence that they have -- that they
21 are more satisfactory.
22 For example, as you all know from reading the
23 Kenneth Crump article, there are problems with using NOELs,
24 and I won't go into that.
25 So good.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
4
1 DR. RHOMBERG: Should I start?
2 CHAIRMAN FROINES: Yes, please.
3 DR. RHOMBERG: Thank you all for having me.
4 Since I haven't participated in your meetings
5 before, I'm at a little bit of a disadvantage of knowing
6 about exactly where you are and what your questions are on
7 this, so I'm hoping to keep this rather informal. If I'm
8 going off in the wrong direction, something that's too
9 elementary or too beyond what you thought about, just tell
10 me and I'll try to make some on-the-fly adjustments.
11 I've titled my talk, "Some Old Problems and Some
12 New Approaches in Noncancer Risk Assessment."
13 CHAIRMAN FROINES: Lorenz, let me stop you just
14 for a second.
15 DR. RHOMBERG: Sure.
16 CHAIRMAN FROINES: I think if the panel wants to
17 ask questions during the course of the presentation, that's
18 fine.
19 But I think we'll hold questions, George, from
20 members of the audience.
21 But after Lorenz has had -- has finished then I
22 think we'll also give your folks a chance to ask questions
23 if they have them.
24 DR. RHOMBERG: Fine.
25 Well, as I said, this is a title I've chosen.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
5
1 Notice it doesn't say solutions.
2 I think that we have some problems in noncancer
3 risk assessment that are longstanding problems and there are
4 people that are trying to grapple with them. And I think
5 what I'm really going to be talking about is the grappling,
6 rather than exactly what the solution is and what's to be
7 done.
8 First, let's sort of start out where are we, what
9 is the solution, what is the source of these old problems.
10 You know, really the basic approach to noncancer
11 risk assessment has been, one, to assume that these are
12 endpoints with thresholds, somehow that there is variation
13 and sensitivity among the population, and that's why you get
14 a dose response, that at some doses some will respond and
15 others will not, and at higher doses more will respond and
16 less will not, because of variations in sensitivity of their
17 individual thresholds.
18 That's a toxicologically reasonable thing for a
19 lot of kinds of endpoints, but again that's something that
20 probably could be thought about more in terms of underlying
21 rationale.
22 And then the other thing is the basic quantitative
23 approach that's been taken. It's really a safety
24 assessment, rather than a risk assessment kind of approach,
25 that dates back to the Lehman and Fitzhugh paper in 1954
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
6
1 that really established this method, and with minor changes
2 or embellishments it's really gone on since then in more or
3 less the same way.
4 What Lehman and Fitzhugh did, they were working
5 for FDA, and they wanted to have a method for assessing the
6 safety of food additives. Here were compounds that were
7 added to foods in very small amounts, and so you really were
8 just asking the question of can you really show that this is
9 well below any levels that would can cause toxicity, and
10 therefore we are fairly sure that any kind of thresholds for
11 human toxicity that would be present for these compounds
12 that we are well below that.
13 And they argued that well 100-fold below a NOEL in
14 animals, a no effect level, in animal tests is probably
15 sufficient for that.
16 And they argued it mostly on the basis of
17 experience. We don't really see anything too much worse
18 than this. And they did mention that this could, in some
19 sense, account for the fact that humans and animals might be
20 different in the doses that give them toxic effects and
21 moreover that individual humans are different from one
22 another in their tolerances, and somehow that this 100-fold
23 could take care of all of that.
24 They really weren't all that explicit about saying
25 10-fold for this or 10-fold for that, but just 100-fold
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
7
1 ought to be enough.
2 By and large, that methodology is continuing
3 today. And here's what it looks like in a more modern
4 guise, but it's basically the same thing.
5 This is what the EPA does. They call their output
6 from this a reference dose or RfD. You take a NOEL from
7 animal tests, the most sensitive sex and species if you've
8 got a variety of endpoints, and you usually want to have
9 them to allow for the fact that different endpoints and
10 different sensitivities can pop up in different species.
11 And then you divide by a series of uncertainty factors,
12 which are usually a factor of ten, but not invariably, and
13 you can argue about their values. And you might also divide
14 by something called a modifying factor or a data quality
15 factor to further hedge the allowable dose, if you think
16 that that's necessary, considering the quality of the
17 database, the quality of the data that go into it, the
18 completeness and so on.
19 And basically what this amounts to is modifying
20 the Lehman and Fitzhugh methodology to, one, recognize more
21 reasons, more extrapolations that need to be done and
22 therefore adding some more factors for them. And, two, a
23 little bit more explicitly dividing the overall margin that
24 we have there into pieces that are sent out representing
25 various factors.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
8
1 So each of those factors is an extrapolation in
2 some sense that we have to make from animals to humans in
3 assessing the safety.
4 And because we are uncertain about that
5 extrapolation, we sort of divide by ten to account for that
6 in some sense. What we mean by that is something I'll get
7 into in a minute.
8 These are traditional ones. UFA for animal-human
9 extrapolation, because we're using animals we want to know
10 about human doses.
11 UFH for what is often called average human to
12 sensitive human. As we'll talk about more, that's a little
13 bit of an oversimplified way to do it, but it's meant to
14 account for the fact that in the human population there may
15 be some that are very sensitive and we want to protect them
16 as well, so we divide by ten again.
17 Subchronic to chronic. If you only have a
18 subchronic test and you want to know about what would be
19 safe for chronic lifetime human exposure, as is often the
20 application here, you might want to add something for that.
21 And if you have a test that doesn't actually
22 identify a NOEL you might want to say, well, we'll use a
23 factor in for that.
24 So the point is that you've added these things by
25 explicitly looking at the extrapolations that you want to do
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
9
1 or the uncertainties in those extrapolations that you want
2 to account for.
3 Well, what's wrong with this methodology? It's in
4 some sense served us well, and often you get those
5 statements, well, we've been doing it for 50 years, and we
6 don't really have too many disasters on our hands, so maybe
7 it's okay.
8 But there are some difficulties.
9 One is that we keep thinking of new things that we
10 might want to be aware of or careful about, and the one
11 that's being talked about these days is risk to children.
12 And there's a mandate in the Food Quality
13 Protection Act, for instance, that says, well, you have to
14 argue whether you do or don't need yet another factor of ten
15 for children as being yet more sensitive than even the
16 sensitive humans accounted for in UFH. Or do you?
17 Another problem is that, you know, these factors
18 of ten, where do they come from? There's a famous saying if
19 we had 12 fingers they would be factors of 12. They're just
20 order of magnitude adjustments that we don't really have a
21 great idea about whether they're big enough or too big or
22 small enough or too small, or whatever, and we would like to
23 try to apply some data to those to see if in fact this kind
24 of safety that we presume is actually being accomplished.
25 Another problem is that in some sense if these
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
10
1 represent the variable quantities, then it would be unlikely
2 to be extreme on all of those quantities at once. We're
3 sort of assuming that we are by multiplying those things
4 together, and some way we'd want to take into account that
5 fact.
6 So we have a problem here in that things are
7 conservative to some degree, because we have allowed for
8 uncertainties and variations in case to case that may or may
9 not exist in our particular case.
10 We have a hard time saying quantitatively how
11 uncertain we really are.
12 So one set of problems.
13 And another set of problems has to do with what
14 we're asking risk assessment to do nowadays. We're asking
15 something very different than Lehman and Fitzhugh were
16 asking. Lehman and Fitzhugh had manifestly small exposures
17 that we just wanted to screen to see whether they were safe,
18 but there was no sense in which we were then going to start
19 jacking up the exposures to these things up to the safe
20 level.
21 But once you say you can determine what a safe
22 level is, there's a slippery slope here and we're slipping
23 down it rapidly over time. Well, I suppose it's taken us 40
24 years, but we've continued to slip down it. And the slope
25 goes like this.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
11
1 Well, here's a screening, sort of a one-way
2 screening kind of pronouncement. Below this level we're
3 really fairly sure that we're safe, but above it we're not
4 saying we're unsafe, we're just saying we're progressively
5 less sure that we're safe.
6 That doesn't really matter if you always have low
7 exposures and you just want to make the argument this is
8 okay.
9 But once you think you can do that, then you start
10 to say, well, how much could I jack up the exposure, if
11 you're not applying that to food additives, but to something
12 else, how much emission will you allow from this factory,
13 how much cleanup will you require at this toxic waste site
14 or something like that until you get up to a level of risk
15 that will be acceptable. So you can start titrating risk so
16 to speak.
17 That puts a very different interpretation on this
18 kind of level.
19 Another thing is that we are then, because we're
20 doing those things, we are then talking about exposures that
21 are sometimes in the range of this RfD or maybe even above,
22 and we say, well, we don't say that there is a risk above
23 the RfD, or the acceptable daily intake, or whatever like
24 that, we're just less sure that there isn't one.
25 And yet the interpretation of that becomes very
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
12
1 important. The more we're asked, for instance, to do things
2 like cost-benefit analysis, you have to say what is the
3 benefit of this regulation in terms of noncancer cases of
4 toxic effects avoided. How do you put any kind of measure
5 on that, on being a little bit more sure than you were at a
6 lower dose that nothing will be -- it will be little bit
7 less sure that nothing will happen, but you don't really say
8 how much is going to happen.
9 So in some sense there's really no dose response
10 analysis in here. There's no estimation of risks.
11 And more and more we're being asked to do that
12 kind of stuff.
13 So in some sense this methodology isn't serving us
14 as well as it used to, because the kinds of questions we're
15 asking about it, are asking it to cover, are different.
16 Here's a quick summary of some of the problems.
17 The first one I didn't even mention, I should just
18 mention that briefly, the nature of the NOEL. I guess
19 you've talked about this before. NOELs have to be one of
20 the doses that was tested and therefore the experimental
21 design that was chosen by the experimenter has a big
22 influence on what is determined as the NOEL. It's not the
23 same thing as a threshold, and it's really more about the
24 experiment than it is about the toxicology of the compound.
25 Some of that can be gotten around by using the
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
13
1 benchmark dose procedure that I'm sure you've talked about.
2 I mentioned are the uncertainty factors of ten too
3 big or too small. There's a problem of compounding
4 conservatism if each one is slightly conservative and you
5 multiple them all together, you get a very conservative
6 result at the end.
7 Here's another one I didn't mention. When you
8 have pharmacokinetic data or something else that in some
9 sense is accounting for some of the extrapolation and is
10 being specific about some of the uncertainty that you have
11 in one of these extrapolations, how do you then reduce the
12 uncertainty factor to account for the fact that you have
13 more information, if this is to account for uncertainty and
14 is to help you with the extrapolation and you have another
15 way of doing that better, somehow you should get rid of at
16 least some of that uncertainty factor. We don't really have
17 a good idea of how much of it, because we haven't really
18 explicitly said what we think we are accomplishing with
19 those factors.
20 In the end the conservatism is unclear and it
21 isn't consistent. Is this a very conservative procedure or
22 is it a slightly conservative procedure, and maybe not even
23 conservative enough, and we don't really have a good way of
24 making that argument other than by polemics.
25 Then as I say, this question of use and risk
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
14
1 management, a one-way conclusion about safe that doesn't
2 help us when we want to talk about exposures in the ranges
3 where there might be risks that start to emerge and we want
4 to look at costs and benefits or we want to say what we're
5 accomplishing with our regulations.
6 So what do we need then?
7 We need a risk measure, not just a safety line,
8 but a risk measure that will say something about what's the
9 probability of responses at high doses, outcomes and the
10 probabilities of occurrences and presumably also something
11 about their severity. That's a tough one, though.
12 Characterization of uncertainty. We want to try
13 to separate our estimation of these uncertain quantities in
14 our characterization of that uncertainty from our decisions
15 of conservatism about what to do in the face of that
16 uncertainty. It's really a question of separating risk
17 assessment and risk management. The way we have it now,
18 it's sort of built in. We define a safe level as something
19 that has these safety concerns built in and if those are
20 seen as excessive or not strong enough, somebody doesn't
21 have a real easy way of disentangling what we've estimated
22 about the compound from these risk management concerns, and
23 we want to be able to do that and the way to do that is to
24 fully characterize the uncertainty in the extrapolations.
25 We want a means to incorporate case specific data,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
15
1 pharmacokinetics and endpoint-specific pharmacodynamics,
2 meaning models of the toxic effects when we have particular
3 understanding of a compound, rather than just treating it in
4 this generic way, and that somehow reduces some of those
5 factors that we have to apply, but we don't have a good way
6 of saying how, and so all together we want a sound framework
7 for how to do it.
8 Well, what can done to do that? As I said, I
9 wasn't really going to have solutions, but I'm going to show
10 you sort of a tentative idea towards a solution that was
11 proposed a few years ago by some of my colleagues at
12 Harvard, Sandra Baird and John Evans and some others.
13 And basically what they were suggesting in a paper
14 that's in Human and Ecological Risk Assessment is to think
15 of it in the problem this way.
16 These uncertainty factors are extrapolations, the
17 magnitude of which varies from one chemical to another, from
18 case to case, and we can get empirical data on how that
19 variation happens. So if we can now describe a distribution
20 of the magnitude of the adjustment that's needed for each of
21 those factors, a statistical distribution, and then work
22 through all of those and call them adjustment factors now,
23 rather than uncertainty factors, because the uncertainty
24 arises in the question of where in the distribution you are
25 for the particular chemical you're looking at at the moment.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
16
1 As I say, an adjustment that varies from case to
2 case.
3 So you can say let's try to get empirical
4 characterizations of those adjustments and then work through
5 the joint distribution of all of those to come up with -- PT
6 there is population threshold, meaning a dose that for a
7 whole population should be below the level it's causing an
8 effect.
9 And we don't know that number for sure, but we can
10 say something about the distribution of where it would be
11 and how sure we want to be to be below it, it's something we
12 can do as a risk management decision.
13 And if you try to do that, and they tried to do
14 this with the way that they admitted was sort of
15 preliminary, because getting these distributions is not at
16 all easy, as a lot of our -- the following part of the talk
17 will be about, but if you do this, you get some sort of
18 estimates of how they do this. I'll let you read the paper
19 and see how they actually propose to do all of these
20 things.
21 This is the actual result they got for acetone. I
22 can't remember these. Acetone. Where you work through
23 those distributions with a Monte Carlo kind of approach and
24 they got this distribution of what would be the likely
25 population threshold and the arrow there shows where the EPA
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
17
1 RfD was.
2 And it's sort of where you would hope and expect
3 that it would be in that it's sort of conservative compared
4 to the whole of the distribution, with not too much chance
5 of not being stringent enough, but a fair probability of
6 being more stringent than would may be necessary, according
7 to this kind of approach.
8 Well --
9 DR. GLANTZ: What percentile is that in the
10 distribution?
11 DR. RHOMBERG: I don't know offhand, but it's
12 somewhere close to a fifth. So if you want to say this is
13 sort of in the 95th, you know, percent, lower bound on
14 something, you'd be pretty close, actually.
15 DR. GLANTZ: Wouldn't that say that the current
16 approach actually is working? Because you go through this
17 more complicated model and you end up about where you would
18 hope to be with the simple-minded approach.
19 DR. RHOMBERG: I think it suggests that it was
20 working for the purposes that it was trying to do. But as I
21 was trying to suggest to the earlier slides and the earlier
22 discussion, some of our -- we have some new purposes as
23 well, for which it wouldn't serve so well. But for the
24 purpose of finding a reasonably conservative, but not unduly
25 conservative value, that's probably good for protecting,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
18
1 probably, most of the people. Yeah, it worked pretty well
2 in this case.
3 There's a caveat there, though, and that is that
4 in coming up with these distributions they were highly
5 influenced by the existing factors of ten and the rationales
6 for them. They didn't really have a good way of getting
7 good empirical data on all of these distributions, so in
8 some sense they're just building in those distributions
9 again, and it's not too surprising they got some of them
10 back.
11 So this is an interesting approach.
12 I'd like -- I like it as something you can do. Of
13 course it is technically difficult, but it isn't without
14 some further issues and I just want to point out some things
15 that it doesn't do so well.
16 One is in some sense it's mixing uncertainty and
17 variability. We have uncertainty and we're talking about
18 one compound here at a time for which we're doing this for
19 acetone, or whatever, and we're saying, now, well, there's
20 this distribution of what could be the relative toxicity of
21 acetone in the animals that were tested, which I don't know
22 offhand what they were, and in humans.
23 But, you know, one value from that distribution
24 applies to everybody, because that's sort of an
25 animal-to-human thing, we're talking about one species of
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
19
1 animals and one species of humans, so we're taking one value
2 from that.
3 On the other hand, our adjustment for -- I pointed
4 to the wrong one before --
5 DR. GLANTZ: Excuse me. Wait. You just lost me.
6 I thought -- I don't see how you're taking one
7 value. I thought the whole point of this was that you're
8 considering the whole distribution.
9 DR. RHOMBERG: You're considering the whole
10 distribution because you don't know which one value in that
11 distribution applies to acetone, but if you were sort of
12 omniscient, you would say this is a distribution over all
13 chemicals of relative toxicity in animals and humans, and
14 acetone is one of those chemicals and so it's somewhere in
15 that distribution. Not knowing where it is, we choose, we
16 use the whole distribution, but, if we knew, it would be one
17 particular value.
18 DR. GLANTZ: So you're saying then -- see, I had
19 assumed that each of the distributions applied to the
20 chemical of interest. You're not saying that?
21 DR. RHOMBERG: No, it doesn't. It's on the, as I
22 say, it's on the logic that the needed adjustment for each
23 of these factors varies from one chemical to another. And
24 we can get empirical data on how that happens among
25 chemicals and then for a particular chemical we don't have
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
20
1 any such data, we don't know where it is, so we have to say,
2 well, we'll assume it's drawn from the same distribution
3 that applied to the other chemicals.
4 But whatever it is, it's one of those things and
5 so in some sense the one value applies to everything.
6 That's true for that one, but in a way you're
7 making the point that I'm trying to make here, which is that
8 in this factor for sensitive humans, human sensitivity does
9 vary amongst everybody for the one chemical and so everyone
10 isn't the most sensitive human. So, in fact, the human
11 population that we're interested in assessing the risk that
12 was going to be spread through this distribution and the
13 whole distribution will apply, and we're mixing those things
14 together and maybe it would be a good idea to keep them
15 apart.
16 Dale will have a lot more to say about that in a
17 minute.
18 There are some other things.
19 The subchronic to chronic and the LOEL to NOEL
20 distributions, how you would actually do those empirically
21 there's some real questions about that I won't get into now.
22 What I'd really like to say is, though, this is a
23 good start, but then the question is how do you empirically
24 get those distributions, and how do you think about them.
25 So what I'd like to spend the rest of my time on
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
21
1 is some questions about thinking about those distributions.
2 Let's start with the animal-to-human distribution.
3 Now, we have a factor of ten. We say take the
4 animal NOEL dose, or better maybe a benchmark dose, and
5 divide it by ten to apply to humans. Why do we do that?
6 What do we think we've done by that?
7 In my quizzing people about this, I've come up
8 with three kinds of answers that people give of what they
9 think they have done when they have done that.
10 And this diagrammatically shows those three
11 answers.
12 What I plotted here on the log scale across the
13 bottom is the human-to-animal ratio of equally toxic doses,
14 so that a factor of .1 means that humans will have the same
15 response to a dose as a tenth as much as animals, and ten
16 means they can have ten times more of the dose, on the dose
17 scale that you're using here, before they have the same
18 response.
19 And then on the vertical axis is relative
20 frequency over chemicals. And there are some people who
21 say, well, I think that we do this factor of ten because
22 there's -- chemicals vary from one to another in what the
23 relative toxicity is. There's some that humans are more
24 sensitive to, some that they're less sensitive to. Maybe on
25 average in doses in milligram per kilogram per day are
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
22
1 equally sensitive.
2 But we want to allow for the fact that humans
3 might be particularly sensitive for this chemical, and so a
4 factor of ten, that UFA there, that I've drawn as sort of a
5 factor of ten down from the one-to-one line, represents
6 basically making sure out in the tail of that distribution.
7 Okay.
8 But obviously on average that's very conservative,
9 because if on average the right answer is that they're about
10 equally toxic on these dose scales, that sometimes humans
11 are actually less sensitive and yet we're assuming that
12 they're ten times more sensitive. You could be up to, say,
13 100-fold conservative on this. On the average you'd be
14 about 10-fold conservative or so.
15 The second idea is that, no, humans really are
16 more sensitive in milligram per kilogram per day basis than
17 animals. This is what we do in cancer risk assessment with
18 the so-called surface area scaling rule of the three-fourths
19 power of body weight scaling rule is now used at EPA, where
20 we're saying that at least on a milligram per kilogram per
21 day basis, humans would have a smaller does to get the same
22 response. There's a systematic difference that we're
23 actually having to do.
24 So some people say, well, from mice to humans on
25 that surface area scaling is about a factor of 13 or 12,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
23
1 from rats to humans is about a factor of six or seven, so
2 ten is sort of an approximation of that. We're making that
3 correction the way you do with surface area scaling of doses
4 to get equally toxic doses in animals and humans.
5 Under that interpretation you need that whole
6 factor of ten, and maybe it isn't quite enough, just to get
7 the equally toxic level in humans that you had in the
8 animals. It's not conservative at all, it's just getting
9 you to the dose that has the same effect.
10 And then you allow, well, there's still variation
11 in that from chemical to chemical. So sometimes maybe it
12 isn't as bad as all of that, this tail, but sometimes it's
13 maybe even worse. So in some fairly large fraction of cases
14 that factor isn't enough to allow for the animal-human
15 variability.
16 A third kind of conception of this factor of ten
17 is -- well, before I get on to B here, obviously the whole
18 distribution has just shifted over by a factor of ten.
19 That's the idea is that's where the center of it is.
20 The third idea is that in some sense you're doing
21 both A and B. You say, well, humans really are more
22 sensitive than animals, and chemicals really do vary from
23 one another to one another, but when you combine those, a
24 factor of ten is enough to make sure that you've cut off
25 most of the tail, so that the distribution is shifted over
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
24
1 and it has some spread, but ten is still considered enough
2 to cover both of those effects.
3 Okay. As I say, you talk to different people and
4 you have different ideas about what you thought you were
5 doing.
6 And a lot of the discussion about what to do with
7 the uncertainty factors, and some of the work that's been
8 done elsewhere, not at Harvard, but elsewhere, where you try
9 to turn the factor of ten into a log normal distribution for
10 which 10-fold is a 95 percent level, they're basically
11 trying to do something like this, except that they actually
12 draw it here starting at one and assuming that there are no
13 chemicals for which humans are more sensitive than animals,
14 and it's only a degree of how much less sensitive they are
15 with 95 percent of chemicals being 10-fold or less.
16 DR. BLANC: Did you say that backwards?
17 DR. GLANTZ: I'm totally lost.
18 DR. RHOMBERG: I may have said it backwards.
19 DR. BLANC: What you mean is that there's no
20 chemicals for which humans are not ten times as sensitive as
21 animals?
22 DR. RHOMBERG: Yeah. I think that's what I mean.
23 I wish I could really draw it.
24 DR. BLANC: We're not ten times more resistant
25 than animals or --
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
25
1 DR. RHOMBERG: Right. We're never more resistant
2 than animals. We're rarely more than ten times more
3 sensitive.
4 CHAIRMAN FROINES: Rarely more than ten times.
5 DR. RHOMBERG: Right. There are few chemicals.
6 If we assume that we're ten times more sensitive than
7 animals, we will cover most of the cases. That's what I was
8 trying to say. I don't know what I actually said.
9 DR. BLANC: That is what you said, it's just that
10 it's a double negative, and I'm not quite sure I understand
11 how that curve would be drawn, because you're assuming,
12 therefore, that everybody -- that 10-fold would be good
13 enough, but you're not saying that you could draw the
14 distribution the other way. What would that distribution
15 really look like? Is that one of those up there?
16 DR. RHOMBERG: I'm not sure I understand your
17 question.
18 You could draw it the other way out here and say
19 humans are usually more resistance on this dose scale.
20 DR. HATTIS: You're referring to Paul Price's
21 work.
22 DR. RHOMBERG: Right.
23 DR. HATTIS: Paul Price draws it so that
24 essentially it would look like it would start from here and
25 then trail off down here, so it would drop to zero here.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
26
1 DR. RHOMBERG: Right. There would be nothing on
2 this side, and there would be not very much on this side of
3 one, so it would be a distribution.
4 DR. BLANC: I see. Thank you.
5 DR. RHOMBERG: With all that, why was I saying all
6 of that?
7 DR. GLANTZ: Excuse me. You've like totally
8 confused me.
9 DR. RHOMBERG: Oh, dear.
10 DR. GLANTZ: It's fine.
11 DR. BLANC: He's often confused.
12 DR. GLANTZ: I'm often confused.
13 Anyway, could you just back up to B and just kind
14 of go over that again, and then C, and how they compare to
15 A, and make it like the dumbest possible terms.
16 DR. RHOMBERG: Oh, dear.
17 B is the idea that humans really are on average
18 10-fold more sensitive than animals. So you need that
19 10-fold because we're 10-fold more sensitive.
20 DR. HATTIS: The reason why in part is that we
21 process things slower than animals. So that if you've
22 gotten a given amount of milligram per kilogram in you, we
23 take a lot longer to get rid of it, more or less as the body
24 weight to the record.
25 DR. GLANTZ: But then why is this, if I'm like
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
27
1 boring everybody with these questions, stop me, but then why
2 does -- do you have the distribution dropping down about at
3 one? I mean, why are you putting the distribution where you
4 are?
5 DR. RHOMBERG: Well, I don't have a reason to put
6 it in any particular place, because I don't know -- this B
7 doesn't say anything about the width of that distribution,
8 but it just acknowledges that there is some width to it.
9 But basically in A you're correcting for, you're
10 allowing for the uncertainty. You're allowing for the
11 spread of the distribution. But you figure it's centered on
12 one.
13 B, you're looking for the centering. You say it's
14 centered on one-tenth, but you don't really say anything
15 about the spread of the distribution. But even though you
16 haven't said anything about it, it's still there, which
17 means that in some sense there will be some chemicals that
18 are yet worse than that and some chemicals that aren't as
19 bad as that.
20 DR. GLANTZ: Let me try and restate what I think
21 you're saying and correct me if I'm wrong.
22 In A what you're saying is the humans and the
23 animals are equally sensitive, and that's why you have the
24 distribution centered on one?
25 DR. RHOMBERG: Right.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
28
1 DR. WITSCHI: Something got lost, and this is the
2 scaling. The top one assumes on a mig per kilo basis, and
3 the second one the scaling should be done according to body
4 surface, which then, according to the fact you have in your
5 next thing, would bring it back to ten. So a dose of one
6 milligram per kilo in a mouse would correspond to a dose of
7 .1 milligram per kilo in man. That's if you scale by the
8 body surface.
9 DR. GLANTZ: Okay. But that's getting me even
10 more confused.
11 I just want to be one step at a time.
12 A, you're basically saying that the humans and
13 animals are on average equally sensitive.
14 DR. RHOMBERG: Right. On the dose scaling that
15 you have chosen to draw.
16 DR. GLANTZ: The thing on the horizontal axis is
17 the dose to get a given effect.
18 DR. RHOMBERG: Right.
19 DR. GLANTZ: So what you're saying in B then is
20 that the humans are ten times more sensitive than the
21 animals.
22 DR. RHOMBERG: Right.
23 DR. GLANTZ: Because one-tenth of the dose creates
24 the same effect.
25 DR. RHOMBERG: Right.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
29
1 CHAIRMAN FROINES: Peter is saying that's because
2 of inherent differences between the animal and the human.
3 DR. GLANTZ: Right. Well, that's another
4 question.
5 But then so then what's the point you're making
6 when you say there's an uncertainty?
7 Oh, I see. You're saying that uncertainty factor
8 of ten in B takes into account the fact that the
9 distribution is shifted like that?
10 DR. RHOMBERG: Right.
11 These are two fundamentally different ideas about
12 what you've accomplished with your factor of ten.
13 The first one is that you've allowed for the fact
14 that there are some compounds for which humans are
15 particularly sensitive and you want to make sure you catch
16 all those.
17 The second one is you say all chemicals on average
18 humans are 10-fold more sensitive and you want to correct
19 for that to get to the same point, basically the equally
20 toxic dose.
21 But there's moreover distribution around that that
22 you haven't taken into account when you've applied the
23 factor of ten. You've just dropped that issue from your
24 analysis.
25 DR. GLANTZ: So then what does C do then?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
30
1 DR. RHOMBERG: C says, well, we think that there's
2 some of both of those in there, that humans maybe really are
3 more sensitive than animals, but not by a factor of ten, by
4 something less, and there is variation from chemical to
5 chemical, so the distribution is shifted some, but that
6 nonetheless this tail is such that the factor of ten cuts
7 off most of it, that we've covered most of the chemicals in
8 a factor of ten.
9 DR. GLANTZ: But in B it wouldn't?
10 DR. RHOMBERG: Right.
11 Most people, and here when they are pressed to
12 something like C, but the problem with that is that then
13 that factor is partly a correction, a systematic correction
14 for extrapolation, the size of which we don't really have a
15 good handle on how big it is. And partly an allowance for
16 uncertainty.
17 And once we try to get empirical about this, we
18 will want to not just -- these are mental models in these
19 graphs. We want to get some real data to try to sort out
20 which of these things we're really doing, and what I'm
21 trying to point out is that we're sort of confused about
22 what we thought the thing meant in the first place, and
23 therefore what we are accomplishing by those things, and
24 therefore what kind of data would answer those as empirical
25 distributions.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
31
1 DR. GLANTZ: Just continuing disrupting you. So
2 in getting back to what Pete was saying, what you're saying
3 is the reason for going from A to B in this thing has to do
4 with the dose scaling, is that what you're saying?
5 DR. RHOMBERG: Yeah. I mean, one of the ways that
6 you can handle that is if you think there's that systematic
7 effect you try to take into account by dose scaling, and if
8 you have done that, and then plot these doses in the scaled
9 doses, then you're getting back to that.
10 So one way to handle that would be to scale the
11 doses and then apply the factor of ten. And then you get
12 back to the A conception.
13 I'll probably go over this really quickly here. I
14 just wanted to point out the different kind of dose scaling
15 methods that are usually used to point out that for cancer
16 we usually use surface area scaling in the past. EPA is now
17 talking about body weight to the three-fourths power
18 scaling, which is rather similar, but that noncancer we've
19 traditionally talked about body weight scaling, and is that
20 a good idea.
21 And I just wanted to show a few little things to
22 at least cast some doubt in your minds on whether that's a
23 good idea.
24 This is an understudied area about scaling and
25 noncancer.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
32
1 CHAIRMAN FROINES: Are you going to say anything
2 about the two-thirds to three-fourths?
3 DR. RHOMBERG: A little bit later, but I'm not
4 going to say lots.
5 CHAIRMAN FROINES: That's okay.
6 DR. RHOMBERG: There's a lot to talk about here.
7 CHAIRMAN FROINES: Dale, are you?
8 DR. HATTIS: Not much.
9 DR. RHOMBERG: I'll be happy to come back sometime
10 and talk about that, but I will talk about it some, unless
11 I've completely run out of time, which seems likely right
12 now.
13 On the top, this is work from Sandy Baird, and
14 it's plotting the fraction of the time that the various
15 kinds of experiments that you do in noncancer standard data
16 set end up being the most sensitive sex and species.
17 DC stands for dog chronic, RC, rat chronic, RR is
18 rat reproductive.
19 Basically when you have all of these studies in
20 dog, rats and mice and rabbits and so on like that, most of
21 the time the dogs are the most sensitive species. This
22 should be telling you something. The bigger the species is
23 the more often it's the most sensitive. Somehow that's
24 suggesting that maybe we're not doing the dose scaling
25 right. If larger species are systematically more sensitive
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
33
1 and the data sets we see and the animals, doesn't that
2 suggest that some sort of better scaling that allows for
3 larger sensitivity in bigger species might be in order.
4 If you use the surface area scaling in noncancer
5 the way it isn't traditionally done, and then redecide which
6 is the most sensitive sex and species, it's much more evenly
7 divided amongst the different experiments that you do.
8 CHAIRMAN FROINES: Lorenz, can you, Ray or Elinor,
9 can you folks in the back see this or do we need to dim the
10 lights? You can see?
11 FROM THE AUDIENCE: You can see okay. I can't
12 read the details.
13 CHAIRMAN FROINES: Peter, see if you can lower
14 them a little bit.
15 DR. RHOMBERG: I apologize, but in fact the
16 details, you know, I'm using this as an illustration as a
17 general point.
18 CHAIRMAN FROINES: That better?
19 FROM THE AUDIENCE: Yeah.
20 DR. RHOMBERG: I want to stand out of people's
21 way. Well, okay. Sorry.
22 The point is that the dog is usually the most
23 sensitive species, if you do scaling in the traditional
24 milligram per kilogram per day way, and if you do it by
25 surface area scaling, it's much more evenly divided amongst
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
34
1 the various species, suggesting that there's maybe some
2 systematic effect there that we haven't taken into account.
3 Another relevant piece of information is the paper
4 by Travis and White --
5 DR. GLANTZ: Is the surface area scaling got a
6 power in it? The two-thirds or --
7 DR. RHOMBERG: Yeah. Two-thirds power of body
8 weight.
9 The other piece of evidence is this paper by
10 Travis and White in 1988 in risk analysis, where they looked
11 at 23, I think, antineoplastic drugs. In these you have a
12 good measure of human toxicity, because you give these drugs
13 up to the level that are toxic in humans, clinically, as
14 part of treatment, and you can also get toxic levels in
15 animals from tests, and they basically -- this takes a
16 little bit of explanation to see what this graph is, they
17 actually made allometric regressions to see what power body
18 weight optimally scaled the animal toxic dose up to humans
19 and back and forth, among species, what equalized the toxic
20 doses among species the best, what power of body weight.
21 For each chemical they plotted that power, so
22 here's the power of body weight that is suggested by each
23 chemical and here are the chemicals.
24 You can see that, yeah, there's variation among
25 from chemical to chemical that we're talking about, but
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
35
1 they're varying around the line that's the .75 power of body
2 weight. I think the actual number here is .74.
3 DR. WITSCHI: Actually, I object to using this
4 paper as evidence for this, because the original paper who
5 brought this up was written in about 1963, by, among others,
6 Fryrick, and this is the first paper where the thing from
7 the body scale it comes. It's an old paper, about early
8 '60s. And those guys looked at all the toxicity of
9 anticancer drugs and the then available toxicity in mice.
10 They came up with the scale in fact of two-thirds. So
11 that's not new.
12 DR. RHOMBERG: Oh, no. First of all, this is
13 already 1988. It's already not new. But it wasn't new when
14 they did it, you're right. Pyrite did it back in the '60s.
15 They just did it again with more data, more
16 compounds.
17 And since then, there have been compounds added to
18 this.
19 The point I want to make here is that here is the
20 suggestion for noncancer toxicity. This is noncancer. It's
21 not acute, but it's sort of five- or seven-day dosing,
22 subacute or whatever you want to call it, that the
23 three-fourths power of body weight is the weight to scale
24 these, not milligram per kilogram.
25 So maybe our assumption that doses are equally
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
36
1 noncancer toxic in humans and animals in milligram per
2 kilogram per day needs to be revisited.
3 DR. WITSCHI: There's no reason why it's surface
4 is more appropriate, because surface is related to blood
5 flow and blood flow is related to metabolism.
6 DR. RHOMBERG: Right. I have lots of talks, I've
7 written big papers on this myself about why such scaling
8 ought to apply.
9 I guess I will just refer to all of those right
10 now.
11 I just want to acknowledge the fact that there is
12 a question here about how the best way to do that is.
13 And I want to very briefly treat this here.
14 Basically, what Travis and White were doing was using almost
15 an allometric equation, where you relate the unscaled
16 factor, the size of the dose in milligrams necessary to
17 produce a toxic effect, or milligrams per day actually, to
18 some power of body weight where W is body weight.
19 And basically it's easier if you take the log of
20 through the both sides of this equation and you get a
21 straight line on a log log plot, where the log of the
22 feature is equal to a constant plus the slope of the line
23 times the log of body weight.
24 So basically I want to address the question what
25 does scaling doses really mean.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
37
1 Here's an example of scaling doses badly. What I
2 mean by scaling badly, I'm saying we do one of those log log
3 plots here, here's the milligram per day, otherwise unscaled
4 by body weight, toxic dose, that's equally toxic, the
5 species A, B, C, D, E, F, G, plotted against the body mass
6 of those species.
7 And you can see that this line that you are saying
8 let's scale it this way is inappropriate in that it's
9 systematically overpredicting here and systematically
10 underpredicting here. So that the actual line is flatter
11 than that.
12 What if you invoked this as the way you want to
13 scale the doses, that's a dose scaling theory that you would
14 invoke by this line.
15 DR. GLANTZ: Where did that line come from?
16 DR. RHOMBERG: You have it from a theory. You say
17 I want to do milligram per kilogram per day scaling, because
18 I think that that's the right thing to do.
19 DR. GLANTZ: So that would be the line for
20 milligram per kilogram scaling, is that what you're saying?
21 DR. RHOMBERG: It's close to it. Yeah, well.
22 Yes, it is the milligram per kilogram scale. It's got a
23 slope of one.
24 The slope of that line will be the power of body
25 weight, so it's got a slope of one.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
38
1 So if that's --
2 DR. FUCALORO: Isn't that 1.2 to the 1.2 power? I
3 mean, I'm just reading. It doesn't look right to me.
4 DR. RHOMBERG: It is 1.2. I exaggerated this a
5 little to make it look good from the back of the room, okay.
6 But say somebody says I want to scale doses by the
7 1.2 power of body weight, because I think that's the right
8 thing to do and the point is that any such theory
9 corresponds to a statement about what the slope of this line
10 is. The slope of this line is now going to be 1.2. Okay.
11 And if you've chosen badly your scaling, then you
12 get this kind of systematic bias around things.
13 Moreover, if you look at the deviations off of
14 this line from A down to the line here and C down to the
15 line and G up to the line and so on, those deviations, those
16 residuals around the line are the variations from species to
17 species in what's the toxic dose.
18 And so what we're aiming at now is some sort of
19 characterization of how different a dose's toxicity is from
20 one species to another, because we want to take one of those
21 species and extrapolate to another, namely, humans.
22 The point is that a badly chosen line will have
23 big residuals and therefore it will have a big variation in
24 the equally toxic doses when scaled according to this power
25 of body mass that you're invoking here, the 1.2 power of
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
39
1 body mass.
2 Moreover, it will be a bias such that the small
3 species are at this end of the distribution and the big
4 species are at that end of the distribution, because these
5 residuals are on this side, and those residuals are down
6 under on that side.
7 So basically that shows that there will be a bias
8 in here, so the distribution is wider than is necessary when
9 you're scaling badly.
10 And also it introduces bias in that it's not just
11 going from any species to any other species. If you're
12 going from a small species to a big species, you get
13 different kind of span here than from any pair of species.
14 So if you -- basically what I'm saying is if you
15 make an empirical distribution of relative toxicity in, say,
16 mice and rats and mice and dogs, or whatever like that, and
17 you say we'll just use milligram per kilogram scaling for
18 now, on the assumption that that's the right thing to do,
19 the choice of doing that scaling is making that distribution
20 of relative toxicity amongst species wider than it needs to
21 be and it is ruining your assumption that you want that you
22 want to be able to draw from that randomly when you are
23 doing the Monte Carlo analysis.
24 So you really have to settle the scaling question
25 before you do this.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
40
1 Here's scaling well the same data and now I've fit
2 this line by regression, and it's an artificial example so
3 it comes out to be the .75 power of body weight. I say,
4 okay, now if we realize this, and of course for any
5 empirical set of data you might want to actually do the
6 regression, and see what's the best thing. It's got the
7 properties of regression, the residuals are minimized. And
8 assuming that this is really linear on this log log scale,
9 they will also be unbiased in terms of low here and high
10 there.
11 So you get a narrower distribution. In other
12 words, you're more certain about your cross-species
13 extrapolation, you're optimally certain about that for any
14 kind of scaling, and you're also unbiased, and any other
15 solution will be less optimal. You'll increase your
16 apparent uncertainty and will introduce bias.
17 So this is basically an argument why you have to
18 settle the cross-species scaling question before you really
19 grapple with even an empirical distribution of relative
20 toxicity across species.
21 Now, I've done some work on trying to do some of
22 this empirical stuff on one or two issues ago in risk
23 analysis on LD 50s, these acute single oral doses that are
24 lethal.
25 And I did lots and lots of species. This is just
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
41
1 showing rats and mice. And you get this ratio of the log,
2 and it shows that, yes, it does differ over species. There
3 are 4,600 and some chemicals in here. It's sort of peaky
4 compared to a real normal distribution, but in a way this is
5 a common log, so it's basically saying the factor of ten
6 either way is really pretty much covering this.
7 And in fact that works pretty well for other
8 species pairs as well.
9 You can argue that this now is evidence that sort
10 of concept A that I had earlier on the graph that we were
11 arguing about is applying for this kind of noncancer
12 toxicity, namely, single dose lethality.
13 It also suggests that these are doses expressed in
14 milligrams per kilogram that maybe this is evidence that
15 milligram per kilogram scaling is working for noncancer.
16 But you will notice that now there's a conflict in
17 my results on single oral lethal doses, LD 50s. I find
18 milligram per kilogram scaling working very well to predict
19 across species.
20 Travis and White found milligram for three-fourths
21 power to work well for the antineoplastic agents.
22 DR. BLANC: Yeah, but you're looking at a mouse to
23 a rat. They're really very similar in terms of their --
24 assume this area of the body weight times --
25 DR. RHOMBERG: I'm just showing that --
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
42
1 DR. BLANC: What happens when you do mouse from
2 monkey?
3 DR. RHOMBERG: Same thing. Although there are as
4 many data for monkey. But I did mouse, rat, monkey, dog,
5 hamster, pig.
6 DR. BLANC: How about mouse to dog, do you have
7 that?
8 DR. RHOMBERG: I don't have it in my -- but the
9 same pattern held across all pairs of species, no matter --
10 I did a regression of the size of the difference with the
11 size of the difference in body weight. If there's scaling
12 by other than the one power of body weight, you should get a
13 trend that the acceptability of this milligram per kilogram
14 scaling would get worse and worse, the more different the
15 body weight was between the species compared, and there was
16 no trend.
17 If you look at all the species together in a big
18 regression it comes out to be the 1.01 power of body weight.
19 This is in risk analysis. I'll let you read the
20 paper rather than go on about it. But I'm just showing this
21 one, because I have a lot of data from mouse to rats, but I
22 have it for ten species in all pair-wise comparisons.
23 DR. GLANTZ: Just, I just want to make sure I
24 understand this. So what you're doing is the thing, the
25 distribution is over different chemicals, right?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
43
1 DR. RHOMBERG: Yes.
2 DR. GLANTZ: You take chemical A and you find the
3 LD 50 in the rat and in the mouse and you repeat the ratio.
4 DR. RHOMBERG: Right.
5 DR. GLANTZ: And then you take the chemical B and
6 you do the same thing and those are --
7 DR. RHOMBERG: Right.
8 If you want to see that, I do have a quick extra
9 slide. Here are the actual data where I plot the LD 50 in
10 the mouse versus the LD 50 in the rat. You see, yeah, there
11 is variation, but a one-to-one line corresponding to
12 milligram per kilogram per day scaling works really well,
13 whereas the prediction of milligrams per three-fourths --
14 per kilogram per three-fourths per day, you know, clearly is
15 missing the middle of that cloud of points and is
16 systematically wrong, although there is variation around it
17 and there's the endpoints on either side of that line as
18 well.
19 It's only when you got a lot of data that you can
20 really tell things apart for species that are so similar in
21 body weight.
22 But I have such plots for all of my comparisons.
23 Well, I wanted to just call attention to this sort
24 of contradiction between this result from milligram per
25 kilogram on single oral lethal doses and Travis and White's
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
44
1 results of three-fourths power of body weight.
2 What are we going to do for scaling for noncancer?
3 Are we going to do it by milligram per kilogram or we going
4 to do it by three-fourths power? Why are those two things
5 different? Is one of us wrong?
6 And my suggestion, and this is sort of speculation
7 now, is that, no, we're both right, but there's a
8 fundamental difference between single dose scaling and
9 repeated dose scaling.
10 And I want to talk about why I think that might be
11 here briefly.
12 And if you don't follow this, I'm sorry, because
13 it gets a little bit complicated. You have to sort of have
14 thought about this a while to sort of see what I'm getting
15 at.
16 This is a simplified hypothetical model, but it's
17 a real model and that there are real equations underlying it
18 that will have the properties that I'm describing.
19 Basically what I'm suggesting is that for a really
20 severe single dose endpoint like acute one dose lethality,
21 the real issue is how much reserve in some sense does the
22 body have to fight off this onslaught? And what if it seems
23 reasonable to suppose that the reserves that you have are
24 proportional to your body size. It's sort of your standing
25 level of capacity or ability to fight something off or
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
45
1 absorb something without harm.
2 So basically --
3 DR. HATTIS: Or just to dilute it.
4 DR. RHOMBERG: Or just to dilute it, right,
5 exactly, to the point where the concentration is below that
6 which is going to cause a harmful effect.
7 If that's the case, then a human and a mouse
8 should have about the same milligram per kilogram size to
9 the volume that they're diluting it by dose for a single
10 lethal dose, even though the human, because of the slower
11 physiological approaches of humans, will do that slower,
12 absorbs it slower, and then there's recovery as this
13 depleted whatever it is or damage starts to repair itself.
14 Now, I built in a repair or recovery function
15 that's proportional to the loss from the control level in
16 here.
17 So the process takes a little longer in time, but
18 nonetheless how far you dip into your well of reserves is
19 the same depth, and in some sense the same milligram per
20 kilogram per dose should be equally toxic.
21 If you then scale by the three-fourths power of
22 body weight, the human could get off easier, because it has
23 to dip in that well less deeply.
24 DR. FUCALORO: Excuse me. I'm missing it. The
25 concentration is the concentration of what?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
46
1 DR. RHOMBERG: Compounds somewhere in the body.
2 DR. FUCALORO: Somewhere in the body.
3 At time zero tell me what happens.
4 DR. RHOMBERG: At time zero you say gavage the
5 animals or you give the single oral dose to humans, however
6 you do it, I guess you wouldn't gavage, but something
7 equivalent, and it's absorbed. And as it's absorbed it's
8 now depleting something or causing some damage. I'm just
9 called -- I'm sorry, I should have pointed out, I'm talking
10 about concentration of some substance in the body that is
11 depleted with interaction with --
12 DR. FUCALORO: Because it looks like just the
13 opposite. Not the concentration of the toxin, but some
14 concentration of a important chemical the body needs that is
15 depleted and then the body recovers.
16 DR. RHOMBERG: Yes. I'm sorry. My apologies.
17 I'm trying to rush here and I'm leaving out details. And
18 that's important.
19 The model is something like glutathione depletion
20 here, but it doesn't have to be that literally. It can be
21 any kind of damage where going down is sort of damaging,
22 reducing some level of something that you need that
23 recovers.
24 DR. WITSCHI: A model is nice, but we have lots of
25 what happens after an LD 50 in mice. But I would wonder how
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
47
1 many data we have for after LD 50 in man. How many
2 experiments do we have? And now we're talking about some
3 fundamental differences because the most trivial one of
4 those is dose rate, and you only can compare what happens
5 after an LD 50 in mouse, what happens in man if you have
6 some real data, but you do not have.
7 DR. RHOMBERG: Well, there are some real data for
8 Travis and White for the antineoplastic agents. They aren't
9 single doses. They are multiple doses.
10 In fact, the upshot of my previous analysis on LD
11 50s, for which there are like four human LD 50s in this
12 database, suggests the sort of milligram per kilogram
13 scaling and that seems to be in conflict with these results
14 for multiple dose things on antineoplastic agents that
15 Travis and White has.
16 So what I'm trying to do is a conceptual model,
17 which I admit is conceptual and doesn't have data to say
18 what could be going on here to reconcile those two empirical
19 results. The empirical results both exist.
20 DR. GLANTZ: Now, would your graph of the human,
21 if you carried the time scale out, it would look like the
22 mouse, just slower?
23 DR. RHOMBERG: Yes. It would be just seven times
24 slower on the time scale. And if you do that trick of
25 dividing time by the fourth power of body weight, those two
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
48
1 curves would look the same, for those familiar with that.
2 I just want to contrast that then with a repeated
3 dose scenario.
4 DR. FRIEDMAN: Before you leave that, I'm not sure
5 I understand. Are you saying that this shows that the
6 milligrams per kilogram scaling works? This is a situation
7 where it works?
8 DR. RHOMBERG: It would work, yes.
9 DR. FRIEDMAN: Why does it work if the mouse
10 recovers so much quicker and the human recovers so much
11 shorter?
12 DR. RHOMBERG: This is on the hypothesis that it's
13 somehow that the depth of depletion of this depleted
14 substance or the amount of damage that's caused is what the
15 result of lethality.
16 DR. FRIEDMAN: The distance that it drops down is
17 the key point?
18 DR. RHOMBERG: Right. So if you drop down to 50
19 percent of your background level of this glutathione, or
20 whatever like that, you have a toxic reaction. If you don't
21 dip that far down, you don't.
22 I admit, it's hypothetical, but, as I say, I'm
23 trying to come up with a construct that would explain these
24 results and then it becomes a hypothesis to follow.
25 And so the subsequent part of the hypothesis, what
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
49
1 happens on multiple dosing and the idea here is that in
2 multiple dosing the issue is fundamentally different, it's
3 not your reserve capacity that you carry around with you
4 every day ready for whenever it may be challenged, but
5 rather your ability to recover. In some sense what you have
6 is with repeated dosing some damage every day, and then in
7 some sense recovery every night if you want to sort of split
8 it into sequential parts. Obviously, they're going on
9 simultaneously.
10 So then the question is if humans are slower to
11 recover, why would that be, because of the slower
12 physiological processes and larger species that draws out
13 their time scale vis-a-vis rats and mice.
14 Then a daily dose of a certain amount in a mouse
15 that it can just recover from every day, so that it just
16 barely avoids dipping down or below that level that it's
17 toxic or just barely grazing that level of depletion that
18 can't be sustained.
19 The daily dose has to be in proportion to that
20 recovery rate, and humans with their slower recovery will
21 have the same amount of depletion every day and less
22 recovery every night, are going to have a bigger effect.
23 DR. FUCALORO: We'd have difficulty, really, I
24 mean, I can see what you're drawing there clearly, but these
25 results from your model, your mathematical model, how you
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
50
1 have your equations for depletion and recovery; correct?
2 DR. RHOMBERG: Yes.
3 DR. FUCALORO: And we haven't seen those. I
4 suppose you have some somewhere.
5 DR. RHOMBERG: Yeah, but they're hypothetical.
6 DR. FUCALORO: I know. And you can change those
7 models. I don't know if they're kinetic, first order
8 kinetic types or all sorts of types of modeling for
9 depletion.
10 DR. RHOMBERG: It's first order absorption, first
11 order reaction to the compound with this endogenous
12 compounds being depleted.
13 DR. FUCALORO: That's what I was thinking.
14 DR. RHOMBERG: It's got a recovery function that's
15 proportional to the amount of depletion.
16 DR. FUCALORO: Right.
17 DR. RHOMBERG: But that operates more slowly, and
18 all these things are operating more slowly in humans than in
19 rats.
20 DR. FUCALORO: I guess what I'm saying is that by
21 showing the difference between the three-quarter scaling, I
22 mean, but by using different rate constants and so on, you
23 can get almost anything, any types of curves.
24 DR. RHOMBERG: I suppose that's true, but two
25 things.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
51
1 One is this is referring back to the empirical
2 results, the Travis and White, and the stuff on the LD 50s,
3 that need to be reconciled somehow or other, so it's not
4 just that I'm making this up. There's a phenomenon that
5 needs to be explained. If this isn't the explanation, then
6 we need another one.
7 CHAIRMAN FROINES: Let me interject here.
8 I'm a little worried about time, because we've
9 been going an hour now. And we had thought this would be a
10 little briefer.
11 So would the panel hold their questions and let
12 him finish, and then we can have questions.
13 Go ahead, Gary.
14 DR. FRIEDMAN: My only point is if we don't
15 understand the point --
16 CHAIRMAN FROINES: Absolutely.
17 DR. FRIEDMAN: Then I feel it's important to
18 interrupt, just to get some quick clarification.
19 CHAIRMAN FROINES: If there's a point of
20 clarification, but otherwise I think it would be easier if
21 he could finish a conceptual.
22 DR. RHOMBERG: Okay.
23 DR. HATTIS: Let me help.
24 CHAIRMAN FROINES: No, no, no, Dale. Sit down.
25 Sit down. You'll get your turn.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
52
1 DR. RHOMBERG: Okay. I just want to, if I could,
2 just quickly follow through with this argument, and I hope
3 that you will see the point.
4 The point is that if you scale the doses by the
5 three-fourths power of body weight for each day's dose, and
6 in some sense the recovery process is scaling that way as
7 well, then you're keeping the recovery ability in balance
8 with the damage that's caused every day.
9 So in some sense you have this contrast. With a
10 single dose the issue I'm proposing with a single dose the
11 issue is how deeply do you dip into your well of reserves.
12 And the well is about the same depth in humans as it is in
13 animals, and so you get a milligram per kilogram dose to do
14 that equally.
15 Over the long term, the issue is how much on the
16 well analogy, how much do you draw out of the well every day
17 compared to the rate at which stuff is seeping back in to
18 replenish the water that was drawn away. And I'm saying
19 that in humans that seeping back, that recovery,
20 replenishment, is slower systematically and therefore the
21 amount of stuff that you can draw out every day would be
22 smaller as well.
23 So if that's something like that were true, that
24 would explain this result.
25 Moreover, the things that I rely on in making the
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
53
1 equations for these models are patterns of differences
2 across species that people think exist. It's not that I'm
3 making them up in order to make this phenomenon happen. I'm
4 making the rates scale, the three-fourths power body weight
5 of which there's tons and tons of data to suggest that's a
6 reasonable way to approach this problem.
7 So let's just leave it at that. I'm trying to
8 just sort of suggest a reason for this, but I don't want to
9 detract from the original message, which was we somehow have
10 to decide how to do the scaling for noncancer before we can
11 get this empirical distribution across species, and so we
12 have to grapple with this issue, and it's not an easy one,
13 and the answer for single doses might be different than the
14 answer for multiple doses.
15 So that was the main point.
16 Oh, boy. I'm going to skip over this one
17 entirely.
18 I want to briefly talk about the sensitivity among
19 humans and the question of dose response and this will be a
20 good lead-in to Dale, because that's what he's going to be
21 talking about mainly.
22 All this was just about the animal-to-human
23 factor, but another one of those factors that we always have
24 is the question about average humans to sensitive humans or
25 something like that.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
54
1 And I just want to point out that we have this
2 notion here that the whole reason why there's a dose
3 response is because there's variation in sensitivity or in
4 capacity in some sense among the population.
5 So here I'm basically just showing this
6 diagrammatically. You have a population of varying reserve
7 capacity that is now sort of eaten away to some degree by
8 exposure to a compound and it's eaten away more by bigger
9 exposures than smaller exposures, so that when you draw
10 samples from that distribution to test at various doses, at
11 higher doses you've exceeded the capacity of a larger and
12 larger fraction of the population, and that's why you get a
13 gradual dose response curve in terms of percent responding,
14 rather than a step function which says doses below this are
15 okay and doses above this are not, even if there's a
16 threshold for individual behavior. And we sort of have this
17 model for what's going on underneath here.
18 The point I want to make is that if you remember
19 way back when I started I said one of the things we need is
20 a risk measure. You have to be able to do dose response
21 analysis for noncancer and not just safety analysis.
22 We also want to get a distribution of sensitivity
23 among humans for the factor UFH. And I just want to point
24 out in some sense those are the same things, that on the
25 assumption that this kind of process is responsible for the
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
55
1 dose response, that somehow looking at that process, tracing
2 out the shape of the distribution, because, after all, this
3 curve is basically the cumulative distribution of the
4 distribution of individual reserve capacities here.
5 Basically, in doing one, you're doing the other.
6 And that's why I think it would be good to take
7 that factor, and rather than just building it into this
8 equation as another distributed uncertainty factor, to look
9 at it explicitly as a dose response function and look at
10 empirical data on sensitivity among humans, of which Dale
11 has a lot, as a way of looking at the shape of the dose
12 response curve in humans.
13 The question then is there's a lot of questions
14 about how do you figure the properties of that dose response
15 curve, its location and so on like that from the animal
16 data. That's part of the development that we're still
17 working on in this question.
18 One of the things you can do is to say, well, we
19 will just use the dose response curve in animals, but one of
20 the difficulties there is that, you know, we know that
21 humans are more variable amongst themselves than animals.
22 They're genetically heterogenous, they have different
23 lifestyles, young and old, other exposures and so on.
24 And so basically the slope of the dose response
25 curve we get in animals ought to be steeper than the one we
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
56
1 get in human, because it's reflecting a narrower
2 distribution of variability, and so in some sense we have to
3 decide what are we going to do about the fact that sort of
4 for a given center of a dose response curve, it's going to
5 go down more slowly in animals -- in humans than we observed
6 in animals.
7 And one of the answers to that is to look at the
8 human data directly, which Dale will talk about.
9 Well, this took longer than I thought, but I hope
10 that I've been able to get across some of the ideas here.
11 The point is that you can imagine approaching this
12 problem by empirical distributions of these uncertainty
13 factors, but once you start to look at the individual
14 factors and look at what data you would apply to them, some
15 of these other kinds of problems about dose scaling and like
16 that pop up, so it shows it to be a complicated problem, but
17 nonetheless I think getting back to his original motivations
18 is one worth pursuing for the benefits that we would have to
19 make noncancer risk assessment be able to better answer the
20 tasks that we're trying to have it do for us.
21 Thanks.
22 I would say any questions, but there have been
23 plenty of questions already.
24 DR. WITSCHI: Just what's the discrepancy, you
25 said, between the Travis data and your data? They scale on
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
57
1 a body surface basis and you find better congruence by
2 scaling on a weight basis?
3 DR. RHOMBERG: Yes.
4 DR. WITSCHI: There's one thing you note and you
5 mentioned this about having research and so on, but the
6 toxicity that comes from cancer drugs, they're really going
7 to people who are already sick.
8 DR. RHOMBERG: Yes.
9 DR. WITSCHI: These are not normal people,
10 otherwise they wouldn't have gotten the cancer drugs, so
11 they cannot necessarily be compared to healthy animals in
12 which you do the LD 50s.
13 DR. RHOMBERG: That's a good point.
14 In fact, the measure of toxicity is hard to make
15 exactly congruent.
16 You know, there's a series of discussions on that
17 point in the Travis and White paper, and one of the things
18 they bring up is that, well, these are animals in their
19 cages without any kind of treatment, and the humans,
20 although they were sick, were also in a clinical setting
21 where they were being monitored very carefully. In some
22 sense those are offsetting kinds of biases.
23 But it's true that it's hard to make any of these
24 comparisons really strictly.
25 I would only point out that it wasn't simply mice
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
58
1 and humans in the Travis and White analysis, but they also
2 had other species as well that were done more or less on a
3 comparable basis, and humans were the only ones that were
4 actually in a clinical setting.
5 DR. BYUS: Does it work for glutathione? I mean,
6 you brought up that example. It would be a lot more
7 comforting to me if you had an example like glutathione
8 mediated toxicity to show that this scale is better.
9 DR. RHOMBERG: There are some data on glutathione
10 depletion time courses in animals.
11 DR. BYUS: Are they faster at recovering than
12 humans?
13 DR. RHOMBERG: I'm not sure. I haven't found any
14 data yet on humans for this, but I'm sure there must be.
15 DR. BYUS: There must be plenty.
16 DR. RHOMBERG: Right.
17 DR. BYUS: It would be more comforting to me if --
18 DR. RHOMBERG: If you had a real example.
19 DR. BYUS: You had an example. And you brought
20 that up, that's why I am asking.
21 DR. RHOMBERG: Right. I made the thing as a sort
22 of general argument, and not simply about glutathione, but
23 glutathione is something that you sort of think as operating
24 sort of like my general argument.
25 DR. BYUS: But still, a mechanism that's known and
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
59
1 there are a lot of mechanisms known for a lot of drugs and a
2 lot of toxic agents, if you took, somebody did, work through
3 that mechanism as opposed to a different mechanism, maybe
4 that may be the answer.
5 DR. RHOMBERG: Right. I haven't published that
6 model yet, because I don't have a good example to stick it
7 with, and just as a model it does seem arbitrary. Its main
8 attraction to me is that it explains this phenomenon and
9 it's sort of in accord with what people sort of think is
10 going on. So it's a fruitful hypothesis, but at this point
11 that's what it is.
12 CHAIRMAN FROINES: If there are no questions from
13 the panel, let's ask DPR and OEHHA and others if they want
14 to make comments.
15 DR. MARTY: I just had one question.
16 This is Melanie Marty from OEHHA.
17 Do you have these data sets where you can look at
18 nonlethal endpoints? The lethality part bothered me,
19 because there's obviously no recovery from the lethal
20 effects. That might be more useful to look at a nonlethal
21 single dose endpoint to try to compare with the more
22 subchronic or chronic exposures.
23 You know, and also obviously the chemotherapy was
24 not meant to be lethal. It may have been, but they try not
25 to kill you with the drug.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
60
1 DR. RHOMBERG: But it was very severe toxicity.
2 It's a good point. It would be interesting to
3 look at things of different severity and also at chronic as
4 well as the short-term things. When we're talking about the
5 repeated dose studies, it's really only four or five days,
6 it's not a lifetime or anything like that.
7 The difficulty is finding data that are taken
8 systematically enough that you can actually generate these
9 databases that you need to do this kind of analysis.
10 The reason I did lethality is because there were
11 tons of data on LD 50s all organized and luckily LD 50,
12 lethality means the same thing in most labs, whereas
13 hepatotoxicity might not, and so it's really hard to find
14 data that are really comparable in different species.
15 Even though we have a test system that we use a
16 lot of the time, it's hard to find it in different species.
17 We've got a system that we run in mice and that we don't do
18 the equivalent endpoint in rats or in rabbits, because we're
19 already doing it in mice. So finding the data to answer
20 this kind of thing is a real challenge and sometimes you can
21 say in the whole world of the literature, you ought to be
22 able to pull out those things and organize them, but it's
23 been a task that's bigger than I've had funding, let's say,
24 to try to accomplish.
25 CHAIRMAN FROINES: George, last question.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
61
1 DR. ALEXEEFF: George Alexeeff with OEHHA.
2 I think there's a couple of things.
3 I think you tried to discuss some of the issues
4 regarding the uncertainty factors that you think can be
5 better quantitated, because I think that's kind of the areas
6 you're focusing on, because there's lots of other areas that
7 we use those uncertainty factors for, that maybe they can be
8 quantitated.
9 Like, for example, the severity, the effect that
10 we see in the animal versus what we might see in the human,
11 or the effect that we see in the animal and the effect we
12 might see in the human, so part of the safety factor is for
13 that, because we're looking at maybe a severe effect in the
14 animal, and we're thinking maybe the humans might be
15 responding in less severe.
16 Or the fact that study design of the animal study
17 is just really poorly designed and we're trying to make up
18 for the fact that we're just not really that confident in
19 the actual data set because they're not as well structured
20 as the cancer studies where you have large groups of
21 animals.
22 And the other thing I thought was that those
23 distributions that you showed where you had the Baird paper,
24 I know she called them adjustment factors or something like
25 that, but really to me I just don't see whether or not
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
62
1 uncertainty distributions, because really if they're not
2 based on actual data, if they're basically modeled
3 distributions for that particular uncertainty factor, then
4 really they're just uncertainty distributions.
5 DR. RHOMBERG: I wouldn't disagree with -- taking
6 your points in reverse order -- I wouldn't disagree. I
7 think that might have been a better term for it.
8 To the first point, you're right, there are other
9 uncertainties that aren't covered here and it's one of the
10 challenges. As I say, you've got these factors that you
11 have vague notions about what you think you've accomplished
12 with them, and everything that needs accomplishing, somebody
13 is sort of saying, well, some part of that factor of ten is
14 to account for this, and when you try to pars these things
15 out and apply data to some of them and realize that others
16 are yet unaddressed, that is when it gets challenging, but
17 the only solution is to really think about what it was that
18 you're trying to do with those factors, what you think they
19 represent, what's in there and what's not, so that you can
20 disassociate them later on.
21 Good point.
22 CHAIRMAN FROINES: Okay. We're going to take a
23 five-minute break before Dale starts, to give the
24 stenographer a chance to rest her wrists. So let's -- but
25 it is going to be very brief.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
63
1 (Thereupon a short recess was taken.)
2 CHAIRMAN FROINES: Can we get started again.
3 My strategy when we're not reviewing documents is
4 to try and get the meeting finished before lunch.
5 And when we're reviewing documents, clearly that
6 takes us longer, often, especially when we're doing more
7 than one.
8 But when we're having a meeting like this, which
9 is more catch-up, if we can push ahead we can finish in a
10 reasonable time so everyone can leave.
11 And our next speaker is Dr. Dale Hattis, who is at
12 Clark University, and the panel knows Dale from the diesel
13 workshop when he presented the results of his work on
14 estimating risk for diesel exhaust.
15 DR. HATTIS: I'm going to try to follow on to
16 Lorenz's talk with -- Lorenz gave you a good deal of the
17 problem and I'm going to give you a straw man, quote,
18 solution, unquote, and how you approximately you get there,
19 although I don't want to be -- I want to be clear that I'm
20 not going to be presenting you with tests of the solution,
21 but basically a concept of how we tell whether our NOEL
22 uncertainty factor procedures or something else that we can
23 come up with are doing what we could reasonably hope that
24 they would do.
25 And basically the idea behind that is by making
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
64
1 some kind of a quantitative specification for the RfD, if
2 only to provide a benchmark for trying to assess whether
3 it's working or not.
4 So basically this is what the title, which is
5 slightly different than what it appears in the program, a
6 quantitative definition, and my basic hope, this is
7 basically the outline of the talk.
8 First, we'll talk a little bit about what we have
9 to gain by this kind of thing. Lorenz has already talked
10 about that to a degree. The hope is that the hundredth
11 anniversary of the Lehman and Fitzhugh paper is rapidly
12 approaching in the year 2054, and we need to build, with the
13 rate at which these things do change, we do need to start
14 now to prepare ourselves with some sort of a replacement, so
15 that we'll be ready when that day comes.
16 I will then talk to you a bit about the
17 difficulties and costs of trying to make a quantitative
18 definition of the RfD.
19 And then I'll specify the elements of this straw
20 man and I'm going to suggest to you is a starting point.
21 And then the requirements for a viable system,
22 eventually.
23 So among the benefits are, first, that we stop the
24 lying. Basically the implication is made with the current
25 RfD approach that we are talking about population thresholds
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
65
1 and a dose that we expect will have no effect in a diverse
2 human population.
3 And I think that there is some hope that that's
4 right some of the time, but I think that as a general
5 manner, I don't believe in population thresholds. I think
6 that there is good reason some of the time to believe in
7 individual thresholds, because the general mechanism that
8 people have for these kind of traditional toxic effects is
9 the overwhelming of homeostatic systems, and that it follows
10 that every individual might have a different capacity for
11 absorbing and counteracting a small perturbation of
12 homeostatically controlled processes so that out of that
13 comes the idea that individuals have different amounts of
14 reserve capacity, and different people therefore might get
15 the effect or not when their individual reserve capacity is
16 depleted for counteracting a particular kind of perturbation
17 of one physiological parameter or another.
18 But if we have a broad population distribution of
19 reserve capacities including some people who are very
20 marginal or submarginal for a particular things, like my
21 87-year-old father had an episode of congestive heart
22 failure about a week ago, his heart wasn't pumping enough
23 blood to keep his lungs clear of fluid.
24 If you exacerbate that kind of problem in his
25 cardiovascular system a bit, he's going to be marginally
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
66
1 worse off than he was in the absence of that perturbation.
2 So with a diverse population with different
3 amounts of reserve capacity, and in some sense subminimal
4 for the excellent function, I think that one should believe
5 that there are effects.
6 And by trying to specify what these distributions
7 are in much more empirical ways, we have a chance of
8 approaching what the real truth is likely to be.
9 The second is the idea is to reconcile what we do
10 in noncancer assessment with to some extent with cancer
11 assessment, rather opposite to the way that EPA is now
12 wanting to do it. EPA wants to merge noncancerous -- some
13 parts of cancer risk assessment into noncancer ones. I
14 think that we ought to make the noncancer ones more
15 quantitative, like the cancer potency assessment.
16 Provide a basis to quantitively assess risk for
17 input to policy decisions. Lorenz mentioned this already.
18 Partly for those times where you want to juxtapose the cost
19 and benefits of policies to control exposures, and, second,
20 to facilitate judgments of the equity and fairness of the
21 burden of health risks and benefits potentially imposed on
22 different subgroups. You can only do that if you talk in
23 some more quantitative terms about, well, how large do you
24 think that burden is likely to be, how often.
25 Finally, this would allow some comparable analyses
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
67
1 of uncertainties among exposure and toxic potency,
2 potentially leading to some value of information analyses
3 that allows you to have a chance of directing your resources
4 to the places where you have in fact larger uncertainties,
5 and those that might change your decision making.
6 There are serious disadvantages and cost to trying
7 to go this direction.
8 Among those are there some need for the experts to
9 assess and publicly defend past choices of acceptable
10 intakes and risks, and I can just see that you spent 20
11 years developing a series of regulatory actions based upon
12 good faith analyses of data, and now we're going to revisit
13 that whole 20 years' worth of work. That's a cost.
14 Second, there's some difficulty of social
15 acceptance of finite risks. There's a need for explicit
16 decision making on uncomfortable tradeoffs that is not posed
17 by the NOEL safety factor procedure in the same way, that
18 the NOEL safety factor pretends at least that you have no
19 risk, and I just think it's not -- at least sometimes not
20 likely to be right, but nevertheless it's a cost to face
21 those tradeoffs squarely.
22 Third, of course, numerical expressions will lead
23 some to imagine that the estimates of risk are more precise
24 than they are. These estimates of risk that we can now
25 produce, or are likely to be able to produce in the next
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
68
1 decade or so, are going to be highly uncertain. We need to
2 be able to represent that uncertainty in a fair way to our
3 audience.
4 Fourth, there is going to be significant
5 controversy over some technical details that are important
6 technical details. What sorts of distributions of human
7 susceptibility should we assume based upon our limited data?
8 Okay. We will not have specific measurements in
9 as many as ten to the fifth people. We're going to have to
10 use data on some more limited sample of those to be able to
11 project to a larger universe. We need to have rules about
12 how to do that. There's no way to avoid that.
13 The elements of this proposal are tentatively, I
14 would suggest, that the RfD for cases where you want an RfD,
15 that is a dose that you're going to tentatively accept
16 without thinking about benefits and other things, as the
17 more restrictive value of either the daily dose rate that's
18 expected with 95 percent confidence to produce less than one
19 in 100,000 excess risk incidence over background, okay. So
20 in increments of background of a minimally adverse response
21 in a standard general population.
22 So the A relates to a standard general population,
23 ten to the minus fifth risk, which resembles what's been
24 thought of for California as the standard for notification
25 for a much more serious outcome that is cancer, so this
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
69
1 would be rather more protective in the sense that we're
2 talking about not cancer, about a minimally adverse
3 response. Okay.
4 B, to take into account the possibility that there
5 might be some recognizable minority of people with a lot
6 more sensitivity. In those cases, basically, we want to say
7 there might be a daily dose that's expected, again with 95
8 percent confidence, to produce less than a one in 1,000
9 excess risk incidence, excess incidence over background of a
10 minimally adverse response in this definable sensitive
11 subpopulation.
12 So we don't want to, even if there's a one in
13 10,000 group of people out there that we know about to be
14 highly sensitive, we don't want to cause them to have,
15 without thinking about it very seriously, to have a
16 particularly large burden. And basically you can play
17 around with these numbers as you like to reflect the social
18 consensus, but I think this is a starting point for
19 calculations, okay, to be made that I think is not grossly
20 inconsistent with what you might think of as policies, as
21 the precedence from the cancer side.
22 Now, how would we actually do this? Okay.
23 First, let's say we're starting with animal data,
24 which is often the case. Start with people data, that
25 avoids some of the problems, but not all of them.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
70
1 First, we would --
2 DR. GLANTZ: I just had one quick question.
3 DR. HATTIS: Yeah, sure.
4 DR. GLANTZ: Why did you pick 100,000 in a
5 thousand?
6 DR. HATTIS: I picked 100,000 because of the
7 analogy with Prop 65 standards. That's the one number that
8 I'm aware of in California law, and it seemed like it wasn't
9 as extreme as one in a million, and basically that's --
10 well, that's the answer.
11 And why one in a thousand, well, I didn't want to
12 get too high, because if you get started up to one percent,
13 then you start to look at, say, okay, you start looking in
14 the faces, 1,000 is still not too many -- well, it's my
15 guess, and I would defer to our elected representatives the
16 final policy choice of what the number ought to be.
17 But that was my guess as to what would be a
18 sensible recognition of the possibility of minority
19 subpopulations.
20 DR. BLANC: But by simple long division, if I
21 understand correctly, even if the standard did protect one
22 in 100,000, and if the people in the sensitive population
23 were a hundred times more responsive than the people in the
24 general population, then the same standard would still
25 protect them at a one in a thousand.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
71
1 DR. HATTIS: If they were only one in a thousand
2 of the population, if the sensitive population were no more
3 common than one in a thousand in the population. That's why
4 I said that if that second requirement under this tentative
5 scheme would only kick in if you have a substantially rarer
6 group that was -- see, if you have a group that's as common
7 as one in hundred --
8 DR. BLANC: No, they could still be as common as
9 one in a hundred, couldn't they, if -- and be a hundred
10 times more sensitivity and wouldn't you still be protecting
11 them --
12 DR. HATTIS: No. If it were as common as one in a
13 hundred, which is ten to the minus two, and they were a
14 hundred times more sensitive, then essentially that would
15 put us in general population risk of one to ten to the
16 fourth, which would break the general population standard.
17 So they have to be both rare and extraordinarily
18 more sensitive in order for that --
19 DR. BLANC: And you could balance them -- you
20 could make your own calculations if they were a thousand
21 times more responsive, but they were only --
22 DR. HATTIS: Sure. Right.
23 So most of the time the general population
24 standard is going to govern, but this second case was put in
25 there just so that in case we have something, someone that's
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
72
1 really very -- quite a bit more sensitive and very rare,
2 that we could take that into account.
3 CHAIRMAN FROINES: The ten to the minus fifth
4 value standard was first defined by Governor Deukmejian in
5 the mid '80s, and so that was essentially a policy decision.
6 There's nothing -- it's not cast in stone.
7 DR. HATTIS: Moreover, for technical purposes it
8 would be easier if the standard were higher, but I don't
9 think that we want to make that -- we don't want to choose
10 technical convenience over public health policy, even though
11 the ten to the minus fifth makes things inconvenient for us.
12 Let's start with some putative animal data on a
13 particular kind of response. One of the things, obviously,
14 that gets done under both the existing procedure and my
15 suggestion is that you need to define some candidates
16 anyhow, for what the critical effect might be and the
17 critical studies that show that effect.
18 And I would depart from the standard procedure at
19 this stage a little bit in that I would want to keep in the
20 analysis multiple candidates if they have a reasonable
21 chance of being significant contributors to an overall risk
22 or effect. So it's conceivable that you could have
23 multiple effects that would sum up to an overall risk that
24 you would care about.
25 Second, for each effect I want to choose a dose
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
73
1 response family for projection, and there are two broad
2 categories of these families.
3 One is the one we've been talking about
4 implicitly. The homeostatic system overwhelming idea that
5 leads to individual thresholds for response, but where the
6 responses of a mixed population give you a population dose
7 response that tells you something about the distribution of
8 thresholds in that population. Okay. That's one category.
9 But there is maybe another category of mechanism
10 that you could use where, in fact, what you're doing is
11 perturbing some parameter in animals that has some analogy
12 with some human parameter for which you have good human data
13 on its individual significance or its group significance for
14 ultimate risk.
15 Examples of this are, for example, sperm counts as
16 a mediator of male fertility reduction or other sperm
17 quality parameters.
18 We have all kinds of data in animals showing
19 testicular atrophy and the point at which the dose response
20 in the animals for reductions of sperm determine.
21 I wouldn't then want to go and assume that the
22 fertility response in people mirrors the fertility response
23 in animals, because the physiology and the reproductive
24 physiology are very different. And moreover we have good
25 human data, at least we have some human data, on how often
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
74
1 people with different amounts of sperm counts with standard
2 sets of partners succeed in achieving conception per month,
3 and then basically we can use those data.
4 So once we have the sperm count effect modeled in
5 the animals, we would then want to go to the people to
6 interpret the significance of that.
7 There's several other examples of that as a
8 possibility of fetal growth inhibition as a predictor of
9 birth weight reduction and strong associations between birth
10 weights and infant mortality.
11 For example, this upper curve here is a curve that
12 shows the relationship between birth weights and infant
13 mortality. That's the probability that the baby dies in the
14 first year of life. It's a very strong and more or less
15 continuous relationship between these parameters extending
16 over orders of magnitude. Okay.
17 So notice that nearly all the infants that are --
18 that are 500 grams or less die, this is in 1980, actually
19 some of them survive now, but nevertheless there is no sharp
20 defining distinction between those who meet the conventional
21 definition low birth weight and those that are slightly
22 above that. It's over a broad continuous range, it's a
23 little bit worse to be born a little earlier and a little
24 lighter. Okay.
25 Probably not because babies directly die of being
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
75
1 too light, but because their birth weight reflects some --
2 is a proxy for some developmental state, some developmental
3 steps that they have taken to protect themselves against
4 either infections early in infancy or forgetting to breathe
5 or something of that sort.
6 But nevertheless we have responses in animals that
7 are fetal growth inhibition. We have fetal growth
8 inhibition in humans based upon data on very small amounts
9 of cigarette smoking and things of that sort that suggests
10 that this is a very sensitive parameter.
11 So I would say that for things that operate this
12 way with this risk factor mode, we should be going to those
13 kinds of modes, rather than the sensitivity, partly because
14 we have the advantage of taking into account human data on
15 the relationships between these intermediate parameters and
16 difficult to directly -- things that are difficult to
17 directly measure that we care about.
18 So anyhow other examples of this are destruction
19 of neurons or oocytes that don't regenerate, changes in
20 cardiovascular risk factors, blood pressure, serum
21 cholesterol, that sort of stuff, where we have human data on
22 prospectively collected on the relationships between these
23 intermediates and risk, and we don't have to be very cheery
24 about that.
25 There is another slide. Right.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
76
1 Third step is to select a specific analytical
2 approach for the animal dose response data, now going back
3 to the threshold branch of the tree.
4 We want to have some -- we want to select some
5 point of departure for projection, and because of the
6 phenomenon that Lorenz illustrated in some of his slides, I
7 don't trust the human variability to be the same as the
8 animal variability. My guess is that as a general matter
9 the animals are more uniform than the human, the exposed
10 human population is likely to be, because the animals tend
11 to be relatively genetically uniform, they're treated
12 usually at an early adult age, they are treated under
13 conditions where you exclude lots of infections and moreover
14 if you look at the, at least the animal lethality data that
15 have been the basis of Sandy Baird's distributions and other
16 things, they are lot more narrow than the distributions of
17 sensitivity that I infer from the human data.
18 Now, my data are not for lethal effects usually,
19 so there is a possible difference in the comparison that I
20 can make between distributions of sensitivity for very
21 severe effects, which are typically measured in the animal
22 systems, and distributions of sensitivity for such things as
23 methyl choline sensitivity, the amount of methyl choline
24 that it takes to cause you to be -- to decrease in your
25 lungs the amount of air that you can exhale in one second by
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
77
1 20 percent or things of that sort.
2 Anyhow, I've got a lot of measurements in people
3 of not very severe effects to moderately severe effects, and
4 I don't have much in the way of human LD 50, although
5 there's just a little bit. There's lethality information.
6 And I'm -- but in any event I don't think that
7 it's a general matter we should be assuming that the animal
8 distribution of sensitivities is similar to the human. I
9 would rather go and use a generic assumption based on my
10 human experience and use the animal data only to get an idea
11 of what the chemical can do and what the -- in general the
12 potency of the chemical in the animal systems.
13 So therefore I think it's just as well to go from
14 something like an ED 50 for my point of departure for the
15 projection of the human response.
16 For biomarker mediated effects, the idea is to
17 model the biomarker response in animals as a continuous
18 function, and then translate to people before the
19 interpretation of the ultimate effect parameter in quantal
20 form.
21 Fourth, then I need to define my needed
22 adjustments, as Lorenz has just done. These adjustments
23 are -- it often happens that the study that was conducted in
24 the animals is not exactly the right length compared to the
25 human exposure, that's the acute chronic projection usually,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
78
1 and the dynamics of elimination repair in people.
2 There's incompleteness of the database, not all of
3 the studies have been done, and that was the part of
4 Lorenz's presentation.
5 There is an animal-to-human factor where one can
6 make a general expectation that on average maybe for chronic
7 effects you have this metabolic scaling factor of body
8 weight to the three-quarters, but there's clearly
9 variability among that, and any chemical that I take is a
10 random draw from that distribution about that factor and we
11 need to say okay to that.
12 And then finally there's severity effect and human
13 variability where the human variability gets done on the
14 basis of how many standard deviations do I need to get from
15 the ED 50 and projected ED 50 in people down to this ten to
16 the minus fifth type of incidence.
17 And anyhow, basically we combine these basically,
18 it's different adjustment factors with a Monte Carlo
19 simulation which is basically a fancy way just to draw
20 random samples from each of these things with some caveats,
21 right.
22 We want to worry about whether there are
23 dependencies among these, the different factors, but
24 probably there ought not to be in general.
25 And then calculate how many of our -- how often do
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
79
1 we expect, based upon the empirical distributions of each of
2 these factors done for the kind of chemical, hopefully, and
3 kind of response that we're studying, do we expect it to
4 be -- do we expect our adjustments to achieve our goals.
5 So that basically the 95th percentile confidence
6 that I referred to is based on this kind of Monte Carlo
7 simulation, hoping that we can develop empirical
8 distributions or some other distributions that reflect our
9 real uncertainty in each of these areas that we need to
10 cover from a particular set of data.
11 Now, what are the requirements for anybody to take
12 this kind of thing seriously? To be a viable replacement
13 for the current RfD I think a numerical definition needs to
14 be a plausible representation of risk management values. It
15 needs to be estimatable with no greater amount of chemical
16 specific information than is traditionally collected or
17 could be easily augmented with existing things.
18 It needs to be subjected to a series of
19 comparisons with existing RfDs, not because the existing
20 RfDs are a gold standard, but because the first question
21 that any risk manager is going to ask you is, well, how is
22 this going to change my overall standards as a general
23 matter, and we have to have, in order to even be considered,
24 we have to have that comparison made with enough cases where
25 we are pretty sure we know the general tenor of the changes
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
80
1 that are being suggested.
2 And, finally, it needs -- and this is actually the
3 part of the benefit of the alternative, is it needs to be
4 able to accommodate more advanced types of technical
5 information.
6 One of the difficulties with the current system is
7 that there's no way to use variability information in a very
8 orderly way anyhow, and therefore this hasn't been a very
9 great incentive to collect it.
10 This kind of alternative procedure should allow
11 you to plug in your empirical factors in place of a default
12 set of distributional factors based upon empirical
13 distributions from other chemicals, and therefore have a
14 predictable way to use the new kind of information on
15 comparative pharmacokinetics, et cetera, in the thing.
16 I mentioned where there is a plausible
17 representation of risk management values issue.
18 I already mentioned that part of the way I got to
19 the starting point here was an analogy with the standard for
20 notification for cancer hazard under California's
21 Proposition 65.
22 The minimally adverse severity level specified
23 makes this somewhat more protective, given the similar
24 incidence level specified for a less serious outcome, that
25 is if we do the calculations correctly.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
81
1 Estimatable with no greater chemical specific
2 information than traditionally collected.
3 Obviously, we need some generic assumptions on the
4 distributional forms for variability and uncertainty. I've
5 built, spent a fair amount of time building some of the
6 basis for this in the paper I presented in Bologna last
7 September that I think you have up here. I would, as I
8 said, be happy to inflict some of those data on you, but I'm
9 paying attention to your injunction, John, to not present
10 large tables of logarithms until provoked.
11 The basic calculations, however, assumed some log
12 normal variability among people. I think there's some
13 reason that we might want to relax that in some cases,
14 because at least there's some evidence in the data that
15 there might be some amount of bimodality in the log normal
16 distribution some of the time.
17 I can't resist --
18 DR. WITSCHI: Excuse my ignorance, but log normal
19 distribution has a tail only on one side, right?
20 DR. HATTIS: Yeah. At least much more extended
21 tail on one side than the other.
22 DR. BLANC: Which side?
23 DR. HATTIS: The side that's the bad side.
24 DR. BLANC: The sensitive side?
25 DR. HATTIS: Yeah.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
82
1 DR. BLANC: It's on this side?
2 DR. WITSCHI: Wouldn't it be the other side, the
3 resistance side?
4 DR. HATTIS: In principle it's possible, but I'm
5 viewing it the other way, but I'm trying to figure out why I
6 think it's -- basically, if you look at in log space it's
7 completely symmetrical. And I'm viewing it as --
8 DR. FUCALORO: You mean more symmetrical?
9 DR. HATTIS: Well, in theory it's completely
10 symmetrical, but at least on average, yeah.
11 What I have done is to arrange my data in always
12 in the same orientation with the worst values of the
13 parameter, the values of the parameters that would lead to
14 greater risk on one side of the thing than the other.
15 And when I do that, at least for the, for example
16 for the pharmacokinetics parameter, I would put low volume
17 of distribution on the worst side and low clearance rates on
18 the worst side, and bigger area under the curve of a
19 concentration times time product on the worst side.
20 And so in that way I get -- I do all the log --
21 the analyses in log space, so that I only translate them
22 back into linear space.
23 DR. WITSCHI: So you're a pessimist.
24 DR. HATTIS: You always get -- the problem is
25 that -- yeah. Right. Anyhow, you basically -- right.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
83
1 I don't have -- I have to think about that
2 question, because I'm obviously dancing around it, a proper
3 response to your question. Basically if you look at it in
4 log space there's no problem about which side, as long as
5 you keep the orientation respect to sensitivity consistent
6 on the database.
7 Anyhow, one can do technical sensitivity
8 calculations and maybe some uncertainty calculations
9 assuming mixtures of two or more log normal distributions
10 and figure out what the impact of a risk analysis would be
11 of that.
12 The most easily quantified uncertainty is
13 reflected in the observed variability among chemicals of log
14 normal variability estimates for overall sensitivity in my
15 case or for in Lorenz's case the other adjustment factors.
16 There's also, I think, on the scientific side some
17 potential for learning the mechanistic related groupings
18 predicted with different amounts or forms of variability for
19 either chemicals that are processed in different ways or
20 chemicals that have different kinds of actions in the body.
21 I'm finding for example that my chronic systemic
22 neurotoxicants have rather less variability than such things
23 as acute responses to locally acting things.
24 I'm finding more variability in that case than in
25 the chronic systemic neurotoxicants.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
84
1 I'm worried about whether I'm also comparing
2 different severities of effect. It's possible that I have
3 more -- less variability for very severe effects and more
4 variability for less severe types of effects.
5 But all of that is subjected -- is subject to
6 empirical study, essentially, and so what I'm suggesting
7 that one of the benefits of this general proposal is that it
8 opens up the regulatory system to a sense to influence to
9 being informed by empirical information.
10 And with that, I think I'll -- I will talk, if you
11 like, a little bit more about how I come up with my
12 estimates of variability and what they are, but I think at
13 this point I think I can open it up to some questions.
14 CHAIRMAN FROINES: Do you have any examples of --
15 DR. HATTIS: Yes, I do.
16 CHAIRMAN FROINES: -- outcomes that you actually
17 have done?
18 DR. HATTIS: Right.
19 CHAIRMAN FROINES: Why don't you give one just as
20 an --
21 DR. HATTIS: All right. Yeah. Now, this is a way
22 of plotting things essentially that turns a log normal
23 distribution into a straight line. Essentially what I'm
24 doing is plotting in this, some of the time I'm plotting the
25 Z score. I didn't invent this method. This is called
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
85
1 probit analysis. This was invented in the '30s, right.
2 This is not something that I came up with myself.
3 This is essentially the Z score of the percent
4 response essentially and the Z score essentially is the
5 number of standard deviations that you have to go from the
6 mean of a distribution to have the area under the curve of
7 the distribution correspond to the percent of the people who
8 respond at a given level.
9 So essentially if ten percent of the -- if five
10 percent of the people respond with -- to a given
11 concentration of chromium with skin sensitivity, then we say
12 that five percent of the people have thresholds below that
13 concentration.
14 And if 50 percent of the people responded some
15 higher concentration, we say that that corresponds to a Z
16 score of zero.
17 The five percent level corresponds to a Z score of
18 minus 1.64, which is essentially in a normal distribution.
19 If you go 1.64 standard deviations below the mean,
20 then five percent of the people will be even farther from
21 the mean than that, below that.
22 Anyhow, so if I plot this kind of Z score
23 transformation versus the log of the chromium concentration
24 that it gave responses in this group of tested people,
25 basically this is 102 tested, people were tested. 54 of the
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
86
1 them responded to the highest dose level, just a little bit
2 over a Z score of zero, okay, so just a little bit more than
3 the Z.
4 What you see is that the points can be arranged in
5 a way that is reasonably close to a straight line, which
6 means that these data are reasonably -- suggest a reasonably
7 log normal distribution of the individual thresholds of
8 these people for this kind of response.
9 The same -- and basically the slope of that line
10 relates to the amount of individual variability that you
11 have.
12 In this case, the slope is rather shallow and
13 corresponds to a lot of variability. Basically for those of
14 you who think in log normal statistics, this is a geometric
15 standard deviation of ten. Okay.
16 What that means is that 95 percent of the
17 population would have the threshold spread out over four
18 orders of magnitude, essentially from a 100-fold less than
19 the median, minus two standard deviations, to a 100-fold
20 more than the median.
21 So this is an example of a response, and this is
22 based on individual clinical measurements. So I'm not
23 talking about anything that's a little faky here. This is
24 talking about clinical measurement that is quite highly
25 variable.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
87
1 What basically most of the things are not as
2 variable as that, but there are quite a few things that are
3 almost as variable. Among the things that are almost as
4 variable -- let me get a couple of them. This I'm choosing
5 partly because it's a large study, and I'm told that one of
6 you is a lung physician.
7 This distribution, this is a distribution of --
8 this is a plot in the opposite direction. This orientation,
9 the shallow slope, means less variability. But don't get
10 too confused about that.
11 This is the log normal plot of the concentrations
12 that cause, I believe, a 20 percent increase -- 20 percent
13 decrease in FEV 1, the amount of air that you can breathe
14 out in one second, down in a very large epidemiological
15 study over 5,000 smokers with mild to moderate air flow
16 obstruction.
17 And what you see again is a pretty decent
18 correspondence of the points to the theoretical predictions
19 of this straight line, which represents the log normal
20 distribution.
21 You see a pretty good amount of variability. This
22 is -- this is the log -- the geometric standard deviation
23 would be 10 to the .6, which is about, doing it quickly in
24 my head, six or seven or so. And that's, you know, that's
25 enough variability so that if you wanted to get from a five
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
88
1 percent effect level to a ten to the minus fifth effect
2 level, you need about 3.3 standard deviations to do that.
3 3.3 standard deviations is three times .6, meaning
4 1.8 orders of magnitude. So you need something like, oh,
5 50-fold to do the job that you would expect the 10-fold
6 safety factor to do if what you want the 10-fold safety
7 factor to do is to get you from a five percent effect level
8 consistent with a no observed effect, to ten to the minus
9 fifth effect.
10 So that's essentially the implication of that kind
11 of variability.
12 DR. BLANC: Say that statement one more time.
13 DR. HATTIS: Right.
14 DR. BLANC: This is not the key statement you're
15 trying to make through all of this?
16 DR. HATTIS: Yeah. What I'm saying is that the
17 thrust of Lorenz's and Sandy Baird's stuff is that several
18 of the uncertainty, the tradition uncertainty factors do
19 appear to be conservative, that is that they build in some
20 more protective factor than would be needed, at least for
21 the average chemical, which is what was the intent after
22 all. But this one may not be as a general matter.
23 DR. BLANC: Because the range that you would have
24 to go, to go from five percent, which would be the no effect
25 level in humans to --
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
89
1 DR. HATTIS: To a ten to the minus fifth.
2 DR. BLANC: To ten to the minus fifth was how many
3 logs?
4 DR. HATTIS: That often you need --
5 DR. BLANC: In that particular case.
6 DR. HATTIS: In that particular case you might
7 need about 50-fold, rather than 10-fold. You might need
8 another factor of five or so, if the population is really
9 log normal. Okay.
10 DR. BLANC: Wouldn't we be going a 100-fold,
11 because wouldn't we be going from the lowest to the values
12 in the no effect level, which was assuming -- okay --
13 DR. HATTIS: Okay. Some of the time you have --
14 you don't have perfect log normal distributions. Some of
15 the time you have -- now, this is a classic case of a
16 genetically determined variability. This is a distribution
17 of -- this is from the original paper, propyl thiouracil
18 taste sensitivities. Now this is not an adverse effect,
19 okay. So I must say that. But this is a different kind of
20 response. Each of these blocks from minus three to minus
21 two, four of those represent one order of concentration of
22 propyl thiouracil. But what you do see is that there is now
23 we look like there's actually two distinct subpopulations.
24 But nevertheless there's still variability in each of the
25 subgroups.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
90
1 DR. GLANTZ: What are the open bars?
2 DR. HATTIS: The open bars -- the open bars are
3 the total population. The shaded bars are the things that
4 he characterized as tasters and nontasters.
5 And I -- one of my pet peeves is how biologists
6 have a tendency to dichotomize. I mean, I do believe that
7 there's two humps to this distribution, but by saying that
8 these are the tasters and these are the nontasters, you're
9 giving the wrong impression. These are the more sensitive
10 and the less sensitive people. Okay. It's not that --
11 DR. FUCALORO: Biology has a tradition of
12 taxonomy.
13 DR. HATTIS: I like taxonomy. I think taxonomy is
14 good. I think the overuse of dichotomization is a disease.
15 I mean, there are people who talk of the normotensives and
16 the hypertensives.
17 DR. FUCALORO: It's either a bird or a dinosaur.
18 DR. HATTIS: That's right. And you either have
19 chicken pox or you don't.
20 But it's not true that you either have -- right?
21 But so there are some things where you should dichotomize, I
22 think, but you don't dichotomize everything.
23 And you do have to make decisions about who you
24 test for -- who you treat for blood pressure, but it isn't
25 true that there is -- blood pressures are a broad continuous
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
91
1 distribution, and to draw an arbitrary line and say these
2 people are qualitatively different than those people is just
3 without basis. It's only based upon my habits and
4 prejudices. Anyhow.
5 DR. BYUS: You should have been here for lead.
6 You would have loved that discussion.
7 Don't write that down.
8 DR. HATTIS: This isn't the first time that I've
9 encountered this problem, actually, which is why I'm
10 venturing to be pejorative about a whole class of people
11 that includes myself.
12 Anyhow, nevertheless, what I'm trying to say as a
13 general matter is that sometimes we have to worry about
14 bimodality of the distribution.
15 But it's often the case that we have variability,
16 particularly for these less severe responses that clearly
17 spans a great deal more in the order of magnitude.
18 With that, I'll turn it back, or if you want I
19 will again go into a little bit more detail about how I'm
20 trying to figure out, reconstruct distributions of overall
21 variability. Basically I've got a database of something
22 like, it's over 200 observations now, and each observation
23 is a data set or a number of data sets where I'm looking at
24 variability in human populations.
25 The human populations are not always ideal for
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
92
1 projections in the way we're trying to do it. Sometimes
2 they're restricted to normal healthy adults. In fact,
3 rather often they're restricted to normal healthy adults,
4 often young, normal, healthy young adults of one gender, and
5 there are other difficulties, but nevertheless the idea is
6 by putting together variability observed for a number of
7 portions of the pathway between external exposure and
8 biological response, classifying the biological responses
9 one way or another, I'm trying to develop the empirical
10 distributions that one would need to say, okay, for this
11 kind of chemical, for this kind of response, what should we
12 expect.
13 But absent the subcategorization we can use the
14 existing distribution for some beginning calculations.
15 CHAIRMAN FROINES: George or people in the
16 audience have questions to ask now?
17 DR. ALEXEEFF: George Alexeeff with OEHHA.
18 I was thinking back to the earlier part of your
19 presentation and the issue about different endpoints. And
20 one of the things that we struggled with with our acute
21 document was different endpoints with different severities.
22 So, for example, reproductive toxicity that might
23 occur at a slightly higher dose than respiratory irritation.
24 So theoretically we have two different sensitive
25 subpopulations. One is for the respiratory irritation.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
93
1 Let's say we can have asthmatics, but then for the
2 reproductive effects or depending on how you define pregnant
3 women, or potentially pregnant woman as a sensitive
4 subpopulation.
5 So I'm just wondering how one can look at your --
6 the calculation. I was thinking of the one in a thousand
7 versus the one in 100,000, and if one then has different --
8 you're talking about carrying forward the critical effects
9 of multiple critical effects, if your thought was to then be
10 able to calculate differential impact to different
11 subpopulations.
12 I don't know if I explained it. But to me that
13 seemed to be maybe a power of carrying forth those critical
14 effects.
15 DR. HATTIS: I think it's simpler to keep them
16 separate and do the complete calculation in parallel terms,
17 because you don't know at the early step, you could have a
18 difference of severity which would require you to have an
19 adjustment factor that you needed to determine later for
20 that to adjust back to a minimal adverse severity level, if
21 that's the way you're doing it.
22 Or, you know, at some point you might even say if
23 we can't adjust the severity, say we've got a quantal type
24 phenomenon, either you've got a cleft lip or you don't, and
25 there is no such thing -- or at least I guess both -- then
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
94
1 you might say we don't want to tolerate as much as one in
2 100,000 of that. You might want to have a stronger -- we
3 want to be more confident that as a general matter that you
4 don't have that effect or you want to have an equal level of
5 confidence for a smaller incidence than --
6 DR. ALEXEEFF: Well, I was thinking less in terms
7 of adjusting for severity and more simply that calculation.
8 I don't know what the percentages of how one would define
9 the sensitive subpopulation for reproductive effects. If
10 you're including all women of reproductive age or only those
11 women that are currently pregnant at the time.
12 And the same thing on the asthmatic side.
13 You would have different proportions instead of --
14 I mean, it could go anywhere from one in two, or one in
15 three, to maybe a much -- if you assume women of
16 reproductive age might be one in four of all people. I
17 don't know what the amount is, but some very high number.
18 And if that could end up impacting or actually
19 causing a lower level than you would for the respiratory
20 effects, even though it's more mild, because you have a
21 larger population that could be impacted.
22 That's actually I was wondering if you had done
23 those kind of calculations where you had differential -- I
24 don't know if I'm making sense -- but differential impacts,
25 differential sizes of the sensitive subpopulations that
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
95
1 could then end up causing the actual final effect level to
2 overlap.
3 DR. HATTIS: I think that that's an interesting --
4 the short answer to your question is that I haven't tried to
5 do that. I think that that's where we need to have some
6 serious dialogue with the risk managers and the affected
7 communities as to exactly what the rules are for defining
8 sensitive subpopulations of that kind.
9 And I don't have a quick and easy answer to
10 suggest that, after all, I'm just a poor technical person, I
11 don't do risk management.
12 I do think that it's -- I think that one of the
13 key things that a policy person would say is that if you can
14 narrowly define a population, and it's easy for them to do
15 ex ante, if it's easy for the person to say I'm pregnant,
16 I'm an albino, and therefore I know that I'm much more
17 sensitive than you typified in your calculation. It might
18 be reasonable to deal with them.
19 I mean, but I don't know exactly how far you take
20 that. I don't know how fair that is. I mean, you have
21 people who are very highly sensitive to iron in the diet and
22 that came up as an issue in the allowing of iron
23 supplementation of bread.
24 So I don't know exactly how to do that.
25 CHAIRMAN FROINES: Thanks, Dale.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
96
1 Lorenz.
2 DR. RHOMBERG: Lorenz Rhomberg.
3 Dale, I was wondering have you ever given any
4 thought to the notion that you might want to adjust the ten
5 to the minus fifth or ten to the minus whatever level that
6 you are trying to protect, based on the size of the
7 population exposed, so that ten to the minus fifth is very
8 stringent if you're only exposing four or five people. The
9 chance that one of those might be among them is very small,
10 but if you're exposing the whole population of the US, then
11 it's not very stringent at all.
12 You can imagine something that says I'm going to
13 protect sort of with some assurance the most sensitive or
14 next to most sensitive in the group or something like that,
15 and it could have it slide depending on how many were
16 exposed.
17 DR. HATTIS: These calculations are all in a
18 context of an equity framework. Is it unfair to impose this
19 as a general summary thing before we talk about tradeoffs,
20 essentially.
21 I think that there's another whole set of
22 calculations of where one is juxtaposing cost and benefit
23 where the size of the population exposed in affected matters
24 centrally. And so that I think for purposes of the equity
25 discussion, I think maybe it's most significant to talk
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
97
1 about these individual risk things, but for purposes of
2 saying how do we best direct our resources, how do we set
3 priorities for social intervention, there's certainly room
4 for the size of the population to enter into that whole
5 other set of discussions about juxtaposing cost and benefit,
6 which I think is its own thing.
7 I'm not absolutely sure whether that I want to
8 automatically allow as an equity matter.
9 Certainly at the point where you get to very large
10 risks, as much as one in ten, you start to get into the
11 realm of criminal law. So you can't, regardless of the size
12 of the population, you can't go as high as that.
13 And I think it certainly matters whether you're
14 exposing one in a million people or ten people to a ten to
15 the minus six or ten to the minus fifth risk. I think
16 that's more a discussion I guess of the social priority that
17 should be allocated to those.
18 DR. FUCALORO: Weren't your comments based upon a
19 population large enough that you reached a sampling is equal
20 to probability, essentially? That the population is so
21 large that when you say one in 100,000, at least a million
22 people or more to -- you're certainly not talking about a --
23 DR. HATTIS: That was my general idea, but still
24 I'm wondering, you know, if you have a population of a
25 hundred houses in a neighborhood of a facility of some sort,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
98
1 would you want to do a different calculation, would it be
2 equitable to expose them to a greater amount of risk just
3 because they're relatively fewer of them.
4 I think in an equity terms maybe not.
5 On cost-benefit terms, then that certainly does
6 matter.
7 So I think it matters what's the social policy
8 framework that is governing the particular choice.
9 CHAIRMAN FROINES: Thank you, Dale, and Lorenz.
10 Thank you very much.
11 (Applause.)
12 CHAIRMAN FROINES: I think we would have to
13 restructure the nature of this panel to deal with the issues
14 that Dale just raised. We'd have, since we're now mixing
15 risk management and risk assessment --
16 DR. HATTIS: This is firmly in risk management,
17 but the point is I'm presuming to raise it as a technical
18 person is that I'm convinced that the risk management
19 community will only work -- will only start to think about
20 it if the technical community opens up some choices for
21 them.
22 CHAIRMAN FROINES: We have three more items on the
23 agenda. One is the MOE REL approach, one is the findings
24 and one is the update.
25 I think all of those are going to go relatively
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
99
1 quickly.
2 So what I think we should do, if everybody agrees,
3 is to take a ten-minute break for the stenographer and then
4 to go until we finish those three items and then call it a
5 day, if that's acceptable. But I think she would like to
6 have another break.
7 (Thereupon a short recess was taken.)
8 CHAIRMAN FROINES: Melanie and Jay Schreider.
9 There he is.
10 DR. MARTY: I think I was tasked with just a very
11 brief overview of what we do for reference exposure levels.
12 The panel has heard all of this in the
13 deliberations on the technical support document for
14 determining acute reference exposure levels. So I'm going
15 to be very brief.
16 OEHHA has developed a number of acute and chronic
17 reference exposure levels, and you've seen the acute and
18 you're going to see the chronic.
19 We used either the NOAEL divided by uncertainty
20 factor approach, which you just heard a whole lot about, or
21 the benchmark dose divided by uncertainty factor.
22 We have used uncertainty factors in a way that's
23 quite similar to the US EPA approach.
24 Essentially the uncertainty factors used with the
25 benchmark concentration approach vary whether the study
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
100
1 subject is animals or humans or sensitive human population,
2 and factors have been used for intraspecies variability, as
3 well as interspecies variability.
4 The uncertainty factors used with the NOAEL
5 approach are fairly similar to that used with the benchmark
6 dose approach, although we have a couple of other issues and
7 that is we may not have a NOAEL, we may be stuck with LOAEL
8 and we have to do that extrapolation when there's one of
9 those uncertainty factors would account for extrapolating
10 from a low observed adverse effect level to a no observed
11 adverse effect level.
12 Also I should throw in for chronic reference
13 exposure levels sometimes you don't have a chronic study,
14 you have a subchronic study, so there's an additional
15 uncertainty factor that would be incorporated into your
16 reference exposure level calculation.
17 CHAIRMAN FROINES: What is that value?
18 DR. MARTY: Generally ten for subchronic to
19 chronic.
20 Also for the NOAEL approach it's going to the
21 uncertainty factor you use is going to depend on whether
22 it's an animal study or a human study or a study in
23 sensitive humans.
24 This table is taken out of the acute reference
25 exposure level document, basically gives an idea of the size
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
101
1 of the uncertainty factors used for the different
2 uncertainties.
3 Essentially if you start out with an animal study
4 and you use the benchmark concentration where you have a
5 regression analysis from -- to get you to the benchmark
6 concentration for a five percent response rate, we would use
7 a 10-fold intraspecies species variability factor and a
8 3-fold interspecies uncertainty factor, for a total of about
9 30.
10 If you start out with a human study, but it didn't
11 involve sensitive individuals, for intraspecies we have used
12 either three or ten.
13 We use ten in the case of formaldehyde, even
14 though it was a benchmark concentration, because there was
15 good evidence for a wide variability in the response to the
16 irritancy produced by formaldehyde.
17 And then if you have a human study that was done
18 in sensitive individuals, then we have used an interspecies
19 variability factor as low as one.
20 Next slide, Jim.
21 DR. BLANC: How would you comment on that in terms
22 of the data we heard earlier that for example for certain
23 human sensitivities, maybe 50 would be better and that the
24 range should be from three to 50 instead of from three to
25 ten?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
102
1 DR. MARTY: Yes. That is a possibility. And we
2 did see that with the -- Dale showed a slide of people
3 reacting on a patch test to hex chrome and in my mind that
4 is the type of response where you would anticipate seeing a
5 lot of variability because, correct me if I'm wrong, but
6 it's essentially a hypersensitivity response.
7 So I think that we have been using ten, and we're
8 still using ten, and for some endpoints we are probably not
9 being health protective enough.
10 DR. BLANC: George.
11 DR. ALEXEEFF: George Alexeeff with OEHHA.
12 Probably the example that we have is the
13 formaldehyde example, because in that case, although Melanie
14 just quickly went through it, we took a benchmark
15 concentration which we already calculated the response, at
16 least within the study population, the five percent response
17 rate, and then on that we add an additional 10-fold
18 uncertainty factor. The reason for that is because we're
19 looking at all the other formaldehyde data, there was a very
20 very wide distribution of effects, still some individual
21 effects reported at some lower levels. So that seemed to be
22 an example that might be consistent with what Dale was
23 saying where it's a very very broad range.
24 DR. MARTY: I think a lot more research needs to
25 be done on variability by endpoint.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
103
1 CHAIRMAN FROINES: But the implication of what
2 Dale was saying is that to the degree that data does exist,
3 you could incorporate that data into an estimate of your
4 uncertainty factor. You're not -- nobody said it has to be
5 as small as ten. Nobody said it has to be ten.
6 DR. MARTY: Right.
7 CHAIRMAN FROINES: So to the degree that issue has
8 merit in terms of existing data, and Dale has 300,000 log
9 probit graphs that he can give you, at least that's what I
10 think he said. It's something to consider in terms of
11 certain chemicals as they come up.
12 DR. GLANTZ: Actually, I think Dale has three
13 times ten to the fifth graphs.
14 DR. ALEXEEFF: Actually, I think that what Lorenz
15 was saying that part of the original ideas of these
16 uncertainty factors was when you went from animal studies
17 you had animal factor and the human factor, and it could be
18 that part of the animal factor was really helping to deal
19 with the human variability, but just overall the factor was
20 a hundred.
21 So I think the situation that we might be
22 concerned with are those where we're basing it actually on
23 human data and the variability within humans might be much
24 broader than ten, because we don't have any other factors
25 we're adding in. So if there were ways that we can develop
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
104
1 or understand what would those situations be, that would
2 greatly help us in terms of making sure we're providing
3 protective levels.
4 DR. MARTY: This table is also taken partly from
5 the truncated table from our acute REL document.
6 When we use the NOAEL method, we have both
7 interspecies and intraspecies uncertainty factor, and they
8 have generally been ten.
9 If you have a human study, then obviously you
10 don't have an interspecies adjustment that needs to be made,
11 but you may want to make an intraspecies adjustment,
12 particularly if your study does not include sensitive
13 subjects.
14 We also have the case where we may have to
15 extrapolate from a low observed adverse effect level, and we
16 have done an analysis of the subset of chemicals where we
17 looked at the LOAEL to NOAEL ratios and for mild effects the
18 95th percentile was six. So we ended up using that for mild
19 effects.
20 For other than mild effects, we used a factor of
21 ten.
22 DR. FUCALORO: Excuse me. Define mild effect.
23 DR. MARTY: We have a whole definition in that
24 document.
25 DR. FUCALORO: I know you do.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
105
1 But that is -- that's essentially the type of
2 effect, I mean, as opposed to being just a low dosage, am I
3 correct?
4 DR. MARTY: Right. The severity of the endpoint.
5 Anyway, in sum, that's what we've been doing to
6 develop reference exposure levels.
7 CHAIRMAN FROINES: Move ahead.
8 DR. MARTY: Jay Schreider is going to talk about
9 what DPR has been doing.
10 DR. SCHREIDER: My name is Jay Schreider. I'm
11 from Department of Pesticide Regulation.
12 And I think what we do is many respects very
13 similar to what OEHHA does. We tend to use in a lot of
14 instances a margin of exposure approach which is similar to
15 what US EPA does, in least in the Office of Pesticide
16 programs, and that's frequently because we are looking, or
17 at least initially, was the focus was on occupational
18 exposure, and that we have a no observed adverse effect
19 level over the exposure, and that's contrasted with the REL,
20 or in our case what we sometimes use is the reference
21 concentration, where you have the no observed adverse effect
22 level over a specific uncertainty factor.
23 The endpoints of the studies obviously would be
24 the same, that is there's no difference in how we're
25 treating the endpoint or how we're judging the endpoint.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
106
1 One difference you can see at the top where we --
2 again the margin of exposure approach or what used to be
3 called the margin of safety.
4 With a reference concentration we employ a
5 breathing rate correction or breathing rate for along with
6 the uncertainty factor to adjust for breathing rates between
7 adults and children and then also between the experimental
8 animals and man, assuming a higher breathing rate for
9 children and that frequently as a result becomes the driving
10 factor in the risk assessment.
11 DR. GLANTZ: Can I just ask a question.
12 Would I be correct in saying that the -- what
13 you're calling the reference concentration and what OEHHA is
14 calling the reference exposure level are the same? Is that
15 true? No? Okay. Except for the breathing rate.
16 DR. SCHREIDER: Except for the breathing rate they
17 would be the same.
18 DR. GLANTZ: Is that --
19 DR. SCHREIDER: I mean, there may be differences
20 on individual basis on the uncertainty factor, but the ideas
21 would be the same.
22 DR. GLANTZ: Well, let me ask some more precise
23 questions, because assuming that you did the uncertainty
24 factors the same way, and if you didn't take breathing rate
25 into account, then what you're calling the reference
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
107
1 concentration would be the same as what OEHHA is calling the
2 reference exposure level?
3 CHAIRMAN FROINES: No. Because in the chronic
4 document OEHHA uses breathing rate adjustments. So that
5 depends on whether we're talking about acute versus chronic.
6 DR. GLANTZ: Okay.
7 CHAIRMAN FROINES: In this, he must be referring
8 to a chronic.
9 DR. SCHREIDER: We would use the breathing rate
10 regardless of whether we're talking about acute, subchronic
11 or chronic. We apply a lot of our data or no effect levels
12 from acute studies, and we would still use the breathing
13 rate correction.
14 DR. GLANTZ: Does the breathing rate correction --
15 the breathing rate correction makes the reference
16 concentration bigger or smaller?
17 DR. SCHREIDER: Lori.
18 DR. LIM: This is Lori Lim from DPR.
19 I think we need to show another transparency.
20 DR. GLANTZ: Am I getting ahead of you here?
21 DR. LIM: It's just a secret.
22 DR. GLANTZ: The secret transparency.
23 DR. LIM: Just in case.
24 This shows exactly what I think you're talking
25 about in comparison, whether you adjusted it or not.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
108
1 DR. SCHREIDER: The only difference would be going
2 over the reference exposure at the top, and the reference
3 concentration at the bottom where we had got the first
4 adjust after the animal study which was using six-hour
5 exposure, adjust that for 24 hours, and for the breathing
6 rate of the animal, and then convert that for the breathing
7 rate of the child at the bottom, and we would come up with a
8 reference concentration of 40 micrograms per cubic meter,
9 versus taking the concentration from the animal study and
10 dividing it by the uncertainty factor would give reference
11 exposure of 122 micrograms per cubic meter.
12 DR. GLANTZ: So I'm just trying to harmonize here.
13 Anyway, this is in honor of Rick Becker.
14 So if OEHHA was given the same NOEL, okay, would
15 you come up with the same numbers?
16 DR. ALEXEEFF: Well, there's --
17 DR. GLANTZ: If not, could somebody explain to me
18 the difference.
19 DR. ALEXEEFF: There's a couple of different
20 things going on in this slide, and it almost -- we almost
21 have to have a couple more slides just to sort of tease a
22 couple things out.
23 For example the six-hour to 24-hour adjustment
24 there, okay, that is a weighted adjustment. So if we were
25 making a 24-hour level, we would do the same adjustment. So
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
109
1 that's right there a factor of four. Okay. That seems to
2 be kind of -- so if you were -- the top one is basing it on
3 a six-hour reference exposure level and the bottom one is a
4 24-hour reference exposure level.
5 Is that right, Lori, just to be sure?
6 So in one sense that adjustment is kind of
7 confusing. The other adjustments, although they're showing
8 how they make the calculations, which is what the purpose of
9 this slide is, but to make the -- understand the question
10 you're -- to answer the question you're asking, if we were,
11 like I say, we would always, if we were looking for a
12 24-hour level, then we would make the six over 24 adjustment
13 just like they did. So what our number would look like was
14 basically 122 divided by four. So it actually would be like
15 30. So that would be the actual difference in this case if
16 we were both coming up with 24-hour levels, our number would
17 be 30 something and theirs is 40.
18 CHAIRMAN FROINES: How do we --
19 DR. GLANTZ: I'm now totally confused.
20 DR. ALEXEEFF: There's two adjustments happening
21 in here, which is their standard practice.
22 One is the time adjustment, because they're
23 developing a 24-hour standard.
24 The other one is the breathing rate adjustment.
25 You asked a question about the breathing rate
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
110
1 adjustment. In order to understand that, how that impacts
2 it, you have to sort of ignore the time adjustments.
3 So I was trying to normalize the time adjustment
4 issue, which we would also do if we -- remember, we had that
5 concentration time formula? So it would depend upon the
6 exact situation what the exact calculation was.
7 CHAIRMAN FROINES: But, George, but that's a
8 specific point. You're attempting to develop acute RELs for
9 a one-hour exposure and you're making a determination that a
10 one-hour exposure is precisely an acute exposure.
11 Now, for any time averaging that's done, there
12 should be a reason for it.
13 So the question I think Stan is getting at in
14 terms of the six divided by 24 hours is what's the
15 toxicologic basis for that particular adjustment.
16 Is that what you're saying?
17 DR. ALEXEEFF: That's the kind of adjustment we
18 would make, let's say, in our cancer documents or our
19 chronic documents do that exact same adjustment. It's just
20 that regular averaging you would be considered concentration
21 times time to the concentration of the N where N is one.
22 CHAIRMAN FROINES: I understand.
23 DR. ALEXEEFF: Just the regular averaging time in
24 that case.
25 So this is assuming sort of a, what do you want to
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
111
1 call it, lineal extrapolation over concentration to the nth
2 power, where N equals one. And we would do that kind of
3 calculation if we were doing adjusting from like an
4 occupational exposure.
5 DR. FUCALORO: So in other words, six hours a day
6 of exposure, that all you're doing, right?
7 DR. LIM: What we are assuming is that person
8 would be exposed for 24 hours, so we amortize it to 24
9 hours.
10 DR. FUCALORO: What's the concentration, because
11 you ultimately have to meet the standard of milligrams -- I
12 got myself a little confused.
13 You really ultimately have to get the standard of
14 micrograms per kilogram per day.
15 DR. LIM: Maybe I should just take it through
16 slowly. It's kind of complicated.
17 The whole conversion is analogous to dietary
18 experiment where we have certain amount of food, but the
19 animal is not going to take it all, so you have to put in
20 food consumption rate.
21 So in this case we're saying we're putting in
22 breathing rate factor into consideration.
23 So the animal is exposed to NOEL in terms of air
24 concentration in this amount.
25 The rat was breathing at this amount and for six
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
112
1 hours and you amortize it over 24 hours, the dosage for that
2 animal for the no effect level in terms of milligram per
3 kilogram per day is this level.
4 So when we back calculate it for the child to have
5 the same dosage, breathing at this rate and putting the
6 factor of ten, what air concentration should it be that it
7 cannot be higher than this particular air concentration.
8 DR. FUCALORO: So 40 micrograms per meter cubed
9 and a child breathes six hours a day will put him or her at
10 the 2.9 milligrams per kilogram per day, 24 hours.
11 CHAIRMAN FROINES: Let's go ahead.
12 DR. SCHREIDER: Next one.
13 In terms of the uncertainty factors that we use,
14 there again are similar and I'd like to point out though
15 that these are adult uncertainty factors so that when
16 obviously there is more information that either humans are
17 less sensitive or more sensitive than assumed for the
18 10-fold or that the variability may be more or less, that
19 would be factored in, so that again the starting from a no
20 effect level or no observed adverse effect level from an
21 animal study, we would use an intraspecies uncertainty
22 factor of ten to get to humans and then intraspecies factor
23 of 10-fold again for as a default for the human variability.
24 The second case would be where we have the no
25 effect level for human study, and then in that case we would
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
113
1 go back to what we're talking about before, and use the
2 interspecies uncertainty factor of ten.
3 That actually may be coming more of an issue as we
4 start seeing more data submitted where in fact no effect
5 levels are generated in humans and we're starting to see
6 some of those with some of the organophosphates, with those
7 studies done to generate a no observed effect level for
8 cholinesterase inhibition, so that that starts becoming an
9 issue.
10 CHAIRMAN FROINES: Tony.
11 DR. FUCALORO: I was just wondering, I don't
12 understand the heading for that where you say the
13 uncertainty factors used by DPR in calculation of MOE or
14 reference concentration. Nowhere did I see in your
15 equations MOE using an uncertainty factor.
16 Am I missing something?
17 DR. SCHREIDER: Yeah. When we calculate, let's
18 say we calculate an uncertainty factor of 40 for a given
19 situation, that's compared with -- I'm sorry I wasn't clear.
20 That was compared with a benchmark, and that benchmark would
21 be uncertainty factor. In that case it goes from an animal
22 study to benchmark of a uncertainty factor of 100.
23 Would then be for management to determine whether
24 or not the adequacy of the -- of that exposure situation.
25 DR. FUCALORO: It's not MOE, you're just talking
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
114
1 about a reference exposure.
2 DR. LIM: In that --
3 DR. FUCALORO: Maybe I'm missing. I don't see
4 anywhere where an MOE uses that.
5 DR. LIM: Precisely. It is not in the equation.
6 What I meant to say is that in our discussion in our
7 documents in the risk appraisal section we usually generally
8 say have some discussion where the default ten would be
9 adequate or not for interspecies or intraspecies. That's
10 all I meant. It didn't mean that it would physically be in
11 the equation.
12 CHAIRMAN FROINES: This is a very important point.
13 This is an extremely important point, precisely because the
14 MOE in the DEF document is the basis for the determination
15 of whether something is a toxic air contaminant or one of
16 the bases.
17 But there is no explicit incorporation of
18 uncertainty factor in the MOE calculation. I'm being very
19 clear. I'm saying there is no explicit incorporation.
20 There may be an implicit incorporation, but that's what
21 we're trying to get you to tell us about.
22 DR. GLANTZ: See, and I just want to jump on that,
23 because I had been -- I mean, this is actually very helpful
24 because I'm beginning to understand these distinctions
25 finally.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
115
1 But the problem with using the MOE is this point
2 about the lack of an explicit consideration of the
3 uncertainty factor, because, as you know from attending
4 these meetings, a huge amount of discussion goes into the
5 question of how to handle the uncertainties, and by not
6 explicitly accounting for that, what you're doing when you
7 come up with an MOE, let's say you come up with an MOE of
8 9600, which is the DEF number. Okay.
9 Well, that could be a very big margin of -- if
10 there was not a high level of uncertainty associated with it
11 or it could be not a very big margin if there was a lot of
12 uncertainty.
13 So that I think what we need to be shooting toward
14 is actually going more toward the idea of the REL in these
15 pesticide documents where the REL that we come up with and
16 sort of bless, explicitly accounts for the uncertainty
17 factors, and then if you want to talk about a margin of
18 exposure based on the REL, that I think would make sense
19 because then what you're doing is you're coming out and
20 saying the margin of exposure is a thousand, what you're
21 saying is that the exposure levels are a thousand times
22 below the REL.
23 CHAIRMAN FROINES: Let me make a point about that.
24 DR. GLANTZ: Because, see, the problem you have
25 with this -- this is what confused me about this all along,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
116
1 the problem that you have with these, the margin of
2 exposures is that you end up with an apples and oranges
3 problem, because you're not taking into account the fact
4 that one margin of exposure may be associated with something
5 with a lot of uncertainty and another one might be
6 associated with something where there isn't a lot of
7 uncertainty.
8 CHAIRMAN FROINES: I want to follow up.
9 DR. GLANTZ: And then I have another question.
10 CHAIRMAN FROINES: I want to make one point about
11 this.
12 The child MOE in the DEF document is 9600.
13 Now, part of the problem with that is the
14 perception. Perception weighs heavily in all this.
15 So 9600 looks to be very safe. So everybody says,
16 wow, we don't have to worry about DEF because the number is
17 9600.
18 Well, now we have a different NOEL for DEF than we
19 did when this document was written, which is now .6 and so
20 when you take the .6 versus the 12.2, and then if you add an
21 uncertainty factor of 100 in, it turns out that your MOE
22 incorporating the uncertainty factor into the NOEL
23 calculation turns out to be 20. That's a very small number.
24 And one then has to say when you have that number
25 of 20 is the public appropriately protected at that level.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
117
1 And I think we might all agree that they are and
2 you might disagree that they are.
3 And the fact of the matter is once when you start
4 to get these numbers of 9600, it's because if you divide by
5 a hundred, you're down to 96. And therein lies the problem,
6 I think, everybody is concerned about.
7 It is precisely that by not incorporating the
8 value of 100 into an uncertainty factor calculation it gives
9 an appearance of being different than in fact it really is,
10 let alone the science that we can talk about.
11 DR. SCHREIDER: Yeah. I think there would be a
12 couple points.
13 One is that we're moving in the direction of doing
14 the RELs, uncertainty for these documents, generating a
15 reference concentration.
16 And also I think part of this may be due to how we
17 started out, which was, aside from air toxics of having a
18 set concentration or set pesticide and starting with a
19 concentration on occupational exposure and saying, okay, is
20 exposure from all routes adequate or not adequate, and some
21 of this may be residue of that, and certainly the FQPA we're
22 still having to look at aggregate risk.
23 But I think certainly we're moving in the
24 direction of generating both values.
25 And part of it is also do we generate it
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
118
1 numerically, which you're suggesting explicitly and
2 numerically, or do you handle it in the risk appraisal
3 section where it's described.
4 And I think I clearly hear you say, no, ought to
5 be really cut and dry so that it doesn't give the impression
6 even if the --
7 CHAIRMAN FROINES: I've been putting words in your
8 mouth, because I'm assuming as you go along in your
9 presentation you're going to say that the MOE calculation
10 incorporates somehow what you haven't really said yet, the
11 100-fold factor.
12 So why don't you go back to where you were.
13 DR. GLANTZ: Before you do, let me continue
14 interrupting here.
15 Because I'm just trying -- I'm just trying to
16 understand how what DPR and OEHHA are doing different,
17 because my goal is in the end to have you using the same
18 procedures, the same definitions.
19 DR. BYUS: The definitions.
20 DR. GLANTZ: The best definitions available and
21 what's good for -- but when you look at the uncertainty
22 factors, which you just you had up, or you recited, I went
23 back and looked at the slides that Melanie presented, and it
24 looks to me that you're using the same uncertainty factors
25 that OEHHA is. Is that true?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
119
1 DR. SCHREIDER: Basically.
2 DR. GLANTZ: Okay. Basically?
3 DR. SCHREIDER: We look at a given date set --
4 DR. GLANTZ: Let me be more precise. The default
5 values that you're using are the same; is that correct?
6 If there's data to not use a default factor,
7 neither you -- both you and OEHHA would use the data. But
8 in the absence of specific information you're using the same
9 uncertainty factors and you're using uncertainty factors for
10 all the same things too, right?
11 DR. SCHREIDER: Right.
12 DR. GLANTZ: If I handed a NOEL to Melanie or to
13 you and said go compute the reference exposure, the
14 reference exposure level, the reference concentration, you
15 both get the same number. Is that correct?
16 DR. SCHREIDER: Correct.
17 DR. FUCALORO: But I guess the problem, and maybe
18 I'm missing some of what you're saying, I think this is what
19 you're saying, John, is that the MOE, the REL is independent
20 of what's actually out in the environment. It's something
21 that's a pure figure that people have worked through and
22 working through statistics and studies, whereas MOE of
23 course has exposure in it.
24 And I guess, am I correct in assuming this is what
25 you object to, essentially? Am I correct?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
120
1 And that -- and the difference between them
2 algebraically and I think that's obvious in your first
3 slide, is the MOE is equal to the REL times the factor times
4 UF over exposure. I mean, that's just algebraic.
5 So the answer to your question is you can always
6 calculate, if they use the same set of uncertainty factors.
7 I mean, I don't know, is there anything more
8 complicated?
9 I do I think agree with Dr. Froines in saying that
10 the REL is a much purer number in the sense that it doesn't
11 have the effect of exposure which can change, after all,
12 from area to area and from year to year.
13 DR. SCHREIDER: Correct.
14 DR. FUCALORO: Is that all there is to the
15 discussion or am I missing something? It seems --
16 CHAIRMAN FROINES: There's two parts of the
17 discussion, I think.
18 One part is let's assume we did what you said,
19 which is I think what we should do. I think we should take
20 the REL as -- you should establish an REL and then look at
21 your exposure as the second factor.
22 DR. FUCALORO: That's what I thought you said.
23 CHAIRMAN FROINES: Then the question becomes what
24 limit, what definitions do you put on your exposure and what
25 constitutes -- is an exposure equal to the REL, is that
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
121
1 considered safe, not safe. In other words, is there an
2 exposure uncertainty that we then apply?
3 They're using the 95 percent upper confidence
4 limit on the distribution and so one could argue that that's
5 a reasonable number and adds some conservatism and so one
6 could take the upper 95 percent confidence limit as the
7 exposure number and divide that into the REL, and if you
8 have one that would seem like a reasonable value.
9 And so but in each case all I'm saying is each
10 case you have explicitly defined what you're talking about.
11 Whereas here you're not. You've got this exposure
12 which actually includes their uncertainty factors built into
13 that, so it's not explicit.
14 DR. FUCALORO: When they get an MOE of 9600, you
15 have to ask yourself how does that compare to uncertainty
16 factors and I think that's therein lies a bit of confusion.
17 It's much cleaner to use the REL, since --
18 CHAIRMAN FROINES: We take the current, we take
19 the current NOAEL for DEF and go through the numbers, then
20 it turns out that the value that you calculate is 20. And
21 then you have to make a judgment about how important you
22 think 20 is.
23 DR. BYUS: They just corrected it. I kept looking
24 at the numbers. I still don't get it. I was going to ask
25 you --
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
122
1 CHAIRMAN FROINES: They add up now. The number
2 should be -- the denominator should be 303.5, right?
3 DR. GLANTZ: Nanograms.
4 DR. BYUS: I thought there was another --
5 CHAIRMAN FROINES: So go ahead. We're now
6 disrupting your presentation.
7 DR. SCHREIDER: Not at all.
8 When we calculate the margin of exposure that does
9 not in itself, that is in effect then compared to what would
10 be the appropriate uncertainty factor or total safety
11 factor. That's done in the risk appraisal section where,
12 depending on what words were used, it may be that because
13 this was derived from a low effect level, not a no effect
14 level, we should have an uncertainty factor that would
15 normally be a benchmark of 10-fold for LOEL to NOEL. Those
16 are the sorts of things that would go into a verbal
17 description in the risk appraisal section and, again, that
18 typically was from what we were asked to do for overall
19 assessments of the chemical is the exposure to this chemical
20 acceptable or not and to give information to risk management
21 to make that decision.
22 MR. GOSSELIN: We can't hear anybody.
23 CHAIRMAN FROINES: Can you get closer to the
24 microphone.
25 DR. SCHREIDER: Can you hear me now, Paul?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
123
1 MR. GOSSELIN: Yes. I don't know if people aren't
2 speaking into the mikes or not.
3 DR. SCHREIDER: I'm now speaking into it.
4 And I think the other point is that in the case of
5 the air toxics that we're moving in the direction of
6 calculating the RELs as a reference concentrations. That's
7 going to be a standard practice.
8 And, Paul, you can correct me if I'm wrong.
9 Paul?
10 MR. GOSSELIN: Yes.
11 DR. SCHREIDER: I was indicating that we are in
12 fact for the air toxic compounds going to be generating
13 reference concentrations or RELs.
14 MR. GOSSELIN: That is correct.
15 DR. SCHREIDER: So that that will be generated in
16 addition to the margin of exposure calculation and part of
17 that is so that we can then combine the exposure from all
18 the routes of exposure, whether it's going to be inhalation
19 in the case of the air toxics, but there may also be
20 occupational, depending on how the pesticide is used around
21 the home, maybe dermal exposure, so that those can be
22 combined to try and determine overall margin of exposure to
23 that compound.
24 So to that extent, the reference concentration
25 would not be applicable -- or be particularly suited to that
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
124
1 overall concentration or the overall calculation of what the
2 overall exposure was.
3 So I think we are going to be moving in the
4 direction and we'll be including both values.
5 DR. MARTY: I think I need to add something to the
6 discussion that might clear it up a little more.
7 The way we use a reference exposure level in
8 site-specific risk assessment in the hot spots program, that
9 reference exposure level is compared to the modeled ground
10 level concentrations that are predicted using air dispersion
11 models and information about the facilities' emissions.
12 The ratio of the ground level concentration to the
13 reference exposure level is called the hazard index for that
14 chemical.
15 If the hazard index is one or higher, it triggers
16 risk management in the hot spots program.
17 So that would almost be equivalent to an MOE of
18 100 or 1,000, depending on the uncertainty factors that were
19 put into the reference exposure level.
20 CHAIRMAN FROINES: I missed something. I didn't
21 get what you said.
22 DR. GLANTZ: I got it. We'll explain it to you
23 later, John.
24 DR. FUCALORO: We're getting cruel now.
25 DR. BYUS: It's the coffee.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
125
1 CHAIRMAN FROINES: Coffee, and people are getting
2 hungry.
3 DR. ALEXEEFF: George Alexeeff.
4 Let me explain it a little bit differently.
5 DR. FUCALORO: George always fills the breach.
6 DR. ALEXEEFF: When we are developing a reference
7 exposure level, we're developing a concentration with the
8 uncertainty factors built in. So it's simply a risk
9 assessment value. And we are assuming that the uncertainty
10 factors we apply are appropriate, that there are risk
11 assessment issues issued on variability or taking into
12 account subpopulations.
13 So our reference exposure level is calculated as a
14 risk assessment value.
15 That value is then used by a risk manager, let's
16 say, they look at their exposure, and then they look at the
17 ratio of that. If the ratio is greater than one, that is to
18 say the REL over the exposure, if it's greater than one --
19 actually it's flipped. You're right. If the ground level
20 exposure over the reference exposure, if it's greater than
21 one, that means your exposure is exceeding the reference
22 concentration, and then the risk managers have to think
23 about what they -- if they need to do anything.
24 So that is analogous to the MOE.
25 What my understanding of the MOE, and this might
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
126
1 help in the discussion, is that it's basically a tool more
2 for the risk manager where the REL is basically a risk
3 assessment product, a risk assessment output. And I think
4 that's kind of one difference.
5 They're setting up a system so the risk manager,
6 as I said, DPR can make a decision as to whether it's a
7 toxic air contaminant or there's some management issue to
8 do.
9 I don't know if there's -- as opposed to MOE
10 actually being a discrete risk assessment issue.
11 CHAIRMAN FROINES: I have two questions.
12 One of which is if, George, if your ratio was .1,
13 then would the risk managers worry about that or would that
14 be basically he would, he or she would consider it
15 negligible risk? In other words, where does the negligible
16 risk enter?
17 DR. ALEXEEFF: They would consider it negligible
18 risk.
19 Usually it's under one, and depending upon the air
20 districts, how far below one is kind of their decision.
21 It's a little bit on the accuracy of how they measure. Some
22 say .5, some might be .99. I'm not sure. Depends on the
23 different districts.
24 If it's below one, it's considered negligible.
25 CHAIRMAN FROINES: Let me make my second point,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
127
1 because -- that was just bookkeeping.
2 The more important point is this panel has
3 historically disagreed with DPR on the determination of a
4 compound as a toxic air contaminant. That is, we've always
5 believed a chemical should be determined a toxic air
6 contaminant irrespective of the level of exposure. And
7 that's a fundamental disagreement which got us into some
8 very contentious discussions in the '80s, and Stan remembers
9 them well.
10 DR. GLANTZ: Just for the record, I was just
11 observing John being contentious.
12 CHAIRMAN FROINES: That's the -- so there is a
13 major underlying history and disagreement.
14 Now we're into the situation where we're not
15 debating that issue for at least for the moment, but the MOE
16 comes in.
17 And I think that this discussion comes up because
18 of this notion that once you get into the exposure as being
19 determining the determination of whether this is a toxic air
20 contaminant, then I still would feel better if we had an
21 explicit risk assessment process that then the exposure
22 determination became a next step in.
23 DR. GLANTZ: Well --
24 CHAIRMAN FROINES: And so I don't agree with the
25 notion of having it be essentially the third piece. What
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
128
1 you have is the risk assessment, the exposure assessment and
2 then this hazard characterization, and then the risk
3 appraisal. And so that the uncertainty factors get brought
4 in your process as the risk appraisal, which is at the very
5 end of this document, and that's the problem.
6 DR. BYUS: I agree. I didn't realize it at all
7 until you were just going over it.
8 I'm not sure what the MOE real advantage of it.
9 Who uses the MOE?
10 CHAIRMAN FROINES: US EPA.
11 DR. SCHREIDER: US EPA.
12 DR. GLANTZ: This is one more way that we can do
13 better than they do.
14 FROM THE AUDIENCE: I think I know --
15 CHAIRMAN FROINES: Excuse me. I think that what
16 should happen is if your leadership wants you to come up and
17 speak to it, he should have you come up, but I'd rather you
18 don't call up and come walking up.
19 FROM THE AUDIENCE: The MOE is based on 100, not a
20 one. That's billion in 100, MOE 100 acceptable.
21 DR. GLANTZ: Well --
22 CHAIRMAN FROINES: That didn't help.
23 Dale, what you were going to say?
24 DR. HATTIS: I think that part of the difference
25 is in fact, I think, perceptual. And the very fact of
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
129
1 having a big number being exhibited as the difference
2 between the exposure that's okay or not okay in the exposure
3 that exists, has a policy implication, to put it in the most
4 neutral way I can.
5 And that one of the ways in which this can be
6 confusing to people is if in fact some of that number is in
7 fact what should really be an adjustment factor for real
8 average differences between the exposure circumstance in the
9 experimental animals and exposure circumstances --
10 CHAIRMAN FROINES: I think that's particularly
11 important to the lay public when you take -- when they're
12 looking at this thing and they see something called a no
13 effect level, a no effect level, and then you divide it by
14 some exposure and you get 10,000, the average layperson
15 would say let's go home and forget this chemical.
16 And that perception is a problem, because the
17 reader is first seen as the no effect level, well, if you're
18 10,000 a no effect level, you must be safe. So that the
19 person who doesn't understand some of these more complicated
20 elements is going to have a very different perception and
21 it's going to create a level of belief that the compounds
22 are safe, that may or may not be entirely accurate.
23 DR. FUCALORO: To beat a dead horse of course, a
24 MOE of a million does not tell you whether or not the
25 compound is very toxic. I mean, it doesn't matter. Really
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
130
1 the REL tells you how toxic it is and then you look at
2 ambient concentrations and see if you want to control this.
3 I mean, we're saying the same thing, the algebra
4 is very simple. We're just wondering, I guess, as a panel,
5 I don't know how much influence we have, whether or not
6 documents can be recast in such a way as to introduce
7 something like REL and in order to talk about the inherent
8 toxicity of a material, rather than looking at the effect on
9 the population, given some ambient concentration that people
10 think we have in California.
11 MR. GOSSELIN: Might I interject? This is Paul
12 Gosselin. Can everyone hear me?
13 DR. FUCALORO: We sure miss you.
14 MR. GOSSELIN: Miss you too. I'd rather be there
15 than here.
16 You know, I caught the tail end of this and I
17 think Jay said it's our intent to have our documents be in a
18 format that OEHHA -- you're used to seeing from OEHHA.
19 So that's the direction we're moving in. It's our
20 intent to recraft our documents in more of an REL type
21 presentation versus an MOE.
22 CHAIRMAN FROINES: Stan.
23 DR. GLANTZ: I think that's wonderful.
24 What I would -- what I'd like to suggest we do,
25 just to formalize this, is that we say to DPR that we like
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
131
1 the documents to come forward with the REL in them, and also
2 since you, in order to better express what we know about the
3 exposure compared to the REL is also include the hazard
4 index, which OEHHA uses for -- in their things.
5 And then if you want to also put the MOE in
6 because you like it, I don't mind if it's there.
7 But I think what -- so I would suggest that the
8 panel actually say that's how we want the documents
9 presented, which I don't think is that different from what
10 you're saying, Paul.
11 MR. GOSSELIN: No. I would actually say that if
12 you want to move to endorse our efforts to do that --
13 DR. GLANTZ: Okay.
14 DR. FUCALORO: I second.
15 DR. GLANTZ: I so move. And Fucaloro seconded.
16 CHAIRMAN FROINES: Do you want language or --
17 DR. GLANTZ: I think I'll make a motion, if you'd
18 like a motion, and that is that we direct or that the
19 pesticide documents have an REL presented in them, a hazard
20 index to provide a better measure of the risk, the magnitude
21 of the risk, and in also the controlling for exposure, and
22 also that OEHHA and DPR use the same uncertainty factors,
23 same default uncertainty factors, in that analysis, so that
24 we will standardize things.
25 Obviously, if there's -- that's the end of my
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
132
1 motion.
2 And then I'll just add the comment that if there
3 are data that would lead you to something beside the default
4 uncertainty factors, that obviously you use the data driven
5 uncertainty factors.
6 But so that's the motion I'm making.
7 DR. FUCALORO: My second stands.
8 CHAIRMAN FROINES: So I think, let's just take --
9 is there any discussion?
10 MR. GOSSELIN: I would add that the uncertainty
11 factor issue, suggest that it might be taken out, because
12 one of the things we do when we go to a peer review with
13 OEHHA is that those issues could and should come up during
14 that process, and that if we do have legitimate reasons to
15 be different, we can explain that in our comments and our
16 report back.
17 But I think that categorically taking a blanket
18 statement that the two institutions are going to immediately
19 bridge scientific perspectives, I think it's probably kind
20 of a leap.
21 DR. GLANTZ: Well, I don't agree. Well, that's
22 okay, jump.
23 DR. WITSCHI: Do we have a discussion by a
24 non-panel member on a motion that's made by the panel?
25 DR. GLANTZ: Okay. By a what?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
133
1 DR. BLANC: It's useful to -- Paul Blanc here. I
2 think it's useful to hear Paul's perspective on this, but
3 now that we're heard that perspective, I think it would be
4 useful to hear from the panel members, and useful for you to
5 hear from the panel members their take on that.
6 I think what you're going to hear, though, is from
7 people that the consensus here is that OEHHA and your agency
8 should not come to us with different uncertainty factors for
9 the same chemical.
10 I think that OEHHA's already stated that there
11 will be some variability within their own agency on certain
12 uncertainty factor adjustments, varying between three and
13 ten, depending on the chemicals. So that does leave you
14 some leeway, but I don't want to be in a situation, and I
15 think the other panel members feel the same way, that for
16 the same chemical, let's say for some reason you were coming
17 to us with a document related to chlorine as a biocide, we
18 wouldn't want to have you using an uncertainty of three and
19 OEHHA using an uncertainty factor of ten for that same
20 chemical. And similarly if the range of uncertainty factor
21 from OEHHA varies from three to ten for sensitive
22 populations, let's say, we wouldn't want you to come in with
23 a one that had adjustment of two or 1.5.
24 CHAIRMAN FROINES: If one can define a reason why
25 an uncertainty factor should differ.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
134
1 DR. BLANC: Within that range.
2 CHAIRMAN FROINES: Within that range, then they
3 could petition and give a scientific argument.
4 But I think that the general principle is that
5 reasonable people should reasonably agree on uncertainty
6 factors.
7 DR. GLANTZ: Yeah. I think we spent a lot of time
8 discussing this issue in the context of the acute reference
9 exposure document. And I'm pretty comfortable with the
10 uncertainty factors that we came up -- the default, I want
11 to keep saying default, uncertainty factors that we came up
12 with there.
13 And I mean the numbers which were presented here
14 by DPR as the defaults are in fact the same ones that OEHHA
15 is using anyway. So I don't think that should be a
16 controversial point.
17 Again, obviously if there's data that would lead
18 you to conclude that one -- that you shouldn't use the
19 default number, then you should use the data. The data is
20 always better than the default.
21 CHAIRMAN FROINES: This is -- but Stan's right,
22 that's a good point, because where -- you know where our
23 history is. The studies that DPR uses are often these
24 industry studies that have a kind of mysterious quality to
25 them because we never see them and so they --
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
135
1 DR. BYUS: We can see them if we want to.
2 CHAIRMAN FROINES: I know. If we go to the
3 library, I understand.
4 But the point is that DPR may, based on those
5 studies, decide that they want to establish a NOEL based on
6 study X from Eastman Kodak, for example.
7 George has been working with data on the same
8 chemical, but he's been doing it all with the peer reviewed
9 literature, for example, just as a hypothetical situation.
10 He decides that his NOEL is based on a different
11 set of data than what DPR decides to do.
12 But it seems to me if that's the case, then we
13 really do need to work to get the two groups reading the
14 same studies to make -- and they shouldn't be making
15 different determinations based on different scientific
16 studies, which means that what they decide to use for the
17 NOEL should be reasonably consistent or else we ought to
18 find out why it's not consistent.
19 Does that make sense?
20 MR. GOSSELIN: I agree with that.
21 DR. FUCALORO: Let's call the question.
22 CHAIRMAN FROINES: So we'll call the question, all
23 in favor raise your hands.
24 (Panel members raise hands.)
25 CHAIRMAN FROINES: So that passes.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
136
1 DR. GLANTZ: Unanimously.
2 CHAIRMAN FROINES: Unanimously.
3 Now, there's one other issue that we haven't
4 talked about, but I think we're not going to talk about
5 today, but it is really an issue of major magnitude.
6 DR. GLANTZ: Before you do that, just to pound one
7 last nail into this coffin, I don't think this is
8 controversial, but I would appreciate it if the OEHHA and
9 the DPR people would just get together and look over the
10 default uncertainty factors and the definitions and all that
11 stuff that we spent so much time talking about in the
12 context of the REL document, and just come back at the next
13 meeting and just say to us, we looked at them, and we are
14 agreed on the definitions and we're agreed on the
15 uncertainty factors.
16 I mean, that's what you've already said, but I
17 just want to make absolutely sure that nothing is falling
18 through the cracks.
19 It should be a short conversation, but I just want
20 to be absolutely positive that everybody is on the same
21 page.
22 If you just check, if you guys could just get
23 together, look over that and come back to the next meeting
24 and say, yeah, we did that, and everything is under control.
25 I think it is.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
137
1 DR. ALEXEEFF: We can certainly do that with
2 regard to the, let's say, the acute document, which has
3 already been processed through.
4 I don't know what we do if we have a difference,
5 but let's presume we don't.
6 But then the chronic document will be different.
7 There's additional uncertainty factors in the chronic
8 exposure issues. So we'd almost have to do it again at that
9 point or --
10 DR. GLANTZ: Well, I think then you should,
11 because what I want to see happen is a standard protocol
12 that everybody is using. And so if we need to revisit that
13 in the chronic document, we should. And I would hope in
14 developing that that you'll talk to DPR so that when that
15 comes forward, any issues will have been resolved by the
16 time it comes to us. And if they haven't been, then we can
17 help resolve them.
18 DR. ALEXEEFF: And then along similar lines, the
19 DPR develops seasonal values, which we don't develop, so
20 that would be some discussion of that at some point.
21 CHAIRMAN FROINES: That's what I'm coming to.
22 DR. ALEXEEFF: Three different types of values
23 that we develop, and depending on some slight modifications
24 of those uncertainty factors in those cases.
25 MR. GOSSELIN: I was raising an invisible hand
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
138
1 here.
2 One thing, a couple things is that I agree with
3 Stan on the point, because I, you know, I walked away after
4 the discussion on the acute documents and my staff reviewed
5 it, and largely acknowledged how well written the document
6 is, how clear, and actually pointed to a need for us at DPR
7 to go back, take a look at what our existing policies are
8 and actually, in light of the acute documents, start to
9 document what our policies are and start to follow down
10 similar paths.
11 I think if particularly since both agencies are
12 going to be dealing with the panel, if there's a process to
13 prepare similar document on our chronic endpoints, I would
14 offer that we'd like to get involved with that in tandem
15 with OEHHA early on, and with the panel, to participate in
16 that.
17 But I think one of the things I walked away from
18 the acute document is that we're going to be -- we are
19 actually going to start internally to go through a process
20 of identifying what policies we need to revisit and sort of
21 recraft and then start sending those out to OEHHA, the
22 panel, and getting peer review on them before we finalize
23 them.
24 CHAIRMAN FROINES: Good. Thanks, Paul. That's
25 very helpful.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
139
1 The issue I want to comment about that the final
2 issue about the MOE that is in many respects by far the most
3 crucial, because coming up with a NOAEL is not an easy task,
4 but it's certainly a doable task, but I think that the
5 actual determination of, quote, what we mean by exposure is
6 a major task that's very difficult.
7 And we have talked about it, Lyn Baker is here,
8 and we've talked about it in the past. Roger Atkinson is
9 now on the committee and hasn't been part of all of those
10 discussions, but I think it will be interesting when we take
11 it up more.
12 And so I think that one of the key issues is going
13 to be how we actually define protocols for doing exposure
14 measurements, what are we trying to measure, are we trying
15 to do dermal, are we trying to measure metabolite, I mean,
16 rather breakdown products, air chemistry. What are we
17 actually trying to measure and coming up with these exposure
18 values.
19 And so I make that as a major issue which we need
20 to develop speakers and actually have probably a half day
21 session talking about the whole issue of exposure.
22 And so we'll defer it for now, but I think it's a
23 fundamental issue that we have to deal with on this one.
24 So let's go over, if there's nothing else more on
25 this, we can -- I just want to bring you up to date.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
140
1 Does everybody have the DEF findings now?
2 These findings have taken a while and they've
3 taken a while because we were cavalier at a meeting not long
4 ago where we went through, and you can tell everybody had
5 had a lot of coffee, because everybody was jumping -- if you
6 look at the transcript, everybody was jumping into the
7 discussion and people held positions very strongly and
8 wanted changes made and so on and so forth.
9 And then that -- but we didn't write everything
10 down that everybody was saying, so we then had to go back to
11 the transcript and went through the transcript to see what
12 the changes in fact were, and when we looked at the
13 transcript to make the changes, we realized that there were
14 enormous inconsistencies in what we had done, and especially
15 in relationship to the document itself.
16 Well, I think that we have basically addressed all
17 the inconsistencies and all the problems, and so I think the
18 draft that you have is one that we would propose go forward.
19 I should say that a very major series of
20 discussions occurred because what we do at these meetings is
21 we always work with OEHHA and say -- and on lead Stan went
22 and took about three hours going through line by line
23 recommending changes. If everybody remembers that.
24 And OEHHA sat there and made the notes down and
25 they then dutiful went out and made the changes and seemed
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
141
1 at least in appearance's sake happy to do that. We never
2 know what was inside their head, but they were smiling.
3 DR. GLANTZ: At least some of them were happy.
4 CHAIRMAN FROINES: Some of them were.
5 And they went and made the lead changes.
6 DR. FUCALORO: You should see the little dolls
7 they have of you with pins in it.
8 CHAIRMAN FROINES: But we moved so quickly on the
9 DEF document that the transcript reads that we want to do
10 this and we want to do this and we're going do this and
11 we're going to do this, and no place in that transcript does
12 it ever -- does Paul Gosselin's voice ever come in and say,
13 okay, we'll do that.
14 So we left with the situation where DPR really
15 hadn't been part of the discussion in a way that OEHHA is
16 normally part of the discussion when we're making changes.
17 And I think one rule out of that is when we go
18 through and make changes, we have to get agreement on the
19 part of the agency as we go through just to make sure that
20 everybody is on the same page. And if Paul or other
21 representatives have major problems, that's when they should
22 state them.
23 So DPR then had to consider whether they wanted to
24 make the changes while I was rewriting the findings.
25 For example, on the designation of DEF as an
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
142
1 NOAEL, as the values for the NOAEL, and so on and so forth,
2 and Paul and I finally resolved it yesterday at 7:30 in the
3 morning, and he agreed and DPR agreed, that those changes
4 would be acceptable to DPR, and they will go back to their
5 document and make the changes that now are -- that will make
6 their document consistent with our findings.
7 So that's the outcome. It's, I think, a very good
8 outcome. It avoided a lot of potential controversy.
9 So they feel that they were able to make those
10 changes and they felt comfortable with them as a matter of
11 science.
12 Paul, did you hear what I said?
13 MR. GOSSELIN: Yeah. Actually, I got called out
14 on a phone, but I heard the last couple minutes and, yeah, I
15 largely agree. I think part of it too is that I think over
16 the past year just the rebuilding of a relationship between
17 us and the panel, you know, on better communication of
18 getting at least through the DEF document process.
19 CHAIRMAN FROINES: Okay. So I think these
20 findings, as they currently exist, are acceptable to DPR,
21 are acceptable to the panel, I mean, except people are
22 welcome to make suggested changes, but this is what will go
23 forward. And I think the process worked out, although it
24 wasn't as easy as everybody thought it would be.
25 There is one thing I do want to say, everybody now
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
143
1 has the cholinesterase document, which I downloaded.
2 DR. GLANTZ: Just a point. Do you want us to vote
3 on these or anything?
4 CHAIRMAN FROINES: Yeah. I'll come back to that.
5 Let me make one comment.
6 I think one thing that is very important is that
7 we made a decision on defining a number, and that number we
8 defined as a no observable adverse effect level for DEF.
9 I do want to make the point that from a matter of
10 science that we did not define policy on blood plasma or RBC
11 cholinesterase. We said that in the case of DEF that the
12 plasma and RBC cholinesterase is a reasonable surrogate for
13 an adverse effect. That's the decision we made.
14 But I just want to emphasize that we'll have to
15 make that decision on every chemical that comes up, because
16 it's not going to be true that plasma cholinesterase is an
17 adverse effect in every case. It's going to be defined by
18 the toxicokinetics of the compound for the most part.
19 And I feel very strongly about that.
20 So we haven't made, as far as I'm concerned, we
21 haven't made a policy decision. We have made a
22 determination based on the science associated with this
23 particular pesticide, and the next pesticide that comes up,
24 we'll have to revisit the issue of plasma cholinesterase in
25 the context of looking at the science.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
144
1 And we're going to have to ask DPR to do a much
2 better job in the future in developing the toxicokinetics of
3 these pesticides, because they are going to differ very much
4 in terms of whether or not they're reaching the brain,
5 whether or not they are having an effect in the brain, and
6 whether what you measure the plasma has anything to do with
7 what's happening in the rest of the body.
8 I'm waiting for Paul, who's just got a --
9 DR. BYUS: I'm not sure I agree with John.
10 DR. BLANC: Is there some reason that you'd rather
11 not share that you're making this statement? Am I missing
12 something?
13 CHAIRMAN FROINES: About what?
14 DR. BLANC: Why you're making that statement.
15 CHAIRMAN FROINES: I just want to make clear that
16 we, at this point, have not made a policy decision about
17 blood cholinesterase as an adverse effect, that we've made a
18 decision about blood cholinesterase relative to DEF.
19 DR. BLANC: You think it has implications for how
20 we might approach other --
21 CHAIRMAN FROINES: No, I personally think that
22 every chemical is going to be different. And some
23 chemicals, a plasma cholinesterase inhibition may have no
24 significance whatsoever and in some cases it may be an
25 adverse effect, but I think you can only determine that by
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
145
1 looking at the science associated with the determination.
2 DR. FUCALORO: Didn't Hanspeter point out the
3 problem with plasma cholinesterase two times ago saying that
4 it's a -- I hate to use the term -- generic effect, that
5 regardless of how it's depleted it can have, for sensitive
6 people, some very negative health effects.
7 I'm not a toxicologist. I just -- you were
8 talking about the surgery or something.
9 DR. WITSCHI: This was a famous case was king of
10 Morocco in '56 or something like this, who didn't tolerate
11 the anesthetic for a trivial operation because he had a
12 deficient plasma cholinesterase and he never woke up out of
13 the narcosis, and that's how people became aware.
14 It also happened in Switzerland, that's why I
15 know.
16 CHAIRMAN FROINES: Well --
17 DR. BLANC: We'll revisit that. I just was
18 curious if there was something I was supposed to be getting
19 that I wasn't getting.
20 CHAIRMAN FROINES: No. I just think that we may
21 find that clearly a plasma cholinesterase is a surrogate for
22 something else going on and in a quantitative sense you can
23 have lots of plasma cholinesterase inhibition going on
24 without necessarily brain cholinesterase being --
25 DR. BLANC: I'll just say that, John, because most
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
146
1 cholinesterase inhibitors in commercial use are lipid
2 soluble, I think that it would be the unusual situation I
3 think that you're describing, and that I do view what we did
4 as precedent setting, and I do think that DPR should not get
5 the misimpression that cholinesterase inhibition is going to
6 be presumed benign until proven otherwise.
7 So, clearly, every time that we take a decision it
8 is not a blanket policy for all chemicals.
9 On the other hand, it would be, I think,
10 misinterpreting the thrust of this group were DPR to somehow
11 think that this was an aberrancy in terms of cholinesterase
12 inhibition.
13 CHAIRMAN FROINES: I think that's fine. All I'm
14 saying is that I think as a matter of policy we haven't made
15 that decision.
16 DR. BLANC: That's always true, isn't it?
17 CHAIRMAN FROINES: No. We could define policy if
18 we chose to.
19 DR. BLANC: But we didn't.
20 CHAIRMAN FROINES: But we haven't though.
21 DR. BLANC: No.
22 CHAIRMAN FROINES: I sat down with eight
23 toxicologists and pharmacologists and discussed this at UCLA
24 recently and there wasn't a single person who in giving some
25 examples from some certain organophosphates would have
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
147
1 called that an adverse effect whatsoever.
2 So I think that there's very strong scientific
3 points of view that may differ very widely on this issue.
4 I think we have to look at it as the issues come
5 forward to us.
6 And everybody may vote to call it an adverse
7 effect, but I think we have to look at it as an individual
8 situation rather than a generic situation.
9 DR. BLANC: That's fine. I'm just saying that in
10 general it's going to be the burden of argument is going to
11 be -- have to be that it's not an adverse effect, in
12 general. It's a very suspect outcome variable.
13 CHAIRMAN FROINES: Yeah, I agree.
14 And I also think that what you find in the plasma
15 is different than what you find in the red blood cell and
16 that raises some other issues. So that we'll have to talk
17 about the whole ball of wax as it comes forward.
18 Finally, the pesticide documents.
19 DR. FRIEDMAN: Do you want discussion of the DEF?
20 CHAIRMAN FROINES: Oh, yes.
21 DR. FRIEDMAN: I obviously have not read the whole
22 thing, since we just received it, but I feel there's
23 something explicit missing from the conclusions.
24 If you read 31 and 32 on page five, it says for
25 something to be a toxic endpoint, pesticides with risk
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
148
1 greater than ten to the minus seven should be identified as
2 such, and then it just in 32 it gives the levels which are
3 obviously greater than ten to the minus seven, or some of
4 them.
5 But I'd like to see that 32 have a statement,
6 therefore DEF is a toxic air contaminant. I don't think we
7 ever say that explicitly.
8 DR. GLANTZ: That's on No. 34.
9 DR. FRIEDMAN: I didn't see that. Sorry. I
10 missed that.
11 DR. GLANTZ: I move we adopt them.
12 CHAIRMAN FROINES: I thought there was one other
13 question somebody had. I thought Peter had a question.
14 DR. WITSCHI: No.
15 CHAIRMAN FROINES: You're okay.
16 DR. WITSCHI: Okay.
17 DR. FRIEDMAN: I just feel a little uncomfortable
18 getting this document and saying -- with no chance to look
19 at and saying we should adopt it. Could you just at least
20 run through the major changes that were made?
21 CHAIRMAN FROINES: Sure. No problem at all.
22 The reason you're just getting it is we just
23 resolved it and everybody -- the vote has been taken on this
24 issue. So what you're really saying is what you'd like to
25 know is --
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
149
1 DR. BLANC: Does this reflect all the comments as
2 you best as you could adjudicate.
3 CHAIRMAN FROINES: Yeah.
4 The important factor is originally the major
5 difference -- no, there are two major differences.
6 One is Roger Atkinson was quite concerned with
7 discussion of a number of elements of it, and particularly
8 around air chemistry and he agreed to provide information to
9 DPR that they could incorporate in their document and we can
10 make that available what he wrote to you. It all
11 strengthens the document.
12 DR. FRIEDMAN: Where does it appear in the
13 document or what in the document was changed?
14 CHAIRMAN FROINES: Going to go in this document.
15 DR. FRIEDMAN: I see. It isn't in --
16 CHAIRMAN FROINES: No.
17 DR. FRIEDMAN: Okay.
18 CHAIRMAN FROINES: And so the document -- so the
19 document will now have Roger's input.
20 The second major area was that the value -- look
21 under 18, originally DPR had proposed a NOEL of 12.2
22 milligrams per cubic meter.
23 We recommended that two changes. One that the
24 NOEL not be a NOEL, it be an NOAEL. And we also recommend
25 that that be 2.4 milligram per cubic meter based on blood
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
150
1 cholinesterase inhibition.
2 So if you'll look under 18, those values have been
3 incorporated. So now the report identified an acute air
4 NOAEL of 2.4 milligrams per cubic meter. So that represents
5 the major change that occurred.
6 Then there was discussion about the NTE, the
7 neurotoxic esterase, which -- and Paul was commenting on
8 that and I was. And basically we've changed that so that it
9 now states the inhibition of NTE in sensitive species is a
10 biomarker that correlates with the induction of OPIDN, you
11 know, delayed neuropathy. So that change has been made.
12 And we've taken out the sort of the implication
13 that NTE is a mechanistic finding rather than a biomarker.
14 And I think that's important.
15 Then one of the interesting things that happened
16 was under 26 we added a sentence which says the SRP requests
17 DPR and OEHHA to evaluate the hen delayed neurotoxicity
18 model and determine how it may be used to perform
19 quantitative risk assessments, because as of now we don't
20 have the -- we have not incorporated the NOAEL from the hen
21 feeding study into the document as a NOAEL. And the
22 question is how might we do that.
23 And that's important also, because this issue is
24 going to come up again and again to the degree that we deal
25 with compounds that display a delayed neuropathy.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
151
1 And finally on 27 I took out an entire paragraph
2 and just put this in, which I think is very important. And
3 that reads the NOAEL of .4 milligram per kilogram day, based
4 on brain cholinesterase inhibition for the rat oral toxicity
5 study, is close to the NOAEL of 0.6 milligrams per kilogram
6 day based on blood cholinesterase inhibition for the rat
7 inhalation study.
8 I take that as being a very important sentence,
9 precisely because it shows that there is a quantitative
10 relationship between the brain cholinesterase and the blood
11 cholinesterase inhibition. And so that the dose response
12 for the brain and the dose response for the blood are in the
13 same ballpark, and so one feels more comfortable defining
14 the NOELs and NOAEL, given the proximity of those values.
15 So that's pretty much it. Those are the major
16 points.
17 So the document has now, as I say, it has a
18 significantly lower NOAEL and that changes the MOE to what I
19 calculate this morning as about MOE of 20, if you
20 incorporated the -- no, I think it is --
21 DR. BYUS: I have three at 7,000 for No. 21.
22 CHAIRMAN FROINES: What?
23 DR. BYUS: It's written down as 3,000 to 7,000,
24 the MOE. No. 21.
25 CHAIRMAN FROINES: Well, we'll have to change
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
152
1 those. I just didn't do that calculation. So I'll do the
2 calculation and we'll change it.
3 DR. FUCALORO: Just two things.
4 First, in item 32 is just a repeat of 23, and
5 that's what you intended, is that correct? It's a
6 conclusion and really it's just the same, identical.
7 And the other is the MOEs that you're prepared to
8 change, I'm not quite sure I know why. Maybe you should
9 talk about it.
10 CHAIRMAN FROINES: The MOE of -- let's see here.
11 In the document the MOE -- well, let me go back and check,
12 because in the document the MOEs range from 9600 to much
13 higher values. So that 3,000 to 7,000 may reflect the new
14 values, and I'll just double check that.
15 When I say it gets down to 20, that's if you
16 included a 100-fold uncertainty factor in the calculation.
17 DR. FRIEDMAN: Is there some reason that the
18 complete was crossed out, the bottom of 33?
19 CHAIRMAN FROINES: Yeah. I feel -- this is my
20 judgment call. I feel that this document still has
21 significant limitations, that if we could it would be good
22 to see a rewritten document that was organized differently
23 that dealt with some of these issues like the hen values.
24 And I think there's things missing from this document. And
25 we're not going to be able to correct them and we'd like to
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
153
1 move forward on this.
2 So I felt that the panel's finding would be a more
3 accurate statement of the way we think about this if we said
4 that the document represents a balanced assessment of our
5 current scientific understanding, because I frankly don't
6 think this document represents a complete assessment of our
7 scientific understanding.
8 DR. FUCALORO: And I would say just one more
9 thing, John, I don't mean to harp on this.
10 CHAIRMAN FROINES: No, no, it's fine.
11 DR. FUCALORO: I guess I actually do mean to harp
12 on it.
13 In item 21 where you're thinking of incorporating
14 uncertainty factors into their computed MOEs, that really
15 lends confusion, because the MOEs are not defined with
16 uncertainty factors.
17 CHAIRMAN FROINES: I'm just going to check these
18 numbers to see if they're okay.
19 DR. FUCALORO: Okay. We may want to put in RELs,
20 which are fine with me, I have no problem, and they're
21 purely computational.
22 DR. GLANTZ: The RELs are in the findings.
23 DR. FUCALORO: They're in the findings, right.
24 CHAIRMAN FROINES: No, Gary is exactly right, and
25 Tony is. Any changes are -- this is your document. So I'm
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
154
1 not trying to speed us through.
2 DR. BYUS: I just have one question. It says in
3 No. 20, so the NOAELs include an uncertainty factor. Is
4 that -- this is what it says, the NOAEL, it incorporates
5 100-fold uncertainty factor to address potentially increased
6 sensitivity in humans, and then we have the MOE, which is
7 the NOAEL divided by the exposure.
8 Does the NOAEL include the uncertainty factors or
9 not? Someone care to answer me?
10 CHAIRMAN FROINES: This document, what are you
11 referring to? 20?
12 DR. BYUS: I'm referring to 20, yeah. It says the
13 NOAEL value is the air concentration below which there is no
14 anticipated health risk. It incorporates a 100-fold
15 uncertainty.
16 CHAIRMAN FROINES: I think that this is a good
17 point. I think what you mean here is that the it actually
18 refers to the air concentration level. Am I not correct?
19 DR. SCHREIDER: The reference concentration level
20 incorporates the 100-fold --
21 CHAIRMAN FROINES: That's right. The NOAEL
22 doesn't. And you're absolutely right the it is referring --
23 I'll correct it.
24 DR. FUCALORO: It's referring to the REL, right?
25 CHAIRMAN FROINES: It's referring to the air
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
155
1 concentration level.
2 DR. FUCALORO: The REL.
3 CHAIRMAN FROINES: No, it's referring to the air
4 concentration level.
5 DR. FUCALORO: Which is the 8.8 micrograms per
6 meter cubed?
7 CHAIRMAN FROINES: Listen, it says an air
8 concentration --
9 DR. FUCALORO: I see, I see, I see.
10 CHAIRMAN FROINES: Then you've got that statement
11 about the REL in parentheses, so that the it below is
12 referring to the air concentration level.
13 DR. BYUS: That's --
14 CHAIRMAN FROINES: We'll clean it up.
15 DR. FUCALORO: Equivalent to the usage to an OEHHA
16 REL.
17 I'm reading like mad.
18 DR. FRIEDMAN: So I understand that if we do vote
19 to adopt this that you will go through it word by word and
20 make the changes that incorporate some of the things that
21 have been brought up today or anything else that you might
22 find that's incorrect?
23 CHAIRMAN FROINES: Yeah. We'll correct any other
24 errors that we find, but it's better to find them now
25 because we won't have to go back.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
156
1 DR. FUCALORO: In any event, when you do clean it
2 up and if we do vote on it today, you will send us a final
3 copy so we can take a look at and see if there are any
4 problems.
5 CHAIRMAN FROINES: This is it. Let's do it today.
6 I don't want -- I'll send you a copy which has been sent
7 forward. I don't want -- do you guys realize how long we
8 have been pursuing this?
9 DR. BLANC: Also, how many hours you must have
10 spent to go through this thing. I'm sure it must have been
11 incredible.
12 CHAIRMAN FROINES: The problem with it was is when
13 we got the first findings, we got SRP findings that DPR had
14 helped craft and we had OEHHA findings and they were
15 completely different. So it's actually been a number of
16 different elements to it.
17 But so let's try and find everything we can right
18 now.
19 DR. ALEXEEFF: Dr. Froines.
20 CHAIRMAN FROINES: Yes.
21 DR. ALEXEEFF: George Alexeeff.
22 Just going back to Dr. Byus' comment, I think the
23 sentence he pointed out still needs to be corrected, because
24 it is -- I don't think -- I think you are referring to the
25 next sentence, which required clarification of the word it,
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
157
1 but the previous sentence is also incorrect. It needs to be
2 revised. So because the previous statement says the NOEL
3 value is the air concentration below which.
4 CHAIRMAN FROINES: You're right.
5 DR. ALEXEEFF: That has to be changed. It should
6 be --
7 CHAIRMAN FROINES: REL.
8 DR. ALEXEEFF: Yes. The REL or the air
9 concentration standard, whatever is being used.
10 DR. BYUS: I'm not sure what you mean. It doesn't
11 sound right.
12 My other question is about the MOEs on No. 21. Do
13 we want to say the MOEs are significantly above the target
14 value for prompting regulatory action? Is that a
15 informative statement, are we making that finding? Do we
16 think -- is the MOE is not much value in determining that
17 statement. That's what I'm asking.
18 DR. FUCALORO: Because the statement like that
19 implies we have some uncertainty factor in mind in order to
20 make that statement.
21 CHAIRMAN FROINES: We should take it out for
22 another reason. That's a risk management statement.
23 DR. BYUS: Take it out. That's not our finding.
24 CHAIRMAN FROINES: Our job is not to prompt -- our
25 job is to hopefully prompt regulatory action, but not to say
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
158
1 that we're prompting regulation. So that's good.
2 DR. FUCALORO: Let me ask what's on some people's
3 minds as they're furiously reading this. I read this
4 furiously too.
5 Do we -- you would like to get a motion and
6 passage of this document with all the changes noted prior to
7 my comments, and perhaps after my comments, but we don't
8 have any mechanism for mail ballot after seeing the final
9 document or maybe that is much too cumbersome.
10 CHAIRMAN FROINES: We can do whatever you like.
11 DR. FUCALORO: Or we can wait until next time.
12 CHAIRMAN FROINES: No, no, no.
13 DR. FUCALORO: You refuse to do that.
14 DR. GLANTZ: I'd like to suggest that maybe we
15 take a break, let people read it carefully for ten minutes
16 or something, and then make any final changes and vote on
17 it.
18 CHAIRMAN FROINES: Okay.
19 DR. GLANTZ: We should just stop and let everyone
20 go through it and just deal with it, because this has
21 dragged on for a very long time, and I think John has done a
22 very good job, but there may be a few dregs of things that
23 weren't edited properly, but I think just take, maybe take a
24 break, we've done this before, and let people just read
25 through it so they don't have to be listening to whatever
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
159
1 everyone else is trying to say, and then just go through it
2 and finish it.
3 CHAIRMAN FROINES: I would rather finish it today.
4 DR. FUCALORO: Okay. I'm not going to insist. I
5 just wanted to, if people were not speaking up I wanted --
6 DR. GLANTZ: That way people will be able to just
7 stop and --
8 DR. FUCALORO: I meant to crystallize a solution
9 to what we're going to do, and this seems reasonable to me.
10 CHAIRMAN FROINES: There's no -- these are -- I
11 marked four places where we made changes and all of them are
12 quite reasonable changes and they're changes we would like
13 to have made. So that there are -- and they're changes that
14 we clearly missed the last time we went through this
15 document.
16 DR. FUCALORO: Let me make one -- let me go one
17 step further.
18 I understand what you're saying, Stan, but are
19 people prepared to vote now? If people are prepared to vote
20 now, then let's scrap the break and get on and finish up the
21 meeting.
22 DR. FRIEDMAN: I think, may I propose that we vote
23 now, then you'll send us the final document with the
24 understanding that if there's some minor things in the final
25 thing that you send us, after you've made the recalculations
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
160
1 and corrections that you will --
2 CHAIRMAN FROINES: I'll go with one caveat. I'll
3 send you the document and we will give you a return date on
4 comments, which will be within 24 hours or --
5 DR. FUCALORO: Make it 48, but certainly a date
6 and time certain.
7 CHAIRMAN FROINES: No matter how many hours we
8 give you, you'll do them towards the end of those hours.
9 DR. FRIEDMAN: Sometimes we don't even get -- we
10 may not get the mailing. We may be out of the office for a
11 day.
12 CHAIRMAN FROINES: We'll fax them.
13 DR. FRIEDMAN: We may be out of the office for a
14 day or two.
15 DR. FUCALORO: There's natural variability in
16 human behavior.
17 CHAIRMAN FROINES: We've gone this far. We'll
18 give you 72 hours. At the end of 72 hours if you haven't
19 got them back, it's over. And that's a major compromise.
20 You guys all owe me now on the 72-hour time.
21 DR. BYUS: We owe you.
22 Could on 24, could we define what NOEL is as
23 opposed to NOAEL, just so that we're clear we mean NOEL?
24 DR. FUCALORO: One is Christmas.
25 DR. BYUS: Right, I know. We define NOAEL, we
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
161
1 should define NOEL.
2 DR. FUCALORO: Adverse and not adverse.
3 With that in mind, a motion to approve is in
4 order, Mr. Chairman?
5 I move to approve.
6 DR. GLANTZ: Second.
7 CHAIRMAN FROINES: All in favor?
8 You notice I didn't ask for any discussion.
9 (Panel members raise hands.)
10 DR. GLANTZ: Can we have 72 hours of discussion,
11 please?
12 DR. FUCALORO: By yourself.
13 CHAIRMAN FROINES: The final thing --
14 DR. GLANTZ: Just for the record it was a
15 unanimous vote.
16 CHAIRMAN FROINES: The final issue for us is to
17 hear from DPR about their schedule. And I actually have
18 three questions about that.
19 One, is I'd like to hear from DPR about their
20 schedule. I actually have three questions about that. One
21 is I'd like to hear from DPR about their schedule. I'd like
22 to hear from Melanie about her schedule. And I'd like to
23 know what we're doing in May.
24 MR. GOSSELIN: You want me to go first?
25 CHAIRMAN FROINES: Sure.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
162
1 MR. GOSSELIN: We were scheduled to bring methyl
2 parathion back in May. There's a question of staff on
3 actually having the document completely rewritten for final
4 consideration. We may still achieve that, but, if nothing
5 else, we're going to have the lion's share of the issues
6 that were presented during the workshop and the issues
7 raised by the panel to be able to be presented then and also
8 a recalculation of REL format and a more complete emphasis
9 on blood and brain cholinesterase inhibition, so we can at
10 least schedule the discussion of the methyl parathion
11 document for May.
12 CHAIRMAN FROINES: Okay.
13 DR. BYUS: Pardon me, Paul, this is Craig Byus.
14 We are going to present it in May? You are going
15 to do it in May? Is that what you said?
16 I just talked to Ruby yesterday and she thought it
17 was going to be difficult to get done in May.
18 MR. GOSSELIN: No. And she sent me an e-mail.
19 It's kind of a question as to whether it's going to be in a
20 final format. But I think even if it's not, I think she'll
21 be prepared to talk about the issues that she's worked on,
22 because I think if we don't -- if we're not able to finalize
23 it in May, I think just given the lapse in time from, I
24 think it was November, that it would probably be good to use
25 that meeting to bring the document back up, go over the
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
163
1 changes that have occurred since then, and get prepared to
2 wrap it up in June.
3 DR. BYUS: I think that's good. I had a nice
4 conversation with her about it yesterday. She was just
5 worried she wouldn't be able to get it to me in time to look
6 at. I think my discussions with her are very positive and
7 what she's doing, I think, is very good, incorporating some
8 new data and also doing the REL calculations, as well as
9 calculating it on a variety of different parameters is a
10 good one.
11 If she doesn't get it to me -- I told her, I
12 encouraged her, correct me if I'm wrong, to try and bring it
13 up to the panel in May if possible. At least we can go over
14 some of the issues. If she gets it to me too late I don't
15 have time to read the entire document in detail, we can
16 still do that and bring it back up in June.
17 CHAIRMAN FROINES: I have a question about that.
18 Lesley, if anybody knows, Ray, Lesley, Elinor, is
19 methyl parathion one of the 14 EPA risk assessments that
20 were just released?
21 MR. GOSSELIN: Yes, it is.
22 CHAIRMAN FROINES: It is one of them.
23 So we clearly need to get to the panel that EPA
24 risk assessment, so they have that to look at before they
25 get the DPR document.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
164
1 MR. GOSSELIN: We'll attempt to get the latest
2 version from EPA and get that out.
3 CHAIRMAN FROINES: Okay. And has OEHHA written
4 comments on your document yet?
5 MR. GOSSELIN: Yeah. Those were presented back in
6 November, and our response to those, we've already gone
7 through that step.
8 I believe OEHHA has drafted up their findings.
9 But what I would suggest is that at this point we
10 work on probably just time wise to start taking a look at
11 OEHHA's findings and sort of the current shape of the report
12 besides crafting up closure to the document.
13 CHAIRMAN FROINES: You've drafted your -- Melanie,
14 I'm sorry. You've drafted your methyl parathion findings?
15 DR. MARTY: Yes.
16 DR. ALEXEEFF: George Alexeeff with OEHHA.
17 We've drafted them, and we've approved them
18 internally. We just have not yet transmitted them to ARB or
19 DPR at this point. But they're very similar to the previous
20 findings we had before.
21 CHAIRMAN FROINES: So I think that we -- Paul,
22 we're going to plan to have methyl parathion in May.
23 MR. GOSSELIN: Except that I think staff wise I
24 don't think we're going to have the rewritten report ready
25 in time. It might be tight. But I think if nothing else it
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
165
1 would be a good point to at least go over the document where
2 changes will be in the document and start to go over some of
3 the findings, and then bring closure in June.
4 CHAIRMAN FROINES: Okay. And I think Craig
5 needs -- Craig is the lead, and so he needs to see where
6 you've gotten to as soon as possible, because if he has any
7 major problems, that could be a sticking point.
8 MR. GOSSELIN: Yeah. That's why I'm trying to
9 build in, and I don't think trying to squeeze all this in
10 May may work out, to give Craig time to -- a chance to go
11 over the document and bring up some final issues he might
12 have, but I think by the May meeting we would have reached
13 that point in that discussion to at least start having a
14 broader discussion with the panel.
15 CHAIRMAN FROINES: Yeah. Just I don't want to
16 take it up if we're not going to be able to really go
17 through something that we can bring to closure in our minds.
18 We don't want to create another one of these rocks rolling
19 up and down the hill again.
20 MR. GOSSELIN: I know. I was just thinking if we
21 try to bring this entire document up to the panel all at
22 once at the end, let's say we resolve all the issues with
23 the lead, the panel may bring issues up and the findings may
24 have some further discussion, and we may need to come back
25 in June anyways.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
166
1 CHAIRMAN FROINES: What will then follow and when?
2 MR. GOSSELIN: You mean the next step? I think
3 it's going to take maybe about two weeks or couple weeks to
4 get some of the rewrite and amendments to the report.
5 Then that's going to be transmitted to Craig and
6 for review.
7 And then there will be some final interaction and
8 hopefully we will resolve all the issues.
9 And then the document itself with appropriate
10 flagging of the changed issues will be brought to the entire
11 panel.
12 CHAIRMAN FROINES: No, yeah, I hear that. But I'm
13 saying in terms of other pesticides besides --
14 MR. GOSSELIN: Oh, oh. Molinate is, that document
15 is talked about in December is being significantly
16 rewritten. It's going to take some time, so we're looking
17 summer to having that document be brought back in a new
18 format.
19 The other issue is MITC. We have the initial
20 draft document that we're going to start initiating
21 discussions with the lead on.
22 CHAIRMAN FROINES: Who is our lead on MITC? Roger
23 and Peter?
24 DR. WITSCHI: Yeah.
25 CHAIRMAN FROINES: You're the lead.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
167
1 DR. WITSCHI: I'm not going to be available in May
2 or June.
3 CHAIRMAN FROINES: This is something that is going
4 to come up in the summer.
5 So you're talking about a draft MITC document
6 sometime over the summertime?
7 MR. GOSSELIN: You mean once we get through the
8 leads? Probably, yeah.
9 CHAIRMAN FROINES: And then what follows MITC?
10 MR. GOSSELIN: Molinate will be brought back in
11 the summer. We've already gone through one cycle with that.
12 And I think next up is azinphos methyl.
13 CHAIRMAN FROINES: And that will be in the fall?
14 MR. GOSSELIN: Yes.
15 CHAIRMAN FROINES: And then what?
16 MR. GOSSELIN: Offhand I don't have the rest of
17 the list with me.
18 CHAIRMAN FROINES: Okay. I'm interested in trying
19 to see if we can look at those 14 organophosphate risk
20 assessments and think about whether or not we can approach
21 them as a group and avoid ourselves, if there was some basis
22 that you folks and OEHHA could review those risk
23 assessments, and they were good and we could deal with
24 various elements that are missing that are necessary for
25 California, that we could think in terms of having 14
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
168
1 pesticides move forward based on the US EPA risk assessments
2 but --
3 MR. GOSSELIN: What I could do is have my staff
4 get together with OEHHA, pull together those 14 US EPA risk
5 assessments. And would you want us to initially go over
6 sort of the status and format of them with a lead?
7 CHAIRMAN FROINES: Yeah. I think what we need to
8 do is to see if we can approach them collectively. I mean,
9 George gives us 120 compounds at one time, so 14 should be a
10 piece of cake.
11 The problem is, I'm looking at Lyn Baker, there
12 may -- what information is available on them in terms of
13 pesticide use and air monitoring may be one of the gaps that
14 we have to worry about, and there may be others, but since
15 we have risk assessments already prepared, it behooves us to
16 take a look at those 14 and see how we can make that process
17 move more efficiently.
18 MR. GOSSELIN: We've been working closely with EPA
19 on these 14 and they match up in many respects with what
20 we're working on.
21 If you want to get a lead or two, we can sort of
22 evaluate the status of the 14 with OEHHA and the lead and
23 report back in May.
24 CHAIRMAN FROINES: Okay.
25 MR. BAKER: Dr. Froines, Lyn Baker with ARB.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
169
1 Just thought you might be interested to know that
2 we have done monitoring for eight of those 14.
3 CHAIRMAN FROINES: Well, that's good. I mean, if
4 we can do something to bite off a big chunk of pesticide,
5 that would really be good, because the anticipated 12
6 clearly never worked out. And so but if EPA has done a good
7 job on the risk assessment, and I'm not presuming that they
8 have, they may be of some value.
9 Melanie, what are you doing in May and June and
10 July?
11 DR. MARTY: I hear my staff laughing.
12 We have two more hot spots related documents that
13 we need to bring to the panel.
14 One of them is the technical support document for
15 determining chronic reference exposure levels. And we
16 mentioned that that has 120 chemicals or so in there
17 summarized with RELs.
18 What we wanted to do to make it more palatable is
19 to bring forth the methodology in the first 40 or so
20 chemicals first.
21 CHAIRMAN FROINES: First what?
22 DR. MARTY: First bring forth the methodology in
23 the first 40 chemicals and then feed in the rest of the
24 chemicals over time.
25 And we hope to have the methodology in the first
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
170
1 40 chemicals to the panel in June. So that's to you for
2 review.
3 We also have the exposure assessment and
4 stochastic analysis technical support document which we want
5 to bring forward to the panel this summer. I'm guessing the
6 earliest we would bring that to you would be July.
7 CHAIRMAN FROINES: Do we have a June meeting,
8 Bill?
9 MR. LOCKETT: Yes.
10 CHAIRMAN FROINES: Do we have a July meeting?
11 MR. LOCKETT: No.
12 CHAIRMAN FROINES: Good. Let's not.
13 This is -- we're meeting with reasonable
14 frequency. July would be a good time to take a break.
15 Maybe even August.
16 DR. MARTY: We would anticipate that the rest of
17 the chronic RELs would come to you September-ish,
18 November-ish. Something like that.
19 We also have a Governor's Executive Order to
20 evaluate ethanol as an oxygenate in gasoline, which we need
21 to do an assessment with reference to gasoline without
22 oxygenate and gasoline with MTBE, and that document needs to
23 be peer reviewed, and we were thinking that you all would be
24 the peer reviewers.
25 We have an absolute deadline of December 31st to
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
171
1 get that to the Environmental Policy Council.
2 CHAIRMAN FROINES: What about Elinor and I put our
3 hats on and MTBE hats on and go back to work.
4 Are you going to look at combustion products?
5 DR. MARTY: Yes.
6 CHAIRMAN FROINES: And atmospheric chemistry
7 products?
8 DR. MARTY: Yes.
9 CHAIRMAN FROINES: Because the key question has to
10 do with what are we going to think about in terms of acid
11 aldehyde and PAN, at least.
12 DR. ALEXEEFF: Actually it's a fairly
13 comprehensive analysis of which the Air Resources Board is a
14 large part.
15 There's also a component from the Water Resources
16 Control Board in terms of breakdown products.
17 It hasn't been worked out exactly how the peer
18 review is going to work. Clearly something at least the Air
19 Resources Board and our parts would certainly fit with the
20 activity that this panel does in terms of the atmospheric
21 chemistry breakdown, the types of health effects we're
22 looking at.
23 But we're not sure how it's fitting with the Water
24 Board issue and how we're going to get that component
25 reviewed.
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
172
1 Along similar lines, there's a question in terms
2 of understanding ethanol and its byproducts and relative
3 toxicity. There's a question as to how much we have to
4 compare it with the existing fuel, which is the MTBE fuel.
5 So under that circumstance we may also bring MTBE
6 to the panel to review, which we have a document that's
7 already been prepared.
8 And there's also the UC report which can also be
9 relied upon.
10 Simply to nail down the reference levels and the
11 potencies, so that when the calculations are done with the
12 ethanol document, we have agreement previously on what
13 numbers we're comparing them to.
14 CHAIRMAN FROINES: I assume that you would have
15 to -- what would you do? Modify your PHG document on MTBE
16 to bring it to the panel? That has the quantitative risk
17 assessment.
18 DR. ALEXEEFF: Right.
19 CHAIRMAN FROINES: Our document doesn't have a
20 quantitative.
21 DR. ALEXEEFF: It would be the PHG document, and
22 that's our Public Health Goal water document. And that
23 document was written with the idea that we may have to have
24 an air risk number, so it has an air analysis as well.
25 But the advantage of the UC document is that it
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
173
1 does incorporate, I believe, does incorporate, is my
2 understanding, the Air Resources Board information on MTBE
3 as well as the self-generated information on MTBE that UC
4 did in terms of combustion analysis.
5 So I think you'd have a full exposure assessment
6 without having to wait until we wrote another document, is
7 what I'm trying to think of.
8 I'm not -- that is something that's going to be
9 worked out with us and the Air Resources Board and the Water
10 Board. We may be bringing you those documents later this
11 year as well.
12 CHAIRMAN FROINES: We did an exposure assessment
13 but those data is a little bit out of date at this point.
14 Elinor, what's the last date that we have dated
15 from ARB?
16 MS. FANNING: Gosh, I know we had some '96 data
17 and I believe -- I can't remember if we ever got the first
18 half of the '97 data incorporated into that document. We
19 had issues with sort of getting, compiling it. I would redo
20 that exposure assessment.
21 CHAIRMAN FROINES: Are we going to see the chronic
22 document in May or June? I'm sorry.
23 DR. MARTY: We'll bring it to the panel in June,
24 but that's -- you guys are going to get it in June for your
25 review. If I'm remembering correctly, the meeting is set up
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
174
1 for the 16th. So you won't have had enough time to review
2 everything by the time the meeting comes around.
3 CHAIRMAN FROINES: So we have nothing from OEHHA
4 right now in May.
5 And we have methyl parathion in May.
6 And it seems to me that if there's a problem with
7 methyl parathion such that Craig doesn't think it's ready to
8 come to the panel, then we should not have a May meeting and
9 take up everything in June. Makes more sense.
10 So Craig needs to be the litmus test on whether or
11 not we should plan to take up methyl parathion in May.
12 DR. BLANC: The June meeting is in Los Angeles.
13 CHAIRMAN FROINES: I guess. I don't know.
14 So right now we will --
15 DR. GLANTZ: You've got a lunch offer over here,
16 Craig.
17 CHAIRMAN FROINES: Right now we assume that we
18 will have a meeting in May on methyl parathion, but it's
19 also possible there won't be a meeting in May, and in June
20 we'll take up methyl parathion and the chronic RELs.
21 DR. BLANC: I would like to make a motion that we
22 adjourn.
23 CHAIRMAN FROINES: Got a second?
24 DR. FUCALORO: Did you make the motion or just
25 like to?
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
175
1 DR. BLANC: I'm making the motion that we adjourn.
2 DR. FUCALORO: I second.
3 CHAIRMAN FROINES: All in favor.
4 Thanks, everyone.
5 (Thereupon the meeting was adjourned
6 at 1:45 p.m.)
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345
176
1 CERTIFICATE OF SHORTHAND REPORTER
2
3 I, JANET H. NICOL, a Certified Shorthand Reporter
4 of the State of California, do hereby certify that I am a
5 disinterested person herein; that I reported the foregoing
6 meeting in shorthand writing; that I thereafter caused my
7 shorthand writing to be transcribed into typewriting.
8 I further certify that I am not of counsel or
9 attorney for any of the parties to said meeting, or in any
10 way interested in the outcome of said meeting.
11 IN WITNESS WHEREOF, I have hereunto set my hand
12 this 21st day of April 1999.
13
14
15
16
Janet H. Nicol
17 Certified Shorthand Reporter
License Number 9764
18
19
20
21
22
23
24
25
PETERS SHORTHAND REPORTING CORPORATION (916) 362-2345