OSHA: Proposed Standard For Indoor Air Quality: ETS Hearings, September 26, 1994
OSHA: Proposed Standard For Indoor Air Quality: ETS Hearings, September 26, 1994
UNITED STATES DEPARTMENT OF LABOR
OCCUPATIONAL SAFETY AND HEALTH ADMINISTRATION
PUBLIC HEARING
PROPOSED STANDARD FOR INDOOR AIR QUALITY
Monday
September 26, 1994
Interstate Commerce Commission
Washington, D.C.
The above-entitled matter came on for hearing, pursuant to notice, at 9:46 a.m.
BEFORE: HONORABLE JOHN VITTONE
Administrative Law Judge
AGENDA
PAGE
Neal L. Benowitz 1194
Questions:
Ted Grossman 1245
John Rupp 1311
Jim Dinegar 1340
Debra Janes 1345
Ms. Sherman 1346
Wayne Ott, Ph.D. 1351
Questions:
John Rupp 1376
Ms. Sherman 1434
Peggy Jenkins, M.S. 1442
EXHIBITS
EXHIBIT NO. IDENTIFIED RECEIVED
29 1244 1244
30 1311 1311
31 1316 1316
32 1376 1376
33 1443 1443
P R O C E E D I N G S
(9:46 p.m.)
JUDGE VITTONE: Good morning. I hope everybody had a good weekend. We are going to resume this morning. We have three witnesses scheduled for today. Dr. Neal Benowitz, Mr. Wayne Ott, and Ms. Peggy Jenkins.
My understanding is that the OSHA team will lead off with Dr. Benowitz. Is Dr. Benowitz going to be using a slide or some kind of presentation?
DR. BENOWITZ: Slide projection.
JUDGE VITTONE: Slide projecting. Okay.
Dr. Benowitz, do you want to come forward over to this area here. This is where the witnesses will be, right there.
Dr. Benowitz, would you state your name for the record, your affiliation, and who you're representing today, if anyone?
NEAL L. BENOWITZ
DR. BENOWITZ: I'm Neal Benowitz. My affiliation if University of California San Francisco, and I'm here representing OSHA.
JUDGE VITTONE: Okay. You supplied a statement for the record on August the 18th, 1994?
DR. BENOWITZ: Yes.
JUDGE VITTONE: All right, sir. If you're prepared to go forward with your presentation, you may do so now.
DR. BENOWITZ: Yes, I would.
I would like to start out and just give a little bit of my background. I'm currently professor of medicine at University of California, San Francisco, and Chief of Clinical Pharmacology and Experimental Therapeutics, which is part of our Department of Medicine. I practice medicine at San Francisco General Hospital.
I have board certification in internal medicine, on clinical pharmacology and medical toxicology. The bulk of my research, through my academic career has been studying the effects and pharmacology in nicotine in humans. I've done work in that area for about 20 years.
That work has resulted in over 250 publications and 40 or 50 book chapters, the bulk of which have related to effects or pharmacology of nicotine in humans.
I was a scientific editor of the 1988 Surgeon General's report on nicotine addiction. I've been involved in other Surgeon General's reports on smokeless tobacco, environmental tobacco smoke, and ethnicity issues.
I've also served on the Scientific Advisory Board of EPA on their risk assessment of environmental tobacco smoke, and I recently served on the Institute of Medicine Committee that dealt with preventing tobacco or nicotine addiction in youth.
What I would like to talk about this morning is the issue of the validity of cotinine as a biomarker of human environmental tobacco smoke exposure.
This first slide is objectives. Cotinine is the main proximate metabolite of nicotine and has been widely used as a marker of environmental tobacco smoke exposure, both in smokers and in non-smokers.
The use of cotinine is a biomarker for environmental tobacco smoke exposure has been criticized, and what I would like to do is review the basis for the use of cotinine for that purpose and address various criticisms that have questioned its validity.
The objectives of my presentation are to establish that cotinine is a valid quantitative marker of the level of environmental tobacco smoke exposure; and, second, to establish that measurement of cotinine levels in human blood, saliva or urine is the most specific available marker of environmental tobacco smoke exposure.
Two: Nicotine is present in environmental tobacco smoke. Some of the background that's important before we talk about cotinine as a biomarker is some of the issues surrounding nicotine, it's parent, and its validity as a marker of environmental tobacco smoke.
It's known that the tobacco in a cigarette contains from 6 to 12 milligrams of nicotine. On average, a smoker absorbs systemically 1 milligram of nicotine, and that is almost irrespective of what the nominal smoking machine determined yield shows.
This is because individuals tend to adjust their smoking behaviors to take in a certain amount of nicotine that satisfies their nicotine addiction.
75 percent or more of nicotine in tobacco is emitted into the air as sidestream smoke, so that's 3 milligrams or more per cigarette.
It has been well shown that nicotine is present in the air in work places where there is smoking occurring, and concentrations of nicotine are much different in environments where people smoke or where people do not smoke.
Let me go on, next, to provide some of the basic parameters regarding nicotine in tobacco smoke.
Slide 3: sidestream tar to nicotine ratio is not substantially affected by a brand of cigarette.
One criticism has been that nicotine levels are not representative of levels of other tobacco toxins in ETS. It is known that ETS is present primarily in the vapor phase.
I say nicotine is present primarily in the vapor phase of ETS. There are other toxic components that are part of the vapor phase and still others that are part of the particulate phase, which has also been assessed as tar.
I think it's important to recognize that there are harmful chemicals in both phases and main culprits in causing cancer or heart disease or other injuries from environmental tobacco smoke are not known, so we really don't know which phase is the optimal.
I should also say nicotine is most likely not a major culprit in causing disease, and so nicotine is serving primarily as a marker for other potential toxins and environmental tobacco smoke.
One issue of concern is whether nicotine represents environmental tobacco smoke from different brands of cigarettes -- high yield cigarettes, low yield cigarettes, filter cigarettes, non-filter cigarettes, ventilator, non-ventilated cigarettes.
On this slide, I've summarized results of two studies that have looked at this question, where sidestream smoke was captured, and concentrations of tar and nicotine were measured in sidestream smoke.
What I have done is looked at the ratios of tar to nicotine in sidestream smoke.
I've expressed the ratio as a means. I've also shown the range for you on this slide, and the coefficient of variation.
The coefficient of variation is a term that means the standard deviation of a series of measurements divided by the mean, and this is a term that's widely used as a measure of variability. I'll be coming back to this.
In studies of Adams and Rickert, which looked at different kinds of cigarettes, the mean ratio of tar to nicotine was relatively constant, you can see the average was 5.2 in one case, 5.8 in the other case. The range is fairly narrow.
The Rickert study had one value at 8.1, but that's really an outlier. Most of the values were actually closer to the range of 5 to 6.
Most importantly, coefficient of variation is about 12 or 13 percent. This, when looking at biological measures, is quite a small coefficient of variation, and I think, by most standards, we would say that these studies indicate that sidestream smoke, tar to nicotine ratios, are comparable in different kinds of cigarettes.
In Rickert's study, he specifically looked at the correlation between tar to nicotine ratio in sidestream smoke and mainstream nicotine or tar yield, and they found no correlation.
He also looked at ventilated and non-ventilated cigarettes and found no correlation either.
So I would conclude from these data that nicotine is representative of environmental tobacco smoke generated from different brands and types of tobacco.
Next slide 4: Time-weighted nicotine exposure reflects exposure to other components of environmental tobacco smoke.
Another criticism is that the ratio of nicotine to other ETS components varies over time with the aging of smoke, and therefore nicotine is not an indicator of other components.
As I mentioned previously, nicotine in ETC exists primarily in the vapor or gaseous phase. There is evidence that nicotine and particulate concentrations decline at different rates, depending on surface material and other characteristics of the environment.
There is also a problem in that there are particulates that exist from other sources, besides environmental tobacco smoke, and that can bias ratios of nicotine to particulates, particularly at low ETS concentrations. This is something that Dr. Hammond is going to address in great detail later on.
So, in fact, if one measures at any particular point in time after single exposure, one can see changes in nicotine to tar ratio. However, people are exposed to ETS not at single points in time but over long periods, either in the work place, say over 8 hours, and what is most appropriate is to really look at time-weighted exposures which would deal with repeated generation of smoke and decline of nicotine and particulate concentrations over time.
When that is done, when time-weighted samples are used rather than point samples, there is fairly good consistency of nicotine to particulate ratios.
Slide 5: Nicotine and air sampled over time reflects exposure to other ETS components.
These are some data that were reviewed in a paper by Repace and Lowery. Again, some of these data Dr. Hammond will talk about later.
This slide shows results of three studies which looked at the relationship between respiratory suspended particles or RSP, to nicotine in the air of homes and workplaces.
The first paper by Hammond involved 47 home measurements; Meisner measured 21 workplaces, and Nagda 61, I believe, aircraft. When these air samples were taken over time, so we're looking at time-weighted averages and the relationship of RSP to nicotine is examined, there is relative consistency with a ratio that varies from 9 to 10.
Again, Dr. Hammond will talk about this later, but my conclusion, looking at this is, when the exposure to ETS is measured over time, as is relevant to human exposure, that the level of nicotine in air reasonably well represents the concentrations of other components of ETS, such as particulates.
This is Slide 5A, showing structures of nicotine and proximate metabolites.
Now that we have talked about the issue of nicotine in air as being a reasonable marker of ETS that's being present, I'd like to then go on to talk about assessing human exposure.
This slide shows the structure of nicotine in your upper left and shows the main proximate metabolic pathways. The one of concern today is cotinine. You can see that cotinine, basically, consists of having oxygen added to one of the -- to the five-member component or ring of nicotine. So that's the proximate pathway.
The other one is nicotine oxide, which is really a minor metabolite.
On average, about 70 or 80 percent of nicotine goes through cotinine, and then, as I'll show you in a second, cotinine is also further metabolized to other compounds, but it is important that, in the vast majority of people, most nicotine goes to cotinine. So it's usually between 70 to 80 percent.
This Slide 5B: These are some quantitatem metabolism data from work in my laboratory that address the fate of nicotine in a quantitative sense in humans.
The circles on the outside -- it starts with nicotine in a box in the middle, and then cotinine in the box in the middle, and then the circles on the outside are what is recovered on average in the urine, from a given dose of nicotine.
So you can see, again, about 80 percent of nicotine goes to cotinine. That is excreted, to some degree, in the urine, so about 13 percent of a nicotine dose is excreted as cotinine.
Cotinine is then converted to trans 3 prima hydroxy cotinine, which is then also excreted in the urine part and also conjugated. Cotinine is conjugated as metabolite, as is nicotine, to some degree, also.
Now, if one looks at what's in the urine, the 3 hydroxy cotinine, is the major metabolite that's found in the urine. However, there is considerable variability in the extent of metabolism of cotinine -- to 3 hydroxy cotinine -- so some people convert a lot and some people convert relatively little.
In that nicotine exposure, results of nicotine going into the bloodstream, and then liver metabolism, is most reasonable to consider the major proximate metabolic pathway as most likely being most representative of nicotine dose, so I believe that cotinine measured in the plasma is probably the best metabolic marker of nicotine intake, again, because, there's relatively a high percentage of nicotine converted to cotinine.
You will hear other speakers say that cotinine is not the major metabolite in the urine, and therefore cotinine is not worthwhile measuring. I think that that is not true. Again, I would restate that the proximate metabolite best reflects the parent and the real issue is the relationship of urinary cotinine, say, to plasma cotinine, and we'll talk about that in a minute.
Slide 6: Cotinine in various biological fluids.
Let me say, because I haven't explained this, nicotine is the main marker in environmental tobacco smoke that we're trying to measure. We don't measure nicotine as a biomarker because nicotine has got a relatively short half-life of 2 to 3 hours, such that with intermittent exposures, levels rise and fall, and so there's quite a bit of variation in the nicotine levels throughout the day.
Cotinine, as its proximate metabolite, has a half-life averaging of about 16 hours, so it is formed from nicotine and then tends to persist for a long period of time and basically gives an integrated value offer previous nicotine exposure.
That's the reason by cotinine has been widely used as a marker of tobacco exposure. Cotinine levels are relatively constant over time.
These are some representative data. There are many studies, and I have chosen this study of Jarvis in the U.K., where nicotine levels were measured by gas chromatography, which is a specific assay, and here are reported mean concentrations in nanograms per milliliter of cotinine and plasma, saliva, and urine.
He studied 94 smokers, 54 non-smokers with environmental tobacco smoke exposure, and 46 non-smokers without environmental tobacco smoke exposure.
If you look at your right-hand column, the values start with plasma. Smokers, more than 100 times that of non-smokers in general. So it's clear that cotinine can distinguish smokers by nonsmokers.
If you then compare nonsmokers with ETC exposure and without, there's also about a 2-fold difference here. So there are significant differences looking at mean values, between people with ETS exposure that's self-reported and without.
A number of studies have shown that plasma cotinine concentration is responsive or is increased in people who have a history of ETS exposure compared to those who don't.
Salivary concentrations are also shown here. Saliva is a widely-used marker because saliva is easier to get. It's not invasive.
A number of researchers have looked at the ratio of cotinine and saliva and plasma, and have shown that they're pretty well correlated with correlation co-efficients of .9 or greater, and salivary concentrations tend to be slightly higher than the plasma concentrations, but they are approximately the same.
It is assessed for two reasons.
One is that it is not invasive, and two is that the concentrations of cotinine are higher in the urine than in plasma, so it's easier to measure urinary concentrations.
In general urine concentrations are about five to six times greater than plasma concentrations. There is some degree of variability expected on the basis of urine flow rates, and perhaps urinary pH to a small degree, but again, in Jarvis' study where he looked at the correlation within individuals of plasma to urine concentrations, he found that there was an excellent correlation.
So you will see studies that report urine concentration, saliva concentration, or plasma concentration of cotinine as markers of environmental tobacco smoke exposure, and these are the general parameters that we'll be looking at.
What I would like to go on to now are some of the pharmacokinetic issues used in translating a cotinine level to a particular intake of nicotine from environmental tobacco smoke, and so I'd like to spend a little time talking about pharmacokinetics.
Slide 6A is a beaker model of human pharmacokinetics. What I'd like to do is to represent the human body, for simplicity sake, as a beaker that has a certain volume of fluid in it, and this volume can be thought of as the blood stream, or the blood stream and body tissues in which a drug is dissolved or binds.
On the top panel you can see that if you dose a certain amount of drug with that syringe, put it into this beaker, mix it up very well, and then take a sample from that beaker to measure the concentration, the concentration will be stable, because you put it in, and there's none leaving.
If you look at a concentration over time, you would get a curve, as shown on your upper right which basically shows a concentration is flat over time. So that would be a situation where there is no drug metabolism or clearance. Now the body does have mechanisms for getting rid of drugs.
The pharmacokinetic process of getting rid of drugs has been termed, "clearance," and clearance is, in effect, a spigot function. You can see on the bottom curve that there's a beaker that has a spigot, and a certain amount of fluid drops through that spigot, and if we assume that this is a recirculating spigot, so that all the fluid goes back into the volume so that the volume does not change, and if we also assume that whatever comes through that spigot is totally cleansed of drug, so that's the drug elimination pathway, then the flow rate through that spigot is the term that's called clearance.
In a sense, it's the volume of blood that's flowing per minute, say in milliliters per minute, of blood that's cleansed of a drug, and that's the term that we use to explain metabolic rates of drugs, we talk about clearance of a drug.
Now clearance is not the same thing as elimination rate; you'll see why that's important later on. If you wanted to know how many milligrams of drug is being eliminated from this beaker, what we would do is to measure the concentration of fluid in the beaker, then measure the spigot flow rate, and then multiply the two together, so that so many mils per minute of fluid or blood are being cleared, and each mil contains "X" number of nanograms or micrograms of drug, and therefore the total elimination rate would be micrograms or nanograms per minute.
The elimination rate is a product of this clearance times the concentration. I don't have a slide here to show this, unfortunately, but another critical concept is that of steady state.
Say instead of giving a single does into the beaker with the syringe, you constantly drip a certain amount of drug into the beaker over time, and you keep on dripping in so that levels in the beaker will rise and rise and rise, and at a certain point, they become constant.
They become constant when the rate of the drug going in is the same as the rate of the drug going out, which means that the dosing rate is equal to the elimination rate, which I said was equal to the product of the clearance times the concentration. That steady state, and I'll come back and emphasize this again, has three factors that are operative only.
It's really the dosing rate, the concentration, and the clearance rate. Any given concentration at steady state is determined only by the dosing in rate and the clearance rate out. It doesn't matter what the volume of this beaker is or any other characteristics. So that's an important steady state concept. And I'll explain more in a few minutes about why that's important.
Slide 6B shows the concept of half life, which I'll be talking about as well. If you go back to this beaker, and you give a certain dose of the drug, you allow clearance to occur, and you measure levels over time, you'll see as you see in the curving line that these levels will decline or fall over time.
If you were to plot it on a logarithmic scale, which is in your upper right-hand corner, you'll see that this line becomes straight. This straight line on a logarithmic scale means that a given percentage of the drug in the body is eliminated for any given fixed interval of time, which means, say we talk about half life.
Half life is the time it takes for half the drug in the body to be eliminated. So say if the half life for cotinine is 16 hours, that means that once you have a certain amount of cotinine in the body, it takes 16 hours for half of it to be gone, 16 hours for half of that to be gone, and 16 hours for half of what's left to be gone, etc.
So half life gives us some idea of how fast the body gets rid of a drug and also gives us some information of accumulation of a drug.
Slide 7: Pharmacokinetics of nicotine and cotinine are similar in smokers and nonsmokers. We know a lot about nicotine and cotinine kinetics and metabolism in smokers, and a concern has been raised about the comparability of nicotine and cotinine metabolism in smokers versus nonsmokers.
This is important if we're going to use some of our knowledge about smokers to apply to nonsmokers. Some early reports using small doses of radioactive nicotine suggested that there were different rates of metabolism, such that smokers metabolized nicotine faster than nonsmokers.
We have done quite a bit of work in this area, and our approach has been quite different from that used previously. We have wanted to study metabolism of nicotine and cotinine in smokers while they're smoking, as well as nonsmokers.
What we have done is to synthesize deuterium labeled nicotine and cotinine, so that's a non-radioactive stable isotope. We've shown in preliminary experiments that these labeled nicotine and labeled cotinine are metabolized in exactly the same way as natural nicotine, and I should say that our experiments with l-nicotine, which is natural nicotine, the sort that's found in tobacco.
I will also point out that some of the early studies with low doses of radioactive compounds used d&l-nicotine, racemic nicotine, so it has some of the unnatural nicotine.
We know from other substances, as well as from cotinine itself, that d&l-nicotine and cotinine may be metabolized at different rates, so I really don't think those studies are representative of humans.
In any case, in our experiment we looked at smokers and gave infusions of labeled nicotine, and we gave them infusions at two doses: We gave them a dose comparable to what they might get from smoking cigarettes, and then we gave them a very low dose, a dose would be the same as what we could give safely to nonsmokers, and then looked at a variety of pharmacokinetic parameters, as shown in this slide.
We looked at half life, volume of distribution, which is sort of the space or volume of the fluid and tissue inside the beaker, and total clearance.
We found that the half life of nicotine in smokers was 157 minutes, and 122 minutes in nonsmokers. The volumes were basically the same. The clearance was actually slightly faster in nonsmokers, at 1319 mils per minute versus 1085 mils per minute in smokers.
These values are pretty close, but most importantly, they're totally different from the earlier studies of Kyerematen, which suggested that smokers metabolized nicotine much faster, and I'll explain why I think it's different in a moment.
I would say that for practical purposes nicotine is metabolized and handled in a very similar rate in smokers and nonsmokers. I can't show you the same exact set of data for cotinine, but I can give you results of other data where we've infused cotinine to smokers and nonsmokers and basically found that half life is almost identical and that clearance is very similar.
I would say that two other investigators have looked at cotinine clearance rates in smokers and nonsmokers, and they all seem to be pretty much the same. So from my data and those of some of the other investigators of cotinine, I conclude that there are minor differences if any in the clearance and metabolism of nicotine and cotinine in cigarette smokers and nonsmokers.
Now the question is why would the early Kyerematen data have shown a much longer half life in nonsmokers? I can't say for sure, but one factor certainly is the fact that they were using unnatural nicotine, which is not appropriate for tobacco studies.
The second thing is that they used very low doses, and in nonsmokers, there may be dose-related binding of nicotine to certain body tissues which then gets slowly released, and so a long half life of nicotine might be due not to a difference in metabolism but slow release in body tissues. I believe that their data do not accurately reflect clearance and really are looking at probably distributional funny.
Another issue which I can't show you data on yet, but we'll have some data soon, is the pattern of metabolism of nicotine and cotinine in smokers versus nonsmokers.
As I've said before, 70 to 80 percent of nicotine is converted to cotinine. If there's concern about the use of cotinine in nonsmokers, it would be that cotinine is somehow overestimating the exposure. So the concern you would have in terms of metabolism is do nonsmokers make more cotinine from nicotine and therefore overestimate cotinine.
But when you're starting from 80 percent, the maximum you could go is from 80 percent to 100 percent, which is not likely, because other people have measured metabolites in nonsmokers and have shown that there are other metabolites, nicotine as well.
So the maximum error that could play a role if patterns of metabolism are different would be about 20 percent, which would not create any substantial error.
I think that the metabolism of nicotine and cotinine is likely to be quite similar in smokers and nonsmokers. The pharmacokinetics are similar, and therefore, we're justified in using data from smokers to do dose estimations in nonsmokers.
Slide aL: I just want to show you an example of some of the data from our study in smokers and nonsmokers that I've just described. These are blood levels of labeled nicotine that were given by intravenous infusion.
The solid circles are smokers, the open circles represent nonsmokers. If you look at the low dose where smokers and nonsmokers got exactly the same dose, those are the lower two lines, you can see that nicotine levels are virtually superimposable in smokers and nonsmokers. So there's no evidence whatsoever that they handle nicotine in a different way.
Slide 8: At steady state, the blood concentration is determined by exposure rate and clearance.
One of the assumptions that we'll be making in interpreting cotinine values is that cotinine levels represent relative steady state levels with steady state exposures.
Again, let me go back to this key concept. At steady state, as shown on the top line of this equation, the "d" which is the dosing rate is equal to the elimination rate.
That's the definition of steady state: what's going in is what's going out. We know that elimination rate is equal to the product of the clearance times the blood level at steady state.
If we rearrange this equation, which is shown on the bottom of this slide, you can see that a blood concentration at steady state is determined only by two factors: by the dosing rate, and by the clearance.
Slide 9: Prolonged half life after low-level exposure does not affect the validity of the use of cotinine in steady state exposure conditions.
Now, the reason why I've made such a big point about the steady state business deals with a criticism that the elimination rate of cotinine in nonsmokers after particular exposure to nicotine is longer, substantially longer, than that which is seen in smokers.
Therefore, the claim has been made that measuring cotinine in nonsmokers would overestimate nicotine exposure. Well, that is not true at steady state.
As I've said before, one explanation for this funny could be that with a single dose in a nonsmoker, there is avid binding to body tissues and very slow release, and that could produce a long half life, which is not elimination determined, but actually released from body tissues.
At steady state conditions, as would be the case with regular ETS exposure in the workplace, the relationship between exposure rate and blood concentration is determined only by clearance, not by tissue binding, and the data showing very long half lives after a single exposure in nonsmokers are totally irrelevant.
They do not reflect clearance, they do not reflect the dose-concentration relationship, and those data have no bearing on the assessment of steady state exposures.
Slide 10: Cotinine is a quantitative biomarker of daily nicotine intake. I would like now to consider the quantitative aspects of cotinine as a biomarker, and the question of inter-individual variability. Now the details behind the derivation of the equation, as I'm going to show you, are in my statement, so I won't bore you more than I have already with these pharmacokinetic issues by explaining them all, so I'm going to show you the first and last parts.
The basic assumption is that at steady state, the generation rate of cotinine, which is generally how much cotinine is being generated from nicotine in the body per time, is equal to the dosing rate of nicotine, how much nicotine is being taken in per time, times the percent of conversion of nicotine to cotinine.
In other words, you get a certain amount of nicotine in, a certain percent gets converted to cotinine, and if you know those two factors, you know how much cotinine is entering the body per unit of time. What we want to find out by the use of a biomarker is the intake rate of nicotine.
So we need to know two things for that:
The generation rate of cotinine and we need to know the percent conversion of nicotine to cotinine. In studies we've been doing in large populations, we've been giving labeled nicotine and cotinine simultaneously to two peopl