Tobacco Documents Online

Topica-l Report NCI/S&HP/ORNL #60 SITE VISIT II TO BATTELLE PACIFIC NORTHWEST LABORATORIES RAT MODEL INHALATION BIOASSAY Richland, Washington, September 14-15, 1977 12-30-77 R. A. Jenkins and B. E. Gill Tobacco Smoke Research Program Bio/Organic Analysis Section Analytical Chemistry Division Oak Ridge National Laboratory Oak Ridge, Tennessee 37830 Interagency Agreement (ERRA-NIfl/NCI) No. 40-485-74, Part II Internal (ORrdL•).~Contract Charge No. 3390-0224 Intended for informal communication with project management only. Confidential until published or released by author.

SITE VISIT II TO BATTELLE PACIFIC NORTHWEST LABORATORIES RAT MODEL INHALATION BIOASSAY R. A. Jenkins and B. E. Gill Summar . Data collected on a second site visit to BNW suggest that the smoke dose which the Code 13 group retains, when normalized for body weight, is at least as large as that which Code 13 beagles retain in other bioassays. The SEB IV rats probably retain a 50110 greater dose than comparable dogs. The overall quality of the exposures at BNW continues to be very high. Measured CO and C02 concentrations in the exposure chamber atmosphere suggest that the Code 32 and 27 groups are exposed to the targeted levels of smoke gas phase constituents. The relative concentration of some selected constituents in the particulate phase of the chamber atmosphere is the same as that generated by the Code 29 variant on-site. Whether significant differences in smoke chemistry exist between smoke generated at BNprd and that generated at ORNL still remains open to question. An optical particulate sensor was used to continuously monitor smoke concentrations in the exposure chamber. Results of this monitoring indicate that the animals deplete a vastly larger fraction of the smoke from the later puffs of the cigarette than they deplete from the earlier puffs. Introduction. A second site visit to the tobacco smoke inhalation bioassay being conducted at Battelle Pacific Northwest Laboratories (BNW) in Richland, Washington was made on September 14-15, 1977. The primary purpose of the visit was to obtain samples to aid in chemical definition of the inhalation exposure and to provide data for comparison with other NCI-sponsored bio- assays. Data was also obtained to provide a measure of the smoke dose retained by the experimental animals. In addition, an optical particulate sensor, developed under contract with the Council of Tobacco Research USA, Inc., was used as a continuous monitor of smoke concentration in the exposure chamber. Data was obtained to determine the practicability of using such a sensor to more accurately determine the quantity of smoke particulates depleted by chronically exposed rats." Experirnental. Details concerning the exposure protocol and various analytical procedures are described in the Topical Report covering the first site visit to BCaW (NCI/S&HP/ORNL #41 ). Thi s was our fi rst opportuni ty at BNW to measure the levels of both particulate (TPM, nicotine) and gas phase (carbon monoxide

and carbon dioxide) constituents which are actually present in the exposure chamber. These constituents are collected by the method referred to as total chamber withdrawal.. In this p rocedure, the normal chamber face plate- containment tube assembly is replaced with a pair of flat, solid face plates, on one of which is mounted a large bore sampling port. Machine puff volumes are then adjusted to reflect any changes in effective chamber volume. On the sampling port are mounted a Cambridge filter pad, evacuated Saran gas sampling bag, and three-way stop-cock assembly, the third arm of which is connected to a 1500-m1 manually operated syringe. After a puff is delivered to the exposure chamber, smoke is drawn rapidly through the filter pad (which collects TPsi and nicotine) and into the syringe. The stopcock positioning is altered and the smoke gas phase (diluted with air) is ex- pelled into the gas sampling bag. The process is repeated for each puff. An aliquot of the gas phase is transferred to a 125-ml glass gas sampling tube. The sample is then returned to ORNL for analysis of CO and C02 via gas-solid chromatography. Good visualization of the two constituents is achieved using 80/100 mesh Amberlyst 15 ion exchange resin in the Ni-form as the support-stationary phase. As mentioned previously, the CTR-sensor was used to continuously monitor particulate concentration in the exposure chamber. The sensor is described in detail elsewhere (1). Briefly, the sensor consists of a light emitting diode - phototransistor package. Infrared light from the LED is backscattered off smoke particulates and is registered by the phototransistorr. The ensuing resistance change is converted to a voltage drop, amplified, and is displayed on a strip-chart recorder. The sensor head is placed in a port on a modified faceplate on the exposure chamber. In order to prevent stray room light from affecting the sensor response, a black cloth hood is placed over the exposure platform once the animals are in place. As smoke is drawn into the chamber, the sensor response rapidly climbs above its baseline value. Then, as the animals respire and deplete smoke from the chamber, the sensor response falls in a nearly linear manner. When smoke is expelled from the exposure chamber, response rapidly returns to baseline. The summation of the absolute magnitude of the decrease in response.o~er each puff of the cigarette, divided by the summation of the absolute magnitude of the initial response, is taken as the fraction (percentage) of the smoke depleted from the chamber. The magnitude of response is somewhat dependent on the size of the particles which are scattering the light. Thus, in a dynamic system such as cigarette smoke,

the response may be expected to change as the size of the aerosol droplets grow. However, for the dilute smoke present in the exposure chamber, coagulation rates are reduced, and the maximum response increase over the 30-second exposure period has been determined (without animals present) to be about 3 %. General Observations. The ORNL site visit team was impressed with the over- all high quality of the inhalation exposures being conducted at BNW. The technical staff appears to be aware of the importance of attention to detail, and great care is taken to insure the continued quality of the exposures. The exposure systems are kept very clean and are well maintained. It is our opinion that the animals being exposed are'receiving as reproducible a smoke dose as is possible to administer, given the inherent limitations of the exposure system and cigarette variation. From a smoking standpoint, the environmental conditions in the exposure wards were acceptable, but slightly drier (44% relative humidity at 73.5°F) than desirable (60% RH at 74°F). BNW staff are compensating for the low humidity by storing cigarettes in a humidor until immediately prior to smoking. Relative humidity in the humidor was measured to be 66%. Table I compares cigarette static burn rates for an SEB IV variant (Code 04) smoked in both wards with that for the 04 variant under analytical conditions. (Values given in all tables are means ± standard deviations.) The burn rates under exposure conditions are only slightly elevated, and are probably due to the ventilating air flow in the wards. The E3Nld values are comparable to those encountered at other NCI-sponsored bioassays. In general, environmental conditions in the BNW study can be considered good. Machine puffing parameters were determined on this visit, just prior to the daily adjustment of puff volumes. Puff durations in one ward (Rni 328) were short by about 10%, averaging 1.8 seconds per puff. Puff volumes in this room were slightly high, averaging 37.3 ml. In the other ward, puff r,~~? durations averaged 1.9 seconds, puff volumes averaged the specified value ~ of 35 ml . Q Q •` CO Results and Conclusions. Onc~Tab1e II are tabulated individual values for TPM and nicotine deliveries of cigarettes smoked on the Maddox-ORNL exposure systems at EiNGJ. These samples are considered to be machine input samples. The values for the Code 27 cigarettes exhibit the greatest variation. However, since the relative standard deviation of the nicotine:TPFI ratio for the Code

27 smoke is less than 10%, the variation in nicotine delivery is probably the result of variation in total smoke delivery, rather than variation in cigarette nicotine content. Table III compares average per cigarette deliveries of cigarettes smoked under exposure conditions at BNW (from Table II) with those of cigarettes smoked under analytical conditions at ORNL. Despite the slightly higher static burn rate at f3CiW, Code 29 and 27 cigarettes smoked there averaged somewhat higher deliveries of TPM and nicotine. The Code 29 (SEB-IV) cigarettes smoked at BNW show a substantially increased level of nicotine. Since the nicotine:TPM ratio is so much greater at BNW than that at ORNL (0.075 vs 0.056), a non-representative sampling for cigarettes smoked at BE3W may be indicated. For the Code 27 variant, the nicotine:TPM ratio determined under both analytical and exposure conditions is essentially the same (0.052 vs 0.054). This may suggest that the exposure systems smoking the Code 27 variant may be just generating slightly more smoke. Table IV summarizes chamber depletion data for the variants as measured on this visit. Nicotine values are expressed in terms of the nicotine concentration (in pg/ml) in the atmosphere available to the rats. Because of some sampling difficulties, not as many valid samples were obtained as would have been desired. The chamber depletion data suggest that the Code 27 and SEB-IV exposure groups deplete nearly the same fraction of the available smoke. The fraction of Code 13 smoke depleted appears somewhat lower. Fiow- ever, the absolute amount of nicotine in the Code 13 grab sample is so low that analytical error makes this value less reliable. Table V compares the average amounts (per cigarette) of TPM, nicotine, CO, and COz actually found in the exposure chamber (as measured by the total chamber withdrawal method) with those generated under analytical smoking conditions. Several points should be noted. First, all of the measured TPM values are slightly lower in the chamber samples. Because of the scatter in the data, not all of the differences are statistically significant, but they do suggest that some of the more volatile constituents may be evaporated off of the filter pad by the relative,ly large amount of air drawn across it during sampling. In all casers; the nicotine content of the chamber is some- what greater than that which would have been predicted from analytical smoking. This gives further support of the hypothesis that the exposure devices are producing a greater level of nicotine on-site. The most impor- tant comparisons from Table V are the CO and C02 values. The data indicate

that for the SEB IV and Code 27, as much carbon monoxide and carbon dioxide get into the chamber as is generated analytically. Thus, the Code 32 (SEB-IV) and 27 animal groups are exposed to as much CO and C02 as would have been predicted from analytical smoking. The chamber contents values for Code 13 are about 20% lower than the analytical values. It is also important to note that the CO:CO2 ratios for the contents of the exposure system are very close to those found under analytical conditions. This indicates that, despite differences in puffing profiles, the effective puff volumes through the fire cone of the cigarette and the combustion efficiencies for the two systems are very similar. Table VI compares nicotine values for the smoke collected at the input to the exposure chamber with those of smoke actually removed from the chamber and those which would have been predicted from the "early" (3 second) 20-ml chamber grab sample used to define the initial smoke concentration to which the animals are exposed. The predicted nicotine values are calculated from a measured amount of nicotine collected by drawing a known volume (20-m1) across a filter pad, a measured number of puffs, and an assumed effective chamber volume of 350 ml. That the amount of nicotine actually found in the chamber is relatively close to that collected at the chamber input in- dicates that very little smoke is lost going into the chamber. (This is confirmed by visual observation.) That there is such a discrepancy between the amount of nicotine predicted to be present and that found suggests that the assumption of a 350-rrrl effective exposure chamber volume is grossly incorrect, and in some cases may be low by as much as a factor of 2-3. This tends to support the hypothesis that rapid dilution of the smoke into the animal containment tubes occurs simultaneously with the introduction of smoke into the chamber. One of the major criticisms of the chamber depletion method of estimating smoke dose has been that the apparent decrease in smoke concentration in the exposure chamber (as measured by the so-called 27-second grab sample) could be due to either depletion by the animals (inhalation or fur deposition) or slow dilution of smoke into the containment tubes (i.e. an increase in the effective vo,lrame of the exposure chamber). The data presented on Table VI'seem to suggest that this dilution occurs much more rapidly than may have been anticipated. The chamber depletion method is probably a valid procedure for measuring the relative depletion of smoke from the exposure atmosphere. However, an additional, independent method

for determining the absolute amount of smoke available to the animals (such as the method of total chamber withdrawal) must also be employed to define the exposure. - On this site visit, some preliminary data was obtained in order to determine whether the aforementioned optical particulate sensor could be used as a continuous smoke concentration monitor to instrumentally determine the amount of smoke depleted by the animals during chronic exposure. The sensor has the advantages of being totally non-invasive and requiring no additional chemical analyses to determine smoke depletion. Furthermore, it provides a continuous readout, alerting the operator to any leaks or other malfunctions in the exposure system. It is not practical to include copies of sensor response traces in this report. However, several things are evident from these continuous readouts of smoke concentration. First, the well documented increase in the amount of particulates generated with increasing puff number is quite apparent. Over the course of the cigarette consumption, the amount of Ti'M per puff increases by about a factor of two. Secondly, the fraction of available smoke which the animals deplete varies dramatically with puff number. At the earliest puffs, the rats deplete only about 10% of the available smoke. At the later puffs, the animals may deplete as much as 70% of the smoke available on a given puff. Since smoke chemistry can change as a cigarette is consumed, this alteration in the fraction of smoke depleted can be important in terms of the chemical composition of the smoke dose which the animals retain. Table VII compares estimated exposure chamber smoke depletion for three different sets of animals per code as calculated from nicotine analyses of the 3-second and 27-second grab samples of the chamber atmosphere with that calculated from the continuous smoke concentration readout of the CTR optical sensor. Calculations from the sensor response indicate that each exposure group depletes about the same fraction of available smoke. The sensor response suggests that this fraction is slightly less than that cal- culated from the nicotine grab samples. This is probably because the nicotine grab sample values are not corrected .fer the °'chamber depletion" which are a result of the sampling itself. The early (3-second) 20-ml grab sample can potentially remove as much as 6,°% of the smoke in the chamber. In general, it appears that the sensor can do as well as a grab sample procedure in determining the amount of smoke depleted from the exposure chamber.

Of course, one of the most important chemical parameters of any tobacco smoke bioassay is the dose which the experimental aniinal retains. Table VIII lists the estimated maximum possible weekly dose of nicotine, per animal, for the three experimental groups at BiV4t. The method for estimating dose no longer relys solely on the nicotine chamber grab sample. Instead, the chamber grab sample is used to measure only the fraction of available smoke depleted. The absolute amount of smoke present is determined via the method of total chamber withdrawal. The values on Table VIII are considered maximum possible doses, since apparent depletion can be a result of impaction of smoke particulates on chamber walls or deposition on the fur of the animal, in addition to deposition in the animals'.respirator tracts. If only half of the particulates depleted actually deposit in the respiratory tract (i.e. the true dose is only half that listed on Table VIII), then, in terms of a body weight normalized dose, the Code 13 rats would retain about the same normalized dose as do beagles exposed to Code 13 cigarettes in other NCI bioassays. The SEB IV group at BNW would retain a 50% greater normalized smoke dose than a comparable group of beagles. One of the important functions of bioassay monitoring is to determine the extent to which the chemical composition of the smoke which the animals are offered is an artifact of the exposure methodology. Tables IX A-C com- pare the concentration of several selected particulate phase constituents in smoke generated by the exposure systems at E3PlW with that of smoke from ci garettes sampl ed from the bi oassay, returned to ORNL, andd smoked under analytical conditions. Comparison is made between both the absolute level of constituents and relative levels, normalized for nicotine delivery. In terms of relative levels, the analytical smoking produced some significantly different canstituent concentrations from that of exposure system inputs. Relative concentrations of constituents in the Series IV smoke condensate (2) are closer to those obtained under BTdW exposure system smoking, suggesting that the analytical values may be higher than anticipated, rather than the exposure values being lower. For Code 29, the relative concentrations in the exposure input samples are not substantially different from those determined from'Si~te Visit I, whereas the analytically generated smoke appears to be significantly different. Generally, values obtained for exposure system smoking for the three Codes appear to be consistent for both site visits. This is certainly desirable. Whether the particulate

phase chemistry being obtained on-site is substantially different 'from that obtained from analytical smoking remains an open question. Table X compares the smoke particulate phase chemistry of the Code 29 smoke generated by the exposure system at BPdW with that which is actually found inside the exposure chamber. The relative constituent concentrations are very similar, suggesting that the particulate phase of the smoke which the animals breathe is chemically identical to that which is generated by the cigarettes. That is, there is apparently no rapid selective deposition of smoke constituents inside the exposure chamber. References 1. Higgins, C. E., T. M. Gayle, and J. R. Stokely, °Sensor for Tobacco Smoke Particulates in Inhalation Exposure Systertts," submitted to Beitrage zur Tabakforschunc~, October, 1977. 2. Griest, trl. H., et al., "Final Smoke and Condensate Data for the Fourth Series of Experimental Variants," NCI/S&HP/ORNL Topical Report No. 52.

TABLE I Comparison of Code 04 Static Burn Rate: Exposure Condi ti ons at BP#19 vs Analyti ca1 Smoki ng Coneii ti ons at ORNL Site Burn Ratea mm/min Burn Rateb mg/min Rm 328, BNW 5.18 ± 0.11 73.4 ± 1.9 Rm 332, BNW 5.16 ± 0.15 71.9 ± 2.3 Analytical Smoking, ORNL 4.75 ± 0.24 67.0 ± 3.2 aBurn rate given in terms of actual length of cigarette burned. bBurn rate given in terms of actual weight of cigarette burned.