Tobacco Documents Online

Authors: Date: Departments: Divisions: Previous Reports: Proj. No./Name: Document Number: No. of Pages: Title: OBJECTIVE: Roger A. Jennings James C. Walker Walter T. Morgan John H. Reynolds IV Apri19, 1991 Biochemical/Biobehavioral Technical Support Biobehavioral Statistical Support/Al BIOBMM91-005; BIOBMM 91-008; BIOBMM91-010 502 / Product Understanding BIOBMM 91-026 14 R000999 RJaR&D SCIENTIFIC INFORMATION SERVICES LIBRARY EFFECT OF CHEMICAL STIMULANT CONCENTRATION ON THE PERCEPTION OF DRAW To evaluate the effect of different concentrations of aerosolized ethyl alcohol, a model stimulant, on the perception of puff resistance using a measurement system which incorporates controlled delivery of aerosols through a pressure drop (p.d.) standard, recording of puff behavior, and ballot ratings of sensory attributes. SUMMARY: The effect of an oral stimulant (ethyl alcohol) on puffing behavior and perceived puffing resistance was determined in 24 subjects. Aerosols of water or water solutions of 4 concentrations of ethyl alcohol (ETOH) were generated by an ultrasonic nebulizer and puffed through glass pressure drop standards (PDS) having a nominal pressure drop of 100 mm water (at a flow rate of 17.5 ml/sec air). A PDS of 100 mm water, through which only air was puffed, served as mental reference point. This PDS was sampled just prior to puffing on the test PDS. Perceived puff resistance showed a significant decrease with increasing concentrations of ETOH. Measured mean draft showed a significant decreasing linear effect with increased ETOH concentration, especially for concentrations greater than 25%. This indicated a modification to puffing behavior during the puff in response to the perceived mouth sensation. These data, in concert with results from two previous studies, indicates that in a simplified model cigarette, one of the sensory cues that determines perceived resistance to a puff is the chemosensory stimulation of the oral cavity. STATUS: This study is complete and is the third in a series of studies with similar objectives. td O G7 © Cb Cl+ KEY WORDS: draw; perception of draw; puff; aerosol; draft; flow; pressure drop standard; chemical; resistance; ethyl alcohol; 64-17-5; human; sensory

INTRODUCTION Draw or "lit resistance" is used routinely to evaluate product acceptance. The perception of draw during a puff is known to be affected by a number of product manipulations (Reynolds et al., 1983; Greene, 1987a; Teot ,1987; Green, 1987b; Gignac, 1989a; Smith, 1989; Liu, 1990; Gignac,1989b; Gordin, 1990; Gignac, 1990). These studies have also shown that the perception of draw is an important sensory component of the smoking experience. Thus it is important to understand the bases for this perception so that it can be optimized for different products. Three studies by Jennings et al. (1991a, b, c) demonstrated that oral stimulation can affect both the perceived resistance and puffing behavior. Perceived resistance was shown to be inversely related to the degree of chemosensory stimulation provided during the puff. The studies by Jennings et al. strive to determine how chemical-draft interactions modify perceptual ratings of puff resistance. The ultimate goal of these studies is to identify the role of perceived draw in smoker preferences and to suggest ways to manipulate the perception of draw beneficially. The present study was designed to further explore the effect of an oral stimulant on the perception of puff resistance in a large group of both male and female smokers by means of a proven measurement system which incorporated controlled delivery of aerosols through pressure drop standards (PDS), objective measures of puff behavior, and subjective measures of key sensory attributes. The primary focus of this study was to determine if decreasing concentrations of ETOH would produce decreased perceptions of mouth feel and reductions in perceived resistance. We had previously shown that aerosolized 100% ethanol could significantly lower perceived puff resistance and that the sensory cues that drive perceived puff resistance occur during a puff and are localizable to the mouth. Furthermore, these same sensory cues drive puffing behavior. METHODS Subjects. Twenty-four subjects (12 male and 12 female) from Bowman Gray Technical Center were recruited for this study. Each was an experienced smoker of either Premier or an ultra-low tar brand. Ages ranged from 29 to 47. All subjects were free of respiratory ailments or allergies at the time of testing. Each subject was informed that he/she would be puffing on a device which would yield an aerosol of ethyl alcohol. Subjects were asked to sign a Consent Form and to refrain from smoking, drinking or eating at least 30 minutes prior to the test session. This experiment employed the apparatus and methods described earlier by Jennings et al. (1991a, b). Briefly, the system consisted of the components described below. Chemical Aerosol Generator. The test device for this experiment was composed of a Puff- Profiler probe-PDS' connected to the ultrasonic nebulizer. An aerosol was produced in the headspace of the nebulizer unit by a high frequency sonicator. Before each trial, the ' Pressure Drop Standards. Glass multi-capillary PDSs were obtained from Celanese Corporation, P.O. Box 32414, Charlotte, N.C. 28323-2414. A PDS with 100 mm water pressure drop (at a flow rate of 17.5 ml/sec air) was used in both the aerosol generator and as a reference. The actual pressure drop through the 100 mm PDSs was confirmed by RJRT QA Technical Services. When not in use, PDSs were cleaned with ethanol, vacuum dried and stored in sterile polystyrene test tubes. © 2

experimenter activated the nebulizer and presented the probe-PDS unit to the subject through a hole in the partition. After the subject took a puff on this device, the nebulizer was turned off. Puff-Profiler System. A Puff-Profiler computer was used in the present experiment. This device was developed in the Biobehavioral Division to measure human puffing behavior. The system consists of a customized PC circuit, microprocessors, and application software configured to process analog input from two highly sensitive pressure transducers linked to puff-profile probes'. Prior to the beginning of an experiment, the Puff-Profiler computer was calibrated. The calibrator device measures the pressure transducer outputs (voltage) at two different pressures and uses these voltages to compute calibration curves from which a pressure is assigned to each voltage. In addition, the Puff-Profiler computer was set to report puff volume, puff duration, mean flow, peak flow, mean draft, and peak draft for each puff. This information was sent to a printer after each puff. All puff parameters were retained in computer memory and written to disk at the end of each experimental session. Test Chemical. Dehydrated 200-proof punctilious ethyl alcohol (U.S. Industrial Chemicals Co., Tuscola, Ill) and a dilution series of 25, 50, and 75% (v/v), was used as the chemical stimuli. Sterile filtered deionized water was used as the diluent and control. Ballot. Psychophysical responses to "puff resistance" and "intensity of mouth feel" were entered directly into an Apple IIe computer by the subject using an electronic mouse and a 0,! Basic program. The program displayed the attribute scales and stored the subject's responses. This program was a modified subset of a larger program which controls an CD apparatus for the automated measurement of the responses of humans to odorants (Walker et C) al., 1990). After a puff was taken on the test device, the subject clicked the mouse button to access the first question, "puff resistance". Using the mouse, the subject was able to position CD a marker along the unstructured line scale at the point corresponding to the perceived puff resistance. The reference terms Easier, Ref, and Harder related to the puff resistance of the ~ test device relative to one 100 mm reference PDS whose perceived resistance corresponded to C7, the Ref position on the ballot scale. A click of the mouse button identified the selected puff resistance and advanced the computer program to display the next question, "Intensity of CI" Mouth Feel". This attribute had a range from "None" to "Extreme" to verify the degree of a chemical stimulus. The unstructured line scale of both psychophysical ratings had a range of values from 0 to 40. Ballot ratings were written to disk, converted to ASCII and transferred c:3 via a computer program (Kermit ) to a PC-DOS system. CJ Test Room. The perception of draw measurement system was assembled and used in a 586 n ft3 clean-room located in Room 103, Building 611-13, Floor 1, West Laboratories, Bowman Gray Technical Center. The room had a laminar airflow of 1000 cfm, providing 102 air changes per hour. Approximately 80% of the air was recirculated and 20% imported from ambient air outside the test room. Temperature was maintained at 72°F. The white metal- walled room was illuminated by three 48-inch florescent bulbs behind Lucite' diffusers. Sound-proof entryway doors provided excellent insulation from external noises. Experimental Procedure. Each subject was tested individually. During the experiment, the subject faced a partition flanked on the left by a TV monitor, which played taped entertainment, and on the right by a computer monitor and electronic mouse. The partition, 2 A detailed explanation of how the puff profile probes work can be found in Reynolds, Norman and Gordon (1983). 3

which separated the subject from the experimenter, contained one permanently positioned glass PD6 (reference standard) having a nominal draft of 100 mm water. A 2" diameter hole between these fixed PDSs was used to pass the test standard -connected to the nebulizer- to the subject during each trial. Five nebulizers were positioned behind the partition and filled with either 5 ml of distilled water or one of 4 concentrations of ethyl alcohol. Glass PDSs, fitted with puff-profiler probes and sterile cardboard mouthpieces, were placed in the labeled holes of a vacuum manifold. Constant vacuum on the PD6s not in use removed liquid from their inner walls. Prior to the beginning of the test session and before entering the test chamber, the experimenter showed the subject a nebulizer with the PD6 and puff-profiler probe attached (see Appendix 1) and explained that "the nebulizer produces an aerosol which travels through the PDS to the mouth during a puff." After review of a consent form explaining the study and use of ethyl alcohol, the subject was asked to sign the form and was escorted into the test room and seated in front of a partition. An explanation of the experimental procedure was given to the subject3. Before the session began, the subject was asked to puff on the reference standard and to practice answering ballot questions with the electronic mouse. Once the subject understood his/her role in the experiment, the experimenter closed the room's entryway door and took his seat on the opposite side of the partition. The beginning of the trial was signalled by the appearance of the test standard through the hole in the partition. The experimenter held the test device in position and the subject took one puff on the 100 mm reference PDB. This was done in order to establish a mental picture of a "reference" resistance. The subject then took a puff on the test device and immediately rated the two sensory questions. The first question, "puff resistance", had a 40- point unstructured line scale with the labels Easier, Ref, and Harder. The "Ref" label was anchored midway along the line and corresponded to the resistance of the reference PDS. The second question, "mouth feel", also had an unstructured line scale with the labels none and extreme on either end. C) C) Experimental Conditions. This experiment was conducted using 5 conditions presented to ® 3"You will be comparing the puff resistance of the test device that I project through the hole in the partition with the reference standard that is permanently fixed in the board in front of you. When you see the test device coming through the hole, you are to take one normal puff from the reference standard then one normal puff from the test device. Next, go to the mouse, click the button once to access the first ballot question, and move the mouse to position the plus symbol on the point along the horizontal line scale that corresponds with the degree of resistance felt during your puff on the test device. This resistance rating should be compared to the resistance felt while puffing on the reference standard. Notice that the line scale is labeled at one end with the word 'Easier' and the other end with the word 'Harder', and the term 'Ref' anchored midway along the line. 'Ref' corresponds to the resistance felt when puffing through the fixed reference standard. To enter your response to the question, click the mouse button. This records your response and moves you to the next question which concerns 'Mouth Feel.' This question indicates if you felt the ethyl alcohol. Rate 'Mouth Feel' relative to your life's experiences from none to extreme. The feeling of grain alcohol or straight vodka would be 'extreme' for most people. Click the mouse button to enter your response. This concludes a trial. There will be 15 trials in this experiment. I will verbally guide you through the first trial. During trials that expose you to ethyl alcohol, I will ask you to take a small sip of water and swallow after you have answered the ballot questions. Please remain seated during this session. Please do not talk unless something is wrong or you forget what to do. Please relax and enjoy the nature movie which will be playing throughout this session." n 4

24 smokers (Table 1). The order of presentation of each block of 5 conditions was randomized within each test block for each subject. In this manner, each subject experienced each of the 5 conditions during the first test block before proceeding into the second and third block. The inter-trial interval was approximately 120 seconds. During each trial, the subject took a puff on the reference PDS and one puff on the test PDS. The subject completed the trial by answering the computer-generated questions. Table 1 Condition 1 2 3 Test conditions presented 3 times each during a 15-trial session. Nominal Pressure Drop (mm water) Aerosol Content Description 100 Water Test Standard Control 100 25% ETOH Test Standard 100 50% ETOH Test Standard 100 75% ETOH Test Standard 100 100% ETOH Test Standard 'Ci C) Data analysis. Data originated from three sources: 1) Puff-Profiler computer (puff &-p parameters); 2) Apple Computer (ballot scores); and 3) Lotus spreadsheets (randomized © condition sequences and subject information). These data were merged and uploaded to the VAX where they were analyzed statistically using SAS. A multivariate approach (repeated C) measures analysis of variance) was employed. Since each concentration level was evaluated ~ during a single session for each subject, ratings of the different levels might be expected to be correlated. Accordingly, the data were averaged over the three replicates and levels of ETOH ~. concentration were compared using the repeated measures analysis. Hypotheses of overall differences between the levels and more specific hypotheses of linear effects of concentration were tested in this framework. Hypotheses of higher order effects were also tested for ~.3 completeness, but greater emphasis was placed on the tests for linear effect because they indicate correlations with the concentration levels. Because of the variation observed in C-3 earlier experiments, a significance value of 0.10 was used in these tests. Definitions. This report makes reference to draft, "pressure drop" or p.d. in three contexts: 1) nominal p.d. (the p.d. produced by a multi-capillary glass standard when air is passed through it at a volume flow rate of 17.5 cc/sec); 2) actual p.d. (the measured p.d. resulting from a subject puffing on a p.d. standard); and 3) perceived p.d. or "resistance" (the ballot rating, made by a subject, of perceived resistance while puffing on a p.d. standard or the test device). RESULTS The repeated measures analysis of variance showed no significant effects of sex, usual cigarette brand, or order of test condition presentation on the response variables (ballot ratings and puffing parameters). There was no evidence that subject differences accounted for the effects observed for the ballot ratings or puff parameters. Repeated measures analysis C) 5

of variance was performed to test hypotheses of overall differences (that at least one test condition was different from the other four) and of linear effects (that changes in subject response was consistent from one ETOH level to another). The analysis showed significant discrimination between the ETOH concentration levels for both overall difference and for the more specific linear effect hypotheses, for all of the ballot responses and some of the puff behavior responses. Table 2 presents significance levels for the observed effects. Table 2 Significance levels for all observed effects. Values are from a repeated measures ANOVA. The first value refers to the subject response to each ETOH level where "subject-by-ETOH level" constitutes the error term (univariate approach). The second value is the significance when the multivariate approach is used. Three blocks of 5 ETOH levels were analyzed. There were 24 subjects. A value of 0.10 or less was considered significant for this analysis. Values meeting this criterion are in boldface. Response Variable Overall Linear Puff Resistance' 0.00/0.00 0.00/0.00 Puff Resistance2 0.00/0.00 0.00J0.00 Mouth Feel 0.00/0.00 0.00/0.00 Puff Volume 0.35/0.10 0.14/0.23 Puff Duration 0.50/0.66 0.09/0.19 Puff Peak Draft 0.80/0.93 0.26/0.41 Puff Mean Draft 0.02/0.22 0.00/0.02 Puff Peak Flow 0.38/0.22 0.07/0.08 Puff Mean Flow 0.59/0.65 0.58/0.59 Peak:Mean Draft3 0.02/0.17 0.00/0.00 Peak:Mean Flow3 0.01/0.00 0.00/0.00 WoW 0.17/0.55 0.02/0.09 'Raw ballot ratings; 2 Normalized ballot ratings; 3Calculated ratio; 4Calculated by multiplying the mean draft times the mean flow (discussed further in this section). 6

Effect of ETOH concentration on ballot ratings of mouth feel and perceived puff resistance Ballot ratings of mouth feel increased significantly with increasing concentrations of ETOH in the aerosol (Figure 1). Ballot ratings of puff resistance showed a significant decrease with increasing concentration, both on a raw and subject-adjusted (normalized) basis (Figures 2 and 3), indicating that the increased mouth sensation resulted in a perceived reduction of resistance, even though the nominal p.d. was held constant. 30.0 T Extreme g 25.0 Y 20.0 s ~ 15.0 ~ 10.0 I 5.0+ . None 0.0 ~ 0 .'30 T 25 50 75 Ethanol concentration (x) T A I T 100 T\ 1 • 1 15 0 25 50 75 100 Ethanol concentration (X) 0.2 0.0 Difficult 1 ~ i Easy 0 25 50 75 100 Ethanol concentration (X) Figure 1. Mean LSEM) mouth feel rating from computer-generated ballots. Arrows indicate positions on the electronic ballot corresponding to "None" and "Extreme" oral sensations felt during the puff. Figure 2. Mean (+SEM) perceived puff resistance rating from computer-generated .~ ballots. Arrows indicate the direction along the ballot line scale corresponding to ~ "Easy" and "Difficult or Harder" perceptions a, of puff resistance. The resistance of the 100 mm reference standard was anchored © midway along the 40-point scale. C:) 0% 0% Figure 3. Mean (±SEM) perceived puff ~~ resistance rating from computer-generated c:D ballots. Averages are from data that was normalized prior to averaging. No significant differences were observed between this treatment and mean scores in Figure 2. 7

Effect of aerosol content on puffing behavior While figures 4 and 5 show that both puff volume and duration decreased slightly on average as ETOH concentration increased, the change was not found to be significant, either on an overall basis, or in a test for strict linear effect. Similarly, peak draft did not change significantly with increased ETOH concentration, but mean draft showed a significant decreasing linear effect with increased ETOH concentration, especially for concentrations greater than 25% (see Figures 6 and 7). This indicates a modification to puffing behavior during the puff in response to the perceived mouth sensation. 160 - ~ 120 1 ~ 100 ~ eo so 40 20 3.0 2.6 2.6 2.4 2.2 2.0 1.a 1.6 1.4 1.2 1.0 0.6 0.6 900 so0 700 600 500 400 ~0 200 100 I I I I 1 0 25 50 75 100 Ethanol concentration (X) f ~-~ I I I I I 0 25 50 75 100 Ethanol concentration (X) I I I i I 0 25 50 75 100 Ethanol concentration (f) Figure 4. Puff-profiler measures of mean (+SEM) puff volume at each ETOH concentration. The y-axis was scaled according to the observed minimum and maximum puff volumes. Figure 5. Puff-profiler measures of mean (+SEM) puff duration at each ETOH concentration. The y-axis was scaled according to the observed minimum and maximum puff durations. Figure 6. Puff-profiler measures of mean (+SEM) puff peak draft at each ETOH concentration. The y-axis was scaled according to the observed minimum and maximum puff peak drafts. 8

Figure 7. Puff-profiler measures of mean (+SEM) puff mean draft at each ETOH concentration. The y-axis was scaled according to the observed minimum and maximum puff mean drafts. 200 - 100 •---s--.--. i 0 25 50 75 100 Ethanol concentration (X) In Figure 8 and 9, the increase in peak flow rate from 0% to 25% ETOH concentration was marginally significant (p=0.09), and the apparent decrease in mean flow rate (from 25% to 50% ETOH) was not significant. 100 90 eo 70 so 50 40 30 20 0 25 50 75 100 Ethanol concentration (X) Figure 8. Puff-profiler measures of mean c) (-+SEM) puff peak flow at each ETOH level. © The y-axis was scaled according to the observed minimum and maximum puff C) peak flows. C) Figure 9. Puff-profiler measures of mean (±SEM) puff mean flow at each ETOH level. The y-axis was scaled according to the observed minimum and maximum puff mean flows. I 1 I 1 1 0 25 50 75 100 Ethanol concentration (x) 9

Table 3 summarizes the puff parameter data presented above in Figures 4 through 9. Only the trends shown for mean draft are significant. Table 3 Summary table showing directional change, at each ETOH concentration, of each puff parameter relative to response to water aerosol. Additional response variables were calculated to explore further the effects of ETOH concentration. These were ratios of peak-to-mean draft, peak-to-mean flow, and work, calculated as mean draft times mean flow (see Figures 10 through 12). Both peak-to-mean ON draft and peak-to-mean flow showed strongly increasing effects with Increased ETOH concentration (p=0.0005 and 0.0002, respectively). These results are compatible with the 0 hypothesis of consistent peak draft applied at the beginning of the puff, followed by an C) increasing reduction in draft applied for the remainder of the puff due to increasing mouth feel, resulting in the observed reduction in mean draft. In other words, the relation increased C) because while the numerator was relatively constant, the denominator became smaller with increased ETOH concentration. The flow rate ratio change simply followed the draft ratio C) change. The decrease in work with increased ETOH concentration shown in Figure 12 was n, marginally significant (p=0.10) and reflects further the reduction of mean draft with increased ETOH concentration. CY'` 1.6 e ~ C 1 7 I . . t i JI 1 t 1.6 v 1 i 1.5 0 1 1.4 0 25 50 75 100 Ethanol concentration (X) Figure 10. Mean (±SEM) peak-to-mean draft ratio at each ETOH level. 10