Tobacco Documents Online

I I I I I I I I I I I I I I I I Ps~choonarm~ccioeti ! 1~9? 11,. g Psychophannacolob ( Sprinser-`'zr1aQ 1993 Nicotine or tar titration in cigarette smoking behavior? A Hasenfratz, B. Baldinger, and K. Battig Comparative Physiology and Behavioral Biology Laboratory. Swiss Federal Institute of Technology. ETH-Zentrum. Turnerstr. 1. CH-8092 Zurich. Switzerland Received August 10. 1992 Final Version January 11. 1993 Abstract. A significant problem in assessing the relative relevance of nicotine and tar yield for compensatory smoking after switching from high to low yield cigarettes is that nicotine and tar yield are highly intercorrelated across conventional cigarettes and that the tar/nicotine ratios vary only within a modest range. A better differenti- ation between the impacts of nicotine and tar yield was expected by comparing in a laboratory experiment a new low nicotine; medium tar cigarette ("Next") with conven- tional low nicotineilow tar (ultra-light) cigarettes and with medium nicotinertnedium tar cigarettes with respect to nicotine absorption and physiological effects. Twelve females, habitually smoking medium type cigarettes ( _ 0.7 mg nicotine) participated in the study. Neither the number of cigarettes smoked under field conditions nor the puffing behavior during the laboratory experiment differed between the three types of cigarettes. In the laboratory. Next produced only very small increases in plasma nicotine and changes in cardiovascular or EEG measures. whereas the effects of the medium cigarettes were in the expected range and those of the ultra-light cigarettes about halfway in between. The nicotine absorp- tion nicotine yield and the CO absorption/CO yield ratios were similar for Next and the habitual cigarettes, but about twofold higher for the ultra-light cigarettes. This suggests that gustatory and olfactory sensations, which are supposed to be more dependent on tar than on nicotine yield, may play a greater role for the regulation of smoking behavior than hitherto believed. Key words: Cigarette smoking - Tar/nicotine ratio - Smoking behavior - Compensation - Cardiovascular It has been widely documented that the absorption of nicotine and CO through cigarette smoking does not parallel machine smoking yields for CO and nicotine. Nicotine yields between 0.7 and 1.2 mg were seen to result in similar plasma levels of nicotine, but smoking ultra- Correspondence to: M. Hasetifratz light cigarettes with around a fourfold lower yield in nicotine resulted in only about twofold lower nicotine plasma levels (Hofer et al. 1991). In contrast, increasing the nicotine yields or requiring forced puffing does not result in adequate increases of plasma nicotine. This phenomenon of "up titration" for light and "down titra- tion" for strong cigarettes has generally been attributed to nicotine as the pharmacologically active ingredient of tobacco smoke (for a review see Nil and Battig 1989). The most important arguments were provided by studies in- vestigating the effects of nicotine receptor blockade on puffing behavior. Rose et al. (1989) found that eight smokers, smoking with a smoke mixing device after the application of various levels of mecamylamine, increased their self-administered nicotine at the highest doses (10 and 20 mg) of mecamylamine. Using similar doses of mecamylamine, Nemeth-Coslett et al. (1986) found that the higher doses increased the number of puffs per ciga- rette, puffs per session and the CO uptake, and decreased the intervals between cigarettes and between puffs within a cigarette in eight freely smoking subjects. There are, however, also some observations that do not fit into this concept. Evidence that not only nicotine but also the tar content might play an important role was presented in several experiments. Battig et al. (1982) in- vestigated the puffing behavior of 43 female and 67 male smokers and found that, for the women, partial correla- tion procedures suggested nicotine to be more important in determining puffing behavior than CO and condensate yield. However, there were also no women smoking the vt strongest cigarettes (1.3-1.7 mg nicotine yield) in tha study. A similar analysis of 55 smokers was done by Sutton et al. (1982), who found that the total puff volume was a more important determinant of peak blood nicotine concentration than the nicotine or tar yield of the ciga- rette. On the other hand, since tar yield was the most important factor for the total puff volume, they concluded that the smokers were compensating for a reduced de- livery of tar by puffing and inhaling a greater volume of smoke. A different approach was used by Stepney (1981), where 19 smokers switched from their habitual brand to a I

I I I I I I I I I I I I I I I I I I low tar low nicotine and to a low tar. medium nicotine cigarette at intervals of 3 weeks. Analvzing '4-h urine levels of nicotine and cotinine as well as several puffing parameters. he found that both low tar brands were "oversmoked" relative to the subjects' usual medium tar cigarettes. He therefore proposed that tar yield was more important in determining puffing behavior than nicotine vieid. Adopting a similar design. Armitage et al. (1988) observed that smoking intensities decreased progressively from low tar to "maintained nicotine" to middle tar cigarettes. although the amount of nicotine absorbed from maintained nicotine and middle tar cigarettes was similar to and significantly greater than the levels reached with the low tar cigarettes. Thus, they concluded that smokers may adjust their smoking behavior to achieve not only an acceptable nicotine level in the bloodstream but also an acceptable tar level, which seems to be important for sensory satisfaction. However, a common problem for most of these studies is that the nicotine and tar yields of cigarettes are highly intercorrelated (Nil 1991) and that the tar/nicotine ratios vary only within a moderate range for the usual com- mercial cigarettes (approximately between 5 and 15). A better dif~erentiation between nicotine and tar yield is offered with the new brand "Next." which was available from a test market in Florida. USA. This cigarette has a nicotine yield lower than that of "ultra-light" cigarettes but a tar yield comparable to medium yield cigarettes (relations for Switzerland, cf Hofer et al. 1991), which results in a tar/nicotine ratio of 116. In order to-assess the effects of nicotine and tar yield after switching from medium to low yield cigarettes, three types of cigarettes were compared in the present study: a) the habitual cigarette of the subjects with a nicotine yield of 0.7 mg or more, b) an "ultra-light" cigarette, chosen as the cigarette of the same taste category with the lowest nicotine yield, and c) Next. Since all subjects habitually smoked cigarettes with nicotine yields of 0.7 mg and above, compensatory smok- ing behavior could be expected with one or both of the two low nicotine cigarettes, depending on the relative importance of nicotine and tar yield. If nicotine yield were the more important factor for compensation, correspond- ing effects could be expected for both low nicotine cigar- ettes. On the other hand, if tar yield were the more important factor, corresponding effects could be expected from the ultra-light cigarettes only. The analysis of physiological data was introduced in order to assess possible differentiations between the general effects of smoking and the pharmacological actions of nicotine. Materials and methods Subjects Twelve female regular smokers smoking cigarettes with a nicotine yield of at least 0.7 mg participated in the study. Their mean age was 27.9 years (range 22-36) and they reported smoking 21.1 cigarettes per day (range 15-30). On testing days, they were required to abstain from smoking and drinking caffeine-containing beverages upon getting up until the beginning of a session. The subjects were selected from responders to an advertisement in a local newspaper and their fee was SFr. 200. - for completing three test sessions and r:~or iing their cigarette consumption on the dav preceding each of the te,t sessions. Three types of cigarettes were compared in a within subiect com- parison. They differed from each other in nicotine and tar yte(d as summarized in Table 1. The ultra-light cigarettes were of the same brand as the habitual cigarette or. if not possible, of the same type of tobacco tMaryland or American Blend). The third cigarette was the new brand "Next", obtained from a test market in Florida, USA, which has a nicotine yield comparable to the ultra-light cigarettes but a tar yield comparable to the habitual cigarettes. The smoking order of the three types of cigarettes was balanced. PhYsiological parameters Blood pressure. Blood pressure was measured from the control room using an automatic measuring set (arm cuH': Tonomed Electronic. Speidel & Keller AG. Germany) immediately before and after each rest phase. The following signals were continuously recorded with an AT- compatible micro computer and stored on streamer tape for later off-line analysis. ECG. The electrocardiogram was recorded with Beckman Ag,,AgCl electrodes fixed below the middle of the right clavicle, below the last rib on the left and, for the reference electrode, below the last rib on the right. The R-wave peaks of the ECG were detected using an ECG cardiometer (Cardiotronics AG. Stockholm) and digitally recorded. Finger and ear pulse amplitude and arrival time. Miniature photo- sensors were placed at the palmar surface of the distal phalanx of the left ring finger and at the left earlobe. The finger and ear pulse arrival times were computed as the time between the R-peak of the ECG and the point at which the finger or the ear pulse amplitude began to increase. EMG. The electromyogram of the musculus frontalis was recorded with three Beckman Ag/AgCl electrodes arranged in a horizontal line in the middle of the forehead. Respiration. Respiratory amplitude and frequency were registered with a Pneumotrace respiration transducer Imodel 1130, Morro Bay, CA 93442). Body movement. Body movements were measured with four piezo- electrical crystals, centrally installed under the seat. The impulses of the three dimensions were recorded as sum vector. EEG and EOG. EEG activity was recorded with goldcup electrodes from C. (international 10-20 system). Combined ear references with resistances between them were used and a midforehead electrode Table 1. Mainstream smoke yields of the three cigarette types Habitual Ultra-light Next Tar (mg/cig) 10.1 (7.0-14.0) 1.83 (1.0-3.0) 9.3 Nicotine (mg/cig) 0.81 (0.7-1.0) 0.22 (0.2-0.3) 0.08 Tar/nicotine ratio 12.4 (10.0-15.5) 8.33 (5.0-10.0) 116.25 CO (mg/cig) 9.6 (7.9-11.9) 3.3 (2.0-5.5) 9.9 Entries are: means and ranges in parentheses 9046.399501

1\\ I I I I I I servtd :L, Qround. All electrode impedances were kept beiow 5 kOhm. The stgnals were amplified with bandpasses from 0.2 to :5 Hz. EOG acttvtty was monitored with Beckman Ag,/AgCI elec- trodes placed below the left infraorbital ridge and above the left supraorbital ridge. The signal was monitored with a bandpass setting of 0.5-25 Hz Biochemical parameters Nicottne und cortnine. Plasma concentrations of nicotine and cotinine were analyzed by Dr. C. Feyerabend at the Nicotine Laboratory, Poisons Unit. London, by capillary column gas-liquid chromatography (Feyerabend and Russell 1990). Blood samples (10 ml) were collected in heparinized glass tubes and stored at room temperature. After the end of a se:ssion, the samples were cxntrifuged and the plasma was kept frozen at - 70 °C until analysis. The reproducibility was tested with 2 x 15 blind double probes (doubles of the randomly selected original probes). The test-retest correlation coefficients for nicotine and cotinine were 0.976 and 0.999, respect- ively. Carbon monoxide (CO). The CO concentration of the expiratory air was measured using the EC50 Micro Smokerlyur (Bedford Instru- ments, England). When thts was the case. the time interval was recorded and the same procedure restarted with a third 4-min rest phase. Data processing and statistics The blood pressure measurements before and after the rest periods were aggregated to one mean value for each rest pertod. All continuously recorded physiological data were anaiyzed off- line so as to obtain the mean values and standard deviations for each suctx.ssive 10-s period. After a visual artefact control carried out under blind conditions, the 10-s averages were aggregated to means for each 4-min rest phase. From the EEG data, the relative power of the delta (1.0-3.9 Hz)L theta (4.0-7.8 Hz), alpha (7.9-11.7 Hz) and beta bands (11.8-24.7 Hz) as well as the peak frequencies of the aipha and beta bands were determined for each 4-min rest period. These reducxd data sets as well as all other subjective and biochemical parameters were submitted to ANOVAs (BMDP 2V) with the factors type of cigarette (T: habitual, ultra-light. Next), pre- versus post-treatment phases (P) and ordinai number of cigarette (N: first versus second cigarette). For comparisons between the three types of cigarettes post hoc contrast analyses (Rosenthal and Ros- now 1985) were performed. For all significance levels of the analyses of variances Greenhouse-Geisser probabilities were considered where appropriate. I I I I I I I I Subjectiae parameters Subjective ratings of the craving to smoke, taste and strength of the cigarette, enjoyment of and dizziness after smoking were assessed using l00-mm scales presented on a computer screen. The cursor was positioned on the scale using a trackball. [Craving: "How much would you like to smoke now?" (left end (0 mm) labelled with: not at all: right end (100 mm) iabelled with: very much): taste: "How good was the taste of the cigarette?" (not at all good taste-very good tastek strength: "How strofig did you find the cigarette?" (weak-strong); enjoyment: "How much did you enjoy the cigarette?" (not at all-very much): dizziness: "Do you feel dizzy now7" (not at all-very much).] Smoking behavior During the day precxding a test session the subjects had to note the time each time they lit a cigarette and had to coilect the butts for control. During the smoking penod of a test session the total number of puffs, the puffing intervals for each cigarette and the time from the end of the first cigarette until the subject signaled the wish to light the next one were recorded. Procedure Each subject took part in thrae test sessions which were always 1 week apart and which took place in the morning. For the day preceding a session the subjects were given a sufficient supply of the type of cigarette to be smoked in the test session. After the arrival at the laboratory, the subjecu filled out ques- tionnaires, a cannula was inserted into the left forearm vein and a first blood sampie was collected. Then, after a CO measurement the first 4-min rest phase (with blood pressure measurements before and after) started, during which the subjects had to sit quietly with their eyes closed. After that a second CO measurement was done, the subjects completed analog scale ratings and then they staned to smoke a cigarette as naturally as possible. Two minutes after the last puff, the second blood sample was collected, CO was measured, further analog scaie ratings were completed and the second 4-rtvn rest.phase was started (5 min after the last puff), again with blood pressure measurements before and after. After this rest phase, a third CO measurement was done and then the subjects were freee to read women's magazines untii they wished to smoke another cigarette. Results Cell means of all parameters which differentiated between the three types of cigarettes and the correspondent ANOVA F-values and significance levels are presented in Table 2. Biochemical parameters Plasma nicotine increased significantly from pre- to post- smoking (factor P) and this increase was dependent on the type of cigarette (interaction T x P), with the habitual cigarette producing the greatest increase, Next producing only a negligible one, and the ultra-light cigarettes in between. The nicotine absorption/nicotine yield ratios for the first cigarette of a test session (Fig. 1) reached very similar values for Next and for the habitual cigarette, whereas with the ultra-light cigarette this ratio was about two times higher. Plasma cotinine levels, as indicators of nicotine expo- sure on the preceding day, showed a differentiation be- tween the three types of cigarettes similar to that of plasma nicotine absorption (factor T), with the highest levels for the habitual cigarettes, the lowest ones for Next and medium ones for the ultra-light cigarettes. A cotin~te level/nicotine yield ratio was not analyzed because (he' time was too short to achieve steady states in cotinifig levels. *4 For respiratory CO, only a very weak effect of the type of cigarette was obtained (interaction T x P), which was ~ due to the difference between habitual and ultra-light ~ cigarettes [contrast analysis: F(1,22) = 9.22, P< 0.01].,W, However, the CO absorption/CO yield ratios for the first,,= cigarette of a test session (Fig. 1) reached the highest value ~ for the ultra-light, the lowest for Next and an intetinediate ~ one for the habitual cigarettes. 0 I~

= ~6 Table 2. Cell means of dependent ~artables and summar% of atiO% A results I % anable - Cigarette Habitual Lltra-Itght Next ANOVA F-~alues=" Pre Post Pre Post Pre Post T N P T x N T x P ti x P I Binchentica/ parameters: Plasma nicotine I 5.5 '_0.7 4.6 13 0 4-1 5.S 1578** 130.,,*** t 1.43**` 76.57*** ;.94* [ng ml] 2 11.1 22.6 8-t 15.8 -3.3 6.3 Plasma cotinine 1 211.1 304.8 177.5 169.5 115.6 111.4 11.8-i*** 2.77 16.18** 3.83* 1.34 :0.18**' I [ng ml] 2 210.0 209.7 168.4 167.6 108.2 107.9 Respiratory CO I 10.4 15.4 10.4 13.9 10.2 14.1 0.83 68.69*" 95.29s** 1.83 4.56* 0.88 [ppm] '- 14.1 18.8 12.8 16.4 12.6 17.3 I Phrsiological parameters: Heart rate i 68.3 79.9 65.9 75.3 68.3 69.9 6.61•* 1.00 51.68*'* 7.88" 28.54*:* 6.13' [bpm) 2 72.4 77.3 69.8 75.2 67.1 68.1 Systolic blood 1 100.3 101.0 99.4 104.4 102.5 101.4 0.66 1.04 9.21' 1.20 4.03' 0.02 pressure [mm Hg] 2 100.0 103.8 103.5 104.3 101.3 100.5 I Diastolic blood 1 71.6 76.4 73.5 74.7 72.8 70.3 0.67 0.47 0.01 0.35 5.89'* 0.93 pressure [mm Hg] 2 71.1 72.3 73.4 73.5 74.0 68.0 Finger pulse 1 150.9 77.5 130.4 80.0 149.6 117.1 4?7* 53.17"* 77.22"* 2.31 5.06* 16.48" amplitude [units] 2 90.8 53.2 93.1 68.3 108.6 96.8 I Theta power [°o] ! 19.9 18.6 20.1 18.6 19.6 19.8 3.26 4.18 16.19'* 1.93 4.82' 0.05 , 19.4 18.4 19.3 18.2 19.9 19.7 Beta power [%] 1 32.8 35.9 319 34.4 33.0 32.9 4.60' 1.99 25.13'*' 0.02 5.75" 0.36 , 33.9 35.5 32.9 35.1 33.3 33.1 I Subiecrire parameters: Cravtng [°o] 1 80.6 31.6 79.7 44.7 71.9 44.9 0.32 0.22 64.49"' 0.94 3.53' 2.48 2 68.5 40.3 70.2 40.5 73.9 48.6 Taste [°°] 1 77.6 48.7 29.8 8.33** 0.00 0.67 2 79.2 46.0 31.4 I Strength [%] 1 59.8 21.9 27.5 14.83"' 0.27 0.34 2 57.5 18.5 29.4 Enjoyment [%] 1 70.4 50.8 37.1 5.26' 0.14 4.29' 2 75.5 45.9 40.1 I Dizziness [%] 1 45.5 26.8 9.2 7.01" 5.42' 5.93" 2 23.4 12.4 13.3 I I I I I I I I Entries are: cell means. F-values and significance levels. Abbreviations: T: Type of cigarette (habitual/ultra-light/Nextk N: Number of cigarette in a session (1 '_): P: Pre-versus post-treatment phase. Significance levels: 'P < 0.05: "P < 0.01: "'P < 0.001. The F-values of the triple interaction T x N x P were omitted as they did not reach significance except for plasma nicotine [F (2. 22) = 5.70'] Pht•siological parameters Among the peripheral physiological parameters assessed. only heart rate, blood pressure and finger pulse amplitude differentiated significantly between the three types of cigarettes. Pre- to postsmoking increases in heart rate and diastolic blood pressure were greatest for the habitual cigarettes, somewhat smaller for ultra-light cigarettes and nearly absent (heart rate) or even negative (DBP) for Next. However, in contrast to diastolic blood pressure, the pre- to postsmoking increase of systolic blood pressure after the first cigarette was greater for the ultra-light than for the habitual cigarettes. From the other parameters only a few significant smoking effects were obtained. These included a decrease in finger and ear pulse amplitude [P: F(1.11) = 5.57, P< 0.05] as well as a decrease in respiration amplitude (F (1,11) = 6.13. P < 0.05]. Although these effects tended to differentiate between the three types of cigarettes, only the decrease of finger pulse amplitude produced a corre- sponding significant interaction. The heart rate increase/nicotine yield ratio showed a result very similar to that obtained for the nicotine ab- sorption/nicotine yield ratio. with the ratios being similar for Next and the habitual cigarettes but about twofold higher for the ultra-light cigarettes. A similar picture was obtained for the EEG param- eters, although the pre- to postsmoking effects differed significantly for theta and beta power only. Whereas theta power decreased from pre- to postsmoking similarly for the ultra-light and habitual cigarettes, almost no changes were observed for Next, and a similar result was also obtained for the beta power increase. Subjective parameters Significant differences between the three types of ciga- rettes were obtained for craving, taste, strength, enjoyment and dizziness. Whereas the habitual cigarettes produced the greatest pre- to post-treatment decrease of craving and the highest ratings for the other parameters, Next and th~,' ultra-light cigarettes produced smaller and similar ratin~ 4114, Smoking behavior ~ ~ The number of cigarettes, the mean and the relatiWZ standard deviation of the mean intercigarette intervto An ~ W

I I I I I I I i I I I I I I I Nicotine absorption/nicotine yield so- so= 840~ E ~aoLa =20- habituai ultra-light NEXT CO absorption/CO yield habitual ultra-light NEXT Fig. 1. The nicotine absorptioninicotine yield ratio and the CO absorptioniCO yield ratio (means ± SE) of the ultra-light ciga- rettes are about twofold higher than those of NEXT and the habitual cigarettes. Significant differences (paired t-test compari- sons. df = 11) are indicated ("P < 0.01; 'P < 0.05) during the day preceding the test session (same type of cigarette as during the test session) revealed no significant differentiations between the three types of cigarettes (Table 3). The puffing interval, the number of puffs per cigarette and the latency until the subject wished to smoke another cigarette also failed to significantly differentiate between the three types of cigarettes. Discussion The aim of the present study was to determine the relative relevance of nicotine and tar yield for compensatory Table 3. Cell means of smoking parameters and summary of ANOVA results smoking after switching from high ~ ield to low yield cigarettes. Whereas Next produced almost no increase in plasma nicotine and no..changes in cardiovascular and EEG parameters, the effects of the habitual cigarettes were in the expected range as found earlier (Hasenfratz et al. 1990) and those of the ultra-light cigarettes were about halfway in between. This was the case for increases in heart rate. blood pressure, peripheral vasoconstriction and electrocortical arousal and a decrease in respiration amplitude. The analysis of the nicotine absorptioninico- tine yield, the CO absorption/CO yield and the heart rate increaseinicotine yield ratios revealed that tar yield seemed to be more important than nicotine yield, since the smokers of the present experiment compensated only while smoking the low tar~low nicotine cigarettes (ultra- light) but not the medium tar/low nicotine cigarettes (Next). All three ratios were similar for Next and the habitual cigarettes but about twofold higher for the ultra- light cigarettes. Further evidence was obtained from plasma cotinine levels. Although a single day of free smoking was rather short to achieve steady states. the plasma cotinine levels differed significantly between the three cigarette types. One recent study by Robinson et al. (1992) used a very similar cigarette with a tar/nicotine ratio of 123.3 (7.4 mg tar, 0.06 mg nicotine). Investigating heart rate and EEG effects, they found that, even though the nicotine yield was very small, the smoking of the test cigarette produced increases in plasma nicotine which were sufficient to produce small heart rate increases but not to affect EEG parameters significantly. A further remarkable result was that the test and the control cigarette were smoked very similarly, suggesting that these smokers also did not compensate for reduced nicotine. Finally, preliminary results of a similar study from this laboratory (not yet published), which assessed the puffing behavior while smoking the same three types of cigarettes, revealed that the mean and total puff volumes were comparable for Next and habitual cigarettes but con- siderably greater for the ultra-light cigarettes. The compensation for a reduced tar yield suggests that taste and odor of a cigarette may be important factors in determining smoking behavior. Rose and Levin (1991) concluded in their review that burning tobacco produces Cigarette Habitual Ultra-light Next T N T x N Smoking parameters during the sessions: Puffing interval [s] 1 20.4 17.8 18.0 2.39 0.66 0.27 2 20.7 18.0 19.3 Puffs/cigarette 1 12.3 12.3 12.8 0.22 0.07 0.90 2 12.0 12.7 12.2 Latency until 28.6 25.4 26.6 0.44 - next cigarette [min] ~ Smoking parameters during the presession dayr. Intercigarette 54.2 50.1 50.4 0.41 interval [min] Rel. SD [%] 71.5 69.8 73.7 0.09 CT. N cigarettes 19.8 19.3 19.3 0.04 W Entries are: cell means and F-values. Abbreviations: T: Type of cigarette (habitual/ultra- ~ light/Next); N: Number of cigarette in a session (1/2); P: Pre- versus post-treatment phase C.it C 1~1 I

I I I I I ~ I I I I I I I =~6 thou~.ands of chemicals. which toeether with nicotine pro~tde critical sensory components important in the maintenance of smoking via a strong conditioned associ- ation between the sensory aspects of smoking (putative conditioned stimulus) and the pharmacological effects of nicotine tputative unconditioned stimulus). This sugges- tion was based on experiments IHerskovic et al. 1986: Rose 1988) which showed that the throat scratching stimulus of smoke may be an essential component for smokine satisfaction. The~importance of taste and smoke in smoking be- havior is also underscored by recent investigations of averaged EEG potentials evoked by discrete nicotine vapor stimuli. Hummel et al. (1992b) observed that nico- tine vapors elicited burning, stinging and odorous sen- sations that differed in time course, threshold dose levels and topographical distribution of the evoked potentials. The same laboratory (Hummel et al. 1992a) further ob- served that hedonic ratings of the S( - ) stereoisomer differed between smokers and nonsmokers, whereas this was not the case for the R( + ) stereoisomer. These studies not only underline the importance of smell and taste for smoking behavior but they also call for increased interest in the multitude of nicotinic as well as non-nicotinic chemosensory signals produced by tobacco smoke. Arguments that nicotine administration per se does not necessarily reduce the craving to smoke have been reported from several studies. Nicotine chewing gum (Hughes et al. 1984), transdermal nicotine administration (Rose et al. 1990), and intra-nasal nicotine administration (Fertig et al. 1986) all failed to abolish totally the craving for cigarettes. On the other hand, Rose and Hickman (1988) reported that after the inhalation of citric acid. which elicited airway sensations resembling those of ciga- rette smoke. the craving for cigarettes was lower than after control inhalations of air. In a further study (Levin et al. 1990), they found that the satisfaction obtained from the citric acid aerosol was highly correlated with the reported intensity of throat sensations. Thus. under the conditions of the present experiment, which allowed good differentiation between the nicotine and the tar yield effects of a cigarette, a reduction in tar yield appeared to be more important for compensatory smoking behavior than a reduction in nicotine yield. Acknowledgements. We thank Dr. C. Feyerabend for analyzing the blood samples. Mr. P. Schmid for technical assistance, Mrs. B. Strehler for editorial assistance and the Swiss Association of Cigarette Manufacturers for financial support. References Armitage AK, Alexander J, Hopkins R. Ward C (1988) Evaluation of a low to middle tarjmedium nicotine cigarette to maintain nicotine delivery to the smoker. Psychopharmacology 96: 447-453 Battte K. Buzzi R. Nil R 119821 Smoke cteid oi ciaarette> tn i nutiin_ behavtor tn men and women, P<ychopharmaculoc% -F !1U 14S Fertue JB. Pomerleau {)F. Sanders B i 19x6i tiicottnz-prouuccd antinociception m-ihtttimally deprtved smokers and t.v-smul.en. Addict Behav 11 : 239-248 Feyerabend C. Russell MAH (1990) A rapid gas-liquid chromato- graphic method for the determination of cotinine and nicotine in biolozical duids. J Pharm Pharmacoi 42:450-353 Hasenfratz 10. Nil R. Battig K 119901 Development of central and peripheral physiological smoking effects over time. Psychophar- macoloev 101:359-365 Herskovic JE. Rose JE. Jarvik ME (1986) Cigarette desirability and nicotine preference in smokers. Pharmacol Biochem Behav 24:171-175 Hofer 1. Nil R. Battig K (1991) Nicotine yield as determinant of smoke exposure indicators and puffing behavior. Pharmacol Biochem Behav 40: 139-149 Hu¢hes IR. Hatsukami DK. Pickens RW. Krahn D. Malin S. Liunic A (1984) Effect of nicotine on the tobacco withdrawal syndrome. Psychopharmacoiogy 89:82-87 Hummel T. Hummei C. Pauli E. Kobal G(1992a) Olfactory discrim- ination of nicotine-enantiomers bv smokers and nonsmokers. Chem Senses 17:13-21 Hummel T. Livermore A. Hummel C. Kobal G(1992b) Chemosen- sory event-related potentials in man: relation to olfactory and painful sensations elicited by nicotine. Electroencephalogr Clin Neurophysiol 84 : 192-195 Levin ED. Behm F. Rose JE (1990) Development of a citric acid aerosol as smoking cessation aid. Drug Alcohol Depend 25:273-279 Nemeth-Coslett R, Henningfield JE. O'Keeffe MK. Griffiths RR (1986) Effects of mecamylamine on human cigarette smoking and subjective ratings. Psychopharmacology 88:420-425 Nil R (1991) A psychopharmacological and psychophysiological evaluation of smoking motives. Rev Environ Health 9 [2]: 85-115 Nil R. Battig K (1989) Smoking behavior. a multivariate process. In: Ney T, Gale A (eds) Smoking and human behavior. Wiley, Chichester. pp 199-221 Robinson JH. Pritchard WS. Davis RA (1992) Psychopharma- cological effects of smoking a cigarette with typical "tar" and carbon monoxide yields but minimal nicotine. Psychopharma- cology 108:466-472 Rose JE (1988) The role of upper airway stimulation in smoking. I.n: Pomerleau OF. Pomerieau CS (eds) Nicotine replacement: a critical evaluation. Liss. New York, pp 95-106 Rose JE. Hickman C (1988) Citric acid aerosol as a potential smoking cessation aid. Chest 92:1005-1008 Rose JE, Levin ED (1991) lnter-relationships between conditioned and primary reinforcement in the maintenance of cigarette smok- ing. Br J Addict 86:605-609 Rose JE. Levin ED, Behm FM. Adivi C. Schur C (1990) Transdermal nicotine facilities smoking cessation. Clin Pharmacol Ther 47:323-330 Rose JE. Sampson A. Levin ED. Henningfield JE (1989) Mecamyla- mine increases nicotine preference and attenuates nicotine dis- crimination. Pharmacol Biochem Behav 32:933-938 Rosenthal R, Rosnow RL (1985) Contrast analysis: focused com- parisons in the analysis of variance. University Press. Cambridge. Stepney R(1981) Would a medium nicotine, low tar cigarette be less hazardous to health? BMJ 283:1292-1296 Sutton SR. Russell MAH, lyer R. Feyerabend C. Saloojee Y (1982) Relationship between cigarette yields, puffing patterns, and smoke intake: evidence for tar compensation? BMJ 285: 600-603