Tobacco Documents Online

I I I I I I I I I I I I I I I I I I rn~6 .osi9sazooro ~~*2 p~ American Society for P~xharmacology~gzpenmenul Tberapeutin 0~~~ . Vol. 246, No. 2 Prvued in U.S.A. ,Cigarette Brand-Switching: Effects on Smoke Exposure and §~oking Behavior ,,"S p. ZACNY and MAXINE L STITZER .,4WVr,&-ti3 of Psychiatry, The Johns Hopkins University School of Medicine, Baltimore, Maryland ,WW {or publication May 13. 1988 ySTRACT ,W study examined the effects of cigarette yield (Federal Trade ,4rrrossion-detennined deliveries of nicotine, tar and CO) on wo biological exposure to smoke constituents and smoking ynevqrs. Smokers (N = 10) of high-yield cigarettes were xrtqW in random order among five different commercially „debie cigarette brands with nicotine yields of 0.1, 0.4, 0.7, t.1(aftered brand) and 1.0 (usual brand) mg and smoked each oqrette type for 5 days while a wide variety of assessments ,.re performed. Steady-state cotinine and CO levels were sub- wioaAy lower after 5 days of smoking ultra-low yield cigarettes poorwae, 152 ng/ml; CO, 25 ppm) than when smoking usual- wrd high-yield cigarettes (cotinine, 252 ng/ml; CO, 38 ppm). bosh CO and nicotine boost (acute exposure) were related to Od. However, relative between-yield differences in all nicotine and CO exposure measures were smaller than predicted from Federal Trade Commission yield ratings. Substantial yield-re4ated alterations were observed in smoking behavior. Subjects smoked more cigarettes and took larger and more closely spaced puffs when smoking low- as compared with high-yield cigarettes. The amount of tobacco burned per day was similar across all yield conditions. However, filter vent-blocking of ultra-low yield ciga- rettes did not appear to occur on a consistent basis. Subjective reports indicated poor acceptability of lower-yield cigarettes. We conclude that switching to lower-yieki cigarettes brings about substantial afterations in smoking behavior which are at least partially responsible for the observed biological compensation associated with these cigarettes. Currently, king-size filter cigarettes which are available in t5s United States have a range of nicotine yields from less than 0.1 to 1.2 mg, a range of tar yields from less than 1 to 17 mg md a range of CO yields from less than 1 to 18 mg (Federal 'Ihde Commission, 1985). This wide range of yields is accom- Kahed primarily by varying the extent to which the cigarette 13lter is ventilated, and to a lesser degree by varying the size Od type of the filter, the amount of tobacco in the rod and the Pcreity of the paper surrounding the tobacco (United States Fn)lic Health Service, 1981). King-size filter cigarettes with a oeatine yield of 0.7 mg or less and a tar yield of 10 mg or less +N generally labeled as "light", which conveys the marketing h4lication that they are safer than higher-yield cigarettes. Ro*ever, there is a considerable amount of evidence which ftests that lower-yield cigarettes are not safer than high- 7k1d cigarettes. One line of evidence comes from studies in which biological t4*SUre to smoke constituents are measured in samples of ftokers consuming usual-brand cigarettes with a range of ,&ftir'id for publication November 30, 1987. T~i+ work was supported by Research Grant CA 37786 from the Nttmnal ~= Inatitute and Institutional Tnin~ Grant T 32 DA 07209 from the ~+0A+1 Institute on Drug Abu.e. different yield characteristics. Numerous usual-brand smoking studies of this type have shown either no or weak relationships between cigarette yields and smoke exposure levels, as mea- sured by smoke constituent markers such as nicotine, cotinine, CO or thiocyanate (e.g., Benowitz et a1.,1983; Ebert et a1.,1983; Hill et aL, 1983; Maron and Fortmann, 1987). A second line of evidence comes from brand-switching stud- ies, in which smokers of high-yield cigarettes are switched to lower-yield cigarettes while their smoke exposure levels are measured. Upon switching to lower-yield cigarettes, smoke exposure levels have remained unchanged (Benowitz and Jacob, 1984; Kanzler et al, 1983; Ossip-Klein et al, 1983; Robinson et aL, 1983), or reduction in biological exposure to one or more smoke constituents has been substantially less than predicted from the cigarette package yields (e.g., Ashton et a, 1979; Benowitz et al., 1986; Turner et aL, 1974; West et aL, 1984). One explanation often put forth regarding the poor relation- ship between cigarette yields and observed smoke exposure levels is that smokers may compensate for low-yield character- istics by smoking more low-yield cigarettes. Indeed, in two recent usual-brand smoking studies, smokers of lower-yield cigarettes consumed significantly more cigarettes than did smokers of high-yield cigarettes (Benowitz et aL, 1986; Maron ~EV1ATi0NS: CO, cait~on morwxide; RTD, resistance to draw; FTC, Federal Trade Cortunission; ANOVA, ar>afysis of varianoe; A.B., aftered ~d U.B. usua! brand. 619

I I I I I I I I I I I I I I I I I I 620 Zacny and Stitzer and Fortmann, 1987). Also, several brand-switching studies have shown that cigarette consumption increases when smokers are switched to lower-yield cigarettes (Ashton et aL, 1979; Benowitz et aL, 1986; Robinson et aL, 1983; Turner et aL, 1974). However, other usual-brand smoking studies (e.g., Ebert et aL, 1983; Folsom et aL, 1984; Gori and Lynch, 1985; Russell et aL, 1980) and brand-switching studies (e.g., Kanzler et aL, 1983; Russell et aL, 1982; West et al., 1984; Woodman et aL, 1987) have not found increased cigarette consumption with lower- yield cigarettes. Another explanation often put forth regarding the poor re- lationship between cigarette package yields and observed smoke exposure is that smokers compensate for low-yield character- istics by altering their smoking topography (cf. Gritz, 1980; Henningfield,1984; McMorrow and Fozz,1983; Moss and Prue, 1982; United States Public Health Service, 1984). In a series of recent studies done in our laboratory, we identified three pa- rameters of smoking topography that influence the uptake of nicotine and CO from cigarettes: 1) number of puffs (Chait et aL, 1985); 2) volume of puffs (Zacny et.al., 1987); and 3) filter vent-blocking (Zacny et aL, 1986). Relatively few studies have actually assessed these detailed smoking topography parame- ters across cigarettes with a full range of yield characteristics currently available to smokers. Available evidence does suggest that smokers draw more and/or larger puffs from lower-yield than from high-yield cigarettes (e.g., Epstein et aL, 1981; Ossip- Klein et aL, 1983; Tobin and Sackner, 1982; Woodman et aL, 1987). Filter vent-blocking is the act of obstructing filter vents of lower-yield cigarettes with one's fingers or lips. Although filter vent-blocking can markedly increase smoke constituent uptake, the incidence of this behavior among low-yield smokers is unknown. In the only known studies to assess incidence of filter vent-blocking, 41% of a small sample of smokers (N = 39) whose usual-brand cigarettes were ventilated, and 25% of a small sample of smokers (N = 16) who were switched to ventilated cigarettes, were judged to block at least some of the filter vents (Kozlowski et aL, 1982; Robinson et aL, 1983). Inhalation parameters (i.e., inhalation volume and lung expo- sure duration) have not been sensitive to cigarette yield alter- ations (Tobin and Sackner, 1982). No study to date has measured simultaneously the full spec- trum of smoking behaviors (i.e., cigarette consumption, puffing and inhalation topography, filter vent-blocking) and smoke exposure levels across commercial cigarettes with a full range of yields including ultra-low yields. Such a study would 1) provide descriptive data concerning acute and chronic smoke exposure levels associated with different-yield cigarettes and 2) allow a direct examination of behavioral mechanisms which might be at least partly responsible for any observed discrep- ancies between package yield predictions of exposure and mea- sured uptake of tobacco smoke constituents. Accordingly, we switched habitual smokers of high-yield (nicotine, 1.0 mg) cigarettes, in random order, to five different cigarette brands, including an ultra-low yield (nicotine, 0.1 mg), a low-yield (nicotine, 0.4 mg), a medium-yield (nicotine, 0.7 mg) and two high-yield (nicotine, 1.0-1.1 mg) brands. Smokers were given 5 days of exposure to each cigarette type, during which time their daily cigarette consumption and chronic biological exposure levels were monitored. Two of these days, subjects attended laboratory smoking sessions so that their smoking topography and acute smoke exposure levels could be assessed. „a3j_ Methods Subjects. Subjects were 10 cigarette smokers (five males and 1$i females): mean age, 36.1 years (range 19-50); mean years smoking, 1U (range 6-38); mean number of cigarettes smoked per day, 30.5 (ream 20-50). Their usual-brand cigarettes were filtered and nonmenthole~ with an average nicotine yield of 1.0 mg. Subjects, recruited t~ newspaper advertisements, were in good health, had a negative pr,* nancy test and reported no medication usage or drug/alcohol abt., Subjects were informed before inclusion into the study that 1) *0 study involved smoking different brands of nonmentholated, filt., cigarettes and 2) the brands might be higher or lower in conatitu'. yields than their customary brand. Cigarettes. Subjects were exposed to five brands of king-size (14 mm) cigarettes during the study. The brand names and specific aW rette characteristics are listed in table 1. One of the brands w.e the subject's usual brand. The other four brands, Now. Vantage G'ka Lights, Marlboro Lights and Camel, represent a range of yield cbareo- teristics currently available to smokers of filter cigarettes in the UyAy States. These particular brands were selected following a preiiminary survey of available brands because they had similar RTD characterw tics. Cigarette RTD is a variable that may affect smoking topogra* (Dunn, 1978; Zacny et aL, 1986). Cigarette RTD of 200 cigarettes froa each brand was measured at the Tobacco and Health Research Instiuae (Lexington, KY) by use of a Filtrona Pressure Drop Tester (Ameriras Filtrona Company; Richmond, VA). For each of the brands, ciganft weight, filter weight and tobacco weight were determined by taking 5.w cigarettes from each of four packs and weighing them with and withoa filters. Twenty cigarettes from each brand were smoked on a PhipQr and Bird analytical smoking machine (Phipps and Bird, Incorporated: Richmond, VA) at the Tobacco and Health Research Institute accroed. ing to a standardized protocol; i.e., 35 cc puffs, 2 sec in duration, everf 58 sec until 3 mm in front of filter overwrap. This analysis was doee to determine the average number and total volume of puffs that an taken from each of the cigarette brands during FTC analytical smoking machine procedures. General Procedures Study design and schedule. During the 5-week study, subjeGa smoked four experimental cigarette brands and their usual brand for 3 consecutive days each, with order of conditions determined by a Lada square design. Subjects reported to the laboratory on a Thursday ac Friday morning to begin each smoking condition. They reported aa Monday or Tuesday afternoon of the following week for data collectiat and again on Tuesday or Wednesday morning of that week to end t6e condition. On the 2 days separating study conditions, subjects smoW their own cigarettes. On the first day of each condition, subjects wae given a supply of their assigned cigarette brand sufficient to lait 1 week. Subjects were told 1) they could smoke as many or as few d these cigarettes as they pleased throughout the study week. 2) thwf were to return all unsmoked cigarettes to the experimenter at the ead of each study week; and 3) they were to smoke only the cigarettes thd were supplied to them by the experimenter. Two subjects (J. W. Aod S. B. U.) repeated the ultra low-yield (nicotine, 0.1 mg) condition aMt admitting that they had smoked their usual-brand cigarettes occamo, ally during the first exposure. Data from their second ezposure 90 included in analysis. Laboratory smoking aseessment. On the first day (Day 1) +ld final (Day 6) day of exposure to each condition, subjects attended a 100-min laboratory smoking-asaessment session, during which tlsf" smoked two cigarettes of their assigned brand while behavioral and biological exposure measures were collected. Thus, laboratory smoking data were collected for the first exposure to each experimental type and after chronic exposure to each assigned brand of ezperim cigarettes. To standardize presession smoking deprivation, subj80 were instructed to abstain from smoking overnight, before the moraiat session. During the sessions, subjects were seated,in a room b0us"ng the smoking measurement equipment, with an experimenter prexnt 1° I

I I I I I I I I I I I I I I I I I I I al piocedtues and monitor data colleetion imenl One f Zper ~ e . o was smoked using a plastic holder that allowed measure- ~~~rai puff parameters, including puff volume. The other ~,,, Was smoked without a holder. The two laboratory smoking ~~,,,,"y were separated by 60 min. Subjects took, on average, one ~~'"~' when smoking from the holder as compared to smoking ~~ the holder. This holder effect was similar across cigarette yield ~. Cs no Holder x Yield interactions were noted). Therefore, g60 oZiiy the magnitude of effects and not functional relationships 00 yield conditions that were influenced by the holder. Additional AW of the holder will be reported elsewhere. ...&g Topography Measurement psoIIg. Three puffing parameters were measured during smoking- ,awgment sessions conducted with the holder 1) number of puffs; 2) aw-puff interval (temporal period from the offset of one puff to the MW of a subsequent puff); and 3) puff volume (the amount of smoky jdrawn from the cigarette rod into the mouth during a puff). The ,,,kwlogy used in measuring these puffing topography variables has ,w described in detail elsewhere (Zacny et aL, 1987). ltespiration. Three postpuff respiration variables were measured.: ;l uthalation volume (amount of smoky air inhaled after a puff); 2) mialation volume (amount of air exhaled after the postpuff inhala- Usr and 3) lung exposure duration (temporal period between inhala- Nq onset and end of exhalation). The respiration parameters were sasared with a respiratory inductive plethyamograph (Respitrace; ,.oinvasive Monitoring Systems, Inc.; Ardsley, NY). The technology sod in measuring these respiration variables has been described in buil elsewhere (Zacny et aL, 1987). Oaily Cigarette Consumption Daily cigarette consumption was monitored by having subjects 1) noocd the time of day that each cigarette was smoked; and 2) save iras from all cigarettes smoked. For each study day, the number of cp:ettes smoked as recorded by the subject was compared to the snmber of butts she/he collected that day. Typically the self-monitored .roking count deviated from the butt count by no more than one ap:ette. If there was a deviation between the two counts, the larger et the two numbers was considered the cigarette-per-day measure for tba day. , Cgarette Butt Measures Weight of tobacco burned. The weight of each spent cigarette 1*et returned from natural environment smoking was subtracted from 16 weight of an unsmoked whole cigarette of that brand (see table 1). Dod+ weight of tobacco burned per cigarette and total amount of tobacco bmed per day were determined. Proportion of the tobacco rod burned. This measure was calcu- Yhd by dividing the weight of tobacco burned per cigarette by the total wKht of tobacco in an unsmoked cigarette (see table 1), and multiply- 1K the obtained number by 100. This provided a measure of smoking i.uasity independent of the amount of tobacco present in the rods of ftrent brands. Veat-blocking analysis. When smokers block filter vents of ultra- 60 Yield cigarettes with their fingers or lips, characteristic stain laterns can be detected on the spent filters which are distinctly ilterent from stain patterns seen when vents are not blocked during (Rorlowald et oL, 1980). We analyzed spent filters from the ~trt low-yield cigarettes smoked outside of the laboratory to aasesa ~6ether or not filter vent-blocking had occurred. The cigarettes wen saiiPed by the experimenter (J. P. Z.) to one of four categories: 1) "I* unblocked (filter periphery completely white); 2) vents blocked (dher completely brown); 3) vents partially blocked (filter periphery light brown with dark brown spot in the middle); 4) vents questionable (Iome, but not all of the filter periphery stained light brown). For each 14bct, eight spent filters were randomly selected from both the 2nd md the 5th day's collection of cigarette butts for analysis by a second 4ter who was blind to the initial ratings of the experimenter. Filters Cigarette Brand•switching 621 from the 0.4 and 0.7 mg nicotine-yield cigarettes could not be analyzed using the four stain pattern categories because it was impossible to distinguish between unblocked and partially-blocked vents. Biochemical Exposure Measurement Procedures Plasma nicotine and cotinine. Seven milliliters of blood was drawn immediately before and 1 min after the last puff from each of the two cigarettes smoked in laboratory sessions. Nicotine and cotinine levels in the plasma were determined by gas chromatography (Jacob et aL, 1981). Nicotine exposure from the smoking of a single cigarette, i.e., nicotine boost, was measured by subtracting the precigarette nic- otine level from the postcigarette nicotine level. The mean of the four cotinine samples obtained during the Day 6 (morning) smoking-assess- ment session of each study week constituted the cotinine measure for that study week, and was representative of chronic nicotine exposure from a given brand of cigarettes. CO. During collection of lung air samples for CO analysis, subjects exhaled residual air from their lungs, took a deep breath, held the breath for 20 sec then exhaled successively into two 1-liter polyvinyl bags. The second bag, containing alveolar air, was analyzed for CO content, using an Ecolyzer 2000 (Energetics Science, Elmsford, NY). Expired-air CO levels obtained via the above procedure are highly correlated with carboxyhemoglobin levels (e.g., Jarvis et aL, 1980). In the smoking assessment sessions, expired-air samples were obtained immediately before and 2 min after the last puff from each of the two cigarettes smoked. CO exposure from the smoking of a single cigarette, i.e., CO boost, was measured by subtracting the precigarette CO level from the poatcigarette CO level. An expired-air sample was also ob- tained on the 5th afternoon of each study week, after subjects had smoked an assigned brand for 4.5 days. Subjects were asked to refrain from smoking for 20 mi.n before this session so that CO measurements would not be influenced by recent smoking. The afternoon CO level was representative of daily CO exposure from a given brand of ciga- rettes. Subjective Report Measures Cigarette characteristics rating scale. After each cigarette that was smoked in the laboratory-assessment sessions, subjects were asked to rate nine characteristics of the cigarette. Subjects made their subjec- tive estimations for each of the nine measures by placing a vertical. hatch mark somewhere along a 100-mm bipolar scale. Subjects rated the cigarettes on strength (very weak/very strong), heat (no heat/very hot), harahness (very mild/very harsh), draw (easy/hard), degree of taste (no taste/a lot of taste), type of taste (very bad/very good), satisfaction derived from smoking the cigarette (very unsatisfying/very satisfying), amount of tobacco smoke obtained per puff (mostly air/ mostly smoke) and likelihood of brand purchase (not at all likely/very likely). Smoking withdrawal scale. Subjects filled out a Smoking With- drawal Scale (Hughes et aL, 1984) on the fifth afternoon of each study week, at least 20 min after their last cigarette. At this point in time, they had had 4.5 days of exposure to a given brand of cigarettes. The acale was designed to measure the number and degree of withdrawal symptoms exhibited during smoking abstinence. Subjects were asked to rate the degree to which they experienced each of 16 symptoms (e.g., cigarette craving, irritability, anxiety, difficulty in concentrating and increased eating) during the previous 24-hr period by circling a number from 0 to 3(0, not present; 1, mild; 2, moderate; and 3, severe) for each of the symptoms. The circled numbers corresponding to each of the symptoms were summed to yield a composite withdrawal symptoms- tology score. Data analysis. Six cigarette characteristics were compared using one-way ANOVA: 1) total weight; 2) tobacco weight; 3) filter weight; 4) RTD; 5) number of puffs drawn during the FTC smoking machine procedure; and 6) total smoke volume drawn during the PTC smoking machine procedure. Measures of smoking topography, daily cigarette consumption, cigarette butta, biological exposure and subjective reports were analyzed by one- or two-way ANOVA. Depending on the particular I

I I I I I I I I I I I I I I I I I 622 Zacny and Stitzer variable being analyzed, factors were one or more of the following: cigarette yield, successive ezposure days (days 2-5) and asseayment time (day 1 us. day 6). Day 1 data were excluded from the successive days factor because the smoking-assessment session and the no smok- ing stipulation before the session reduced cigarette consumption for that day. Tukey post-hoc comparison tests between conditions were done when a main effect of yield was obtained. From the vent-blocking analysis, the percentage of spent filters assigned to each of the four stain-pattern categories was calculated for descriptive purposes. Results Cigarette characteristics. Table 1 shows constituent yield and physical characteristics of study cigarettes. In addition to delivering less nicotine, tar and C0, low-yield cigarettes weighed less than high-yield cigarettes [F(4,95) = 305.7, P < .001] and contained significantly less tobacco in their rods [F(4,95) = 779.4, P < .001] but had somewhat heavier filters [F(4,95) = 980.1, P < .001]. Study cigarettes were selected for similar RTD characteristics. Table 1 shows that although these cigarettes did differ significantly on the RTD measure (F(4,950) = 42.3, P<.001], the largest variation between conditions was less than 1 cm of water. The study cigarettes differed signifi- cantly on both number of puffs [F(4,95) = 149.4, P<.001] and total smoke volume [F(4,95) = 146.3, P < .001] obtained via the smoking machine analysis. Particularly notable is the fact that fewer puffs were drawn from the Now and Vantage low- yield cigarettes than from the higher-yield cigarettes in the analysis. Chronic Smoking: Biological Exposure Measures Cotinine. Figure 1(top frame) shows that cotinine levels after 5 days of smoking under each condition were 151.7, 188.4, 220.8, 259.2 and 252.2 ng/ml when subjects smoked cigarettes with nicotine yields of 0.1, 0.4, 0.7, 1.1 (A. B.) and 1.0 (U. B.) mg, respectively [F(4,36) = 8.6, P < .001]. Post-hoc tests revealed that cotinine level in the 0.1-mg condition was signif- icantly lower than levels in the 0.7-, 1.1-(A. B.) and 1.0 (U. B.)- mg conditions and that cotinine levels in the 0.4-mg condition were significantly lower than those in the 1.1 (A. B.)-mg and 1.0 (U. B.)-mg conditions. CO. Figure 1 (bottom frame) shows that afternoon CO levels after 4.5 days of smoking were 24.5, 37.5, 36.6, 35.5 and 37.5 TABLE 1 Ci9arette characteristics yaft ppm when subjects smoked cigarettes with nicotine yi" - 0.1, 0.4, 0.7, 1.1 (A. B.) and 1.0 (U. B.) mg, respectively (pt4., = 8.8, P < .001]. Corresponding CO yields were 2, 5, 11 14 (k B.) and 15 (U. B.) mg, respectively. Mean CO level in th. mg condition was significantly lower than CO levels ip ~ other four conditions. Laboratory Smoking: Biological Exposure Measures Nicotine boost. As shown in figure 2 (top frame), aicauft boosts (pre-minus postsmoking plasma levels) were 5.8, IA.:13.0, 18.8 and 17.3 ng/ml in the 0.1-, 0.4-, 0.7-, 1.1-(A. B,i r 1.0 (U. B.)-mg nicotine-yield conditions, respectively (py4.% = 22.6, P<.001]. Nicotine boost in the 0.1-mg condition ,, significantly lower than nicotine boost from all other conditioa whereas nicotine boost in both the 0.4- and 0.7-mg conditiod were lower than those in the 1.1 (A. B.)- and 1.0 (U. B.).Nt conditions. Nicotine boost did not differ as a function of ~.. sessment day (1 vs. 6). CO boost. As shown in figure 2 (bottom frame), CO boomti (post-minus presmoking CO level) were 4.1, 7.2, 9.4, 8.0 4W 9.2 ppm in the 0.1-, 0.4-, 0.7-, 1.1-(A. B.) and 1.0 (U. B.).al nicotine-yield conditions [F(4,36) = 14.8, P < .001]. Post.ia tests revealed that CO boost from 0.1-mg cigarettes was sipd. icantly lower than boosts from all other conditions and ths boost from the 0.4-mg condition was significantly lower tbn boosts from the 0.7- and 1.0 (U. B.)-mg conditions. CO booa did not differ as a function of assessment day (1 vs. 6). Chronic Smoking: Behavioral Measures Cigarettes per day. Figure 3 shows that subjects smokd an average of 34.3, 31.8, 28.4, 27.1 and 28.6 cigarettes per dty in the 0.1-, 0.4-, 0.7-, 1.1-(A. B.) and 1.0 (U. B.)-mg nicotint yield conditions, respectively [F(4,36) = 6.2, P<.001). Poa hoc tests revealed that significantly more 0.1-mg cigaretras were smoked than 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)-mg cigaretto and significantly more 0.4-mg cigarettes were smoked than 1.1 (A. B.)-mg cigarettes. Cigarette consumption did not vm across successive study days within conditions. Tobacco smoked per day. As shown in table 2, the amouat of tobacco burned per cigarette differed significantly acrar cigarette yield conditions [F(4,36) = 256.0, P<.001]. This is not surprising, however, as these study cigarettes start out with Now y~~ Muboro t.ight Camd ~, Nicotine yieid (mgf 0.1 0.4 0.7 1.1 1.0 Tar yield (mg)Z 1 5 10 16 16 CO yield (mg)s 2 5 11 14 15 Vert4ated fgters Yes Yes Yes No No Cigarette wt. (g) 0.791•.'.`.a 0.908°•`•e 0.946° a 0.979° 0.955 TobaCCo wt. (g) 0.544•.s.°.e 0.654s.`.a 0.731 "•r 0.815° 0.788 FIDer wt. (g) 0247•'"•" 0.254'.0'° 0.215`•a 0.164° 0.167 RTD (cm H2O) 112••0'° e 11.7° 12.1 ° 11.8 12.0 Number of p" 6.78••e.°'s 6 58ae.a 8.02°•a 8.31 8.36 Total smoke voiume (mq' 238•.e.o,a 231 aae 280" 291 291 _ ~ ' nt subjscts smolced Maiib)ro king size t~ter aigarettes: one subject smdced Wruton and one smoked Ralelgh king size agarettes. Values stw~nm are w"'~ Eig ntaverages (Marlboro x 8: Wirwton x 1, Raieiqh x 1). = Fnom Jaruery. 1985 FTC report ' aRlokln9 fnact*le p(GOSdIIe. • SiQrr}fcarttfy d(ferent from 0.4-mg oond'itlon. • S4niAcantfy diHerent from 0.7-mq condition. • SiqnilfCarttly dflxent from 1.1 (A. B.)rtg candition. • Sigrtiffeantly Bfferont firom 1.0 (U. B.}mg cotdition. I

I I I I I I I I I I I I I I I I I 320 1 240 '1 80 Cotinine N bcd ~ ~ od 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) 48 -1 361 0. 24 " a Carbon Monoxide r_'~ r=-,.Z r=, 121 0 11aTdl I ~ 1 I T I I T I t I 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) NICOTINE YIELD (mg) Rp 1. Mean Day-6 cotinine levels (nanograms per rtWiter) and aftemoon *-5 expired-sir CO levels (parts per million) from 10 subjects are sthown as a fiuiction of cigarette nicotine yield (n>Agrams). Brackets ndcate S.E.M. [a sovftantly ddferent from 0.4-mg r,ordtion; b, signif- eantly different from 0.7-mg condition; c, significantiy cifferertt from 1.1 (A. B.)-mg conditiorr d, siglificantty diiferent from 1.0 (U. B.)-rng condi- 1onl. .ery different amounts of tobacco present in the rod (see table 1). The weight of tobacco burned per day, a measure that takes into account the number of cigarettes smoked, showed no 94aificant effect of constituent yield [F(4,36) = 2.3, P<.08], ilthough the total amount of tobacco burned per day was somewhat less for the 0.1-mg cigarettes than for any of the ather constituent yield types. The percentage of available t.o- bacco burned per cigarette differed as a function of constituent l'kld [F(4,36) - 18.9, P<.001] and was greater for 0.1-, 0.4- imd 0.7-mg cigarettes, where about 80% of the tobacco was burned, than for 1.1 (A. B.)- or 1.0 (U. B.)-mg cigarettes, where about 75% of the tobacco was burned. Vent-blocking. In this study, 1631 spent filters from the 0.1-mg condition were rated for hole blocking. Only 0.1% were identified as completely vent-blocked, whereas 72% were rated as unblocked. The remaining 28% were rated as partially blocked (6%) or placed in a questionable category in which the 4eCrimination between unblocked and partially blocked was difficult (22%). Five individual subjects were judged to be c6arly not vent-blockers, as at least 85% of their butts were Wed unblocked. For four other subjects (T. B. L, P. L U., D. 8• E. and J. W. A.), 25% or more of their submitted butts were rated in the questionable category (mean, 47.5%; range, 26- 68%), suggesting that they may have vent-blocked on an incon- k6teat basis. For one subject (A. S. M.) almost all butts were rated as either partially blocked (30%) or questionable (67%), +uggeating more consistent vent-blocking behavior. Ten per- Cigarette Brand-Swltching 623 24 -t 18 12 n Nicotine Boost 12 abcd cd cdT I I . J , 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) Carbon Monoxide Boost 0 11ab c d I bd I I ~ i I 0.1 0.4 0.7 1.1 (AS) 1.0 (UB) NICOTINE YIELD (mg) Fig. 2. Mean nicotine boost (nanograms per milliliter) and CO boost (nanograrns per milliliter) from 10 subjects are shown as a functian of cigarette nicotine yield (milligrams). The boosts were measured in the laboratory smoking sessions. Boost refers to post-minus presmoking nicotine and CO leveis in the piasma and expired eir, respectively. Brackets indicate S.E.M. [a. significantty dffferent from 0.4-mg condition; b, significantiy different from 0.7-mg condition; c, significantly different from 1.1 (A. B.)-rtg condition; d, sk,ytifiprttl)r cSfferent from 1.0 (U. B.)- mg oonditionl• cent (N - 160) of the smoked butts were rated by a second person as a reliability check. The reliability sample was repre- sentative of the larger sample in distribution of original rating categories. The inter-rater reliability ratio (agreements divided by total number rated) was 0.80 with the majority of disagree- ments occurring between the unblocked and questionable rat- ings. Laboratory Smoking: Behavioral Measures Puff number. Figure 4 (upper left frame) shows that average puffs per cigarette were 11.3, 9.9, 11.2, 10.6 and 10.4 in the 0.1-, 0.4-, 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)-mg conditions, respectively [F(4,36) = 1.9, P<.13]. Inasmuch as the number of puffs used to determine smoking machine yields differs across the brands used (see table 1), we also calculated an excess puff measure by subtracting the average number of puffs used in smoking machine tests from the number observed during ad lib smoking in the laboratory. Figure 4 (upper right frame) shows that excess puffs per cigarette were 4.5, 3.2, 3.1, 2.0 and 1.9 in the 0.1-, 0.4-, 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)- mg conditions, respectively [F(4,36) a 5.34, P < .002]. Post- hoc tests showed that the number of excess puffs under the 0.1- mg condition was significantly greater than the number under the 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)-mg conditions. Neither the I

' 624 Zacny and Stnzer , Ctgarettes per Day I I I I I I I I I w m ~ ~ z 40 1 30 20 bcd o 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) NICOTINE YIELD (mg) Fg. 3. Mean number of cigarettes smoker per day from 10 subjects are shown as a function of cigarette nicotine yieid (milligrams). Data from the first day of each study week are excluded. Brackets indicate S.E.M. [b, significantly different from 0.7-mg condition: c, significantly different from 1.1 (A. B.)-mg condition: d, significantly different from 1.0 (U. B.)-4rg condition]. puff number variables nor the following smoking topography variables differed as a function of assessment day (1 us. 6). Puff spacing. Figure 4 (bottom left frame) shows that interpuff intervals were 22.3, 24.7, 26.4, 30.3 and 29.7 sec under the 0.1-, 0.4-, 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)-mg conditions, respectively [F(4,36) = 7.92, P<.001J. Post-hoc tests revealed that 1) puffs were spaced at shorter intervals under the 0.1-mg yield condition than under the 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)- mg conditions; 2) puff spacing under the 0.4-mg condition was shorter than under the 1.1 (A. B.)- and 1.0 (U. B.)-mg condi- tions; and 3) puff spacing under the 0.7-mg condition was shorter than under the 1.1 (A. B.)-mg condition. Puff volume. Figure 4 (bottom right frame) shows that average puff volumes were 64.7, 58.0, 61.1, 52.4 and 53.4 ml in the 0.1-, 0.4-, 0.7-, 1.1 (A. B.)- and 1.0 (U. B.)-mg conditions, ~ respectively [F(4,36) = 15.0, P < .001). Post-hoc tests revealed that puff volume was greater for the 0.1-mg cigarette than for the 0.4-, 1.1 (A. B.)- and 1.0 (U. B.)-mg cigarettes, whereas puff volumes under both the 0.4- and 0.7-mg conditions were greater Ithan under the 1.1 (A. B.)- and 1.0 (U. B.)-mg conditions. Cumulative puff volume. This measure, which represents the total smoke dose pulled from the cigarette, was derived by multiplying number of puffs by average puff volume for each laboratory cigarette smoked. Cumulative puff volumes were 703.6, 544.8, 673.3, 529.6 and 529.9 ml for the 0.1-, 0.4-, 0.7-, i.l (A B.)- and 1.0 (U. B.)-mg conditions, respectively [F(4,36) 7.45, P < .001J. Post-hoc tests revealed that cumulative puff volume in both the 0.1- and 0.7-mg conditions was greater than for the 0.4-, 1.1 (A. B.)- and 1.0 (U. B.)-mg conditions. Respiration parameters. Average inhalation volumes ~ TABLE 2 Vq, %measured by the Respitrace ranged from 670 to 774 ml ac% yield conditions, whereas average exhalation volumes raw from 759 to 844 ml. Average lung exposure time ranged 5.0 to 5.6 sec. None of the respiration measures differed s~ icantly across cigarette-yield conditions. Subjective Reports Withdrawal symptoms. As shown in figure 5, avers, Smoking Withdrawal Scale scores were 9.4, 11.6, 6.4, 3.5 am 2.4 when subjects smoked cigarettes with nicotine yielch q 0.1, 0.4, 0.7,1.1 (A. B.) and 1.0 (U. B.) mg, respectively [F(4,x = 2.8, P<.04). In post-hoc testing, only scores in the 0..t.tit condition were significantly different from scores in the 1.1 t,k B.)- and 1.0 (U. B.)-mg conditions. Cigarette acceptability. Table 3 presents data for nia, rating measures used to assess the subjective reaction to stu* cigarettes. Each of these measures showed significant mm effects of constituent yield (P < .01). In general, lower-yiQ cigarettes were rated as less strong, less hot, less harsh, easw to draw, having less taste, having poorer taste, delivering rel& tively more air than smoke and providing less satisfaction th" the high-yield cigarettes. Subjects indicated that they would he less likely to purchase any of the altered-brand cigarettes a compared to their usual-brand cigarettes. Discussion This study has shown that cigarette yield characteriatio influence biological exposure to nicotine, cotinine and CO wMa smokers switch among cigarette brands with a wide range d yield characteristics. Lower-yield cigarettes, especially the 0.1• and 0.4-mg nicotine-yield cigarettes, tended to be associaad with lower biological exposure levels. However, as will be dir cussed below, the actual reductions in nicotine, cotinine .ad CO exposure were less tharn what would be predicted from FTC yield information. Smoking behaviors were assessed simulta- neously along with biological exposure measures in order 4 determine if some or all of the measured smoking behavim played a role in biological compensation. Indeed, several smoi- ing behaviors were altered in a manner consistent with tbt observed biological compensation. Although FTC yields are not expected to predict abeoW dose exposures to cigarette smoke constituents due to marlw differences in smoking machine methods us. human smokiN behaviors (e.g., smokers exhale part of the drawn puff, whet+O machines capture all the smoke drawn in a puff), cigazVO yields may be expected to predict relative exposure differeaor across cigarette brands if the volume of smoke inhaled remaiee constant (Rickert et al., 1986). Thus, compared to the ugod' brand, high-yield (nicotine, 1.0 mg) cigarettes, nicotine e31W 00 Weight and proportion of tobacco burned dtxing ad Rb cluonic artrokking Ggarette NMr~e YiMd ~ 0.1 rtg 0.4 mg 0.7 mg 1.1(A H.) eg 1.0 (U . 8.) m9 Weight of tobacco burned per oigarette (g) 0.43~'.aa 0.53a.°.e 0.58aa 0.61 ° Q 0.59 Weight of tobacco burned per day (g)' 14.9 16.8 16.4 16.7 16.8 Proporoon of tobacco rod burned 80.4°•° 80.1c•e 78.7 75.6 ~ 74.8 I _ ~ ' Mafn effect af yield canditlon P < 08 ~ ' Sigrrificantty dttferent hnm 0.4-mg cotsdition. • Sigttificantfy difbment from 0.7-mg conditjon. 0 ° siyvficarmy 6fferont from 1.1 (A. B.}rtg oorn7kfon. ~ Signrf4antfy dHferent from 1.0 (U. B.}rtg ooncf'rti0rt. I I UO

1 ~ ~ ~ 1 ~ ~ ~ ~ ~ ~ i ~ ~ ~ ~ ~ ~ ~ Clgarette Brand-Switching 625 Number of Puffs Excess Puffs 14 , 6.0, 12 '1 4 5 ~ . , 2 0: W LU m ~ 10-! ~ co 3.0 ~ ~ z z 0 0 bcd - - ' ~- T . 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) 40 , Inter-puff Interval 80 , Puff Volume 30. N ~ ~ ~. 4: z LU 20-, J 60 U W J J W ~ 10. . 0 bcd Cd C acd T od d 0.1 0.4 0.7 1.1 (AB) 1.0 (UB) NICOTINE YIELD (mg) 0.1 0.4 0.7 1.1 (AS) 1.0 (UB) NICOTINE YIELD (mg) Rq. 4. Mean number of puffs (top left frame), excess puffs (top right frame), interpuff irrterval (bottom left frame) and puff votume (bottom right trme) from 10 subjects are shown as a function of cigarette nicotine yieid (nrMigrarns). Smoking topography data were cflHe~,-ted from the laboratory xnoking sessions. Excess puffs refers to the difference from actual number of puffs taken by the subjects and the number of puffs used in the snolcing machine test. Brackets indicate S.E.M. [a, signifioanUy different from 0.4-rng condition; b, sx~ty different from 0.7-mg condition; c, kgniflcantty different from 1.1 (A. B.)-mg condition; d, significantly different from 1.0 (U. B.)-mg conditionJ. 20 , 15 ~ 5 T T od 0 0.1 0.4 Withdrawal 0.7 1.1 (A8) 1.0 (U8) NICOTINE YIELD (mg) R9. 5. Mon withdrawal symptomology aoores from 10 subjecis are Mbwn as afunction of cigarette nicotine yield (rtillgrarns)• 3ubjects fftd ~ late ~ ~yScaje,each study Brackets ir>dilcate S.E.M. [c, signiflCantly different from 1.1 (A. B.)-rtg condtion; d, signi}E- carttly different from 1.0 (U. B.)-mg condition]. " reductions of 30, 60 and 90% would be predicted when 'mokers switch to brands delivering 0.7, 0.4 and 0.1 mg of nicotine. In the present study, average chronic exposure to nicotine, as reflected in plasma cotinine levels, was reduced by 12,25 and 40% when subjects switched to the lower-yield study agarettes. It is clear that cotinine exposure reduction was subetantially less than predicted by F'I'C yields. Acute nicotine "posure, as reflected in plasma nicotine boost, was reduced by TABLE 3 Cig.roft characteristic ratinqs GSgrans KoOtne YWd 0.1 eg 0.4 mg 0.7 rtg 1.1(A. 8.) rtg 1.0 (U. 8.) eg StrengHl 18.7''o•" 47.6`'° 50.9`'4 73.6 66.4 Heat 15.5''aae 29.5° 33.1 43.4 46.0 Herstx>ess 28.3" 38.4` 35.5` 61.2 48.4 Draw 63.6e.` ° 46.5 30.1 29.7 25.8 Taste arnount 19.0''a°'a 42.0°•" 52.8°•a 71.8 81.1 Taste quality 16.9''e'°'a 47.5° 56.7° 59.6 81.9 Satisfactfon 14.4"'.° s 48.2° 582° 59.3° 87.5 Smoke vs. air 21.1''''GS 53.1°.s 63.8° 78.7 84.1 Liceiy to purchase 7.5''a°.e 39.6° 55.2° 55.4° 85.3 ' SigrWbantly difteront from 0.4-mg condifion. ' SiprdicarNly dffetent from 0.7-nq oonditbn. ~ Sigr~iffCantfy dMfererrt from 1.1 (A. 8.}mg co ~tion. Siqndfpntly dMferent from 1.0 (U. 8.)-rtg conditlon. 25, 38 and 67%. Although these reductions are not in proportion to the yield reductions, they are certainly closer to them than were the reductions in cotinine levels. One possible reason as to why cotinine exposure levels were more poorly related to yield is that a greater number of variables (metabolism and clearance rates and daily cigarette consumption) affect chronic exposure levels. Another possible reason accounting for the poor relation between cotinine levels and yield is that our subjects, outside of the laboratory sessions, may have been smoking high-yield cigarettes in the low-yield cigarette conditions. Indeed, two subjects had to repeat the 0.1-mg condition because they ad-

I I I I I I I I I I I I I I I I I I 626 zacny and stitzer mitted smoking some high-yield cigarettes in this condition. We feel confident, though, that for the most part subjects did smoke the lower-yield cigarettes outside of the laboratory when they were instructed to do so. First, if widespread noncompli- ance to our instructions had occurred, this would have tended to obscure between-condition differences in cotinine levels, yet there was a considerable difference in cotinine levels between conditions. Second, the mean cotinine level measured in the condition in which noncompliance would be most likely to occur, the ultra low-yield (0.1 mg) condition, was actually somewhat lower than those levels obtained from habitual smok- ers of ultra low-yield cigarettes in other studies (Benowitz et aL, 1986; Gori and Lynch, 1983). Compared to the high-yield (nicotine, 1.0 mg) condition, FTC yields (table 1) would predict CO reductions of 27, 70 and 87% when smokers switched to the lower-yield brands used in the present study. Obtained reductions in CO boost were 0, 25 and 55%. Chronic CO exposure levels were even less related to yield: afternoon (d.e., steady state) ezpired-air CO levels of about 36 ppm were measured during chronic smoking under all yield conditions except the ultra low-yield (0.1 mg) condition, in which CO levels were reduced to about 24 ppm. Our finding that CO levels were lower during smoking of the ultra low-yield cigarettes stands in contrast to several reports in which smokers of ultra low-yield cigarettes had CO levels that were not signif- icantly different from those measured in smokers of high-yield brands (Gori et, aL, 1986; Kanzler, et aL, 1983; Maron and Fortmann, 1987; Ossip-Klein et aL, 1983). Discrepancies between yield-predicted and obtained exposure levels may be due at least in part to behavioral changes that are apparent when smokers switch to low-yield cigarettes. In- deed, subjects in the present study significantly increased their daily cigarette consumption when smoking low- and ultra low- yield cigarettes (fig. 3). The increase in cigarettes per day is in agreement with recent reports of heavier smoking among ha- bitual smokers of low-yield cigarettes (Benowitz et al., 1986; Maron and Fortmann, 1987). Increases in daily cigarette con- sumption offset the lower tobacco weights used in low-yield cigarettes (table 1) and resulted in similar total amounts of tobacco burned per day across yield conditions (table 2). This very interesting compensation for total tobacco burned has been noted in two other recent studies (Benowitz and Jacob, 1984; Benowitz et aL, 1986). The present study also showed that low-yield cigarettes are smoked more intensively than high-yield cigarettes. In this study, number of puffs per cigarettes was similar across yield conditions (fig. 4). However, low-yield cigarettes contain less tobacco than do high-yield cigarettes, and when smoked in a standardized fashion by smoking machines, they require fewer puffs than do high-yield cigarettes (table 1). Smokers were able to take the same number of puffs from cigarettes with widely differing yields only by spacing their puffs more closely when smoking the low-yield brands (fig. 4). Other studies have also shown that equivalent or even greater numbers of puffs are taken from low-yield than from high-yield brands (Battig et aL, 1982; Bridges et aL, 1986), supporting the present observation of increased smoking intensity. In addition to puffing more rapidly, smokers in the present study also took larger puffs from low-yield than from high-yield cigarettes. This observa- tion is consistent with other previous reports (e.g., Tobin and Sackner, 1982; Woodman et aL, 1987). Taking larger and "ex- cess" puffs from lower-yield cigarettes most likely narrows the vol, ~ gap in actual amount of smoke which reaches the lungs {% lower- and high-yield cigarettes. Indeed, Woodman et aL (194 using radiolabeled cigarette smoke found that similar am of smoke entered the lungs from a low-yield (nicotine, 0.6 cigarette as from a high-yield (nicotine, 1.4 mg) cigarette Whk larger puffs were taken from the low-yield cigarette. It is interesting to note that smoking behaviors such as ptJ volume changed immediately when subjects in the preft study were exposed to new cigarette brands. This observati4 has been made previously by other investigators (Ashton et 4~ 1979; Woodman et aL, 1987). The immediate changes co* occur because smoking behaviors are controlled passively b, characteristics of the cigarette such as RTD. Alternativ* they could be due to active behavioral compensation by subj4 who are smoking cigarettes with low constituent yields. Dft from the present study support the latter interpretation is RTD was equated across cigarette brands in the present st,* (table 1). It should be noted, however, that behavioral changq may be made in response to reduced overall smoke concentnm tion from filter-vented cigarettes rather than in response ta levels of any particular constituent in the smoke delivered. In the present study, subjects did not alter either inhalatice volume or lung exposure duration in response to cigarette yieid alterations. This is not surprising inasmuch as these respiratory aspects of smoking do not appear to influence nicotine absorp. tion (Zacny et al., 1987). Subjects were more likely to altet those aspects of their smoking behaviors (i.e., volume and spacing of puffs and number of cigarettes) which had a direct impact on increasing nicotine (and other smoke constituent) exposure from lower-yield cigarettes. Filter vent-blocking is a behavioral strategy that can dra- matically increase the amount of smoke obtained from loR- yield cigarettes during natural environment smoking (Zacny d aL, 1986). The present study provided little evidence for con- sistent filter vent-blocking when subjects smoked ultra low- yield cigarettes. These findings are somewhat discrepant with other filter vent-blocking assessment studies which have found that at least a sizable minority of smokers appear to block filter vents (Kozlowski et aL, 1982; Robinson et al., 1983). In tha present study, filter blockade often could not be unequivocally judged from e=aminntion of stain patterns. Given the potency of this behavior in affecting smoke exposure (Zacny et a1, 1986), further refinements in the assessment procedure as well as studies with larger samples of smokers are needed before ae can conclude that vent-blocking is or is not consistently used by smokers to enhance exposure from low-yield cigarettes. Subjects could discriminate clearly between cigarette brands on a variety of subjective dimensions (table 3). Low- and ultra low-yield cigarettes were rated very poorly on measures d strength, harshness, taste, satisfaction and the likelihood thd they would be purchased for use by study subjects. Poor so- ceptability ratings of low-yield cigarettes have been reported previously (Benowitz et aL, 1986) and may reflect or eaplaia the low market share (<5%) captured by these cigarettes 4» Kozlowski, 1987). Subjects also gave lower liking ratings tO alternate brand high-yield cigarettes than to their usual-brand cigarettes with similar yield characteristics. This is consistent with previous reports (Benowitz et aL, 1986; Robinson et aL, 1982) and with the behavioral brand loyalty that is commoa' among smokers. Withdrawal symptoms, in particular cigarette cravin& tended to be exacerbated when subjects smoked lower-yield

I I I I I I I I I I I I I I I 1 I I 108 dFlrettes (fig. 5). This is consistent with the decreased cotinine ,els observed when subjects smoked these cigarettes (fig. 1). one previous study which examined this issue (West et aL, lyg4) failed to observe increased withdrawal complaints in lookers who switched to ultra low-yield (0.1 mg of nicotine yield) cigarettes. However, these subjects, unlike those in the orsent study, were motivated to quit smoking. ' This study has shown that although nicotine, cotinine and CO exposure levels from commercial brand cigarettes are re- lated in an orderly manneF to cigarette yield, yield fails to accurately predict the magnitude of relative exposure differ- ences across cigarettes with a wide range of yield characteris- acs. Furthermore, substantial increases in both frequency and intensity of smoking were noted especially during exposure to cigarettes with nicotine yields of 0.4 and 0.1 mg. These behav- ioral alterations, which suggested compensation for lowered constituent yield, explain at least in part the observation that relative between-brand nicotine, cotinine and CO exposure level differences were substantially less than predicted by FTC yield information. Acknowledgments The authors wish to thank Linda Felch for her assistance in statistical analysis of the data and Dn. Larry Chait and Harriet de Wit for their thoughtful comtoenta regarding the manuscript. References ASH'rON, H., STEPNEY, R. AND THOMPSON, J. W.: Self-titration by cigarette smokers. Br. Med. J. 2: 357-360, 1979. BArrIG, K., Buzzt, R. AND NIL, R.: Smoke yield of cigarettes and puffing behavior in men and women. Psychopharmacology 76: 139-148, 1982. BeNowrrz, N. L., HALL, S. M., HERNU+G, R. I., JACOB, P. T., JoNES, R. T. AND OSMAN, A. L.: Smokers of low-yield cigarettes do not consume lesa nicotine. N. EagL J. Med. 309: 139-142,1983. BeNowrrz, N. L. AND JACOB, P.: Nicotine and carbon monozide intake from high- and low-yield cigarettes. Clin Pharmacol. Ther. 36: 1265-1270. 1984. BENowrrz, N. L., JACOB. P., Yu, L. TAZcoir, R., HALL, S. AND JoNES, R. T.: Reduced tar, nicotine, and carbon monoxide exposure while smoking ultralow- but not low-yield cigarettes. J. Am. Med. Aasoc. 256: 241-246,1986. BwDGES, R. B., HUMBLE, J. W., TuRBEK, J. A. AND REHH, S. Ra Smoking history, cigarette yield and smoking behavior as determinants of smoke expo- sure. Eur. J. Respir. Dis. 69: 129-137, 1986. CHArr, L. D., Russ, N. W. AND GiurnTHs, R. R,: Effects of graded smoke inhalation and subsequent cigarette smoking. Addict. Behav. 10: 273-280, 1985. DUN N, P. J.: The effects of a reduced draw resistance cigarette on human smoking parameters and alveolar carbon monoxide levels. In Smoking Behaviour- Physiological and Psychological Influences, ed. by R. E. Thornton, pp. 203- 207, Churchill-Livinpton, Edinburgh, Scotland,1978. EUxT, R. V., McNABS, M. E., McCustmt, K. T. AND SNOw, S. L: Amount of nirntine and carbon monoxide inhaled by smokers of low-tar, low-nicotine cigarettes. J. Am. Med. Assoc. 250: 2840-2842,1983. EPSTEIN, L. H., OssiP, P. J., CoLEauN, P., HUGHas, J. W. AND Wnsr, W.: Measurement of smoking topography diuing withdrawal or deprivation. Behav. Ther. 12: 507-519, 1981. FtDERAL TRADE COMMISSION REPORT: "Tar," Nicotine, and Carbon Monoxide of the Smoke of 207 Variations of Domestic Cigarettea, February, 1985. FocsoM, A. R., PEcii,+cEx, T. F., DE GAUDEHARIS, R„ LvErxne, R, JACOBS, D. R AND GtLLUSt, Ii. F.: Consumption of "low-yield" cigaxttea: Its frequency and relationship to aanm thiocyanate. Am. J. Public Health 74: 564--668, 1984, GOSt, G. B., BENowrrz, N. L. AND LYNCS C. J.: Mouth veraua deep airways abeorption of nicotine in cigarette smokers. Pharmacol. Biochem. Behav. 25: 1181-1184, 1986. GORI, G. B. AND LYNCH, C. J.: Smoker intake from cigarettes in the 1-mg Federal Trade Commission tar c1ass. ReguL TouooL PharmacoL S: 110-120, 1983. Goat, G. B. AND LYNCH, C. J.: Analytical cigarette yields aa predictors of smoke bioavailability. Regul. ToziooL PharmacoL 6: 314-326, 1985. Gxt1'z, E. R.: Smoking behavior and tobacco abuse. In Advances in Substance Abuse, ed. by M. K. Mello, pp. 91-158, JAI Press, Greenwich, CT, 1980. ~NINGrtELD, J. E.: Behavioral pharmacology of cigarette smoking. In Ad- wncea in Behavioral Pharmacology, ed, by T. Thompson and P. B. Dewa, voL 4, pp. 131-210, Academic Pzda, New York. 1984. wLL, P., HALaY, N. J. AND WYNDEa, E. L.: Cigarette amoking: Carbo:ybe- Cigarette Brand-Switching 627 moglobin, plasma nicotine, cotinine and thiocyanate vs. self-reported smoking data and cardiovascular disease. J. Chronic Dis. 36: 439-449, 1983. HUGHES, J. R., HATSUKAMI, D. K.. PICKENS, R. W., KRAHN. D., MALIN, S. AND LUKNIC, A.: Effects of nicotine on the tobacco withdrawal syndrome. Psycho- pharmacoiogy 83: 82-87, 1984. JACOB. P. T., WILSON, M. AND BENOwITz, N. L.: Improved gas chromatographic method for the determination of nicotine and cotinine in biologic fluids. J. Chromatogr. 222: 61-70, 1981. JARvis, M. J., RUSSELL, M. A. H., AND SAL00JEE, Y.: Expired air carbon monoxide: A simple breath test of tobacco smoke intake. Br. Med. J. 281: 484-485,1980. KANZLER, M., JArFE, J. H. AND NEE, J.: Low nicotine cigarettes: Cigarette consumption and breath carbon monoxide after one year. Clin. Pharmacoi. Ther. 34: 408-415, 1983. KozLOwSKI, L. T.: Less hazardous smoking and the pursuit of satisfaction. Am. J. Public Health 77: 593-541, 1987. KozLOwSicI, L. T., FRECKEIt. R. C., KHOUw, V. AND POPE, M. A.: The misuse of "less-hazardous" cigarettes and its detection: Hoie-blocking of ventilated filters. Am. J. Public Health 70: 1202-1203, 1980. KOZLOWSKI, L. T., RICKERT. W. S., POPE, M. A., ROBINSON J. C. AND FRECKER, R. C.: Estimating the yield to smokers of low tar, nicotine, and carbon monoxide from the "lowest yield" ventilated filter cigarettes. Br. J. Addict 77: 159-165. 1982. MARON, D. J. AND FORTMANN, S. P.: Nicotine yield and measures of cigarette smoke exposure in a large population: Are low-yield cigarettes safer? Am. J. Public Health 77: 546-549, 1987. McMottxow, M. J. AND Foxx, R. M.: Nicotine's role in smoking: An analysis of nicotine regulation. Psychol. Bull. 93: 302-327, 1983. Moss, R. A. AND PRUE, D. M.: Research on nicotine regulation. Behav. Ther. 13: 31-46, 1982. OSSiP-KLEiN, D. J., EPSTEIN, L. H., WINTER, M. K., STU.LER, R., RUSSELL, P. AND DICKSON, B.: Does switching to low tar/nicotine/carbon monwdde-yield cigarettes decrease alveolar carbon monoxide measures? A randomized con- trolled triaL J. Consult. Clin. PsychoL 51: 234-241,1983. PRuE, D. M., SCoTT, R. R., MARTtN, J. E. AND LoMAN, B. D.: Carbon monoxide levels and rates of consumption after changing to low tar and nicotine ciga- rettes. Behav. Rea. Ther. 21: 201-207, 1983. RICKERT, W. S., COLLISHAW, N. E., BRAY, D. F. AND ROBINSON, J. C.: Estimates of mazimum or average cigarette tar, nicotine, and carbon monoxide yields can be obtained from yields under standard conditions. Prev. Med. 15: 82-91, 1986. ROBINSON, J. C., YOUNG J. C. AND RICKERT W. S.: A comparative study of the amount of smoke absorbed from low yield ("less hazssdous") cigarettes. Part 1: Non-invasive measures. Br. J. Addict. 77: 383-397, 1982. ROBINSON, J. C., YOUNG, J. C., RICKERT W. S., FEV, G. AND Kozr.owsiu, L. T.: A comparative study of the amount of smoke absorbed from low yield ("less hazardoua") cigarettes. Part 2. Invasive measures. Br. J. Addict. 78: 79-87, 1983. RUSSELL, M. A. H., JARVIS, M., IYEB, R. AND FEYERABEND, C.: Relation of nicotine yield of cigarettes to blood nicotine concentrations in smokers. Br. Med. J. 280: 972-976, 1980. RuSSELL, M. A. H., SUTTON, S. R., IYER, R., FEYgRABEND, C. AND VESSEY, C. J.: Long-term switchir}g to low-tar low-nicotine cigarettes. Br. J. Addict 77: 145-158,1982 TOBII+, M. J. AND SACKNER, M. A.: Monitoring smoking patterns of low and high tar cigarettes with inductive plethysmography. Am. Rev. Respir. Dia. 126: 258-264,1982. Tuxxes, J. A., SiLL>rrr, R W. ANn BALL, K. P.: Some effects of changing to low-tar and low-nicotine cigarettes. Lancet 2: 737-739, 1974. U.S. DEPA1rTMENT oF HEAL'rH AND HUxAN SERv1cES: The Changing Cigarette: A Report of the Surgeon General, 252 pp. U.S. Department of Health and Human Servicea, Public Health Service, Office of the Assistaat Secretary for Health, Office on Smoking and Health, Publication No. (PHS) 81-50156, Washington, DC, 1981. U.S. DEPARTI/ENT or Ha,ALTH AND HUMAN SEHVICFs: The Health Consequences of Smoking: Chronic Obstructive Lung Disease, A Report of the Surgeon General, 545 pp. U.S. Department of Health and Human Services, Public Health Service, Office of the Assistant Secretary for HeaIth. Office on Smoking and Health, Publication No. DHHS (PHS) 84-50205 Rockville, MD, 1984. WEST, R. J., RUSSELL, M. A. H., JARvis, M. J. AND FEYERASErtD, C.: Does switching to an ultra-low nicotine cigarette induce nicotine withdrawal effecta? Psychopharmacology 84: 120-123, 1984. WOODMAN, G., NEwuAN, S. P., PAvu, D. AND Ct.AwKE, S. W.: Response and acclimatiaation of symptomless smokers on changing to a low tar, low nicotine cigarette. Thoraz 42: 336-341,1987. ZACNY, J. P., Sztrzsa, M. L., BROwN, F. J., YINGLING, J. E. AND GaQaTHS, on R. R.: Human cigarette smoking: Effects of puff and inhalation parameters smoke exposure. tT. PharmacoL Exp. Ther. 240: 554-564,1987. ZAcNY, J. P., STSTZER, M. I.. AND YINGLING, J. E.: Cigarette filter vent blocking: Effects on smoking topography and carbon monoxide exposure. Pharm.coL Biochem. Behav. 25: 1245-1252,1986. Send reprint requests to: Dr. James P. Zacny, Department of Psychiatry, The University of Chicago, 5841 S. Maryland Ave., Chicago, IL 60637. I