Tobacco Documents Online

I A Psychopharniacological and Psychophysiological Evaluation of Smoking Motives Rico Nil I I I I I I I I I I I I I I I I Im Holzerhurd 30 CH-8046 Zurich Switzerland . ABSTRACT Compared to alcohol or opiate dependence, the physical withdrawal symptoms which occur with the cessation of the tobacco smoking habit are relatively weak, although they can produce discomfort. Long-term abstinence rates, how- ever, remain similarly low. This raises the ques- tion about the nature of the strength of this habit. Vvhen evaluating the complex mechanisms of cigarette smoking behavior and its determinants, a surprisingly large variety of pharmacological and nonpharmacological motives emerges. These appear to outweigh the health-related arguments for abstinence in the majority of smokers. An attempt has been made to categorize clas- ses of motives according to their positive or nega- tive reinforcing impacts on the habit. The acute tobacco withdrawal syndrome, problems with weight gain after cessation and the phenomenon of craving are classified as primarily negative reinforcers. Effects of smoking on cognitive func- tions and on "pleasure" are seen as primarily positive reinforcers. In conjunction with stress, the tranquillizing effects of smol:ing seem to have negative reinforcing properties in situations in- volving passive coping and anxiety, whereas smoking may have positive reinforcing effects in situations involving active coping. Reprint address: Rico Nil Im Holzcrhurd 30 CH-SWG Zurich Sµitzcriand VOLL'ME 9, NO. 2. 1991 It is suggested that the memory of these rein- forcing effects of smoking can contribute to the phenomenon of craving. Although substantially reduced after discontinuation of the smoking habit, craving may exacerbate and contribute sig- nificanth to late relapse. INTRODUCTION For almost 30 years tobacco smoking has been recognized as an important health risk factor. This is most particularly true for respiratory dis- eases, but also for diseases of the cardiovascular system. Therefore, in many industrialized countries major efforts have been made to in- stitute smoking prevention and cessation programs. On the other hand, scientific research programs designed to investigate the underlying mechanisms of the smoking habit have been much less prominent. As a result, a large number of programs in and outside of special smolang cessation clinics have been develnped, representing a wide range of ap- proaches and "philosophies," but, with the excep- tion of pharmacological treatment approaches, often lacking a scientific basis. They reach from suggestive treatments to purely pharmacological approaches. The latter mostly include nicotine substitution therapy using alternative routes of administration: nicotine-containing chewing gum /177, 1S'_;, nasal nicotine /17Z, 173/, transdermal nicotine /1, 164/, a nicotine aerosol to be inhaled /145/. Other pharmacological approaches are under discussion, for example, clonidine, an mj- as I

86 RlCO NTL I I I I I I I I I 1 I I I I I agonist which is in use as a pharmacological treat- ment strategy in alcoholics and opiate addicts. Clonidine has been shown to reduce craving and other symptoms of smoking withdrawal such as irritability, anxiety, restlessness, difficulty con- centrating and hunger /50, 138/. Furthermore, antidepressants, in particular specific serotonin uptake inhibitors such as zimelidine, fluoxetine and citalopram, which have been shown to sup- press ethanol consumption in animals and humans /118/, are being discussed as possible new phar- macological tools in smoking cessation /491. These approaches are interesting, not only be- cause they do not substitute nicotine but because they seem to have an anti-craving effect. Moreover, the newer, highly specific substances have minimal side effects. However, although alternative routes of nicotine administration or other pharmacological approaches can significantly reduce smoking withdrawal symptoms and thereby increase the probability of smoking abstinence /82, 156/, there is presently no clear-cut evidence that any one pharmacological or nonpharmacological smoking treatment program can offer a real advantage over the others. Schwartz /183/ analyzed the outcome of smoking cessation studies in the USA and Canada over a 10-year period and found that those smokers who decided to stop smoking without the help of a program reached the highest level of long-term abstinence. Recently, Gmur and Tschopp l51/ concluded, based on an analysis of a suggestive smoking treatment program, that the existence of such a variety of different smoking cessation programs is justified as long as no one method exists which offers a proven increased success rate. It is perhaps this great variety of cessation programs together with the above-mentioned findings of Schwartz which illustrates best the resistance of the smoking habit. Long-term relapse rates for abstinent smokers have been compared with those for alcoholic and opiate addicts, and they turn out to be almost identical /75/, not only with regard to the absolute rates but also in terms of their development over time (see Fig. 1). io, REI.APSE RATE OvER T(ME ,... +H"tRO1N t--~-+SMOKING o---CaLCaHoL Z"4"1 2 3 4 5 6 7 8 9 10 11 12 MONTHS Fig. 1: Relapse curves after cessation. The relapse cur- ves of smokers, opiate and alcahoi dependent persons appear to be almost iderrtical. From /75/. Copyright 1973 by the American Psychdogip! Association. Reprinted by per- mission. THE TOBACCO WTfHDRAWAL SYNDROME Nicotine dependence has been defined as a substance-induced mental disorder because "there is difficulty with quitting due to the exacer- bation of withdrawal symptoms following ter- mination of tobacco use" /see 82/. The American Psychiatric Association has defined nicotine de- pendence and the related withdrawal symptoms /3/. According to this definition, tobacco withdrawal is present when nicotine has been used daily for at least severalweeks and when an abrupt cessation of nicotine use, or a reduction in the amount of nicotine use, is followed within 24 hours by at least four of the following signs: 1. Craving for nicotine 2. Irritability, frustration, or anger 3. Anxiety 4. Difficulty concentrating 5. Restlessness 6. Decreased heart rate 7. Increased appetite or weight gain REVIEWS ON EWiROMMG\TAL I IEiLI1-i I

I I I I I I I I I I i I I I I I I S4f0KJNG 'A0't'fVF'S Is smoking cigarettes a form of drug depend- ence? Does smoking produce signs of physical dependence, or does it merely reflect a strong habit based upon psychological dependence processes? In discussing this problem it is useful to start with a general definition of drug dependence. Jar- vik and Hatsukami /81/ defined drug dependence as "a repetitive self-administration of a chemical substance characterised by an overwhelming in- volvement with the use of the substance, difficulty in refraining from its use and producing potential- ly adverse consequences". Tobacco smoking, involving the repetitive self- administration of a mixture of chemical substan- ces, appears to meet these criteria. The question then arises whether signs of physical dependence, particularly the development of withdrawal symptoms and effects of tolerance, also occur in habitual smokers. These signs may be of impor- tance for the question of drug dependence in general, but possibly also for the continuation of the smoking habit. In 1964, when attention was directed to the possible health consequences of smoking by the U.S. Surgeon General /201/, the question of tobac- co dependence was raised. One of the reasons for the uncertainty about tha.-existence of tobacco dependence was that signs of.physical depend- ence, in particular, a tobacco withdrawal syndrome, had not yet been recognized. In the meantime, major efforts have been made to investigate the effects of smoking cessation. An extensive review together with a detailed ques- tionnaire analysis was presented by Shiffman /188/. Early findings showed that a series of physiological changes can be observed after smok- ing has been discontinued. The perhaps most prominent of these was a decrease in heart rate accompanied by a decrease in diastolic blood pres- sure and a decreased excretion of catecholamines /117/. These findings were interpreted as signs of a general decrease in arousal. EEG studies sup- ported this conclusion in that smoking abstinence was shown to produce a decrease in peak alpha frequency and an increase in slow-wave activity /88/. Shiffman /188/ used a 25-item smoking withdrawal questionnaire and extracted five s- clusters of symptoms, namely physical symptoms, psychological symptoms (mental performance, mood and anxiety), arousal; appetite and the crav- ing for cigarettes. It was seen that the signs of withdrawal emerge rapidly (after 2 hours) and appear to decrease gradually during the first week of abstinence /188/. The analysis revealed that sex and riumber of cigarettes per day are important determinants, in that women and high cigarette consumers developed more severe symptoms than did males and modest consumers. During the past years, the existence of a tobac- co withdrawal syndrome has generally been ac- cepted, and increased efforts have been made to investigate single symptoms and their course over time. The development of smoking abstinence syndromes across the first day of abstinence was investigated by Gilbert and Pope /48/: A craving for cigarettes that increased steadily during the day and a decrease in heart rate were clearly and immediately (within the first 2 hours) present after cessation. Typical peaks in craving were observed after meals, corresponding well with the peaks in smoking frequency during unrestricted smoking control days. Furthermore, hand tremor was reduced and skin temperature higher, and there was a tendency towards more eating. A similar development of craving for a cigarette during the f rst hours of abstinence was observed by Glassmann et aL 1501. A decrease in heart rate after smoking cessa- tion was consistently reported by several authors, and it has been suggested that this effect may persist /210/. A recent long-term study of tobacco withdrawal symptoms with and without nicotine replacement sfiowed that excessive hunger and weight gain also persisted, at least during the 10- week observation period of that study /571. On the other hand, irritability, anxiety and impatience appeared to level off after about 4 or 5 weeks. Craving remained rather resistant, even to nicotine replacement (nicotine-containing chew- ing gum). This study, together with other studies which covered shorter periods /73, 74, 84, 181, 209/, confirms the existence of a tobacco withdrawal syndrome. Finally, it is interesting to note that similar symptoms of withdrawal, though somewhat lower in intensity, were observed in smokeless tobacco users (tobacco sniffing or VO[.L\lE9,\O.t1991 1

1- 1 I I I I I I I i I I I I I 88 RICO N7L chewing) after cessation /65/. Apart from the tobacco withdrawal syndrome, the development of tolerance effects might also be seen as an index of physical dependence. Tolerancc occurs if the effectiveness of a given dose decreases when administered in succession, producing, thus, a need to increase the dose for the same effect. However, in general, it appears that tolerance is not necessarily a sign of dependence and, on the other hand, that the development of tolerance is also not sufficient to define depend- ence /see 81/. Jarvik /78/ reviewed the develop- ment of tolerance to the effects of tobacco and discussed its possible importance with particular consideration of: a) metabolic tolerance (induc- tion of microsomal enzyme formation); b) be- havioral tolerance /see also 194/, and c) pharmacodynamic or tissue tolerance. For the smoker, an important tolerance effect involves the changes in sensory irritation of the respiratory tract produced by cigarette smoke. This sensation, which is aversive for a beginning smoker or a nonsmoker, is even considered to contribute to the pleasurable effects of smoking in habitual smokers /see 168/. Good experimental evidence has accumulated indicating a rapid development and dissipation of tolerance to the cardioaccelerating effects of nicotine. Rosenberg et al. /169/ showed that inter- mittent intravenous nicotine bolus injections produced gradually decreasing cardiovascular responses. Benowitz et aL /17/ were able to demonstrate that there is a strict parallelism of the increases in plasma nicotine and heart rate only within a very limited dose range. Furthermore, tolerance effects for heart rate responses to smok- ing were shown to dissipate to a great extent rather quickly, and the magnitude of tolerance was re- lated to cotinine plasma levels /152/ and also to inhalation behavior, in that consistently high CO absorbers showed greater tolerance effects than consistently low CO absorbers /133/ (see Fig. 2). Most recently, Perkins et aL /147/ also found that a measured-dose nasal nicotine spray produced a smaller increase in heart rate in heavy than in light smokers. PPm 6-1 4- 2-a 0 H• HIGH CO ABSORBERS L• LOW CO ABSORBERS C3 'DEPRIVED / •NONOEPRtVED H L H L A HEART RATE Chtrn. 1P.+n• 5ptn. z 2-y z i i H L H L Fig. 2: Group averages of singie-cigarette CO absorp- tiort (tidal CO boost) and heart rate response (heart rate) for selected, consistently high (H) and low (L) CO absorbing smokers. The cigarettes were smoked in the laboratory at 1 p.m. (over- night-morning deprived compared to non- deprived smoking) and at 5 p.m. (afternoon deprived compared to nondeprived smoking). The neslAts indicate that the heart rate response to smoking increased after deprivatian, in par- ticular in the high CO absorbers, suggesting the dissipation of acute heart rate tolerance during deprivation. From /133/. Copyright 1987 by Springer-Verlag. Adapted by permission. Although there is good evidence that chronic as well as acute tolerance develops for the heart rate response to nicotine, it appears to be more difficult to demonstrate tolerance for other func- tions. If tolerance developed for those effects of smoking which are important for smoking motiva-" tion, one would expect an escalating dose along a,~, smoking career. In fact, smoking intensity seems = to be lower rather than higher among older W~i• smokers with a longer smoking career /13/. ~ ~ O ~ I RL•VlL-1YS 0 \ L•*lVIROX`tGV rAL li£1.L.'Ri

I I I I a I 1 11 I I I I I I I I SMOK1`'G M0'iTVES In sum, a tobacco withdrawal syndrome and tolerance to some effects of smoking are well documented. The psychological and physical symptoms can produce significant discomfort, which may differ considerably in stre,igth between individuals. However, these symptoms do not ap- pear to approach the profile or the dimension of withdrawal effects seen in alcohol or opiate de- pendence. Whereas opiate and alcohol withdrawal include painful physical reactions (see DSM-III-R definitions, American Psychiatric As- sociation /3/), this is hardly the case with the decrease in heart rate or weight gain which are the main physical withdrawal symptoms of tobacco abstinence. What then makes smoking so resis- tant? SMOKING MODELS A comprehensive review of smoking models and the relevant supporting evidence has recently been presented by Ashton and Golding /5/. Generally, it has been increasingly recognized that smoking behavior is more than a simple es- cape-avoidance response to the aversive conse- quences of nicotine withdrawal, as was proposed in the late seventies 178, 171, 179/. It is now ac- knowledged that smoking also has positive rein- forcing effects for the smoker. Positve reinforcers are defined as stimuli (drug effects) producing pleasure or euphoria in a subject who is already in a"normaI" mood state. This definition may be weak, because such terms as "pleasure" and "nor- mal mood state" are difficult to define. Neverthe- less, although these terms depend on subjective judgement, they are contrary to any profile of negative reinforcement. Negative reinforcement is produced by a stimulus (drug effect) which ter- minates subjectivelytmpleasant states, such as dis- tress, or dysphoria or, in the case of opiate dependence, painful physical states. Positive smoking effects have been described in particular for cognitive processes, i.e., vigilance or rapid information processing (see below), whereas pleasure enhancing effects of smoking, for example after a meal or during a party, are anecdotically reported by many smokers, but are VOLtNE 9. NO. 2.1991 89 perhaps more difficult to operationalize. It is beyond the scope of the present review to discuss in depth the evidence which has accumulated over the past years ir.dicating that separate systems underlie the organization of positive and negative reinforcement. Important is that in these studies nicotine has been shown to stimulate the "trigger ione" of positive reinforcement, the ventral teg- mental area of the midbrain, thereby stimulating dopaminergic pathways /see 211 for review; 23, 97, 170/. Furthermore, the psychopharmaco- dynamics of nicotine appears to cover a broad range of possibly reinforcing effects/66/, including stimulation as well as tranquillization. Smoking models have also focused on arousal, proposing that smokers learn to use nicotine to manipulate their arousal level, i.e., to shift from either an aversive state of underarousal or an aver- sive state of overarousal to an optimal level /6, 11, 53/. The reinforcing smoking effect in this context lies in using smoking as a tool to control arousal homeostasis in unpleasant situations (stress, anxiety, boredom, etc.) /5/. However, experimental evidence has shown that smoking (nicotine) has direct and indirect functional impacts not only on arousal but, more generally, also in brain systems involved in govern- ing behavior, in particular in systems of reward/punishment, learning and memory /5, 11/. Consequently, more integrative models of smok- ing behavior (tobacco consumption in general) were called for. The multivariate biobehavioral model proposed by Pomerleau and his group /see 154/ represents perhaps the best accepted model today, with the most integrative understanding of the smoking habit. It proposes, based on central nicotinic pharmacodynamics, that the smoker learns to use smoking (nicotine) to regulate (fine- tune ) the body's normal adaptive mechanisms /see 154/. The model is based on a careful review of the nicotinic effects on endogenous neuroregulators. However, even with this multi- variate model of smoking behavior, it still remains& ZP difficult to explain why ex-smokers relapse, evenl:'-~ ~ long after cessation and, thus, after complete me-.`," tabolic clearance and after overcoming acute.;,,, withdrawal symptoms. c~ ~ ~ ~ ~o

90 RICO NIL I I I I I I I I I p I I 1 I I DETERl11INANTS OF CIGARETTE SMOKING BEHAVIOR 1. Why microanalyses of smoking behavior? A serious problem in the experimental analysis of smoking dependence is the evaluation of the smoke/nicotine dose. As opposed to the con- sumption of other substances where the manipula- tion and the measurement of the consumed dose is relatively simple, smoking behavior is such a complex process, resulting from the interaction between the smoker and his cigarette, that the smoke dose is difficult to estimate. The cigarette alone is a highly complex product delivering a mixture of thousands of substances. Smoking be- havior is complex too, including a great variability in puffing and inhalation behavior, shaping the smoke bolus in the mouth and, finally, inhaling it with a variable amount of expiratory air. The latter represents the most important act for nicotine absorption. Smoking behavior, there- fore, allows the individual tailoring of dosage by means of several adaptive behavioral ' mechanisms. These behavioral mechanisms ap- pear to be rather independent of each other/131/, suggesting that different modes of control seem to operate for the number of cigarettes a smoker smokes, for cigarette puffing behavior, and finally, for respiratory inhalation and alveolar smoke ab- sorption. Furthermore, the relatively high degree of independence of the single measures hampers the direct comparability of studies using different variables to describe smoking behavior. The evaluation of these behaviors, however, offers the possibility of investigating smoking motivation, assuming that factors which increase the intensity of smoking behavior indicate positive "motivators" and that factors which cause a decrease in smoking intensity indicate negative "motivators." The investigation of smoking be- havior appea-s, thus, to be a promising attempt to elucidate better the determinants of smoking motivation. Since the various components and measures of smoking behavior appear to be variable and sen- sitive to a serie.; of factors, it seems warranted to search for their principal determinants in an at- tempt to elucidate smoking motives. 2. Cigarette design The most urgent question concerns the impact of changing cigarette smoke deliveries on smoKing behavior and smoke absorption. The increased relevance of this question stems from the con- tinuous decrease of cigarette smoke yields which has been observed over the last few decades and the parallel tendency toward a somewhat lower health risk from such cigarettes /99I. With respect to smoking behavior, the main findings are that smoking lighter cigarettes leads, if at all, to only a marginal increase in cigarette consumption /44, 192/. On the other hand, single-cigarette smoke absorption, as assessed by biochemical markers, varies enormously between subjects. A decrease in smoke absorption in association with decreas- ing cigarette smoke deliveries is, therefore, dif- ficult to see in cross-sectional studies /13, 18, 35, 175/. However, a closer inspection of such cross- sectional data suggests that low- and in particular ultra-low-yield cigarettes (less than 0.5 mg nicotine yield) might preclude high levels of smoke absorption /18,129,132,174/. The perhaps most comprehensive cross-sectional analysis of smoking behavior and smoke absorption na revealed that 10 to 17% of the plasma nicotine boost could be explained by the nicotine yield of the cigarette, reflecting a slight but significant positive covariation. The great individual variance, - even within smokers of the same cigarette type, together with the absence of a clear relationship between cigarette smoke yields and biochemical markers of smoke absorption, reflects the wide range in which regulatory mechanisms of smoke absorp- tion are operative. An increasing number of cross-sectional as well as switching studies (which require the smoker to switch from one cigarette type to another) have demonstrated compen- satory adaptations in cigarette smoking behavior in response to varying smoke deliveries. In an attempt to relate smoking behavior to cigarettes differing in smoke yield, it was fo that the estimated mouth intake of smo depended clearly on the strength of the cigare" /131/. These relations prevailed in spite of a com- pensatory increase in puff volumes with decre:IP ing cigarette smoke yields, indicating that pt~ ~ 2G46400109_ '=> REVILI~VS ON E~'N'iRO\'~tEMAL I {F11L;Ti

I I I I I I I I I I I I I I I I sHOKJ`c uoTlvFs volume compensation was not complete. A.!- veolar CO absorption remained independent of the CO yield of the cigarettes, suggesting an al- most ideal compensation of smoke absorption at the alveolar level. Respiratory inhalation volumes were positively rather than negatively related to the smoke yield of the cigarettes. This result is difficult to explain and inconsistent with the idea of a compensatory increase in respiratory inhalation volumes with decreasing smoke yields of the cigarettes. It suggests that inhalation volumes do not necessarily represent inhalation efficiency. It was speculated that with strong cigarettes smokers might dilute concentrated smoke by increasing inhalation volumes. On the other hand, they might avoid a further dilution of light-cigarette smoke. Nevertheless, these results supported other failed attempts to relate respiratory inhalation volume to cigarette smoke deliveries or to alveolar smoke absorption /2,105/. Beside the effects of varying cigarette smoke yield values, other factors affecting smoking be- havior have been investigated. A series of ex- perimental manipulations of puffing conditions /119, 120/ provided evidence that the length of the tobacco rod affects puffing behavior. Such studies demonstrated that puff duration, but not puff in- terval, depends on the length of the tobacco rod and on draw resistance, whereas no significant effect of visual control, distance to the burning ember, smoke temperature or smoke deliveries was observed. Nemeth-Coslett and Griffiths /121/ further showed that puff volume, but not puff duration, was sensitive to the pharmacological delivery of smoke, and they also suggested that puff volume and puff duration were insufficient to predict CO absorption. A series of studies inves- tigated puffing behavior along the burning time of a cigarette. It was seen that a change in puff ng behavior can be caused by both the development of satiation of the smoking need and a change in smoke characteristics (smoke composition, smoke temperature, etc.) towards the end of a cigarette /see 212/. Heriing et aL /69/ and Buzzi et aL /22/ found that puff volumes decreased along the burn- ing time of the cigarette in some of the smokers tested. However, along 30 puffs of similar smoke quality (presentation of a new cigarette every third 91 puff), this trend toward smaller puff volumes along the sr'rloking time could not be replicated / 14, 12s/. Taken together, it appears that a whole series of cigarette design characteristics may affect cigarette smoke puffing and inhalation behavior. Furthermore, it was observed that different smok- ing behavior variables remain relatively unrelated /129/. This not only limits the comparability of studies using different smoking behavior vari- ables, but, importantly, it also points towards a multifactorial control of smoking behavior. How- ever, nicotine, on account of its psychophar- macological profile /11/, is generally regarded as the primary reinforcer of smoking behavior. A separate investigation of the role of nicotine as a determinant of smoking behavior is, thus, jus- tified. 3. Nicotine Regarding smoking as a form of nicotine self- administration, it has been suggested that adaptive mechanisms in smoking behavior allow the smoker to satisfy his individual pharmacological and/or nonpharmacological need for smoking even when smoking a lighter cigarette. Such ob- servations have supported the so-called "nicotine titration hypothesis", proposing that each smoker seeks his individual (homeostatic) nicotine dose. A series of attempts have been made to isolate the role of nicotine in determining smoking: a) statistical separation of the effects of nicotine from those of other smoke constituents (tar, CO); b) independent manipulation of nicotine and other smoke components by means of experimen- tal cigarettes; c) nicotine specific pharmacologi- cal manipulations such as presmoking preload with nicotine or the use of nicotine antagonists. a. Statistical separation of effects A few cross-sectional studies attempted to separate statistically the effects of the nicotine delivery of cigarettes from those of tar or CO. Bdttig et al. /13/ used the method of partial correla- tions relating puffing behavior variables to the smoke deliveries of cigarettes and found among I vOLLME 9, \O. ? 1991

92 RlCO `iL I I 1 I I I I ~ I I I I I I I men (n = 67) just one negative relationship be- tween tar yield and total puff volume per cigarette. Similar results were found by Sutton et al. /198/, who used the method of multiple regression analysis: the nicotine delivery of the cigarettes correlated positively and the tar delivery negative- ly with puff volume. Furthermore, the plasma nicotine boost was positively related to tar delivery and negatively related to nicotine delivery. These results were interpreted as evidence for tar rather than nicotine compensation and, thus, against the nicotine titration hypothesis. Evidence for a determining influence of nicotine on smoking be- havior was found in the female subject sample (n = 43) in the Battig et aL /13/ study, in that nega- tive relationships between nicotine yield and both alveolar CO absorption (mixed expiratory air CO boost) and an inhalation efficiency index were found as well as a positive relationship to puff interval. However, the differentiation of tar, nicotine and CO effects by means of multivariate statistical methods should be interpreted with cau- tion, because the tar, nicotine and CO yields of cigarettes are highly intercorrelated. Conse- quently, the results of multivariate analyses are based on rather small amounts of rest variances. It is therefore advisable to verify such results by means of direct experimental manipulations. b. Varying nicotine: tar and nicotirte: CO ratios A number of studies have explicitly tested the hypothesis of nicotine regulation and investigated for this purpose whether smoking behavior would be changed from cigarette type to cigarette type in such a manner as to maintain nicotine intake at a similar level. A detailed comparative analysis and interpretation of these studies has been presented / 129/. A fair degree of smoking for nicotine or incom- plete compensation in smoking behavior in response to changing nicotine deliveries was reported by most of the studies /52, 60, 69, 105, 160, 184, 213/, whereas clear evidence for nicotine regulation through smoking was reported by Jar- vik et aL /83/, Dunn and Freiesleben /34/ and Fagerstrom /38/, who found nicotine regulation to depend on the degree of nicotine dependence. Two well-designed studies, however, found evidence of tar rather than nicotine regulation /4, 193/. Most of these studies rated nicotine as being important for smoking behavior, but they general- ly concluded that compensation for nicotine was incomplete. c. Nicotine-specific pharmacological effects The next set of studies to be considered repre- sents attempts to manipulate smoking behavior by means of acute applications of nicotine-specific pharmacological agents. Three studies inves- tigated the effect of intravenous nicotine applica- tion and revealed somewhat conflicting results. Lucchesi et aL /100/ and Henningfield et aL /67/ found a decrease in cigarette consumption after i.v. nicotine. The latter study reported that nicotine self-administration occurred under a FR- 10 schedule of reinforcement. Kumar et aL /94/ found that smoking behavior remained unaffected by i.v. nicotine but not by previous smoke inhala- tion. The effects of nicotine-containing chewing gum on subsequent smoking behavior were ex- plored by Kozlowski et aL /93/, Russell et aL /177/ and Nemeth-Coslett and Henningfield /122J. These studies generally found a compensatory decrease in smoking behavior intensity, which, however, was hardly reflected in the biochemical markers of smoke absorption. Jarvik et al. /80/ used nicotine-containing capsules and Rose et aL /164/ the method of transdermal nicotine, and both found compensatory adaptations in smoking behavior. Finally, three studies explored the ef- fect of the nicotine antagonist mecamylamine /124, 166, 195/ and consistently found an increase in smoking behavior intensity. Taken together, most of these experiments were successful in affecting smoking behavior by means of the presmoking administration of nicotine, or, even more clearly, by the presmoking administration of the centrally acting nicotine an- tagonist meramylamine. The perhaps most im- portant conclusion of the studies using nicotine preloads is that the route of nicotine application seems to be crucial for its effects. The strongest effects on smoking behavior were seen with the presmoking application of nicotine by means of smoke inhalation, whereas the effects of in- travenous nicotine applications were much less 2C46400111_ RBViL•ti1'S Oti EVb'iRONMENTAL HF.AL.'nt

~~ SMOKING MO'iTVES 43 I I I I I I I I I I I I I I I I clear-cut. Kumar et al. /94/ found no effects with intravenous 5-second pulses of nicotine. Lucchesi et al. /100/ found effects on smoking only with certain temporal injection patterns, and Hen- ningfield et al. /67/ found effects'with a subject- paced procedure of nicotine self-administration. A limitation of the latter study, however, is the small number of subjects (n=3), and a more general criticism is that of the absence of nicotine plasma measurements. 4. Smoke "irritation" and smoke "taste" As described above, the most widely studied manipulations of cigarette design so far have in- volved the lowering of smoke yields. However, evidence has been accumulating which suggests that non-nicotinic factors may also be important in affecting smoking behavior: the incomplete compensation through smoking behavior for nicotine uptake (see above) and the evidence presented by Stepney /193/ and Sutton et a1. i198/ that smokers appear to compensate for tar rather than nicotine. - Furthermore, Rose et aL 1167/ reported that smokers rated their smoking satisfaction as lower and their craving as less reduced after smoking when their upper airways were anesthetized. In another attempt to investigate the significance of smoking-induced airway irritation, it was found that the inhalation of an irritating citric acid aerosol /165/ or a nicotine-free cigarette smoke fraction /163/ was effective in reducing cigarette craving. Jaffe and Glaros r/7/ found two or- thogonal smoke taste factors on the basis of which subjects discriminate between cigarettes, namely a flat-sharp dimension which is highly inde- pendent of nicotine, tar and CO yields and a second, high vs. low nicotine dimension. A study by Nil and Battig /128/ investigated smoking be- havior across taste categories and found an in- crease in smoking intensity from mentholated to dark tobacco to•blond tobacco. This effect, which turned out to be independent of the smoke yield values of the cigarettes, was reflected in puff volume, inhalation time and CO absorption. It was therefore concluded that factors which affect cigarette smoke taste have effects on smoking behavior which are separate from those obtained by comparing smoke yields. 5. Nonmedical drugs In the investigation of interactions between smoking and substances which are widely and regularly used in the general population, it is the pharmacological determinants of cigarette smok- ing behavior that are of particular relevance. m Alcohol A series of investigations found positive rela- tions between the consumption of alcohol and smoking /10, 16,76, 199, 202/. The question arises therefore whether this relationship is the conse- quence of psychopharmacological interactions between the two substances or whether it simply reflects a parallelism of the two consumption habits. Investigations under laboratory conditions were carried out by several groups in both al- coholics and social drinkers in order to elucidate pharmacological interactions. Griffiths et a1. /56/ reported an increased rate of cigarette consump- tion among alcoholics after ethanol consumption. Mello et d /107, 108/, Mintz et al. /113/ and Nil et al, /130/ demonstrated in social drinkers that al- cohol not only increased the rate of smoking but also intensified puffing and respiratory inhalation. Although ethanol appears to be a powerful deter- minant of cigarette smoking, the underlying mechanisms remain unclear. Whereas additive effects of nicotine and al- cohol on cardiovascular parameters were shown by Benowitz et ai: /19/ and by Myrsten and Andersson /116/, antagonistic interactions were reported for reaction time and for skin tempera- ture and finger pulse amplitude /110/. Further- more, Knott and Venables /89, 90/ interpreted their results on EEG variables in terms of both synergistic and antagonistic effects. Recently, Michel and Battig /110/ investigated the separate and combined effects of smoking and alcohol (0.7 gfkg) -)n mental performance in a rapid informa- tion processing task, task-related EEG measures and cardiovascular functions. It was reported that VOI l'ME 9, \U 2.1991 ~

I I I I I I I I I I I I I I I I I I 94 RICO \'IL smoking not only increased information process- ing but that it also counteracted the impairing effects of alcohol on the processing rate. Smok- ing, however, did not counteract the alcohol-in- duced increase in reaction time in this task, nor did it modify components of task-elicited evoked potentials which are thought to represent the cog- nitive part bf information processing (N1, LP [late positivity] and CNV). On the other hand, this study showed that smoking after alcohol caused an activation at the peripheral and cortical levels, expressed as cardioacceleration, an increase in beta power and a reduction of the alcohol-induced increase in alpha power. Further, smoking antagonized the effects of alcohol on the flicker fusion threshold /96/ and on two-flash fusion /200/, whereas there was no inter- active effect on free recall /85/ or on reaction time /90/. Taken together, the results of studies on the interactive effects of alcohol and smoking general- ly favor the concept of an at least partially an- tagonistic interaction. Presently, it is not clear whether alcohol increases the pleasurable effects of smoking, thereby increasing smoking behavior intensity, or whether it decreases smoking satis- faction, resulting in a compensatory increase in smoking. From the results presented by Benowitz et aL /19/, it appears to be clear that ethanol does not interact with nicotine metabolism. tion. The latter finding, however, was interpreted to be independent of the caffeine content. Koz- lowski /91/ found that caffeine depressed nicotine consumption, as evaluated by cigarette butt analysis. Nil et al. /130/ found no consistent effects of caffeine on subsequent cigarette smoke puffing and respiratory inhalation behavior. Finally, Chait and Griffiths /24/ reported decreased cigarette consumption after caffeine in some but not all subjects. Although such behavioral studies have failed to give a clear-cut impression of the nicotine-caf- feine interaction, a recent study by Rose /161/ suggested interactive effects on arousal, in that smoking antagonized subjective arousal after caf- feine. Moreover, caffeine appeared to increase inhalation. Further evidence of an interaction be- tween the two substances comes from phar- macokinetic studies showing that caffeine metabolism is increased among smokers as op- posed to nonsmokers /86/. On the psychophysiological level, additive effects on blood pressure and heart rate were reported /42, 157/. In sum, the lack of consistent and clear-cut results suggests subtle, and perhaps situationally and/or individually defined, interactions between nicotine and caffeine /see 130, 161/. c. Other nonmedical drugs b. Caffeine Caffeine is perhaps the most often consumed psychotropic substance, and its joint consumption with a cigarette is not only often observed, but also reflected in positive correlations between the con- sumption of the two substances /see 76/. Again, the question arises whether the co-consumption of caffeine and nicotine has a basis in psychophar- macological interactions. A series of laboratory studies were carried out to investigate behavioral interactions: Ossip et aL /140/ and Ossip and Epstein /139/ found no effects of coffee drinking on the number of cigarettes smoked during a 1-hour session, whereas Marshall et al. /103, 104/, using a similar experimental set- ting, reported an increased cigarette consump- Two studies using a residential research setting allowing the subjects to self-administer marijuana showed a temporal correlation between cigarette. and marijuana smoking. However, no evidence was found that marijuana smoking systematically affects cigarette smoking /108/. Further, a study by Nemeth-Coslett et aL /123/ extended smoking behavior measures to detailed analyses of puffing and inhalation and found no effects of marijuana smoking. There is evidence that opiates affect smoking behavior. Mello et aL /109/ found in former heroin addicts that they smoked more cigarettes when heroin was available. Bigelow et aL /21/ observed a covariation of methadone dose with cigarette smoking. On the other hand, Karras and Kane /87/ reported tVjVT~ an lopi~d an- ~ REVIL•~YS ON EN1'iRO.1tV;TAL PtFALTit