Tobacco Documents Online

I I I I I I I I I I I I I Figns and Symptoms of Tobacco Withdrawal gihn R. Hughes, MD, Dorothy ffatsukami, PhD • To test the validity, magnitude, and clinical significance of jhe signs and symptoms of tobacco withdrawal defined by bSM-IU, both observed and reported signs and symptoms were jpeasured In 50 smokers during two days of ad lib smoking and f,hen during the first four days of abstinence. Observer and pubject ratings of the DSM-lll symptoms of craving for tobacco, krttability, anxiety, difficulty concentrating, and restlessness jrtcreased after cessation. In addition, bradycardia, impa- pence, somatic complaints, insomnia, increased hunger, and increased eating occurred after cessation. The frequency and {rttensity of these symptoms varied across subjects; however, the average distress from tobacco withdrawal was similar to that observed In psychiatric outpatients. Subjects who had more withdrawal discomfort were more tolerant to the car- diovascular effects of nicotine. Subjects who had more with- drawal discomfort did not have a lower rate of smoking cessation. (Arch Gen Psychiatry 1986;43:289-294) M any signs and symptoms have been attributed to smoking cessation.'-' Seven of these symptoms- craving for tobacco, irritability, anxiety, difficulty concen- trating, restlessness, headaches, drowsiness, and gastroin- testinal (GI) tract disturbances-have been collated and termed tobacco urithdrawal in DSM-111.' The significance of these symptoms is indicated by several surveys in which smokers reported that the onset of withdrawal prompts smoking, deterred attempts to stop smoking, reduced cessation success, and caused relapse to smoking.2 A number of studies have. documented tobacco with- drawal phenomena; however, the validity and generalizabil- ity of these studies may be limited. One of the concerns about the validity of the studies is that most relied on self-reported symptoms that may have been inaccurate. However, objective changes that parallel self-reported withdrawal symptoms have been demonstrated in several well-documented laboratory experiments. In this setting, abstinence induces changes in neurohormone levels (eg, increased levels of epinephrine), neurophysiologic features (eg, increased slow-wave activity on the electroencephalo- gram), cardiovascular function (eg, bradycardia), and in- tellectual performance (eg, decreased vigilance).Ll We are unaware of prior studies that directly observed the more relevant withdrawal phenomena of anxiety, irritability, restlessness, etc. Another concern is that the reliance of ~ most studies on retrospective designs may have allowed a "rationalization bias" to occur. Rationalization bias refers to I I I I the possibility that smokers may exaggerate withdrawal symptoms to rationalize their inability to quit. Systematic' prospective studies in which subjects are observed on a research ward during drug use and then during abstinence would prevent both the problems of self- report and retrospective designs. Such studies are common in the fields of alcohol, sedative, and opiate withdrawal; however, similar studies for tobacco withdrawal have only Accepted for publication Sept 27, 1984. From the Department of Psychiatry, University of Minnesota, Minneapo- W. Reprint requests to Department of Psychiatry, University of Vermont College of Medicine, Burlington, VT 05405 (Dr Hughes). recently been completed by our group' and others.' These inpatient studies confirmed most self-reported withdrawal symptoms. In addition, our study reported that subjects had few withdrawal symptoms as inpatients but many symptoms after discharge to the natural environment.' This result suggests that environmental cues (eg, other smokers) are crucial to the onset of tobacco withdrawal. A final concern with studies on tobacco withdrawal is that several important facets of the disorder have not been well described. For example, reports of the severity, duration, and time course of tobacco withdrawal are quite varied. Also, why some smokers have severe withdrawal symptoms and others have minimal symptoms is unknown. Finally, and perhaps most importantly, whether withdrawal symp- toms influence the ability to quit smoking has not been adequately tested." The present study provides a detailed and systematic description of tobacco withdrawal so that many of, these unanswered questions can be addressed. In addition, the study provides both an objective and a subjective evaluation of tobacco withdrawal in the natural environment. We believe this strategy produced more valid and generalizable results. The results of the study (1) validate most of the DSM-III symptoms of tobacco withdrawal and several other signs and symptoms, (2) demonstrate that tobacco withdrawal induces clinically significant discomfort, and (3) associate tobacco withdrawal with tolerance to nicotinic effects. The results of the study do not confirm the hypoth- esis that tobacco withdrawal reduces success at smoking cessation. SUBJECTS AND METHODS Sample Population This article is based on the results from 50 smokers who were randomly assigned to receive placebo gum during a double-blind study of the effect of nicotine gum on the signs and symptoms of tobacco withdrawal.' In the study, subjects were recruited by a public service announcement. To be included, subjects must have wanted to stop smoking, have had at least one episode of tobacco ' withdrawal and be tobacco dependent as defined by DSM-III.' Procedures Informed consent was obtained after the study was explained to the subjects. Subjects then placed a$25 deposit, which was refunded if they attended all experimental sessions. Next, they completed a smoking history form and four questionnaires to determine their degree of dependence: the Fagerstrom'Iblerance Questionnaire,° the MacAndrews Addiction Scale,' the Reasons for Smoking Scale,' and a rating of the severity of withdrawal symp- toms in the past. Subjects were told to continue to smoke as usual for the next 24 hours. On the next evening, subjects were tested for tolerance to nicotine by recording their heart rate and drawing a blood sample for determination of nicotine level before and two minutes after they smoked one of their own cigarettes in their normal manner. Plasma nicotine level was measured by radioimmunoassay10 by the American Health Foundation (Valhalla, NY).lblerance was calcu- lated as the increase in heart rate divided by the increase in plasma nicotine level and was then corrected for body weight. Next, subjects were counseled about smoking cessation for a few (less than ten) minutes. They were instructed not to smoke after awakening the next morning and to use gum when they had a craving for a cigarette. On the 1st, 2nd, and 4th evenings of abstinence, subjects returned to report any cigarettes smoked and give a breath sample I Arch Gen Psychiatry-Vol 43, March 1986 Tobacco Withdrawal-Hughes & Hatsukami 289

I I I I I I I I I I I I I I I I I I I for analysis of carbon monoxide to verify self-report.° On the last day of the study; they indicated which gum they believed they had received. - Measures A battery of measures of the signs and symptoms of tobacco withdrawal were administered on the two evenings during baseline and the three evenings during abstinence. These measures were based on a review of the literature" and our previous studies.` Subjects were not given any expectancies about the severity of withdrawal symptoms. In addition, they attended individual ses- sions to prevent comparison of symptoms among subjects. Three withdrawal signs were measured. Heart rate was mea- sured after five minutes in the supine, position. Orthostatic re- sponse was determined by measuring the heart rate two minutes after standing and subtracting the supine heart rate. Weight was measured to the nearest 0.1 kg. Several withdrawal symptoms were measured. The DSM-III Table 1.-Sampte Characteristics; CharaolsNsttc Preant Sample (N=50) US Avwayst Demographics Age, yr 38 (10.0) NA# Men 23 (46) (55) Completed high school 49 (98) (61) White collar 31 (62) (35) Earning >$25,000/yr 26 (52) (35) Smoking habits Cigarettes/day 29 (11.5) 24 Nicotine, mg/cigarette 0.9 (0.3) 1.1 Duration of srtwking, yr 21 (9.4) 21 94 tried to quk>3 times 26 (52) (39) *Mean (SD) or number of subjects with the percentage in parentheses. tEstimated from Public Health Service statistics13u $NA indicates not available. criteria for tobacco withdrawal-cigarette craving, irritab anxiety, difficulty concentrating;,restlessness, headache, ness, and GI xract disturbances-=were rated by the subj follows: 0, not present;1, mild; 2, moderate; and 3, severe in In addition, subjects rated fatigue, impatience, hunger, and various somatic complaints (eg, sweating and dizziness) similar scale. Alcohol and caffeine intake were measured by self-report. Insomnia was measured by self-reported sleep quacy, sleep latency, number of awakenings, time awake, and sleep time from a standardized sleep inventory.' Subjects completed the Profile of Mood States (POMS) questionnaire, w produced scores for vigor, anger, fatigue, tension/anxiety dep sion/dejection, and confusion,. as well as a total mood disturbanoa score. ° To verify self-report, each subject had a spouse, relative, or friend rate his or her irritability, anxiety, restlessness, fatigu g drowsiness, impatience, and somatic complaints on a rating scaw identical to that used by the subject. Subjects were instructedtoj recruit only persons who had sufficient time to observe them an~ could rate them every day. Observers were instructed to ignore; self-report of symptoms as much as possible and to base theii ratings on observed changes in the subject's behavior. Obse were not asked to rate cigarette craving, difficulty concentratinlas hunger, eating, or insomnia because pilot work indicated they couldi not accurately rate these symptoms. Data Analysis Initially the representativeness of the smokers studied wsa estimated by comparing their demographics and smoking habita; with those of recent population-based samples of smokers."-" Several sets of analyses of the signs and symptoms were calcu- lated. Both the interrater and intrarater reliability of the subject and observer ratings were determined. The intrarater reliability was calculated by comparing both the observed and reported sign and symptom on the first and second days of baseline wit.h Wilcoxon's matched-pairs signed-ranks test." The interrater re- liability was calculated with Spearman's rank-order correlation coefficient between subject and observer ratings." The validity af each sign and symptom was determined by comparing the meann baseline and abstinence scores with two-tailed paired t tests. The magnitude of each sign and symptom was determined by subt.ract- Median ratings for most prevalent symptoms of tobacco withdrawal syndrome. Craving Severe r Baseline Abstinence Moderate 4- 0-- Mild Not Present Severe Moderate Mild Not Present ~ I I i I 1 I Difficulty Concentrating r Baseline i Abitinencw 1 2 3 4 6 Days 0-0 i I 1 f 1 I I 290 Arch Gen Psychiatry-Vol 43, March 1986 Irritability Anxiety Baseline Abatlnenc. r Baseline Abstinence ^- I 1 i I { I Restlessness Impatience r- Baseline , Abstinencs r- Basehne , Abstinence 0__0 1 1! 1 1 i 1 2 3 4 6 Days I- I 1 i I I I 1 2 3 4 6 Days r Tobacco Withdrawal-Hughes & Hatsukand

I I I I I I I I I I I I I I I I I I ~ Table 2.-Mean (SE) Scores for Subject Ratings of Signs and Sympioms- Baseline Abstinence t Physiobgic signs T Supine heart rate, beatsJmin - 74.2 (1.5) 64.8 (1.3) -7.8t prthostatic response, beats/min 13.8 (0.9) 10.9 (0.9) -2.7# Weight, kg 68.7 (1.9) 69.8 (1.8) NS DSM-!Il symptoms (0-3) Craving for tobac~.ro 1.8 (0.1) 2.3 (0.1) 3.4t Irritability 0.6 (0.1) 1.4 (0.1) 7.3t Anxiety 0.9 (0.1) 1.7 (0.1) 7.9t D'rfficulty concentrating 0.5 (0.1) 1.2 (0.1) 5.2t Restlessness 0.8 (0.1) 1.3 (0.1) 4.6t Headaches 0.2 (0.1) 0.4 (0.1) NS Drowsiness 0.6 (0.1) 0.7 (0.1) NS GastrointesUnal tract problems 0.2 (0.1) 0.2 (0.1) NS Other symptoms (0-3) Fatigue 0.8 (0.1) 0.7 (0.1) NS Impatience 0.7 (0.1) 1.3 (0.1) 6.tt Somatic complaints 0.5 (0.1) 0.7 (0.1) 2.6 Eating/drinking Eating (0-3) 0.3 (0.1) 0.9 (0.1) 52t Hunger (0-3) 0.3 (0.1) 0.7 (0.1) 2.9# Alcohol intake, drinks/day 1.7 (0.3) 1.0 (0.2) -3.1# Caffeine intake, drinks/day 5.4 (0.5) 4.9 (0.4) NS Sleep Sleep adequacy (0-3) 0.9 (0.1) 0.7 (0.1) -3.0$ Latency, min 13.5 (1.8) 13.6 (2.1) NS Awakening, No. 0.9 (0.1) 1.4 (0.1) 1.9§ Time awake, min 11.5 (2.6) 14.9 (2.7) NS Tirneasleep,hr 7.2 (0.1) 6.6 (0.1) -2.7$ 'N- 50. tP<.001. $P<.01. §P<.05. ing the mean baseline score. from the mean abstinence score (except the magnitude of weight gain, which was determined by using the weight on the last day of abstinence). The incidence of each sign and symptom was determined by calculating the percent- age of subjects whose mean abstinence score was greater than (or, in some cases, less than) their baseline score. Significant changes in the time course of each sign and symptom during abstinence were determined by comparing the scores on the first, second, and fourth days of abstinence with Friedman's analysis of vartance.' The total intensity of withdrawal discomfort was determined by subtracting the mean baseline score from the mean abstinence score for the symptoms that were validated and then summing these difference scores. Since the ratings were based on the same ordinal scale and had equivalent variances, Z-score transforma- tions were unnecessary. The determinants of the withdrawal discomfort were determined by relating the value of the determi- nant to the intensity of withdrawal discomfort with Pearson's product-moment correlation coefficients, t tests, or one-way analy- ses of variance. The clinical significance of tobacco withdrawal was determined by ax2 test, which compared the incidence of absti- nence during the last two days of the study between subjects who had severe withdrawal discomfort on the first two days of absti- nence and subjects who had little withdrawal discomfort on the first two days of abstinence. The analyses were calculated with the Statistical Packages for the Social Sc¢ences." " RESULTS Subject Characteristics Of the 142 applicants screened, 42 were rejected, most because they smoked low-nicotine cigarettes (<0.5 mg per cigarette). The education and socioeconomic class of the 50 subjects who were accepted and assigned to receive placebo appeared to be higher than those of the national samples; however, their smoking habits were similar to those.of the national samples (Table 1). None of the subjects-&opped out during the study. During the abstinence period, 18 subjects reported smoking at least one cigarette or had high carbon monoxide values (ie, >10 ppm). These 18 subjects smoked only a few cigarettes during the four days of abstinence (median, 2.5; range, one to eight). The intensity and frequency of withdrawal signs and symptoms in these subjects did not differ from those of the 32 complete abstainers; thus, the data of the incomplete abstainers were included in the following analyses. Reliability The large majority of the variables were stable during baseline (Figure). The only two variables that changed,during baseline, ie, lacked intrarater reliability, were the number of awakenings and minutes awake, both of which increased from the first to second day of baseline (P<.05). Thirty-one of the 50 subjects located an observer who felt confident in his or her rating. Observer ratings of irritability, anxiety, restlessness, drowsiness, fatigue, impatience, and so- matic complaints were all significantly related to their respective subject ratings (intrarater r=.40 to .62; P<.05). Validity Supine heart rate and orthostatic response decreased after cessation (Table 2). Weight had not increased after four days of abstinence. Among the self-reported DSM-III symptoms, craving for tobacco, irritability, anxiety, difficulty concentrating, and restlessness increased after cessation. The DSM-III symptoms of headaches, drowsiness, and GI tract symptoms did not increase after cessation. Self-reported impatience, somatic complaints, hunger, and eating also increased. Alcohol intake decreased, but caffeine intake did not change. Sleep ratings and total sleep time decreased, and number of awakenings increased after cessation. The observer reports confirmed both the positive and negative results from subject ratings; ie, observer ratings of irritability, anxiety, difficulty concentrating, restlessness, impatience, and somatic complaints increased after cessation, but ratings of drows- iness and fatigue did not change (Table 3). Most of the results from the POMS also confirmed the results from subject ratings. Scores on the anger, tension, and confusion scales increased after cessation. These results are consistent with subject ratings of increases in irritability, anxiety, and difficulty concentrating. Conversely, the vigor score on the POMS decreased after cessation. This result is inconsistent with subject ratings of no change in drowsiness. Magnitude Supine heart rate decreased ten beats per minute and the orthostatic response three beats per minute after cessation. Most of the observer and subject ratings of withdrawal symptoms increased one half to one point of a possible three points (Tables 2 and 3). Except for cigarette craving and anxiety, most of the symptoms were rated mild during abstinence. Alcohol intake decreased by less than one drink per day, and time asleep de- creased about 30 minutes. Incidence The incidence of symptoms was similar for subject and observer ratings. The most common signs and symptoms were decreased heart rate, increased weight, and increased irritability, anxiety, difficulty concentrating, restlessness, impatience, and insomnia (Table 4). Of the subjects, 78% reported four or more DSM-1II criteria, and thus tobacco withdrawal would have been diagnosed. Total Withdrawal Discomfort The difference scores of nine subject ratings were summed to produce a reported withdrawal discomfort score: craving for tobacco, irritability, anxiety, difficulty concentrating, restless- ness, impatience, hunger, somatic complaints, and insomnia. These symptoms were chosen because they appeared valid and reliable, were nonredundant, and were rated on a similar scale. The number and intensity of these symptoms varied greatly across subjects. The mean (=SD) number of symptoms that increased after abstinence was 5.6 ~ 2.1. The mean increase in total intensity was 3.8 ± 3.5 of a possible 27 points (ie, nine symptoms times three points per symptom). Difference scores in observer ratings of irritability, anxiety, restlessness, and impatience were summed to produce an observed withdrawal discomfort score. The number and intensity of ob- served symptoms also varied widely across subjects. The mean number of observed symptoms was 2.8=1.2, and the mean increase in total intensity was 2.1±2.0 of a possible 12 points. Arch Gen Psychiatry-Vol 43, March 1986 Tobacco Withdrawal-Hughes & Hatsukarni 291 ~

I I I I I I I I I I I I I I I I Table 3.-_ Mean (SE) for Observer Ratings and Scores on the Profile of Mood States Baseline Abstinence t Observer ratings (n=31) Irritability 0.4 (0.1) 0.9 (0.1) 5.0* Anwety 0.6 (0.1) 1.1 (0.1) 4.0* Restlessness 0.4 (0.1) 0.9 (0.1) 5.0* Drowsiness 0.3 (0.1) 0.3 (0.1) NS Fatigue 0.3 (0.1) 0.4 (0.1) NS Impatience 0.5 (0.1) 1.0 (0.1) 4.3* Somaticcomplaints 0.2 (0.1) 0.4 (0.1) 2.8t Profile of Mood States (n = 50) Vigor 16.1 (0.7) 12.9 (0.7) -4.5* Anger 5.0 (0.7) 6.5 (0.8) 2.5t Tension 8.4 (0.7) 12.0 (0.8) 5.5* Fatigue 7.4 (0.7) 6.1 (0.7) NS Depression 6.2 (1.1) 5.0 (0.9) NS Confusion 5.4 (0.5) 6.7 (0.6) 2.6t Total mood disturbanoe 14.0 (3.1) 23.4 (3.4) 3.7* *P<.001. tP<.05. Time Course The signs and symptoms occurred immediately after cessation and persisted for the four days of abstinence studied (Figure). Only anxiety tended to decrease during the study (P=.08). Determinants Subjects who reported more withdrawal discomfort smoked fewer cigarettes per day, had tried to quit more often, had more intense withdrawal symptoms on their last attempt to quit, smoked soon after arising, rated their first cigarette as the most difficult to do without, and weremore tolerant to nicotine (Table 5). In addition, subjects who believed they had received placebo reported more discomfort than those who were uncertain which drug they had received, who in turn reported more discomfort than subjects who believed they received nicotine gum. Only the number of prior cessations and the degree of tolerance were also related to observer-rated withdrawal discomfort. Clinical Significance Subjects who had intense withdrawal discomfort on the first two days of abstinence were not less likely to abstain completely from smoking on the last two days of abstinence than subjects who had little withdrawal discomfort on the first two days of abstinence. COMMENT Validity of Tobacco Withdrawal A major purpose of this study was to test the validity of DSM-III and other symptoms of tobacco withdrawal in smokers who quit smoking in their natural environment. In the study, the DSM-II1 symptoms of craving for tobacco, irritability, anxiety, difficulty concentrating, and restless- ness increased after cessation. These symptoms are com- mon in surveys of tobacco withdrawal and have been vali- dated in several laboratory studies.`-' Conversely, the DSM-III symptoms of headache, drowsiness, and GI tract disturbance did not increase after cessation. These symp- toms are uncommon in survey studies and have not been well validated.'-' Several signs and symptoms not included in DSM-III occurred after cessation: decreased heart rate, increased eating, increased sleep disturbance, and decreased alcohol intake. Decreased heart rate and increased eating are two of the most prominent and consistent changes observed in prior studies of tobacco withdrawal.'-' Worse sleep was reported by many subjects in the present study; however, sleep laboratory studies report improved sleep after cessa- tion." These contradictory results may be due to the limited Table 4.-Incidence of Signs and Symptoms of Tobacco Wrthdrawal* Si9rvSymptom Reported Observed Physiologic signs Supine heart rate (decrease) 25 (79) Orthostatic response (decrease) 21 (68) WeigM 23 (73) DSM-lll symptoms Craving for tobacco 31 (62) Irritability 40 (80) 22 (72) Anxiety 43 (87) 24 (77) D4ffic:utty concentrating 36 (73) ... Flestlessness 35 (71) 21 (68) Headaches 12 (24) Drowsiness 11 (22) 8 (25) Gastrointestinal tract problems 16 (33) Other symptoms Fatigue 20 (40) 7 (22) Impatience 38 (76) 24 (78) Somatic complaints 24 (49) 7 (23) Eating/drinking Eating 28 (56) Hunger 26 (53) Alcohol intake (decrease) 24 (48) Caffeine intake (decrease) 20 (41) Sleep Adequacy (decrease) 42 (84) Latency (decrease) 25 (50) No. of awakenings 24 (48) Time awake 27 (55) Time asleep (decrease) 31 (62) *Percentage of subjects in whom the signs and symptoms increased unless noted otherwise (N-50 for physiologic signs and reported symptoms; N=31 for observed symptoms). generalizability of sleep laboratory studies or to the de- creased validity of our subjects' self-reports. Subjects in the present study reported decreased alcohol intake, yet sub- jects in the only other study of alcohol intake after smoking cessation reported increased alcohol intake.' We are un- aware of observational studies that might clarify these contradictory results. In summary, we believe the valid signs and symptoms of tobacco withdrawal are craving for tobacco, irritability, anxiety, difficulty concentrating, restlessness, bradycar- dia, and increased eating. Reliability, Magnitude, and Time Course Two findings suggest the tobacco withdrawal syndrome possesses both interrater and intrarater reliability. First, the present study demonstrated that self-reported symp- toms of tobacco withdrawal were observed by others. Second, an earlier study by our group showed that signs and symptoms of tobacco withdrawal were consistent across repeated cessations.n The magnitude of discomfort from tobacco withdrawal has not been previously estimated because prior studies have not used standardized psychological tests. In the pres- ent study, most subjects experienced five or more symp- toms and rated most symptoms as mild. More importantly, the mean scores on the POMS during abstinence (Table 3) were in the 40th to 50th percentile of norms for psychiatxic outpatients.° Thus, during abstinence the subjects were as distressed as the average psychiatric outpatient. The time course and duration of tobacco withdrawal have not been well studied. In prior studies the onset of with- Tnt-- -n wthdraw I-Hughes & H tsukarsi

I I I I I I I I I I I I I I I I I I Table 5.-Subject Characteristics Associated With lntensity of Wrthdrawa! Discomtort* ` Charaeteristie Wfthdrawal Discomfort Subject Observer Rated Rated (N=50) (N=31) Dernographics Age NS Sax NS present smoking habit Cigarettesiday r= -.29* NS Nicotine yield/cigarette NS Depth of inhalation NS past smoking history Age at onset NS Duration NS ... No. of prior cessations r=.37t r=.29* Severfty of past withdrawal r=.28* NS Addiction scales Fagerstrom Latency to first cigarette t= - 2.25* NS Am vs wa smoking NS Am cigarette best t-2.71* NS Difficulty refraining NS Smoke when ill NS Total score NS MacAndrews NS Reasons for smoking Addiction NS Negative affect NS Belief in nicotine vs placebo F=3.83* NS Nicotine levels PrecSgarette NS Increase while smoking NS Tolerance - rs-.33* r=-.28* *P<.05. tP<.01. drawal symptoms is within hours of cessation,z and symp- toms appear to have a diurnal course, with worsening during the evening.Z In earlier studies'~;'-` and in the present study, symptoms appeared to decrease even over the first four to five days of abstinence. The fate of symp- toms after this period is debatable. One study reported a continuing decline in severity,ffi while a second reported an increase in some symptoms in the second week.° In other studies, the total duration of symptoms varied widely both within and between studies (eg, from three days to five years).Z Variability Across Subjects The number and intensity of withdrawal symptoms in this study varied greatly across subjects. Several factors have been hypothesized to explain this variability. One factor often cited is that smokers differ in their degree of depen- dence." Several recent studies have attempted to discover markers of nicotine dependence. This search has been motivated in part by recent findings that smokers who appear to be dependent benefit from nicotine gum treat- ment much more than smokers who do not appear to be dependent."` In the present study, several proposed mark- ers of dependence-multiple failures to stop smoking, severe withdrawal symptoms in the past, smoking soon after awakening, and rating the first cigarette as the most difficult to do without-were 'reiated to-increased with- drawal discomfort. Conversely, th ome st commonly used marker of dependence, ie, the number of cigarettes per day, was negatively related to tobacco withdrawal. Although early reports found that heavy smokers had more with- drawal discomfort than light smokers,'-' several other studies failed to replicate this finding. One possible expla- nation for this failure to replicate is that light smokers are actually the more dependent smokers who have decreased smoking as an alternative to the more difficult option of cessation. Another possible explanation for the failure to replicate is that the number of cigarettes per day is a poor measure of tobacco or nicotine intake.I Consistent with this hypothesis is the report in recent studies that serum cotinine level (a metabolite of nicotine) predicts withdrawal discomfort'; however, in the present study serum nicotine values did not predict withdrawal. Another factor commonly cited to explain the variability in withdrawal discomfort is expectancy. A priori expecta- tions' and "abstinence phobiass26 are said to have a large effect on the withdrawal symptoms from tobacco, opiates, and other drugs. In the present study, subjects who be- lieved they received placebo reported the most intense symptoms, while subjects who believed they received nico- tine gum reported the least intense symptoms. This finding may have been an example of expectancy effects in that subjects who believe they are receiving placebo probably expect to have severe withdrawal symptoms. Gender is another factor thought to influence withdrawal discomfort. Some studies report that women have more withdrawal symptoms from smoking cessation than men.z This possibility is interesting because it may explain why women have less success at smoking cessation. However, other studies, including this one, found no differences in overall withdrawal discomfort between men and women.'-' Cause of Tobacco Withdrawal The mechanism by which abstinence from tobacco in- duces withdrawal symptoms is unknown. One hypothesis is that withdrawal phenomena are due to pharmacologic de- pendence on nicotine similar to dependence of the sedative/ opiate type.' Although the present study did not directly test this hypothesis, two results from the study suggest the hypothesis may be correct. First, in both the present study and a previous study," subjects who were more tolerant had more withdrawal discomfort. This result is consistent with classic theories that tolerance and withdrawal develop concomitantly because they are both manifestations of the development of physiologic dependence.a Second, with- drawal effects were initially the opposite of nicotine effects (eg, increased irritability) and then decreased toward base- line. This "rebound" or "overshoot" phenomenon is con- sistent with dependence of the opiate/sedative type. A third result that is consistent with the nicotine dependence hypothesis comes from several studies of the effect of nicotine on tobacco withdrawal. All of these studies found that nicotine relieves tobacco withdrawal." Another mechanism to account for tobacco withdrawal is that withdrawal symptoms represent a simple behavioral reaction to the loss of a reinforcer.' For example, when animals are deprived of a reinforcer (eg, food) they become irritable, restless, and aggressive and increase their eating and oral behaviors.a These behavioral reactions are quite similar to those observed in tobacco withdrawal. Thus, some of the symptoms of tobacco withdrawal may be due to frustration rather than nicotine deprivation. Significance of Tobacco Withdrawal The clinical significance of tobacco withdrawal is also unknown.L' The importance of withdrawal in maintaining other types of drug use and in preventing cessation of drug use has been widely debated. In the present study, the Arch Gen Psychiatry-Vol 43, March 1986 Tobacco Withdrawal-Hughes & Hatsukami 293 1

I I I I I I I I I I I I I I I I I I I intensity of wit~idrawal symptoms was not related to the ability to refraiff completely from smdking immediately after cessation. Although it is unlikely that withdrawal symptoms would fail to influence short-term abstinence but would influence long-term abstinence, a prospective study of the influence of withdrawal symptoms on long-term smoking cessation would have been a much better test than the present study. Although several retrospective studies have found that successful quitters reported less with- drawal symptoms than unsuccessful quitters,z we are un- aware of any large, prospective trial of the effects of withdrawal on cessation success. The fact that such a study has never been published is a glaring gap in the tobacco withdrawal literature. Validity and Generalizabillty of This Study We believe the validity of our results is greater than that of prior studies for several reasons. First, the study used smokers who wished to quit, a true placebo therapy, measures of withdrawal that were shown to be reliable and had been previously validated,3 and a sample with no attrition. These procedures limited selection, memory, and rationalization biases. Second, the study measured the signs and symptoms both during ad lib smoking and absti- nence; thus, subjects served as their own controls. Third, both the positive and negative results of the study were consistent across several measures: subject ratings, ob- server ratings, and standardized POMS scores. We also believe the results of our study are more gener- alizable than those of prior retrospective surveys or labora- tory experiments. The major reason we believe this is that, to our knowledge, the present study is the first to document observable changes in the behavior of smokers when they quit smoking in their natural environment. Another reason to believe our results are generalizable is that smoking habits of the subjects in the study were quite similar to those of population-based samples of smokers in the United States. Conversely, the generalizability of our results can be questioned because our subjects were volunteers for a nicotine gum study. Such volunteers may have been likely to attribute their smoking to dependence on nicotine and have had a bias to report more severe withdrawal symptoms. Contrary to this hypothesis is the finding that dependent smokers are more influenced by expectancy effects than nondependent smokers'; thus, dependent subjects may be especially influenced by placebo effects and thus report less severe withdrawal symptoms. CONCLUSION In summary, the present study clarified several aspects of tobacco withdrawal, ie, the study defined several valid and reliable symptoms of tobacco withdrawal, demonstrated that tobacco withdrawal produces a significant amount of observable distress, and associated tobacco withdrawal with tolerance to nicotine. In addition, the study reported results on the time course and duration of tobacco with- drawal, the cause of its variability across subjects, and whether withdrawal signs and symptoms influence the outcome of smokers' attempts to quit. We believe further studies on tobacco withdrawal will help determine why smokers find it so difficult to stop. Further, we believe that the study of tobacco withdrawal may be an excellent model of how behavioral and phar- macologic processes interact to induce a mental disorder. This study was supported by a grant from Merrell Dow Pharmaceuticals, Cincinnati, grant DA-0298-01 from the National Institute on Drug Abuse, Bethesda, Md, and funds from the state of Minnesota for psychiatric research. Roy Pickens, PhD, Dean Krahn, MD, Shlomo Malin, MS, Alice Lnlmic, and Saul Shigman, PhD, provided assistance. 294 Arch Gert Psychiatry-Vor 43, March 1986 References 1. Hatsukami D, Hughes JR, Pickens$W: Characterization of toba~ abstinence: Physiological and subjective effects, in Grabowski J, Hall (eds): Pharmacological Adjunett in the Treatment of Tbbacco DependencR; Bethesda, Md, NIDA Research Monograph 53, 19&5, pp 56-67. ~ 2. Shiffman SM: The tobacco withdrawal syndrome, in Krasnegor Nsi (ed): Cigarette Smoking as a Dependence Process• Bethesda, Md, NIDa.? Research Monograph 23, 1979, pp 158-185. 3. American Psychiatric Association Committee on Nomenclature and Statistics: Diagnostic and Statistical Manual qf Mentat Diaorders. ed 3. Washington, DC, American Psychiatric Association, 1980, pp 159-16Q, 176-178. 4. Hatsukami DK, Hughes JR, Pickens RW, Svikis D: Tobacco with- drawal symptoms: An experimental analysis. Psychopharmacology 1984; 84:231-236. 5. Gilbert RM, Pope MA: Early effects of quitting smoking. Paychophan macology 1982;78:121-127. 6. Hughes JR, Hatsukami DK, Pickens RW, Krahn D, Malin S, Luknic A: Effect of nicotine on the tobacco withdrawal syndrome. Psychopharma. cology 1984;83:82-87. 7. Fagerstrom K-0: Measuring degree of physical dependence to tobacco smoking with reference to individualization of treatment. Addict Beher 1978;3:235-241. 8. MacAndrews C: On the possibility of the psychometric detection o[ persons who are prone to the abuse of alcohol and other substances. Addid Behav 1979;4:11-20. 9. Ikard FF, Green DE, Horn D: A scale to differentiate between types of smoking as related to the management of affect. lnt J Addict 1969;4:649-659. 10. Langone J, Gji7ca HB, van Vunalda H: Nicotine and its metabolites: Radioimmunoassay for. nicotine and cotinine. Biochemiatry 1973;12 5025-5oso. U. Hughes JR, Frederiksen LW, Frazier M: A carbon monoxide analyzer for measurement of smoking behavior. Behav Ther 1978;9:293-296. 12. Carskadon MA, Dement WC, Mitler MM, Guilleminault MD, Zarcone VD, Spiegel R: Self-report vs sleep lab findings in 122 drug free subjects with complaints of chronic insomnia. Am J Psychiatry 1976;133:82-88. 13. McNair DM, Lorr M, Droppleman LF: Manual for the Profile of afood States. San Diego, Educational and Industrial'Iesting Service, 1971. 14. Adult Use of Tobacco-1975. US Dept of Health, Education, and Welfare, 1976. 15. Health Conaeque+ecea of Smoking-The Changing Cigarette. US Dept of Health and Human Services, 1981. 16. Smoking and Health-A lierport of the Surgeon General. US Dept at Health and Human Services, 1979, Appendix Al-A29. 17. Siegel S: Nonparametric Statistiea for the Behavioral Sciencea New York, McGraw-Hill Book Co, 1956. 18. Nie NH, Hall CH, Jenkins JG, Steinbrenner K, Brent DH: Statistical Paekagea for the Social Sciences. New York, McGraw-Hill Book Co, 1975. 19. Soldatos CR, Kales JD, Scharf MB, Bixler MB, Kales A: Cigarette smoking associated with changes in sleep difficulty. Science 1980;207: 551-553. 20. Gandet FJ, Hugli WC: Concomitant habit changes associated with changes in smoking habits: A pilot study. Med Timea 1969;97:195•205. 21. Hughes JR, Hatsukami DK, Pickens RW, Svikis D: Consistency of the tobacco withdrawal syndrome. Addict Behav 1984;9:409-412. 22. Wynder EL, Kaufman PL, Lesser RL: A short term follow-up study on ex-cigarette smokers. Am Rea Respir Dis 1967;96:645-655. 23. Shiffman SM, Jarvik ME: Smoking withdrawal symptoms in two weeks of abstinence. Psychopharmacology 1976;50:35-39. 24. Hughes JR: Detning the dependent smoker. Validity and clinical utility. Behav Med Abst 1985;5:202-204. 25. Herning RI, Jones RT, Benowitz NL, Mines AH: How a cigarette is smoked determines blood nicotine levels. Clin Pharmacol Ther 1983;31: 84-90. 26. Pomerleau 0, Fertig J, Shanahans S: Nicotine dependence in ags- rette smoking: An empirically based multivariate modeL Pharmacol Bio- chem Behav 1983;19:291-299. 27. Gritz ER: Smoking behavior and tobacco abuse, in Mello NK (ed): Advances in Substance Abuse. Greenwich, Conn, JAI Press Inc,1980, vo11, pp 91-158. 28. Hall SM: The abstinence phobia, in Krasnegor N (ed): Behavioral Treatment and Analyses of Substance Abuse. Bethesda, Md, NIDA Re- search Monograph 25, 1979, pp 5b-67. 29. Kslant H, LeBlanc AE, Gibbins RJ:lblerance to, and dependence on, some nonopiate psychotropic drugs. Pharmacol Rev 197123:135-191. 30. Hughes JR, Hatsukami DK: Short-term effects of nicotine gum, in t~ COD ~ Grabowski J, Hall S(eda): Pharmacological Adjuncts in the Treatment 4f Tbbacco Dependence. Bethesda, Md, NIDA Research Monograph 53,1985, pp 68-82. ~ 31. Falk JL: The origins and functions of adjunctive behavior. Ani+rt ~ Learx Behav 1977;5:325-335. - 32. Fagerstrom K-0, Storm HG:. The effects of instructions on smoking cessation for individuals with different degrees of dependence. Behaa ~ Psychother 1981;9:310-315. C.C Tobacco Withdrawal-Huahes & Hatsukaml