Tobacco Documents Online

I I I I I I I I I I I I I I I I I Cigarette Craving, Smoking Withdrawal, and Clonidine Abstract. Clonidine, an a-2-adrenergic agonist, significantly reduces opiate withdrawal. Fifteen heavy smokers abstained from cigarettes on three separate occasions and received instead clonidine, placebo, or the benzodiazepine alprazo- lam. Clonidine and alprazolam diminished withdrawal symptoms. The two drugs suppressed anxiety, tension, irritability, and restlessness equally but clonidine had a greater effect than alprazolam on cigarette craving. These observations suggest that noradrenergic activity is a common feature in the pathophysiology of withdrawal and that a special relationship exists between central noradrenergic activity and craving. The a-2-noradrenergic agonist cloni- dine . diminishes the opiate withdrawal syndrome in chronically addicted human subjects (1). Central noradrenergic func- tion had long been implicated in the action of opiates, but the anatomical locus for that interaction remained un- known. In the 1970's, evidence began to accumulate that a major anatomical con- nection between the adrenergic and opi- ate systems existed in the locus coeru- leus. This nucleus accounts for nearly half of the noradrenergic neurons and produces the majority of norepinephrine in the mammalian brain. Its noradrener- gic cells are densely populated with in- hibitory opiate receptors. Enkephalins and opiates as well as a-2-noradrenergic agonists decsease the firing rate of these cells, and abrupt opiate withdrawal re- sults in a marked increase in this firing rate (2). Extensive data have now accu- mulated from both experimental animals (3) and man (4) confirming Gold's origi- nal observation (1), and animal data sup- port the assertion that this diminished withdrawal behavior is related to dimin- ished noradrenergic activity (5). We asked whether this concept of nor- adrenergic involvement in opiate with- drawal could be extended to appetitive behaviors such as smoking. We now report that clonidine alters the acute withdrawal syndrome associated with cigarette smoking and suggest that cen- tral adrenergic overactivity is a common feature in the pathophysiology of with- drawal syndromes seen with a variety of addictive substances, including ciga- rettes, alcohol, and opiates. Volunteers smoking more than 30 cig- arettes per day for at least I year, were recruited to participate in a double-blind crossover study of the effects of cloni- dine on the acute smoking withdrawal syndrome. In addition to a placebo con- trol, a benzodiazepine-like drug, alpra- zolam, was used in a second experimen- tal condition, Alprazolam has been shown to be equally anxiolytic and slightly less sedative than diazepam (6). All subjects were in good health and were drug-free, except for two female volunteers who used medication for birth control. All subjects were instructed to refrain from smoking for 24 hours on three separate occasions. On each occa- sion they were told not to smoke after going to bed and to report without smok- ing at 0830 the next morning. Baseline pulse, blood pressure, and psychological measures were obtained; then one of three treatment regimens was begun. Subjects received clonidine (0.2 mg), al- prazolam (1.0 mg), aqd placebo in one of three randomly assigned sequences. All treatments were given in two divided doses with the second dose given 90 minutes after the first. Pulse and blood pressure, including orthostatic blood pressure, were measured every 90 min- utes. At the same time, subjects com- pleted a series of nine visual analog scales. These scales used a 10-cm line to assess tension, anxiety, irritability, crav- ing (thoughts about or wish to smoke), restlessness, impaired concentration, and sadness (or tearfulness). Subjects also completed similar scales for drowsi- ness and dizziness. At the end of each experimental day, they made a global assessment (on a scale of I to 10) of the I 864 SCIENCE, VOL. 226

Table 1. Mean (± S.E.M.) hour ratings (from an arbitrary analog scale) during hours 2 to 7. Data from hour I are excluded because previous blood pressure studies show no drug effects until hour 2 (16). There were no significant differences across days at baseline. ~ i ~ ~ 1 ~ ~ ~ ~ ~ ~ ~ ~ I ~ ~ ~ ~ l S Treatment Analysis of variance Friedman ca e Clonidine Alprazolam Placebo F P FTS P Anxiety 1.46 t 0.44 1.54±0.46 2.94±0.80 4.53 0.024 4.14 0.12 Irritability 2.20 ± 0.52 2.04t0.77 3.91 ± 0.77 5.34 0.014 7.09 0.03 Craving 3.24 ± 0.67 4.97 ± 0.62 6.03 ± 0.69 9.77 0.001 8.91 0.01 Restlessness 1.60 ± 0,38 1.63 ± 0.37 3.09 ± 0.76 3.69 0.043 4.14 0.12 Concentration 2.17 ± 0.66 1.72 ± 0.52 2.27 ± 0.68 0.05 0.948 0.59 0.74 Sad-tearful 0.83 ± 0.40 0.83 ± 0.30 0.72 ± 0.37 0.04 0.957 3.32 0.19 Tension 1.51 ± 0.35 1.51 ± 0.41 3.12 ± 0.73 5.09 0.016 8.77 0.01 Drowsy 5.06 :t 0.60 4.41 ± 0.70 1.13 ± 0.42 13.10 0.001 13.64 0.001 Dizzy 2.58 ± 0.59 2.04t0.50 1.15 ± 0.35 8.82 0.006 7.95 0.02 degree to which the experimental treat- ment had helped them to do without cigarettes. To avoid crossover effects, we asked subjects to resume their normal smoking pattern after each experimental day. A minimum of three normal smoking days separated experimental days. After ses- sion 3, subjects were offered a 3-week trial with the experimental treatment of their choice. Global ratings of the difficulty in not smoking under each of the three experi- mental conditions were compared by analysis of variance with a randomized block design. The results were further analyzed pairwise by post hoc Studen- tized range tests. This same procedure was used to compare treatment effects obtained on the nine individual analog scales after averaging the scores during Pressure to smoke 10 0 0 8 o 0 9 0 8 0 0 0 0 0 0 a 8 Co c ~ 0 ~ a 0 8 0 4 0 0 ~ ~ 2 0 0 8 8 co 0 0 Clonidine Placebo Alprazolam Fig. I. Individual ratings, mean, and standard errors of the global difficulty in not smoking obtained from subjects after completing all three treatments. 16 NOVEMBER 1984 the period from hour 2 to hour 7 after the first doses. A comparison of linear trends on one specific scale, craving, was also accomplished by the same pro- cedure. Because it was not clear that these scale scores represented normally distributed measurements, they were converted to rankings and the more con- servative Friedman nonparametric two- way analysis of variance performed. Fifteen subjects, 2 men and 13 women, completed all three experimental days. They averaged 34 years of age, smoked an average of almost two packages per day, and had been smoking for an aver- age of 16 years. Thirteen subjects clearly preferred one of the drug conditions over placebo during the acute phase of smok- ing withdrawal. Of the two subjects who did not prefer active medication, one showed only modest withdrawal symp- toms in the placebo condition and could not distinguish between the drug and placebo conditions, in spite of a history of smoking one and a half packages of cigarettes per day. The other subject who did not prefer drug treatment mani- fested marked withdrawal symptoms during the placebo period and experi- enced no relief from either drug. This subject also experienced very little pulse or blood pressure effect from clonidine. Clonidine [Q(28) = 6.35, P < 0.001] and alprazolam [Q(28) = 3.76, P < 0.05] sig- nificantly reduced the difficulty in not smoking (Fig. 1). Of the 13 subjects who found drug treatment effective, 10 pre- ferred clonidine (binomial calculation, P = 0.046). The analyses of the visual analog scores are presented in Table 1. In spite of the subjects' tendency to prefer cloni- dine over alprazolam on the global rat- ings, most individual items did not distin- guish between the two drug treatments. For anxiety, irritability, concentration, and tension, both drugs were clearly better than placebo and essentially iden- tical to each other. Even the two most common side effects, drowsiness and dizziness, occurred to the same degree. Only the scale measuring craving (think- ing about or wishing to smoke) reflected the strong tendency in the global ratings for subjects to prefer clonidine. Here clonidine was significantly more effec- tive than both placebo [Q(20) = 6.18, P < 0.01] and alprazolam [Q(20) = 3.83, P < 0.05]. Alprazolam was not signifi- cantly more effective than placebo. Previous studies of smoking withdraw- al syndrome have shown that craving is the most consistently observed with- drawal symptom (7) and that it tends to be least in the morning and to increase as the day progresses (8). Therefore, we plotted the hourly rating for craving dur- ing the first 7 hours of treatment (approx- imately 0830 to 1530). Craving among tol . 0 7 4 3 o+ 0 4 a Time (hours) Fig. 2. Mean [:t standard error of the mean (S.E.M.)] hourly ratings of craving by treatment. Trend analysis shows mean slope and S.E.M. Clonidine =-0.07 ± 0.13, alprazolam = 0.15 ± 0.11, and placebo = 0.41 ± 0.14. F(2, 26) = 6.68, P < 0.005 (mean square error, 0.1144). Post hoc test for clonidine versus placebo, Q(26) = 5.26, P<0.01. 865

I I I I I I I I I I I I I I I I I I subjects given placebo increased during the course of the day and, as a result, the contrast between clonidine and placebo increased as the day progressed (Fig. 2). Clonidine and alprazolam both re- duced the acute withdrawal syndrome following sudden abstinence in heavy cigarette smokers. However, subjects preferred clonidine, and clonidine had significantly more effect on cigarette craving than alprazolam. This result does not imply that clonidine is a cure for smoking. Even among those smokers who have been successfully withdrawn from cigarettes, a high percentage will return to cigarettes within the next 6 months (7, 9), Nevertheless, any drug that could enable heavy smokers to ab- stain quickly and relatively asymptomat- ically is potentially clinically useful. Our clinical experience suggests that cloni- dine continues to be effective over sever- al weeks; subjects were generally not maintained for a longer period of time. Our experience with alprazolam beyond the brief experimental period was limited and not encouraging. Several studies have been made of the use of clonidine in the treatment of alco- hol withdrawal (10). These studies were initially undertaken in an effort to control hypertension associated with alcoholic withdrawal, but clonidine also seems to suppress symptoms of alcohol withdraw- al. The efficacy of clonidine in ameliorat- ing cigarette, alcohol, and opiate with- drawal and clonidine's powerful inhibi- tion of noradrenergic activity, together with the evidence from animals of norad- renergic hyperactivity in opiate with- drawal, suggest noradrenergic over-ac- tivity as a common characteristic in the pathophysiology of withdrawal syn- dromes. It seems reasonable to ask what clini- cal characteristics these withdrawal syn- dromes share and to consider how simi- larities might relate to the efficacy of a drug that decreases noradrenergic activi- ty. Certainly, the delirium tremens and hallucinations of end-stage alcoholic withdrawal differ from the agitation, muscle cramps, and retching of opiate withdrawal and from the tension and irritability of cigarette withdrawal. Be- fore these end-stage characteristics come to dominate the clinical picture, however, these syndromes share certain characteristics. Wikler suggested that the maintenance of opioid-taking behav- ior depends most on the intense craving for the drug associated with the onset of S withdrawal (11). Similarly, alcohol with- drawal always begins and usually con- sists not of delirium tremens but of "bad nerves" and the "urge" for a drink (11, 12). In our 15 abstinent subjects, as in other smokers who have been studied (7), the most consistent withdrawal symptom is craving, by which we mean a preoccupation with, thoughts about, or an urge for, the habituating substance, not necessarily associated with any physical distress. It is then clear that this craving is a common denominator across these habituations, and we suspect it plays an important role in maintaining the habituation. Although initial contact with ciga- rettes, alcohol, or opiates need not be for the purpose of regulating tension or dis- tress, chronic users come to learn that these substances can diminish such dys- phoria. They also come to learn that with long-term use the absence of these drugs produces dysphoria. It would be easy to assume that craving for a tension-reduc- ing drug develops in the absence of the drug because the addict experiences a rebound dysphoria and craves or seeks the drug to eliminate that dysphoria. But this is not our observation. Craving is the earliest, the most consistent, and the most severe symptom of cigarette with- drawal. It is also specifically responsive to clonidine. That is, clonidine and alpra- zolam both reduced anxiety, irritability, restlessness, and tension, and both pro- duced drowsiness; yet clonidine had sig- nificantly greater effect on craving. Be- cause clonidine had a greater effect on craving than alprazolam, because it se- lectively affects noradrenergic activity, and because this noradrenergic effect of clonidine is more powerful than that of the benzodiazepines (13), we suggest that there is a special relation between noradrenergic activity and craving. We hypothesize that the habituated person learns that modest increments in noradrenergic activity predict higher lev- els of activity (5) that are associated with withdrawal dysphoria. Thus, modest in- crements in noradrenergic function be- come signals to the addict to seek the substance he has come to expect will diminish his impending distress. This link between the central adrener- gic system and craving is additionally appealing because it explains the obser- vation that stress increases cigarette craving. That is, stress increases central noradrenergic activity (14). This height- ened noradrenergic activity, even in the absence of any dysphoric feelings, could result in an increasing urge to smoke. It is unnecessary to limit this craving mod- el to habituating substances. Behaviors that reduce tension and noradrenergic activity, such as binge eating (15), could be habituating because of a similarly learned craving for the behavior. ALEXANDER H. GLASSMAN WYNN K. JACKSON B. TIMOTHY WALSH STEVEN P. ROOSE Department of Clinical Psychopharmacology, New York State Psychiatric Institute, New York 10032, and Department of Psychiatry, College of Physicians and Surgeons, Columbia University New York 10032 BOB ROSENFELD Division of Biostatistics, Columbia University References I. M. S. Gold. D. E. Redmond, Jr., H. D. Kleber, Lancet 1978-II, 599 (1978). 2. D. G. Amaral and H. M. Sinnamon, Prog. Neurobiol. 9, 147 (1977); M. J. Kuhar, Fed. Proc. Fed. Am. Soc. Exp. Biol. 37, 153 (1978); G. K. Aghajanian, Nature (London) 276, 186 (1978). 3. L. F. Tseng, H. H. Loh, E. T. Wei, Eur. J. Pharmacol. 30; 93 (1975); D. R. Meyer and S. B. Sparber, Pharmacologist 18, 236 (1976); J. Ve- tulani and B. Bednarczyk, J. Pharm. Pharma- col. 29, 567 (1977): S. Fielding et al., Pharma- col. Exp. Ther. 207. 899 (1978). 4. T. W. Uhde, D. E. Redmond, Jr., H. D. Kleber, Psychiatry Res. 2. 37 (1980); M. S. Gold, A. C. Pottash, D. R. Sweeney. H. D. Kleber, J. Am. Med. Assoc. 243, 343 (1980); D. S. Charney and H. D. Kleber, Am. J. Psychiatry 137, 989 (1980); A. M. Washton and R. B. Resnick, ibid., p. 1121. 5. A. C. Swann et al., Eur. J. Pharmacol. 86, 167 (1983). 6. K. Rickels et al., Am. J. Psychiatry 140, 82 (1983). 7. S. M. Shiffman, in Cigarette Smoking as a Dependence Process. N. A. Krasnegor, Ed. [National Institute on Drug Abuse Research Monograph 23, HEW Publ. (ADM)79-800, Washington D.C., 1979), pp. 158-184. 8. T. W. Meade and N. J. Wald. Br. J. Preventive Soc. Med. 31, 25 (1977); R. M. Gilbert and M. A. Pope, Psychopharmacology 78, 121 (1982). 9. W. A. Hunt and D. A. Bespalec, J. Ctin. Psy- chol. 30, 431 (1974); W. A. Hunt and J. D. Matarazzo, J. Abnorm. Psychol. 81, 107 (1973). 10. S. E. Bjorkqvist, Acta Psychiatr. Scand. 52, 256 (1975): J. Walinder et al., Drug Alcohol Depend. 8. 345 (1981); A. Nardoni et al., Clin. Ter. 97, 619 (1981): A. J. Wilkins, J. A. Steiner, W. J. Jenkins, Clin. Sci. 64. 64P (1983). 11. A. Wikler, Br. J. Addict. 57, 73 (1961). 12. S. M. Wolfe and M. Victor, in Recent Advances in Studies of Alcoholism, N. K. Mello and J. H. Mendelson. Eds. (Government Printing Office, Washington D.C.. 1972), pp. 188-199: M. M. Gross, E. Lewis. J. Hastey. in The Biology of Alcoholism, B. Kissin and H. Begleiter, Eds. (Plenum, New York. 1974), pp. 191-263. 13. L. Isaac, J. Cardiovasc. Pharmacol. 2(SuppI. 1). S5-S19 (1980): S. J. Grant, Y. H. Huang, D. E. Redmond. Jr.. Life Sci. 27. 2231 (1980). 14, J. Korf. G. K. Aghajanian. R. H. Roth, Neuro- pharmacology 12.933 (1973). 15. S. L. Foote, G. Aston-Jones. F. E. Bloom, Proc. Natl. Acad. Sci. U.S.A. 77, 3033 (1980). 16. S. N. Anavekar, B. Jarrott. M. Toscano, W. J. Louis, Eur. J. Clin. Pharmacol. 23, 1(1982). 8 May 1984: accepted I August 1984 1 866 SCIENCE. VOL. 226