Tobacco Documents Online

I I I I I I I I I I I I I I a I I I PreTalence of Tobacco Dependence and Withdrawal John R. Hughes, M.ll., Steven W. Gust. Ph.D.. and TerrN• F. Pechacek, Ph.D. In a sample of1,0U6 middle-aged male smokers drawn jrom the general population, 90% `\'=905) julfilled DSM-III criteria and 36% (N=362) fulfilled Fagerstrom's crrteria ror tobacco dependence. Among the 875 who had stopped smoking in the past for at least 24 hours, 21 %(N=184) fulfilled DSM-III criteria and 46% (N=403) ful ftlled the authors' own criteria for tobacco withdrawal. Concordance of results among the criteria for diagnosing tobacco dependence and withdrawal was low. These results suggest that the DSM-III criteria for tobacco dependence are overinclusive and that there is little consensus among the definitions of tobacco dependence and withdrawal. (Am J Psychiatry 1987; 144:205-208) R esearchers in the National Institute of Mental Health's Epidemiologic Catchment Area Program reported that the 6-month prevalence of drug depen- dence in the United States is less than 10% (1, 2). However, this survey omitted two common substance use disorders: tobacco dependence and tobacco with- drawal. Tobacco dependence is defined by DSM-111 (pp. 176-178) as 1) continuous use of tobacco for at least 1 month plus 2) either unsuccessful attempts to abstain, development of tobacco withdrawal symp- toms, or continued use of tobacco despite having a tobacco-induced illness. Tobacco withdrawal is de- fined by DSM-III (pp. 159-160) as use of more than 10 cigarettes per day that contain at least 0.5 mg of nicotine and occurrence of at IPzst four of the follow- ing within 24 hours of stopping smoking: craving for tobacco, irritability, anxiety, difficulty concentrating, restlessness, headache, drowsiness, and gastrointesti- nal disturbances. Presented at the 47th annual meeting of the Committee on Problems of Drug Dependence. Baltimore, Md., June 1984. Re- ceived Jan. 8, 1986; revised May 5, 1986; accepted June 30. 1986. From the Departments of Psvchiatrv and Epidemiologv, University ot Minnesota, Minneapolis. Address reprint requests to Dr. Hughes, Behavioral Medicine Service. Department of Psvchiatry, Universitv of Vermont College of Medicme, Burlington, VT 05405. Supported by grants DA-02239, DA-02298, DA-03728, and DA-04066 and a Research Scientist Development Award (DA-00109) to Dr. Hughes trom the National [nsntate on Drug Abuse and by funds for psychiatric research from the State of .Minnesota. Copyright C 1987 American Psychiatric Association. Am J Psychiatry 144:2, February 1987 Thirtr-three percent of Americans smoke (3). The proportion of these smokers who are tobacco depen- dent or experience withdrawal is virtually unknown i4). The few previous studies of the prevalence of tobacco dependence and withdrawal used select pop- ulations (e.g., patients in smoking-cessation clinics) and unvalidated criteria (5); thus, the generalizability and validitv of their results can be questioned. Knowledge of the prevalence of tobacco dependence is important for at least three reasons. First, several recent advances in dependence-based treatments for smoking appear to be effective, e.g., nicotine gum (6), clonidine (7), and brand fading (8). Knowledge of the prevalence of tobacco dependence and withdrawal would help determine the proportion of smokers who might benefit from such treatments. Second, smokers who attend formal treatment programs are usually dependent on tobacco (9). Knowledge of the preva- lence of tobacco dependence would help anticipate the need for such smoking-cessation programs. Finally, indirect evidence suggests that the smokers who have quit aie the less dependent smokers (10). If this selection bias continues, then future populations of smokers should contain higher proportions of depen- dent smokers. Serial determinations of the prevalence of tobacco dependence and withdrawal are necessary to test this hypothesis. For these reasons, we surveyed the prevalence of tobacco dependence and withdrawal, defined accord- ing to accepted criteria, in a population-based sample. METHOD In 1974-1975, 116,980 households in defined cen- sus blocks in the Minneapolis-St. Paul metropolitan area were screened to find subjects to participate in the Multiple Risk Factor Intervention Trial (11). From these households, 30,401 men were screened, but only 703 men were entered into the trial. In 1980. we randomly sampled 5,000 of the households that con- tained ineligible men and found 1,184 men who were smokers and were willing to be in a study. After they gave informed consent, these men were randomly divided into four groups, each of which was given some combination of mailings, phone calls, and book- lets about smoking and brief annual clinic visits. At the 2-year follow-up, 1,006 men were located, and they completed the survey that was the basis for this report. 205

;c?if1(t!,M rt~l t1(!. t\l)\11711OR.?1~ 11 I I I I I I I I I I I I I I 1 I I At this r(>IIMt'-UP. h.y '. ~=t~`1 r tht' ,UnleLts hSJ stoppeci ~me~Kin1-1. ~tncr thr t~~:r~XnracC Or rn0,c Whc~ ,topped did not Jitter across the tour groups, data trom 111 5ublects were used in this anal~-s(s. To determine the prevalence ot tobacco dependence and ~~ lthdraw sl. %~c questl(~ns on ti;L: DSM-III criteria as well as the criteria proposed by Fagerstrom ; 1 2; tor tobacco dependence and our own criteria for tobacco withdrawal 13). We chose the latter two sets ot criteria because, unlike most depen- dence and withdra%val criteria, they have been well validated ;4, 12, 13). The Fagerstrom criteria tor tobacco dependence include 1) the number of ciga- rettes smoked per day, 2) the number of milligrams of nicotine in each cigarette, 3) whether the smoker inhales. 4) whether the subject smokes more in the morning than in the afternoon and evening, 5) whether the subject smokes upon arising, 6) whether the first cigarette of the day is the most desirable, 7) whether it is difficult to refrain temporarily from smoking, and 8) whether the smoker smokes when ill. Our own criteria for tobacco withdrawal require the presence of at least four of the following in a smoker who has stopped for more than 24 hours: the first five of the DSM-III withdrawal s,vmptoms, whic;-L we have already listed, plus increased appetite, impatience, somatic com- plaints (headache, dizziness, tremor, stomach or bowel problems), and insomnia. RESULTS The mean±SD age of the subjects was 51.1'_-6.4 years. Ninety percent (N=905) had completed high school, 58% (N=583) were professionals (occupa- tional classes I-[II) ' 14), and 46% (N=463) earned at least S30,000 a year. The mean-SD number of ciga- rettes they smoked per day was 28.0± 12.8, the mean- SD amount of nicotine per cigarette was 0.9-0.4 mg, and the mean=SD number of years they had smoked was 33.0±7.1. Forty-two percent (N=423) had tried to quit at least three times. These smoking habits are almost identical to those reported in previous surveys of middle-aged men (3) and, in fact, are quite similar to those obtained in survevs of all U.S. adults ;3). Ninety percent (N=905) of the smokers fulfilled DSM-111 criteria for tobacco dependence. Sixty-one percent (N=614) had made an unsuccessful attempt to stop smoking, 21% (N=211) experienced tobacco withdrawal symptoms, and 23% (N=231) continued to smoke despite physical illness caused by smoking. Thirty-six percent (N=362) of the smokers fulfilled Fagerstrom's criteria for tobacco dependence (see table 1). The most prevalent DSM-III criterion was unsuc- cessful attempts to quit. Inhaling and smoking on arising were the most prevalent Fagerstrom criteria. Eighty-seven percent (N=875) of the smokers re- ported that they had stopped smoking for at least 24 hours in the past. The mean=SD length of time since they had last quit for at least 24 hours was 67±86 TABLE 1. Prevaience of Tobacco Dependence According to the Fagerstrom Criteria in 1.006 Smokers i • < Score=() Score=! Score=2 .riterton ane Meth o1 ot Scornnc N °~ N . N % \.;,e1t,Cr ~;r , li;.trettC, •tnoi:eJ :iaY 3av ;.15=n, h-'3=1. -25=' 181 l5 3U1 .30 523 52 \fc nr nicnttne rer :ti_arerte u.y=u, u.y-1, = 1. ..1.2=' .i4.3 54 312 31 151 15 (nhale nevcr=u, ,umenmes= l, alwavs=21 20 2 30 3 956 95 Smoke more in a.m. than fl.m. no=(1. ves= 1 453 45 553 55 5moke upon snstnl: <30 minutes =U. >30 minutes=ll 332 33 674 67 `lost destrabie ~tc.uerte ,others=U. tirst one= 1, 885 88 121 12 Difficult to retrain temporarily (no=O, yes= t) 875 87 131 13 Smoke when ill (no=0, yes=1) 573 57 433 43 'A score of more than 7 is required for a diagnosis of tobacco dependence. months. The mean=SD duration of the last attempt to stop was 69± 139 days. Ninety-three percent (N=814) quit abruptly. Ninety-six percent (N=840) stated that their memory of withdrawal symptoms was accurate. The self-reported accuracy of withdrawal symptoms and the number of withdrawal symptoms were not associated with the length of time since they had last stopped or the duration of the last period of nonsmok- ing. Among those who had quit at some time, 21% (N=184) fulfilled DSM-111 criteria and 46% (N=403) fulfilled our criteria for tobacco withdrawal (see table 2). The corresponding rates for all smokers (quitters and nonquitters) were 18% (N=181) and 37% (N= 372). Craving for tobacco, restlessness, increased ap- petite, and impatience were the most common with- drawal symptoms. The DSM-111 definition of tobacco dependence was not concordant with the Fagerstrom definition of tobacco dependence or with either definition of to- bacco withdrawal (tc<.04) (15). The Fagerstrom defi- nition of tobacco dependence was modestly concor- dant with both definitions of tobacco withdrawal (K=.23 and .16 for DSM-111's and our definition, respectively. The DSM-111 definition of withdrawal and our definition of withdrawal were fairly concor- dant (K=.44). DISCUSSION The different definitions of tobacco dependence and withdrawal produced very different prevalence rates_ ~ and showed little concordance. One possible reason ~ for this lack of agreement is that one or more of the ~ definitions are invalid. We know of no tests of the.~ validity of the DSM-111 definition of tobacco depen- ~ dence. In addition, the fact that this definition classi- ~ fied 90% of the tobacco users in this study as depen- ~ F-i ~ 206 Am J Psychiatry 144:2, February 1987

1 1 I I I I I I I I I I I I iABLE 2. Prevafertee of Tooacca Jvithorawat Symptoms in 875 Smokers Who Had Zit Smokin¢ in the Past t rireril ~ ' ~ kI -111 ;',ed cciba~:,:c, ror a: iea,t sr%rr.ii xcrh. "mc Kr:i mctrr tnan I'l 1IL'1rc.^.C, i J.it l.icarette cc ntatned more tnan t+ ~ me ut n,;omtit .~t le.ist tour < t tnr tc,lluH int imin _'-1 iiuurl ,t stopping smokine •- I ~iu .a Craving tor tobacco h39 Irntabtlitv 324 Anxtetv 368 42 Dtfficuln, concentrating 193 „ Restlessness Headache 438 26 ;n 3 Drowsiness -0 8 Gastrointestinal disturbances ?6 3 authors' own criteria: at least four of the following ~ctrhtn 24 hours of stopping smoking Craving for tobacco 639 3 ; Irritability 324 Anxtetv 368 Dtfficultv concentrating 193 Restlessness 438 5 0 Increased appetite 420 48 Impatience 411 4 7 Somatic complaints 166 19 Insomnia '9 9 dent suggests that it is overinclusive and thus may lack diagnostic discriminabilitv. The only validity test of the DSM-111 definition of tobacco withdrawal that we know of found that three of the eight symptoms did not occur upon cissation of smoking (13). On the other hand, several controlled, prospective studies have tested the validity of the Fagerstrom definition of tobacco dependence and of our definition of tobacco withdrawal (e.g., 4, 6, 9, 10). Of the 11 rests of the Fagerstrom definition of dependence. most have found that it predicts tolerance, withdrawal, smoking to obtain nicotine, relapse, and, most impor- rantly, therapeutic response to nicotine gum ;-+). Like- wise, seven studies (e.g., 4, 13. 16) have found that our own criteria for tobacco withdrawal are reliable, valid, and sensitive to both abstinence from tobacco and relief by nicotine gum. For these reasons, we believe that the prevalence rates given by the Fagerstrom definition of tobacco dependence and our definition of tobacco withdrawal are valid and conservative first approximations, whereas those calculated from the DSM-!II criteria may not be valid. According to Fagerstrom's and our definitions, one- third of smokers are either behaviorally dependent or physically dependent on tobacco (i.e., experience with- drawal symptoms). Since one-third of Americans smoke, approximately one in 10 Americans is pres- ently tobacco dependent, according to our results. In comparison, no psychiatric diagnosis was present in more than 10% of the population in the Epidemiologic Catchment Area Program reports (1, 2). Our results must be considered preliminary, as the generalizabiliry and validity of our data can be ques- tioned. The generalizability is uncertain because we Am J Psychiatry 144:2, February 1987 ' It (,1IE'.S. ('l 5T. .1\D I'E.C.HAc.E.f: •ur~<<eu on" mtdcile-,iUeci male smokers. Although ~i;uat;:)n ,; c.ur prC~.1ICnCL ' raCc's t11 uthCr agC antj ~CX "'roups ts ccrtatnlv needed, previous studies ;e.g., 5. 13) nave round that az;c and sex have little or no effect on ;c)baucc~ Licpendence and wtthdrawal. 1-he %sliciitv w'c)ur r',ttmates of tobacco withdra«al ~an be questioned tor at least two reasons. First. our sata are based on retrospective self-reports. However, the large majortty ot our smokers stated that their memory was accurate. More important, the time since the subjects had last quit smoking was not associated with an increase or decrease in the number of with- drawai symptoms thev reported. Second, we decided to require that a smoker must have stopped smoking for at least 24 hours for withdrawal to be assessed. We made this rule to exclude halfhearted, temporary, or environmentally induced periods of abstinence. Also, many of the symptoms of withdrawal (e.g., insomnia) are not evident until several hours after cessation (5). However, our rule may have biased the results if it excluded smokers who had such severe withdrawal that they could not remain abstinent for 24 hours. We reasoned that, if anything, the rule would produce an underestimate of the prevalence of withdrawal. In conclusion, our data indicate that many, but not all, smokers are behaviorally or physically dependent on tobacco. This result suggests that a considerable minority (20%-50%) of smokers may benefit from formal treatment programs or dependence-based treat- ments. Our data also indicate that the DSM-III criteria for tobacco dependence are overinclusive and are not concordant with other measures of tobacco depen- dence and withdrawal. These results suggest that the DSM-111 criteria for tobacco dependence need to be reformulated. REFERENCES 1. Jtvers JK, Weissman Tischler GL, et al: Six-month prevalence of psychiatric disorders in three communities. Arch Gen Psvchtatrv 1984: 41:959-967 '_. Blazer D, George LK, Landerman R, et al: Psychiatric disorders: a ruraliurban comparison. Arch Gen Psychiatry 1985; 42:651- 656 3. Smoking and Health-A Report of the Surgeon General: DHEW Publication PHS ?9-5066. Washington, DC, US Gov- ernment Printing Office. 1979 4. Hughes JR: Identttication ot the dependent smoker: validity and clinical utility. Behavioral Medicine Abstracts 1984; 5:202-204 :. Shiffman S.I~t: The tobacco withdrawal syndrome, in Cigarette Smoking as a Dependence Process, NIDA Research Monograph 23: DHEW Publication ADM 79-800. Edited by Krasnegor NA. Washington. DC, US Government Printing O,ffice, 1979 6. Hughes JR. Miller SA: Nicotine gum to help stop smoking. JAMA 1984; 252:2855-2858 -. Glassman AH, Jackson WK, Walsh BT, et al: Cigarette craving, smoking withdrawal, and clonidine. Science 1984; 26:864-866 8. Prue DIM, Kraptl JE, Martin JE: Brand fading: the effects of gradual changes to low tar and nicotine cigarettes on smoking rate, carbon monoxide, and thiocyanate levels. Behavior Ther- apist 1981; 12:400-416 9. Russell MAH, Peto J, Patel UA: The dassification of smoking by factorial structure of motives. J Royal Statistical Society, Series A, 1974; 137:313-333 10. Pechacek TF: Modification of smoking behavior, in Smoking 207

I I I I i I I I I I I I I I I I I K(Sf: E• 1C IOk~ FOK .AtURT.AI 1 ; Y 1nd Hr.tltii- \ Rcnc~r; w ~. , eL:un t,c nrr. i!11 IF,\t. I'-rii- -inon I'H'~ ',7- \\ i., Pc . rrnrnent i'rint mc Om". I y-y I l. \lultiPlr Kuk Facror In;cn•. .: m Iri.ri Rrse.tr,h (,n ut,: .\lul- npie Ri~k I-.tctt:r Intenc,.: -.u: Ri., mortltuts re.ultti. I\NI \ I~•'. _ut l~h -i}-- hanLe. .rns Ii. 1'_. Fat;erstrom ;;-U: \teasunnu .:urc•r t I i a.u ie{%nuen,r tt) 1 h. tobacco dependence with rererence to indr.icluaiiwuon ot treac- ment. Addict Behav 1978; 3:'_33-2-41 I~. Hupneti f R. Hanukami I): ~iUns .tnci ~~ mptomti ur ;nha ccirhdr.t«a;. .\rch (,en P.~chi.ur% iaNh: -1;:289-294 14. Hullint_,neau Ali. Redlich F{: Social (:lass and Mental lllnec, \ Commumn ticudc. New YF,rk. /ohn Wiley & Sons, l9iss U)hen I: .-\ xemcrent ot aareement tctr nominal scales. Edura. n mal ana Psvcholc~xical ~teasurement lN6t); 2U:37--tb HuLhc, i R. H.irsukami DK. Pickens R\C'. er al: Effect ot niconnc rn the te bacco «•,rhd,awal +vndrome. PsvehopharmacoiuQ% Berlin 1984; K.3:82-87 1 208 Am J Psychietry 144:2, February 1987