Tobacco Documents Online

I I I I I I I I I I I I I I I I Symptoms of Tobacco Withdrawal A Replication and Extension John R. Hughes, MD; Steven W. Gu.t, PhD; Kelli Skoog; Robert M. Keenan, PhD; James W. Fenwick, PhD • Smak.rs (n -315) who wisMd to quit w.rs randomly as- siqned in a doub(fblind mannsr to groups using sithar nleotlns or Placubo 9um. S.M-nport.d and obs.rrsd s)rmptoms of tobao- co withdrawsl w.r. eoU.eb.d bifora c.ssatlon and at follow-Ups of I to 2 w.Ncs, I month, and e months. S.M-rsport.d andYor obsarvsd anW anz/.ty, craving, difftcufty concentrating, hur- Q.G impatl.nc., and nstl.ssn.as were t!» most prornin.nt symptoms of tobaeeo wtthdrawal. Thaa symptoms had return.d to pnesssat4on l.vsis by I month sxc.pt inenssed w.l9ht, hun- gar, and craving continued tor 6 months In many smoksrs. Nloo- tlns gum d.crsas.d most symptoms, Including erwin9 and hun• psr but not wsi9ht Abatinsnt snakars with mors Intsnse withdrawal wers not morr Ilkafy to rrlapse. Abstinent snaksrs who 9ainsd mora weight were Nss likely to rMapse. (Arch Gan PsyoAlatr)c 1991;4W2-59) The signs and symptoms of tobacco withdrawal listed in DSM-111•R are craving for nicotine, irritability, awdety, difficulty concentrating, restlessness, decreased heart rate, and increased appetite or weight gain. In addition, dysphoria, impatience, and insomnia appear to be valid symptoms.' These symptoms are observable, reliable, and ota magnitude that is clinically signi5cssat.' Furthermore, the results of laboratory studies are concordant with these symptoms; eg, cessation from tobacco inereasea aggressive responding,' is labeled as anxietyiike in animal studies,' impairs performance on cognitive tasks,' increaaes energy (caloric) intake,' in- creases the pleasantness of sweets,' increases false starts in vigilance tasks,' and increases rapid eye movement intensi- ty.' Most of the symptoms appear to be due to nicotine depri- vation as they are relieved by nicotine replacementi and occur with cessation of nicotine intake alone." Aaepud for pubiieatwn February 13,1990. From the Departments d Psyahistry, Prqe2wiov, Fami1F Pr.ctio. (Dr Hurfies), aad Medieat Biwa••u•+ic- (Dr Fen.ieh), Umrerd2r d Vermont. Busiinitms tbe National Iurtit+tL on Drug Abuse (Dr Guet), Betbe.da. M$ and the Departmeau a( P.yhiauy aed Prydwiop (lSs Skoogand Dr Kseaan~ Univernty ol Minnrota. Minneapoli.. Read befm. tb4 wmal meeting ol tAe Arocatioe for Bebarior Aaaky:r, Bo.LM Ha... Noreasb.r 14,19A7. Reprmt rpsrta to Depatment of Prrehiatr! Uni..ratr of Vermoot CoDep ol Y.didne, BurWqtoa. VT 06" (Dr Hugbe.k Many important questions about tobacco withdrawal re- main unanswered: (1) What is the prevalence and severity of withdrawal in the general population of smokers? (2) What is the time pattern and duration of withdrawal? (3) Does a prolonged abstinence syndrome exdst? (4) Why do srnokers vary widely in reported withdrawal severity?(5) Is withdraw- al the major reason most smokers fail when they try to quit smoking? Prior prospective studies have been unable to ade- quately address these questions for several reasons. First, most studies used the 5% to 10% of smokers seen in smoking cessation programs" as subjects. These samples are problematic becsuse smokers who attend such programs ap- pear to be more dependent on nicotine than self-quitteis; ie, they smoke more cigarettes per day, have more prior unsuc- cessful attempts to quit, and score higher on scales of depen- dence.n Second, most prior prospective studies used relatively small sample sizes. Large sample sizes are needed because usually only a subeet of smokers who try to quit actually do so. In addition, large samples are necessary to examine reasons for between-subject differences and to test the ability of withdrawal to predict cessation. Third, most studies were of a short duration. Few proepec- tive studies have examined withdrswal beyond the first 2 weeks of abstinence, yet much relapse occurs after this. The present article reports withdrawal dati from a clinical trial of nicotine gum. The cessation results of that trial have been reported elsewhere.° The trial provided a larger, more generalizable popuLtion with a longer follow-up than prior withdrawal studies and thus provided an opportunity to repli- cate and extend our prior findings on tobacco withdraaaL':" Specifically, we retested the reliability, validity, incidence, predictors, duration. and clinical significance of the symp- toms of tobacco withdrawal. We also retested the effect of nicotine replacement on tobacco withdrawal. SUSJECTa AND MtETtiODs Three hundred QReen'smokers were recruited for a randomizad, ~ two hmi>,y practice eliaio bp ph ~ rsidaa referral and word of mouth. Incluaion aiteria were daily smoldng, a desue to stop smok- ing, DSM-111 criteria tor tob.eeo dependence, no present history of Axis I psychiatric or drug abuse disorder,; and ability to ebew Qum. S2 Ared+ Gen PsvchiRt*v-~~,+ aJl ,?~r+u~,v 1941 Tob.ooo VYOxtromai -F+ iohas.t a!

I I I I I I I I I I I I I I I L TabN 1.-SUbpCS (:zT3faQIRSbCs1' Pre..nt Stvoy Sttb" haDrti cganttes pw CsY 29.2 _ 12.0 29.6 s 10.7 30.2 s 10.6 ZOp N,cotx+m y+ad. mqagantte 0. 72 = 0.30 0.70 z 0.29 Duranon ot unoV+q, Y 1 S.7 _ 10.2 19.7 _ 9.1 19.0 a 10.1 21 % Who tMO 10 oue >3 bmes 39 36 y{ -- 271 Feqerstrdn xnn 5.6 01.5 5.7 = 1.5 7.1 § n 1.2 - cRae7CterlstSt P„e.bo aum (N=10'6) Hlootln. aum (N.21(1) CassarJOn cnnlet (Nsi0D1 a,..reqe uS Smpitae7 4st++oqrsprwcs Aqe.y 38.3:10.3 374_97 36,7_ 10,6 35.+4 •'.MM 41 45 544 SM : Msm.d 62 751 60 S5i % Comoat.o nqn eU+oo+ 95 97 96 ISM : WhRadlu woncM 50 54 78 % wrth noanarwiC ineorne >S30 000 51 721 •vaws a» mean e SO urtmess oawrns>ie rloicaN0. tF+Dm Muqnes etld MabtattaM." SFnom tRe Offroe on Smokx+g and MeapA (.L PieRS. PhO, wrtnrrt oormwrwafbn 0ct1bK 19841 0ata on iroome, oocttpa+yon. ar+0 nwoane Y1efd an not pttlMtted as IMse Cata aff 4 Yeef3 dd. §P<.051or dtlMnce oaw" n praau sridy and oeaaenon GlYrC pP<.05/or odtersnes baw.en pr*see,t study ano avar.qe US smoket 1P<.051or dnfennce Deewwt nrc,ooru st+e ptaeWo Qum qraupa Ezcluaion aitlria were prior use of nicotSne gum, use ou tobscco otber thsn cigarettes, and contraindication to nicotine gum. The demographics and smoking habits of the present sample were simiIar to thoee of our prior clinie study except for fewer quit attempts and a lower dependence score on the Flagerstrorn" Scale (Table 1) (P<.05}. Compared with the average US smoker, our auaple smoked more ciQarettes and had tried to quit more otten (P<.05). Within our .ample, the placebo and nicotine groups were similar except the nicotine group had i higher income and were more licely to be married (P<.0d). Analyses were not corrected torthese demographic and smoking habit dilferences (ie, sez, education, occupational ata- tus, income, and marital statua) because none oithese wen related to the severity of withdrawal outcome (aee belowi Proasdwa At the initial seaaion, subjects completed inforsaed aonsent and a precessation rating of the symptoms of tobacco withdrawal. Subjects named an observer who saw them daily (usually a apoctre or coworlt- er) and were given a similar rating form fot the ob.er.er to oemplete and mail in. Subjects were counseled by the study nune and their phfsidah for 10 minutes each, Qiven a smoking ceaaacion booklet, shown a 13- minute slideftspe about nicotine gum, and given a preac:iption for gum. Subjects were advised to use the eum ae deecribed in the package insert, eg, stop smoking abruptly, ehew ad libitum an crrr- ing, use for not more than 3 montha, and ta~er Quta use. The interven- tion is described in more detail dsewhere. " Subjects were randomly assigned to nieotine or placebo group in a double-blind manner in a21 ratio. The nicotine;um wse the maritet- ed 2-mQ dose (Nieorette, Xarion-!(.rrell-Dow, Cindnaati, Ohio). The placebo had no nicotine but waa tia.ored to tasta similar to nicrotine gum. Subjects obtained jum ad libitum at the pharmacy located at the practice. Subjects attended a follow-up session 1 to 2 weeks after their quit date in which they were eounseled brieDy by the nurse and phrician (10 minutes eaeh). Both this session and the initial ssaion wert doae individua)ly to prevent subjects from ooctpariae withdrawal sytap- toms and gums. At this aaaion, subjects submitted a breath auaple for carbon monoxdde to verify ceaaaboo." Further treatment was not 2iveli. Follow-up forms weee mailed to the subject and their deaignated obssner° at 1 to 2 weeks and 1, a, and 12 months aftter their quit date. If these were not returned withia I week, a s.aoed form was mailed MoR Gen PsyCh+atry-Vo1 b, Juxiary 1891 out If these were not returned, we called subjects up to three times to obtain the same information. The follow-up forms asked about withdrswal symptoms and smoking status. At the 12-month foilow6 up, those who daimed abstinence wer+ paid to return to obtain a breath sampLe for carbon monoxdde" and a saliva sample for mtinina' (a metabolite of nicotine) aad thiat7anate" (a byproduct of tobacco smoke)~ meawtse The observer and subject withdsawal ratin~sciles were identical to those used in our prior outpatient studies. " These seales have been ahown to be valid, reliable, and sensitive to cesaatian and nicotinic e8eeta. 'The salea aaked subjects or observers to rate the subjeety symptoma, with 0 indiating not present:1. mild; 2. modar- ate; and 3, severe. Ratia;e were based on the taat 24 hours. The ratings w.rr eaapleted only by abstinent subjeets, with the exeep- tion that at I-week follow-up, aonabatinent subjects wera asked to ratu tbemselree based on the time not smoking. Weight was queried at 1-and 6-month follow-ap but not at 1-to 2-week follow-up Obeer.- ers were aiked to ignOre seif-reportae much aa possible and baae their ratitq on observed beharior. Observers were not asked to rate several sy 'I mptoasa beeause pilot work indicated they could not do so reliablr Obeertiess were aLo asked to indiaLe whether they thought each day'a ratint was reliabie. Weight was meawred at 1- and 6-month tollow-up by eelf-eeport, which has been found to be accurat.. Data AnaFytl. Although the symptom aoaj.e were ordinal arxl many ot the aeoeee were eJ<ewed with many sero eaores, our large saasple eize permitted parametric analysa. Unc orrected two-tailed P values arm g;.en. Readers may wiah to halve the P ralues considering our a priari hypothe.ee or to taceau the acapubie P value given the number of atatiatial teeta. iiEStlL?'S As..s.m.nt of Tob.eso *'Ithdraw.l The rymptomm of tob.er.o withdrawal were based on ratin;s of withdrawal amont the 106 subjate wfio e.aiwd plaabo gsmt. Valsd- ity~ tlliability, magtQinde, and inddenoe were b.aed on 1- to 2-week data. Self-repotted peeee.aatioa and 1- to 2-week foUow.up w9th- Tobaooo MfdhdrawN'HugN.e ta al S3 I

.r r. r ... .. +.. a. .. ~.. r ... ~ a. ~.r .. .. .. .. ~r. r t ,. M r i s 00 wa s 00 ~d fJ d 1 p0 •L) I s 00 40. 8, 00 v 7w 9 It 00 O w S z w t t zs a • • w O I U p $I ~' zs a N N aa z~ ~. 1~I f~ ~ .fi9966E9V99

I I I I I I I I I I I I I I I I I I scrore, eg, increased from mild to moderate. The magnitude of each symptom was calculated as the simple difference in preeesaation and 1-week scores. Among t'-he self-reported symptoms. hunger and in- creaaed eating occurred in 67% otsubject.s and increased an average of 1.1 of a possible 3 units. Self-reported anger, anziety, ditdculty concentnting, impatience, and restlessness occurzed in 52% to 59% of subjects and increased 0.6 to 0.8 units. Self-reported crsving, drowsiness, insomnia. physical symptoms, and stomachache oc- curred in 28% to 40% of subjects and increased 0.3 to 0.4 uaiu. Observer ratings produced similar incidences and magnitudes. The mean (= SD) number of symptoms that increased wzs 4.7 x 2.2 of a possible nine self-reported symptoms and 1.9 -1.4 of a possible 5ve observed symptona. Sizty-eight percent of subjects reported at least four symptoms and would have been diagnosed as having nico- tine withdrawal in DS.tit-111-R. Relationship Amony Symptoms Withdrawal symptoms were highly correlated. In a correlation matriz among the valid withdrawal symptoras, 26 of 32 correlations were significant (P<.05). Factor analysis indicated four facton with eigenvalues of greater than 1.0: mood disturbance (anger, arudety, difficulty concentrating, impatience. and restlesanesa), somatic com- plainta, other symptoms (hunger, insomnit), and craving. T1m. Coun. Time-course analyses can be done using between-sample or within- subject comparisons. In the between-sample (or croea-sectional) cocn- parison. different follow-up cohorts are compared. This analysis has the advantage of pioviding the better estimate of withdrawal sever- ity for smokers abstinent at that follow-up. Its disadvantage is that, since it compares different self-selected samples, any differences in time course may be due not to the natural history of withdrawal but rather to dilferences in sample charscteristin across those still absti- nent at different foUow-upe. The within-subjects (or longitudinal) analysis compares withdraw- al only within subjects aver time and thus avoids this problem. However, the within-subjects analysis uses only the small selected sample of subjects who wese able to remain abstinent for 6 months. Our time-course results were similar for both the between-sample and within-subjects analyses. We decided to use the within-subjects analyses for statistical analyses because we believed they were more conservative. We decided to use the between-sample results for illustrative purposes because they are more representative af the severity of withdrawal at any one follo+wp. At 1- and 6-month analyses, few slippers were eneountered: thus, only subjects reporting abstinence were included in these analysa. If nonresponders to follow-up are considered smokers, and only obeer.- er-veriSed abstinence is considered, 43 subjects (41%) in the placebo group were abstinent at 1-month follow-up and 20 subjects (19%) were abstinent at 6-month follow-up. By 1- and 6-month follow-ups, many of the ratings had declined even among the smokers who did not quit (dotted line, Figure), suggesting a drift in ratings owing to repeated testing orseasonal effects. Thus, to assess whether any symptoms weit still present at 1 and 6 months, two criteria were used: ie, the symptom had to be greater thin the p:ecessstion score, and the difference score (be- tween the follow-up and precesaatioa) of abstinent smokers had to be greater than that for nonquitters. At 1-month follow-up, the only symptoma that differed from before cessation and from nonquittess were dillSculty concentrating (a dif- ference of +0.3 for abstsiners vs 0.0 for nonquitters), hunger (+ 1.1 vs +0.4), and weight (+ 1.4 kg rs +0.4 )tg) (P<,0R At the 1-month follow-up, we calculated an indicator of self-reported and observed withdrawal discomfort. This was done by summing scores of the nine nonredundant. valid, self-reported symptoma-anger, craving, anu- ety, difIIculty concentrating, hunger, impatience, insotania, physial symptoms, and restlesanesa-and comparing the difference in the precessation and abstinence seores. Neither of these summed scores differed between quitters and nonquitters at 1-month foUow-ap At the 1-month follow-up, the incidence dscores st01 above preces- sation levels for abstainers vs nonquitters was 45% s 29% for difficul- ty concentrating, 69% vs 44% for hunger, and 42% rs 20% for weighL At the 6-month follow-up, all of the symptoms except hunger and weight remained at or below precessation levels. Hunger remained above baarline in the quitters compared with the nonquitters (+0.6 Arch Gen Psychiaay-Vo( 48. JarWary 1991 va -0.1), and weight gain was greater among abetinent subjects compared with nonquuters (+ 5.2 vs + 1.0 kg) ( P<.05), Among the 6- month abetainers, 22'b had gained 9 kg or more vs none of the nonquitters. At the 6-month follow-up; the incidence of scores still above precessasion levela for abatainers and nonquitters was 53% vs 41% for hunger and 85% vs 29% for weight Detwminants of Tobacco Wtthdravral The ability of several possible determinants to prospectively pre- dict withdrawsJ discomfort was tested in exploratory analyses. This was done by relating the value of the preeessation score of the predictor to the increase in the total self-reported withdrawal dix.rom- fort score and the craving score at 1-week and 1-month follow-ups and to weight gain at 1- and 6-month follow-ups by product-moment correiations. z' tests, and one-way analyses of variance (ANOVA). The determinanu tested were age. sex, marital status, education, occupation, income, number of cigarettes per day, nicotine yield of cigarette, self-reported past withdrawal severity, duration of smok- ing. Fagerstmm score for nicotine dependence," smoking to relieve negative affect and craving scores from the'Ibmkins Scale,3' number of prior quit attempts, alcohol and caffeine intake, and history o( alcohol and other abuse problems. None of the items predicted total self-reported or observed withdrawal, craving, or weight gain to any substantial degree at any follow-up (r<.20), except that older smok- ers and those who had smoked longer weighed more at 6-month follow-up (r>.64; P<.003). A similar nonsignificant trend occurred in predicting weight gain at 1-month follow-up. M1Rtlsdrawd and Smoidng Cassat(on The clinicaW significance of tobacco withdrawal was evaluated by three tests of whether increased withdrawal would predict an inabil- ity to remain ab.tinent Firat, the nine valid symptoms that occurred at 1 to 2 weeks and the total withdrawal discomfort did not predict the ability to remain abstinent through the 1-month follow-up. Second, among the three valid symptoms at 1-month follow-up, difficulty concentrating and hunger did not predict ability to remain abstinent, but increase in weight did (P=.02). Abstinent smokers who had gained more weight by 1-month follow-up were mcrr likely to remain abstinent through 6-month follo.wp. At 6-month follow-up, hunger did not predict ability to remain abstinent, but a greater weight gain by 6 months showed a nonsignificant trend to predict continued abstinence through 1-year follow-up. Ethex of Nleotlr. Gum As stated above, in the placebo group (n=105), 84 subjects (80%) were abstinent at 1 to 2 weeks, 43 subjects (41%) were abstinent at 1 month, and 20 subjects (19%) were abetusent at 6 months. Compars- ble figures for the nieotine group (n - 210) were 158 (7596),113 (54%), and 61(29%). In the IIrat week, 88% of those in the placebo group and 85% of those in the nicotine gronp acquired the gum and averaged 7.7 pieces d nicotine gum and 6.8 pieces of placebo gum per day. Gum use dropped off quickly with only 42% of the nicotine group and 29% of the placebo group using the gum for at least 1 month. At 1-month follow- up, gum users chewed a mean ot 7.1 pieces of nicotine gum and 4.6 pieces of placebo gum per day. Fewer than 10% of subjects used nicotine or placebo gum at 6-awntA follow-up. Analyses ot the effect of nicotine gum were limited to the valid withdrawal symptoms at I- to 2-week and 1-month follow-ups. Anal- yses were based on intent to treaL An alternati.e to intent-to-teeat analysis is to compare nicotine gum with placebo gum only among subjects who were using the gum. Since subjects self-seleeted into gam-uaing and non-Vuta-naing groups, such an analysis may be bi- aaed. We ehoee to use the more conservative intent-to-treat anslysest hawe.= analysis of 1-month follow-ap may have produced false- negative results beeauae few subjects were still uaing gum at that follow-up. To determine the effect of niootins, the scores for each of the valid withdrawal symptoms were entered into a 2 x 2 repeated-measures ANOVA with drug (nicotine vs placebo) as the grouping factor and time (preoesaation vs follow-up) as the repeated faetoz Separate ANOVAs were done for the 1- to 2-week and 1-month folkaw-upe. These ANOVAs seuehed for an interaction indieating that the in- TW.ocn W&&a+Mi-Fk,gfws st al 55 I

I I I I I I I 1 I I I I I I I I I I Mean ratrnga for wRttdravai amorg mse- abstu+ertt at 1- monttt toltovruo (for pf~'-ebo. n-43; for ncoone, n=113) and 8-monttt fOJbw'uP (for d+- cebo, n= 201. Six-montft data for nkabne gum a» not given. as rncobne vs ptacebo aompan- son at 6-montn fol Io++wp 1s sus- pect bscause of the small num- ber of subjects using gum at that foilow-up (for placebo. n=6; for nicatine, n-18). A11 SEs ars less than 0.1 U. g • ¢ Irtttabwrty Total S.IS-r.porteC VY&KVnra crease in withdrawal was tus for the niootititt gsoup than for the PLWebo gMPF At the 1-week follow-up, nicotine use decreased sdf-eeported alad observed anQer and impstience and aelf-repoeted aamet3+, cra.ing, hunger, and insomnia (Table 21 Nicotine uae neitber decreased aelt- reported or observed restleaatiae or amdety nor alt-eeported diffi- culty concentrating or physinl snaptoma. At the 1-moath toUaw-epy nicotine use decreased self-t*ported diIDcntty aweentrsting and lam) ger(P<.06) but not weieht In the prior audia, level of dependence on tob.eeo appeared to determine respocue to aicotine Yuum.''Ib determine whether le.e1 a[ dependence predieted relief of witbdrxwal by t:icotitse, we entered total withdrawal disooeniort .cores at 1- to 2-week follow-vp (ia, change ecoree) into a 2 x 2 ANOVA with drug and dependence lao- torr. Dependeaa wsa claasiIIed by the Facerstran Saiq" the Z1m- idna Smoideg for Neqazi.e A>Yeet and Cra.icic acabe, aad by eerer ity of past withdrswai. Noaa ot thae predicted witbdrawal tdi.f by nicotine gua. On. poea%le ezplaaatson of the abs'iity d nicotine to raliew with- drawal is that aubjeets oould ideatify..hether they teeaiwd nieotaie or placebo gum and this identiseatiaa aaused them to eeport low or higb witbdraw:l symptoma-i In bct, 51% of aubjeeia eorteattlr idea• SS Iirch Gert Psyeltistry-bol 48, Jwxnry 1991 Mcoa,. Gum O : Total Obaer-.d wm,araVw we;qM r3,;n 26 Bl 4 26 tiW their gnm,1t96 itteorseetly identiIIed their gnta, and 35% we:e wncertsin. This tate didenti5ntion is areaterthan chance (P<.0011 Tb te.t the abo.e-ataied hypotheaia, we eYaluattd whether identifi- cs>son was related to nicotineti effeet on withdrswsl using a method described prtviowly.* The effect at nicotine `um was independent of kitntitlabon, ej, niootine reiia.ed withdrawal a.ea when we cora- pared tho.e who teaived niootine gum yet believed they had re- cei"ed pLcebo fnm with those who recei.ed placebo Qum yet be• liered they had r.edred niootine sum. L:OMYE1V•r S.M40porSfad SytirPUM+. The results o(the present study consrm all of the DSM-111- R aytnptoma: craving for nicotine, irritability, anxiety, diffi- culty concentrsting, restleaanea., and increased appetite (we queried about craving for tobacco, not nicotine and about hunger, not appetite). In addition, the study con&rms our prior resnlts and those of others that non-DSM-111-R aymp- totas of Withdrawal include impatienee, ineoasnia, and PhYm- cal complainta but not headacba.- Tobaooo WNlxlraM.d-Huphrs et al fS a- 3• 21 t Ditfiqrty Concoevatx+q .' 0 I

I I I I I I I I I I I I I I I I I TaA4.3.-Enecx ot Nlcoons Gum on Se6-moortsa anC Observed YYrtrar=wa+ Syrnptortu• - (tsSIN" ed) Huqn.. ~9sat" t> cro.s arw seJw+.b.. uto(19 (In Pr.ss) s ) w.N ( y~~) DSM-IlI-R symptortr Crmiq fa ncflona + + - - + - - - - ImcaNuty/u,q.r ++(«+) +«(««) + +* +« AruNry « «(-) • «(++) -- pff-Ay concwrtnong * + « - - - - R.etlessnass +(-) ++(+«) ++ -- -- IrXreaaW ippW* + - - - + * - - - - oT1ef tyRlptXRi Imoapency ++(«+) ++(++) ++ Inaorixu - - - - + - - ~ ... ... ... -- ++ •R*wR& wM ooa.nwr ra>X+4s n ptrMwolss: Pltis sgn ndCatss poaRM rssuR: mrxat s9n, npaov. MwYt: antl OouOM pus or mtttu Lgns, mOts /vbuat r.euRs With the exception of craving, the DSM-111-R symptoms occurred in 52% to 67% of subjects, increased by 0.6 to 1.1 units, and appeared to be statistically robust (t=4.4 to 8.7). Several non-DSM-111-R symptoms (eg, insomnia, physical complaints, and stomachache) occurred in 28% to 40% of subjects, increased by 0.1 to 0.4 unit, and were not statistical- ly robust (t-2.8 to 3.9). These values are substantially less than those of the DSM-111-R symptoms. Thus, although these symptoms could be considered valid symptoms, their magnitude and statistical reliability appear to be less than that of the DSM-III-R symptoms. The incidence (31%) and m&gnitude (0.3) of craving for tobacco was smaller than expected. As in our prior study,° self-reported craving scores were high even while smoking. Similar results have -been described with other drugs of abuse.' Prior studies that found high craving ratings during abstinence have often assumed craving would be low during srnoking and assumed high levels of craving after cessation indicate withdrawal. Our results question such an asaump- tion. There are several possible explanations for our small change in craving. One explanation is that our particular measure of craving is insensitive. Other more sensitive word- ings (eg, desire to smoke or missing a cigarette) have been proposed."1 A second explanation is that craving was artiII- cially elevated in anticipation of cessation. However, in our more recent studies, craving ratings taken 2 weeks and 1 day prior to cessation were similar (J.R.Fi., unpublished data, 1990). A third explanation is that craving is maintained by the availability of a drug.' A fourth explanation is that smokers use a different criterion in rating craving during amoking and during abstinence.' Weight increased substantially among long-term quitters, with a mean long-term weigbt gain of 4.3 kg more than nonquitters. More telling is our linding that at 6-month fol- low-up, 85% of those who quit had gained weight and 22% had gained 9 kg or more. These results are consistent with the conclusion of recent comprehensive reviews.' Oburver Ratlng. The present study also replicates our prior finding that increases in anger, anxiety, impatience, and restlesanesa (but not drowsiness) are apparent to others near the smoker-" This finding suggests that these withdrawal symptoms repre- sent actual changes in behavior. Unfortunately, neither we nor others have quantified the degree to which such changes produce functional impairment, eg, at work or with interper- sonal relationships. The present study failed to confirm our prior suggestion that observer and subject ratings are concordant.° In the prior study, the interrater correlations ranged from.40 to.62, whereas they were less than .40 in the present study. The smaller concordance in the present study may be because observer and subject ratings may have differed by 1 to 2 days on the particular day being rated. Despite this lack of concor- dance, observer ratings were sensitive to cessation and nico- tinic effects; thus, observer ratings may have tapped a differ- ent set of behaviors than self-ratings of the same construct. Tirtw Cows. The total discomfort scores and all of the valid observed and self-reported symptoms showed an initial increase and subse- quent decline (Figure). This "time-limited" pattern suggests these symptoms are withdrawal symptoms and not the simple offset of drug effects. ` The mean total discomfort score aad almost all of the symp- toms had returned to precessation values by 1-month follow up; this is consistent with other studies.' However, a sub- stantial proportion (20% to 25%) of subjects continued to have withdrawal symptoma at 1-month follow-up. Increased hunger and weight persisted beyond the 1- month follow-up and returned toward precesaation levels at 6 months. This result is also concordant with results of prior studies' and suggeeata that hunger and weight gain are either tune-limited symptoms that have a long time course or are simple offset effects due to the absence of nicotine. Protracted withdrawal syndromes (ie, lasting 6 months or longer) have been described for alcohol' and opiates.' With the possible exception of hunger and craving, such prolonged tobacco withdrawal did not appear to occur in the present study or prior studies.'*" Craving for tobaaco did drop below precesaation values, yet 74% of abstinent subjects continued to report craving at 6 months after cessation. Although con- tinued reports of irritability, anuety, etc, at 1- and 6-month follow up could be due to the normal everyday stressors of life, continued reports of craving for tobacco could be intu- preted as evidence M a continued abnormality. One explana- tion of prolonged craving may be that stimuli associated with either tobacco use or withdrawal continue to elicit craving and are difficult to extinguiih." Recent empirical work suggests such conditioned effects can occur with cigarettes but are more pronounced during ad h'bitum smoking than during abatinence.' , Ara, Ga, Psya,iatryy-vol 48, ,lanxury 1991 Toeaooo vr/a,draMr-HuqN.a.t al .s7

I I I I I I I I Varlabllhy Across Subocts Atthough some investigators have commented that the severity of tobacco withdrawal varies.widely across subjects, it is unclear that this ftnability is greater than that seen with other drug withdrawal syndromes." Except for weight, we could not account for any of the between-subject variability using a number of previously tested predictors. Withdrawal and Smoking Cessation A major hypothesis of the nicotine dependence theory is that tobacco withdrawal prevents smoking cessation.'°" Based on smokers seen in withdrawal clinics, some investiga- tors have noted most relapses occur well after withdrawal has abated' and have thus challenged this hypothesis. However, studies of self-quitters indicate 60% to 65% relapse in the first month after cessation.°" In the present study, neither the individual symptoms nor total withdrawal discomfort during the first week or month predicted relapse at either 6- or 12-month follow-up. We are aware of only two prospective studies of whether smokers with more withdrawal are less likely to remain abstinent. Our first study did not find that seLf-reported withdrawal severity predicted relapse within the first 4 days after cessation.' A second study by another group reported that depression pre- dicted relapse during 2 to 3 weeks after cessation." Contrary to what many would have predicted, subjects who gained weight had a smaller likelihood of relapse than those who gained little or no weight. One prior prospective study of smoking cessation found a similar result."One expla- nation for this finding is that food deprivation increaaes the reinforcing effects of drugs." Cessation of smoking may de- crease metabolic rate.' If so, then to not gain weight, smokers may have to deprive themselves of food and thereby increase the reinforcing effects of cigarettes taken during minor slips. Effect of Nicotine The effect of nicotine gum on tobacco withdrawal in the present study is concordant with the results of our own" and prior studies (Table 3).' Our recent review of prospective studies of tobacco withdrawal concluded there was clear evi- dence that nicotine relieves increased anger, anxiety, and impatience, marginal evidence for relief of difficulty concen- trating and restlessness, and no evidence for relief of drowsi- ness, insomnia, and physical symptoms.' The results of the present study are quite concordant with this review and extend such conclusions to nonclinic attendees. In the present study, the effect of nicotine on craving and hunger was the largest and most statistically robust of any of our findings. These results contradict the findings of both our own" and several other studies' that nicotine has little effect on craving or hunger. The results are concordant with a recent study." One possible explanation for the discrepant findings is that the effect of nicotine on hunger and craving does not occur in the first 1 to 4 days (when nicotine effects were measured in the negative studies) but takes 1 to 2 weeks (when measured in the positive study) to occur. This explana- tion suggests that nonpharmacologic factors (eg, expectan- cy") control craving and hunger in the first few days after cessation, but later, pharmacologic factors influence craving and hunger. Nicotine gum did not decrease weight gain in the present study. In prior studies, nicotine gum has and has not de- creased weight gain.' Possible explanations for our negative results are that we used subjects who were not clinic attend- ees and who used less nicotine gum than those subjects in prior studies. 58 Arch Gen Psychiatry-Voi 48, January 1991 Validity and Get+era/t?,alrty There are several reasons to believe the pr[3ent study produced a conservative yet valid assessment of tobacco with- drawal: ie, the study used (1) a large sample, (2) previously validated measures, (3) scales sensitive to only large changes in symptoms, (4) withdrawal scores calculated as changes from before cessation, (5) observational measures, and (6) analyses based on intent to treat. There are also reasons to believe our results are more generalizable: ie, our use of (1) few inclusion criteria, (2) a large sample, and (3) a sample that was not composed of cessation clinic attendees. Despite these assets, our study has some limitations. First, although the sample sizes were adequate for the 1- to 2-week analyses, the samples for 1- and 6-month analyses were a smaller selected group of smokers who were able to remain abstinent. Second, the absence of control groups of never-smokers, long-term exsmokers, and continuing smokers not trying to quit limits our interpretations.' For example, without such groups, it is impossible to determine how closely the symp- tomatology of our recent abstainers approach those of long- term exsmokers or never-smokers. We did use relapsed smokers as a control group, but this tactic may have produced biased results as subjects self-select into relapse and abstain- ing groups. Third, our sample volunteered in hopes of receiving nico- tine gum. This recruitment bias may have selected for a sample likely to show withdrawal effects and to be sensitive to nicotinic effects. In addition, although our subjects were medical outpatients and not withdrawal clinic attendees, their demographic and smoking characteristics were more similar to those in clinic samples than the average smoker. A more generalizable test would be to examine withdrawal symptoms among true self-quitters, ie, unselected smokers who quit and receive no intervention. We are presently con- ducting such a study. Fourth, in the first week our subjects averaged only seven pieces of nicotine gum (at a 2-mg dose of nicotine) per day, and this may have decreased the magnitude of nicotinic effects. In addition, most subjects did not use the gum for long and this no doubt decreased the magnitude of nicotine effects at 1- month follow-up. Fifth, although our sample size was sufficient to detect withdrawal and nicotinic effects, it may have been inadequate for analyses of determinants of the variability in withdrawal and of the ability of withdrawal to predict cessation. Both analyses required comparing 20 quitters with 85 nonquitters on more than 20 variables. CONCLUSIONS In summary, the present study replicated and extended to a more generalizable sample our prior findings that a defined group of symptoms occur with cessation of tobacco and are (1) observable, (2) prevalent, and (3) relieved by nicotine, but (4) do not predict who will be able to stop smoking. In addition, the study found that nicotine withdrawal usually lasts 4 weeks, but hunger, weight gain, and craving persist for at least 6 months. This investieation..a+funded by;ranu DA-440H6. DA-03728. and DA-0298. Research Scientist Development Award DA-00109 (Dr •Hughee) from the National In.titute on Drug Abuse, aod grant HL39?2o hom the National Heart. l.ung, and Blood Inatitute. Bethada. Md. Margaret Healy, Roy W. Picicsna. Diane Ramkt. and the pAyaicians and phumatiata at the plymouth and Maiu Clinica of Park-Nicrollet Medial Center, btiaas.poli.. Mian, helped in candtutiag the atudy. Stere Higytina and R,ubert Weat prorided eoeaa+ents on the maass.eript Tobacxo wrthdrawal-Hughes et a!

I I I I I I I I I I I I I I I I I R.+.nw.a 1. Hugh.s JR. Hinta. ST. Hats+t....+ DK. E'Smd of abatinente from tobaeeo: a cineal revxw. In: Kcrla"lc LT. Antua H. Cappell HD. Glaser F. Goodstadt H. Iarael 1F.-Kalant H. Sellers EM, Vi+Cilia J. eds. fisa.arck `ldusxes sn .UeoAol and Drr, P+obl.ma New York, NY: Plenum Press: 1990:10:317398. 2 Cherek DR. Bennett RH, Roaeb. JD. GrsbowsidJ. Human ag¢reesive and non•ag;resaive respondu+r dursn` acute tobaceo abtinence. In: Harria LS, ed. Problems of Drvq Deptndswee 1989- Ptcetediwqs of CA. 61at Anntial Seunafie Sfsehaq. TAs CommUUs on Probl.nu of DmW Ikpendrnes. Inc. WuhinQton. DC: Nauonal Inuatuts on Dru; Abuse Research Monograph Series. 1990:95:435-t38. 3. Harns CM. Emmett-OtfeHby MW, Robinaon NG. Lal H. Withdr+waa from chronic nicotinc substitutes partiaily for the introceptave stimulus pro- duced by pentylenetetn:oL Ptyckopha*maeology.1986:90:86-89. 4. Snyder FR. Davis FC. Hennmg5eld JE. The tobaeeo withdrawal syn- drome: aaseas.d on a computerized tert battery. Drvq ALcokol Dr- pend 19&9:23:259-266. 5. KleaQe. RC. Meyers AW. Winders SE. Deurtniniag the rnsons for Meight gain fotlawnnQ smoldna ees.ation: current IIndin1s, methodoloQieal issuea, and future direetiona for rexateis. Ann Bekav M.d, 1989;4:134-145. 6. Rodin J. Weight cban;e followin` eaojdng enaaioa the rale of food inuke and exercise. Additt B.Aaa.1987;12:303317. 7. Hughes JR. Keenan R. YeUin A. Effect of tohecro wftbdrawal on sus- uirsed uunnon. Addiet Ba1ae. 1989:14:577-b80. 8. Kales J. Allen C, Pteatoa T. Tae T. Kalaa A. Chansea in REM akep atsd dreamina.nth aguetu amoldng and fo8owust withdra.al. PrychopAysalo- gy.1970:7:347348. 9. Hughea JR. Hataukami Di{. Skoog K Phyaeal dependeaee on nieotma Qunm: a plaeebo-eubauuation truL JAd(A.1966:255:32773279. 10. West RJ, Russell MAH. Effects of witbdrawal hom long-term nicotmo gum use. Pryekol ![.d. 1985:15:891-89:i. 11. F'wrs MC, Novotny TE. Flynn W. Maklaa D. Davis R. Smoking eeaaa- tioa data trom the 1986 Adult Use of Tbbaeco Survq: In: Aoki M. fiiaatmaehi S. Tbminap S, eda. Sm4kihp aAd flea/tk 1997. Protwdiwqs af tlw SistJt Wor(d Confer.nee on Smokiap and flealtA. PrirKeton. NJ: Escarpta Medita4 1988:189-194 12 HushesJR. Hataalami DK. Signu and symptonn of tobaaeo witbdrawaL Arch Gen PiyrJtietry. 1986:432S-294. 13. Huc6asJR, GustSW, Keenan RM. FenwiehJW, Hea1y ML. NieoCnf.s placebo gum in Cenerai peaetic.. JAMA. 1989:261:1300-1306. 14. HughaJR. Hatsultami DK. Picksas RW, Kraha D, MaiiaS, LukaieA. EKeetaf nieotine on tba toh.oco witHdrawai syad:cen.. PsycAopAm+eaaotoyp. 19840:82a7. 15. Hughes JR. Miller'S. Niootine gum to help atop amoidns. JAd[A. 1884-952:?8rS2858. 16. Faterntrom K-0. Measuring de;rae of physieal depeadena to tob.r;ao smoking with refetttxe to indiridualitatioa of treatmeat Addid Behas. 197E:3:Zi5-?AL 17. HuehesJR.Kecika'LDoetoraheipfng staokera:reaiworidtaetia.dli+as .1frd.1986;89.293-296. 1d. Hughes JR. Fradern7oen LW, Fraaer M. A earboa mosw:tide analyr+r for measurement of smoking beharite &tias ?ker.1976;90.296. 19. HuthesJR. Utilityo(obeervertepoetso(amoidn;atatva.JStidatAbww 1}.ar. 1990?249-253 20. Jacob P. Wilson M. Bena.its NL NiseoOaa aed mtirdae detaminatioem in biologinl fluida: an improved Nchromatogsaphie method. J CAro+aatopr. 19812?2:61-70. 21. Denwa PM, Daridow B, Baa HE, Jonn EW. A eb.miayi tast Lor smoking exposure. Arut Ea.i+o~s B.a1tA. 1967:1406bd7it 22 Wing RR. Epstein LH. Os.i.p DS. LaPeett RE. ReliabBity and va>id;tF of setf•ttportt and obaes.a+r eatimaw a( reLtire ..igbt Addict B.Aas. 1979:4:133-140. 22. Stunkud AJ. Albanm JY. Tbe aoearaq of aelf-npastad.agha. Am J Clia NYtr. 1981:34:1590-15fY. 24. Ihuird FF, Green DR. Hoea D. A atale (e difUreatinto bat..s types of smokin; as related to ma:ug.mau of aSaet. lat J Addiet.19M14:61M6i. 25. Hughes JR. Krahn D. BliaBsier aad tsa .alidiey of dwbi.biind dm tnala. JClix PsyckoyAar+woeol, 1986*. 138-14Z ?.8. Meyer RE. Mi:vs SM. I%. X.+oen Stimrlns.• Implu+ar,oa./b.a fll.ory ojAddietion. New York. NY: Plenum Ptw: 1979. 27. Knvlowsld LT, Wilkinson DA. Use and anauae of the coneept a/aaanC by akohoi. tobacco and drug researchers. BrlAddicL 1987:82.l1-4b. 28. West RJ. Schneider N. Craving for e+garett.a. Br I Addiet. 1987:82407• 415. 29. West RJ. Kratuier HR. Craving for nYuates and p.yeho.eti.e dryt-L In: Warburtnn D, ed. Comparutsvs Drvq Usa. Patv. Franee: Har.ood Aes- demfc Publiahers. In preas. 30. Gritz ER, KIeeQes RC, Meyers AW. The smoldtg and body weight relatuonship: implieationa for intervention and post.ceasabion weight eontraL Ann Bekav.N.d. 1989;4:149-153. ' 31. IGeaeee RC, Meyers AW, LaVasque ME. Smoki.ni, body weight and their effects on smoidesg beharior. a comprehenaive ee.new of the litsrasute. PrycAol Bull. 1989106:204-Zl0. 32. Gtoss J, Stitser ML. Nicrotine teplaeetaent: ten-week eIl'eets on tobacco withdrawal symptoms. P+yctopkarmaeollopy. 1999:96:331.'i3(. 33. West RJ. Hajek P. Beleher M. Titne cour.e of cigarette .itEdra..l symptoms during four weaka of t:eatmmt with tucocine gum. Addiet Bdhae. 1987:1215. X Wert RJ, Haj.k P, Bekher lt. Tima eoaru of cigarette w4thdrswal sympconu vhile -, nitoans rum PsyeAopAarmacoloyy. 1989:":143145. 3d. DeSao CB. O'DotsncII WE, A14.d LJ. Lop.s CE. Symptoeuta(op m akvbotia s wiau eta`.s of abatioera AkwMol Clitt Esp Ru.19669.506Z12. 36. Martin WR. Jasin.ki DR. PAysioloSial parameters of mosphiae depeoM dena in oon-toleraac., early ab.tia.ap and protracted abstinarA.. J Psy- eAia& Ru.1969!1:9-17. 37. RrJder A. Pp(oid Deperds#et: U.eAawfami aad tYiatineRt. New Yoek NY: Plenum Pse.r 1980. 3g. Ha.a.,ka...: DK. Opland E. Husbe.JR. McCoy P. Krenan R. Conditional ~~~ 'edat.d etimuli amon9 aQuate anwkers. P+ydropAar- 39. KatzJL. Pharmamb;id and beha.ioaal Leton in opioid depmdana ia animala. In: Goldberg SR, Stolerman IP. ads. Behavioral Aaalys(s of Dn+q D+p.+d.*e.. Oriando. Fh Aadentie Prer Ine.1986187J28. 40. Ajhton H. Stepeey R. S+nokieq: PsycAo/agy aad PlaroemolapY. LaeN don. Eailaed: Ta.iRtoelt Pubtiahas Ltd:198Z 41. Mansan GL, Gok4it>; JF. Iat PtyckopAannoealopy of S+aoiA.ng. New York, NY: Cambridge Uai.asity Prat 19B1 42. Huac WA. Bstaett LW, Branch LG. Ralapa ratea in addictioo pro- etama. I Clin Psych. 1971.'27:45b4Gb. 43. Gritz E. Marcva A, Carr C. Unaided amokiageatation: Great Aaxtiaa Stnoheout aed New Ynri Eve qtottat J Psyrho.oe Oneoi.19886217-234. 44. Marlstt GA. Curry S. Gordon JR. A longitudinal analysis af unaided atnoirint tewtion. JCowsrll Clia Psye,Yol.1988:56:715-720. 45. Htajek P. Weat RJ. Beloher M. Seneritp of .ithdra.al symptoms aa a pr.dietor af ontcome af an attempt to quit amoiting. Pryakoi ,1[sd 1989:19i81- 886. 46. HaS Sl[. Giaa~ D. JwAa R'L Smokitq ee,udoa aad weight Pm J Coaarlt Clir Psye/1o41866s4:342446. 47. Careoil ME. Me(aeh RA. Inata..d drns-reintoreed behavior due to food depeiration. In: Thampaon T. Dews PB. BareettJE. eda. Adroarees ioe BeAav- (oral PMarrwaeotogy: Orlando. F1a: Acedemac Press Ix 198k4:47-aE. 4a. Groe.J. Stitz,ar iQ. MalduiodoJ. Nieotaie replaoement effem on poatn asaabon weight pia. J Conarlt Clbs Psye+lot 19A9:57:87-92. 49. Hughas JR. Gulliver SB. Amori G, IG:.nalt G. Fenw+ek J. E.ffat of it~aroedoty aad nieotia. aa emoicas uaaaioR. withctrawal sya+ptama and aelf- administraboatot eieotiae gnm. . 1989;99:486-491. 50. Se3meidar NG. Jarvik ME. N iootiae gnm.a pLeebo ruen: eompareaoea d withdrawal aymptama and sttoeess rwa.Ix Grabowaid J. Hall SM. eda. PAar *reolop(cal AdjrwRt in SwAkir0 Ceeeatiaw. Waahinzton, DC: National Inat')- tmu af Drug Abusa Rea.are3s Moeopaph b3; 19BbSi-101. Dept of Health and Haman Serviees pubfiwooa (ADM) 86-1333. 51. Weat R,>: Jar.i. MJ, Rua.dl 1(AH, Carrutber. YE. Feyaband C. Etfiec of nieotiam eeplaanknt un the egsetta withdrawal syad=oene. Br J Addiet.19@1;79:Y1S219. Arch Ge11 PSycriiaVy-Yol4a, Jarxmy 1991 Tobactao WRhdrawd-Htaglts *t al 59 it