Tobacco Documents Online

I I I I I I I I i I I I I I I I I I Nicotine vs Placebo Gum in General Medical Practice John R. Hughes, MD; Steven W Gust, PhD; Aotxrt M. Keenan; James W. Fenwick, PhD; Margaret L. Heale.y. PhD Three hundred fifteen smokers who attended a family practice dinie and wished to quit smoking were assigned in a random, double•blind manner to roceive either nicotine (2 mg) or placebo gum. Smokers initially received brief advice from a physician and nurse, a slide presentation and written materials (29 to 35 minutes), and a single follow-up visit (12 to 20 minutes) one week after cessa- tion. After corrections for marital status and income,10•ti of those who received nicotine gum and 7'ti of those who received placebo gum reported oontinuout abstinenci for 11 months and passed observer and biochemical verification (this difference was not statistically significant). We conclude that, when uaed in a nonselected group of smokers along with a brief intervention in a general medical practice, the pharmacologic effects of nicotine gum to increase cessa- tion are either smaU or nonexistent NICOTINE gum is an effective aid to smoking cessation R hen used with a be- havioral proQram in smoking cesaation elinics" or, in some instances, with a behavioral program in general medical practicee.' In fact, the Food and Drug Administrstion-approved use for the gum apecifically states that the gum is to be used only with a behsviorsl proQram (Nicorette package iruert, Merrell Dow Pharmaceuticals, Cincin- nati, 1984). How"r, 99.5% of those given the gum are not in such a program but rather receive gum during a brief visit with their physician.' This is be• ause few smokers seen in general praN • F,vn s» o.a"mu a r.rertimy, r.yu+orooy. .na f.^uy r,.eucOlo(Hkiprt),.ne mnrw++.ru.ne S1tS4tCs (De fen+nCk). UnNrMy of M1nywM. aueVfp. ra,. e* Nuow rx,nuu a, on,p ADus. waurqton. DC (a Gus,) .rro,no o.p++u+»rn eM r.yewmr uN- Nrtay Oa M"~fiOla (M+ Kw,.A). Ma rak.+Meowl 1A.04at Fow+oa1oon (Dr N*.iey). AArw..poM wcMx *qusu a+a,Rw a.hr"a., rn.n+600c). Qy 6abor.lory, o.o.nn»M a.syemiry. uh»nRr d wnnax Ce" a M.uu+.. orr S Aoto.cl A. ea• .1V«,. vTosWDr wworr.t (lAXA 1fS9X1:130d1306) tice are willinQ to attend behavioral pro- grams (<2% in one study`). It is unclear.vhether nicotine Qum is effective when used w-ith brief interven• tions.' 'Ibn trials randomly assigned smokers to advice alone vs advice plus nicotine gum and determined lon;-term cessation (Table 1). Six found higher quit rates in the gum Eroup•" and four did not." Three trials showed a statis- tically siQniflcant advantage with the addition of nicotine run.' In the six positive tria3s, the absolute increase in long-term quit rates with use of nicotine aurll ranged fsom 4% to T%,''" except one study found a 16% increise.' Tbe relative increase in quit rates ranged from 1.3 to 2.T times more cessation in the nicotine gum group than in the iroupnot receivin[Nm- Trvo of the previously cited trials" and two other studies" randomly as- signed smokers to receive either nico- tine or placebo gum along with a brief intervention. One study reported 8% higher quit rates with nicotine gum over placebo ium at six-month follo.vup,n but the two studies with one-year fol- low-up periods reported no advantage of nicotine gum over placebo gun." We believed another plaeebo-con- trolled trial of nicotine tum and brief intervention was indicated for two rea- sona. First, the prior positive placebo- controlled trial selected only hiEhly mo- tivated subjects, failed to biochemically verify all abstainers, and failed to report one-year outcome." On the other hand, the prior nsQative placebo-controlled trials' used populations known to be resistant to smoking cessation treat- ment, failed to provide adequate phyai- cian t:rainina, and lacked appropriate folloe+-up. We planned our trial to avoid these problems and to provide results more generalizable to the typical sr>,o- ker/ph)-aician interaction that occurs in a tenersl medical practice. SUBJECTS AND METHODS Wbjeeb Zhree bundred fffteen smokers w re reeruited from two nonacadenlic, fimdy practice clinics of a health maintenance organization. The clinics.Pele located in suburban Minneapolis. Subjects were recruited both by physician referral and by word of mouth. Recruitment oo- curred between November 1984 and June 1985, which was nine to 16 months after nicotine gum began being mar. keted in the United States. Inclusion crit,eria.vere daily cigaretxe smoking, a desire to stop smoking (as indicated by willingness to set a quit date within two weeks), and the ability to ehew Qum. Exclusion criteria we:e prior use of nicotine gum, use of tobacco other than cigarettes, and a eontiaindi- ation to nicotine Qum." Since referrfns 1 1300 JAMA. Marct+ 3. tlt9-WI 261. No. Y Nwsr Gvn n Gensrel PraCbu-Hupna et sl

1 ~ i 1 1 1 i 1 i ~ ~ 1 1 ~ 1 TaDN t.-lor+q•tarm Aostinancy in Rancforn¢b Tnais of Nicotina Gum Wrtfl Bnat Atlv>ty• su o0et, % Aaetlrrett I~ ir 94SeM+9 Mo. N+eanna Nlooune Gum vtt No Gum centrel Nleonn. conao+ Bnesr+TnonocSoo.fyw i'urnonarycon4c 777 to 9 Carr+ows N alll GarwY praet+a 573 3 1 FpemtrUm' Gwwa! Drwbes 145 ... 251't Of FW~ 01 Y" lktMr9rty MaR1 Is1 a~ ft9a « or Ruw++at ar SAauqnnauar V W Sunan t -uri.tr G«+«., p-e~. z99 , 2a a 13u+«y pr.cao. 15Sd Caae+arv Drataoa 99 woek+n• zM swtron a -+a,i.x» Wa+Xsa. 1ei W;,on w or oAnMY prscbN 19M . . . 13 n 10 0 is 4 9 2 9 4 MlOWfM Gum N -I=bO 9Mbh Thonee Sooaty Puknoesary tlmit 902 10 11 Camoe.M «oil, Gse" pn~+oa aJ6 . . . 3 2 Foronara+gY• PueaenaaRr+dxrc 600 JOt3 2n .Ju+uo:k at a!^ G.n.ti praeao. 900 tot s •.• .•• •Parantapa axtt)nuovsy aDSUnant and D~oetwmica~y v~nfr0lzap Si noud. t-'cwtt pravaNnp ratnM than oonanuous aesenae+ea. l1SoCMrtwoaYy mn/ytl in onty a aubMl 01 GalrtMO aba4MlMtl fp<.os by r, t.st No emootwr,icA MnRcsSm Tab1s 2. - SuDject Chanetarisba Mwrtt Swtly Clrrsew IaUe Ilbetlfr (M•tt0) -bOmbe (N.10d) c....tlen CIIRIC", fM.100) Avwpa tla AbWt Sn+okaet Damo9ra)rMa AMu+(sSC).o•.r 57.4 (29.7) Ss.S (s1oa) 361 (sto.9) 'x.IdM 46 41 54 Sa 1b A1art»0 75 in 9o aS x CompNUO nqn su,od 97 9S 96 W xWR1n.ODHar,rdkM 54 b0 7t ,9 With frousNhp10 :tom+. >s3o000 72 s1s 46 se smokx+o naea Moan ( e SDI N. at o94rsn.sW aar 29.9 (a10.7) 29.2 (mttA) 30.2 (sto.1) 20 kMan ( e SD) wmetr yo*G. n+4a0aroft 0.70 (=0.29) 0.72 (z030) 0.M (m0.13) 0.95 Mban (: SD) duraeon d 0100, r 19.3' (29.1) 11.7(e10t) 19.0 (310.1) 21 xwHOe+.esoa+K >J Was 3{ 31 64 27 AMan ( s Sa) Faqwsaomn troDrs 5.7 (:1.5) s.9 (:13) 7.1 (:1.2) NA F tEf4rr+atao~komM Y1aDys• Orpt4pw by 1-iMCr of 1M OdROtt on 8mokinq aed FMa79t afd M The 11oM4 Cats40uwnc•t Of Sniobnp: /WCOM» AOMCapn: A AapoR of Ifa $wpaon GMfMN": NA ete~ rot t1wliDb. ae+.W,a wtw"n nrow.,. ane a.0e0 poupa. physicians knew the inclusion/exclusion criteria, only ten smokers who applied were excluded: four due to prior use of nicotine gum, three due to recent dental work, two due to temporomandibular joint disease, and one due to pregnancy. The sample size was chosen to detect a 10% increase in one-year quit rates. In calculating the necessary sample size, a one-sided a of .05 was used because pri- or data suggested and we were intereat- ed in only one outcome, it, an increase in quit rates with use of nicotine gum. Also, the trial used a 2:1 (nicotine.pla- cebo) randomization scheme for reasons outlined later herein. Using these as- 9umptions, we calculated a sample size of 315 subjects. With this sample size, the power (1- $) to detect a 15Sr vs 5%, 20% vs 10%, and 25% vs 15% difference in quit rates was .90, .79, and .71, respectively. Subjects were middle-aged smokers, evenly divided between men and wom- en (Table 2). Almost all of the subjects had completed high school and most were snarried and in professional house- holds esrning more than $30 000 a year. These cha:acteristics were aimilar to those of patients in a smoking cessation clinic; however, the present sample had fewer men, was more educated, and had a higher occupational statua and income than population-based samples of US adult smokers taken in 1985" and 1986 (estimated from tables provided by J. Pierce, PhD, written communication, 1988). Within the sample, a larger pro- portion of subjects in the nicotine gum group were married and had higher in- comes than in the placebo group (P<.05). Subjects smoked about 1.5 packs of moderately low nicotine cigarettes per day and had done so for about 20 yeara (Table 2~ About one third had tried to quit more than three times previously. These smoking habits were aimi7ar to those in the smoking cessation clinic sample; however, the present sample smoked more cigarettes per day of low- er nicotine yield and had tried to quit more often than the estimated popula- tion-based samples of US amokera." Within the sample, smoking habits did not differ between nicotine and placebo groups. The mean FLgerstrom score for dependence was slightly below the cut- off for classi5cation as dependent amok- ers (ie, 7Y and was somewhat less than that seen in smokers attending a amok- ing cessation clinic.° PfOceQur91t A nurse screened potential subjects over the telephone, told them about the study, and gave them a free appoint- ment to aee their physician and herseli At the initial appointment, the study nurse obtained informed consent and showed a 13-minute 9lideltape about the correct rationale, expectationa, and in- atructions regatdiag nicotine gum use. She also gave the subject a booklet of behavioral atrategiea to combat the habit part of smoking, a booklet on the use of nicotine gum, and a list of public and private smoking cessation clinics, including those at the health mainte- nance organirstion. The nurse was trained in smoking cessation counseling and also gave ten to 12 minutes of indi- vidual advice. Subjects then received aix to ten minutes of physician advice JAMA. AAaren 3. 1909-W( 261, No. 9 N=tw Gum in Ga+era( Praetice-(iuphMs et al 130f ~

I I I I I I I I J 1 I I I I I I I ~ (see later herein). The total amount of intervention was 29 to 35 minutes. All 13 physicians at the two clinics agreed to participate in the study. The physicians were board-certi5ed or -eli- gible family physicians. None of the physicians smoked. One had quit amok- ing recently and was using nicotine gum. None of the physicians had prior tsaining in smoking cessation. The phy- sicians attended two one-hour work- shops to learn about nicotine gum, the protocol for brief physician advice, and the rationale and procedures for the study. The protocol for physician advice is described in detail elsewhere.` In the protocol, physicians (1) elicited from the patient a smoking-induced symptomhll- ness/risk (eg, coughlbronc}sitia/caneer), (2) stated its reversibility with cessa- tion, (3) made a personal request for cessation, (4) asked for a firm quit date within the next two weeks, (5) elicited the biggest fear about cessation (eg, ir- ritability, weight gain, or dif8culty con- centrating) and gave advice about it, (6) wrote a prescription for gum, and (7) scheduled a follow-up appointment one to two weeks after the quit date. At the end of the visit, subjects were asked if their physician covered these points. Tbe slideTtrpe seen by the patients pre- sented the guidelines suggested in the package insert (Nicorette); eg, stop smoking abruptly, chew the gum on de-. sire for a cigarette, chew the gum slow- ly, and so on. Subjects were not given an advised amount of gum per day. They were told that common mistakea were to use too little gum, to start use of gum too late during a craving for a cigarette, and to stop use of gum too soon. Sub- jects were advised to use the gum until they had little or no desire for ciga- rettes. The Nicorette package insert is unclear about the recommended maxi- mum duration of use of the gum (both three and six months are impiied?. Sub- jects were told explicitly that if they were using the gum three months after cessation to start tapering use such that they stopped Cun use by the fourth month. Subjects wen aatigned randomly in a double-blind manner to receive nicotine or placebo gum in s 2:1 ratio. The 2:1 ratio was used to incrEase the ability to recruit subjects in that subjects would have a two out of three chance d reeeiv- ing active drug rather than the typical one out of two chance. Subjects were randomized by adding a fourth random digit between 1 and 9 to their three•digit subject number. One exception to ran- domization aas that if two subjects from the same household were entered, the second subject was assigned to the same gum that the first subject received. 1302 JAAAA. Maroh 3.1pe9-WI 261. No. 9 The nicotine gum was the marketed 2 mg dose (Nicorette). The placebo was a gum without nicotine that was flavored to match the taste and irritancy of nico- tine gum.' Subjects obtained gum ad lib at a pharmacy located in each clinic. T1he gum was dispensed from bins marked 1 to 9. Bins 2, 5, and 9 contained placebo and bins 1, 3, 4, 6, 7, and 8 contained nicotine gum. Pharmacista used the fourth digit of the'subject number to dispense the proper gum. Pharmacista were blind to the contents of the diSer- ent bins. All gum was free. One to two weeks after their quit date subjects attended a follow-up session and were counseled by the nurse and physician (six to ten minutes each). The physician asked about arnokfng status, withdrawal symptoms, and side effects. Further treatment was not scheduled. If the physician saw the patient for oth- er reasons, he/she asked about smoking status and gum use and could give more brief (less than ten minutes) advice. Three months after their quit date subjects were sent a letter that re- minded them they should stop gum use and that described four ways to slowly taper use over the next month; howev- er, subjects had acceaa to gum for 12 months. M.aaur.s At baseline subjects completed a de- mographics and smoking history ques- tionnaire that included the Fagerstrom Tblerance Scale' of dependence and the addiction and negative affect subscslea of the Reasons for Smoking Scale." Sub- jects designated an observer (usually a spouse, relative, or friend) to confirm smoking status and gum use. Follow-up forms were mailed to the subjects and their designated obsetwss one week and 1, 6, and 12 months after their cessation date. If these .vere :iot retur:ud in one week, a second form was malled out. If theae wen not te- turned in one week, subjects and ob- sery s were telephoned up to ten timea to obtain the aame information. At the 12-month follow-up, subjects who claimed sbstinence for at least one month were paid =25 to return to the clinic to obtain a breath sample to deter- mine carbon monoxide content and a saliva sample for cotiaine and tbiocya- nate level to verify the observer report and the self-report. Breath carbon mon- oxide was analyzed via an elect:ochemi- al analyzer,' salivs cotinine via gas chromatography (Labstat Inc, Kitche- ner, Ontario),' and saliva thiocyanate via colorimetry (Division of Epide- miology, University of Mianesoti, Mianespolia).° 2646399"7'75 Tb determine gum use the clinic phar- macist recorded the dates of all pre- scriptions. If subjects did not return for gum for six weeks, they were sent a form asking about glun use and side ef- fects while chewing the gum. Data Analysis Dropouts and subjects unable to be reached were counted as srnokers. In- termediate quit rates (one week and one- and six-month follow-ups) were calculated as point prevalence rates; ie, percentage of subjects quit at the time of follow-up. From 50% to 54% of ob- servers were not able to be reached at the follow-ups; thus, at the intermedi- ate follow-ups, subjects who reported abstinence and had missing observer data were counted as abstinent. One-year quit rates .vae calculated both as point prevalence rates and as continuous cessation rates; ie, pereent- age of subjects quit at 1-, 6-, and 12- month follow-ups. The one-year point pre-mlence and continuous abstinence rates wen calculated four ways based on increasingly stringent criteria: (1) self-reported abstinence only; (2) selt- reported abstinence not refuted by observer; (3) self-reported abstinence not refuted by observer and breath car- bon monoxide level less than 10 ppm'; and (4) self-reported abstinence not refuted by observer, breath carbon monoxide level less than 10 ppm, and saliva cotinine level less than 15 nghaL.' For the latter two criteria, subjects who reported abstinence but failed to report for biochemical verifica- tion were counted as amokers. Pharma- cy data indicated three subjects .vere still using nicotine atua at one-year fol- low-up and had saliva cotinine levels greater than 15 nghnL. Cotinine is a metabolite of nicotine; thus, theae ele- vaLed values were assumed to be due to chewing nicotine guns. For these three subjects, a thiocyanate level less than 1.6 mmoVL was used to validate long- term ceasation.' Finally, quit rates us- ing the most stringent eriteria, ie, biochemically verified continuous absti- nence, were statistically adjusted for differencea between the nicotine and placebo groups in marital status and incoa~e. REStlLtS This article focuses on outcome data. Other articles wi'll descnbe in detail the effects of nicotine on.vithdrawal symp- toma and gum use. Smoking Cessation Point prevalence abstinence rates did not differ between groups at the one- week follow-up visit ('hble 3). At the N4w+m cwrt+ n c«»tai Pr.cac.-ha:of+es., a,

I I I I I ~ I I I I I I I I I I I I Tab$.3 -Subt•ets Aasnn.nt a Escr+ Fo+bw•up vlsit by S.M•rsport.na ocs.N.r R•oon chance. Neither total nor average num- No. (%) ot SuDNct. ADsun.nt G+ovp I Wk / rno 6 ne Ly NXmen4 (N.2t0) 156 (751 113 (S4•) 61 (29') aS (21) PsayOp (H - 105) 79 (75) 43 (4 1) 20 (19) 20 (19) •P<.os at PoCOOne Yf p1".oo ey y1. Tabla s.-SubjKts ADstin.nt .t OnaVsu Folbw•up by Incr>tasinqty $trinq.nt Cnl.ria thought they had received accordin g d d h No. (%) of sub{.ats Aa.un.++t .eM~ r,.r.Wrort ~++t eArMr w+oem+. ruo.a s.04.oa+.a aasan.~+o. 40 taal 21 (x) 34 (1g u(ts) A obadevW 46 m) :o (19) r(1s) ,e (15) Pka ftMM R*nMXM WWW <1o ooR, as (1n is (14) 31 (15) »(Iot r,,u e,,,,,,,,,,, <,snpnrnL 25 (t2!) 9 (1) 23 (1is) • (8) rw. ox,fo. +nMUMN WADA " ... ... 21 (10) > m •CrM.na an cums,unvs pra..ow+q oown it,. MWa tPON or.~+i«+u.,aK,+., uo..•..cbony mnwu»on:.nc oa+entOA. a+R n t•. e-..na t9anoran wNO+v up tinckrda snr" suq.ets wt+o a+.r.a Kun. purn. 10t wrcm a hocy.n.a M..* Iw eun t.s nwnYL tw u+Man.a +or Ooar+. ax«u one- and six-month follow-up visits about 10% more subjects in the nicotine gum group were abstinent than in the placebo group (P<.05). Relapse rates were similar between the two groups; ie, of those abstinent at one month the proportion who had returned to smok- ing by one year was similar in the nico- tine and placebo groups (75% vs 79%). At the one-year follow-up session, 260 subjects (83%) supplied self-report data. Of the remainder, 31 (10%) were dropouts and 24 (8%) could not be con- tacted. The major reasons for dropping out were patient request (12), adverse effects (five), pregaancy (four), moved (four), and tried someone else's gum (four). Dropouts and noncontacts were evenly divided between groups and were counted as smokers. With some of the outcome esiteria, the nicotine gum group had quit rates 2% to 5% greater than those of the pla- cebo group (Table 4). None of these dif- ferences were statistically significant using X'teats aith a one-sided probabili- ty value. The nicotine gum group had amre married and high-income subjects, and these two variables were associated with higher quit rates (see later herein). 'Ib adjust for these ditlerences, eight 2 x 2 tables (nicotine gum vs placebo and smoidng vs abstinence) corresponding to the four income and two marital sta- tus groups were generated. The ad- justed quit rates were determined by calculating the cross product ratio from the tables using Maentel-Haenrxl pro- cedures. When these corrections were applied to the most stringent criteria, the adjusted quit rates were 10% for the nicotine gum group and 7% for the pla- cebo group. The 95% con5dence limits for the difference between nicotine and placebo groups with these adjusted quit rates was - 3% to + 10%. Internal Validity Cheoks Eighty-six percent of subjects re- turned for the one- to two-week follo+w up visit and 6% attended a withdrawal clinic or group. Neither of these statis- tia differed between nicotine and place- bo groups. Eighty-five pemnt of sub- jects in the nicotine group and 88% of subjects in the placebo group obtained gum. Fortyfive percent of the nicotine group and 36% of the placebo group ob- tained more than one box of gum (P, not significant). Seventy-six percent of subjects re- ported that physicians implemented all of the points of the brief advice protocol. The average number of subject. per physician was 24.3 (range, seven to 44). Physicians saw an average of 1.2 new subjects per week (range, 0.3 to 2.0 per week). The rate of one-year continuous cessation among each physician's sub- jects varied from 0% to 44%. This varia- tion is not greater than that expected by ber of patients treated per week was related to the -quit rates obtained by individual physicians. Quit rates for the two clinics and for subjects entered in the fust and second halves of the trial were similar. None of the physician behaviors or clinic quit rates differed by gum gror u sp Atte topping use of the gum, subjects were asked which gum they o t to a me escn prevaoualy. b e d Among all subjects, 51% correctly iden- tified their gum, 14% incorrectly identi- fied their gum, and 35% were uncertain of which gum they had received. 'Ihis rateofidentiScationis grwterthanthat expected by chance (P<.001) Physi- ciana could not identify subjects' gum group and pharmacists could not identi- fy which bins contained active gum. Pr.dietors of Outaome The ability of several characteristics to predict outcome (ie, abstinent or amoking) and the ability of nicotine gum to improve outcome (ie, increase in quit rates with nicotine use) were tested us- ing logistic regression. Degree of de- pendence was tested using the FLger- strom Zblerance Questionnaire' and its seven individual items, the addiction scale of the Reasons for Smoldng Scale," and subjects' global rating of dependence. Smoking characteristies tested included rate and duration of smoking, nicotine yield of cigarette, age started amoking, number of prior quit attempts, and duration of longest quit. Demographics tested included age, sex, marital status, income, and education. Other variables tested included amok- ing to relieve negative affect" and belief that one received nicotine gum.' 'Ib test these possible predictors, each variable and gum assignment (nic- otine vs placebo) was entered into a logistic regression predicting long-term abstinence at one year. In this equation, a main effect for the variable would indicate it predicted ouicomc. A signi5- eint interoction between the variable and gum assignment would indicste that the tlternpeuti.c effut oj nieoti"t (ie, difference in quit rates between nie- otine and placebo gum) differed accord- ing to the level of the variable. Among the 23 items tested, 5ve pre- dicted long-term outcome. Smokera who were married, who had h4her incomes, who smoked lower nicotine yield cigarettes, who believed they re- ceived nicotine gum, or who had no his- tory of alcohol problems were more ldce- ly to be abstinent at one-year follow-up (P<.05k In a multiple logistic regres- sion, these items correctly predicted I JAMA, Mareh 3, tM-WI 261. No. 9 rkcotre Gum in Gerwra! Practie.-M+,pt»s et al 1303

I I I I I I I I I I I r I 85% of the cases as either smokers or abstainers. Two items predicted a differential re- sponse to nicotine and placebo gum. Nicotine gum improved the outcome in subjects with a history of alcohol problems (2% quit with placebo vs 10% with nicotine gum) more than those without this history (18% quit with p1a- cebo vs 20% with nicotine gum; for in- teraction, P w.05). A similar trend oc- curred with self-reported "ditficulty refraining from smoking when prohibit- ed," an item on the Fagerstrom scale.' Nicotine gum increased the quit rates of smokers who reported they had difficul- ty refraining from smoking but did not increase the quit rates among smokers who reported no such difficulty (P=.06). Three negative results bear mention- ing. First, several,' but not all,` studies have found that the Fagerstrom scale for dependence predicted quit rates and retponse to nicotine gum. We did not Snd this. Second, a prior study similar to ours reported nicotine gum was most effective in women who were highly de- pendent." We did not replicate this find- ing. Third, in the present study subjects could identify nicotine and placebo gums. SubjectIs belief in which gum they received was related to outcome. However, subjects' belief did not intlu- ence the therapeutic efficacy of nicotine; ie, the difference in quit rates between nicotine and placebo were similar among subjects who did believe they had received nicotine gum and among those who did not believe they had re- ceived nicotine gum. This negative find- ing replicates a similar finding using withdrawal symptoms as the outcome. Side Effects Side effects more prevalent with nico- tine than placebo gum included burning in the throat (76% vs 53%, P<.01), erue- tation (49% vs 21%, P<.01), hiccups (37% vs 14%, P<.01), dizziness (34% va 17%, P<.01), nausea (33% vs 17%, P-.02), stomachache (31% vs 17%, P=.05), and flatulence (31% va 17%, P -.05). Jaw ache was prevalent but did not differ between groups (68% vs 67%). The remaining symptoms of mouth wres, excessive salivation, and vomit- ing were less common (<30%) and did not differ between groups. Among the five subjects who dropped out due to adverse effects, four were in the nico- tine group. Two of these four reported diarrhea and one each reported dizzi- ness, mouth sores, and ulcer symptoms. COMMENT We conclude that, when used with a brief intervention in a general medicsl 1304 JAMA. March 3, tp89-Wt 261. No. 9 practice, nicotine gum use does not pro- duce a clinically or statistically signifi- cant increase in long-term quit rates above that seen with placebo gum use. We believe this conclusion is valid for several reasons. First, the physicians who prescribed the gum and the nurse who advised patients were trained and, according to subjects, implemented the advice protocol. Second, we ensured that all subjects received the same Food and Drug Administration-approved guidelines by requiring viewing of a slidd tape. Third, almost all subjects in the study used the gum and attended the follow-up session. Fourth, outcome was biochemically verified and identical using several different criteria. We also believe our results are generalizable. We used few incluaion/ exclusion criteria. Our sample differed in some respects from the average US smoker, however, our subjects seemed similar to samples of smokers who wished to stop smoking. Also, the inten- sity of our intervention was the maximal feasible for a primary care practice ac- cording to a survey of local practitioners prior to the trial. One possible criticism of our study is that our sample size was inadequate. Several prior studies have reported in- creases of 1% to 7% in quit rates Rith use of nicotine gum (Table 1), and in- creases in quit rates in the present trial were 2% to 5%. Our sample size had inadequate poair to detect such differ- ences as we a priori set our goal to detect a 10% difference. One could argue that with such a small investment of time and such a benign medication such small in- creases in quit rates are sufficient to justify prescribing the gum. We be- lieved that such small increases were not clinically significant given the cost of nicotine gum, its potential side effects (including dependency:'), and the fact that physician advice alone usually in- creases quit rates by 5%." Finally, a survey of 60 primary care physicians in our area indicated that the majority would not use a medication that in- ¢reaaed smoking cessation rates by 5%. Our results are consistent with three of the four prior studies of nicotine and placebo gum in medical practice that reported no advantage to using nicotine gmn'L" (Table 1). Two of these trials used selected populations; ie, pulmo- nary patients' or smokers who failed a prior program." The subjects in the third study were similar to ours, but this study did not have trained physi- cians give a set intervention or provide an in-person follow-up consultation." Thus, compared with prior negative studies, our study used a more general- izable population and a more structured and intensive adjunctive therapy and still found negative results. Our results seem to contradict those of the Stanford placebo-controlled triat, which reported positive results from nicotine gum use and brief advice." However, our six-month point preva- lence abstinence rates (29% vs 19%) are similar to those of the Stanford trial (30% vs 22%). Thus, one possibility is that with one-year follow-up, the Stan- ford trial will show a loss of advantage for nicotine gum as we found. If the Stanford trial does report continued ad- vantage for nicotine gum at one year, then several differences in our trial and the Stanford trial could explain the dis- crepant results. First, the Stanford tri- al excluded 65% of those eligible and included only thoae who could quit smoking for 48 hours. We had no such inclusion/exclusion criteria. Second, the Stanford trial reported point preva- lence abstinence (ie, those abstinent at the follow-up) and did not biochemically verify all abstainers. We rested our con- clusion on continuous abstinence (those abstinent for 11 months) and biochemi- cally verified all abstinence. Third, the Stanford trial used extensive biblio- therapy as its minimal intervention. We used brief physician and nurse advice plus a single booklet. Fourth, the Stan- ford trial took place in a public health setting. We used a general medical practice. Our results also seem to contradict the 5% to 7% increases in quit rates observed in several nicotine gum vs no- gum trials in general medical practice." This apparent discrepancy could be ex- plained by partitioning the effects of nic- otine gum in these studies to placebo effects (ie, a 2% to 3% increase in quit rates) and pharmacologic effects (ie, an- other 2% to 3% increasel There are several reasons.vhy.ve do not believe this partitioning interpreta- tion is correct. First, although nicotine gum has been shown to have pharmaco- logic effects in pLcebo-controlled etials in withdraaal clinics," this does not necessarily mean that pharmacologic ef- fects will occur in general practice. The pharmacologic effects of nicotine may be efficacious only in the more depen- dent amoker,' and such dependent smokers may be relatively uncommon in general practice.` Also, the pharmaco- logic effects of nicotine may be eff ca- cioua only when given along with an intensive psychological therapy. This type of interaction occurs with other psychoactive dru~s' and perhaps with nicotine gum.n" Second, the parti- tioning interpretation assumes our 3% difference in quit rates between nico- tine and placebo gum is valid. We did Nicom,. Gun, in G.nenl Pracbo.-hkupn.s aof

I I i I V t I I I I I I I ~ I not show this in the present study. Whether our small difference would be shown to be statistically significant with larger sample sizes is debatable, espe- cially in the absence of a prior demon- station of an advantage to nicotine over placebo gum at one-year follow-up. Our results have implications both for the present use of nicotine gum and for further tests of nicotine gum in general practice. We have hypothesized that the therapeutic effects of nicotine are Ismited to when the gum is given to de- pendent smokers along with intensive therapy. If this is true, then the utility of nicotine gum in general practices may be limited by the ability to define depen- dent smokers and the acceptability, m3ilability, feasibility, and cost-effer tiveaess of intensive adjunct therapies.' 91.f..noaa 1. f5Ceat:am Y-0: E15aey d nicotine chewing s um: A re.ie., in Pbmerieau OF, Pameritau CS, Facer.uvm K-o, et al (edah Nicotrws Bepiate. nuwt: A Crftim! Evalratio+t. New York, Alan R Lisa Inc,1988, pp 109-12E. t. Lam WL. Sr.e PC, Sacks HS. et al: Heu.analy- .ia of readomirud conuolled trfala d nicotine chewg usg gum. Lancd 1987x27-29. 3. Tbnneson P. fryd V, Hsnsen ld. et al: Effect of aieotine chewint gum in combmuioa with rsoup counielia= in the cae.uiaa d amoidng. N E+Wi J xed 1988;318:15-18. 4. Hughes JR: Problem_s of nieottne sttm. in Ock- ene Jx (edr. PAa+macolopic Treot+neaV of tbboeea Depeedesee: Prxe.dinpe oJ tAs Rvrtd Conprri+, r'or.f0uk, 79f6. Cambridge,lSae, Institute for the Study of SmoMnj Betuvioe snd Pol'uy. 19a6, pp 111-1<7. i. Jatarozik: iC Vessey H. Fowler C. et sl• Coar trolled trial of th:" different antvmolong intet- ventioru in general practice. Br N.d J 198/; 286:1499-1503. i. Faserat:arn K-0: Effect of nieotine &4wins gum and follow.up apyoietmerfts in physieian- based smolotK cessation. Pnr lled 1981;13 517- 527. 7. Page AR, Wahm DJ. SclJesel RP, it aL• Smoh- ini aeaawon in fami)y practice: The edecta of ad- rice and nicrotine chewing Sum pe:eeription, Addie! Behar 1966;11:4434/8. 9. Ruseell HAH, Herriman R, Stupkton J, at s1 EQeet d nieotine cbswins Cum as an adjutxt te teneral praetiaury adviee min.t smoicm. ar Yed J 19ffir;67:1?E2-1785. 9. Sutton S. Halktt R: Rudoenised triaL of brief indiiidual usitmeat !or emokint wint nieotine gum ia a workplace aettin* Awt J PrOiic HeaW 19i7:77:121a1211. lo. Sutton S. Hatktt R: Smoidns iaterveetion ie tAe.rorkplsce nsiM.ideee.pn.ad aicatiae ebe.- mi tum, Prer Med. ia press. i1. Wilson Dk, Taylee W, Gilbert JR, et ak A tandomised trial d s tamity pdyeieian iaterrentiaft loratnoking ceaestiea JAMA 198l26P1S7o-lb7< lz British Thoracic Sooety: Compar;eon d 1'ar eethads of sawicng .ithdrawpi in patients witb ea+ekinj-reLted dieeaiee: RepoFt by a eubeanmit- tee d tbe ReeeareA Comrnittee d tAe Briti.h Tla z7eic Soeiety. Brlled J 19aS?a6396•694. lt CiepbeU TA, Lroer E. Prescott RJ: Stopping For example, in the present study as well as others,' few smokers (<6%) seen in general practice were willing to at- tend more intensive interventions even when offered free of charge. T'hus, it may be necessary to design medical practices and future studies so that de- pendent smokers can be identified and successfully persuaded to enroll in psy- chological therapies for smoking cessa- tion before nicotine gum use can be shown to be effective in medical settings. In sumrnasy, based on our results and those of three of the four prior placebo- controlled trials,'"11 we conclude that, when used with a nonselected group of smokers and a brief intervention in a general medical practice, the pharmaco- logic effects of nicotine gum (Z mg) to emoidnr Do nieotsne ehe.in< <uat and postal es- couratement add to daetors~ advice? Pr+oetusowe+ 1987;231:114-117. 14. Haaeldewia AS, Blais LW, Sansone C. et a1• Nicotine gum and ael4help manuaJs in amoldng aseation: An evaluatioa in a medical contasL Ad- diet Behar, in pnw. li. Shaughnesay AF. Davis RE, Reeder CE: Nieo- tine cheRinj gum: Effectiveness and the laAuenee d patient eduntion in a family p:actice. J fb+R P+oe! 198725:2btr269. 1i. Fortmam SP, Killen JD. Tekh kJ. et aL- Hiai- mal contact treatment for emokiag eeaaation: A plaeebo-eonttolkd trial of nicotine potaerilex ard ulf-dieected relapse pte.vntion: Initial rceultv of the Stanford Stop Saalcng Projeet IAJfA 1968;280:1575-1680. 17. J.nuozic K. Fowler G, Vea/ey M, et a1• Place- bo controUed trial dnicotirx chewing rum in 0ener a1 practice. Brlfed I 19b4;289:79i-797. 18. HufheaJR, idiilerS: Nieounegumtabelpatop smoking. JA.YA 198/;252=14r285d. 19. 'henda in tobacco use in the United Statea: Appendix A. iw TAe H.oitA Co+wqreraca q/S+nab inp: Nicotine Addittiat: A Report q(f+1e SY+9ewt Cene+nl. Dept of Health and Human Services, 199. pp 561-6dE. fl~esetsots K-0: )(eaenrinrdesree dphyaieal 20. deptndenee to tobaeao smoking with reference to iadiridualvstiaa of treatment. Addid 8ds. 19713215-?Al. i1. Hushea JR, Husukami DK, Pickeee RW, et ak Effect of nicotine oe the tebaece .itAdra.al ern- droiet. PeyulopJ+ar+nacolr~py 196/:31:32a7. tt. Hughes JR. Kattks T: Doctors helping amolo- ers: Real world taetin. Xiwx Med 1986:69•293-296. ?3. Hughes R, Piekens RW, Spring W, et ar lA- wvctioeaa control whether nieotine will sa~t as a eeinfaset J PAa*rraeol Erp 7Tur 19AS=106 l1i f4. Ikard FF Green DE, Heea D A sde to dffa,- erttiate bet.een types d aaoldag as related to arnagement d alleet. lnt J Addiet 19ft.d:6i9-M. SS. Hughes JR, Frnderi7uea LW, Faoer H: A earboa saoewade analyser ter esasuren" d enmidns beha.iee: BrAar 7Aer 19?6,9r290-296. Si. Jeeob P, Wibae 1[. Bene.its NL: Nicotine aed eet&rire determinations in biolWiul fluids: An in- pevred Cu eluonrtogysphie methed. J Ckrwws• topr 1981=241•70. increase amoking cessation are either small or nonexistent. This conclusion is important because the large majority (>99%) of smokers are not screened for nicotine dependence and receive nico- tine gum with a brief intervention.` Tbis study was Au+ded by a grant (DA-040a6) uW a Research Scentiat Development Award (DA- 00109 to Dr Hughes) from the Nationallnesitvte oa Drug Abuse. Washinoon, DC. Herteu-Dow R.- uatels lnn5tute, CirutcruuLi, prwided all gum. We thank the physicians, pharafscieta, aed eta8 of the Plymouth and Main cliniq d tbe Park-Nieoi- 1et ldedinl Certter. Minneapolia, tor tbeir belpful comments and exemplary conduct of the etudy. We also thank Warren Bickel, Brian Flynn, Stawe Hit- Ona, and Rom SeekerWalker for their comments on the aunusafpt and John Pierce, PhD, dtbe US ol6ce oa Smoking end Hea1tA, Roekrflh.l+Ld, !or p:vviding data ttom which rswlts far the aratsge US anoher.a. estimated. 27. Dereen Pk. Davido. B. Bar HE, et aL• A ehemiW test for amolnni exposure. Arch Es.i+oR XealtA 1967:11:B66-S74. !a!. J.rrir YJ, Tuast.U-Pedoe H, f17er+bead C, et a1• Corsperiaon dtseto ased to dtsita;uis6 amok- en from nonsmokers. Am I Pullie Health 19d7;77:1435-143L 29. Husbea JR. Rmlm 1? Blindi+e.a Wd the w4d- ity of double-blind drui trials. J Clis PeyeMerAier. mace1196S;S:1J&1d2. 30. Hall SM, Tunstall C. Ginsbetg D; et al: Niee• tine gum and bebarioral treaueeat A piaeebo-eea trvlled trial. I Coa." Clis PrycAoi 1967Ab.i03- 3~Jaekson PH. Stapietoa JA. Ras.e311LA8. et al: Predictors d auteome in sgerKral prattitiamer in- tt:.esstioe agsitut emo. Pi.r Y.d 19e6i16Zf4- 32. Hajek P, Jackson P, Bekbtr k: Leeg-f.ra aee of .ieotine ebewing Vtn: Oeeursesce, deteVi- nanta, and effect on weight gaia. JAMA 19eDOSD:IS93159i. 33. Hughes JR: Dependence potertia! aed abre liaWy of aicotine replacement tberapw, ia Pr merleau OF, Pomerleau CS, F*erstnm K-0, et al (edak Nieoline Repiacewwwt: A CrUieel Ewlns- tiwn. New York. Alan R Liea iae,19K 9pl61-2'J3. S1. RuseeC kAH.lYilaon C, Tayloe C. et al Eaect af peattitioe*es adriee alsitvt smoiong. B*Y.d J 197i-.2:Z11-ZlS. ii.- Russell 1(AH. Pet J, Patel UA: Tise eLrilk.- tioa daaoiant by teetora! etnad~e d eativ.e, JR Stat Sae 1971;13731ii83. si. Hollon S. Beek AT. Paydtetberapy aad dM 5e~ 1d~SL, Berge ~(ede): H hyeA& tAe.eplr ewd BsAavio. CAawpe: An rwtpirie.l Awslyeia New Yerk, Jo+ha R'akr t Soaa Le,197i, 3~7.~Sebneides NG, Jeraic YE. lbnrtie Al,.t aL Nicotine jwn in smoking ee.eation A piaeebo tea- troUed, dwbie-bliad <si.L Addiat B.4ss 1flS; t?51-26L 3{. Jervia il: Nicotine replacement r eoie tJmspy ee as adimet, in Pbmerksu OF. Peme:l.nt C8, Farntraa 1C-0, et al (adak Nieoti*e Rept.ee- ewn+t: A Critiest Reti.v. New Yeek, AL. R Liu Iuc,19Bl, op 10-1LL .atMA,l1tarett 3. 1989-WI 261. No. 9 N*Xr*Gum in Gef»tai Pnetie.-IiupMs.t aa 1306 ! ' .