Reducing Tobacco Use reimbursed for clinical and pharmacologic treatments to help patients quit smoking (Group Health Associa- tion of America, Inc. 1993; Schauffler and Parkinson 1993). Appropriate reimbursement may be essential to ensuring greater clinical attention to tobacco addic- tion (Schauffler and Parkinson 1993; Fiore and Baker 1995; Kaplan et al. 1995). The Public Health Service-sponsored Clinical Practice Guideline Trenting Tobncco Use and Dependence has recommended that health care professionals use the "five A's" to help their patients quit smoking: (1) nsk about smoking, (2) ndvise all smokers to quit, (3) address willingness to make a quit attempt, (4) assist patients who want to quit, and (5) arrange follow-up visits (Manley et al. 1991; Glynn and Manley 1993; Orleans et al. 1993; Houston et al. 1994; Fiore et al. 2000). These recommendations, based on a comprehensive review of the empirical literature, constitute a proscriptive algorithm for clinical inter- ventions (see the text box). Additional follow-up visits, at increasing inter- vals, with patients who continue not to smoke have been associated with greater long-term abstinence (Kottke et al. 1988; Wilson et al. 1988; Orleans et al. 1991). Patients who have relapsed should be helped to quit again at follow-up visits and subsequent visits. The Five A's T o help their patients quit smoking, clinicians can use the "five As" approach: (1) rusk patients about smoking, (2) odsjise all smokers to quit, (3) ns- scss willingness to make a quit attempt, (4) nssist those who want to quit, and (5) arrnnge follow-up visits with those trying to quit (Glynn and Manley 1993). These brief clinician interventions, which are described in this text box, can be completed within two to three minutes at each visit and have been associated with a cessation prevalence of 5 percent (Glynn 1988) to 8 percent (Kottke et al. 1988). All patients seen in a primary care setting should be routinely asked about their smoking sta- tus. One means of institutionalizing the identifica- tion of smokers is to expand the vital signs to include smoking status (Fiore 1991). Another means is to use stickers or other markers to clearly identify charts and prompt clinicians to help their patients who smoke quit (Cohen et al. 1989b; Ockene et al. 1991). All patients who smoke should be advised to quit. This advice should be clearly stated and per- sonalized. After giving this advice, clinicians should assess whether smokers desire to quit at the present time. Clinicians should provide motivational ma- terials and messages to those not willing to quit. These patients should be asked about smoking and advised to quit at all subsequent visits. Clinicians should assist patients who want to quit. The clinician should work together with the patient to set a date to quit (preferably within two weeks of the clinic visit) and should provide the patient with practical advice about how to quit and self-help materials. Clinicians should determine whether the pa- tient is likely to require adjunctive help and whether the patient is a candidate for pharmacotherapy. Pharmacotherapy should be considered for all pa- tients motivated to make a quit attempt, except in the presence of specific contraindications (Fiore et al. 2000). The choice may take into account previ- ous patient experience, preferences, and other fac- tors (see "Pharmacologic Interventions," later in this chapter). Clinicians should also present other treat- ment options to their patients who want to quit. In particular, patients should be made aware of com- munity cessation resources (such as those offered by the American Cancer Society and the American Lung Association) and of intensive clinical inter- ventions (see "Intensive Clinical Interventions," later in this chapter) available in the community. The primary care clinician, however, should con- tinue to monitor and assist those patients who elect to undergo intensive treatments. Clinicians should arrange for a follow-up visit to discuss smoking cessation within two weeks of the chosen date to quit. Researchers have docu- mented that scheduling follow-up visits or making follow-up telephone calls improves cessation suc- cess (Kottke et al. 1988; Wilson et al. 1988; Ockene et al. 1991, 1992; Orleans et al. 1991). Follow-up visits should be arranged whether the patient has been referred to another clinic or treated by the pri- mary care clinician. Mamgcnw~~t of Nicotine Addiction 203 Surgeon Gerleral's Rrport Modifications in treatment, including a discussion of more intensive efforts, should be considered for relaps- ing patients at each iteration. An area of current active research in minimal in- terventions is the use of computer-tailored messages for individual smokers who want to quit. Computer software that approximates deductive or inductive human reasoning has been proposed as an efficient and cost-effective mechanism for this modality (Velicer et al. 1993). In a large trial of one such system, interac- tive computer reports plus individualized manuals produced higher current abstinence (20 percent) and prolonged abstinence (11 percent) than did standard manuals, individualized manuals alone, or personal- ized counselor calls (Prochaska et al. 1993). Similarly, analyses of two separate controlled trials found that computer-tailored letters generated significantly greater cessation proportions in groups receiving them than in control groups (Strecher et al. 1994). Although these mechanisms have not been extensively evaluated, they are a promising avenue for further investigation. Efficacy Kottke and colleagues (1988) performed a meta- analysis of 39 smoking cessation trials conducted in medical practice settings. Most of these trials involved relatively minimal interventions, but some more in- tensive treatments were included. Participants had a mean of 4.8 (standard deviation = +4.4) contacts with these clinic-based programs. The major conclusion of this analysis was that success increased with the num- ber of intervention modalities employed, the number of health care professionals involved in the effort, and the number of follow-up assessments. Duration of follow-up (as opposed to number of follow-ups) was not predictive of success. Using diverse techniques may be a key characteristic of successful clinic-based smoking cessation programs (Fiore et al. 2000). A suc- cessful program might be one in which face-to-face counseling or advice is given; dates for quitting are set; pamphlets are distributed; reminders by telephone are made; smokers are advised and counseled on quit- ting by physicians, nurses, and other health profes- sionals; and multiple clinic visits or telephone calls are made after the smoker's quitting day. In the meta- analysis by Kottke and colleagues (1988), cessation assistance delivered by nonphysicians tended to be slightly more effective than that performed by physi- cians, but a more recent meta-analysis (Fiore et al. 2000) found no difference in effectiveness between physi- cians and nonphysicians. Both individual and group counseling was effective (Fiore et al. 2000). The meta-analysis by Kottke and colleagues (1988) also suggested, however, that complex interventions are not necessary for clinic-based success. Compared with smokers who received no assistance, smokers who received help consisting of advice only or brief coun- seling had a 13.1-percentage point increase in cessation 6 months after treatment and a 3.8-percentage point in- crease after 12 months. Comparable estimates for smokers whose only treatment was to receive written self-help materials from health care professionals were 1.6 percent at 6 months and 2.0 percent at 12 months. The impact of brief intervention is illustrated in one study by Russell and colleagues (1979), who found that providing advice in a primary care setting produced a biochemically confirmed increase in abstinence of 3.3 percentage points; when smokers were told they would be followed up and when self-help materials were distributed in conjunction with the advice, the resulting one-year increase in abstinence was 5.1 percentage points. Trials postdating the meta-analysis of Kottke and colleagues (1988) have also indicated that brief clini- cal interventions have a small but reliable impact on smoking cessation success (Cummings et al. 1989a; Risser and Belcher 1990; Taylor et al. 1990; Ockene et al. 1991, 1994; Weissfeld and Holloway 1991; Hollis et al. 1993; Strecher et al. 1994). A meta-analysis of seven studies found that physician advice to quit increases cessation by 30 percent (Fiore et al. 2000). The consis- tency of these findings over a considerable time span and in multiple settings lends credence to the useful- ness of minimal interventions. Smokeless tobacco use may be particularly ame- nable to minimal clinical interventions, especially in dental office settings. Oral lesions caused by smoke- less tobacco are quite common among users of these products (Ernster et al. 1990; Tomar et al. 1997) and provide the opportunity for the dentist to point out the direct adverse health effects of smokeless tobacco. Several trials have examined the efficacy of minimal clinical interventions in smokeless tobacco cessation. In a randomized trial conducted in a dental health maintenance office clinic to test a minimal clini- cal intervention, Stevens and colleagues (1995) re- ported significantly higher smokeless tobacco quit rates in the intervention group than in the usual-care group at both 3 months (32.2 vs. 21.3 percent) and 12 months (33.5 vs. 24.5 percent). In a randomized clini- cal trial conducted in private dental offices, Severson and colleagues (1998) also found that a minimal inter- vention significantly increased smokeless tobacco quit rates in the intervention group compared with rates in the usual-care group at 3 months (17.8 vs. 8.8 IO4 Chapter 4 Reducing Tobacco Use percent) and 12 months (10.2 vs. 3.3 percent). A mini- mal intervention trial for smokeless tobacco use among college athletes, which included dental examinations to demonstrate oral lesions, 15-20 minutes of counsel- ing by dental hygienists, and follow-up telephone calls, found that three-month biochemically assayed quit rates were 24 percent in the intervention group and 16 percent in the control group (Masouredis et al. 1997). Relevant Process Measures Although minimal clinical interventions provide smokers with some practical advice about quitting, their primary purpose is to increase smokers' motivation to quit. Specific process measures-such as measures of this motivation-are seldom incorporated into minimal clinical interventions. The nonspecific measures some investigators use do not associate clinical success with changes (such as greater awareness of disease risk or enhanced belief in one's ability to quit). Nonetheless, the available evidence suggests that minimal clinical interventions can enhance smokers' desire and inten- tion to quit (Russell et al. 1979), decrease the number of cigarettes smoked per day (Folsom and Grimm 19871, and increase the number of attempts to quit smoking (Folsom and Grimm 1987; Cummings et al. 1989b; Strecher et al. 1991). In addition, patients have reported that physicians trained to perform more intensive in- terventions are more helpful than physicians without such training (Ockene et al. 1991). Summary Substantial evidence suggests that minimal clini- cal interventions (e.g., a health care provider's repeated advice to quit) foster smoking cessation and that the more multifactorial or intensive interventions produce the best outcomes. These findings highlight the im- portance of cessation assistance by clinicians, who have a unique access to more than 70 percent of smokers each year. Moreover, minimal clinical interventions have been found to be effective in increasing smokers' motivation to quit and are cost-effective (see "Cost- Effectiveness," later in this chapter). However, re- search has not clarified fully the specific elements of minimal interventions that are most important to clini- cal success nor the specific types of changes they pro- duce in smokers that lead to abstinence. Intensive Clinical Interventions Intensive clinical interventions (sometimes called "formal" or "organized" cessation treatments) are multisession counseling programs involving extensive contact between a health care provider and a smoker. The value of intensive interventions has been ques- tioned because they are more expensive and reach fewer smokers than self-help and minimal clinical in- terventions do (Chapman 1985). However, more in- tensive interventions continue to attract interest because they are more successful at helping people quit smoking (Schwartz 1987). Despite their comparatively high cost, they are cost-effective (Elixhauser 1990), and they may be especially well-suited for treating the most addicted smokers (Lichtenstein and Glasgow 1992; Orleans 1993). Intensive clinical interventions may be charac- terized by structure and content. Structural variables include providers' credentials and training; individual, telephone, or group format; session length; total num- ber of sessions; and duration of follow-up. Relatively little research into intensive treatments has been de- signed to assess the effects of different structural vari- ables (Lichtenstein and Glasgow 1992). Increased patient contact results in better outcomes (Land0 1981; Decker and Evans 1989; Lichtenstein and Glasgow 1992; Fiore et al. 2000). In a meta-analysis of research on the nicotine patch (Fiore et al. 1994~1, researchers found that the following counseling features were as- sociated with significant increases in six-month absti- nence rates: counseling being a main reason for clinician-patient contact, at least weekly clinician- patient meetings during the first 4 weeks of treatment, and more than six clinician-patient meetings in the first 12 weeks of treatment. A more recent meta-analysis that was not restricted to nicotine patch studies (Fiore et al. 2000) found that quitting success increased with increasing contact time (up to 90 minutes of total con- tact) and that there was a dose-response relationship between number of sessions and treatment efficacy (Fiore et al. 2000). Thirty to 90 minutes of total coun- seling and four or more sessions were two to three times more effective in producing long-term smoking cessation than no contact controls. This research sup- ports the notion that in general, as the intensity of clinician-patient counseling increases, so does the long- term effectiveness of treatment. Because so little information is available on how structural variables affect intensive treatment outcomes, this section concentrates on a review of con- tent variables. Content refers to the specific informa- tion, materials, and techniques to which smokers are Marlagement of Nicotine Addiction 105 exposed during the course of treatment. The various contents of intensive smoking cessation interventions are not easy to evaluate, partly because the method- ological quality of clinical trials tends to differ across content areas. For example, trials of relatively unor- thodox treatments, such as acupuncture and hypnosis, tend to use shorter follow-up periods than assessments of efforts involving pharmacologic and behavioral treat- ments (Schwartz 1987; Ter Riet et al. 1990); inflated effi- cacy estimates may thus result for unorthodox treatments. These methodological concerns are handled here by limiting the review primarily to studies report- ing outcomes with at least five months of follow-up. Another problem in evaluating the content of intensive interventions is that the evolution of treat- ments over the past 40 years prevents a cumulative assessment of specific intensive interventions. More- over, changing research interests and methodologies make it difficult to integrate findings from over the entire period. For instance, pharmacotherapies have changed greatly during this period and are now in- corporated routinely into intensive treatments. In ad- dition, treatment response may be affected by changes in the nature of the smoking population; for instance, compared with 40 years ago, a higher proportion of today's smokers are women. Methodological and sta- tistical changes have also altered the nature of the stud- ies themselves: sample sizes are larger to increase statistical power, and biochemical confirmation of ab- stinence is now routine, as is the application of the "intent to treat" principle in analyses. Because of these refinements, early cessation research is now often ne- glected, perhaps because it is difficult to integrate with newer work. On the other hand, some apparently ef- fective methods, such as rapid smoking, have often not been evaluated by newer methods. The older lit- erature on such strategies is included selectively in this review. A related problem, complicating the interpreta- tion of relatively recent research, arises from what Lichtenstein and Glasgow (1992) have referred to as a shift from a "clinical" to a "public health" (p. 518) ori- entation among smoking cessation researchers. This shift has resulted in a dearth of theory-driven research into intensive interventions. In fact, one observer has suggested that the long-term research trajectory favors modifying established models over applying innova- tion in the basic approach to treatment (Shiffman 1993b). Recent emphasis on public health has also produced a research climate that favors the evaluation of treatment packages and minimal interventions over treatment components (Lichtenstein and Glasgow 1992). One reason for this shift is the high cost and large sample sizes required to evaluate individual com- ponents. Thus clinical trials rarely allow assessment of a given treatment's independent contribution. Smoking cessation trials now tend to combine specific treatment components into multicomponent interven- tions. Moreover, within the same study, not only may groups receive different treatment packages but the packages may differ in their structural components. Finally, the question of selection bias remains a challenge to interpreting the literature on intensive interventions. Investigators typically recruit highly motivated volunteers to serve as subjects, because the efficacy of intensive interventions can be tested only if the patients under study actually receive the entire treatment. Efficacy estimates derived from this atypi- cal population may not be appropriate for making pre- dictions about the larger population of smokers. The principal types of intensive interventions must be evaluated in the context of these limitations stemming from the nature of the available evidence. Problem Solving/Skills Training Various strategies try to impart to smokers the knowledge and skills necessary to cope with cessation- that is, both to attain and to maintain abstinence when confronted with withdrawal symptoms or the temp- tation to smoke (Marlatt and Gordon 1985; Curry and McBride 1994). This approach (hereafter referred to as problem solving/skills training) springs from the observation that most relapse efforts seem to be asso- ciated with a finite number of factors, such as alcohol use, negative affect (e.g., depression), and the presence of others smoking (Shiffman 1982; Baer and Lichten- stein 1988; Brandon et al. 1990). Problem solving/skills training tries to help people who have recently quit smoking anticipate these "high-risk" situations and learn to cope with them when they arise. Such inter- ventions also train participants to cope with with- drawal symptoms, replace positive reinforcements they had linked to smoking, and meet other challenges that might be encountered during or after an attempt to quit smoking. General problem solving/skills training targets challenges that occur early in the quitting process (e.g., withdrawal discomfort). Because newly abstinent smokers often return to regular smoking (Curry and McBride 1994), one specialized type of intervention teaches skills to help the former smoker maintain ab- stinence (Marlatt and Gordon 1985). These interven- tions also train former smokers to prevent any relapse from becoming a long-term return to smoking. Former smokers are encouraged to view relapses as a normal 106 Chapter 4 Reduciq Tobacco USC part of the quitting process rather than as an indica- tion of failure (Curry et al. 1988). Another type of problem solving/skills training focuses on coping with the immediate negative affects of quitting smoking. The growing body of research on dysphoria (feeling unhappy or unwell) after smok- ing cessation (Glassman et al. 1988; Covey et al. 1990; Brandon 1994; Hall et al. 1994) suggests that strategies that help smokers who have just quit resist negative moods may be particularly successful (Shiffman 199313). However, a recent meta-analysis (Fiore et al. 2000) did not find that interventions that targeted nega- tive affect improved cessation rates. These interven- tions were used with the general population as well as smokers with a history of depression. It is possible that the results might be more positive if the studies were restricted to high-risk populations. Efficacy Because nearly every state-of-the-art smoking cessation program contains elements of problem solv- ing/skills training (Curry and McBride 1994), the tech- nique is difficult to assess as an individual treatment. Some investigators have failed to uncover evidence that this technique increases cessation success relative to comparison groups (Curry et al. 1988; Emmons et al. 1988; Omenn et al. 1988; Minneker-Hiigel et al. 1992; Zelman et al. 1992). Other studies have found benefi- cial effects, but these benefits have often been modest and have come only through protracted treatment (Hall et al. 1984b; Davis and Glaros 1986; Goldstein et al. 1989; Stevens and Hollis 1989). Even in studies that report success in long-term abstinence through skills train- ing, the overall relapse curves for treatment subjects have paralleled those for comparison groups (Glasgow and Lichtenstein 1987; Goldstein et al. 1989; Stevens and Hollis 1989; Mermelstein et al. 1992; Minneker- Hiigel et al. 1992; Gruder et al. 1993). A recent meta- analysis (Fiore et al. 2000) of 104 studies, however, reported that problem solving/skills training increased quitting success by 50 percent. Some evidence sug- gests that problem solving/skills training may be par- ticularly useful for female smokers (Curry et al. 1988), those who smoke fewer cigarettes (Hall et al. 1984b), those who smoke to cope with emotional stress (O'Connor and Stravynski 1982), and those who are less prone to negative affect (Zelman et al. 1992). Although multicomponent skills-training programs have sometimes included information about managing the dysphoria associated with smoking ces- sation (Tiffany et al. 1986; Kristeller et al. 1993), relevant behavioral interventions have only recently begun (Hall et al. 1994). Initial results suggest that such strategies are promising, but these findings re- quire replication and extension. In sum, the evidence on problem solving/skills training suggests a beneficial impact (Fiore et al. 2000). Such training can offer practical strategies about quit- ting and inculcate desired coping skills. Releuallt Process Measures Skills training rests heavily on two assumptions: (1) coping skills will help former smokers remain ab- stinent in the face of temptation, and (2) smokers can be taught these skills. Some cross-sectional research (Shiffman 1984) and skills-training intervention trials (Hall et al. 1984b; Davis and Glaros 1986; Zelman et al. 1992) have suggested that coping strategies help avert relapse. The available evidence also indicates that patients given skills training acquire coping skills (Hall et al. 1984b; Davis and Glaros 1986; Zelman et al. 1992), and there is evidence that the level of skill acquisition predicts long-term abstinence (Zelman et al. 1992). Although the results of one trial suggest that coping skills are not retained for very long (Davis and Glaros 1986), consistent self-monitoring of smoking during treatment is associated with longer-term maintenance (Kamarck and Lichtenstein 1988); this finding suggests the importance of behavioral characteristics that fos- ter maintenance. One of the goals of skills training is to encourage relapsed former smokers to renew their efforts to quit smoking. Curry and colleagues (1988) found evidence that smokers who had received skills training were more likely to try quitting again if they relapsed. Rapid Smoking Rapid-smoking strategies typically require that smokers inhale deeply from a cigarette about every six seconds until they become nauseated. In theory, this aversive conditioning transforms the subject's perception of smoking from a pleasurable activity into an unpleasant one, thereby making it easier for smok- ers to give up cigarettes. Medical complications produced by rapid smok- ing can include elevations in heart rate, blood pres- sure, and carboxyhemoglobin blood levels as well as electrocardiogram abnormalities (Horan et al. 1977). Because of these potential problems, candidates for rapid smoking should be selected carefully (Lichtenstein and Glasgow 1977). Older persons and persons with cardiovascular or pulmonary conditions are generally excluded from rapid-smoking strategies, Surgeorl Grrwr~l's Report but some evidence suggests that rapid smoking can be conducted with these persons if appropriate pre- cautions are taken (Hall et al. 1984a). Efficacy The 1988 Surgeon General's report on smoking and health (USDHHS 1988) reviewed the literature on rapid smoking and reached two conclusions: (1) al- though its effectiveness is variable when used alone, rapid smoking yields moderately high long-term ab- stinence success (40 percent of subjects were abstinent 6-12 months after treatment) when incorporated in multicomponent behavioral interventions, and (2) aux- iliary treatment factors, such as patient expectations, patient-therapist rapport, and admonitions not to smoke between sessions, can influence how success- ful rapid-smoking strategies are. Few rapid-smoking trials have appeared since the 1988 report. The mid-1980s advent of pharmacologic treat- ments for smoking cessation greatly reduced research interest in rapid smoking. Pharmacologic aids, such as nicotine gum, appear as efficacious as rapid smoking (Zelman et al. 1992) and are probably more acceptable to smokers and program administrators. Nonetheless, the doubling of long-term success associated with rapid smoking (Fiore et al. 2000) suggests that it may remain an option for smokers who are unable to quit through other methods and for whom such aversive condition- ing is acceptable. Relevant Process Measures Rapid smoking is intended to produce aversive conditioned responses to stimuli associated with smok- ing (USDHHS 1988). The technique reliably produces tachycardiac responses to cigarettes, and the magnitude of these responses is directly related to treatment out- come (Tiffany et al. 1986; Zelman et al. 1992). More eas- ily observable variables, such as the number of cigarettes smoked during a rapid-smoking session or the degree of nausea reported by patients, have not been shown to be consistently related to outcome (USDHHS 1988). Other Aversive-Smoking Strategies Three other techniques intended to produce aver- sion to cigarettes have been investigated: satiation therapy, rapid puffing, and focused smoking. Con- cern over the safety of rapid smoking (Horan et al. 1977) was partly responsible for investigation of these alternative aversion techniques. Some evidence sug- gests that they are less unpleasant and less risky than rapid smoking (Glasgow et al. 1981; Tiffany et al. 1986). Satiation therapy requires that patients smoke many more cigarettes per day than they normally do, usu- ally about twice as many (Best et al. 1978). Rapid puff- ing is similar to rapid smoking, but patients are instructed not to inhale cigarette smoke (Tiffany et al. 1986). Focused smoking requires patients to smoke for an extended period of time at a normal rate while concentrating on the negative sensations smoking pro- duces (Lowe et al. 1980). Efficacy Satiation therapy alone produces relatively little cessation success (15 percent at one year) (Land0 19821, but the technique may be more effective when incor- porated into multicomponent programs (USDHHS 1988). Focused smoking and rapid puffing produce long-term abstinence rates that are equivalent to, or slightly lower than, those produced by rapid smoking (USDHHS 1988; Fi;rp et al. 2000). Because these tech- niques do not appear to result in significant tachycar- disc responses (USDHHS 1988), their efficacy is probably accounted for by mechanisms other than aversive conditioning. Cue Exposure Cue exposure therapy is based on the premise that smokers become conditioned to certain cues or contextual signals correlated with smoking behavior. When persons who have recently quit smoking are exposed to these cues, they are motivated to begin smoking again (Rohsenow et al. 1990-91; Brandon et al. 1995). In cue exposure therapy, persons trying to quit smoking are repeatedly exposed to these signals in a therapeutic context in which smoking is prohib- ited; the resulting reduced association between smok- ing and previous cues is hypothesized to reduce some of the temptation for relapse that former smokers will face in the natural environment. Because cue exposure therapy has produced promising results with other addictive disorders (Monti et al. 19931, several researchers have suggested that such strategies be developed for smoking cessation (Hodgson 1989; Heather and Bradley 1990). These strategies may be particularly important for women, whose responsiveness to nicotine replacement therapy appears to be less than that of men (Perkins 1996). Women may be less controlled by nicotine and more influenced by nonnicotine factors (sensory stimuli, en- vironmental factors) (Perkins et al. 1999) and may there- fore respond better than men to behavioral approaches. 108 Chyter 4 Efficacy Studies conducted to date that have evaluated cue exposure have failed to find significant differences in outcome between cue exposure and comparison interventions (Lowe et al. 1980; Raw and Russell 1980; Gtitestam and Melin 1983; Corty and McFall 1984). However, clinical research on cue exposure for smok- ing cessation is sparse, and interpretation of most ex- isting trials is hampered by methodological flaws (Brandon et al. 1995). Relevant Process Measures Environmental associations with cigarette smok- ing can be strong enough to provoke the desire to smoke (Herman 1974; Rickard-Figueroa and Zeichner 1985; Tiffany and Hakenewerth 1991). These provoked responses may affect treatment outcome (Niaura et al. 1989). However, because cue reactivity has not been assessed in existing clinical trials of cue exposure therapy, it is impossible to determine whether such interventions extinguish motivational responses to smoking-related cues. Nicotine Fading Nicotine fading is based on the assumption that withdrawal symptoms will be lessened through a gradual reduction of nicotine intake (Foxx and Brown 1979; McGovern and Lando 1991). Nicotine fading can be accomplished either by progressively switching to brands of cigarettes yielding less nicotine or by using a series of graduated filters (McGovern and Lando 1991). Once the lowest nicotine level is reached, ces- sation is attempted. Nicotine fading should be distin- guished from cigarette fading, in which the number of cigarettes smoked per day is gradually reduced. Cigarette fading has generally not been shown to be an effective smoking cessation technique; participants generally reach a level beyond which they find it diffi- cult to reduce cigarette consumption (Land0 1993; Fiore et al. 2000). Efficacy Foxx and Brown (1979) reported that 4 of 10 sub- jects who tried nicotine fading had quit smoking at 18 months, but subsequent investigations have found more modest long-term results (usually around 20 percent) (Beaver et al. 1981; Lando and McGovern 1985; Burling et al. 1989). Some evidence suggests that nico- tine fading can increase abstinence success indepen- dently within a larger smoking cessation program (Burling et al. 1989). In a community setting where participants were allowed to select their treatment, about 25-30 percent of those who chose multicompo- nent interventions containing nicotine fading achieved long-term abstinence (Land0 et al. 1990; Lando 1993). Brand switching and graduated filters have produced equivalent outcomes (McGovern and Lando 1991). Cinciripini and colleagues (1995) found that 44 per- cent of persons using a combined nicotine fading and skills-training package were abstinent from nicotine at one year, a proportion significantly higher than that produced by matched conditions. Relevant Process Measures Nicotine fading is presumed to exert its effects by gradually weaning smokers from nicotine, thereby reducing withdra>val symptoms. Reductions in nico- tine intake and Mithdrawal indexes are thus the pro- cess measures of primary importance to nicotine fading. One early study suggests that nicotine fading reduces the severity of withdrawal symptoms (West et al. 1984a,b). The process measure of reduced nicotine intake is problematic, because smokers' nicotine consump- tion seldom matches a given brand's machine-rated nicotine yields (McMorrow and Foxx 1983). Smokers are able to compensate for reduced nicotine yield by adjusting how they smoke-by inhaling more strongly, holding smoke in longer before exhaling, inhaling more frequently, or smoking the cigarette closer to its high-yield butt (Benowitz et al. 1983; Kozlowski et al. 1988). Smokers can also compensate for nicotine fad- ing by blocking the air inlet holes on the filters that are used to decrease nicotine intake (McGovern and Lando 1991). The best available evidence indicates that although nicotine consumption is indeed reduced by nicotine fading, the extent of these reductions is smaller than would be expected (i.e., based on machine rat- ings); apparently, some compensatory smoking occurs (Land0 1993). For example, one study (McGovern and Lando 1991) compared two nicotine fading regimens, brand switching and graduated filter use, each of which was designed to reduce nicotine intake by 80 percent by the final stage. Each regimen significantly reduced nicotine consumption but by far less than 80 percent: brand switching reduced intake by 42.5 per- cent and graduated filters by 55.2 percent. Lando and McGovern (1985) suggested that nico- tine fading increases smokers' self-efficacy by provid- ing them with a series of concrete steps that are mastered before cessation. Self-efficacv does increase during the fading process (McGo\.ern al;d Lando 1991), although no more than with comparison treatments (Burling et al. 1989). Moreover, increased self-efficacy has not been shown to predict treatment outcome for nicotine fading (McGovern and Lando 1991). Motivational Rewards Strategies that use motivational rewards are rooted in operant conditioning theory. These efforts are designed to provide reasons for remaining absti- nent to smokers who have just quit-reasons more tan- gible and immediate than the important but delayed outcomes that typically motivate cessation attempts (e.g., improvements in health). In a typical motiva- tional rewards intervention, the provider collects a deposit from each participant at the outset of treatment and refunds a portion of this sum at each follow-up assessment at which the participant demonstrates ab- stinence (Paxton 1983). Other variations of this tech- nique have used nonmonetary rewards (Land0 1982), punished smokers for every cigarette smoked (Murray and Hobbs 19811, instructed participants to reward themselves for abstinence (Tiffany et al. 19861, and rewarded participants who had reduced their carbon monoxide levels (Stitzer and Bigelow 1985). Curry and colleagues (1991) used a theoretical framework that tested intrinsic motivation (personalized feedback) against extrinsic motivation (financial incentive). Ab- stinence at 3 and 12 months was two times higher in the intrinsically motivated groups. Efficacy When used alone, inotivational rewards foster relatively high abstinence success in the short term, but these gains do not appear to be durable (Antonuccio et al. 1992). Participants often return to smoking after the term of the contract expires (Paxton 1980,1981). At- tempts to prolong .abstinence by varying factors such as duration and frequency of reward have generally been unsuccessful (Paxton 1981, 1983). Multicompo-. nent treatments using motivational rewards have some- times fared better than comparison treatments, but these comparisons are generally confounded by other factors (Jason et al. 1990; Lando~et al. 1990) and may lead to type II errors. A meta-analysis of 62 studies comparing components of behavioral controls found that motiva- tional rewards (contingency contracting) did not sig- nificantly alter long-term cessation rates (Fiore et al. 2000). In the final results of the Minnesota Heart Health Program, the failure of community education methods (which included motivational rewards for smoking cessation) to produce results that exceeded secular trends is an important demonstration of the difficulties in evaluating such modalities (Land0 et al. 1995). Relevant Process Measures The process measures most relevant to this strat- egy are presumably motivational; making rewards contingent on abstinence should increase a smoker's resolution to remain abstinent. However, motivational measures have been neglected in research on this intervention. Many programs require participants to administer their own rewards or punishments. Evalu- ations of these strategies should routinely assess how well participants take on this responsibility; to date, evaluations have not made this assessment. Social Support Social support intervention> try to ease the smok- .ing cessation process by enlisting the support of sig- nificant persons in smokers' lives (extratreatment social support) and by providing support from clini- cians (intratreatment social support). Both strategies may range from intense' and pervasive to relatively minimal and limited. Intensive extratreatment social support may train participants to elicit aid and sup- port of family and friends, whereas training clinicians to communicate caring, concern, and encouragement increases intratreatment social support. Increasing the cohesiGeness of smoking cessation groups can enhance both forms of social support (Hajek et al. 1985; Lando and McGovern 1991). At the basic level, the simple use of a group rather than an individual format can be viewed as a social support intervention. Efficacy Strategies that add social support to fiharmaco- logic treatment appear to significantly increase long- term quit rates compared to treatments without social support, although some intensive interventions have reported mixed results (Glasgow et al. 1986; McIntyre- Kingsolver et al. 1986). A recent meta-analysis of 19 studies (Fiore et al. 2000) reported that interventions to increase social support in the smoker's environment increase long-term cessation by 50 percent. A meta- analysis of 50 studies (Fiore et al. 2000) reported that within-treatment social support increased cessation by 30 percent. -The importance of intratreatment social support may well be reflected in the finding that indi- vidual and group counseling are both much more ef- fective than no contact interventions (Kottke et al. 1988; Fiore et al. 1996). Rcducirlg Tobmro Usr Relevant Process Measures Studies of intensive social support interventions have regularly included measures of smokers' per- ceived support. These investigations have found that the amount of support a smoker perceives is directly related to outcome (Malott et al. 1984; Glasgow et al. 1986; McIntyre-Kingsolver et al. 1986; Gruder et al. 19931, but the trials have typically failed to find evi- dence that the support itself has increased this per- ception (Malott et al. 1984; Glasgow et al. 1986). In one study that found social support intervention to be effective, the strategy was itself associated with an in- crease in received support (Gruder et al. 1993). More- over, this increase in support was statistically related to the differential outcome. Because support measures have rarely been incorporated into the evaluation of group treatments for smoking cessation, little is known about whether group formats enhance perceived sup- port and about what influence such support has on treatment outcome (Hajek et al. 1985). Weight Control Most people who quit smoking gain weight (Klesges et al. 1989), and this effect may be greater for women than for men (Williamson et al. 1991; Fant 1996). This effect has been hypothesized to result from nicotine's ability to modify various mechanisms in the central nervous system that regulate body weight (Schwid et al. 1992; Perkins 1993). Apprehension about weight gain may serve as a barrier to cessation at- tempts, especially among young women (Gritz et al. 1989). Cessation strategies that address this barrier have only recently begun to be assessed. Efficacy Two important trials have examined the contri- bution of a weight control component to a multicom- ponent smoking cessation program. One study (Hall et al. 1992) compared a specialized weight control pro- gram with both a nonspecific weight control program and a standard program. Patients in the specialized group learned behavioral self-management, reduced their caloric intake under the direction of a dietitian, and received an individualized activity plan from an exercise counselor. Patients in the nonspecific group attended several group sessions devoted to discuss- ing weight-related issues. Results showed that par- ticipants in both of these weight control programs were less likely to be abstinent after one year (21 percent success for both groups combined) than participants treated with the standard protocol (35 percent success). Another study (Pirie et al. 1992) examined the ef- fects of adding nicotine gum, weight control counsel- ing, both, or neither to a standardized smoking cessation program in a sample of women who had indicated that they were concerned about postcessation weight gain. After 12 months, the group that added nicotine gum to the standard program had much greater success (44.4 percent had quit smoking) than the groups that added weight control counseling to the standard package (27.8 percent success for the group that added weight con- trol only and 27.6 percent success for the group that added both weight control and nicotine gum). How- ever, the standard package alone was the least success- ful program (19.4 percent had quit smoking) and was viewed by participants as less appealing than the weight control component (Pirie et al. 1992). A meta-analysis of six studies (Fiore et al. 2000) that looked at the effect of dieting and physical activ- ity on smoking cessation did not find that these inter- ventions increased cessation success. A recent single study (Marcus et al. 1999) found that vigorous physi- cal activity increased quit rates. Relevant Process Measures Weight gain has not been a consistent predictor of smoking relapse (Gritz et al. 1989), and it has pre- dicted abstinence as well (Hall et al. 1986; Gritz et al. 7989; Hughes et al. 1991b). Nonetheless, actual con- trol of weight is an important process measure for weight control interventions-the primary purpose of which is relapse prevention-because they explicitly assume that preventing weight gain will boost absti- nence rates (Hall et al. 1992; Pirie et al. 1992). Neither published trial of weight control interventions found differences in weight gain among abstinent subjects across treatment conditions (Hall et al. 1992; Pirie et al. 1992). One of the studies (Hall et al. 1992) found evidence for lower caloric intake in specialized weight control interventions, especially among women, but failed to find differences in activity levels across treat- ment conditions. In sum, despite the intuitive appeal of weight control interventions to promote smoking cessation, there is mixed evidence relating such inter- ventions to cessation success (Fiore et al. 2000). Hall and colleagues (1992) suggested that such interventions may interfere with cessation. However, Marcus and colleagues (1999) found that a vigorous exercise inter- vention increased quit rates while contributing to weight management. Pharmacotherapies, including bupropion sustained release (SR) and nicotine gum, may help to delay weight gain after cessation (Emont and Cummings 1987; Doherty et al. 1996; Jorenby et al. 1999). Hypnosis Efficacy Some smokers try hypnosis therapy to help them quit (Schlvartz 1987). Strategies for hypnosis interven- tions include direct hypnotic suggestions to quit, sug- gestions intended to produce aversion to smoking, and training in self-hypnosis to reinforce formal treatment (Simon and Salzberg 1982). Efficacy The methodological shortcomings of hypnosis research make it difficult to estimate the value of this therapy for smoking cessation (Schwartz 1987). Re- viewers have noted that, in general, hypnosis is not very effective when used alone, but it may be useful as part of a multicomponent intervention in which subjects see a therapist many times (Holroyd 1980; Schwartz 1987). In methodologically sound studies, hypnosis often fails to outperform comparison tech- niques, such as self-help strategies (Rabkin et al. 1984; Lambe et al. 1986). Hypnosis techniques may work best for the relatively small proportion of people highly susceptible to hypnosis (Barabasz et al. 1986; USDHHS 1988). Since the late 198Os, there have been only two trials of hypnosis in smoking cessation, with incon- clusi\-e results. Johnson and Karkut (1994) conducted an uncontrolled clinical trial of hypnosis plus aversion treatment and reported about 90 percent abstinence at three months. A similar uncontrolled study of 226 smokers reported a 23-percent abstinence at two years (Spiegel et al. 1993). A recent review of hypnosis by the Cochrane group (Abbot et al. 2000) found insuffi- cient evidence to support hypnosis as a treatment for smoking cessation. Releoar; t Process Measures Appropriate process measures for studies of hypnosis are those that assess the various means of hyp- notic induction and the motivational changes that are presumed to accrue from them. Because measures have rarely been collected, little is known about the mecha- nisms of hypnotic treatments for smoking cessation (Holroyd 1980; Schwartz 1987; USDHHS 1988). Acupuncture The typical acupuncture treatment for smoking cessation involves the insertion of needles or staples into the outer ear, but a number of other techniques ha1.e been investigated (Schwartz 1988). The most commonly cited rationale for using acupuncture is that it relie\,cs the discomfort of nicotine withdrawal. The available evidence suggests that acupunc- ture is no more effective in smoking cessation than placebo treatments (Schwartz 1987). For example, Schwartz (1988) reviewed eight studies in which acu- puncture at a theoretically appropriate site was con- trasted with acupuncture at a placebo site. Only one of these studies found greater success among partici- pants undergoing the procedure with theoretically appropriate sites (MacHovec and Man 1978). A recent meta-analysis of five studies (Fiore et al. 2000) found that acupuncture was no more effective than placebo. Relevant Process Measures Acupuncture is commonly presumed to exert its effects by easing tobacco withdrawal. At present there is no evidence that acupuncture is capable of relieving withdrawal symptoms associated with smoking cessa- tion (Clavel et al. 1987; Schwartz 1987; USDHHS 1988). Summary of Intensive Clinical Interventions Intensive programs serve an important function in the nation's efforts to reduce smoking, despite the resources the programs demand and the relatively small population of smokers who use them. Such pro- grams may be particularly useful in treating smokers who find it most difficult to quit. Because intensive smoking cessation programs differ in structure and content, evaluation is often ham- pered by variation in methodology and by a lack of research addressing specific treatment techniques. Because few studies have chosen to isolate single treat- ments, assessment of the effectiveness of specific ap- proaches is difficult. Nonetheless, skills training, rapid smoking, and both intratreatment and extratreatment social support have been associated with successful smoking cessation. When such treatments are shown to be effective, they are usually part of a multifactorial intervention. Little clear evidence has implicated par- ticular psychological, behavioral, or cognitive mecha- nisms as the agents of change. The specific impact of intensive interventions may be masked by the efficacy of several multicomponent programs, some of which have achieved cessation proportions of 30-50 percent (Land0 1993). Thus, in their positive effect on smoking cessa- tion and long-term abstinence rates (Kottke et al. 1988; Fiore et al. 1994a), intensive interventions seem little different from other forms of counseling or psy- chotherapy. With intensive interventions, as with counseling, it is difficult to attribute the efficacy to Reducing Tobacco Use specific characteristics of the interventions or to spe- cific change mechanisms (Luborsky et al. 1975; Elkin et al. 1989). Pharmacologic Interventions At first look, nicotine replacement therapy ap- pears to be the treatment of a disease with its cause. The rationale, however, is well established. Observa- tions on the beneficial effects of nicotine replacement in abstinent smokers were first made in 1967 (Lucchesi et al. 19671, and the process has its medical precedent in the use of methadone for opiate dependence. Nico- tine use, in the form of 10 or more cigarettes a day, provides continuous neuroexposure (Benowitz 1993). The resulting tolerance and physical dependence pro- duce classic withdrawal symptoms (USDHHS 1988). As Benowitz (1993) has summarized, "Nicotine re- placement therapy serves primarily to break the daily addiction cycle by relieving withdrawal symptoms, thereby facilitating behavioural modification that is necessary for permanent smoking cessation" (p. 158). However, as will be discussed later in this chapter, re- cent data suggest that nicotine replacement may be effective without behavioral support or counseling. A number of candidate delivery systems have now been extensively evaluated with clear and consistent results. In addition, nonnicotine pharmacotherapies for treat- ment of tobacco use are now available. Nicotine Polacrilex Nicotine polacrilex (nicotine gum) was approved by the Food and Drug Administration (FDA) for use as an aid to smoking cessation in a 2-mg dose in 1984 and in a 4-mg dose in 1994. The nicotine in the gum is bound to an ion-exchange resin. Chewing the gum liberates the nicotine, which is absorbed through the buccal mucosa. Currently, both doses of nicotine polacrilex are approved for use as over-the-counter preparations by adults. The package insert instructs patients to use the gum as needed with the constraint that they not exceed a daily dose of 20 pieces of 4-mg gum or 30 pieces of 2-mg gum. Efiicacy With more than 50 studies on its efficacy, nico- tine gum is the most extensively investigated pharma- cologic treatment for smoking cessation. This body of research has been summarized by several major meta-analyses (Lam et al. 1987; Cepeda-Benito 1993; Silagy et al. 1994; Tang et al. 1994). The most recent meta-analysis (Fiore et al. 2000) is summarized in Table 4.3. All meta-analyses found the gum to be effective in helping smokers quit. Lam and colleagues (1987) performed a meta- analysis of nine randomized, controlled trials of the 2-mg nicotine gum. These authors performed sepa- rate analyses on the trials conducted in specialized smoking cessation clinics and on those conducted in general medical settings. In the specialized clinics, ces- sation success was greater with nicotine gum than with placebo gum. In general medical practice settings, however, nicotine gum was no more successful than placebo gum; both types of gum were more successful than usual care. The authors suggested that partici- pants at the specialized cessation clinics had greater success because such participants may have been more motivated to quit and may have received more inten- sive adjuvant behavioral support than those at the generalized settings. The authors also speculated that patients who seek treatment in specialized clinics may be more physically dependent on nicotine and thus more likely to benefit from nicotine replacement than the average patient seen in a general medical clinic. Cepeda-Benito (1993) performed a meta-analysis of 33 trials of the 2-mg gum. As in the review by Lam and colleagues (19871, the trials were categorized ac- cording to whether the adjuvant behavioral support was intensive or brief and according to whether the control group used placebo gum or no gum. Pooled estimates of efficacy were derived for short-term (O-8 weeks after treatment) and long-term (12 2 2 months) outcome measures within each category. Effect sizes were not systematically related to the type of control treatment used but were related to the intensity of be- havioral support provided. When used in intensive interventions, the gum was associated with greater abstinence success than the control treatments at both long-term and short-term follow-up. When used in brief behavioral interventions, however, the gum out- performed the control interventions only at short-term follow-up. The author concluded that nicotine gum is an effective aid to smoking cessation but questioned its long-term value in the absence of adjuvant psycho- social support. In the context of a larger review of available nico- tine replacement therapies, Tang and colleagues (1994) performed a meta-analysis of 28 randomized, controlled trials of the 2-mg gum and 6 randomized, controlled trials of the 4-mg gum. The authors found that among participants recruited through advertise- ments to attend specialized cessation clinics, the 2-mg gum was associated with an ll-percent increase in success over control treatments. However, among Mmmger?rtwt of Nicotirw Addiction 2 13 smokers who were directly invited to participate in a general smoking cessation trial conducted by a non- specialist physician, the 2-mg gum increased absti- nence success by only 3 percentage points over control conditions. Consistent with the analysis by Lam and colleagues (1987), the authors suggested that these findings reflect (1) the greater motivation of the smok- ers who referred themselves (i.e., responded to adver- tisements instead of being directly invited), (2) the greater degree of nicotine dependence in the self- referred group, and (3) the more extensive encourage- ment and more detailed instructions provided by therapists in the specialized settings in which the self- referred smokers were treated. Six of the 28 trials of the 2-mg gum (Fagerstrcm 1982,1984; Jarvik and Schneider 1984; Areechon and Punnotock 1988; Hughes et al. 1989b; Jensen et al. 1990) reported abstinence success as a function of nicotine dependence as assessed by the Fagerstrijm Tolerance Questionnaire (described later in this chap- ter). The authors aggregated these data and found that the 2-mg gum improved cessation success by 16 percentage points among smokers scoring high (indicating considerable nicotine dependence) on the Table 4.3. Meta-analyses of efficacy (estimated odds ratio and abstinence rates) for seven pharmacotherapies used in tobacco dependence treatment Pharmacotherapy Number of study groups Bupropion SR' (n = 2i) Placebo Bupropion SR Nicotine gum, 2 mg (n = 13) Placebo Nicotine gum Nicotine inhaler (n = 4) Placebo Nicotine inhaler Nicotine nasal spray (n = 3) Placebo Nicotine spray Transdermal nicotine (the nicotine patch) (n = 27) Placebo Transdermal nicotine Clonidine (n = 5) Placebo Clonidine Nortriptyline (n = 2) Placebo Nortriptyline 2 4 16 18 4 4 3 3 28 32 6 8 3 3 Estimated Estimated odds ratio abstinence rate (95% CI') (95% CI) 1.0 2.1 (1.5, 3.0) 1.0 1.5 (1.3,1.8) 1.0 2.5 (1.7, 3.6) 1.0 2.7 (1.8,4.1) 17.3 30.5 (23.2,37.8) 17.1 23.7 (20.6, 26.7) 10.5 22.8 (16.4,29.2) 13.9 30.5 (21.8,39.2) 1.0 1.9 (1.7,2.2) 1.0 2.1 (1.4, 3.2) 1.0 3.2 (1.8, 5.7) 10.0 17.7 (16.0, 19.5) 13.9 25.6 (17.7,33.6) 11.7 30.1 (18.1,41.6) *Confidence interval. `SR = sustained release. iNumber of studies. Source: Fiore et al. 2000. 2 24 Chayter 4 ReducilTg Tobacco Use questionnaire but produced only a 2-percentage point increase among smokers whose scores indicated low levels of nicotine dependence. When data from the 4-mg gum trials (Puska et al. 1979; Kornitzer et al. 1987; Tonnesen et al. 1988a,b; Blondal 1989; Hughes et al. 1990a) were aggregated, the influence of nicotine dependence paralleled that seen in trials using the lower dose. Among smokers highly dependent on nicotine, those who used the 4-mg gum had a 21-percent greater success at cessa- tion than those using the 2-mg gum. In contrast, among smokers low in nicotine dependence, those who used the 4-mg gum had an l&percent lower success than those using the 2-mg gum. Highly dependent participants using the 4-mg gum had a 35-percent greater success than those using the placebo gum, but this comparative improvement was only 5 percent greater among less dependent participants. Tang and colleagues (19941 concluded that nico- tine gum is an effective aid to smoking cessation and suggested that its efficacy is a direct function of the dependence of the smoker. On the basis of their re- view of other nicotine replacement therapies (includ- ing the nicotine patch), the authors concluded that the 4-mg gum is the most effective form of nicotine re- placement for highly dependent smokers. Silagy and colleagues (1994) examined 42 nico- tine gum trials in their meta-analysis of nicotine re- placement interventions. To compute effect sizes for each analysis, the authors combined data from the longest follow-up assessments (mainly 12 months) from available trials, regardless of gum dose or type of control treatment. Across all 42 trials, 42 percent of participants using nicotine gum quit smoking, whereas only 18 percent of participants in the control groups, who used either placebo gum or no gum, succeeded in quitting. The pooled odds ratio (OR) for the gum- to-control comparison across all trials was 1.61 (95 percent confidence interval [CI], 1.46-1.78). Differ- ences between gum and control conditions did not vary according to the intensity of adjuvant behavioral support. Fiore and colleagues (1990) conducted a meta- analysis of 13 randomized controlled trials of 2-mg nicotine gum therapy with at least five months of fol- low-up (Table 4.3). Nicotine gum treatment was asso- ciated with a 50-percent increase in quit rates (23.7 percent quit rate vs. 17.1 percent) in the control group. There were too few studies done in the over-the- counter setting to allow meta-analysis of the over-the- counter effect of nicotine gum. Taken together, these meta-analyses suggest that nicotine chewing gum is an effective aid to smoking cessation. This conclusion continues to be borne out as evidence continues to accumulate. In an ongoing project, Silagy and colleagues (1999) have been regu- larly searching medical databases for new nicotine re- placement trials, recalculating effect sizes as new data sources are identified, and frequently publishing the updated meta-analyses. In the most recent edition of this meta-analysis, the pooled gum-to-control OR was estimated at 1.63. That in most settings nicotine- containing gum is associated with greater cessation success than placebo gum suggests that the gum's ef- ficacy is due to its pharmacologic properties. Some evidence indicates that the efficacy of the 2-mg gum depends on the presence of intensive adjuvant behav- ioral support. The meta-analysis by Silagy and col- leagues (1994) suggests that nicotine gum may be beneficial even without intensive adjuvant therapy. In this analysis, however, because 2-mg and 4-mg gum studies are combined, definitive conclusions about the efficacy of either dose alone in the absence of behav- ioral support cannot be drawn. This finding under- scores the importance of selecting those smokers for whom nicotine gum is likely to be beneficial. The avail- able evidence suggests that traditional measures of nicotine dependence may be a useful basis for select- ing gum candidates. Both doses of the gum appear to be of greater value to smokers who are more depen- dent on nicotine. The 4-mg gum may be particularly effective for the most dependent smokers. RelevaA Process Measures Nicotine gum is presumed to exert its effects by replacing a portion of the nicotine that smokers usu- ally obtain through smoking; in therapy, the gum ame- liorates aversive tobacco withdrawal (Benowitz 1991; Hughes 1993). Some evidence suggests that nicotine gum reliably reduces some withdrawal symptoms. Patients receiving the 2-mg nicotine gum have consistently reported having less total withdrawal dis- comfort than patients treated with placebo gum (Jarvis et al. 1982; Hughes et al. 1984,1989a, 1991b; Gross and Stitzer 1989; Hatsukami et al. 1991). However, studies have found that withdrawal severity is not consistently related to smoking relapse (West 1992; Hughes 19931, and the withdrawal suppression produced by nicotine gum appears to be somewhat independent of its effi- cacy. Moreover, the suppression reported seems to accrue through the lessening of a relatively small sub- set of withdrawal symptoms (Hughes et al. 1990b). The 2-mg gum consistently alleviates symptoms such as Sqeo)~ Grtreral's Report anxiety and irritability but does not appear to reliably ameliorate craving, hunger, sleep disturbance, or dif- ficulty concentrating (West et al. 1984a,b; Gross and Stitzer 1989; Hughes et al. 1989a, 1990a; Hatsukami et al. 1991). One trial (Hughes et al. 1990a) has found that the 4-mg gum was no more effective than the 2-mg gum either in suppressing total withdrawal se- verity or in relieving any of the individual symptoms of withdrawal. Future research must explore whether these counterintuitive findings are a result of poor measurement of withdrawal severity or whether other mechanisms explain how nicotine gum produces clini- cal success (Hughes 1993). Effect on Postcessation Change in Body Weight Evidence suggests that the 2-mg gum is capable of delaying, but not preventing, postcessation weight gain. Early in the cessation process, smokers given the 2-mg gum tend to gain less weight than smokers treated with placebo gum (Gross et al. 1989). During this period, weight gain among the 2-mg gum users is inversely related to the amount of gum used (Emont and Cummings 1987; Fagerstrom 1987; Killen et al. 1990a; Nides et al. 1994). However, differences in weight gain between smokers using the 2-mg gum, using placebo gum, and using no gum (Gross et al. 1989; Nides et al. 1994) disappear when follow-up is conducted after gum therapy has ended. Relatively little is known about the weight- related effects of the 4-mg gum. Early trials did not show it to diminish weight gain any more than either the 2-mg gum (Kornitzer et al. 1987; Tonnesen et al. 1988a) or the placebo gum (Puska et al. 1979; Tonnesen et al. 1988a). These trials, however, tended to use dif- ferent weight measures and more distal end points than the typical trial with 2-mg gum, and one trial used a mixed-dose regimen (Tonnesen et al. 1988a). A more recent study, however, reported that nicotine gum sup- pressed weight gain with greater suppression occur- ring with the 4-mg dose (Doherty et al. 1996). Analysis of salivary cotinine showed that smokers who replaced a greater percentage of their baseline cotinine levels gained less weight. Side Effects and Likelihood of Inappropriate Use Common side effects reported by the 2-mg gum users include mouth soreness, hiccups, indigestion, jaw ache, and unpleasant taste (American Medical As- sociation [AMA] 1993; Tang et al. 1994). Most of these symptoms are relatively mild and transient, and many can be resolved by correcting the user`s chew- ing technique. Symptoms observed less frequently (in less than 2 percent of patients) include irritability, lightheadedness, headache, excessive salivation, and anorexia (AMA 1993). Moreover, absorption of nico- tine from the gum is highly dependent on the pH of the mouth (Henningfield et al. 1990). Because nico- tine is inactivated by an acidic environment, patients are urged to refrain from eating or drinking anything but water for 30 minutes before using the gum. Ap- proximately lo-25 percent of successful abstainers con- tinue to use the gum for one year or longer (Hajek et al. 1988; Hughes 1988; Hughes et al. 1991a). Although discontinuance of use should be encouraged, contin- ued use confers a substantial reduced health risk com- pared to a return to smoking. The 4-mg gum appears to have similar side effects, but it may produce slightly more dyspepsia and hiccuping than does the 2-mg gum (Tonnesen et al. 1988a,b). Transdermal Nicotine In 1991, the FDA approved the use of transdermal nicotine patches as an aid to smoking cessation. Nico- tine patches contain a reservoir of nicotine that diffuses through the skin and into the wearer's bloodstream at a constant rate. Patients are usually instructed to apply one patch each day. Specific dosing regimen may vary. All currently marketed brands are designed to deliver approximately 0.9 mg per hour of nicotine over the weaning period. Most are intended for 24-hour wear and deliver 21-22 mg of nicotine; one is intended for waking hours wear (16 hours per day) and deliv- ers 15 mg of nicotine. Full-strength patches typically produce serum nicotine levels similar to trough levels of serum nicotine in moderate to heavy smokers (Mulligan et al. 19901. On July 3, 1996, the FDA ap- proved the transdermal nicotine patch for over-the- counter sales at a dose of 15 mg for use as part of a comprehensive behavioral program of smoking ces- sation, although the FDA's proscription does not pro- vide a clear statement of the constituents of such a program. Since that time, all varieties of nicotine patches have become available over the counter, some as "house brands." Efficacy Several meta-analyses of the efficacy of the nico- tine patch have been published (PO 1993; Fiore et al. 1994~; Gourlay 1994; Silagy et al. 1994; Tang et al. 1994; Fiore et al. 2000). Each meta-analysis has concluded that the patch is an effective aid to smoking cessation. PO (1993) combined data from 11 nicotine patch trials and found that persons using the nicotine patch had greater cessation success than persons using a 12 6 Chapter 4 Reducing Tobacco Use placebo patch. This finding held for both short-term follow-up (3-10 weeks; combined OR = 3.10 J95 per- cent CI, 2.65-3.621) and long-term follow-up (6-12 months; combined OR = 2.26 [95 percent CI, 1.80- 2.861). Gourlay (19941 pooled the results of six trials and found that the nicotine patch produced greater cessation success than a placebo patch at all follow-up assessments (2-3 months, 6 months, and 12 months; all pooled ORs were between 2.2 and 2.4 I95 percent CI, 1.6-3.41). Tang and colleagues (1994) conducted a meta-analysis of six patch trials. Overall, at long-term (12-month) follow-up, persons using nicotine patches had a 9-percent (6-13 percent) greater success at ces- sation than did persons using placebo patches. Nico- tine patches were found to be more effectivre among self-referred subjects than among invited subjects and slightly more effective among smokers \vho were more dependent on nicotine. Silagy and colleagues (199-l) combined data from nine patch trials and found that at long-term (12-month) follow-up, nicotine patches were associated with a combined OR of 2.07 (95 per- cent CI, 1.64-2.62) when compared with control con- ditions (placebo patches or no patch). Secondary analyses indicated that the patch's relative efficacy was not affected by the intensity of adjuvant support. Fiore and colleagues (1994~) examined 17 nicotine patch tri- als and found a combined OR of 2.6 (95 percent CI, 2.2-3.0) at the end of the treatment and 3.0 (95 percent CI, 2.4-3.7) at 72-month follow-up. More intensive ad- juvant support was found to produce higher absti- nence rates at six months (26.5 vs. 19.5 percent for low-intensity interventions) but did not increase the relative advantage of nicotine patches over placebo patches. The 16- and 24-hour patches were found to be equally effective. Neither weaning nor extending treatment beyond eight weeks was found to improve outcome. A recent meta-analysis (Fiore et al. 2000) of 27 studies reported that transdermal nicotine increased long-term cessation by 90 percent (Table 4.3). A meta- analysis of three studies reported that over-the-counter nicotine patch use increased successful long-term cessation by 80 percent (Fiore et al. 2000). These meta-analyses strongly indicate that the nicotine patch is an effective aid to smoking cessation. This conclusion is buttressed by the findings of a con- tinuing, regularly updated review of the existing re- search literature on transdermal nicotine (Silagy et al. 1999). In the most recent release of this evolving meta- analysis, Silagy and colleagues (1999) found a pooled patch-to-control OR of 1.84 (95 percent CI, 1.60-2.10). The data continue to suggest that 16- and 24-hour patches are equivalent in efficacy, that there is no ad vantage associated with weaning or tapering of patch dose, and that the relative efficacy of the patch is fairly independent of the intensity of adjuvant therapy. Nico- tine patches have been consistently found to outper- form placebo patches regardless of dosing regimen and in a variety of investigational settings. For example, a study of "real-world" use of the patch-based on a follow-back of older persons who had filled patch prescriptions-produced a self-reported cessation pro- portion of 29 percent at six months (Orleans et al. 1994). The patch is more effective than placebo treatment when paired with only brief support, and it is associ- ated with the higher long-term success when paired with more intensive counseling or behavioral interven- tions (Fiore et al. 1994b). Though the nicotine patch does increase success rates when used with minimal formal counseling, many nicotine patch clinical trials invrolve frequent follow-up assessments. Such contacts might boost success rates obtained with the patch. In support of this possibility, Jorenby and colleagues (1995b) found that the combination of nicotine patch treatment plus frequent assessments produced follow- up outcomes equivalent to the nicotine patch plus in- tensive behavioral therapy. Further assessment of this issue is important, as frequent follow-up contact does not usually accompany nicotine patch use outside of clinical trials (Cummings et al. 1994; Swartz et al. 1995). A meta-analysis of three studies of over-the-counter nicotine patches, however, indicated that patch therapy was superior to placebo (Fiore et al. 2000). Effects on Discomfort of Nicotine Withdrawal Some evidence suggests that the nicotine patch reduces overall measures of nicotine withdrawal dis- comfort (Daughton et al. 1991; Transdermal Nicotine Study Group 1991; Jorenby et al. 19961, but this find- ing has not been consistent (Abelin et al. 1989; Tannesen et al. 1991; Merz et al. 1993). Use of the nico- tine patch has been repeatedly found to reduce the craving for cigarettes (Abelin et al. 1989; Rose et al. 1990; Tonnesen et al. 1991; Transdermal Nicotine Study Group 1991; Merz et al. 1993; Sachs et al. 1993; Westman et al. 1993; Fiore et al. 1994b; Levin et al. 1994; Jorenby et al. 1996), but other symptoms of nicotine withdrawal are affected less reliably (Palmer et al. 1992). In a study designed to clarify the impact the patch has on with- drawal symptoms, the patch reliably reduced craving, anxiety, and irritability but did not alleviate depressed mood, restlessness, or sleep disruption (Jorenby et al. 1996). The authors noted that with or without the patch, most withdrawal symptoms disappeared within three to four weeks. Effect on Postcessation Change in Body Weight Nicotine patches can attenuate postcessation weight gain while they are in use (Abelin et al. 1989; Sachs et al. 1993; Jorenby et al. 1995a; Dale et al. 19981, but this short-term effect has not always been observed (Rose et al. 1990; Tonnesen et al. 1991; Transdermal Nicotine Study Group 1991; Fiore et al. 1994b). More- over, studies that follow up effects after treatment has ended have not found that persons who used the nico- tine patch gained less weight than those who used a placebo patch (Tonnesen et al. 1991). Side Effects and Likelihood of Inappropriate Use Most side effects of nicotine patch use are rela- tively mild; less than 5 percent of patients need to dis- continue patch therapy because of side effects (Hughes and Glaser 1993). Minor skin irritation at the patch site is reported by 30-50 percent of patch users and can be relieved by moving the patch to another site. Insomnia is reported by l-23 percent of patch users (AMA 1993). Comparatively rare side effects include headache, dizziness, fatigue, gastrointestinal distress, sweating, limb pain, and palpitations (Palmer et al. 1992). Studies have found little evidence that people will inappropriately use transdermal nicotine systems (Palmer et al. 1992; Hughes 1993; Jorenby et al. 1995b). The risks associated with using the nicotine patch during pregnancy are largely unknown. Nicotine it- self poses risks to the fetus, including neurotoxicity (Slotkin 1998), and pregnant women should first be encouraged to quit without pharmacotherapy. Because exposure to nicotine through maternal use of the patch probably poses less danger to the fetus than does con- tinued maternal smoking (Hackman et al. 19991, how- ever, nicotine replacement therapy may be indicated for pregnant women who are unable to quit smoking (Benowitz 1991; Lewis and Fiore 1994). However, if a decision is made to use nicotine replacement therapy during pregnancy, the physician should consider moni- toring blood nicotine levels, using doses at the low end of the effective range, and choosing intermittent de- livery systems (such as nicotine gum) (Fiore et al. 2000). The issue is under active investigation. Continued smoking while using the patch may be a significant problem. In an observational study of self-reported patch use, almost one-half the respon- dents stated that they smoked while using the patch; 20 percent of the respondents did so every day (Or- leans et al. 1994). A small number of adverse cardio- vascular events were reported in patients who continued to smoke while using the patch. When these events received much attention from the popular press, several analyses, including one by an FDA advisory committee, have documented no association between nicotine replacement therapy and cardiovascular events even in patients who continue to smoke inter- mittently (Working Group for the Study of Transdermal Nicotine in Patients with Coronary Ar- tery Disease 1994; Joseph et al. 1996; Benowitz and Gourlay 1997; Mahmarian et al. 1997). Caution should be used, however, for patients with acute cardiovas- cular disease (immediately post-myocardial infarction or in the presence of serious arrhythmias or serious or accelerating angina pectoris). Relevant Process Measures Like nicotine gum, the nicotine patch is intended to reduce tobacco withdrawal symptoms Palmer et al. 1992; Glover 1993b; Hughes and Glaser 1993). Al- though the nicotine patch appears to reduce with- drawal severity, particularly craving for cigarettes, withdrawal suppression may or may not be respon- sible for the patch's efficacy (Hughes 1993). For ex- ample, one trial failed to reveal reliable differences in withdrawal severity between persons using nicotine patches and those using placebo patches (Merz et al. 1993); the trial nevertheless found that participants who used the nicotine patch were nearly twice as likely to quit smoking. Another trial employing two doses of transdermal nicotine found that the higher-dose patch produced significantly greater cessation success than the lower-dose patch, even though both doses provided about the same amount of relief from with- drawal symptoms (Transdermal Nicotine Study Group 1991; Hughes 1993). Clearly, other potential mecha- nisms of the patch's action, as well as the action of nico- tine replacement therapy in general, need to be explored. Nicotine Nasal Spray Nicotine nasal spray was approved for prescrip- tion use in the United States in March 1996. The spray consists of a pocket-sized bottle and pump assembly, which is fitted to a nozzle designed for insertion into the nose. Each metered spray delivers 0.5 mg of nico- tine to the nasal mucosa. The recommended dose is 1 mg, or one 0.5-mg spray per nostril, as needed (Sutherland et al. 1992). Efficacy A number of clinical trials have assessed the effi- cacy of the nicotine nasal spray as an aid to smoking cessation. Sutherland and colleagues (1992) found that Reducing Tobacco Use 26 percent of participants given nicotine nasal spray were abstinent after one year, compared with only 10 percent of participants given placebo. Hjalmarson and colleagues (1994) found similar results in a placebo- controlled trial; at one-year follow-up, abstinence rates were 27 percent and 15 percent, respectively, for par- ticipants given active spray or placebo. Schneider and colleagues (1995) again replicated this effect, finding continuous abstinence rates of 18 percent and 8 per- cent among participants given active or placebo spray. Another study (Blondal et al. 1997) did not find a sig- nificant difference in abstinence rates between active spray and placebo groups at one year (25 vs. ;17 per- cent); active spray was associated with higher absti- nence rates at six months and earlier in this trial. Recently, Blondal and colleagues (1999) provided all participants in a second trial with active nicotine patches, then studied the incremental efficacy of add- ing nasal spray therapy to the patch regimen in a double-blind, placebo-controlled fashion. Results showed that participants given the active spray were more likely to be abstinent -after one year than partici- pants given placebo (27 vs. 11 percent). Participants given active spray had a higher rate of abstinence than participants given placebo a full six years after the start of treatment 06 vs. 9 percent), but this effect was only marginally significant. Taken together, the results of these studies suggest that nicotine nasal spray is an aid to smoking cessation. A meta-analysis by Silagy and colleagues (1999) reported a pooled spray-to-control OR of 2.27, and a recent meta-analysis (Fiore et al. 2000) reported an OR of 2.7 (30.5 percent long-term abstinence rate) (Table 4.3). Effect on Discomfort of Nicotine Withdrawal Evidence regarding the nicotine nasal spray's effects on nicotine withdrawal discomfort is sparse. The results of two studies suggest that the spray may be useful for coping with craving, but may not be ef- fective in alleviating other withdrawal symptoms. One study (Sutherland et al. 1992) found that, compared with participants using placebo spray, participants treated with nicotine spray reported having less total withdrawal discomfort during the 48 hours immedi- ately after smoking cessation and reported less crav- ing for cigarettes during this period. After 48 hours, however, the two groups reported equivalent levels of withdrawal discomfort and craving. When craving did arise, the nicotine spray was consistently rated more effective than the placebo spray. The other study (Hjalmarson et al. 1994) found that during the first 48 hours of smoking cessation, users of nicotine spray reported somewhat less severe withdrawal discomfort than placebo users, but this effect was not statistically significant. The severity of craving was found to be similar across both groups, but the nicotine spray was more helpful in quelling craving than the placebo spray was. Other clinical tri- als have not reported comparisons between active and .placebo spray groups with regard to withdrawal mea- sures (e.g., Schneider et al. 1995; Blondal et al. 1999). Effect on Postcessation Change in Body Weight The limited evidence available suggests that the nicotine nasal spray may be capable of delaying, but not preventing, postcessation weight gain. In one of the trials (Sutherland et al. 19921, participants were allowed to use the spray they were assigned for as long as one year.. Weight effects in that study differed as a function of duration of spray use: abstinent subjects who had continued to use the nicotine spray for the entire year of the study had gained significantly less weight than subjects still using the placebo spray. However, change in body weight was equivalent for abstinent patients who had stopped using either type of spray during the year. Another study (Hjalmarson et al. 1994) failed to find any statistically significant differences in weight gain between participants using nicotine spray and those using placebo spray. The authors observed, how- ever, that participants still using nicotine spray at the 12-month follow-up tended to gain less weight than both participants continuing to use a placebo spray and participants who had stopped using the nicotine spray before that time. Side Effects and Likelihood of Inappropriate Use Unpleasant side effects are common with the nasal spray. Between 75 and 100 percent of nasal spray users reported experiencing irritant effects, such as runny nose, sneezing, throat irritation, nasal irritation, watering eyes, and coughing (Sutherland et al. 1992; Hjalmarson et al. 1994; Schneider et al. 1995). Some authors have reported that these sensory irritation ef- fects are actually viewed as desirable by many smok- ers and have suggested that they may help bridge the gap between cigarette smoking and nicotine replace- ment (Glover 1993a; Schneider 1993). Less common side effects, present in 15-25 percent of users, include nausea, sweating, headache, dizziness, and cold hands and feet. Because the spray rapidly delivers nicotine to the user, the potential for inappropriate use (e.g., using more often or at a higher dose than recommended) is high. The results of both clinical trials lend some cre- dence to these speculations. Sutherland and colleagues (1992) found that 43 percent of abstinent study par- ticipants who had been given the nicotine spray chose to continue using it for the entire year of the study; moreover, mean plasma nicotine concentrations in- creased over the follow-up period among participants who continued to use the spray. Participants in the trial conducted by Hjalmarson and colleagues (1994) were explicitly encouraged to begin weaning them- selves from the spray (whether nicotine or placebo) after three months. Nonetheless, 30 percent of absti- nent participants who had been given the nicotine spray continued to use it after one year. Schneider and colleagues (1995) required that participants in their trial use the spray daily for six weeks, then allowed participants to use spray for up to six months postcessation as needed. Thirty-two percent of par- ticipants given active spray continued using it daily for six months, compared with 13 percent of partici- pants given placebo. The authors also reported that some continuous abstainers assigned to active spray reported being concerned that they were dependent upon the spray at six months postcessation. However, a substantial proportion of these individuals remained abstinent many months after drug weaning. Tonnesen et al. 19931. Each inhaler contains enough nicotine for approximately 300 puffs. Smokers are in- strutted to use between 6 and 16 inhalers per day. Efficacy A handful of published trials have examined the efficacy of the nicotine inhaler as an aid to smoking ces- sation. Tonnesen and colleagues (1993) found that 17 percent of participants randomized to active inhalers had quit smoking at six months, compared with 8 per- cent of participants given placebo. Corresponding rates at one year were 15 vs. 5 percent. Schneider and col- leagues (1996) found active-placebo abstinence rates of 17 vs. 9 percent and 13 vs. 8 percent at six months and one year, respectively. These differences were not sig- nificant in the Schneider trial, although active inhalers were superior to placebo at all follow-ups through three months postcessation. Hjalmarson and colleagues (1997) found continuous abstinence rates of 35 percent and 28 percent for active inhaler users at 6 and 12 months, compared with 19 percent and 18 percent, respectively, among placebo users. Active-placebo comparisons were statistically significant at all follow- ups in this trial. The most recent edition of a regularly updated meta-analysis of nicotine replacement prod- ucts (Silagy et al. 1999) found an inhaler-to-control pooled OR of 2.08, and another recent meta-analysis of four studies (Fiore et al. 2000) reported a pooled OR of 2.5 (Table 4.3). Relevant Process Measures Nicotine nasal spray, like other nicotine replace- ment products, is intended to aid smoking cessation by relieving withdrawal symptoms. Although the spray has been found effective in promoting cessation, its circumscribed impact on total withdrawal severity suggests that withdrawal relief is not itself responsible for the spray's usefulness. The spray's documented ability to alleviate craving may be what makes it an effective smoking cessation treatment. More research is needed to advance definitive conclusions about the Taken together, the results suggest that the nico- tine inhaler is an effective aid to smoking cessation. However, the findings of Schneider and colleagues (1996) suggest that the inhaler may be most useful for producing initial abstinence and that additional inter- ventions may be needed to prevent relapse among users of the inhaler. Effects ou Discomfort of Nicotine Withdrawal spray's mechanism of action. Nicotine Inhaler Limited information is available regarding the effects of the nicotine inhaler on nicotine withdrawal symptoms. Two studies (Schneider et al. 1996; In May 1997, the FDA approved the nicotine in- haler for prescription use. The inhaler consists of a plastic tube, about the size of a cigarette, that contains a plug impregnated with nicotine. Menthol is added to the plug to reduce throat irritation. Smokers are instructed to puff on the inhaler as they would on a cigarette. An average puff delivers approximately 13 ,ug of nicotine (about 1/80th the amount of nicotine contained in an average cigarette puff), which is ab- sorbed primarily by the buccal route (Glover 1993a; Hjalmarson et al. 1997) showed that active inhaler use was associated with decreased craving during the first several days of the quit attempt but not thereafter. Hjalmarson and colleagues (1997) assessed a wide ar- ray of withdrawal symptoms across the cessation at- tempt, but did not find any effects of active inhalers on these other than the fleeting effects on craving. However, this may have been influenced by a floor effect, as mean withdrawal scores were very low in both groups across all assessments. 220 Chapter 4 Reducing Tobacco Use Side Effects and Likelihood of Inappropria te Use The most common side effects associated with inhaler use are throat irritation and coughing. These are reported by between 20 to 50 percent of active in- haler users and are less common among placebo inhaler users (Tonnesen et al. 1993; Schneider et al. 1996; Hjalmarson et al. 1997). Other less common side effects include nausea, bad taste in the mouth, dizzi- ness, gastrointestinal disturbances, and oral burning or smarting. Few (O-9 percent) active inhaler users have withdrawn from clinical trials or stopped using the inhaler because of side effects. The potential for inappropriate use appears to be fairly low, with between 2 to 16 percent of active inhaler users continuing to use the device at six months postcessation in clinical trials allowing unrestricted inhaler use (Tonnesen et al. 1993; Schneider et al. 1996; Hjalmarson et al. 1997). Effect on Postcessation Change in Body Weight Two placebo-controlled inhaler trials have exam- ined postcessation weight gain (Tonnesen et al. 1993; Hjalmarson et al. 1997). Neither study found evidence that active inhaler use prevented or reduced weight gain among successful quitters. Relevant Process Measures The nicotine inhaler is thought to act by reliev- ing withdrawal symptoms (Glover 1993a; Leischow 1994), but little published evidence to date supports this contention. It is often suggested that the inhaler may be effective because it more closely resembles smoking than other pharmacotherapies do, replacing some of the orosensory and behavioral aspects of smoking (Glover 1993a; Tonnesen et al. 1993; Leischow 1994; Schneider et al. 1996; Hjalmarson et al. 1997). Schneider and colleagues (1996) asked partici- pants to rate their assigned inhalers relative to their usual brand of cigarettes in terms of sensory effects, preference, and satisfaction. Results showed that par- ticipants given the active inhaler rated their devices more highly than did participants given placebo. How- ever, the absolute magnitude of the ratings revealed that the inhalers did not compare very favorably to cigarettes in either group. The mechanism of action of the nicotine inhaler would seem to require further scrutiny. Bupropion Bupropion is an atypical antidepressant that is believed to work by blocking neurotransmitter reuptake in noradrenergic and dopaminergic sites in the central nervous system (Ascher et al. 1995). Anec- dotal reports of spontaneous smoking cessation in patients prescribed bupropion for depression, coupled with a growing appreciation of the importance of nega- tive affect and clinical depression in smoking mainte- nance (Hall et al. 1994; Piasecki et al. 1997) have recently stimulated clinical investigations of a sustained-release bupropion preparation as an aid to smoking cessation. These investigations led to the approval of a smoking cessation indication for bupropion by the FDA in 1997. The typical dosing regi- men for smoking cessation consists of 150 mg sustained-release bupropion per day for three days, followed by 150 mg twice a day thereafter. Therapy is initiated one to two weeks before the target quit date and is generally continued for three months. Two large-scale clinical trials of bupropion's ef- ficacy as a smoking cessation aid have been published to date. Hurt and colleagues (1997) compared three doses of bupropion (100 mg, 150 mg, and 300 mg) with placebo. Abstinence rates in the 150-mg and 300-mg groups were significantly higher than those of the pla- cebo group at 12 months. All active treatment groups were found to have higher abstinence rates than the placebo group at earlier end points. Jorenby and col- leagues (1999) studied active and placebo patches and active and placebo bupropion in a 2 x 2 factorial de- sign. Abstinence rates after one year showed no dif- ference between patch-only and placebo groups (16 percent and 15 percent, respectively). Both placebo and patch treatments were associated with higher ab- stinence rates when given with bupropion. Thirty percent of the bupropion-only group (150 mg twice a day) were abstinent at 12 months, whereas 36 percent of participants given active patches and bupropion were counted as abstinent. A recent meta-analysis (Fiore et al. 2000) of two studies reported a pooled OR of 2.1 and an estimated abstinence rate of 30.5 percent (Table 4.3). Thus, the available evidence suggests that bupropion is an ef- fective aid to smoking cessation, and that it may im- prove quit rates over those observed with conventional nicotine replacement therapies, although further stud- ies will be needed to demonstrate such efficacy. Effect on Discomfort of Nicotine Withdrawal The evidence concerning bupropion's ability to suppress withdrawal symptoms is somewhat mixed. Hurt and colleagues (1997) found that their groups using 150 mg and 300 mg reported withdrawal Mnnngcment of Nicotirrr Addictior~ 121 symptoms that were equivalent to those reported by placebo participants. Individuals assigned to the IOO-mg group, however, reported withdrawal that was significantly worse than that among either the placebo group or the other bupropion groups. The authors sug- gested that this effect may have arisen because the lOO-mg dose produced side effects similar to with- drawal symptoms but was not strong enough to re- duce true withdrawal symptoms. Jorenby and colleagues (1999) found that all three groups receiv- ing active treatments compared with the placebo group reported reduced withdrawal. The group given both active patches and active bupropion reported the most consistent withdrawal relief. Further research is needed to characterize the reliability and magnitude of bupropion effects on withdrawal symptoms. Relevant Process Measures Although nicotine replacement therapies are strongly predicated on the assumption that nicotine will relieve withdrawal symptoms, withdrawal relief represents only one of several rationales for using bupropion as a smoking cessation aid. One hypoth- esis is that bupropion may selectively reduce depres- sive symptoms after cessation. However, both trials mentioned previously excluded individuals with cur- rent major depression. Both clinical trials (Hurt et al. 1997; Jorenby et al. 1999) also included multiple as- sessments of postcessation depressive symptomatol- ogy, and neither found any differences among treatment groups on these measures. These findings suggest that bupropion does not work through its an- tidepressant effects per se in relatively healthy clinical trial participants. Bupropion moderates dopaminergic activity in the central nervous system, and dopaminergic circuits are known to play a role in drug reinforcement (Nutt 1997). This raises the possibility that bupropion may exert its effects by replacing positive reinforcement associated with smoking (Hurt et al. 1997). To date, there is no evidence directly bearing on this hypothesis, and it is clear that this process is not easily studied in clinical trials. Laboratory-based pharmacokinetic and neuroimaging studies should be performed to explore this hypothesis. Effects of Postcessation Change in Body Weight Hurt and colleagues (1997) found evidence for a dose-response effect among continuous abstainers, suggesting that participants given the highest doses gained less weight after quitting. Moreover, the dis- parities between treatment groups in terms of weight gain increased across time while medication was dis- pensed. At six-month follow-up, 17 weeks after par- ticipants went off the assigned medication, no differences in weight gain were observed. These com- parisons were limited to a small subsample of continu- ous abstainers. In the Jorenby and colleagues (1999) trial, members of all active treatment groups tended to gain less weight than did placebo participants over the first seven weeks of cessation. Weight gain suppression was greatest for the combined patch- bupropion group. However, none of the groups dif- fered in weight gain after seven weeks after quitting. Together, the results of these trials suggest that bupropion treatment may delay, but not prevent, postcessation weight gain. Side Effects In both clinical trials, two side effects were re- ported more commonly among participants given bupropion than among those given placebo. Dry mouth was reported by 10 to 15 percent of bupropion users, and insomnia was reported by about 30 to 40 percent of bupropion users. Bupropion may increase the risk of seizure and is thus contraindicated for in- dividuals who are seizure prone, such as individuals with a history of alcoholism or alcohol abuse, eating disorder, seizure disorder, or using MAO inhibitors. No seizures were reported in either clinical trial, but participants with risk factors for seizure were excluded from each before enrollment. Clonidine Clonidine is a centrally acting cc,-adrenergic agonist that dampens sympathetic nervous system activity. Clonidine is most commonly used in the man- agement of hypertension; it has not been approved by the FDA as an aid to smoking cessation. Clonidine is available for prescription in oral and transdermal forms; both of these preparations have been investi- gated in smoking cessation trials. Smokers using clonidine as an aid to smoking cessation are generally started on the drug several days before quitting and are maintained on a fixed daily dose for several weeks. Efficacy Covey and Glassman (1991) conducted a meta- analysis of nine early trials of clonidine for smoking cessation. They found that persons given clonidine were more successful at quitting than those given a pla- cebo (OR = 2.36). Five of the nine trials assessed out- come after the therapy was discontinued; only one 122 Chnpter 4 Rrducing Tobacco Use (Glassman et al. 1988) showed a significant overall ad- vantage for clonidine. Clonidine trials using adjunc- tive behavioral therapy were associated with greater relative success (OR = 4.2) than were trials in which treatment essentially consisted of dispensing the drug (OR = 1.7). Trials using transdermal clonidine produced somewhat greater relative success (OR = 3.2) than did trials using oral clonidine (OR = 2.2). The two trials that analyzed efficacy according to sex found clonidine to be much more effective, relative to placebo, among women (OR = 11.0) than among men (OR = 0.9). There is no obvious explanation for this finding. Since the Covey and Glassman (1991) meta- analysis, several large-scale clonidine trials have ap- peared (Prochazka et al. 1992; Glassman et al. 1993; Hilleman et al. 1993; Niaura et al. 1996). These studies indicated a therapeutic effect for clonidine, with some evidence suggesting that clonidine was more effective among women (Glassman et al. 1993; Hilleman et al. 1993) and among those most dependent on nicotine (Glassman et al. 1993). A recent meta-analysis (Fiore et al. 2000) of five clinical trials reported a pooled OR for long-term effectiveness of 2.1 (25.6 percent abstinence rate) (Table 4.3). In these studies, the clonidine dose ranged from 0.1 mg to 0.75 mg per day and was delivered either orally or transdermally. Because of the side effects, the lack of a specific dosing regimen, the prob- lems with abrupt discontinuation of the drug, and the lack of FDA approval, clonidine has been recom- mended as a second-line agent for smoking cessation (Fiore et al. 2000). Effect on Discomfort of Nicotine Withdrawal An early report (Glassman et al. 1984) that clonidine could reduce tobacco withdrawal symptoms, especially craving, spurred the initial investigations of clonidine's usefulness in smoking cessation. Since that report, evidence for this effect has been mixed. Clonidine- and placebo-treated patients have had equivalent levels of withdrawal severity (Wei and Young 1988; Franks et al. 1989; Gourlay et al. 1994). Studies have fairly consistently found that clonidine diminishes the specific symptom of craving (Glassman et al. 1984; Ornish et al. 1988; Prochazka et al. 1992; Gourlay et al. 1994), and some studies have found some effects on withdrawal symptoms, such as anxiety and irritability (Ornish et al. 1988; Prochazka et al. 1992). Side Effects Unpleasant side effects are commonly associated with clonidine use (Gourlay et al. 19941, and as many as 25 percent of patients may discontinue clonidine therapy because of them (Covey and Glassman 1991). The most frequently observed symptoms are dry mouth, fatigue, and dizziness. Local skin irritation is common with transdermal clonidine therapy. The in- cidence of side effects appears to be dose dependent (Gourlay et al. 1994). Care must also be taken to dis- continue clonidine gradually to prevent rebound hy- pertension. No published clinical trials have assessed the effect of clonidine on postcessation weight gain. Relevant Process Measures Clonidine is presumed to exert its effects by ame- liorating withdrawal discomfort (Glassman et al. 1984; Franks et al. 1989). Although a few studies have found that clonidine reduces withdrawal discomfort, find- ings from a well-designed, large-scale multicenter trial (Prochazka et al. 1992) have suggested that this effect does not necessarily lead to greater abstinence. Nortriptyline Nortriptyline is a tricyclic antidepressant that blocks reuptake of norepinephrine and serotonin. As with clonidine, smoking cessation is not an FDA- approved indication for nortriptyline; its primary indication is for the treatment of depressive symptoms. It is a prescription medication and is available in ge- neric form. In smoking cessation studies conducted to date, treatment was initiated 2-4 weeks before the target quit date with gradual titration of dose. Efficacy Two studies have assessed the efficacy of nortrip- tyline for smoking cessation. Hall and colleagues (1998) conducted a 2 (nortriptyline vs. placebo) x 2 (his- tory vs. no history of major depression) x 2 (cognitive behavioral vs. health education therapy) trial that pro- duced a 24-percent sustained abstinence rate in nortrip- tyline users compared with 12 percent in the placebo group. There was no difference in cessation rates as a function of previous history of major depression. In a straight comparison of nortriptyline to placebo, Prochazka and colleagues (1998) found cessation rates at six months of 14 percent in participants given nortriptyline and 3 percent in participants given pla- cebo. A meta-analysis (Fiore et al. 2000) of these two studies reported a pooled OR of 3.2 and a 30.1-percent abstinence rate (Table 4.3). Both studies provide clear evidence of nortriptyline's therapeutic effect. Surgeon Gcrzeral's Report Effect on Discomfort of Nicotine Withdrawal The Hall and colleagues (1998) study assessed both nicotine withdrawal symptoms and negative af- fect in the first eight days following the target quit date. There were no significant differences between the drug therapy groups on nicotine withdrawal severity, sug- gesting that as with many of the other smoking cessa- tion pharmacotherapies, withdrawal relief may not be the primary mechanism of action. The negative affect measure, however, increased in the first three days in the placebo group and declined in the nortriptyline group. This suggests that a negative affect assessment may be more sensitive to some of nortriptyline's thera- peutic effects than a conventional nicotine withdrawal symptom scale. Side Effects Tricyclic antidepressants are known to produce a number of side effects, including sedation and vari- ous anticholinergic effects. In the smoking cessation studies, commonly reported side effects included dry mouth (64-74 percent), lightheadedness (49 percent), shaky hands (23 percent), and blurry vision (16 per- cent) (Hall et al. 1998; Prochazka et al. 1998). Other Antidepressants and Anxiolytics Investigators have begun to explore the poten- tial use of other antidepressants and anxiolytics as pharmacologic aids to smoking cessation, because population-based epidemiologic samples have found that depression and anxiety are associated with ciga- rette smoking (Breslau et al. 1991; Kendler et al. 1993). Research has also shown that smokers with a history of depression are more likely to experience depressive symptoms (Covey et al. 1990) and to relapse after quit- ting (Glassman et al. 1988; Anda et al. 1990) than are smokers without such a history. Some anxiolytics (Glassman et al. 1984; Hilleman et al. 1992) have been shownto ameliorate symptoms of tobacco withdrawal, and preliminary smoking cessation trials using anti- depressants (Edwards et al. 1989) and anxiolytics (Hilleman et al. 1994) have yielded encouraging re- sults. Among the drugs that have been studied or hypothesized to be useful for smoking cessation are buspirone hydrochloride, doxepin hydrochloride, and fluoxetine hydrochloride. Although promising, this avenue of research is not yet developed enough to permit the multipart discussion given to other phar- macologic agents in this chapter. Summary of Pharmacologic Interventions Abundant evidence confirms that both nicotine gum and the nicotine patch are effective aids to smok- ing cessation. The efficacy of nicotine gum may de- pend on the amount of behavioral counseling with which it is paired. The 4-mg dose may be the better pharmacologic treatment for heavy smokers or for those highly dependent on nicotine. The nicotine patch appears to exert an effect independent of behavioral support, but absolute abstinence rates increase as more counseling is added to patch therapy. Nicotine nasal spray and nicotine inhalers are effective aids for smok- ing cessation, although their mechanisms of action are not entirely clear. All nicotine replacement therapies produce side effects, but these are rarely severe enough that patients must discontinue use. Nicotine nasal spray appears to have greater potential for inappro- priate use than other nicotine replacement therapies. Nicotine replacement therapies, especially the gum and the patch, have been shown to delay but not pre- vent weight gain. All nicotine replacement therapies are thought to work in part by reducing withdrawal severity. The available evidence suggests that they do ameliorate some elements of withdrawal, but the relationship between withdrawal suppression and clinical outcome is inconsistent. Bupropion is the first nonnicotine pharma- cotherapy for smoking cessation to be studied in large- scale clinical trials. Results suggest that bupropion is an effective aid to smoking cessation. In addition, bupropion has been demonstrated to be safe when used jointly with nicotine replacement therapy. In the only direct comparison with a nicotine replacement product, bupropion achieved quit rates about double those achieved with the nicotine patch. Bupropion appears to delay but not prevent postcessation weight gain. The available literature contains inconsistent evidence regarding bupropion-mediated withdrawal relief. Bupropion does not appear to work by reduc- ing postcessation depressive symptomatology, but its mechanism of action in smoking cessation remains unknown. Further research is needed to characterize bupropion's central nervous system effects, particu- larly to assess whether the drug partially replaces smoking-related positive reinforcement. Evidence suggested that clonidine is capable of improving smoking cessation rates. Clonidine is hy- pothesized to work by alleviating withdrawal symp- toms. Although clonidine may reduce craving for cigarettes after cessation, it does not consistently ame- liorate other withdrawal symptoms, and its effects on weight gain are unknown. Unpleasant side effects are common with clonidine use. 124 Chapter 4 Antidepressants and anxiolytics are potentially useful agents for smoking cessation. At present, only nortriptyline appears to have consistent empirical evi- dence of smoking cessation efficacy. However, tricy- clic antidepressants produce a number of side effects, including sedation and various anticholinergic effects. Large-Scale Public Health Programs The shift in recent years from a clinical to a pub- lic health perspective in smoking cessation research has led to an increased emphasis on developing and evaluating cost-effective strategies that can be tvidely disseminated (Lichtenstein and Glasgow 1992). This emphasis is reflected in the proliferation of research on self-help manuals (see "Self-Help Manuals," ear- lier in this chapter and "Community Programs," later in this chapter) and on media- and community-based interventions (Flay 1987; Gruman and Lynn 1993). As is true for self-help strategies, media-, worksite-, and community-based programs have promise because they can potentially reach many smokers who may try to quit without formal, face-to- face assistance (Fiore et al. 1990). Moreover, some evi- dence suggests that less educated smokers profit from media campaigns at least as much as more highly edu- cated smokers do (Macaskill et al. 1992). (Other large- scale interventions-educational [Chapter 31 and social [Chapter 7]-are discussed separately.) Investigators have evaluated an array of such programs, but methodological variations across the individual trials have hampered comparisons among studies (Flay 1987; Schwartz 1992). Moreover, meth- odological challenges compromise how research on these programs may be interpreted. For instance, on- going coverage of smoking and its health consequences in the general media may alter the effect of research- based media information. Similarly, secular trends and events that could individually affect large populations of smokers (e.g., the introduction of a new nicotine replacement product) may alter the impact-and complicate the assessment-of media campaigns conducted around the time of such events. Such chal- lenges may account for the inconsistencies seen in this area of research. Media-Based Programs Media used to transmit smoking cessation mes- sages have included television (Brannon et al. 1989; Korhonen et al. 1992; Mudde and De Vries 1999), ra- dio (Farquhar et al. 1990; COMMIT Research Group 19911, the telephone (Ossip-Klein et al. 1991; Pierce et al. 1992), newspapers (Cummings et al. 1987), and the mail (Gritz et al. 1992; McFall et al. 1993). The intensity of media-based programs has var- ied greatly, and these variations may be related to pro- gram success. For example, one study (Gritz et al. 1992) evaluated a minimal mail-based intervention. The in- vestigators mailed self-help smoking materials to a sample of nonvolunteer women who smoked and who belonged to a health maintenance organization. The intervention had no impact; at no point during the 1% month follorv-up period were women ivho had re- ceived the materials more likely to quit smoking or report changes in their moti\.ation to quit than women who had not. In contrast, a more intense media cam- paign evaluated in another study (Orleans et al. 1991) yielded encouraging findings, albeit among treatment volunteers. The investigators tested the impact of add- ing telephone calls from a smoking cessation counse- lor to an intervention that mailed self-help manuals to the volunteers. After 16 months, abstinence from smoking was reported by 23.0 percent of the volun- teers who had received adjuvant telephone counsel- ing and by 15.2 percent of those recei\?ng the self-help materials alone. Mass media campaigns of intermediate intensity, such as televised programs (Flay et al. lY8Y), gener- ally produce modest increases in abstinence-increases that fall short of the moderate effect of telephone coun- seling found among volunteers (Orleans et al. 1991). The influence of intermediate-intensity interventions is difficult to determine precisely, because the results of individual trials may be affected by the peculiari- ties of the specific communities in which they are tested and (as previously discussed) by concurrent changes in secular attitudes toward smoking behavior. These problems are compounded by the designs of communitywide and mass media programs frequently failing to include matched control communities for com- parison. Although more intensive interventions appear to increase cessation over time (Flay 1987), the absence of well-controlled experimental media trials limit any conclusions about a dose-response relationship for media-based programs. The content of various media-based programs can be divided into three categories: (1) programs that present information about the negative health effects of smoking and exposure to secondhand smoke and attempt to motivate smokers to quit; (2) programs that promote the performance of simple cessation-related activities, such as calling a hot line, requesting self- help materials, or enrolling in a smoking cessation contest; and (3) programs that mimic intensive clini- cal interventions (Flav 1987). In general, informational or motivational campaigns can be effective in chang- ing smokers' attitudes, but the effect of such campaigns on behavior is not clear, in part because of the paucity of well-controlled trials that yield a consistent pattern of findings. Research suggests that other types of cam- paigns have greater potential than informational pro- grams to influence smoking behavior, especially if the campaign has multiple components and intense ex- posure (Flay 1987; CDC 1996,1999b; Pierce et al. 1998). Worksite Programs For many years, advocates for tobacco control have been enthusiastic about worksite-based programs, because worksites appear to furnish an ideal setting: a contained audience, an opportunity for smoker partici- pation, an environment in which to convey coherent and consistent messages, and an opportunity to tie in- dividual smoking cessation to overarching institutional policy. Much of the early work in this area provided some justification for the enthusiasm (USDHHS 1986; Glasgow 1987; Fielding and Piserchia 19891, but more recent data, described later in this section (Glasgow et al. 1995; Sorensen et al. 19961, give pause. The main components of smoking cessation efforts in the workplace are nonsmoking policies and specific assistance for cessation attempts (Gruman and Lynn 1993). The evolution of worksite smoking policies, in- timately tied to concerns about the health effects of en- vironmental tobacco smoke (ETS) (Eriksen 1986; USDHHS 1986), is described in some detail in Chapter 5. Although early assessment suggested that restric- tive policies had little effect on smoking outside of work (Glasgow 1987; Rigotti 1989; Tager 1989), most recent studies have demonstrated either reductions in daily consumption of cigarettes (Stillman et al. 1990; Borland et al. 1991; Jeffery et al. 1994) or increases in smoking cessation (Stave and Jackson 1991; Patten et al. 1995; Longo et al. 1996). As described in Chapter 5 (see "Clean Indoor Air Regulation"), there is persistent movement toward increasing restrictions in public workplaces. The strategies for smoking cessation within workplaces are largely those discussed earlier in this chapter: self-help, physician's advice, and formal treat- ment (Gruman and Lynn 1993). As of 1989, about one- half of worksites that sponsored cessation activities offered self-help materials (Fielding and Piserchia 1989). Although initial dropout rates were high, 20-26 percent of participants had quit smoking by 6-12 months after the worksite programs had begun (Orleans and Shipley 1982; Glasgow 1987). Such proportions compare favorably with those observed in general populations. Physician's advice to quit smoking was a component of only about 15 percent of the company programs, but in a number of studies, this modality seemed to exert an effect similar to that observed in general populations: 15-30 percent of par- ticipants had quit smoking at the one-year follow-up (Gruman and Lynn 1993). The programs offering for- mal treatment appeared to produce results at the worksite that were similar to those found for such pro- grams outside the workplace. A special feature of worksite cessation programs is the opportunity to provide incentives, such as com- petitions. Several studies have documented some ef- ficacy in this approach. For example, in one study, 33 percent of participating workers and 25 percent of all workers remained abstinent at work (Glasgow 1987). In a second study, the use of a competition was associ- ated with significantly greater success at quitting than was reported for persons not participating in the com- petition (Klesges et al. 1988). In a review of incentive programs, from 15 to 60 percent of participants quit smoking; the average was around 40 percent (Gruman and Lynn 1993). Some disadvantages of incentives are that (1) determining the award may be difficult, (2) employees may falsely claim cessation, and (3) non- smokers may feel slighted (Fiore et al. 1996). On a population basis, incentives have not been found to be effective. In these settings, incentives may be most attractive to smokers who were going to attempt quit- ting in any case (Chapman et al. 1993). In contrast, a trial of the Take Heart program, which involved 26 heterogeneous worksites, a low-cost intervention, random assignment, and use of worker and management steering committees, failed to pro- duce short-term improvements in smoking cessation that exceeded the secular trend (Glasgow et al. 1995). These results were particularly disheartening in view of the methodological strengths of the study and the diversity of the workplace settings. The authors offer a number of potential reasons for the lack of impact: the cessation activities may have been inappropriate; the behaviors may have been more resistant to change than previously assumed; workers may have had in- sufficient "ownership" of the project; secular trends may have been so strong that they canceled out a mod- est effect; the variability among worksites may have been too great; and, in general, worksite programs may not work. Similar negative findings were observed by Sorensen and colleagues (1996) in an even larger trial of 111 worksites randomized to sites receiving or not receiving the cessation program. The Working Well Trial involved more than 28,000 workers in 16 states and compared seven-day abstinence, six-month Reclucirlg Tobacco Usr abstinence, and changes in smoking prevalence for both types of \%rorksites. Changes occurred in the di- rection hypothesized, but they were small and non- significant; for example, the six-month abstinence rate was only 1.5 percent higher in the program group. Similarly, the program sites showed a nonsignificant trend toward greater adoption of smoking bans. The authors observed that the overall cessation proportions at both types of sites compared favorably with those in other worksite programs. The lack of difference may have resulted from the higher than expected cessation at control sites, which is a phenomenon reflecting a general increase in antismoking awareness. These studies postdate recent reviews of worksite cessation efforts. Several early reviews expressed op- timism about the value of worksite programs but did not provide a quantitative assessment (Hallett 1986; Bibeau et al. 1988). In a detailed meta-analysis of 20 worksite programs involving 34 comparisons, Fisher and colleagues (1990) found that the mean weighted effect size was significantly positive and that an aver- age of 13 percent of participants had quit smoking af- ter treatment. Although modest, these effects provide some quantitative basis for the enthusiasm for worksite programs. The addition of the two recent large projects (Glasgow et al. 1995; Sorensen et al. 1996) may well alter the meta-analytic balance. Although the worksite setting has aforemen- tioned features favorable to large-scale programs (in- cluding the importance of adding to a generalized reduction in exposure to ETS), the strategy cannot be recommended without qualification. Nonetheless, the role of such activities, perhaps enlightened by further targeted research, may be important in multicompo- nent efforts at smoking cessation. Community Programs Results from a number of long-term trials of communitywide programs have recently appeared. (See Chapter 7 for a more detailed discussion of these projects in the context of approaches used in the 1990s.) These trials typically incorporate mass media strate- gies into larger health education programs. Some, such as the Stanford Five-City Project (Farquhar et al. 1990), the Minnesota Heart Health Program (Perry et al. 1992; Luepker et al. 19941, and the Pawtucket Heart Health Program (Elder et al. 1986; Carleton et al. 19951, have been aimed at modifying smoking, as well as other risk factors for cardiovascular disease. Final reports suggest that these trials have met with little success in promoting smoking cessation. The Stanford Five-City Project (Farquhar et al. 1990; Fortmann et al. 1993) tested an intensive multi- media approach, including television, radio, newspa- per, and mass-distributed printed materials. All materials contained information about modifiable risk factors for cardiovascular disease. The average resi- dent of a community receiving the program was ex- posed to more than 500 educational episodes over the course of the five-vear program. By the end of this period, smoking pievalence-the only risk factor on which an impact could be demonstrated-had declined 13 percent more in the program communities than in the control ones. The Minnesota Heart Health Program failed to demonstrate an appreciable impact (Land0 et al. 1995). The Pawtucket Heart Health Program had little impact on smoking behavior; its first attempt at a smoking cessation program prompted only 11 smokers to quit (Elder et al. 1986, 1987). The final results con- firmed the lack of impact (Carleton et al. 1995). One ambitious community project-COMMIT (Community Intervention Trial for Smoking Cessation)-focused on smoking cessation and on policy strategies to reduce prevalence (COMMIT Re- search Group 1991; Gruman and Lynn 1993). In 1986, the NC1 began COMMIT, the largest randomized smoking intervention trial in the world. The design of COMMIT included 11 pairs of matched communities- 10 from across the United States and 1 in Canada. One community from each pair was randomly selected to be the site in which volunteers and local agencies car- ried out COMMIT's 58 mandated program activities. Designed to augment existing community-based efforts to reduce smoking, these activities occurred between 1988 and 1992. The primary end point for COMMIT was smok- ing cessation among heavy smokers. Main goals in- cluded increasing the priority of smoking as a public health issue, increasing the community's ability to in- fluence smoking behavior, strengthening the community's existing economic and policy factors designed to discourage smoking, and fortifying social norms and values that stressed nonsmoking (Gruman and Lynn 1993). Main strategies included training health care providers to routinely assess and manage nicotine dependence, working with community insti- tutions and private organizations to create smoke-free environments, increasing the availability and visibil- ity of smoking cessation services, and using the mass media and schools to educate communities about the dangers of tobacco use. Results of COMMIT indicate that even intensive community-based programs may not have a demon- strable impact on smoking behavior (COMMIT