because of tobacco withdrawal and that the improvement in performance occurred because smoking relieves tobacco withdrawal (Schachter 1979; Silverstein 1982). This latter interpretation assumes that overnight deprivation induces withdrawal; although this assumption has not been tested directly, withdrawal effects can occur after only 12 hours of deprivation (Hughes, Higgins, Hatsukami 1990). Ideally, studying smokers before initiation would allow comparison of this baseline with before and after a smoking episode. As this is impractical, one solution has been to add a control group of nonsmokers (Hughes, Higgins, Hatsukami 1990). For example, smokers performed better after smoking and the same as nonsmokers in several studies of errors on a vigilance task (Taylor and Blezard 1979; Hughes, Keenan, Yellin 1989; Lyon et al. 1975; Heimstra et al. 1980; Tong et al. 1977; Tarriere and Hartmann 1983; Keenan, Hatsukami, Anton 1989) and a tracking task (Lyon et al. 1975) (Figure 1, upper panel). The effect was attributed to relief of withdrawal. One study provided evidence for enhancement of performance from smoking inde- pendent of reversing withdrawal. Wesnes and Warburton (1978) reported a pattern consistent with enhancement when errors on vigilance tasks were studied (Figure I, lower panel). Other indirect evidence can be used to test the withdrawal relief versus enhancement models. Two studies reported enhancement of tracking or motor skills when smokers were not deprived (Parrott and Winder 1989; Hindmarch. Kerr, Sherwood 1990; Larson, Finnegan, Haag 1950; Pomerleau and Pomerleau 1986). Several studies have examined the effect of cigarette smoking or nicotine administration on the performance of nonsmokers (Dunne, MacDonald, Hartley 1986; Hindmarch, Kerr, Sherwood 1990; Wesnes, Warburton. Matz 1983; Wesnes and Revel1 1984; West and Jarvis 1986: Wesnes and Warburton 1984). In two studies, the improvement in nonsmokers was similar to that of deprived smokers (Wesnes, Warburton, Matz 1983; Wesnes and Revel1 1984). One study reported performance to be similar between deprived smokers and nonsmokers (Warburton 1990). Finally, nicotine appears to improve the perfor- mance of animals not previously exposed to nicotine (Clarke 1987; Emley and Hutchin- son 1984). In summary, the results of studies to assess if smoking increases performance through withdrawal relief or by direct enhancement appear contradictory. One possible ex- planation of this discrepancy is that smoking may increase performance through both withdrawal relief and direct enhancement. The specific mechanism that is operative may vary not only among smokers but also within smokers across situations. Variability in Withdrawal Whereas the necessary and sufficient condition to establish dependence is repeated exposure to the drug, other factors may exacerbate nicotine withdrawal symptoms. Although several investigators have commented on the variability of postcessation symptoms, it is unclear that this variability is greater than with other drug withdrawal syndromes (Hughes, Higgins, Hatsukami 1990: US DHHS 1988). The results of retrospective and postcessation studies on self-reported withdrawal symptoms (e.g., hunger, restlessness, or inability to concentrate) among smokers who have a greater 526 60 50 - 2 NONSMOKERS 2 $j 40 - c- 5 3 e 30 I I I I I I 1 2 3 4 5 6 HOURS 0.9 r NONSMOKERS 2 DEPRIVED SMOKERS F: VJ 0.6 1 1 I I I I I I 10 20 30 40 50 60 70 80 MINUTES FIGURE I.-Uppc~ pu~wl: Performance on a meter (i.e., visual) vigilance task SOURCE: Heinhm CI al. IW). f,o~~~r~puwI: Performance on the continuous clock task, a visual vigilance task nicotine intake are inconclusive (Goldstein, Ward, Niaura 1988; Hughes, Higgins, Hatsukami 1990: Shiffman 1979: US DHHS 1988; Williams 1979). Withdrawal effects. including weight gain, have not been found to differ consistently by gender or age (Hughes, Higgins, Hatsukami 1990). Several studies have suggested that expectancy influences the effects of abstinence; that is, some individuals may amplify. deny, or misattribute their withdrawal symptoms (Barefoot and Girodo 1972; Gottlieb et al. 1987: Hughes and Krahn 1985; Hughes et al. 1989). According to the misattribution model, at times the individual can "mistake" withdrawal symptoms for other possible events. For example, in one study a labeling mistake was made when individuals were told that a placebo they were taking was alleged to have side effects similar to the effects of cigarette withdrawal (Barefoot and Girodo 1972). Three direct tests of expectancy have been published (Gottlieb et al. 1987; Hughes and Krahn 1985; Hughes et al. 1989). In one study, subjects in a double-blind trial of nicotine polacrilex gum were asked if they thought they had received nicotine or placebo gum. Those who believed they had received placebo gum had more abstinence discomfort than those who could not differentiate what they had received; this latter group had more discomfort than those who thought they had received the nicotine polacrilex gum (Hughes and Krahn 198.5). Because this study used post hoc ratings, it is unclear that the belief in which gum had been received modified the level of abstinence effects, or that the level of abstinence effects modified the belief of which gum had been received. Two experimental trials have manipulated instructions and thereby directly tested if expectancy influences abstinence effects. The first study randomly assigned smokers to a 2x2 design of contrasting instructions; subjects were told that they received either nicotine polacrilex gum or placebo gum, and actually received either nicotine polacrilex gum or placebo gum (Gottlieb et al. 1987). Most of the measures of abstinence effects were unchanged by instructions or by actual drugs. The physical symptoms and stimulation scores on the Shiffman-Jarvik Withdrawal Scale were less only on some days in the group told they were receiving nicotine than in the group told they were receiving placebo. A second study used a similar design and found that abstinence symptoms were fewer among those who received nicotine polacrilex gum than among those who received placebo gum. but found no effect of instructions (Hughes et al. 1989). In summary, the seemingly valid proposition that abstinence effects are in- fluenced by expectancy has not been completely supported by empirical tests. Abstinence effects have been hypothesized to be greater in more dependent smokers. However, the scales for dependence used to test this hypothesis vary according to whether they are quantifying physical dependence (withdrawal), behavioral depen- dence (desire for tobacco or tendency to relapse), or dependence on tobacco or on the nicotine in tobacco (Hughes 1984). The Fagerstrom Tolerance Scale (TQ) is the most widely used dependence scale (Fagerstrom 1978). TQ consists mostly of items that refer to behavioral dependence on tobacco. The total TQ score predicted total abstinence discomfort in one study (Fagerstrom 1980) and weight gain in another study (Tonnesen et al. 1988). However, two detailed studies failed to indicate that TQ 528 predicted weight gain (Emont and Cummings 1987) or self-reported withdrawal symptoms. The Reasons for Smoking Scale has two scales relevant to the dependence con- struct-the addiction scale and the negative affect scale (Ikard, Green. Horn 1969). Neither of these has been shown to predict weight gain (Boss& Garvey. Costa 1980). self-reported withdrawal (Hughes and Hatsukami 1986). or relief by nicotine polacrilex gum (Hughes and Hatsukami 1986). Russell's Smoking Motivation Questionnaire has a subscale for dependence (Russell, Peto. Pate1 1974). In one study, the scale predicted total abstinence discomfort and irritability but did not predict restlessness, depression. hunger. or inability to con- centrate (West and Russell 1985). Another measure somewhat related to dependence includes the sevoerity of abstinence discomfort in the past, which appears to predict self-reported abstinence (Hughes and Hatsukami 1986). Other generic scales, such as the MacAndrews Scale for Addiction (MacAndrew 1979) and Eysenk Personality Questionnaire (Eysenk and Eysenk 1975). do not predict abstinence discomfort and weight gain (Bosse. Garvey. Costa 1980). Although one study found that self-reported smoking for stimulation predicted abstinence effects (Niaura et al. 1989), an earlier study had found no such relationship (West and Russell 1985). In summary, the evidence that any dependence scale predicts abstinence effects is quite limited. Further tests that use scales that more specifically determine physical versus behavioral dependence and dependence on nicotine versus tobacco may provide more informative data. Timecourse of Withdrawal Several recent studies produced concordant results on the timecourse of nicotine withdrawal. Most signs and symptoms of nicotine withdrawal are readily detected within 24 hours (Hughes, Higgins, Hatsukami 1990). Previous studies have suggested that abstinence effects can occur even sooner. for example, within 2 hours (US DHHS 1988). These studies have measured effects during smoking and 2 to 6 hours post- smoking: it was noted that 2 to 6 hours after smoking. self-ratings of performance were worse than during smoking. Several investigators have interpreted the scores during smoking as representing baseline and the postsmoking scores as representing withdrawal. However. as discussed earlier. an alternate interpretation is possible: the scores 2 to 6 hours po\tsmoking represent baseline scores and the scores during smohing represent the acute effects of smoking (Hughes et al. 1990). The results of several prospective studies indicate that the \ignj and symptoms of nicotine withdrawal peak in the first I to 2 days following cessation (Gumming\ et al. 1985: Hughes and Hatsukami 1986: West et al. 1983: Shil`fmun and Jarvik 1976: Schneider. Jarvik. Forsythe 19X3) and la\t about I month (Grit/. Carr. Marcus 1990: Cummings et al. 1985: Gross and Stitrer 1989: Hughe\ 1990: Hughes t`t al. 1990: Lawrence. Amoedi. Murray 19x2: West. Hajek. Belcher 19X7). For each of IO weehs. Gross and Stitzer ( 1989) recorded symptoms of quitters and found a peak during the first week and a return to bu\eline 3 to 4 ueeh\ po\tcessation. Snyder. Davi\. and Henningfield ( I YXY) trached perfomiance on several ta\h\ over IO day\. Impairment in performance peahed at I to 2 da) \. and performance on mo\t ta\h\ I-eturned to baseline during the IO day\: however. performance on mne task\ &;I\ still impaired after IO days. A study by Gumming\ and colleague\ ( IYXS) included 33 subject\ uho hept a daily record oCX withdrawal symptoms. AI 21 days. few subjects were reporting withdrawal symptoms. with the exception of a11 occasional desire for a cigarette. A fourth study (Hughes IYYO) provided a less-detailed timecour\e but included group\ of never smokers. ex-smokers. and continuing smokers. The withdrawal scores ot abstinent smokers at I month were equivalent to their baseline score4 and to those of never smokers and continuin g smokers (Hughes IYYO). Although the average withdrawal symptom score returned to baseline at I month, 45 percent of \ub,jectr reported symptoms still above precessation levels at I -month followup (Hughe\ IYYO). Further followup of these subjects indicated that their withdrawal score\ had returned to baseline or below baseline by 6 month\ postcessation. Craving. hunger. and ueight gain are exceptions to the I -month duration: they may continue at least through the first h-months after cessation (Gritz. Carr. Marcus 1990: Hughe\ 1990; Hughes et al. IYYO: West. Hajek. Belcher 1987). With cessation of other drugs. a prolonged withdrawal syndrome has been postulated (Martin and Jasinski IY69). There is no evidence of a prolonged nicotine withdrawal syndrome. In fact, scores on withdrawal scales appear to decrease below prece\\ation levels at followup (Figure 2): that is. positive mood changes occur after Ions-term abstinence from smoking (Chapter I I. see section on long-term psychological and behavioral consequences and correlates of smoking cessation) (Grit/. Carr. Marcus 1990: Gross and Stitzer 1989: Hughes IYYO: Hughes et al. IYYO). Withdrawal as a Cause of Relapse Seven recent studies have examined nicotine withdrawal as a predictor of rclap\e. that is. whether smokers with severe withdraual are more likely to relapse. Five studies found that some withdrawal symptoms predicted relapse at some points in time (Gritz. Carr. Marcus IYYO; West. Hajek. Belcher IYYO: Hughes IYYO: Killen et al. IYYO: Swan et al. IYXX). The two studies that did not indicate such a relationship examined the ability of withdrawal to predict abstinence at very earl!, followup (Hughes and Hut- sukami 19X6) or very late followup (Hughes et al. IYYO). In the five positive studie\. mood changes. such as depression and anxiety. were the more common predictors. However, both across and within the studies. there \va\ no con\i\tent or clear grouping of symptoms predicting wlithdrawal at specific points in time. One common findin? wa\ that the number of symptoms appeared to be a predictor (Gritz. Carr. Marcus I YYO: Hughes 19YO). For subgroups of smokers. such ;I> more dependent mohers. withdrawal may be an especially important factor in relapse. hut this relationship ha\ not been demonstrated. Postcessation weight gain ha$ often been hypothesized to be a major cause of relapse. especially among women (Hall. Ginsberg. Jones 1986). Contrary to several (I /~/`ior.i hypotheses. three prospective studies have found that more weight gain predicted les\ relapse (Duffy and Hall I Y90: Hall. Ginsberg. Jones I YX6: Hughes et al. 1990). There 8 6 I I I I I I I I I I I BL 1 2 3 4 5 6 7 8 9 10 0 l- 24 WEEKS FIGURE 2.-Self-reported withdrawal discomfort among abstinent smokers SOURCE: Groa and St~trer I YXL): Hughe\ I IWO). was no gender difference in this prediction in any of the three studies. This finding is further supported by a study in which women who reported eating more in the first 4 days ofcessation were more likely to be abstinent at 6-month followup (Guilford 1966). One explanation for the weight gain-relapse finding is that food deprivation increases the reinforcing effects of drugs (Carroll and Meisch 1984). Cessation of smoking may decrease metabolic rate (Perkins, Epstein, Pastor 1990); if this is true, to avoid weight gain, smokers may deprive themselves of food and thereby increase the reinforcing effects of cigarettes smoked during periods of relapse. In summary, this recent evidence shows that smokers with more severe withdrawal symptoms are more likely to relapse. However, these results should not be misinterpreted. First, prediction is not equivalent to causality: withdrawal symptoms may predict relapse, not because they cause relapse, but because they are associated with some other variable, such as degree of dependence. Second, those symptoms that predict the occurrence of relapse and the timing of relapse-very early (<2 days), early (2-10 days), or later (IO-30 days)-vary across studies. Third. although studies have shown that withdrawal is an early predictor of relapse, these studies have not shown that withdrawal predicts eventual outcome (i.e., long-term abstinence). Summary Strong evidence indicates that smokers who stop smoking experience a nicotine withdrawal syndrome that includes the short-term consequences of anxiety, irritability. frustration, anger. difficulty concentrating, and restlessness. These symptoms general- ly occur within 24 hours and subside after about 1 month. Smokers also report strong cravings or urges to smoke when they are not smoking; this symptom will persist among some former smokers. Hunger and weight gain may also persist longer than I month. Abstinence does not appear to affect short-term caffeine intake. However, it does increase caffeine metabolism, which may mimic or potentiate symptoms of nicotine withdrawal. There are conflicting data on the short-term effects of smoking abstinence on alcohol intake. However, the data suggest that smokers attempting permanent smoking abstinence experience decreased alcohol intake. Research on the effects of smoking abstinence on performance indicates that abstinence impairs performance on attention tasks. This impairment may persist for at least 7 to IO days and is relieved by nicotine replacement. Other more complex types of tasks as well as memory and learning have not been clearly shown to be impaired by abstinence. The relation of improvement in attention tasks with nicotine may be due either to withdrawal relief or to performance enhancement; findings are consistent with both models. However. evidence more strongly suggests withdrawal relief from receiving nicotine. Variability in tobacco withdrawal symptoms resembles that observed for other drug withdrawal syndromes. Several studies have suggested that expectancy influences withdrawal effects. However, this has not been completely supported by empirical tests. Although abstinence effects have been hypothesized to be greater in more dependent smokers, the evidence is conflicting. Recent data indicate that smokers with more severe withdrawal symptoms are more likely to relapse. However, no symptoms or groups of symptoms consistently predict relapse at any given point in time. LONG-TERM PSYCHOLOGICAL AND BEHAVIORAL CONSEQUENCES AND CORRELATES OF SMOKING CESSATION Introduction Most long-term studies of self-quitters or smokers taking part in treatment programs only include data on smoking behavior or smoking status (Adesso 1979: Gordon and Cleary 19X6; Orleans and Shipley 1983; Shipley. Rosen. Williams 1982); followup measures of psychological and behavioral consequences are rarely included. Thu\. although former smokers represent a large and growing segment of the U.S. population (Volume Appendix). the long-tern1 psychological and behavioral consequences of smoking cessation have not been well studied. Very few studies of former smoker-j have employed prospective or longitudinal designs: rather. most have used retrospective or cross-sectional designs. In the typical retrospectivse study. subjects are asked whether after quitting or during their experience of trying to quit, they were more or less nervous. irritable. depressed. sedentary. or health conscious than before quitting. While relevant to the experience of a person abstaining from tobacco, retrospective studies potentially suffer from several limita- tions, including the absence of information about baseline group similarities or differ- 532 ences and the problem of recall bias. (See Chapter 2 for a discussion of methodologic problems.) Successful former smokers may minimize or fail to recall their difficulties or exaggerate their prowess (Heinold et al. 1982): recidivists may exaggerate withdrawal problems to justify their relapse (Graham and Gibson 1971). Cross- sectional studies do not permit the establishment of comparability at baseline. Con- clusions from the data are therefore limited, often identifying the correlates of cessation rather than the consequences. Both consequences and correlates of cessation will be discussed in this Section. Most prospective studies of smoking cessation sequelae have been conducted with smokers participating in formal treatment programs rather than with smokers quitting on their own (Hughes, Higgins. Hatsukami 1990). Treatment participants may differ in several ways from self-quitters. In a recent review of findings concerning short-term withdrawal effects, Hughes, Higgins. and Hatsukami ( 1990) noted that self-quitters had fewer and less severe withdrawal symptoms than treated quitters; they noted. as did Schachter (1982), that clinic populations may include a higher proportion of hardcore, highly dependent smokers. On the other hand, treated quitters may learn new coping skills such as relaxation, self-reward, or exercise and gain additional support for their initial quitting efforts. Therefore, their short-term postquitting experiences may not be representative of the 90 percent of former smokers who quit on their own (US DHHS 1988; Fiore et al. 1990). Thus. in drawing conclusions from studies of participants in treatment programs, it is important to be aware of the possible differences between these two populations of abstainers. Mood, Anxiety, Perceived Stress, and Psychological Well-Being Tobacco use has often been described as a maladaptive response to. or a way to cope with, life stress and a way to regulate negative affect (Tomkins 1966: Billings and Moos 1981: Ockene et al. 1981: Orleans 1985; Abrams et al. 1987). Smokers often believe that smoking helps them cope with stress and anxiety (Ikard, Green, Horn 1969). Thus. in addition to the stress of separation from cigarettes (Tamerin 1972). abstaining from cigarettes potentially could make the smoker feel less able to cope with stress (Abrams et al. 1987: Marlatt and Gordon 1985) and thereby constitute a biologically based source of stress (Grunberg and Baum 1985). If the quitter feels unable to cope with stress without cigarettes, perceived stress may increase, and self-efficacy may decrease. resulting in heightened anxiety and an overall negative shift in well-being. Alterna- tively, Cohen and Lichtenstein (in press) have hypothesized that for smokers who want to quit smoking, continued smoking may prove more stressful than cessation. and quitting smoking may result in a more positive self-appraisal and heightened feelings of self-esteem and personal competence. Similarly, other researchers have proposed that smoking may cause negative self-evaluations and feelings of guilt and helplessness among smokers who want to quit. so that quitting would result in an overall long-term improvement in mood, self-image, and 5elf-esteem (Frerichs et al. I98 I : Knudsen et al. 1984: Schwartz and Dubitzky 1968). Possible long-term changes in anxiety levels after quitting might also reflect quitting- related changes in physiologic stres\ reactivity (Abrams et al. 1987). To the extent that smoking contributes to excess physiologic stress reactivity and more ready arousal to anxiety (Emmons et al. 1986; Williams, Hudson, Redd 1982: US DHHS 1988). cessation might lead to stable reductions in general anxiety. Several models have been proposed to understand the possible long-term conse- quences of smoking cessation for depression or dysphoria (Frerichs et al. I98 I ; Hughes 1988; Hughes, Higgins, Hatsukami 1990; Tamerin 1972). Studies of withdrawal effects have found depressed mood or dysphoria to be a common, transient withdrawal effect, partly reflecting multiple pharmacologic effects of nicotine abstinence (Backon 1983; Hughes, Higgins, Hatsukami 1990; US DHHS 1988). Covey, Classman, and Stetner (in press) found that smokers with a history of major depression had more severe symptoms of depression 2 weeks after a behavioral treatment for smoking than those without such a history. However, some theorists have proposed that for smokers who want to quit, quitting could result in improved mood. well-being, and self-esteem (Frerichs et al. 198 1). Research Results Five cross-sectional studies have compared former smokers with continuing smokers or relapsers on measures of mood, affect. anxiety, and psychological well-being (Abrams et al. 1987; Giannetti, Reynolds, Rihn 1985; Orleans et al. 1983; Pederson and Lefcoe 1976; Pomerleau, Adkins, Pertschuk 1978). Of these live studies, three found no differences between these groups, and two found differences demonstrating more healthy outcomes for former smokers. Pederson and Lefcoe ( 1976) compared 46 former smokers, mostly self-quitters who had not smoked cigarettes for 1 year or longer, with 46 current smokers volunteering for treatment. These researchers found no differences on Jackson Personality Inventory scales that included measures of anxiety and self-esteem. Likewise, Pomerleau, Adkins. and Pertschuk (1978) used the Symptom Checklist (SCL-56) as a 2-year followup measure of dysphoria among 60 smoking cessation treatment participants and found no differences between quitters and continued smokers. Mean duration of smoking abstinence was not reported. Giannetti, Reynolds. and Rihn ( 1985) compared 47 former smokers who had been abstinent for at least 6 months with 35 current smokers hospitalized for cardiovascular disease and found no differences in "habits of nervous tension." In the only study to employ multiple self-report, physiologic, and observer measures, Abrams and colleagues (1987) found no significant differences between 22 former smokers (mean abstinence approximately 2 years) and 22 relapsers on the State-Trait Anxiety Inventory, but did find that former smokers reported significantly less anxiety and had significantly lower heart rates in response to simulated smoking-related stressors. In a study of worksite health screen participants, Orleans and colleagues (1983) compared 525 long-term former smokers who had been abstinent for more than I2 months (mean abstinence = approximately 9 years) with 856 current smokers and found that the long-term former smokers had significantly better age- and sex-adjusted scores on the Health and Nutrition Examination Survey (HANES) General Well-Being Index, including its anxiety and depression subscales, and on the Framingham measures of anger symptoms and anger internalization. However, there were no'differences on 534 these measures between current smokers and recent ex-smokers, those who had been abstinent for less than 12 months. Prospective longitudinal studies of smokers who become former smokers or remain continuing smokers are needed to establish whether any differences between former and current smokers existed prior to quitting, especially since baseline or "prequitting" measures of psychological well-being and self-esteem have been found to predict success in quitting smoking (Hall et al. 1983; Ockene et al. 1982: Schwartz and Dubitzky 1968; Straits 1970; West et al. 1977). The few prospective studies (Table 2) that have been conducted have either documented no significant change in psychologi- cal factors from baseline among former smokers. or no difference in the magnitude of change for former and continuing smokers. or have indicated improvements for former smokers. None of these studies demonstrated long-term negative psychological chan- ges for former smokers. Two of the prospective studies found no significant changes in a variety of mood and psychological measures from a prequitting baseline to long-term followup among former smokers and no significant differences between quitters and continuing smokers in the magnitude of such change. Pertschuk and coworkers ( 1979) asked 24 participants in a nonaversive cognitive-behavioral treatment to complete pretreatment and 2month followup ratings of psychological functioning. These researchers found no significant changes in stress, affect, symptoms of psychological distress, or utilization of psychiatric treatment as indicated by need for psychotropic medication or mental health services, Changes from baseline to followup were not evaluated separately for quitters and nonquitters. but these groups did not differ on 4-month followup ratings. Emmons and associates ( 1986) studied the effects of smoking cessation on cardiovascular reactivity to stress among quit-smoking clinic participants and found no significant changes from baseline to a 6-month followup among 16 abstainers or 8 relapsers. However, this study noted that an average weight gain of 5 pounds among abstainers may have masked improvements in reactivity scores. Because weight was related to baseline and followup cardiovascular measures, it is possible that in each of these studies, treatment assisted quitters in avoiding persistent unwanted side effects. Two studies of nicotine withdrawal effects that extended measurement beyond 4 weeks of abstinence have yielded no evidence for a withdrawal syndrome beyond 4 to 5 weeks (Hughes, Gust, Pechacek 1987; Gross and Stitzer 1989). These studies, reviewed in detail by Hughes. Higgins, and Hatsukami (1990). found that adverse postquitting changes in levels of anxiety. restlessness. impatience, irritability, and dysphoria peaked during the first 2 weeks after quitting. returned to baseline or below-baseline levels by 4 weeks. and remained at those levels at IO- to 36-week followups. Gross and Stitzer ( 1989) studied 40 smokers who quit after a j-session cessation class and maintained biochemically validated smoking abstinence for 10 weeks while using nicotine polacrilex gum or a placebo. Subjects completed weekly ratings of withdrawal symptoms, including symptoms of psychological distress such as irritability. anxiety. and impatience. Weekly followup ratings were adjusted for baseline ratings and baseline smoking rate. For the 20 placebo subjects. mean ratings for irritability. anxiety . and impatience increased from baseline to the first postquit weeh. returned to baseline TABLE 2.-Prospective studies of quitting-related changes in mood, anxiety, stress reactivity, perceived stress, self-image, and psychological well-being Reference Sample 5ile Type of study Findings Strengths or limitation\ Pertachuk CI al. (IY7Y) 24 smokmg cessation clinic participants Emmons et al. ( IYXh) Groaa and Stiller (IYXY) 24 smoking cessation clinic participant\ 40 abstamer5 using nicotine polacrilex gum or n placebo following a 3-sesston treatment Stress. affect. psychological distress. and utilization of psychiatric treatment were assessed at the start of treatment and 2 mo posttreatment Cardiovascular reactivity No significant pre- to poattreattnent (SBP. DBP. HR) in response to change for abstainers (N=16) in cognitive and physical stressors mean SBP, DBP. or HR. and no were assessed I wk prior to difference in amount of change treatment and 6 mo after between abstainers and recidivist\ treatment (N=U) A IS-item withdrawal symptom measure was completed weekly for 10 postquit weeks For placebo subjects. rated symptoms of psychological distrtx (irritability, anxiety. impatience) increased from baselme to first postquit week. returned to baseline by week 4. then declined below baseline initially, stabilizing after 5 wk; scores for active gum u\erh declined below baseline initially. stabilizing after 3 wk at below-baseline levels No significant pre- to posttreatment change in self-reported anxiety. depression, anger, irritabiltty, appetite loss, msomnta, hopelessness. dtfficulty concentrating, apathy, use of psychotropic medication Although pohttreatment scores did not dtfferentlnte abstainer\ (N=l6) and recidivists (N=X). thebe groups were not compared on pre- to posttreatment changes Only abstainen had a significant weight increase during the following period: thih may account for lack of reductton in cardiovascular reactivity Self-reported abstmencr biologically confirmed and baseline score\ and baseline smokmp rate uxd as covariates, but no control for repeated measurement TABLE 2.--Continued Sample we Typr of study Findings Strength\ or limitation\ I'dlacKh (IW7) after :I contact trc:ttmcnt with phyhictan xivicc and active nicotine polacrilex or placebo gum subject\ mted 5 withdrawal \ymptomh relevant to mood and psychological functioning (anger. anxiety, difficulty concentrating, impatience. restlehsnrsh) Among abstinent suhtects. these ratings peaked at l-7 wk postquttting, returned to bawlme by I mo. and declined further to below-baseline at 6 mo Below-baseline 6-mo ratings among nonquitters suggest a drift in mensures due to il repeated testing effect 35 participant5 m u whsation clinic for mokw with chronic cardiopulmonary diwaw 72 ex-\moher\ (N=7 mo ab\tinrnt) who had quit during the year follwinp a worhde health wreen (49 at comp;mie\ with health promotion programs, 73 at control companir\) POMS was adminiwred before md h mo after treiltmrnt HANES well-being, anxiety, and deprr\\ion scalc~ and the Framingham anger symptom 4c;ilc'~ wcrc ximinl5terrd 31 ;L hawlinc health scrrrn and I-4 r followup A meuwx of total mood di\turhancr (anger/irritability + tension + anxiety + fatigue + confusion + d~pr~\\ion/dejectiotl - vigor) at 6 mo w;t\ \ignificmtly negatively correlated with mohing reduction: parallel Ggnificant relations were noted for the ~caics anger/irritability and ten\ion/vnxiety Significant hawline to I-yr improvements in the HANES well-being and deprewon wale\ werr observed for new cx-wwher\ at treatment \ttes only: no changes tn Framingham anger mwwrc~ were ohwrved Analy\e\ controlled for pretrcatmcnt mwhurc~ Analyws controlled for uyz. wx. burlme UIUCI. and tlurntion of nhatinrnce: compxiwn\ with never moher\. long-term fommcr \mohtw. or reciclivlst\ at twittment \ite\ were not cwdKtrd levels by week four, then continued to decline, stabilizing at below-baseline levels by week six. There were significant interactions between use of the gum and the weeks during which it was used for each of these symptoms, with nicotine polacrilex gum significantly suppressing postcessation ratings only during the first 4 to 5 weeks after quitting. The authors concluded that several of the most disturbing aspects of the tobacco withdrawal syndrome appear to resolve within 4 to 5 weeks after quitting (Gross and Stitzer 1989). Although findings suggest positive changes over baseline for these recent quitters, below-baseline 6- to IO-week scores may reflect the effects of the initial treatment or a repeated-testing effect. In a similar study of the effects of nicotine polacrilex gum on tobacco withdrawal. Hughes, Gust, and Pechacek (1987) studied 3 15 smokers for 6 months after a minimal contact treatment involving brief physician counseling, instruction in nicotine polacrilex gum use, and prescription of nicotine polacrilex gum or a placebo. At a pretreatment baseline, and again at I- to 2-week. l-month, and 6-month followups, subjects rated six withdrawal symptoms related to mood and psychological functioning including anger, anxiety, difficulty concentrating, impatience, and restlessness in addition to four others--craving. hunger, insomnia, and physical symptoms. For 75 subjects abstinent at 6 months, of whom 57 used nicotine polacrilex gum and I8 used a placebo. ratings for anger, anxiety, difficulty concentrating. restlessness, and im- patience peaked at the I- to 2-week followup, returned to baseline at 1 month, then dipped to below-baseline levels at 6 months. Subjects receiving nicotine polacrilex gum compared with those using placebo reported smaller increases from baseline to I- to 2-week and I -month ratings for most withdrawal symptoms, but nicotine polacrilex gum effects were not explored at the 6-month followup because too few subjects continued using the gum. However, 6-month ratings were lower on many symptoms even among 240 nonquitters, suggesting a drift in ratings due to a testing effect. In fact, the only symptom change from baseline, which differentiated quitters and nonquitters at 6 months, was that quitters had a greater increase in hunger than did nonquitters (p I2 months) scored significantly higher on personal control than current smokers (~~0.01). A followup conducted 1 year later showed a significant (~~0.0 I ) increase toward internal control among 72 smokers who had quit since baseline (mean abstinence, 7 months). 542 Conversely, Orleans and colleagues (1983) found a significant shift toward more external health locus of control of similar magnitude among 30 individuals who had been former smokers at baseline. but who had relapsed by the l-year followup. A similar pattern was reported by Horwitz, Hindi-Alexander. and Wagner (1985) who followed 2 19 participants in a single-session hypnosis treatment over a I -year period. These researchers found a significant shift (p. There is some support to suggest that treatment method may have a differential effect on an increase in internal locur, of control orientation. Coping and Self-Management Skills The relation of abstinence from cigarettes to a generalized improvement in the extent and use of coping and self-management &ills has not been studied. To the extent that stopping smoking results in an individual'\ acquirin g or strengthening general]! ap- plicable stress-coping and temptation-copin g &ill\. long-term benefit4 of ab\tinnics might be expected to include the gencrali/ed use of such skills. Ho\se\er. no \tudic\ have assessed whether increases in feneralilrd \tre+copin, ~7 \hill\ occur ;I\ ;\ cons- quence of cessation. Longitudinal studies have not included prequitting and postquit- ting measures of generic copin p strategies. A brief review of the relation of coping to smoking cessation and maintenance of abstinence may help to provide direction for this line of needed research. Shiffman and Wills (1985) have developed a conceptual framework of coping that distinguishes stress-coping skills, that is. skills used to cope wnith general life stressors. and temptation-coping skills, or skills relevant forcoping with a situation in which there is a specific temptation for substance use or an urge to smoke. Folkman and Lazarus ( 1988) defined stress-coping as constantly changing cognitive and behavioral efforts to manage specific external and internal demands that are appraised as taxing or exceeding the resources of the person to maintain an appropriate balance between environmental demands and resources available to the individual to meet those demands. Temptation coping can be separated into what smokers do when faced with the immediate tempta- tion to smoke and anticipatory coping or the strategies smokers use to maintain commitment to abstinence and prevent temptation (Shiffman and Wills 1985). To the extent that smoking constitutes a maladaptive response for coping with stress and negative affects such as anxiety, depression. anger, frustration, loneliness. or boredom (Abrams et al. 1987: Marlatt 198Sb.c: Ockene et al. 198 I ), the former smoker must find alternative strategies for coping. The use of healthy all-purpose coping strategies such as self-reinforcement. assertive behavior, social support, relaxation, and exercise has proven important to success in maintaining abstinence in some studies (Ashenberg, Morgan. Fisher 1984: Grunberg and Bowen 1985; Marlatt 1985~: Shif- fman 1982). However, two large worksite studies demonstrated no differences between current and former smokers in the self-reported use of healthy and unhealthy techniques for coping with stress (Blair et al. 1980; Orleans et al. 1983). In support of the importance of coping skills. Katz and Singh ( 1986) found that 77 former smokers who had abstained for 6 months or more (mean 6.7 years) had significantly higher scores on the Rosenbaum Self-Control Schedule (a self-report measure of individual differences in applying \elf-control or coping methods) than 52 smokers recruited for a quit-smoking treatment. "Self-cured" and treated former smokers did not differ on this measure. The inves- tigators concluded that former smokers may have succeeded because they possessed better self-coping skills initially. The same interpretation could be applied to the study by Abrams and associates (1987) in which 22 former smokers (mean abstinence 22 months) exhibited better observer-rated skills to resist the temptation to smoke than did 22 recidivists in simulations involving interpersonal smoking triggers. Shiffman ( 1982) found that former smokers w,ho reported using cognitive and behavioral strategies to cope with smoking temptations vvere less likely to relapse. These few studies support the conclusion that use of skills to cope with stress and with temptations or urges to smoke seem to be more prevalent among former smokers compared with current smokers. 544 Social Support and Interpersonal Interactions Research has not addressed how smoking cessation influences the level of general or quitting-relevant social support available to the quitter or how cessation affects the quality of the individual's interpersonal interactions. Research on social support processes has focused on examining baseline or posttreatment measures of social support as predictors of quitting success (Graham and Gibson 1971: Lichtenstein. Glasgow, Abrams 1986; Mermelstein et al. 1986; Ockene et al. 1982: US DHHS 1989). Several studies have demonstrated that successful quitters had significantly fewer smokers in their social networks at baseline than did continuing smokers (Eisinger I97 I ; Graham and Gibson 197 I ; Ockene et al. 1982). Others have demonstrated that the quitter's success stimulated quitting by others. especially spouses (Suedfeld and Best 1977). A few studies are relevant to the investigation of cessation effects on social support. A large-scale. cross-sectional and longitudinal worksite study (Orleans et al. 1983) found no differences among current smokers, former smokers. and never smokers at baseline in satisfaction with personal relationships and interpersonal communication or in satisfaction with coworker relationships. However, at l-year followup. 72 baseline smokers who had quit (mean abstinence, 7 months) showed a significant decline from baseline in satisfaction with coworker relationships (p 3 meals/day on weekdays 3 meals/day on weekends Avoid\ wachs weekday\ Avoid\ wackh weehends Haa changed dret for health 220% above desirable weight Preventive care 6.1 1% h 36.3 30.5' h I.3 3.1dh 12.9 Il.h"h 1.7 4.X" h 30.3 25.x" h 48.6 14.3 24,s 21.0 35.0 21.9 Digital rectal exam (ever) Blood stool test (ever) Proctov2opic exam lever) WOMEN Alcohol conwmptlon Drink\ beer tS/wk Drinks 23 heers/epiwde Drinkc wine tS/wk Drink\ 23 glasse\ wme/epi\ode Drinhr liquor ZS/wk 23 drmks/epi\ode Dietary practices 3 meals/day on weehdays 3 meaNday on weekends Avoids bnacks weekdays *Avoids snack!, weekend\ Has changed diet for health 20% above desirable weight Preventive care Digital rectal exam (ever) Blood stool test (ever) Proctoscopic exam (ever) Pap smear (within year) Breast self-exam (withm yr) Breast exam (monthly) Mammogram lever) SY.5 3X.6 24.0 0.9 17.1 1.3 7.0 0.7 13.7 so. I 14.2 26.6 22.6 3x.7 `1.3 .56.X 67.4' h 37.9 36.2" h 20.x 27.2" h 39.7 13.5" h 34.x 30.3" h 51.5 57. 3X.5 16.7" h Former \moher\ Current smoker\ % % 17.1' 51. IL 1.7` `0.2' 1.1' 15.1' 32.X' 3S.l' `6.SL 33.6' 26.3' 33.X' 66.X" h 4-l.Yd h 27.7" h 2.3" h 17.zh 1.3" h IO.`)" h 2.7" 14.1h 4Y.Sh 4 I .x* h 76.9 `3.4" Cl 49.0" h ?4.Xh s9.4 33.`)' 2 I S)' 3.0' 32.7' I .9' 17.X' 2.7' x.0' X.5' 29.4' 26.X 23.6 34.5' 20.3' 60.6' 35.7' `I.1 40.7' 31.0 52.1 35.1' TABLE K-Summary of data from 1987 BRFSS, behaviors of former smokers and current smokers aged 18 and older .AdJu\ted odd\ ratio\ Behavior Former \mvhrr\ relatl\,e to never \moher\ Current \mohen relative to never \moher\ Former smoker\ relative to current \mokerr MEN .Alcohol conwmptwn Any alcoholimo 2.5 drinks/episode 260 drmk\/mo Drinkmg and drivmg I .75" 2.1 IJ 0.82" I .67" 2.63' 0.63' I .7S" 3.0?" 0.58" I .4-v' I .YY" 0.71' Weight/diet/exerci\e Obese (BMI)' Obese (Met. Llfe)d Trying to lose pounds More exercise Eating fewer kcal Physical actwity Sedentary I .os 0.63" I .6x" I .06 0.6-1" 1.63" I.??' 0.63" 1.92" O.YX 0.X3" l.l7h 0.x+ O.XZh I .w l.lOh 0.69" 1.57" 0.9 lb I .43" 0.64" Preventive care ~_- Cholesterol te\t Flu shot part month 1.77.' 0.04 1.34" I .OY o.X7h I .Ih" Other Use ST l.7JJ o.84h `.OYJ Use seatbelt O.YZh OX" I .60" WOMEN Alcohol consumption Any alcohollmo 25 drmks/epi\ode 260 drmk\/mo Drinhing and drivmg 2.07" I .X6' 2.xX" I .X7" 2.w 3.35" s .w' 1.92" 0.87" OX" 0.52 0.65" Weight/diet/exercr\e Obese (BMI)' Obese (Met. Life 1" Trymg to lose pounds More exercise Eating feuer kcal Physrcal activity Sedentary O.YX 0.96 1.19" I .07 0.97 1.17" 0 X6" 0.63" 0.65" 0.75" 0.72" 0.96 0.X1" 1.7-t.' 1.59" 1.52" I .60" I .4x" 0.99 I .JS'l 0.69" Preventive care Cholesterol test Flu shot past month I.15 0.95 I.1 I" O.Ylh 1 .os I .os 550