واکنش پذیری استعمال دخانیات در سراسر آزمایشات انقراض انبوه: عاطفه منفی و اثرات جنسیت

Abstract Designing and implementing cue exposure procedures to treat nicotine dependence remains a challenge. This study tested the hypothesis that gender and negative affect (NA) influence changes in smoking urge over time using data from a pilot project testing the feasibility of massed extinction procedures. Forty-three smokers and ex-smokers completed the behavioral laboratory procedures. All participants were over 17 years old, smoked at least 10 cigarettes daily over the last year (or the year prior to quitting) and had expired CO below 10 ppm at the beginning of the ~ 4-hour session. After informed consent, participants completed 45 min of baseline assessments, and then completed a series of 12 identical, 5-minute exposure trials with inter-trial breaks. Smoking cues included visual, tactile, and olfactory cues with a lit cigarette, in addition to smoking-related motor behaviors without smoking. After each trial, participants reported urge and negative affect (NA). Logistic growth curve models supported the hypothesis that across trials, participants would demonstrate an initial linear increase followed by a decrease in smoking urge (quadratic effect). Data supported hypothesized gender, NA, and gender × NA effects. Significant linear increases in urge were observed among high and low NA males, but not among females in either NA subgroup. A differential quadratic effect showed a significant decrease in urge for the low NA subgroup, but a non-significant decrease in urge in the high NA group. This is the first study to demonstrate gender differences and the effects of NA on the extinction process using a smoking cue exposure paradigm. Results could guide future cue reactivity research and exposure interventions for nicotine dependence.

Introduction Cue exposure treatment is a behavioral intervention based on associative learning theories. These theories provide a framework for understanding smoking cessation and relapse and have demonstrated robust research support that illustrates how previously neutral, conditioned stimuli (CSs) elicit conditioned responses (CRs) such as drug urges and approach behaviors. With nicotine dependence, CSs can include external stimuli (e.g., smoking paraphernalia, other smokers) and internal stimuli (e.g., negative affect). Reactivity (CR) to drug-related CSs is a valid and reliable predictor of smoking cessation and relapse (Brandon et al., 1995 and Drummond et al., 1995). To date, researchers have successfully investigated a variety of smoking-related cues (CSs) that can elicit reliable and meaningful measures of smoking behavior. Behavioral laboratory studies of cue reactivity have used extroceptive, tobacco-specific cues such as exposure to a lit cigarette (LaRowe et al., 2007 and Sayette et al., 2001), interoceptive cues such as negative affect (Tiffany & Drobes, 1990), broader contexts presented through guided imagery or the use of smoking confederates (Drobes & Tiffany, 1997), and the experience of nicotine deprivation (Bailey et al., 2010 and Payne et al., 1996). The potency of CSs in motivating smoking can be inferred from animal studies of nicotine self-administration. For example, (Caggiula et al., 2001) found that cues previously paired with nicotine were more effective than nicotine itself at reinstating nicotine self-administration in rats. Cue exposure treatment aims to reduce smokers' reactivity to smoking cues by facilitating the extinction process. Extinction, or the reduction of CRs over time, occurs over repeated presentations of drug-related CSs in the absence of unconditioned stimuli (e.g., ingesting the drug and experiencing its effects). The process of extinction happens naturally during the course of a smoker's quit attempt: if the smoker continues to be exposed to smoking cues and does not smoke, s/he will experience fewer, less intense urges to smoke (CRs) over time. Studies of cue exposure interventions have demonstrated post-treatment decreases in drug cravings and consumption (Childress et al., 1986, Drummond and Glautier, 1994, McLellan et al., 1986, Monti et al., 1987, O'Brien et al., 1990, Rankin et al., 1983, Sitharthan et al., 1997 and Stasiewicz et al., 1997) even though long-term clinical outcomes remain modest. A previous meta-analysis of cue-exposure treatments for substance abuse found that cue-exposure for tobacco cessation was less successful in promoting cessation compared to cue exposure treatment for other drug dependencies (Conklin & Tiffany, 2002). We posit that factors known to relate to smoking cessation outcomes could influence variability in smoking cue extinction performance. Examining individual differences in cue reactivity over repeated exposure trials is warranted as it may help improve the clinical applications of extinction-based behavioral treatment methods. Two plausible factors that could affect cue extinction performance include gender and negative affect (NA). Despite well-established gender differences observed in tobacco cessation and relapse studies, we are not aware of any studies that have examined gender differences in response to smoking cue extinction procedures. A small body of research on gender differences in cue reactivity in nicotine and other drug research has shown conflicting results (e.g., Elman et al., 2001, Field and Duka, 2004, Fox et al., 2006, Franklin et al., 2004, Lynch et al., 2002, Monti et al., 1993, Nesic and Duka, 2006 and Sterling et al., 2004). Current research examining gender differences in reactivity could be informed by growing evidence that males and females respond differently to various reinforcing effects of tobacco smoking (see, Perkins, Donny, & Caggiula, 1999) even though the physiological effects of smoking are similar for males and females (e.g., increased heart rate, blood pressure, decreased skin temperature.) For example, males experience greater positive reinforcement from the pharmacological properties of nicotine than females, perhaps explaining why men have greater smoking cessation success with nicotine replacement therapy than women (Cepeda-Benito et al., 2004 and Wetter et al., 1999). Conversely, women experience greater reinforcement from the subjective effects of smoking, such as the social contexts in which they smoke and relief from NA (Borrelli et al., 1996, Eissenberg et al., 1999, File et al., 2001, Leventhal et al., 2007 and Xu et al., 2009), perhaps explaining in part why women respond more favorably than men to smoking cessation medications that relieve NA (Gonzales et al., 2002 and Scharf and Shiffman, 2004). Also, women may have a greater drug urge than men in response to experimentally induced negative moods (Rubonis et al., 1994). Niaura and colleagues (Niaura et al., 1998) showed that women had a greater reactivity to negatively valenced scripts, implicating potential mediating affects of NA on smoking cue reactivity among women. However, we are not aware of any studies that have examined the effects of gender and non-induced variability in NA on smoking urges using repeated cue exposure methods. Examining the potential differential influence of NA between males and females during the cue extinction process may help guide the development of improved behavioral treatments for smoking cessation. The objectives of the current investigation were to test the influence of gender and NA on change in reported smoking urge reactivity to smoking cues over time during an analog cue exposure procedure with massed extinction trials. We tested three hypotheses: 1) we hypothesized a gender effect on urge over time such that males would demonstrate greater urge reactivity to smoking cues than females over time; 2) we hypothesized that greater NA would influence greater urge reactivity over time; and 3) we hypothesized there would be a gender × NA interaction over time such that the difference between high and low NA groups would be more pronounced among females compared to males during initial increases in urge reactivity and later reduction in reactivity.

. Results Forty-three participants completed the massed extinction trial session. One eligible smoker was not included in the analyses because of voluntarily withdrawing from the study after Trial 4. The final sample included 51% male, 54% African American, and 51% unemployed participants with a mean age of 49 ± 11 years old. Participants smoked daily for an average of 20.56 ± 10.83 years, and had a mean, baseline negative affect score of 17.44 ± 5.96. There were no significant differences between males and females on smoking status, psychosocial, or demographic variables except that males, on average, were older than females (t = 2.56, P < .02). There was no change in withdrawal-related symptoms between the first and last smoking cue exposure trials, except for the energy score. Participants reported less energy (greater fatigue and drowsiness) at Trial 10 compared to Trial 1 (t = − 3.30, P < .01), however, energy and urge at Trial 10 did not correlate (r = .25, P = .12). 3.1. Cue reactivity test A t-test of urge difference scores (exposure trial urge–resting baseline trial urge) was calculated to compare urge reactivity between the neutral cue exposure trial and Trial 1 (the first smoking cue exposure trial.) Urge difference scores were significantly greater following the smoking cue exposure trial than the neutral cue trial (t = 3.63, P = .001). 3.2. Urge response across repeated, non-reinforced smoking cue exposure trials In Table 1, “fixed effects” represent the group or average beta coefficients, and variance components represent the variation (and covariation) of subject-specific beta coefficients around these fixed effects. In the preliminary model (not shown,) urge at baseline, time and time2 (time × time) were predictors and an unstructured variance–covariance matrix was postulated for variance components. Change in − 2 Log Likelihood (− 2LL) with an accompanying smaller AICC indicated that an “independence” structure was sufficient. In other words, the three “covariance components” could be dropped from the model. The resulting model (Model 1) had a significant linear (P = 0.01) and quadratic (P = 0.01) trend for time confirming the prediction that urge would first increase and then decrease across the cue extinction trials. Table 1. Multilevel level (hierarchical) logistic growth curve models. Fixed effects Estimate Std error DF t value P Type 3 effect Num DF Den DF F P Model 1: − 2LL = 279.34, AICC = 299.77 Intercept − 1.24 0.85 41 − 1.47 0.15 Urge baseline 1.62 0.40 386 4.07 <.0001 Urge baseline 1 386 16.56 <.0001 Time 0.65 0.25 42 2.59 0.01 Time 1 42 6.72 0.01 Time2 − 0.06 0.02 42 − 2.60 0.01 Time2 1 42 6.76 0.01 Variance components Wald 95% CI Intercept 13.49 7.27 5.87 56.75 Time 0.21 0.15 0.076 1.81 Time2 0.001 0.002 0.0002 2.02 Model 2: − 2LL = 278.34, AICC = 294.62 Intercept − 1.18 0.76 41 − 1.55 0.13 Urge baseline 1.49 0.34 427 4.35 <.0001 Urge baseline 1 427 19.47 <.0001 Time (male) 0.99 0.30 427 3.25 0.001 Time × gender 2 427 7.67 0.004 Time (female) 0.33 0.29 427 1.15 0.25 Time2 (male) − 0.07 0.03 427 − 2.19 0.03 Time2 × gender 2 427 4.61 0.03 Time2 (female) − 0.05 0.03 427 − 1.54 0.12 Variance components Wald 95% CI Intercept 11.05 5.38 5.13 38.72 Time2 0.002 0.001 0.001 0.011 Model 3: − 2LL = 279.36, AICC = 297.72 Intercept − 1.28 0.84 41 − 1.53 0.14 Urge baseline 1.59 0.40 384 4 <.0001 Urge baseline 1 384 16.01 <.0001 Time (low NA) 0.66 0.30 384 2.18 0.03 Time × NA 2 384 3.84 0.02 Time (high NA) 0.78 0.33 384 2.38 0.02 Time2 × NA 2 384 3.84 0.02 Time2 (low NA) − 0.07 0.03 384 − 2.30 0.02 Time2 (high NA) − 0.06 0.03 384 − 1.92 0.06 Variance components Wald 95% CI Intercept 13.13 7.13 5.68 56.04 Time 0.23 0.16 0.08 1.87 Time2 0.001 0.002 0.0002 9.74 Model 4: − 2LL = 272.58, AICC = 295.11 Intercept − 1.13 0.78 41 − 1.44 0.16 Urge baseline 1.52 0.39 383 3.98 <.0001 Urge baseline 1 383 15.82 <.0001 Time (male, low NA) 1.04 0.37 383 2.86 0.005 Time × gender × NA 4 383 3.16 0.01 Time (male high NA) 0.95 0.36 383 2.67 0.008 Time (female, low NA) 0.35 0.32 383 1.08 0.28 Time (female, high NA) 0.62 0.36 383 1.76 0.08 Time2(low NA) − 0.08 0.04 383 − 2.31 0.02 Time2 × NA 2 383 4.08 0.2 Time2 (high NA) − 0.06 0.03 383 − 1.9 0.06 Variance components Wald 95% CI Intercept 11.07 6.20 4.69 50.30 Time 0.20 0.16 0.07 2.12 Time2 0.001 0.001 0.0002 1.39 Table options 3.3. Gender effect on urge over time To assess whether the curvilinear trajectory was moderated by gender, Model 2 included two interaction terms: gender by linear time and gender by quadratic time. These interaction terms were significant (P = 0.004, P = 0.03, respectively) based on F-tests of “Type 3 Effects” (omnibus tests). 1 The t-tests for the group-specific slopes indicated that females had relatively flat trajectories (no significant linear or quadratic trends), whereas males had a significant linear increase (P = 0.001) and then significant decrease (P = 0.03) in urge over time (see Fig. 1). Goodness of fit indices did not improve with the addition of other demographic, smoking, or psychosocial characteristics. Urge to smoke across trials for males and females. Fig. 1. Urge to smoke across trials for males and females. Figure options 3.4. NA effect for urge over time To determine whether NA moderates the urge trajectories, Model 3 was created with NA entered in interaction with the linear and quadratic terms for time. There was a significant NA linear effect (P < .0001) as well as quadratic (P = 0.02) effect over time. The t-tests of the slopes demonstrated that Low NA had a significant effect on both linear and quadratic trends (P = 0.03, P = 0.02, respectively), whereas high NA had a significant effect on urge only on the linear trend (P = 0.02) (see Fig. 2). The addition of other demographic, psychosocial, or smoking characteristics did not improve goodness of fit indices. Urge to smoke over trials for high or low negative affect (NA). Fig. 2. Urge to smoke over trials for high or low negative affect (NA). Figure options 3.5. Gender × Negative affect × time interaction To assess whether NA could explain the differences in the trajectories for gender, initially Model 4 was created with two, 3-way interaction terms. The “type 3” F-tests revealed significant interaction between Gender and NA for linear time (P = 0.003) but not for quadratic time (P = 0.08). As a result, the final Model 4 consisted of a 3-way interaction for linear time (P = 0.01) and only a 2-way interaction for NA by quadratic time (P = 0.02). The t-tests for the slopes within each subgroup created by the interaction terms indicated a significant linear increase in urge for males in both the high and low NA subgroups, but not for females in either NA subgroup. Finally, as seen in Model 3, the low NA subgroup had a significant decrease in urge (P = 0.02), where the decrease was not significant (P = 0.06) in the High NA group (see Fig. 3). The addition of other demographic, psychosocial, or smoking characteristics did not improve model fit. Urge to smoke over trials for males and females with high or low NA. Fig. 3. Urge to smoke over trials for males and females with high or low NA.