Tobacco withdrawal symptoms mediate motivation to reinstate smoking during abstinence

Abstract

Withdrawal-based theories of addiction hypothesize that motivation to reinstate drug use following acute abstinence is mediated by withdrawal symptoms. Experimental tests of this hypothesis in the tobacco literature are scant and may be subject to methodological limitations. This study utilized a robust within-subject laboratory experimental design to investigate the extent to which composite tobacco withdrawal symptomatology level and 3 unique withdrawal components (i.e., low positive affect, negative affect, and urge to smoke) mediated the effect of smoking abstinence on motivation to reinstate smoking. Smokers (≥10 cigarettes per day; N = 286) attended 2 counterbalanced sessions at which abstinence duration was differentially manipulated (1 hr vs. 17 hr). At both sessions, participants reported current withdrawal symptoms and subsequently completed a task in which they were monetarily rewarded proportional to the length of time they delayed initiating smoking, with shorter latency reflecting stronger motivation to reinstate smoking. Abstinence reduced latency to smoking initiation and positive affect and increased composite withdrawal symptom level, urge, and negative affect. Abstinence-induced reductions in latency to initiating smoking were mediated by each withdrawal component, with stronger effects operating through urge. Combined analyses suggested that urge, negative affect, and low positive affect operate through empirically unique mediational pathways. Secondary analyses suggested similar effects on smoking quantity, few differences among specific urge and affect subtypes, and that dependence amplifies some abstinence effects. This study provides the first experimental evidence that within-person variation in abstinence impacts motivation to reinstate drug use through withdrawal. Urge, negative affect, and low positive affect may reflect unique withdrawal-mediated mechanisms underlying tobacco addiction.

Figure 1

Individual mediational models. Separate mediational model sets were conducted for each withdrawal symptom mediator. a = path from the independent variable (i.e., abstinence status) to the mediator. b = path from the mediator (i.e., either positive affect, negative affect, or smoking urge) to the dependent variable (i.e., latency to smoking initiation—motivation to reinstate smoking) after controlling for the effect of the independent variable. a × b = the indirect effect of the independent variable on the dependent variable that occurs through the mediator, which equals the product of the “a” path and “b” path. c′ = the direct effect of the independent variable on the dependent variable that is not carried through the mediator, which equals the effect of the independent variable on the dependent variable after controlling for the mediator. The total effect = the sum of the indirect effect (a × b) and direct effect (c′), which equals the effect of the independent variable on the dependent variable.

Figure 2

Combined mediational models. A single mediational model set was conducted, which included all mediators simultaneously modeled. a = path from the independent variable (i.e., abstinence status) to the mediator. b = path from the mediators (i.e., the positive affect, negative affect, and smoking urge entered as simultaneous predictors) to the dependent variable (i.e., latency to smoking initiation—motivation to reinstate smoking) after controlling for the effect of the independent variable. a × b = the indirect effect of the independent variable on the dependent variable that occurs through a mediator controlling for the mediation occurring through the other two variables, which equals the product of the “a” path and “b” path for that specific mediator. c′ = the direct effect of the independent variable on the dependent variable that is not carried through the mediators, which equals the effect of the independent variable on the dependent variable after controlling for all three mediators. The total effect = the sum of the indirect effect (a × b) of the three mediators and the direct effect (c′), which equals the effect of the independent variable on the dependent variable.