Ntp055.indd

Nicotine & Tobacco Research, Volume 11, Number 6 (June 2009) 722–729
Original Investigation
Nicotine withdrawal symptoms following a
quit attempt: An ecological momentary
assessment study among adolescents
Rinka M. P. Van Zundert , Emiel A. Boogerd , Ad A. Vermulst , & Rutger C. M. E. Engels Abstract
Introduction
Introduction: The present study describes growth curves of Although research on smoking has focused mainly on initiation
withdrawal symptoms among 138 daily smoking adolescents and escalation of use, more recent research has been devoted to
before, during, and after a quit attempt.
the area of adolescent smoking cessation. Teenagers experience great diffi culties when attempting to refrain from smoking, and Methods :
Participants reported their levels of withdrawal approximately 90% – 95% of adolescents who make an unaided symptoms (craving, negative affect, and hunger) three times a day over a period of 28 days: 1 week prior to and 3 weeks follow- Sussman, 2002 ). Participation in adolescent cessation programs does not seem to improve success rates ( Garrison, Christakis, Ebel, Wiehe, & Rivara, 2003 ; Leatherdale, 2006 ). More insight Results : All withdrawal symptoms were quite stable at a into the natural history of adolescent quit attempts is needed to
relatively low level during the 5 days prior to the quit day. At
tailor psychosocial or pharmacological treatments to this group Day 8, withdrawal symptoms (especially craving) increased substantially. A significant decrease in symptoms was visible during the week following the quit day, and within 2 weeks Studies among adults have demonstrated that withdrawal postquit, both abstinent and relapsed adolescents had revert- symptoms can predict failure to quit smoking ( McCarthy, Piasecki, ed to levels comparable to those during the prequit period. Fiore, & Baker, 2006 ; Piasecki, Fiore, & Baker, 1998 ). Withdrawal The course over time for craving and hunger were best de- symptoms refer to a set of physical and mental discomforts that scribed by a quadratic term, and a linear model best suited emerge when individuals abstain from smoking, such as cigarette negative affect. Individual intercepts and slopes of the growth craving, irritability, restlessness, insomnia, anxiety, depression, in- curves were used to predict abstinence during the last week creased appetite, and poor concentration (see Hughes, 2007 , for a of the study and at the 2-month follow-up. Analyses revealed review). Withdrawal symptoms typically increase strongly during that higher levels of craving at the beginning of the prequit the fi rst week of deprivation, after which they revert gradually to a week and on the target quit day (intercepts) decreased the level that is the same as or lower than baseline levels ( Hughes, 1992 ; odds of being abstinent during the last week of the study. In Jorenby et al., 1996 ; Piasecki et al., 1998 ; for an exception, see addition, the quadratic term for hunger predicted abstinence Shiffman et al., 1997 ), and individual characteristics of the course during the last week. Finally, among all three symptoms, of withdrawal symptoms over time can predict cessation outcomes none of the growth model characteristics predicted absti- in adults ( McCarthy et al., 2006 ). Although we know that adoles- cents also experience withdrawal symptoms during smoking depri-vation and that they report similar symptoms ( Prokhorov, Hudmon, Discussion : The fi ndings generally suggest that smoking cessa-
Cinciripini, & Marani, 2005 ; Smith, Cavallo, McFetridge, Liss, & tion among daily smoking adolescents does not largely depend on how their withdrawal symptoms evolve over time after history of withdrawal in adolescents following a quit attempt and achieving abstinence. related the course of withdrawal over time to relapse outcomes. Rinka M. P. Van Zundert, M.Sc., Behavioural Science Institute, Corresponding Author:
Radboud University Nijmegen, The Netherlands Rinka M. P. Van Zundert, M.Sc., Behavioural Science Institute, Emiel A. Boogerd, M.Sc., Behavioural Science Institute, Radboud Radboud University, PO Box 9104, 6500 HE Nijmegen, The University Nijmegen, The Netherlands Netherlands. Telephone: +31-0031-24-3612816; Fax: +3100-31- Ad A. Vermulst, Ph.D., Behavioural Science Institute, Radboud 24-3612776; E-mail: [email protected] University Nijmegen, The Netherlands Rutger C. M. E. Engels, Ph.D., Behavioural Science Institute, Rad- boud University Nijmegen, The Netherlands doi: 10.1093/ntr/ntp055Advance Access publication on May 7, 2009Received July 17 , 2008 ; accepted January 28 , 2009 The Author 2009. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: [email protected] 722 Nicotine & Tobacco Research, Volume 11, Number 6 (June 2009)
For several reasons, we cannot assume that neither the use of daily measures of withdrawal has two major advantages course of withdrawal symptoms over time nor its association over retrospective ratings (especially those with large time inter- with cessation outcomes as observed among adults are by def- vals between the targeted experience and time of reporting). It inition identical for adolescents. First, adolescence is a devel- reduces the susceptibility of reports to recall bias and it provides opmental phase in which impulsivity, novelty seeking, and suffi cient data to be able to assess within-person developmental suboptimal decision making are considered to be normative traits ( Chambers, Taylor, & Potenza, 2003 ), and in which indi-viduals are still in the process of maturing and developing self- The present study describes the elevation and shape of the refl ective and introspective skills ( Beyth-Marom, Fischhoff, growth curves of the withdrawal symptoms craving, negative af-Jacobs-Quadrel, & Furby, 1991 ; Steinberg & Cauffman, 1996 ). fect, and hunger during the periods preceding and following a In addition, both hormonal processes and social, cognitive, target quit date in a large sample of 138 adolescent daily smok- and environmental infl uences that are uniquely associated ers. We also tested whether individual growth curve estimates with adolescence contribute to adolescents ’ moods being dif- differed as a function of postquit smoking and sex and whether ferent from those of young children and adults ( Buchanan, individual estimates predicted abstinence during the last week Eccles, & Becker, 1992 ). Adolescents may therefore experience of the monitoring period and at the 2-month follow-up. Age, more variable and more intense moods, more variable energy sex, and baseline nicotine dependence were entered into the levels, more restlessness, and more anxiety than individuals at Given these characteristics, it has been postulated that the similarities in withdrawal symptoms displayed by adolescents and adults might be attributed to the phase of adolescence itself
rather than to cessation ( Prokhorov et al., 2005 ). Although Participants
Prokhorov et al. (2005) suggest that this is true to some extent
A total of 272 Dutch daily smoking adolescents contacted the (since individual withdrawal symptoms did not differentiate ef- research team in response to advertisements and articles about fectively between never-smokers and former light smokers), the study that were published or displayed in newspapers, on symptoms for former smokers were reported retrospectively Web sites, and in community centers. Upon contacting the re-and duration of time since cessation was not taken into account. searchers, interested candidates were sent a detailed study de- However, the psychophysiological characteristics and states that scription. After reading the detailed information, 189 of the 272 typify adolescence might affect withdrawal in other ways. Nega- candidates still wanted to participate and were assessed for eligi- tive affect among adults, for example, seems to diminish over bility through a telephone screening. To qualify, candidates had time to levels lower than those reported prior to quitting among to be aged 15 – 19 years, smoke at least 1 cigarette/day, be highly successful quitters, and to persist at equal or even higher levels motivated to quit, and not be currently enrolled in a cessation among those who fail. Moreover, a fairly robust within-subject program. Participants aged 15 years had to have been smoking association was found between negative affect and smoking daily for at least 1 year to be eligible (this criterion did not apply lapses after a period of abstinence among adults ( Kassel, Stroud, for those aged 16 – 19 years). Four candidates were not accepted & Paronis, 2003 ). Since adolescence is characterized by increased because they had already quit smoking and nine candidates and more variable levels of negative affect, it is possible that ces- withdrew before entering the study, leaving 176 adolescents who sation does not instigate substantial increases in negative affect were enrolled offi cially. Of the 176 participants, 19 stopped or that, if it does, it might not necessarily provoke relapse among completing assessments before or right after the quit day and 16 adolescents as they may be more used to (and thus more toler- dropped out during the fi rst postquit week. For these patterns of ant of) mood changes. This hypothesis is supported in part by missing data, we concluded that they were not at random and preliminary fi ndings from Smith et al. (2008) , who assessed therefore should be excluded from our analyses. Three addition-withdrawal symptoms weekly for 4 weeks among adolescent al subjects were excluded because they achieved 24-hr absti-smokers participating in a pilot cessation intervention. They did nence for the fi rst time a week after the target quit day, which is not fi nd symptoms of depressed mood, irritability, diffi culty problematic for the present analyses since the actual quit day concentrating, appetite, sleep problems, and anxiety on the quit was set to the target quit day. The fi nal sample, thus, consisted of day to differ signifi cantly from the baseline assessments. Craving 138 subjects. In the logistic regression analysis, 12 subjects could and restlessness, however, increased signifi cantly after cessation. not be included because they failed to return the baseline ques- In addition, none of the symptoms predicted the likelihood of tionnaire and 4 subjects were excluded because their smoking lapse during the treatment, although this fi nding may have been during the last week could not be verifi ed. Half of the 38 subjects the result of the small sample size. Nonetheless, there seem to be who were excluded from the original sample of 176 individuals clear indications why the course of tobacco withdrawal symp- did not return the baseline questionnaire and, thus, were not toms over time and its association with cessation outcomes included in the following attrition analysis. t Tests indicated that might differ between adolescents and adults. those excluded were marginally more nicotine dependent, t (136) = 2.04, p = .053, and reported a higher average of number of ciga- In the present study, data were collected using ecological rettes smoked per day at baseline, t (136) = 2.04, p = .015, com- momentary assessment (EMA), which uses intensive repeated pared with those who were retained in the analyses. Excluded sampling to gather data on momentary states of individuals in subjects did not seem to differ from those included according to real-time and real-world contexts ( Shiffman, Stone, & Hufford, age, sex, and whether they quit in the past 12 months. 2008 ). Participants reported on their withdrawal symptoms and smoking behavior three times a day during 4 weeks: 1 week pri- The fi nal sample ( N = 138) consisted of 49 male (35.5%) or to the quit attempt and 3 weeks after the quit attempt. The and 89 female (64.5%) adolescents, who were receiving or had Nicotine withdrawal symptoms in adolescents
recently received regular education: 5.7% received lower voca- subscales include items that cover all symptoms of the smoking tional education, 39.7% received intermediate vocational edu- withdrawal syndrome as listed in Diagnostic and Statistical Man- cation, 13.5% received high school education, 12.8% received ual of Mental Disorders, Fourth Edition ( DSM-IV ), except for preuniversity education, and 16.3% were in college (12.1% un- insomnia. Although craving is not one of the DSM-IV criteria, known due to missing baseline questionnaire). Ages were dis- it is considered to be an essential component of the abstinence tributed as follows: 15 (2.2%), 16 (32.4%), 17 (30.1%), 18 withdrawal syndrome (e.g., Hughes, Higgins, & Bickel, 1994 ) (15.4%), 19 (18.4%), and two persons had just turned 20 (1.5%; and was included in the daily assessments. The WSWS has good M = 17.2 years, SD = 1.2). The average age at onset of daily construct validity with high reliabilities for the three symptoms smoking was 14.3 years ( SD = 1.5), and the average length ( West, Ussher, Evans, & Rashid, 2006 ), and the shortened version of time during which adolescents smoked daily was 2.9 years of 10 items has been applied successfully in prior EMA research ( SD = 1.6). Regarding quit attempts in the past 12 months, 14.5% on smoking relapse ( McCarthy et al., 2006 ). Two WSWS items had not tried to quit or cut down their smoking, 31.9% had tried were used to measure craving, six items to measure negative af- to quit smoking but not to cut down, 15.9% had tried to cut down fect, and two items to measure hunger at each daily assessment. their smoking but not to quit, and 29.0% had made attempts to The 10 items could be answered on a Likert scale ranging from cut down and quit smoking (8.7% unknown). Among those who 1 = strongly disagree to 5 = strongly agree . A high score on the reported having made a quit attempt, the average number of items, thus, represented high levels of withdrawal symptoms. prior quit attempts in the past 12 months was 1.9 ( SD = 1.2). Prior to performing the analyses, we examined the factorial Most participants smoked between 11 and 20 cigarettes/day. validity and internal validity of the three subscales of the WSWS. For each assessment (28 days × 3 times a day = 84), we con- Procedure and design
ducted a confi rmatory factor analysis. The WSWS had a fi rm factorial validity with high principal loadings (mean values var- All 176 participants from the original sample were sent a baseline ied between .74 and .92) and an adequate model fi t (mean fi t questionnaire, which was generally completed 1 week prior to the values were c 2 (17) = 34.54, p = .000, comparative fi t index = onset of the monitoring period. If needed, participants also re- .978, root mean squared error of approximation = .103). To de- ceived a letter for their schools that requested the school board’s termine the internal consistency of the withdrawal scales across cooperation. For each individual, the period of monitoring always all assessments, we computed alphas for each assessment started on the fi rst Monday following the telephone screening. and averaged these. This resulted in an average alpha of .88 Participants were monitored for 4 weeks: 1 week prior to the tar- ( SD = .04) for craving, .75 ( SD = .04) for negative affect, and .66 get quit day, during which time they were instructed to smoke ( SD = .09) for hunger. All three factors were intercorrelated: ad libitum , and 3 weeks after the target quit day. For all partici- craving and negative affect: r = .54; craving and hunger: r = .42, pants, the assigned target quit day was the eighth day of the study. and negative affect and hunger: r = .31. During the monitoring period, participants were asked to re- spond to an Internet-based survey (on any desktop or laptop com- Nicotine dependence
. Nicotine dependence at baseline
puter available) three times per day during the following intervals: was assessed using a multidimensional measure of nicotine in the morning (to be completed between 10 a.m. and noon), in the dependence for adolescents, which has good psychometric afternoon (3 – 5 p.m.), and in the evening (8 – 10 p.m.). Participants properties ( Kleinjan et al., 2007 ). This composition was de- were not required to specify their location when they completed the rived from the modifi ed Fagerström Tolerance Questionnaire assessments. The survey always contained the same questions and ( Fagerström & Schneider, 1989 ) and the Hooked on Nicotine took about 3 min to complete. Printed paper diaries with identical Checklist ( DiFranza et al., 2000 ). The combined 11 items of questions were provided in case participants were unable to access the three subscales included aspects of emotional and physical the Internet; these were to be submitted online as soon as partici- symptoms of dependence (irritation, anger, restlessness, and pants had access to the Internet again. Participants who failed to the like when abstaining or smoking less) and behavioral complete a questionnaire within the designated sampling window symptoms of nicotine dependence (e.g., intensity of smok- were sent a text message to remind them. Participants were not re- ing). The scale was composed with the standardized values quired to initiate reports of smoking or withdrawal symptoms out- (range = 1 – 4) since response categories were not the same for side the fi xed intervals (event-contingent recording). Participants each item. Cronbach’s alpha was .79. The average level of base- received the fi rst instructions on the study during the telephone line nicotine dependence was 2.6 ( SD = 0.48). screening interview and instructions on the use of the Internet-based survey through E-mail. Participants were phoned by the re-search team twice during the course of the study to ask how they Outcome variables
were doing and to ensure compliance. Two months after the end of For the present analyses, we examined abstinence during the last the monitoring period, participants completed an online follow-up week of the study and at the 2-month follow-up. A dummy coded survey. Participants received remuneration of 40 Euros for comple- variable was constructed with 0 indicating that smoking occurred tion of the baseline questionnaire and the 4 weeks of monitoring during the past 7 days of the monitoring period (even if only one and an additional 10 Euros upon completion of the follow-up. cigarette) and 1 representing abstinence. Similarly, at follow-up, participants were asked whether they were currently abstinent, with response choices 1 = “ I have reverted to smoking as much as Measures
when I started the study, ” 2 = “ I have cut back on smoking (com- Withdrawal symptoms . We used 10 items from the Wisconsin
pared to when I started the study), ” and 3 = “ I have quit smoking Smoking Withdrawal Scale (WSWS) to determine adolescents ’ entirely. ” Response choices 1 and 2 were grouped and given the daily levels of the withdrawal symptoms craving, negative affect, score 0, which indicated that they had reverted to smoking, and 3 and hunger ( Welsch et al., 1999 ). The negative affect and hunger was recoded as 1, representing abstinence. Nicotine & Tobacco Research, Volume 11, Number 6 (June 2009)
Data analyses
Latent growth curve analysis
We used piecewise linear growth curve modeling (LGCM) in We fi rst examined whether a linear or quadratic trend best fi t Mplus ( Muthén & Muthén, 1998 – 2006 ) to specify growth mod- the data. All prequit intercepts deviated signifi cantly from zero els of craving, negative affect, and hunger, containing an inter- ( Table 1 ), but the prequit slopes were not signifi cant, as can be cept as well as linear or quadratic terms during both prequit seen in Figure 1 , which depicts the best fi tting growth curves. (Days 3 – 7) and postquit (Days 8 – 21) periods. The fi rst two days of the week prior to the quit attempt (Days 1 and 2) were omit- For postquit craving and hunger, a quadratic model de- ted from the growth curves because the withdrawal scores on scribed the data best, with intercepts that deviated signifi cantly those days were much higher relative to the other prequit days. from zero and with signifi cant negative slopes and positive qua- This is a common issue in EMA studies since participants need dratic terms. For the course of postquit negative affect, a linear to complete the questions a number of times before their re- function was most suitable, with a signifi cant intercept and sig- sponses are valid. Since the last week (Days 22 – 28) was used to nifi cantly declining slope. The model fi t indices evidenced good determine one of the study outcomes (smoking during last to excellent fi t of the models (see Table 1 ). week), the growth curves did not include this week. For those who did not achieve 24-hr abstinence on the target quit day, the Translated to more descriptive terms, we can say that prequit actual quit day was set to Day 8. The three daily assessments withdrawal levels were highly stable and that, on Day 8, with- drawal symptoms (especially craving) increased substantially. A strong decrease in symptoms was visible during the week fol- Because the distributions of the symptoms were some- lowing the quit day. During the 2 weeks after the target quit day, what skewed and leptokurtic and assumed random missing negative affect and hunger decreased monotonically to a level values, we used the robust full information maximum like- comparable with the beginning of Day 8, and craving showed a lihood estimator. In addition, a large proportion of the steeper decrease. As can be seen from Figure 1 , the three symp-sample reported smoking after achieving abstinence (68.8% toms followed the same overall pattern, although craving was in the first week after the target quit day, and 40.6% in the most salient in its elevation on the quit day and its curvature. second week after the target quit day), which is likely to af-fect the growth curves. If we would compose different We also tested whether the growth curve parameters differed groups on the basis of their smoking after achieving 24-hr as a function of post-abstinence smoking and sex (only the sig- abstinence, the groups would be too small to perform nifi cant results are reported). Prequit intercepts of craving were LGCM, and statistical power would be jeopardized. To ob- signifi cantly higher among those who smoked during the third tain some indication of how growth curves might differ as a study week compared with those who were abstinent that entire function of post-abstinence smoking, we performed t tests week, t (120) = − 2.23, p = .027. In addition, those who smoked to compare the growth parameters of the postquit growth during the third study week displayed marginally stronger nega- curves between the following two groups: (a) those who did tive slopes for hunger, t (120) = 1.91, p = .058, and stronger qua- not smoke on any day during the second study week (which dratic estimates for hunger, t (120) = − 1.98, p = .050. As for was the first week after cessation) and (b) those who smoked differences between sexes, girls had on average a marginally high- on at least 1 day during that week. We performed similar er postquit intercept of negative affect, t (120) = − 1.86, p = .066. t tests for groups that were distinguished on the basis of their smoking in the third week of the study. Differences Predicting abstinence
between boys and girls in prequit and postquit estimated Among the subsample of 122 subjects, 51.6% were verifi ed to be abstinent during the last study week. At follow-up, 32.5% re-ported current abstinence. Age, sex, and nicotine dependence Next, logistic regression analyses were conducted to test were not signifi cantly related to abstinence during the last week whether individual estimates of symptom trajectories (prequit or with abstinence at follow-up ( Table 2 ). and postquit intercepts, slopes, and quadratic terms) predicted abstinence during the last week of the monitoring period and Of all individual growth curve parameters, only the prequit and abstinence at follow-up. Since slopes and quadratic terms auto- postquit intercepts of craving and the postquit slope and quadratic matically have very low SD s, odds ratios ( OR s) obtained in the term of hunger predicted abstinence during the last study week. logistic regression analyses are likely to be excessively large. We Higher intercepts of prequit craving, or a higher general craving avoided this problem by using standardized values of the growth level across all prequit days, decreased the odds for abstinence dur- curve estimates in the logistic regression analyses. Finally, age, ing the last week. For hunger, a lower slope and higher quadratic sex, and nicotine dependence were included as covariates for term predicted failure of abstinence during the last week. This means that those who reverted to their prequit levels of hunger fast-est were less likely to be abstinent later on. We found no effects of growth curves estimates on abstinence at the 2-month follow-up. Although participants were instructed to smoke ad libitum dur- Discussion
ing the fi rst week of monitoring, eight participants (5.8%) quit smoking before the target quit day. The majority of participants The main objective of the present study was to describe within- reached 24-hr abstinence on the target quit day (73.2%) and person variability in withdrawal symptoms and its association with 14.5% on the next day. The remaining participants (4.3%) smoking cessation in a sample of daily smoking adolescents who reached 24-hr abstinence between Days 10 and 12. embarked on a serious quit attempt. All withdrawal symptoms Nicotine withdrawal symptoms in adolescents
Table 1. Intercepts, slopes, and model fi t indices of craving, negative affect, and hunger
( N = 138)
(craving, negative affect, and hunger) increased on the designated likely to be abstinent during the last study week. Finally, among all quit day. The course over time for craving and hunger were best three symptoms, none of the growth model characteristics pre- described by a quadratic term, and a linear model best suited nega- dicted abstinence at the 2-month follow-up. tive affect. Within 2 weeks postquit, both abstinent and relapsed adolescents had reverted to levels comparable with those during The elevation of all three withdrawal symptoms on the des- the prequit period. Higher levels of craving during the prequit ignated quit day is in line with fi ndings among adults ( Hughes, week and on the target quit day (intercepts) decreased the odds of 1992 ; Jorenby et al., 1996 ; McCarthy et al., 2006 ; Piasecki et al., being abstinent during the last week of the study. The prequit and 1998 ; for an exception, see Shiffman et al., 1997 ) and with pre- postquit slopes of craving did not predict abstinence during liminary results among adolescents ( Smith et al., 2008 ). Craving the last week. Growth parameters of negative affect were not asso- appeared to be the most salient symptom in its elevation and ciated with chances of being abstinent. For hunger, it appeared that curvature, which is in accordance with the study on adolescents those who reverted to their prequit levels of hunger fastest were less by Smith et al. (2008) . It is also in line with the consistent report by adolescents of craving being the most salient and severe symptom in general ( Colby, Tiffany, Shiffman, & Niaura, 2000 ). A comparison of the craving growth curve from the present study with the one among adults as reported by McCarthy et al. (2006) , who used an identical craving scale, shows that both adult and adolescent craving levels remained quite stable during the prequit period but that adolescents seemed to revert to their baseline craving levels more quickly. It seems plausible that those who reverted to smoking after achieving abstinence expe-rienced relief of craving, which may account for the relatively quick overall decline. However, given that the postquit slopes did not differ between those who reported postquit smoking and those who did, it seems that adolescents who successfully quit were not bothered by elevated craving for long. This may explain why the rate of decline did not predict abstinence. It is interesting that both prequit and postquit intercepts of craving had a signifi cant effect on abstinence during the last study week, whereas the shape of the course over time did not. Thus, Figure 1.
Estimated growth curves of craving, hunger, and negative for craving, this seems to suggest that how adolescents enter the quitting process is more important than the process itself. This Nicotine & Tobacco Research, Volume 11, Number 6 (June 2009)
Table 2. Cessation outcomes as predicted by individual characteristics and individual
estimates of the growth curve analyses ( N = 122)
Smoking during last week ( N = 122) Smoking status at follow-up ( N = 126) Note. The estimates for the effects of intercepts, slopes, and quadratic terms are from the multivariate analyses in which age, sex, and nicotine dependence were included as covariates. Prequit and postquit predictors were not included in the analyses simultaneously. explanation is contradicted somewhat by our fi ndings that base- Whereas adults ’ levels of negative affect remained stable across the line nicotine dependence did not predict abstinence. It is possible 3 weeks after cessation (the slope coeffi cient was positive but not that an elevation in symptoms had already taken place in the 1 signifi cant), the postquit slope among adolescents was signifi cant week between completion of the baseline questionnaire and the and negative. Thus, as with craving, adolescents seem to revert to start of the monitoring period. Since the prequit period was very their baseline levels of negative affect more quickly than adults. Ap- short (7 days, of which only 5 were included in the growth parently, trying to quit does not instigate intense negative affect curves), the present effects of craving intercepts might refl ect an- among adolescents, and the elevation and pace of the subsequent ticipatory mechanisms (cf., McCarthy et al., 2006 ). Taking this decrease in symptoms do not seem to provoke relapse either. This one step further, it may be less crucial to target craving once 24-hr may be the result of adolescents in general being subject to more abstinence is achieved but is rather essential to decrease craving variable and intense moods and more anxiety than young children levels before the attempt is started. Although nicotine replace- and adults ( Buchanan et al., 1992 ). Alternatively, although moods ment therapy (NRT) has been found to reduce levels of craving may be more variable and intense during adolescence, those predis- among adults ( Hughes, Shiffman, Callas, & Zhang, 2003 ), prior posed to depressive feelings may be more sensitive to mood chang- studies have found little support for the effi cacy of NRT among es, and changes in negative affect may be linked more closely to adolescents ( Hanson, Allen, Jensen, & Hatsukami, 2003 ; Killen abstinence among adolescents vulnerable to depressive mood. et al., 2004 ; Moolchan et al., 2005 ). More research is needed to examine how NRT could be improved and to explore alternative Prequit levels of hunger were low and stable, and they resem- treatments to decrease prequit craving among adolescents. bled prequit levels of hunger among adults ( McCarthy et al., 2006 ). Although adults and adolescents seem to experience a similar The target quit day peak in negative affect was less pronounced modest increase in hunger during the quit day, the postquit course than that of craving, and the prequit and postquit intercepts and over time appears to be different. Whereas the postquit slope slopes of negative affect did not predict treatment outcomes. The among adults showed a marginally signifi cant linear increase over observation that levels of both prequit and postquit negative affect time, the trend among adolescents was a signifi cant decline and were relatively low and showed only a modest elevation is in line was quadratic. However, this difference in results might be ex- with prior fi ndings among adolescents ( Smith et al., 2008 ). How- plained by postquit smoking since those who had been smoking ever, the postquit part of the growth curve for negative affect did during the third study week displayed stronger quadratic terms show discrepancies from that among adults ( McCarthy et al., 2006 ). for hunger. In other words, those who had reverted to smoking Nicotine withdrawal symptoms in adolescents
experienced faster declines in feelings of hunger. The latter fi nding at the 2-month follow-up) compared with other studies among also provides an explanation of why the quadratic term of hunger adolescents ( O’Connell et al., 2004 ; Sussman, 2002 ). predicted abstinence during the last week since those who smoked during the third week were more likely to be smoking during the Alternatively, the high abstinence rates could have been a re- sult of attrition. Those who dropped out or who were excluded from the present analyses had signifi cantly higher levels of base- Previous fi ndings by Smith et al. (2008) indicated that the line daily smoking and were more dependent on nicotine. Al- course of withdrawal over time among adolescents who achieved though higher levels of nicotine dependence did not affect abstinence differed for boys and girls, but we found little evidence cessation outcomes, those who drop out of studies in which of such relationship. One exception was that girls had marginally smoking abstinence needs to be achieved often can be considered higher postquit intercepts of negative affect than boys, which unsuccessful quitters (e.g., Smith et al., 2008 ). Our ability to gen-is plausible considering that the literature consistently indicates eralize the present results to all daily smoking adolescents may be that females are more vulnerable to depressive symptoms restricted somewhat. These limitations notwithstanding, the ( Piccinelli & Wilkinson, 2000 ), and this also applies to Dutch present study has revealed new insights into the natural history adolescents ( Engels, Finkenauer, Meeus, & Dekovic, 2001 ). None- of the course of withdrawal symptoms over time and the associa- theless, girls did not seem to be at additional risk since the inter- tion of this course with cessation outcomes in adolescents. cepts of negative effect were not associated with cessation outcomes, and sex as an independent covariate did not predict abstinence. In sum, the fi ndings generally suggest that an adolescent’s abil- ity to quit smoking successfully does not depend a great deal on This research was supported by a grant from the Dutch Asthma how withdrawal symptoms evolve over time after achieving absti- Foundation and a fellowship grant to Rutger Engels from the Netherlands Organization of Scientifi c Research . nence. We should be cautious, however, in declaring withdrawal symptoms to be less important to the adolescent cessation process. First, the present fi ndings do not indicate whether day-to-day vari- Declaration of Interests
ations in withdrawal might predict lapse or relapse the next day (as has been demonstrated for craving among adults; Shiffman, Paty, Gwaltney, & Dang, 2004 ). Up to now, almost no research has been devoted to the dynamic effects of withdrawal symptoms among adolescent smokers who are in the midst of a quit attempt (except Acknowledgments
for an exploratory study among 13 adolescents by The research presented in the present paper was carried out at Bartolomei, Colby, & Kahler, 2008 ). Second, in interpreting the pres- the Behavioral Science Institute, Radboud University Nijmegen, ent results, we must recognize several study limitations. For instance, participants who were categorized as abstinent during the last week of the monitoring period may have been identifi ed as smokers by other assessments. The likelihood that this occurred is restricted References
since compliance was high and only 7.1% of postquit assessments (not days) were not completed, but it remains a possibility. Beyth-Marom , R. , Fischhoff , B. , Jacobs-Quadrel , M. , & Furby , L. ( 1991 ). Teaching decision making to adolescents: A critical Furthermore, not all participants managed to achieve 24-hr review . In J. Baron & R. V. Brown (eds.), Teaching decision mak- abstinence on the designated quit day. Even though each person’s ing to adolescents (pp. 19 – 61 ), Hillsdale, NJ : Lawrence Erlbaum actual quit day was set to Day 8, and the postquit growth models started with the actual quit day for each person, those who did not quit until after the target quit day may have differed in initial levels Buchanan , C. M. , Eccles , J. S. , & Becker , J. B. ( 1992 ). Are adoles- of withdrawal from the ones who succeeded. Those who quit later cents the victims of raging hormones: Evidence for activational continued to smoke between the target quit day and their actual effects of hormones on moods and behaviour at adolescence . quit day, which is likely to infl uence the course of withdrawal. Psychological Bulletin , 111 , 62 – 107 . We also must consider a possible reactivity effect associated Chambers , R. A. , Taylor , J. R. , & Potenza , M. N. ( 2003 ). Develop- with the intensive self-monitoring that is a core component of mental neurocircuitry of motivation in adolescence: A critical EMA research designs. Intensive self-monitoring has been ar- period of addiction vulnerability . American Journal of Psychiatry , gued to reduce the problem behavior that is being targeted to be changed ( Shiffman et al., 2008 ). This may occur particularly Colby , S. M. , Tiffany , S. T. , Shiffman , S. , & Niaura , R. S. ( 2000 ). when people want to change their behavior and are able to exert Are adolescent smokers dependent on nicotine? A review of the control over it, which is the case for smoking cessation. Although evidence . Drug and Alcohol Dependence , 59 , 83 – 95 . studies have found little support for the concern that EMA pro-duces signifi cant reactivity ( Shiffman et al., 2008 ), it remains DiFranza , J. R. , Rigotti , N. A. , McNeill , A. D. , Ockene , J. K. , Sav- possible that the intensive self-monitoring infl uenced the way ageau , J. K. , St Cyr , D. , et al. ( 2000 ). Initial symptoms of nico- adolescents dealt with withdrawal symptoms and that it con- tine dependence in adolescents . Tobacco Control , 9 , 313 – 319 . tributed to the rates of abstinence found in the present study. We must keep this possibility in mind since abstinence rates were Engels , R. C. M. E. , Finkenauer , C. , Meeus , W. , & Dekovic , remarkably high (51.6% during the last study week and 32.5% M. ( 2001 ). Parental attachment and adolescents’ emotional Nicotine & Tobacco Research, Volume 11, Number 6 (June 2009)
adjustment: The role of interpersonal tasks and social compe- Mermelstein , R. ( 2003 ). Teen smoking cessation . Tobacco Control , tence . Journal of Counseling Psychology , 48 , 428 – 439 . Fagerström , K. O. , & Schneider , N. G. ( 1989 ). Measuring nic- Moolchan , E. T. , Robinson , M. L. , Ernst , M. , Cadet , J. L. , Pickworth , otine dependence: A review of the Fagerström Tolerance W. B. , Heishman , S. J. , et al. ( 2005 ). Safety and effi cacy of the nico-Questionnaire . Journal of Behavioral Medicine , 12 , 159 – 182 . tine patch and gum for the treatment of adolescent tobacco addic-tion . Pediatrics , 115 , e407 – e414 . Garrison , M. M. , Christakis , D. A. , Ebel , B. E. , Wiehe , S. E. , & Rivara , F. P. ( 2003 ). Smoking cessation interventions for ado- Muthén , L. K. , & Muthén , B. O. ( 1998 – 2006 ). Mplus user’s guide lescents: A systematic review . American Journal of Preventive Medicine , 25 , 363 – 367 . O’Connell , M. L. , Freeman , M. , Jennings , G. , Chan , W. , Greci , L. Gwaltney , C. J. , Bartolomei , R. , Colby , S. M. , & Kahler , C. W. S. , Manta , I. D. , et al. ( 2004 ). Smoking cessation for high school ( 2008 ). Ecological Momentary assessment of adolescent smok- students impacts evaluation of a novel program ing cessation: A feasibility study . Nicotine & Tobacco Research , Modifi cation , 28 , 133 – 146 . Piasecki , T. M. , Fiore , M. C. , & Baker , T. B. ( 1998 ). Profi les in Hanson , K. , Allen , S. , Jensen , S. , & Hatsukami , D. ( 2003 ). Treat- discouragement: Two studies of variability in the time course of ment of adolescent smokers with the nicotine patch . Nicotine & smoking withdrawal symptoms . Journal of Abnormal Psychology , Tobacco Research , 5 , 515 – 526 . Hughes , J. R. ( 1992 ). Tobacco withdrawal in self-quitters . Jour- Piccinelli , M. , & Wilkinson , G. ( 2000 ). Gender differences in nal of Consulting and Clinical Psychology , 60 , 689 – 697 . depression: Critical review . British Journal of Psychiatry , 177 , 486 – 492 . Hughes , J. R. ( 2007 ). R. Measurement of the effects of abstinence Psychology of Addictive Prokhorov , A. V. , Hudmon , K. S. , Cinciripini , P. M. , & Marani , S. Behaviors , 21 , 127 – 37 . ( 2005 ). “ Withdrawal symptoms ” in adolescents: A comparison of former smokers and never-smoker Hughes , J. R. , Higgins , S. T. , & Bickel , W. K. ( 1994 ). Nicotine Research , 7 , 909 – 913 . withdrawal versus other drug withdrawal syndromes: Similari-ties and dissimilarities . Addiction , 89 , 1461 – 1470 . Shiffman , S. , Engberg , J. , Paty , J. A. , Perz , W. , Gnys , M. , Kassel , J. D. , et al. ( 1997 ). A day at a time: Predicting smoking lapse Hughes , J. R. , Shiffman , S. , Callas , P. , & Zhang , J. ( 2003 ). A meta- from daily urge . Journal of Abnormal Psychology , 106 , 104 – 116 . analysis of the effi cacy of over-the-counter nicotine replace- ment . Tobacco Control , 12 , 21 – 27 . Shiffman , S. , Paty , J. A. , Gwaltney , C. J. , & Dang , Q. ( 2004 ). Imme-diate antecedents of cigarette smoking: An analysis of unrestricted Jorenby , D. E. , Hatsukami , D. K. , Smith , S. S. , Fiore , M. C. , Allen , smoking patterns . Journal of Abnormal Psychology , 113 , 166 – 171 . S. , Jensen , J. , et al. ( 1996 ). Characterization of tobacco with-drawal symptoms: Transdermal nicotine reduces hunger and Shiffman , S. , Stone , A. A. , & Hufford , M. R. ( 2008 ). Ecological weight gain . Psychopharmacology , 128 , 130 – 138 . momentary assessment . Annual Review of Clinical Psychology , 4 , 1 – 32 . Kassel , J. D. , Stroud , L. R. , & Paronis , C. A. ( 2003 ). Smoking, stress, and negative affect: Correlation, causation, and context Smith , A. E. , Cavallo , D. A. , McFetridge , A. , Liss , T. , & Krishnan- across stages of smoking . Psychological Bulletin , 129 , 270 – 304 . Sarin , S. ( 2008 ). Preliminary examination of tobacco withdraw-al in adolescent smokers during smoking cessation treatment . Killen , J. D. , Robinson , T. N. , Ammerman , S. , Hayward , C. , Nicotine & Tobacco Research , 10 , 153 – 159 . Rogers , J. , Stone , C. , et al. ( 2004 ). Randomized clinical trial of the effi cacy of bupropion combined with nicotine patch in the Steinberg , L. , & Cauffman , E. ( 1996 ). Maturity of judgment in treatment of adolescent smokers . Journal of Consulting and adolescence: Psychosocial factors in adolescent decision making . Clinical Psychology , 72 , 729 – 735 . Law and Human Behavior , 20 , 249 – 272 . Kleinjan , M. , Van Den Eijnden , R. J. J. M. , Van Leeuwe , J. , Brug , Sussman , S. ( 2002 ). Effects of sixty six adolescent tobacco use J. , Otten , R. , & Engels , R. C. M. E. ( 2007 ). Factorial and conver- cessation trials and seventeen prospective studies of self-initiated gent validity of nicotine dependence measures in adolescents: quitting . Tobacco Induced Diseases , 1 , 35 – 81 . Comparing one- and multidimensional approaches . Nicotine & Tobacco Research , 9 , 1 – 10 . Welsch , S. K. , Smith , S. S. , Wetter , D. W. , Jorenby , D. E. , Fiore , M. C. , & Baker , T. B. ( 1999 ). Development and validation of the Leatherdale , S. T. ( 2006 ). School-based smoking cessation pro- Wisconsin Smoking Withdrawal Scale . Experimental and Clini- grams: Do youth smokers want to participate in these programs? cal Psychopharmacology , 7 , 354 – 361 . Addictive Behaviors , 31 , 1449 – 1453 . West , R. , Ussher , M. , Evans , M. , & Rashid , M. ( 2006 ). Assessing McCarthy , D. E. , Piasecki , T. M. , Fiore , M. C. , & Baker , T. B. DSM-IV nicotine withdrawal symptoms: A comparison and ( 2006 ). Life before and after quitting smoking: An electronic Psychopharmacology , 184 , diary study . Journal of Abnormal Psychology , 115 , 454 – 466 .

Source: http://devpsychopathologyru.nl/wp-content/uploads/2012/12/2009-Nicotine-withdrawal-symptoms-following-a-quit-attempt-An-ecological-momentary.pdf