Psychopharmacology

, Volume 183, Issue 1, pp 1–12

Effect of sustained-release (SR) bupropion on craving and withdrawal in smokers deprived of cigarettes for 72 h

Authors

    • GlaxoSmithKline
  • Stephen T. Tiffany
    • Department of Psychiatry, School of MedicineUniversity of Utah
  • Lisa Squassante
    • GlaxoSmithKline
  • Stefano Milleri
    • GlaxoSmithKline
  • Luigi Ziviani
    • GlaxoSmithKline
  • Alan Bye
    • GlaxoSmithKline
Original Investigation

DOI: 10.1007/s00213-005-0145-x

Cite this article as:
Teneggi, V., Tiffany, S.T., Squassante, L. et al. Psychopharmacology (2005) 183: 1. doi:10.1007/s00213-005-0145-x

Abstract

Rationale

Sustained-release (SR) bupropion enhances quit rates of smokers, generally decreases tobacco withdrawal, and in some studies, reduces craving.

Objective

Investigate the effects of SR bupropion on craving and withdrawal during cigarette abstinence.

Methods

Twenty three smokers participated in three 17-day periods composed of 14 out-patient days followed by 3 (72 h) in-patient days. During the out-patient days, subjects received SR bupropion, placebo, or no drug. During the in-patient days, subjects were abstinent from cigarettes on two occasions while receiving either SR bupropion or placebo and smoked freely during the other occasion. SR bupropion was titrated over the first three out-patient days followed by a fixed dose (300 mg/day) for 14 days (including the three in-patient abstinence days). Cigarette craving, withdrawal, and selected physiological measures were assessed repeatedly over the 72-h periods.

Results

During the 72-h periods, craving intensity was significantly lower with free smoke and SR bupropion than with placebo, and significantly lower during free smoke than during SR bupropion. Overall withdrawal symptoms were significantly lower with free smoke than with either placebo or SR bupropion. Among individual withdrawal symptoms (excluding craving), appetite increase was significantly reduced during SR bupropion compared to placebo. During placebo and SR bupropion, craving intensity displayed a circadian pattern that was different from that observed during free smoke.

Conclusions

SR bupropion reduced craving and appetite increase during smoking abstinence. These findings support the hypothesis that craving and withdrawal symptoms may be controlled by distinct central nervous system pathways.

Keywords

SR bupropionSmokingCravingWithdrawal

Introduction

Cigarette smoking, a leading cause of death and disease in the Western world, is a major priority of public health services (World Health Organization 1997). The 1988 United States Surgeon General's Report (US Department of Health and Human Services 1988) recognized nicotine addiction as the principal mechanism responsible for the maintenance of smoking behavior (Henningfield et al. 1995; Stolerman and Jarvis 1995; Watkins et al. 2000).

The Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) of the American Psychiatric Association (1994) describes the withdrawal syndrome among the criteria for nicotine dependence. The nicotine withdrawal syndrome is characterized by negative affect, insomnia, difficulty concentrating, restlessness, decreased heart rate, and increased appetite and weight. Craving was excluded from withdrawal symptoms in DSM-IV primarily because it was believed that craving and withdrawal might not have comparable profiles during abstinence (Hughes et al. 1984, 1991, 1994; Hughes and Hatsukami 1992). However, craving is recognized as a prominent characteristic of nicotine dependence (Gritz et al. 1991; Breslau et al. 1992; UNDCP and WHO Informal Expert Committee on the Drug-Craving Mechanism 1992; Miyata and Yanagita 2001), and both craving and withdrawal are thought to contribute to the maintenance of chronic smoking and smoking relapse (Brandon et al. 1990; Glassman 1993; Killen et al. 1991; Killen and Fortmann 1997).

The primary pharmacologic approaches recommended for smoking cessation are nicotine replacement treatments (NRTs) and sustained-release (SR) bupropion (Ferry 1999; Fiore et al. 2000; Silagy et al. 2004). While NRTs are substitutes for the nicotine content of cigarettes, SR bupropion (Zyban, GlaxoSmithKline) is an antidepressant with neurophysiological mechanisms that differ from the NRTs (Ferris and Cooper 1993). The clinical efficacy of NRTs is commonly attributed to their ability to reduce both craving and withdrawal (Benowitz 1993; Hughes and Glasela 1993; Fiore et al. 1994; Hughes 2000; West 1992).

In contrast, SR bupropion significantly reduces withdrawal symptoms but has not consistently had effects on craving (Hurt et al. 1997; Jorenby et al. 1999; Shiffman et al. 2000; Jamerson et al. 2001; Tashkin et al. 2001; Durcan et al. 2002; George et al. 2002; Tønnesen et al. 2003; Dalsgrað et al. 2004). The absence of a consistent craving effect with SR bupropion is intriguing, as some researchers have shown that reduction in craving, but not withdrawal intensity, is prospectively associated with successful smoking cessation or maintained abstinence (Curry and Mc Bride 1994; Kenford et al. 1994; Killen et al. 1991; Killen and Fortmann 1997).

Compounds that ameliorate short-term craving and withdrawal symptoms are not always effective in improving smoking cessation rates compared to placebo (West et al. 1991; Hilleman et al. 1992; Klesges et al. 1995; Schneider et al. 1996). Such findings raise the question of whether a pharmacologic treatment for smoking cessation may have a mechanism of action with a simultaneous or differential impact on craving and withdrawal symptoms, or even not related to effects on craving and withdrawal symptoms (West 1992; Leischow et al. 1997; Tiffany et al. 2000).

Evaluations of the impact of medications on craving and withdrawal symptoms are complicated by numerous variations in methodology and assessments across studies (Hughes et al. 1984; Schneider and Jarvik 1984; Schneider et al. 1984; The Transdermal Nicotine Study Group 1991; Teneggi et al. 2002). In a recent laboratory study designed to address the assessment and methodological limitations of previous research (Teneggi et al. 2002), we investigated the impact of NRT patches (NRT-p) on craving and withdrawal symptoms in abstinent smokers not intending to quit. Smokers participated in three study conditions, with each condition lasting for 72 h in a controlled research unit. During one of these conditions, smokers were allowed ad lib access to cigarettes. During the other two conditions, smokers remained abstinent for the 72-h period, during which they were given placebo or NRT-p. The study showed a reduction of craving, but not of withdrawal symptoms, during NRT-p compared to placebo conditions and a circadian rhythm for craving during smoking abstinence (NRT-p and placebo), but not during a comparable period of unrestricted smoking. These findings suggested that craving and withdrawal symptoms may be differentially sensitive to NRT-p administration, and that craving intensity in abstinent smokers can vary systematically over a day. In the present study, we used the same procedures developed by Teneggi et al. (2002) to investigate the effects of SR bupropion on craving and withdrawal symptom profiles in abstinent smokers.

Materials and methods

Design

This study used a double-blind (for the SR bupropion and placebo sessions), randomized, crossover design. Healthy subjects participated in three 17-day periods composed of 14 out-patient days (from Days −14/−1) followed by 3 in-patient days (72 h, from 8 a.m. Day 1 to 8 a.m. Day 4). During the out-patient days, subjects received SR bupropion or placebo or no drug. During the in-patient days, subjects participated in three different conditions: two 72-h (from 8 a.m. Day 1 to 8 a.m. Day 4) nonsmoking periods, during which subjects received SR bupropion or placebo tablet administration, and one 72-h (from 8 a.m. Day 1 to 8 a.m. Day 4) smoking period, during which subjects were allowed to smoke ad libitum (Fig. 1). There was a washout of at least 12 days between the end of each in-patient and the start of each out-patient study period. Each subject had to attend the clinical pharmacology unit of GlaxoSmithKline in Verona (Italy) on eight occasions: a prestudy screening visit, at the beginning of each of the three out-patient periods, during the three 72-h in-patient periods, and a poststudy follow-up visit. A screening visit was performed within 21 days of admission to the research facility for the first out-patient study period. Subjects entered the research facility by 7 pm on the evening before the start of each in-patient period (Days 1–4) when assessments of craving, withdrawal, and other variables were administered. A follow-up visit was conducted 7–14 days after the last in-patient period (Fig. 1). The total duration of a subject's participation was approximately 12 weeks from the time of the screening visit to study discharge.
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0145-x/MediaObjects/213_2005_145_Fig1_HTML.gif
Fig. 1

Study diagram

Subjects

A total of 23 healthy volunteers, 12 males and 11 females, were recruited. Inclusion criteria established that subjects were at least 18 years old; had smoked an average of 15 cigarettes or more a day for the past year; were not intending to quit smoking; had Fagerström Tolerance Questionnaire (FTQ) (Fagerström 1978) scores of at least 7; were healthy as determined by physical, neurological, and psychiatric examination [including an assessment of medical history, electrocardiogram (ECG), and laboratory studies]; were able to read, comprehend, and write in the language of the investigator; and completed a written informed consent prior to study participation. Exclusion criteria included any history of drug allergies, drug, or alcohol abuse; use of long-term medication; use of any drug within the previous 4 weeks; current use of any NRT; alcohol intake during the 2 days prior to the study or illness at the time of the study; and previous participation in any clinical study during the 6 months before the first study day. Regulatory and ethical committee approval was obtained before the start of the study. The study was conducted in accordance with the Declaration of Helsinki.

Treatment

Drug medications were SR bupropion tablet (Zyban) or placebo. They were administered as one tablet (150 mg/day or placebo) on the first 3 days of the out-patient periods (from Days −14/−12). Then, two tablets (300 mg/day or placebo) were administered, one in the morning and one in the afternoon (at least 8 h apart), during the remaining 11 days of the out-patient periods and during the 3 days of the in-patient periods, for a total of 17 consecutive days (Fig. 1). SR bupropion tablets were of the Italian-marketed product Zyban 150 mg (GlaxoSmithKline SpA). Placebo tablets were manufactured by GlaxoSmithKline SpA according to internal specification.

Procedure

Subjects who met inclusion criteria were admitted to the unit on the evening before the start of the in-patient periods (Day 1) and remained until 6 h after the end of each in-patient period assessment on Day 4. During each in-patient period, subjects had to refrain from alcohol, avoid consumption of coffee (except for breakfast and lunch), avoid any prescribed concomitant medication and any over-the-counter preparations, refrain from strenuous exercise, and refrain from all recreational drugs (“drugs of abuse;” urine tests were performed during the study to monitor compliance with this restriction). When subjects received SR bupropion or placebo tablets, they quit smoking by 8 a.m. on Day 1 and remained abstinent until 8 a.m. of Day 4 of the in-patient period. When assigned to the free smoke condition, subjects continued to smoke ad libitum during the in-patient period.

Measurements and evaluations

The measurements were taken at various time intervals through the in-patient study periods. Time zero (0), around 8 a.m., was near the beginning of the enforced abstinence, 30 min after the subject's last cigarette. Self-report questionnaires were administered at 8 a.m., 11 a.m., 2 p.m., and 8 p.m. on Day 1; at 8 a.m., 2 p.m., and 8 p.m. on Days 2 and 3; and at 8 a.m. on Day 4 to evaluate craving and withdrawal symptoms.

The Questionnaire of Smoking Urges (QSU) Brief (Cox et al. 2001), used to assess craving/urge to smoke, consists of ten items scored on a Likert scale from 1 (I strongly disagree) to 7 (I strongly agree). The average total score ranges from 1 to 7. Higher scores represent increased craving. In addition to the average total score, the QSU Brief can be decomposed into two craving factors. Factor 1 represents desire and intention to smoke along with anticipation of pleasure from smoking, whereas Factor 2 represents an urgent desire to smoke and the anticipation of relief from negative effect and nicotine withdrawal. Previous research shows that the QSU Brief displays excellent reliabilities and is significantly associated with multiple smoking-relevant variables (Cox et al. 2001).

The Smoking Withdrawal Questionnaire (SWQ) (Shiffman and Jarvik 1976), used to assess withdrawal symptoms, consists of 25 items scored on a Likert scale from 1 (very definitely) to 7 (very definitely not). The scale yields five factors (Craving, Psychological Discomfort, Physical Symptoms, Stimulation–Sedation, and Appetite), with scores ranging from 1 to 7, with higher scores representing lower levels of withdrawal (to facilitate the interpretation of these factors in this paper, the SWQ data were inverted in the analyses and figures, with higher scores representing higher levels of withdrawal).

The Smoker Complaint Scale (SCS) (Schneider and Jarvik 1984; Schneider 1994), used to assess withdrawal symptoms, consists of 20 items scored on a Likert scale from 1 (very definitely not) to 7 (very definitely). The sum total score ranges from 20 to 140. Higher scores represent higher levels of withdrawal symptoms. Reliability and validity of these two questionnaires have been discussed by Patten and Martin (1996a).

Saliva samples for nicotine and cotinine levels were collected at 8 a.m., 2 p.m., and 8 p.m. on Days 1, 2, and 3, and at 8 a.m. on Day 4. Saliva samples were analyzed for nicotine and cotinine levels using liquid chromatography mass/mass (LC-MS/MS) assay at limit of quantification (LOQ) of 5 ng/ml, with 1 ml sample. Breath carbon monoxide (CO) measurements were performed at 8 a.m., 2 p.m., and 8 p.m. on Days 1, 2, and 3 and at 8 a.m. on Day 4 using an EC50 Micro III Smokerlyzer (Bedfont Instruments).

Statistical analyses

It was calculated that a sample of 21 subjects yielded a power of 80% to detect a difference between treatments of 0.5 in craving and 10 in withdrawal scores. These differences were deemed clinically significant and corresponded to a medium effect size (ES ∼0.4). Sample size was calculated assuming a residual variability of 0.54 for craving and 9.7 for withdrawal, as observed previously in the same experimental setting. The two-sided α level was set at 0.05. Total scores and factors of craving and withdrawal symptoms, as well as single items and sum of craving- and noncraving-related items of SCS, were first summarized for each subject by the weighted mean over each 72-h period. This was calculated as the area under the measurement–time curve, using the linear trapezoidal method divided by the time over which the measurements were taken. The weighted means of each total score, factor, item of interest, and the measures at the start of in-patient study period (8 a.m. on Day 1) were then analyzed by analysis of variance (ANOVA) for a crossover study design, taking into account subjects, periods, and treatments. Estimates of least square (LS) mean values were evaluated for each treatment, and treatment differences were estimated on a pairwise basis. Effect sizes (Cohen 1988, 1992) of treatment differences were calculated. Due to the exploratory nature of the analysis, no correction for multiplicity of tests was performed. Total scores and factors of craving and withdrawal symptoms were also analyzed separately for each treatment as a function of time of day and to capture any linear and circadian trends (as a quadratic component) across successive 24-h intervals. Measures of saliva cotinine concentration (log-transformed) and exhaled CO at the start of an in-patient study period were compared between treatments by means of ANOVA for crossover.

Results

Subjects

A total of 23 subjects, with an average age of 34 (±13), who smoked an average of 23 (±6) cigarettes a day, and had an FTQ score of 8 (±1), participated in the study. Data for the analysis came from 20 subjects who completed all three in-patient periods and 1 subject who completed two of three in-patient periods. Three subjects did not complete the study: two withdrew for personal reasons during the placebo out-patient period, and one withdrew because of nausea and headaches during the SR bupropion out-patient period.

Below, we will examine the following: firstly, relevant measures at the start (8 a.m.) of the in-patient study periods; secondly, whether placebo differs from free smoking and SR bupropion conditions; thirdly, whether SR bupropion and free-smoking conditions differ; and finally, key physiologic measures.

Measures at the start of in-patient study periods

The averages by treatments at the start of in-patient period (8 a.m. on Day 1) of self-report measures and biological indices of smoking are shown in Figs. 2, 3, and 4 and Figs. 5 and 6, respectively. There were no significant differences between the treatments on any of the self-report measures and for initial levels of saliva nicotine (not shown). In contrast, SR bupropion was associated with significantly lower (p<0.01) levels of saliva cotinine (Fig. 5) compared to placebo and free smoke. CO levels were significantly lower (p<0.01) during free smoke compared to SR bupropion and placebo (Fig. 6).
https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0145-x/MediaObjects/213_2005_145_Fig2_HTML.gif
Fig. 2

Questionnaire on Smoking Urges Brief (QSU Brief): mean craving total score

https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0145-x/MediaObjects/213_2005_145_Fig3_HTML.gif
Fig. 3

Smoking Withdrawal Questionnaire (SWQ): mean craving score (Factor 1)

https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0145-x/MediaObjects/213_2005_145_Fig4_HTML.gif
Fig. 4

Smoker Complaint Scale (SCS): mean withdrawal sum total score

https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0145-x/MediaObjects/213_2005_145_Fig5_HTML.gif
Fig. 5

Mean saliva cotinine concentration

https://static-content.springer.com/image/art%3A10.1007%2Fs00213-005-0145-x/MediaObjects/213_2005_145_Fig6_HTML.gif
Fig. 6

Mean exhaled carbon monoxide (CO) concentration; ppm, parts per million

Craving

Questionnaire of Smoking Urges Brief

Table 1 shows craving levels for the average total and factor scores from the QSU Brief. Figure 2 displays the mean 72-h profile of the average total score for each treatment. During placebo, total QSU Brief craving was significantly higher than during free smoke or SR bupropion. Factor 1 was also significantly higher than during free smoke or SR bupropion, while Factor 2 was significantly higher than during free smoke and nearly significant (p=0.065) compared to SR bupropion. During free smoke, craving levels on total QSU Brief and Factors 1 and 2 were significantly lower than during SR bupropion.
Table 1

Summary of statistical key results: means and treatment differences

 

Least square means

Differences (effect size)

Free smoke

Bupropion

Placebo

Free-smoke–placebo

Bupropion–placebo

Free-smoke–bupropion

QSU brief (craving)

 Average total score

3.6

4.4

5.0

−1.4*** (1.0)

−0.6* (0.4)

−0.8** (0.6)

 Factor 1 (anticipation of pleasure of smoking)

4.6

5.4

6.0

−1.4*** (1.1)

−0.6* (0.5)

−0.8** (0.6)

 Factor 2 (anticipation of relief of smoking abstinence)

2.5

3.4

3.9

−1.4*** (0.9)

−0.5 (0.3)

−0.9** (0.5)

SCS (withdrawal)

 Sum total score

44

52

57

−13*** (0.7)

−5 (0.3)

−0.8* (0.4)

 Sum score of craving items

10

14

16

−6.6*** (1.7)

−1.9* (0.5)

−4.7*** (1.2)

 Sum score noncraving items

34

38

41

−6.5* (0.4)

−3.3 (0.2)

−3.1 (0.2)

SWQ (withdrawal)

 Factor 1 (craving)

4.7

5.6

6.0

−1.3*** (1.3)

−0.4* (0.4)

−0.9*** (0.9)

 Factor 2 (psychological discomfort)

4.1

4.4

4.5

−0.5** (0.6)

−0.1 (0.2)

−0.3* (0.4)

 Factor 3 (physical symptoms)

2.0

2.2

2.3

−0.3* (0.3)

−0.1 (0.1)

−0.1 (0.2)

 Factor 4 (stimulation–sedation)

5.1

5.2

4.9

0.2 (0.2)

−0.3 (0.2)

−0.1 (0.0)

 Factor 5 (appetite)

4.1

3.9

4.1

0.0 (0.0)

−0.2* (0.4)

−0.2 (0.3)

QSU Brief Questionnaire on Smoking Urges Brief, SCS Smoker Complaint Scale, SWQ Smoking Withdrawal Questionnaire

*p<0.05; **p<0.01; ***p<0.001

In addition to these overall treatment effects, there was evidence of systematic changes in daily craving levels, with distinctly different patterns across treatment conditions. During placebo, there was a progressive increase for the total QSU Brief craving rating over the first day of abstinence, with the lowest level in the morning and maximum measured intensity at the 8 p.m. assessment. That pattern was reflected in a significant negative quadratic trend (p<0.01). There was no significant quadratic trend on Days 2 or 3 of abstinence. There were no changes for placebo in the average daily intensity of craving over the 3 days of abstinence. During free smoke, total QSU Brief craving ratings achieved the minimum measured intensity around 2 p.m. and maximum intensity around 8 a.m. during each of the three study days. Those patterns were reflected in a significant positive quadratic trend (p<0.001) on each of the 3 days. On the morning (8 a.m.) of Days 2, 3, and 4, craving values of free smoke and SR bupropion were overlapping, while in the afternoon (2 p.m.) of Days 2 and 3, craving values of free smoke were lower than those of SR bupropion and placebo. There were no changes in the average daily intensity of craving over the 3 days of free smoke.

During SR bupropion treatment, there was a progressive increase in the total QSU Brief craving rating over the first day of abstinence, with the lowest level in the morning and maximum measured intensity at the 2 p.m. assessment. This was associated with a significant negative quadratic trend (p<0.01) for craving ratings on the first day of abstinence. There was no significant quadratic trend on Days 2 or 3 of abstinence.

Smoking Withdrawal Questionnaire Factor 1

Table 1 shows the results of the analysis for Craving factor from the SWQ, and Fig. 3 shows the 72-h mean profiles of the Craving factor. Craving levels were highest during placebo, lowest during free smoke, and intermediate to those two conditions during SR bupropion. All differences between conditions were significant.

Across each 72-h period, craving levels assessed by the SWQ Craving factor (as with the QSU Brief) showed systematic daily variations, with different patterns in each treatment condition. On the first and second day of abstinence in both the placebo and SR bupropion conditions, craving levels were at the lowest intensity in the morning, with a progressive increase during each day. There was a significant negative quadratic trend on Days 1 and 2 (p<0.001 and p<0.05, respectively) but not on Day 3 of abstinence. During free smoke, a circadian pattern with a positive quadratic trend (p<0.001) was evident on each of the three study days. This effect was opposite in direction to the pattern displayed during placebo and SR bupropion. Craving achieved the minimum measured intensity around the 2 p.m. assessment during each of the three study days, without changes in the average intensity across each of the 3 days.

Craving-related Smoker Complaint Scale items

We explored individual items of the SCS questionnaire to identify whether differences between conditions were evident on particular SCS items related to craving (Table 1). The sum of the three SCS craving items (“craving a cigarette,” “missing a cigarette,” and “urge to smoke”) produced a pattern of effects identical to that observed for the QSU Brief and Craving factor (Factor 1) of the SWQ, with significant differences between treatments. As with the QSU Brief and SWQ Craving factor, craving intensity, assessed by SCS items, showed systematic daily variation across each 72-h period. During Day 1 of abstinence in both the placebo and SR bupropion conditions, the craving items score was at the lowest level in the morning, with a progressive increase during each day. There was a significant negative quadratic trend (p<0.001) on Day 1, but not on Days 2 and 3 of abstinence. During free smoke, a circadian pattern with a positive quadratic trend was evident on two out of three craving items (p<0.001) on Days 1 and 2, but not on Day 3.

Withdrawal

Smoker Complaint Scale

Treatment comparisons for the SCS sum total scores are summarized in Table 1. Figure 4 shows the mean 72-h profile of the SCS sum total scores for the three treatments. Withdrawal was significantly less intense during free smoke than during placebo or SR bupropion. Although withdrawal was marginally stronger during placebo than during SR bupropion, this difference was not statistically significant (p=0.125).

As with craving, we explored individual items of the SCS questionnaire to identify whether differences between placebo, free smoke, and SR bupropion were restricted to particular SCS items. During placebo, the five items related to psychological discomfort (“anxiety,” “irritability,” “negative mood,” “restlessness,” and “annoyance”) showed higher scores (p<0.05) compared to free smoking (but not to SR bupropion). During placebo, the appetite item score was not different when compared to free smoke, while it was increased, although not significantly (p=0.073), in comparison with SR bupropion. During free smoke, the appetite item score was significantly higher (p<0.05) than during SR bupropion.

Smoking Withdrawal Questionnaire Factors 2–5

Table 1 summarizes results of withdrawal factors 2–5 from the SWQ. Psychological Discomfort (Factor 2) was significantly more intense during both the placebo and SR bupropion conditions relative to the free smoke condition. Physical Symptoms (Factor 3) were significantly more intense during placebo than during free smoke. Stimulation–Sedation (Factor 4) did not show significant differences across any of the treatment comparisons. Appetite (Factor 5) showed no difference between placebo and free smoke, while it was significantly more intense during placebo than during SR bupropion. Thus, among these noncraving factors, only Appetite appeared to be sensitive to the SR bupropion administration.

Effect sizes

Table 1 also shows the effect sizes (Cohen's d) associated with the treatment comparisons across the self-report measures of craving and withdrawal. The craving effect sizes across all craving-related measures (QSU Brief, SWQ Factor 1, and SCS craving items) associated with the abstinence effect (i.e., free-smoke–placebo) averaged 1.3, which, by statistical conventions, is considered to be large (Cohen 1988). The withdrawal effect sizes (i.e., self-report measures from the SCS and the SWQ devoid of craving items) for the abstinence effect averaged 0.3, which is considered small. The average effect size for the SR bupropion effects (SR bupropion–placebo) was around 0.4 (medium) for craving measures and around 0.2 (small) for withdrawal measures.

Physiological measures

Figure 6 shows that, for both SR bupropion and placebo, levels of exhaled CO dropped to very low levels (<6 ppm) after 12 h of smoking cessation. In contrast, CO reached very high levels during the free smoke with a clear circadian profile, with the highest value in the evening and the lowest in the morning. Examination of mean saliva nicotine concentration across days for each condition revealed a parallel pattern of results (not shown). As expected, there were higher levels of nicotine during free smoke than during SR bupropion or placebo. During SR bupropion and placebo, the nicotine levels dropped with a progressive reduction to a very low level (<25 ng/ml) within 6 h after cessation of smoking. As with nicotine, there were higher levels of saliva cotinine during free smoke than during SR bupropion or placebo (Fig. 5). With SR bupropion and placebo, cotinine dropped with a progressive reduction to a very low level (<30 ng/ml) within 54 h after cessation of smoking.

Discussion

This study showed that, relative to a placebo condition, SR bupropion significantly reduced the intensity of craving across all craving-relevant measures (QSU Brief, SWQ Factor 1, and SCS craving items). In contrast, SR bupropion had no significant impact on overall withdrawal intensity, although it did produce some reduction of the SCS total score. With regard to individual withdrawal symptoms, there was evidence of significant differences in favor of SR bupropion on reduction of appetite increase, but not on other signs and symptoms.

The findings that SR bupropion reduced craving but not withdrawal symptoms (except reduction of appetite increase) relative to the placebo condition stands in contrast to results reported by Shiffman et al. (2000). The Shiffman et al. study and the present research shared several features. Subjects did not intend to quit smoking and did not receive psychological counseling. They were exposed to nearly identical dose regimens of SR bupropion during both the preabstinence and abstinence phases. Measures of craving and withdrawal were also obtained at multiple intervals across the 3 days of abstinence for SR bupropion and placebo conditions. Shiffman et al. (2000) reported that the 300-mg dose of SR bupropion had no effect on craving levels but reduced some symptoms of smoking withdrawal, namely, depression, irritability, and difficulty concentrating.

Most importantly, Shiffman et al. (2000) found very little evidence of abstinence-induced craving in their study, whereas the levels of abstinence-induced craving we found were clearly large, with average effect sizes for the placebo vs free-smoke comparison across craving measures approximately 1 or higher. The difference in craving between the present study and the Shiffman et al. (2000) research can be attributed to several factors. Notably, although the subjects were of the same age across the two studies (34 years), the subjects in the Shiffman et al. research were heavier smokers than those in the present experiment (35 vs 23 cigarettes/day and 9 vs 8 on the FTQ score, respectively). As a consequence of their relatively heavier smoking, it is possible that craving levels among smokers in the Shiffman et al. (2000) study were at the ceiling and somewhat less sensitive to treatment. Moreover, it is possible that the procedures and instruments used in our research were more sensitive to the effects of SR bupropion on craving than those used by Shiffman et al. (2000). For example, we assessed craving with the QSU Brief, a 10-item scale (Cox et al. 2001) and the SWQ Factor 1 (Shiffman and Jarvik 1976), a 6-item scale. Both scales exhibit excellent reliability and are associated with considerable validation data. Shiffman et al. (2000) used a nonpublished and unvalidated 6-item scale, and it is possible that this scale does not exhibit the psychometric sensitivity of the QSU Brief or the SWQ Factor 1 scale. Moreover, Shiffman et al. (2000) measured craving effects as the difference between baseline craving and craving during abstinence. As difference scores are always less reliable than the component scales from which they are derived (Cohen and Cohen 1975), the ability of this craving score to detect placebo and SR bupropion effects during abstinence would be restricted relative to more reliable craving measures.

Although several researchers have also reported that bupropion does not alter craving levels (George et al. 2002; Hurt et al. 1997; Jorenby et al. 1999; Jamerson et al. 2001), many others have found significant effects of SR bupropion on craving (Dalsgrað et al. 2004; Durcan et al. 2002; Tashkin et al. 2001; Tønnesen et al. 2003). Moreover, researchers have found that bupropion reduces the impact of craving on relapse (Durcan et al. 2002) and the report of craving as a problem during abstinence (Tønnesen et al. 2003; Tonstad et al. 2003).

Shiffman et al. (2000) also found evidence of SR bupropion effects on some withdrawal symptoms (depression, irritability, and difficulty concentrating) but not on appetite. A nearly opposite pattern of effects was obtained in the present research. These disparities might also be attributed to differences in levels of smoking across the samples and in the use of different measures and procedures to assess nicotine withdrawal. In our research, we assessed withdrawal symptoms using the SCS (Schneider and Jarvik 1984; Schneider 1994) and SWQ Factors 2–5 (Shiffman and Jarvik 1976), while Shiffman et al. (2000) used unvalidated item sets representing DSM-IV criteria for nicotine withdrawal (American Psychiatric Association 1994). The impact of abstinence on withdrawal in our study was relatively smaller than on craving. Differences in the content, reliability, sensitivity, and dimensionality of the withdrawal instruments across studies might account for the differential pattern of the effects of SR bupropion on withdrawal.

Other studies have also shown that SR bupropion reduces withdrawal symptoms (Hurt et al. 1997; Jorenby et al. 1999; Tashkin et al. 2001; Jamerson et al. 2001; Tønnesen et al. 2003; Dalsgrað et al. 2004). Most of these studies also reported a reduction in weight gain with SR bupropion compared to placebo (Hurt et al. 1997; Jorenby et al. 1999; Ahluwalia et al. 2002; Dalsgrað et al. 2004; Tashkin et al. 2001;Tonstad et al. 2003; Tønnesen et al. 2003), which is consistent with our finding of reduction of appetite increase. All of these studies were conducted with smokers who were out-patients, intending to quit, receiving psychological counseling, and treated for a relatively long period (from 7 to 12 weeks). In contrast, we tested in a laboratory setting, and our subjects were not intending to quit smoking, not receiving psychological counseling, and participating in a predefined, short-term abstinence (72 h). Any or some combination of these factors may have contributed to our failure to observe SR bupropion effects on withdrawal symptoms other than appetite. This hypothesis is supported by the fact that some researches have found that motivation to quit and psychological counseling can have a favorable influence on smoking withdrawal symptoms and abstinence rates (Prochaska and Goldstain 1991; Sciamanna et al. 2000; Gorin and Heck 2004).

The overall findings suggest that evaluations of the impact of medications on craving and withdrawal symptoms are complicated by numerous variations in methodology and assessments across studies. Relevant variations may be the scales used, the frequency of assessments of craving and withdrawal symptoms, the intensity of craving and withdrawal symptoms, the out-patient vs laboratory-based nature of the studies, or other fundamental differences in study design (Hughes et al. 1984; Schneider and Jarvik 1984; Schneider et al. 1984; Patten and Martin 1996a,b; Shiffman et al. 2000, 2004; Teneggi et al. 2002). As an example of this, laboratory-based studies of withdrawal and craving are often limited to short periods of sustained abstinence, and they rarely include a within-subject control condition of ad lib smoking. In the absence of this control condition, it is difficult to unambiguously attribute craving and withdrawal changes to nicotine abstinence.

Our results are compatible with the hypothesis that smoking craving and withdrawal are distinct entities controlled by separate processes with different sensitivity to drug treatments. This hypothesis is in line with a wide body of experimental evidence that associates drug craving with the alteration of dopaminergic pathways, the so-called “reward circuit” (Balfour 1994; Corrigal et al. 1992; Di Chiara and Imperato 1988; Di Chiara 2000; Koob 2000; Miyata and Yanagita 2001), while drug withdrawal is mainly attributed to the alteration of nondopaminergic pathways involving noradrenergic systems and/or a corticotropin-releasing factor (Miller and Gold 1993; Koob 2000, 2003).

Craving profiles of QSU Brief, the SWQ Factor 1, and the SCS craving items showed that, during abstinence (with placebo and SR bupropion), craving was lowest in the morning of Day 1 and then increased progressively over the course of the day, with the maximum levels in the evening. This same pattern was repeated on Day 2 for the SWQ Factor 1, but not for the QSU Brief and SCS craving items. These cyclical variations in craving were not affected by the administration of SR bupropion. Across the 72-h assessment period, craving levels in the free-smoke condition also varied systematically over each successive 24-h period; craving was highest in the morning, declined to its lowest level around 2 p.m., and then increased at the evening assessment. This cyclical pattern was opposite to the daily patterns observed in the placebo and SR bupropion conditions on Days 1 and 2.

The circadian changes observed in this research during abstinence were similar to those reported previously from this laboratory (Teneggi et al. 2002). Higher levels of craving in abstinent smokers during the evening hours may arise from one or a combination of factors, including the presence of cues strongly associated with evening smoking, the great disparity between levels of evening nicotine usually achieved during ad lib smoking, and nicotine abstinence or endogenous processes controlled by circadian rhythms. Although exposure to smoking cues was somewhat diminished on the research unit relative to a real-world situation, subjects were exposed to cues associated with evening smoking such as the evening meal, evening television shows, or time of day stimuli. Previous research with abstinent smokers has shown that cue-elicited craving and circadian increases in craving are not reduced by nicotine administration (Davies et al. 2000; Tiffany et al. 2000; Teneggi et al. 2002). However, the craving patterns obtained during placebo across our two laboratory studies are in line with previous findings from Schneider and Jarvik (1984).

The circadian changes observed in this study during the free-smoke condition may also be due to the joint impact of countervailing forces. First, as the subjects smoked over the course of the day, their nicotine levels may have risen sufficiently to counteract their craving associated with overnight nicotine abstinence. Indeed, examinations of saliva nicotine and exhaled carbon monoxide revealed that nicotine and CO levels were lowest in the morning and peaked at the 8 p.m. assessment. However, subjects in the free-smoke condition might not have been able to achieve sufficient levels of nicotine to attenuate the rise in craving over the day driven by cues associated with evening smoking or endogenous circadian processes (Davies et al. 2000; Tiffany et al. 2000; Teneggi et al. 2002).

Circadian changes found in the present study during free smoke were not observed in the study performed previously by this laboratory (Teneggi et al. 2002). This difference across studies may be due to the fact that, in this study, subjects were somewhat heavier smokers as suggested by the prolonged numbers of years of smoking (17 vs 11), the higher FTQ scores (8 vs 7.8), the greater number of daily cigarettes (23 vs 20), and the higher average craving intensity during the placebo condition (5 vs 4.5). Consequently, subjects in this study may have self-administered a higher level of nicotine during the morning of the free-smoke condition, with greater reduction of craving score. These smokers may have also been more sensitive to nonnicotine-craving forces during the remaining part of the day, with a resultant increase in craving (Davies et al. 2000; Tiffany et al. 2000; Teneggi et al. 2002). The identification of the factors that control circadian patterns in craving levels during ad lib smoking and abstinence require further investigation.

Withdrawal profiles of SCS sum total scores showed that, as with craving, there was a circadian pattern, with a progressive increase in withdrawal over Day 1 during abstinence conditions. During the free-smoke condition, there was also a circadian pattern, and withdrawal intensity was significantly lower than during the placebo and SR bupropion conditions. This difference was greatest at around 2 p.m. but was slightly reduced during the remaining part of the day. The similarities between craving and withdrawal profiles can be attributed to the fact that the SCS is a composite measure, with craving items contributing substantially to the total score. As with craving, circadian changes in withdrawal seen during free smoke were not observed in our previous research (Teneggi et al. 2002). However, the withdrawal and craving patterns obtained during placebo across our two laboratory studies are in line with previous findings from Schneider and Jarvik (1984).

The pharmacological mechanisms of SR bupropion responsible for its therapeutic efficacy and other main behavioral effects (i.e., reduction in craving, withdrawal, appetite, and weight gain) are not completely known. Initial preclinical studies have shown that SR bupropion operates primarily as a norepinephrine reuptake inhibitor with a somewhat more modest effect as a dopamine reuptake inhibitor (Cooper et al. 1994; Horst and Preskorn 1998). More recent in vitro studies indicate that SR bupropion also produces a noncompetitive blockade on nicotinic acetylcholine receptors (Slemmer et al. 2000) and enhances serotonin levels (Dong and Blier 2001). Other authors have suggested that bupropion may influence N-methyl d-aspartate (NMDA) receptor systems through a normalization of plasma levels of agmatine, an NMDA receptor antagonist (Halaris et al. 1999). Finally, recent in vivo human positron emission tomography studies have shown that SR bupropion and its metabolites block dopamine transporters in specific brain regions, resulting in increased levels of dopamine (Meyer et al. 2002; Learned-Coughlin et al. 2003). Thus, SR bupropion appears to have complex effects across several neurotransmitter systems that may be involved in smoking dependence. The extent to which any or a combination of these effects is responsible for the impact of SR bupropion on craving levels and smoking cessation requires further investigation.

The significantly initial (8 a.m.) lower levels of saliva cotinine (Fig. 5) during SR bupropion, compared to placebo and free smoke, can be attributed to the lower number of cigarettes smoked by subjects in the SR bupropion condition over each of the last 4 days preceding the in-patient part of the study relative to the other two conditions (SR bupropion, 20.8; placebo, 22.4; and free smoke, 22.5). These SR-bupropion-induced differences in smoking rates were in line with the finding that the initial intensity of craving was lower (although not statistically significant) during the SR bupropion condition compared to the free smoke and placebo conditions.

Initial levels of exhaled CO during SR bupropion and placebo were significantly higher than initial CO levels during free smoke (Fig. 6). One possible source of this difference might be that subjects smoked a higher number of cigarettes or smoked cigarettes with greater intensity before the first CO assessment during SR bupropion and placebo in an attempt at compensatory nicotine loading before abstinence.

In conclusion, this short-term study on smokers not intending to quit and not receiving psychological counseling showed that during abstinence, SR bupropion significantly reduced the intensity of craving and appetite increase compared to placebo. A previous short-term clinical study (Shiffman et al. 2000), similar in design and method to present research, showed opposite results. However, our findings were consistent with the reduced craving and reduced weight increase in clinical studies of patients attempting to quit smoking. This suggests that smoker populations, methods of assessment, study designs, and treatment durations may influence these behavioral endpoints. Bupropion SR, as NRT-p, reduces craving in a short-term laboratory setting and is therapeutically effective in smoking cessation trials. Bupropion SR produced similar profiles of smoking craving and withdrawal symptoms as found in our short-term clinical study of NRT-p, with craving appearing as a more sensitive marker of drug effects than withdrawal symptoms. These findings support the hypothesis that craving and withdrawal symptoms may be controlled by somewhat distinct central nervous system pathways. Further investigations are needed to evaluate this proposal and to more clearly establish the relative impact of SR bupropion on craving and withdrawal. The short-term clinical setting we used to assess craving and withdrawal symptoms during smoking abstinence has been proven to be sensitive to different pharmacological treatments for smoking cessation (NRT-p and SR bupropion). Finally, this short-term clinical setting revealed circadian patterns in craving and withdrawal across two distinct studies. These results support the continued use of this experimental procedure as a suitable model for investigating the short-term effects of potential new smoking cessation aids.

Acknowledgements

The research was funded by GlaxoSmithKline, Italy.

Copyright information

© Springer-Verlag 2005