|Home | About | Journals | Submit | Contact Us | Français|
We have previously demonstrated that acute smoking abstinence is associated with lowered reward motivation and impaired response inhibition. This prospective study explores whether these impairments, along with withdrawal-related symptoms, recover over three months of sustained abstinence.
Participants completed a 12-hour abstinent baseline assessment and were then randomly allocated to quit unaided or continue smoking. All were re-tested after 7 days, 1 month and 3 months. Successful quitters’ scores were compared with those of continuing smokers, who were tested after ad libitum smoking.
Goldsmiths, University of London.
33 smokers who maintained abstinence to 3 months, and 31 continuing smokers.
Indices previously demonstrated in this cohort of smokers to be sensitive to the effect of nicotine vs. acute abstinence: reward motivation (SHAPS, CARROT, Stroop); tasks of response inhibition (antisaccade task; CPT) and clinical indices of mood (HADS), withdrawal symptoms (MPSS) and desire to smoke.
SHAPS anhedonia and reward responsivity (CARROT) showed significant improvement and plateaued after a month of abstinence, not differing from the scores of continuing smokers tested in a satiated state. Mood, other withdrawal symptoms and desire to smoke all declined from acute abstinence to 1 month of cessation and were equivalent, or lower than, the levels reported by continuing, satiated smokers. Neither group showed a change in CPT errors over time whilst continuing smokers, but not abstainers, showed improved accuracy on the antisaccade task at 3 months.
Appetitive processes and related affective states appear to improve in smokers who remain nicotine-free for 3 months whereas response inhibition does not. Although in need of replication, the results tentatively suggest that poor inhibitory control may constitute a long-term risk factor for relapse and could be a target for intervention.
Acute abstinence from regular smoking is associated with somatic symptoms, low mood [1,2] motivational and cognitive disturbance [3-7] and impaired inhibitory control [8-10]. Which of these dysfunctions develop as a consequence of smoking, or are constitutional deficits that predate regular smoking, remains uncertain. If the former is the case, and if neuroadaptations are reversible, then such impairments should normalise with continuing abstinence; if the latter is the case, they would not be expected to improve even with protracted abstinence. Since the time course and trajectory of such impairments may affect the ability of the smoker to maintain abstinence, it is useful to chart their natural history.
Very few prospective studies have included measures of cognition. In one recent study in which mental arithmetic was assessed post-cessation , scores returned to baseline (pre-quit) levels within ten days of quitting. However, Gilbert and colleagues  noted reduced accuracy on the rapid visual information processing (RVIP) task and EEG deactivation in abstinent smokers which failed to return to either pre-quit baseline, or smoking control levels after 31 days.
A handful of cross-sectional studies have compared current, ex, and non-smokers on indices of reward motivation and cognition; in some, impairments were demonstrated by current smokers but not ex-smokers [13, 14], whilst in others the opposite was found . Whilst this may reflect the particular tests used, a recent EEG study  found hypoactivation of the pre-frontal cortex and anterior cingulate in smokers who had been abstinent for a mean of 11 years.
With respect to abstinence-related affective and somatic symptoms, nicotine withdrawal is characterised by anxiety, depression, irritability, restlessness, difficulty concentrating and desire to smoke [1,2]. Early studies suggested that such symptoms recover within a few weeks [17-19]; however, reported durations in recent studies have ranged from less than two weeks [20,11] to over a month [21-23]. In one of very few studies which have tracked withdrawal symptoms for longer periods, Piasecki and colleagues [24,25] found considerable heterogeneity in the profiles of symptoms presented by different individuals (75% of whom used a nicotine patch), and approximately 40% showed an increase in the severity of withdrawal symptoms after 30 days of abstinence, . In a related study they found comparable levels of depression, anhedonia and somatic features (e.g. decreased appetite, lack of motivation, impaired sleep) in both current and ex-smokers (abstinent for at least 6 months) relative to never-smokers .
Prospective studies exploring the time course of abstinence-related responses are, however, methodologically complex [26,22]; a major weakness is selective dropout which can seriously bias conclusions. For example, high relapse rates may distort mean scores in the residual sample if relapse is more likely in those who experience the greatest severity of symptoms or cognitive impairment in the early stages of abstinence. Relatedly, successful abstainers may be characterised by atypical levels of motivation, or may differ in other important ways from the broader population of smokers; hence their experience may not generalise to all smokers. In studies which lack a randomly assigned ‘continuing-to-smoke’ control group, the interpretation of findings may be limited by practice or familiarity effects.
The present study explored changes in withdrawal symptoms, mood, reward motivation and response inhibition over 3 months of continuous smoking abstinence, comparing smokers randomly allocated between ‘quit’ and ‘continue-to-smoke’ conditions.
Smokers were recruited through adverts in local newspapers, radio stations, colleges, libraries and pharmacies in the South East London area. All were aged between 18 and 65 and smoked ten or more cigarettes a day and within the first hour of waking. Exclusion criteria included current diagnosis of psychiatric or neurological condition, regular use of prescription or class A recreational drugs, and pregnancy. All participants gave written informed consent and the study was approved by Goldsmiths Ethics Committee.
This was part of a larger study [5,8]. 145 smokers (57% female; 78% Caucasian) were assessed on two occasions after overnight (12 hour) abstinence (confirmed by breath CO levels < 11ppm), once after receiving a 4-mg NiQuitin lozenge and once after receiving a placebo lozenge (order counterbalanced; procedure double-blind).
Immediately after the second session participants were randomly allocated in a 3:1 ratio between ‘quit’ and ‘continue-to-smoke’ conditions. Those in the quit group were provided with written information and advice but were prohibited from using NRT or other medications. Assessments were repeated at 7 days, 1 month and 3 months, and any quitters who reported more than one lapse since the last session or who had a salivary cotinine level of more than 20 ng/ml were classified as relapsers. Participants allocated to the ‘continue to smoke’ condition were required to smoke a cigarette just before each follow-up testing session to ensure that they were tested in a satiated state.
All participants were offered incremental financial incentives for participating in each successive assessment, subject to their compliance with the requirement to remain abstinent or continue smoking. Participants did not receive any payment until their final assessment. The maximum payment was £150 for those who attended all 5 sessions.
Demographic information included age, gender, and years of education.
The full assessment battery is detailed in Dawkins et al. . Here we focus on only those indices found in the full sample to be impaired during acute abstinence (i.e. in the placebo relative to the nicotine lozenge condition). Specifically:
Each variable was subjected to repeated measures ANOVA with the between-subjects factor of GROUP (abstainers vs. continuing smokers) and the within-subjects factor of OCCASION (acute abstinence as assessed in the baseline placebo lozenge condition; 7 days; 1 month; 3 months). Thus whilst both groups were in the same 12-hour abstinent state at baseline, in subsequent assessments the continuing smokers were nicotine satiated having smoked immediately prior to the assessment session whilst the quit group were completely nicotine free. Since we were specifically interested in comparing the two groups’ patterns of change from one occasion to the next, a priori contrasts compared scores at adjacent time points (i.e. baseline vs. one week; one week vs. one month; one month vs. three months). Where relevant, post-hoc t-tests were also used to compare groups at specific time points, to supplement these contrasts.
Given the high rate of relapse in the quit group (see below) which begins to undermine the initial effects of randomization, we have conducted additional 2×4 ANOVAs for each variable based on a better-matched sub-group of successful abstainers and continuing smokers. Since pre-baseline salivary cotinine levels were shown to be the single most important clinical predictor of abstinent status in this sample (withdrawal symptoms, mood or dependence as assessed by the FTND were not significant predictors) , we have used this variable as a basis for matching. Thus a cut-off of 330 ng/ml was employed which resulted in the loss of three successful abstainers who were outliers at the high end, and eleven continuing smokers. These additional sub-analyses based on 29 quitters and 20 smokers matched for salivary cotinine level (quitters mean: 189.5; smokers mean: 179.5; t < 1, ns) are also reported in brief to supplement the main findings.
Of the 145 participants who completed the baseline assessments , 107 were randomised to quit; of these 33, (31%) maintained cotinine-verified abstinence at all follow-ups. Of the 38 participants allocated to continue smoking, 31 (82%) attended all follow-ups. All analyses are based on these two subgroups. Some data were missing for some variables, usually due to calibration or technical problems. Thus, sample sizes vary between analyses, from 26 to 33 abstainers and 25 to 31 continuing smokers.
Baseline (acute abstinence) scores for the 33 successful abstainers vs. the 31 continuing smokers were compared for all clinical and experimental variables via independent samples t-tests (and chi-square for sex ratio). The groups did not differ in sex ratio, age, years of education, number of cigarettes smoked per day, dependence (FTND), pre-baseline salivary cotinine levels, or any of the experimental variables (see Table 1).
Mean scores are shown in Figure 1.
There was a significant decline in symptom severity from baseline (acute abstinence) to one week (F1,61 = 8.0, p < 0.01), this being more pronounced in the continuing smokers (GROUP × OCCASION: F1,61 = 4.4, p < 0.05). From one week to one month, a significant GROUP × OCCASION interaction (F1,61 = 4.4, p < 0.05) reflects a continued improvement in the abstainers versus a slight worsening in the continuing smokers. Symptoms remained stable from one to three months (no main effect of OCCASION nor an interaction with GROUP: F1,61 < 1, ns, in both cases). The groups did not differ significantly in the severity of their symptoms at any individual follow-up (t62 < 1.5, ns, in all cases). The pattern of findings was unchanged when we repeated the analysis using the better matched sub-sample.
Subjective desire to smoke is shown in Figure 2.
The decline in scores from baseline to one week was significant (F1,62 = 12.5, p = 0.001) and similar in both groups (no OCCASION × GROUP interaction: F1,62 < 1, ns). A further significant decline from a week to a month (F1,62 = 19.7, p < 0.001) was qualified by a significant GROUP × OCCASION interaction (F1,62 = 4.1, p < 0.05) with abstainers’ scores reducing more. There was no overall further reduction from one to three months, nor any interaction with GROUP (F1,62 < 1, ns, in both cases).
Abstainers’ scores were lower than those of the continuing smokers at all follow-up points; this difference was just short of significance at one week (t62 = −1.95, p < 0.06) but highly significant at both one and three months (t62 = 4.2 and 5.2, p < 0.001). The pattern of findings was unchanged when we repeated the analysis using the better matched sub-sample.
Anxiety scores are shown in Figure 3.
ANOVA confirmed a significant reduction from baseline to one week (OCCASION: F1,61 = 5.7, p = 0.02) with no GROUP × OCCASION interaction (F1,61 < 1, ns). There was, however, a significant GROUP × OCCASION interaction from one week to one month (F1,61 = 4.9, p < 0.05) with abstainers showing a further reduction in anxiety but continuing smokers an increase. There was no significant change from one to three months (neither a main effect nor interaction: F1,61 < 1, ns).
Although the graph suggests that anxiety was higher in abstainers than continuing smokers at one week and vice versa at a month, the groups did not differ significantly at any follow-up point (t62 < 1, ns, in all cases).
These findings were essentially replicated in the better-matched sample, although the GROUP × OCCASION interaction from one week to one month now fell short of significance (F1,46 = 2.78, ns)
These scores are shown in Figure 4.
Mood improved significantly from baseline to one week (F1,62 = 7.7, p < 0.01) and this did not interact with GROUP (F1,62 < 1, ns).
A significant GROUP × OCCASION interaction from one week to one month (F1,62 = 6.5, p = 0.01), reflects an improvement in the abstainers but if anything a worsening in the continuing smokers. At this point, abstainers were significantly less depressed than continuing smokers (t62 = 2.1, p < 0.05). From one to three months there was neither a main effect of OCCASION nor a GROUP × OCCASION interaction (F1,62 < 1, ns, in both cases), though slight changes meant that they no longer differed significantly at this point (t62 < 1, ns).
Analyses using the better-matched sample of abstainers and continuing smokers again replicated these findings although the main effect of OCCASION from baseline to one week now fell short of significance (F1,47 = 2.17, ns)
Anhedonia scores are shown in Figure 5.
There was a significant improvement in hedonic tone (i.e. a reduction in SHAPS anhedonia scores) from baseline to one week (F1,59 = 4.9, p < 0.05), not qualified by an interaction with GROUP (F1,59 < 1, ns). There was no subsequent change (all main effects and interactions: F1,59 < 1, ns). At no time point did the two groups differ significantly from each other (t59 < 1, ns, in all cases).
Analysis of the smaller, better-matched sample revealed a similar pattern of results although the significant improvement from baseline to one week was now lost (F1,44 = 1.7, ns).
Reward responsivity increased significantly from baseline to one week (F1,59 = 3.87, p = 0.05), with no GROUP × OCCASION interaction (F1,59 < 1, ns). There were no further changes across consecutive points (F1,59 < 1, ns, for both main effects and interactions), and at no follow-up did the two groups differ from each other (t59 < 1, ns, in all cases; see Figure 6).
The same pattern of findings emerged in the better-matched sample but again, the improvement from baseline to one week now fell short of significance (F1,45 = 1.0, ns).
There was neither a significant main effect of OCCASION, nor an OCCASION × GROUP interaction, for the contrasts between any two consecutive occasions (F1,61 < 1, ns, in all cases; Table 2). The pattern of findings was unchanged when we repeated the analysis using the better matched sub-sample.
Accuracy fluctuated markedly from one occasion to the next (see Figure 7). An improvement from acute abstinence to one week was significant (F1,50 = 10.9, p < 0.01) and greater in the continuing smokers (OCCASION × GROUP: F1,50 = 6.1, p < 0.02). A subsequent decline in accuracy between one week and one month was also significant (F1,50 = 4.1, p < 0.05) and did not interact with GROUP (F1,50 < 1, ns). There was a near-significant improvement from one to three months (F1,50 = 3.4, p = 0.07), again not interacting with GROUP (F1,50 < 1, ns). At no individual follow-up did the groups differ significantly from each other (t50 < 1, ns, in each case). Given the complexity of this pattern, scores at baseline were contrasted directly with those at three months, revealing a significant GROUP × OCCASION interaction (F1,55 = 6.1, p < 0.05) in which continuing smokers had improved highly significantly (t27 = 4.0, p <0.0001) whilst abstainers had not (t28 < 1, ns).
Analyses of the smaller, better-matched groups again revealed a significant improvement from acute abstinence to one week (F1,36 = 4.0, p < 0.05) whilst all other significant main effects and interactions now fell short of significance (F1,36 < 2.5, ns in all cases). However, the critical interaction from baseline to 3 months remained significant (F1,36 = 4.11, p < 0.05).
There were no main effects of OCCASION nor any GROUP × OCCASION interactions in the full, (F1,50 < 1, ns, in all cases; mean scores given in Table 2) or the better-matched, sub-sample.
Consistent with previous reports [17-19], self-reported withdrawal symptoms, anxiety and depression steadily improved over one month, whilst in continuing smokers they remained relatively stable across three months. Anxiety and depression scores were well within the ‘normal’ (non-clinical) range for both groups even during acute abstinence (baseline).
Abstainers’ marked decline in desire to smoke contrasted with the stability shown by continuing smokers’ across the three follow-ups. This profile is consistent Gilbert et al.'s findings that craving dropped steadily over the first month of quitting, differing statistically from the scores of continuing smokers by days 16−22 . In the present study ratings of desire to smoke, mood and physical symptoms were made towards the end of the one-hour testing session; thus although continuing smokers had smoked immediately prior to the session they may have began to experience withdrawal by this point. Since participants reported smoking an average of 18.5 cigarettes per day (just over one per waking hour), their ratings within the session are likely to correspond to their experience at many points during a normal day.
Both abstainers’ and continuing smokers’ SHAPS anhedonia scores were lower at the one-week follow-up than during acute abstinence, and then remained stable to 3 months. The two groups did not differ at any point, and their mean scores fell within the normal range on all occasions. Pomerleau et al.  found self-reported anhedonia in ex-smokers (abstinent for 6 months or more) to be comparable with that in current smokers; both groups were more anhedonic than never-smokers. Thus although the present data suggest that any initial abstinence-related decline in hedonic tone rapidly recovers to the levels smokers achieve by smoking, in the absence of a matched group of non-smokers it is unclear to what extent it truly normalises.
Turning to reward motivation, whilst acute abstinence was characterised by lower baseline reward responsivity on the CARROT, this improved equally in the two groups by one week. Since performance in the continuing smokers was presumably enhanced by their recent nicotine use (consistent with our previous experimental findings [3-5]), the parallel improvement by the abstainers is therefore consistent with natural recovery. However practice/familiarity effects may also have contributed to the observed improvements in both groups.
With regard to attentional bias to pleasure-related words, there was little change in either abstainers or continuing smokers. In fact, there was no evidence that acute abstinence reduced attentional bias to pleasure versus neutral words in the subgroup of successful quitters described here, contrasting with our findings in the larger sample of 145 smokers . This may reflect the lower statistical power yielded by this smaller sample, or alternatively a relative lack of motivational blunting during acute abstinence in smokers who are able to maintain prolonged abstinence. However it is notable that this same subgroup did exhibit diminished reward responsivity on the CARROT during acute abstinence.
Finally, abstainers showed no improvement on two indices of response inhibition (antisaccade accuracy and motor errors on the continuous performance task [CPT). Although neither group showed a change in CPT errors over time, continuing smokers did show improved accuracy on the antisaccade task once they had resumed smoking reflecting our previous finding . Thus deficits in one index of inhibitory control – oculomotor response inhibition - do not appear to recover across three months of abstinence. These findings parallel Neuhaus et al's observation of abnormally low activation of frontal regions in long-term abstainers,  and are consistent with the existence of long-term deficits in response inhibition. These may reflect underlying neural dysfunctions which predate regular smoking or which are acquired during the course of regular smoking which fail to recover within the first few months of cessation. In either event, it could be an indicator of susceptibility to relapse, either because it correlates with a difficulty in inhibiting habitual smoking behaviour, or because smokers smoke in part to increase control over other their impulsive tendencies. Interventions to aid smoking cessation might therefore usefully focus on strategies for enhancing inhibitory control over the longer-term as well as in the initial stages of abstinence.
This study is not without its limitations. First, although an impressive one-third of smokers who were allocated to the quit group managed to abstain for 3 months, this nevertheless resulted in a fairly modest sample size of only 33 abstainers. Second, given that two-thirds of those allocated to the quit group subsequently relapsed, introduces the possibility that the self-selected abstainers showed reduced severity of impairment during acute abstinence. Nevertheless, we have shown in this sample that withdrawal symptoms, mood and dependence (on the FTND) did not significantly predict relapse whilst pre-quit salivary cotinine level did . That our supplementary sub-analysis of successful abstainers and continuing smokers matched for cotinine levels confirmed the results of the full sample, affords us greater confidence in these findings. Third, it is possible that this group of successful abstainers had better emotional regulation ability for genetic or other reasons, or experienced fewer environmental stressors during abstinence, both of which were not assessed here.
Overall, the findings tentatively suggest that appetitive processes and related affective states do improve over three months of abstinence in those smokers who are able to remain nicotine-free whilst response inhibition may not. Nevertheless, the sample size and findings remain modest and should be treated with caution until they can be replicated in a larger and better matched sample.
This research was funded by the National Institute of Drug Abuse (NIDA); grant code 3527427).