Subjective symptoms of nicotine withdrawal, as defined in the DSM-IV classification, include irritability, anxiety, restlessness, impatience, depressed mood, and difficulty concentrating. In this study, short-term overnight abstinence elicited significant subjective withdrawal symptoms in moderate smokers, but was not associated with significant changes in CBF in ventral striatum, frontal cortex, or the thalamus. Although there was no main effect of withdrawal on CBF, a significant correlation was observed between changes of blood flow in thalamus and subjective symptoms: the greater the reduction in CBF, the more severe were the subjective symptoms. While we cannot entirely exclude a mechanistic role for the ventral striatum, it may be more complex than can be gleaned from blood flow studies performed following a brief overnight abstinence. In contrast, the administration of nicotine (6 mg nicotine gum in this study) is a sufficiently potent pharmacological dose to produce a change in ventral striatal blood flow. Our findings are consistent with a known role of ventral striatum mediating acute effects of drug reward.
While the rewarding effects of abused drugs including nicotine clearly involve the mesolimbic dopamine system, the same understanding of brain systems mediating the aversive properties of withdrawal has not been established. There have been many neuroimaging studies of the effects of acute nicotine administration, but none that specifically investigated withdrawal. One difficulty is in defining the onset and duration of withdrawal. Researchers have observed a ‘peak’ in negative withdrawal symptoms after the first week of abstinence, with a return to baseline by the fourth week. West et al (1989)
found this time frame to be consistent across 147 subjects, but there may also be a ‘late’ withdrawal at approximately 8 weeks (Piasecki et al, 1998
). We chose to investigate acute overnight withdrawal to afford consistency across subjects, maximize compliance, optimize image registration in a naturalistic setting, and based on studies showing that blood nicotine levels reach a trough the morning after overnight abstinence, corresponding to the greatest pharmacodynamic and subjective responses associated with the first few cigarettes of the day (Benowitz, 1988
). Among smokers, whether the favorite cigarette of the day is the ‘first morning’ or ‘after dinner’ depends on the degree of nicotine dependence and pattern of smoking. Infrequent ‘chippers’ preferred the after-dinner whereas the most nicotine dependent smokers preferred the first morning cigarette (Jarvik et al, 1993
). Inconsistent behavioral or physiological results may be related to either the type of smoker or the high variability in the temporal profile of withdrawal. Although nicotine withdrawal can be severe, in most cases it is not. It has been suggested that smokers who experience chronic, low-grade withdrawal are more likely to relapse than those whose symptoms abate quickly (Hughes, 2001
), underscoring the important role that duration plays in predicting relapse. Few studies have examined physiological correlates of intermediate to long-term withdrawal. Gilbert et al (2004)
compared EEG recordings before and after smoking cessation in individuals randomized to quit or not quit. There was more low-frequency EEG activity in the quitters compared to non-quitters that began immediately (day 3) and persisted for 31 days after quitting. The cognitive correlate of this low-frequency activity is a generally reduced arousal state. Since thalamic and reticular nuclei rich in cholinergic activity are involved in general arousal, it is interesting to note that in our study changes in thalamic CBF correlated significantly with changes in subjective withdrawal (ρ = 0.8, p
Our finding of an increase in CBF in the ventral striatum after nicotine administration is inconsistent with other reports. Previous CBF studies have used [15
O-PET to measure CBF following smoking or nasal spray in overnight abstinent smokers. Zubieta et al (2005)
found relative decreases in CBF in nucleus accumbens in nicotinized compared with de-nicotinized cigarettes. Rose et al (2003c)
used a principal component analysis to define three factors (frontal, striatal, and reticular systems). In their study, nicotine had no effect on blood flow in the striatal system (defined as caudate, putamen, and posterior cingulate) and reduced flow on the frontal system (ventral striatum, prefrontal cortex, and anterior cingulate). While the grouping of different brain regions to define a ‘factor’ may uncover important brain circuits, it makes it very difficult to compare to other studies including ours. A recent study showing dopamine is released in ventral striatum in nicotine-dependent persons who smoked compared to those who did not, nonetheless, provides evidence of ventral striatal activity (Brody et al, 2004
). Possible explanations for differences between previous PET studies and our MR study include differences in the route of nicotine administration, the use of individual vs
multi-subject group averaging, ROI compared with voxel-wise analysis, and differences in spatial resolution leading to different partial volume averaging effects.
Previous studies also reported an increase in thalamic blood flow and metabolism following nicotine (Zubieta et al, 2005
; Domino et al, 2000b
), a finding that we did not replicate. It is notable that Rose et al (2003b)
demonstrated a complex, inverted-U-shaped, dose-dependent, hemisphere-dependent, response to nicotine in the reticular factor to which the thalamus contributed. Thus, in this well-designed, constrained study, the regional blood flow response to nicotine were demonstrated to be rather complex, dependent on dose, hemisphere, and nonlinear.
The strengths of this study are the within-subject design and use of a novel, relatively non-invasive method to measure CBF using DSC imaging. The ventral striatum which is considered the center of the reward pathway consists of a medial shell and central core with limited afferent input to these subregions. The borders are difficult to delineate, even at a histochemical level (Haber and McFarland, 1999
). Given its small volume and anatomic uncertainties, we believe that an ROI approach to ventral striatum, as used in this study, affords the greatest accuracy. A disadvantage of using ROIs is that it is possible to miss changes in other brain regions. If one were primarily interested in larger cortical areas, then whole brain whole brain approaches such as volumetric brain morphometry (VBM) could be considered. We deliberately chose to maximize spatial resolution of our CBF maps on the targeted region, which necessitated sacrificing full-head coverage. Given the incomplete brain coverage and required spatial smoothing, VBM would not be an option in this study. Limitations of this study are the small sample size and lack of placebo control for nicotine. We had adequate power to detect approximately 10% difference in CBF. Although our primary intent was to study withdrawal from nicotine, the finding of increased ventral striatum CBF after nicotine gum will have to be confirmed with future placebo-controlled studies of larger sample size. Another limitation is the lack of nicotine blood levels. For each subject, abstinence and the withdrawal syndrome was confirmed with self-report, reduction in exhaled CO, and withdrawal measurements. While less nicotine is absorbed by chewing gum as compared to smoking, significant differences in heart rate, blood pressure, and alleviation of withdrawal symptoms after nicotine gum suggest that levels were sufficient to effect both physiological and subjective changes.
In summary, this is the first paper that specifically investigated whether subjective nicotine-withdrawal symptoms were associated with CBF changes in neural structures known to be involved positive reinforcement. The time over which withdrawal symptoms emerge, even when as brief as overnight, require absolute measurements which cannot be done using blood-oxygen-level-dependent fMRI techniques. Further, methods that use no radioactivity would be preferable to those involving radioisotope since measurements were repeated four times. Using MR DSC methods, we demonstrated a small, significant effect of nicotine on striatal CBF, but failed to find a main effect of withdrawal. There was correlation between changes in withdrawal and changes in thalamic CBF, however, suggesting that the thalamus may play a more complex if subtle role in the withdrawal state.