Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Cigarette smoking and gray matter brain volumes in middle age adults: the CARDIA Brain MRI sub-study


Cigarette smoking has been associated with dementia and dementia-related brain changes, notably gray matter (GM) volume atrophy. These associations are thought to reflect the co-morbidity of smoking and vascular, respiratory, and substance use/psychological conditions. However, the extent and localization of the smoking-GM relationship and the degree to which vascular, respiratory, and substance use/psychological factors influence this relationship remain unclear. In the Coronary Artery Risk Development in Young Adults CARDIA cohort (n = 698; 52% women; 40% black participants; age = 50.3 (SD = 3.5)), we examined the associations of smoking status with total GM volume and GM volume of brain regions linked to neurocognitive and addiction disorders. Linear regression models were used to adjust for vascular, respiratory, and substance use/psychological factors and to examine whether they modify the smoking-GM relationship. Compared to never-smokers, current smokers had smaller total GM volume (−8.86 cm3 (95%CI = −13.44, −4.29). Adjustment for substance use/psychological – but not vascular or respiratory – factors substantially attenuated this association (coefficients = −5.54 (95% CI = −10.32, −0.76); −8.33 (95% CI = −12.94, −3.72); −7.69 (95% CI = −6.95, −4.21), respectively). There was an interaction between smoking and alcohol use such that among alcohol non-users, smoking was not related to GM volumes and among alcohol users, those who currently smoked had −12 cm3 smaller total GM, specifically in the frontal and temporal lobes, amygdala, cingulate, and insula. Results suggest a large-magnitude association between smoking and smaller GM volume at middle age, accounting for vascular, respiratory, and substance use/psychological factors, and that the association was strongest in alcohol users. Regions suggested to be most vulnerable are those where cognition and addiction processes overlap.


Cigarette smoking has been associated with higher risk of cognitive decline and dementia in older adults1,2,3,4. Although the mechanisms are not fully understood, these associations are predominantly thought to arise from accumulated smoking-related damage to cardiovascular and respiratory processes in the brain1,2,5,6. Parallel to its established impact on vascular and respiratory functions, tobacco dependence is the most prevalent substance-dependence disorder and frequently co-occurs with psychiatric conditions such as mood and substance use disorders7,8,9. For instance, among persons with alcohol abuse disorder, smoking prevalence has been estimated to be as high as 90%10,11.

In line with its documented effects on cardiovascular health, some imaging studies have reported that smoking is related to compromised vascular and circulation processes in the brain, including changes in perfusion and white matter (WM) lesions4,12,13,14. One consistent finding from imaging studies is that smokers have smaller total brain volume4,12,15, and smaller gray matter (GM) volume in particular, compared to never-smokers6,16. Interestingly, several results show that this association may be differentially localized16,17, suggesting that more direct links between smoking and specific brain structures could be underlying the associations of smoking to functional outcomes such as addiction and cognitive functioning.

However, findings on which brain structures may preferentially be affected by smoking are heterogeneous. The discrepancy could be explained by methodological limitations. One limitation of published studies is that they were mostly conducted in small (often fewer than 100 subjects)15,16 and selective samples (based on participants with several morbidities)18,19. Further, studies in community-dwelling individuals have focused mainly on older populations and thus may have had limited ability to disentangle the primary brain changes associated with smoking from those resulting from general accumulated damage such as advanced forms of cardiovascular disease4,5,20. Another main limitation is that prior studies included limited adjustment for important contributing factors, such as cardiovascular, respiratory, and substance use/psychological factors17. Given their high co-occurrence and complex relationships with smoking and GM, the role of these factors warrants careful investigation1,2,5,11,21.

In this study, we used a community-based sample of middle-aged adults to assess the associations of smoking with total brain GM volume and with volumes of candidate brain regions linked to cognitive impairment and addictive behaviors, while assessing the role of vascular, respiratory, and substance use/psychological factors. We first adjusted for these factors and then tested whether they modified the smoking-GM relationship. We hypothesized that GM volumes would be smaller in smokers, particularly in smokers with additional vascular risk factors and addiction-related behaviors.

Materials and methods

Study sample

The Coronary Artery Risk Development in Young Adults (CARDIA) study is a population-based longitudinal study of the determinants of cardiovascular disease in young black and white adults22. At baseline (1985–1986), the study recruited 5115 participants aged 18–30 years randomly selected by telephone numbers from census tracts in four US cities (Birmingham, Alabama; Chicago, Illinois; Minneapolis, Minnesota; and Oakland, California); the sample was balanced by race (black, white), sex, and education (high school degree or less, higher than high school). Participants were followed for 25 years (with examinations at year 2, 5, 7, 10, 15, 20, and 25 of follow-up); 72% (n = 3499) of the surviving baseline cohort participated in the 25th year examination.

At the 25th year examination, a sub-sample took part in the CARDIA brain magnetic resonance imaging (MRI) sub-study. That sub-study enrolled participants from three CARDIA field centers (Birmingham, Minneapolis, and Oakland) and aimed to achieve a balanced sample within sex-race groups. Exclusion criteria were a contra-indication to MRI or a body size too large for the MRI tube bore. In total, 719 subjects participated in the MRI sub-study.

CARDIA is annually approved by the institutional review boards at each site (the Institutional Review Board of the University of Minnesota, the Kaiser Permanente Northern California Institutional Review Board, University of Pennsylvania Institutional Review Board, The University of Alabama Birmingham Institutional Review Board, and the NIH Office of Human Subjects Research Protection for the Intramural Research Program, National Institute on Aging); separate approval was obtained for the brain MRI study. Each participant provided written informed consent at each exam12.

Brain MRI data

Brain MRIs were acquired using 3-T MR scanners (Oakland: Siemens 3 T Tim Trio/VB 15 platform; Minnesota: Siemens 3 T Tim Trio/VB 15 platform; and Alabama: Philips 3 T Achieva/ platform). In collaboration with the MRI reading center at the University of Pennsylvania (Section of Biomedical Image Analysis, Department of Radiology), each MRI field center performed the MRI scans and transferred the data to the reading center, according to standardized procedures and following standard quality assurance protocols, previously described12. Brain tissue volumes were estimated from the sagittal 3D T1 sequence (Plane Sagittal Coil 12channel File name 3D T1 MPRAGE: Tr 1900 Te 2.89 Fov 250 mm, thickness 1 mm slices 176 slices, Base Res 255, Phase res 100%, Matrix 256 × 256 NSA 1 TI 900 ms Pixel BW 170hz. ETL = 1 Flip = 9).

Image processing, quality control checks, and automated brain tissue volume computations were all performed at the MRI reading center following standardized protocols12,23,24,25,26. First, all imaging data were converted to Neuroimaging Informatics Technology Initiative (NIFTI) format. Initial quality control (QC) was performed through manual inspection of each subject’s T1 scan for motion artifacts and other quality issues. The preprocessing of the T1-weighted scan involved brain extraction (i.e., the removal of the skull, extra-cerebral tissues and cerebellum) using a multi-atlas segmentation method27, correction of image inhomogeneities28, and segmentation of the brain parenchyma into GM, WM and cerebrospinal fluid29. Brain tissue was segmented into anatomical regions by transferring expert defined region of interest (ROI) labels on a standard template image space to the subject T1 space through non-linear registration23. GM and WM were classified into regions of interest according to the Jakob atlas30 and further into normal and abnormal and tissue. Within each ROI, GM and WM volumetric measurements are calculated. Total brain volume was obtained by summing GM and WM volumes; total intracranial volume was the sum of GM, WM, and cerebral spinal fluid volumes.

Here, we assessed the associations of smoking with total GM volume, and as a comparison with total brain tissue and total WM volume. We examined associations with lobar GM to assess whether smoking had a global compared to specific association with GM. We also assessed the relationship of smoking with GM volumes in specific brain structures. These regions have been linked to dementia and cognitive impairment and to addiction disorders and have been implicated in memory, learning and cognition as well as emotion and reward processes: hippocampus (memory and emotion regulation), parahippocampal gyrus (memory encoding and recall), entorhinal cortex (memory), precuneus (memory and cognition), cuneus (visual processing and inhibitory control), thalamus (perception, sensory, and motor processes), caudate (motor, learning, and reward processes), putamen (movement and learning), nucleus accumbens (reward), cingulate (emotion and behavior regulation), insula (emotional experience and perception), and the amygdala (affective processing and emotional memory)21,31,32,33,34,35.

Smoking data

Smoking data were collected with the same interviewer-based questionnaire administered at each study examination. Participants were asked if they have ever smoked regularly (defined as at least 5 cigarettes per week, almost every week) for at least three months. If they answered no, they were classified as never-smoker at that study wave. If they answered yes, participants were then asked whether they currently smoked regularly (at least five cigarettes per week, almost every week) for at least three months; participants were classified as current smokers if they answered yes, and as former smokers if they reported no current regular cigarette smoking.

Mid-life smoking status was based on self-reported status (never, former, current) at year 25. This self-report was verified against prior data from earlier study exams; for most participants, self-reported status was consistent across study waves, and for 57 (8.7%) subjects who reported being never-smokers at Year 25, we re-coded them as former-smokers because they had reported being smokers or former-smokers at multiple previous study exams.


All analyses were adjusted for age, race (black/white), educational attainment (≤high school diploma, >high school diploma), total intracranial volume, and study center. We then incorporated data on several factors combined into three groups, representing three mechanisms and pathways that are hypothesized to act as confounders or mediators of the smoking-GM volume relationship: vascular, respiratory, and substance use/psychological factors.

Vascular risk factors included body mass index, history of hypertension (defined as current or prior occurrence of: diastolic blood pressure ≥90 and/or systolic blood pressure ≥ 140mmHg and/or taking antihypertensive medication), cerebro/vascular events (yes/no variable indicating self-reported occurrence of any of the following: heart attack, angina, heart failure, rheumatic heart disease, mitral valve prolapse, stroke and peripheral vascular disease), history of diabetes (defined following ADA criteria for levels of fasting, non-fasting or postprandial OGTT results, HbA1c percent, or use of anti-diabetes medication), and hypercholesterolemia (total cholesterol ≥ 240 mg/DL and/or use of cholesterol lowering medication).

Respiratory factors were history of respiratory illness (defined as current or prior self-reported occurrence of asthma, tuberculosis, chronic bronchitis, chronic obstructive pulmonary disease, emphysema, or use of medication for asthma or for breathing problems), and forced expiratory volume (FEV1 estimated at 1 sec, in liters, and measured using a spirometer).

Substance use/psychological factors investigated were self-reported lifetime illicit drug use (yes/no for having ever-used any of the following: marijuana, amphetamines, methamphetamines, cocaine, crack, heroin, or pain-killers for non-medical reasons), depressive symptoms (Center for Epidemiologic Studies – Depression; CES-D) scale36, and alcohol consumption, which was based on questions regarding the usual number of alcoholic beverages (wine, beer, or hard liquor) consumed per week; assuming that one drink of beer, wine, or liquor contains 16.7, 17.0, and 19.1 milliliters (mL) of ethanol respectively (as per CARDIA protocol), we computed total consumption per week in mL of ethanol and divided it by 17.24 mL of ethanol per average drink to estimate the usual number of drinks per week. Accordingly, alcohol consumption was classified as no alcohol use (no consumption of alcohol beverages), high-risk alcohol use (based on the sex-specific weekly maximum drinking limits published by the National Institute on Alcohol Abuse and Alcoholism (for men > 14, women > 7 standard drinks/week)), and low-risk alcohol use (>0 and ≤7 drinks for women and >0 and ≤14 drinks for men).

FEV1 measures were collected at the year 20 exam, the study visit prior to the MRI. For all other covariates, we used their values at the time of the MRI (year 25 exam).

In sensitivity analyses, we also included adjustment for cancer history (current or previous occurrence of cancer and tumor) and physical activity (measured at year 25 using physical activity scores expressed in exercise units, calculated based on the reported frequency and intensity of participation in 13 categories of moderate and vigorous recreational sports, exercise, leisure, and occupational activities over the prior 12 months)37. Since results did not change with these adjustments, data are not shown.

Statistical analyses

Of the 719 subjects with MRI data, 698 had complete data on smoking and potential covariates, and 643 had data on FEV1. Total and regional GM volumes were normally distributed. Multivariable linear regression models were used to examine their relationships with mid-life smoking status at the time of the MRI. In all analyses, we adjusted for basic covariates (age, sex, race, educational attainment, intracranial volume, and study centers). We systematically examined the role of other potentially important covariates by using three models, one adjusting for vascular factors, one for respiratory factors, and one for substance use/psychological factors. Then, we tested, using interaction terms, whether these factors modified the associations of smoking and GM volumes and conducted stratified analyses to examine significant interactions. We primarily focused on other vascular risk factors— which often co-occur with smoking and might influence smoking behavior and its outcomes6,12,38— and on alcohol use, given its high co-morbidity with smoking and emerging findings regarding their interactions11,19,39.

Given the high prevalence of self-reported ever-use of marijuana (Supplementary Table S1), we assessed in sensitivity analyses, whether there was a change in results after adjusting for or testing for effect modification by the lifetime use (yes/no) of marijuana and by the lifetime extent of marijuana use (the number of times participants reported using marijuana up to study year 25 (classified into ≤2, 3 to 10, 11 to 99, 100 to 499, and ≥500 times)); we saw no significant difference in the conclusions.

In additional analyses, we evaluated the relationship of indicators of smoking history (cumulative pack-years, age at start of smoking, and years of smoking cessation) and GM outcomes among ever-smokers.

All analyses tested two-tailed hypotheses with a significance level of 0.05. Analyses were completed using SAS version 9.3 (SAS Institute, Cary, NC). Data and statistical codes are available from the CARDIA Coordinating Center:


Compared to never and former smokers, current smokers were younger and had the highest proportion of black participants, the lowest educational attainment, and the highest prevalence of hypertension (Table 1). Other cardiovascular risk factors were not different across smoking groups (Table 1). Current smokers had the highest frequency of respiratory illnesses and the lowest FEV1. All substance use/psychological measures were more frequent in smokers (Table 1): compared to never-smokers, current and former-smokers had higher prevalence of high-risk alcohol drinking and illicit drug use; current smokers had more depressive symptoms.

Table 1 Characteristics of the CARDIA-MRI sample by smoking status at Year 25

Smoking and GM volume

Brain volumes are presented in Supplementary Table S2. Compared to never-smokers, current smokers had significantly smaller total GM volume (−8.86 cm3 (95% CI = −13.44, −4.29); Table 2). There were no differences in total white matter volume across smoking groups; the trend of smaller total brain volume in smokers reflected the association between smoking and smaller GM (Table 2).

Table 2 Smoking status and global brain volumes in the CARDIA brain MRI cohort

Adjustment for vascular and respiratory factors

Adjusting for vascular or respiratory factors slightly attenuated the associations of smoking and total GM volume (with a 6 to 13% change in the smoking-GM volume estimated association; Table 2). There was no indication of interaction between smoking and vascular or respiratory factors in the smoking-GM relationship.

Adjustment for substance use/psychological factors

The smoking-total GM association was substantially attenuated (from −8.86 to −5.54 cm3; 37% change in estimate) when the model included substance use/psychological factors (Table 2).

Similar patterns were found in the lobar (Supplementary Table S3) and candidate region analyses (Supplementary Table S4): Current smokers had significantly smaller GM volume in the temporal, occipital, and frontal lobes, and in the amygdala, entorhinal cortex, cingulate, and insula, compared to never-smokers; these associations were largely attenuated when substance use/psychological factors were taken into account. Former-smokers had a trend for smaller GM volume but they were not significantly different than never-smokers.

Other candidate regions (hippocampus, parahippocampal gyrus, cuneus, precuneus, thalamus, caudate, putamen, and the nucleus accumbens) were not related to smoking status (Supplementary Table S4).

Interactions between substance use/psychological factors, smoking, and brain volumes

In the substance use/psychological factors model, all factors showed associations with GM volume: high-risk alcohol drinking (coefficient = −5.58 cm3 (95% CI = −10.57, −1.14)) and higher CES-D scores (−0.24 cm3 (95% CI = −0.46, −0.02) per unit increase in CES-D score) were associated with significantly smaller total GM; illicit drug use was associated with smaller GM (coefficient = −3.69; p = 0.07).

Illicit drug use and depressive symptoms did not modify the association between smoking and GM volumes (p > 0.3 for interaction terms). There was a significant interaction between alcohol use and smoking indicating that the association is differentially observed in alcohol users (p for interaction of current smoking with high-risk alcohol use = 0.03 and p = 0.07 for interaction with low-risk alcohol use). Details regarding the distribution of sample’s characteristics across alcohol use sub-groups are presented in Supplementary Table S5.

Alcohol use, smoking, and brain volumes

Results of GM volumes across smoking and alcohol use groups (Table 3) show that subjects who don’t use alcohol (irrespective of smoking status) and never-smokers (whether alcohol users or not) do not have different GM volumes (with adjusted mean GM volumes ranging from 516 to 520 cm3). However, the combination of current alcohol (both low- and high-risk use) and smoking use was negatively associated with GM volume (Table 3); therefore, we grouped participants into alcohol non-users vs. subjects reporting any alcohol use (low-risk and high-risk). Among participants who do not use alcohol, smoking was not associated with total or regional GM volumes (Table 4). In contrast, among alcohol users (light and heavy-risk combined), those who currently smoked had −12 cm3 smaller total GM volume, and specifically smaller frontal, temporal, amygdala, cingulate, and insula GM volumes, even after accounting for other substance use /psychological, vascular, and respiratory factors (Table 4). There was a trend toward an association of smoking to smaller occipital GM volume (p = 0.11). Similar to analyses in the total sample, volumes of other candidate structures of interest were not related to smoking in the stratified analyses (p > 0.17; data not shown).

Table 3 Brain gray matter volume across alcohol and smoking groups in the CARDIA brain MRI cohort
Table 4 Smoking status and total and regional gray matter volumes, stratified by alcohol use, in the CARDIA brain MRI cohort

In additional analyses (Supplementary Table S6) investigating the associations of GM volumes to smoking characteristics (cumulative pack-years, age at start of smoking, and years since smoking cessation), the interaction between alcohol use and smoking characteristics were weaker than for smoking status (p-values for interaction terms > 0.09); however, in general, associations between smoking characteristics and GM outcomes were more pronounced in alcohol users. Higher cumulative pack-years and younger age at smoking initiation were associated with smaller total GM (p = 0.11 and p = 0.07, respectively); associations with age at smoking initiation were more pronounced for amygdala and temporal GM volumes (p = 0.06 and 0.02, respectively).


In community-dwelling middle-aged adults, current smoking was associated with smaller total and regional GM volumes. Importantly, results suggest that substance use/psychological —but not vascular or respiratory— factors substantially influenced these associations. In particular, the association of smoking and GM volumes was different in alcohol users and non-users: subjects who did not use alcohol showed no association between smoking and GM volumes, whereas among alcohol users, those who smoked had significantly smaller total and regional GM volumes compared to never-smokers. Regionally, this association was observed in the temporal and frontal lobes and in the amygdala, cingulate, and insula.

In line with other reports mainly in smaller samples40, we found that current smoking was associated with smaller total GM volume but not WM volume6,15,16. This association was specifically observed in alcohol users and the difference in GM volume of −12 cm3 in current smokers compared to never-smokers in this sub-group was large: roughly equivalent to a decade difference in age in our sample (age was associated with 1.2 cm3 smaller volume per year). Results indicate the importance of co-occurring smoking and alcohol use with respect to GM volumes and identify low-risk and high-risk alcohol users in a community-based sample as an “at-risk” subgroup for the potential negative neurocognitive sequelae of smoking.

It has been hypothesized and shown in older cohorts that vascular and respiratory factors may play an important role in the relationship between cigarette smoking and GM5. But in this younger cohort, these factors were not strongly related to concurrent smoking status and did not significantly modify the smoking-GM association. Adjusting for these factors only slightly attenuated the estimates of the smoking-GM association, though greater attenuation was observed in the model adjusted for respiratory factors, notably FEV1; this might be expected, as FEV1 is likely to be a mediating factor of the actions of smoking5. Our results suggest that, at middle-age, substance use/psychological factors are more important for how smoking relates to GM volume. Adding each of depressive symptoms, alcohol use, or illicit drug use alone slightly attenuated the smoking-GM relationship (to −7 or −8 cm3), but together these correlated factors (as seen in Table 1 and Table S5) had a substantial influence on the relationship (Table 2). Importantly, the association of smoking with GM varied depending on alcohol use, accounting for depressive symptoms and illicit drug use (Tables 3 and 4). As was found for smoking, there was no evidence for an interaction between alcohol use and illicit drug use or depressive symptoms with regards to GM volume (data not shown) suggesting that the alcohol-smoking interplay might be distinctively noticeable and important for GM volume.

Findings from studies on alcohol use disorder have shown an intricate interplay between smoking and alcohol use. Smoking has been linked to the severity of alcohol abuse and to a lower success rate of recovery interventions11,41. Alcohol use disorder has been independently associated with several negative functional and structural brain outcomes, including higher risk of dementia, poorer cognitive function, metabolic abnormalities, and smaller global and regional GM volumes11,19,41. Smoking has been found to aggravate some of these outcomes and to suppress metabolic, neuronal, and functional recovery of abstinence interventions11,41. Our results complement these findings in adults with alcohol abuse disorder, and show that both low- and high-risk alcohol users in a community-based sample presented adverse associations between smoking and GM volumes.

Regionally, associations were observed with smaller frontal and temporal GM volume in our sample, a finding consistent with other studies of smoking in both younger and older samples5,16,17,42, and with studies of other substance use disorders11,21,43,44,45. Smoking was also associated with significantly smaller amygdala, cingulate, and insula volumes. Although the exact mechanisms are not clear, functional and structural changes in the insula, cingulate, and amygdala have been widely documented in cigarette smoking and other addictions16,17,35,46,47.We found no associations of smoking with brain structures that have been widely linked to dementia (hippocampus, parahippocampal gyrus, entorhinal cortex, cuneus, and precuneus). There was an association between smoking and smaller entorhinal cortex; but it was not significant after adjustment for substance use/psychological or respiratory factors. Overall, our regional analyses suggest that, at mid-life, smoking seems to be related to emotion and reward/ addiction-related candidate regions and regions at the intersection of cognitive and reward pathways, such as insula, the cingulate, and the frontal and temporal lobes. The frontal, cingulate, and insular regions contain some of the highest densities of nicotinic acetylcholine receptors48. Data from animal models suggest that nicotine can have neurotoxic consequences such as cell loss and synaptic alterations49,50,51. Cigarette smoking also involves exposure to several toxic compounds, including carbon monoxide, free radicals, and free oxygen species, which can lead to cellular and oxidative damage directly affecting neuronal and brain tissue11 and to compromised inflammatory, respiratory, and vascular systems resulting in deficits in blood oxygen and nutrients to the brain11,52. Alcohol use has also been linked to oxidative damage and molecular and cellular impairments11; combined smoking and alcohol use could thus make the brain more vulnerable to each substances’ adverse consequences. It is interesting that we found an interaction with alcohol use and not with other substance use. This might be explained by a larger overlap in the involvement of alcohol and smoking with dopaminergic and γ-aminobutryic acid (GABA) and glutamate systems in several brain regions including the thalamus, cingulate, insula, amygdala, and frontal and temporal lobes53,54. Additionally, unlike illicit drug use, alcohol use is more frequent and legally and socially accepted, which could lead to different profiles of addiction and users. Further, there could be more potential stigma related to illicit drug use that can influence the accuracy of self-reported use; the information on the extent of use was also limited in our study and might not have well captured this behavior. Finally, with the wider acceptability and availability of alcohol, the interaction observed with alcohol use could also be indicative of more contexts and scenarios facilitating a higher cigarette smoking and exposure.

Combined, results of the risk factors models and the regional GM analyses suggest that behavioral and substance-use mechanisms may be central to smoking behavior at younger age and its relationships to GM outcomes at mid-life. These data are in line with results from recent epidemiological and clinical studies indicating that dual concurrent treatment for smoking and alcohol/substance abuse is more effective for long-term abstinence outcomes than treating each addiction separately11,41,55. Importantly, our results suggest that neurobiological links between co-occurring smoking and substance use behaviors with reward-network regions may be underlying the observed synergy between these behaviors in determining the severity of addictive symptoms and relapse.

Given the high prevalence of smoking and alcohol use, their frequent co-occurrence, and the fact that they are modifiable behaviors, our results emphasize the research, clinical, and public health value of further understanding the interplay between these two factors and how it affects brain structures and functions across the life-course and brain aging in late-life.

One limitation of our study is that we cannot exclude the possibility that some brain differences observed between smokers and never-smokers might have preceded and even predisposed to nicotine addiction, due to factors such as early developmental processes related to maternal smoking, a family history of addiction, or exposure to stressors. Our study focused on cigarette smoking and use of other forms of nicotine use was limited (prevalence of cigar, pipe, smokeless tobacco use in our sample was ≤4.3%) or not measured (e.g., e-cigarette and water pipe use) in our sample. The small number of subjects with certain vascular risk factors might have limited the detection of interaction with smoking. Similarly, the low prevalence of the use of specific illicit drugs (other than marijuana use) and the lack of information on actual quantity and duration of use limited our conclusions regarding the role of illicit drug use and of different types of illicit drug use in the relationship of smoking with brain structures. We examined several regions and there is the possibility of false positives. Applying a stringent Bonferroni correction for the 16 regional volumes investigated in the two alcohol subgroups (significance level = 0.0016), the associations with temporal, cingulate, and insula GM remain statistically significant. Overall, the relationships observed were strong and consistent, which supports the overall message that these findings merit further discussion and certainly replication.

In conclusion, in community-dwelling middle-aged adults, current smoking was associated with smaller total and regional GM volumes. These associations were specifically observed in alcohol users, and were localized in the temporal, frontal, insula, and amygdala volumes. Findings suggest important connections between smoking and behavioral co-morbidities and brain regions linked to addiction processes at mid-life. Future research on smoking and brain health, as well as interventions to address smoking and alcohol use, should prioritize subjects who present with both cigarette smoking and alcohol use, as they appear to be at even greater risk for adverse brain outcomes.


  1. 1.

    Anstey, K. J., von Sanden, C., Salim, A. & O’Kearney, R. Smoking as a risk factor for dementia and cognitive decline: a meta-analysis of prospective studies. Am. J. Epidemiol. 166, 367–378 (2007).

    Article  Google Scholar 

  2. 2.

    Brayne, C. Smoking and the brain. BMJ 320, 1087–1088pp (2000).

    CAS  Article  Google Scholar 

  3. 3.

    Weuve, J. et al. Accounting for bias due to selective attrition: the example of smoking and cognitive decline. Epidemiology 23, 119–128 110 (2012). 1097/EDE.1090b1013e318230e318861.

    Article  Google Scholar 

  4. 4.

    Debette, S. et al. Midlife vascular risk factor exposure accelerates structural brain aging and cognitive decline. Neurology 77, 461–468 (2011).

    CAS  Article  Google Scholar 

  5. 5.

    Karama, S. et al. Cigarette smoking and thinning of the brain/‘s cortex. Mol. Psychiatry 20, 778–785 (2015).

    CAS  Article  Google Scholar 

  6. 6.

    Friedman, J. I. et al. Brain imaging changes associated with risk factors for cardiovascular and cerebrovascular disease in asymptomatic patients. JACC Cardiovasc Imaging 7, 1039–1053 (2014).

    Article  Google Scholar 

  7. 7.

    Luger, T. M., Suls, J. & Vander Weg, M. W. How robust is the association between smoking and depression in adults? A meta-analysis using linear mixed-effects models. Addict. Behav. 39, 1418–1429 (2014).

    Article  Google Scholar 

  8. 8.

    Holma, I. A. K., Holma, K. M., Melartin, T. K., Ketokivi, M. & Isometsä, E. T. Depression and smoking: a 5-year prospective study of patients with major depressive disorder. Depress Anxiety 30, 580–588 (2013).

    Article  Google Scholar 

  9. 9.

    Baek, J. H., Eisner, L. R. & Nierenberg, A. A. Smoking and suicidality in subjects with major depressive disorder: results from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). J. Affect Disord. 150, 1158–1166 (2013).

    Article  Google Scholar 

  10. 10.

    Room, R. Smoking and drinking as complementary behaviours. Biomed. Pharmacother. 58, 111–115 (2004).

    Article  Google Scholar 

  11. 11.

    Durazzo, T. C. & Meyerhoff, D. J. Neurobiological and neurocognitive effects of chronic cigarette smoking and alcoholism. Front. Biosci. 12, 4079–4100 (2007).

    CAS  Article  Google Scholar 

  12. 12.

    Launer, L. J. et al. Vascular factors and multiple measures of early brain health: CARDIA Brain MRI Study. PLoS ONE 10, e0122138 (2015).

    Article  Google Scholar 

  13. 13.

    Rose, JedE. et al. PET studies of the influences of nicotine on neural systems in cigarette smokers. Am. J. Psychiatry 160, 323–333 (2003).

    Article  Google Scholar 

  14. 14.

    Durazzo, T. C., Meyerhoff, D. J. & Murray, D. E. Comparison of regional brain perfusion levels in chronically smoking and non-smoking adults. Int. J. Environ. Res. Public. Health 12, 8198–8213 (2015).

    CAS  Article  Google Scholar 

  15. 15.

    Gallinat, J. et al. Smoking and structural brain deficits: a volumetric MR investigation. Eur. J. Neurosci. 24, 1744–1750 (2006).

    Article  Google Scholar 

  16. 16.

    Fritz, H.-C. et al. Current smoking and reduced gray matter volume: a voxel-based morphometry study. Neuropsychopharmacology 39, 2594–2600 (2014).

    Article  Google Scholar 

  17. 17.

    Zhong, J. et al. Voxelwise meta-analysis of gray matter anomalies in chronic cigarette smokers. Behav. Brain. Res. 311, 39–45 (2016).

    Article  Google Scholar 

  18. 18.

    Jørgensen, K. N. et al. Cigarette smoking is associated with thinner cingulate and insular cortices in patients with severe mental illness. J. Psychiatry Neurosci. 40, 241–249 (2015).

    Article  Google Scholar 

  19. 19.

    Durazzo, T. C. et al. Interactive effects of chronic cigarette smoking and age on brain volumes in controls and alcohol-dependent individuals in early abstinence. Addict. Biol. 19, 132–143 (2014).

    Article  Google Scholar 

  20. 20.

    Power, M. C. et al. Smoking and white matter hyperintensity progression: The ARIC-MRI Study. Neurology 84, 841–848 (2015).

    CAS  Article  Google Scholar 

  21. 21.

    Martin-Soelch, C. Neuroadaptive changes associated with smoking: structural and functional neural changes in nicotine dependence. Brain Sci. 3, 159–176 (2013).

    Article  Google Scholar 

  22. 22.

    Friedman, G. D. et al. Cardia: study design, recruitment, and some characteristics of the examined subjects. J. Clin. Epidemiol. 41, 1105–1116 (1988).

    CAS  Article  Google Scholar 

  23. 23.

    Shen, D. & Davatzikos, C. HAMMER: hierarchical attribute matching mechanism for elastic registration. IEEE Trans. Med. Imaging 21, 1421–1439 (2002).

    Article  Google Scholar 

  24. 24.

    Lao, Z. et al. Computer-assisted segmentation of white matter lesions in 3D MR images using support vector machine. Acad. Radiol. 15, 300–313 (2008).

    Article  Google Scholar 

  25. 25.

    Zacharaki, E. I., Kanterakis, S., Bryan, R. N. & Davatzikos, C. Measuring brain lesion progression with a supervised tissue classification system. Med Image Comput. Comput. Assist Interv. 11(Pt 1), 620–627 (2008).

    PubMed  Google Scholar 

  26. 26.

    Goldszal, A. F. et al. An image-processing system for qualitative and quantitative volumetric analysis of brain images. J. Comput. Assist. Tomogr. 22, 827–837 (1998).

    CAS  Article  Google Scholar 

  27. 27.

    Doshi, J., Erus, G., Ou, Y., Gaonkar, B., Davatzikos, C. Multi-atlas skull-stripping. Acad. Radiol. 2013.

  28. 28.

    Sled, J. G., Zijdenbos, A. P. & Evans, A. C. A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans. Med. Imaging 17, 87–97 (1998).

    CAS  Article  Google Scholar 

  29. 29.

    Li, C., Gore, J. C. & Davatzikos, C. Multiplicative intrinsic component optimization (MICO) for MRI bias field estimation and tissue segmentation. Magn. Reson. Imaging 32, 913–923 (2014).

    CAS  Article  Google Scholar 

  30. 30.

    Kabani, N. J., Collins, D. L., Evans, A. C. A 3D neuroanatomical atlas. in 4th International Conference on Functional Mapping of the Human Brain (eds Evans, A. C.) (Organization for Human Brain Mapping, Montreal, 1998).

  31. 31.

    Brand, M., Markowitsch, H. Brain structures involved in dementia. in Competence Assessment in Dementia. (ed Stoppe, G.) 25–34 (Springer Vienna, Vienna, 2008).

  32. 32.

    Desikan, R. S. et al. Automated MRI measures identify individuals with mild cognitive impairment and Alzheimer’s disease. Brain 132(Pt 8), 2048–2057 (2009).

    Article  Google Scholar 

  33. 33.

    de Bruin, E. A. et al. Focal brain matter differences associated with lifetime alcohol intake and visual attention in male but not in female non-alcohol-dependent drinkers. Neuroimage 26, 536–545 (2005).

    Article  Google Scholar 

  34. 34.

    Jones, B. F. et al. Differential regional atrophy of the cingulate gyrus in Alzheimer disease: a volumetric MRI study. Cereb. Cortex 16, 1701–1708 (2006).

    Article  Google Scholar 

  35. 35.

    Goldstein, R. Z. & Volkow, N. D. Drug addiction and its underlying neurobiological basis: Neuroimaging evidence for the involvement of the frontal cortex. Am. J. Psychiatry 159, 1642–1652 (2002).

    Article  Google Scholar 

  36. 36.

    Radloff, L. S. The CES-D Scale. Appl. Psychol. Meas. 1, 385–401 (1977).

    Article  Google Scholar 

  37. 37.

    Auer, R. et al. Change in physical activity after smoking cessation: the Coronary Artery Risk Development in Young Adults (CARDIA) study. Addiction 109, 1172–1183 (2014).

    Article  Google Scholar 

  38. 38.

    Freund, K. M., D’Agostino, R. B., Belanger, A. J., Kannel, W. B. & Stokes, I. I. I. J. Predictors of smoking cessation: the Framingham Study. Am. J. Epidemiol. 135, 957–964 (1992).

    CAS  Article  Google Scholar 

  39. 39.

    Domino, E. F. Tobacco smoking and MRI/MRS brain abnormalities compared to nonsmokers. Prog. Neuropsychopharmacol. Biol. Psychiatry 32, 1778–1781 (2008).

    CAS  Article  Google Scholar 

  40. 40.

    Peng, P. et al. Brain-volume changes in young and middle-aged smokers: a DARTEL-based voxel-based morphometry study. Clin. Resp. J. 11, 621-631 (2015).

  41. 41.

    Durazzo, T. C., Gazdzinski, S. & Meyerhoff, D. J. The neurobiological and neurocognitive consequences of chronic cigarette smoking in alcohol use disorders. Alcohol. Alcohol. 42, 174–185 (2007).

    CAS  Article  Google Scholar 

  42. 42.

    Durazzo, T. C., Insel, P. S. & Weiner, M. W. Greater regional brain atrophy rate in healthy elders with a history of cigarette smoking. Alzheimer’s & Dement.: J. Alzheimer’s Assoc. 8, 513–519 (2012).

    Article  Google Scholar 

  43. 43.

    Weiland, B. J. et al. Substance abuse risk in emerging adults associated with smaller frontal gray matter volumes and higher externalizing behaviors. Drug Alcohol. Depend. 137, 68–75 (2014).

    Article  Google Scholar 

  44. 44.

    Welch, K. A., Carson, A. & Lawrie, S. M. Brain structure in adolescents and young adults with alcohol problems: Systematic review of imaging studies. Alcohol. Alcohol. 48, 433–444 (2013).

    CAS  Article  Google Scholar 

  45. 45.

    Lyoo, I. K. et al. Prefrontal and temporal gray matter density decreases in opiate dependence. Psychopharmacology. 184, 139–144 (2006).

    CAS  Article  Google Scholar 

  46. 46.

    Droutman, V., Read, S. J. & Bechara, A. Revisiting the role of the insula in addiction. Trends Cogn. Sci. 19, 414–420 (2015).

    Article  Google Scholar 

  47. 47.

    Naqvi, N. H. & Bechara, A. The insula and drug addiction: an interoceptive view of pleasure, urges, and decision-making. Brain. Struct. Funct. 214, 435–450 (2010).

    Article  Google Scholar 

  48. 48.

    Topiwala, A. et al. Moderate alcohol consumption as risk factor for adverse brain outcomes and cognitive decline: longitudinal cohort study. BMJ 357, j2353 (2017).

    Article  Google Scholar 

  49. 49.

    Chen, W. J., Edwards, R. B., Romero, R. D., Parnell, S. E. & Monk, R. J. Long-term nicotine exposure reduces Purkinje cell number in the adult rat cerebellar vermis. Neurotoxicol. Teratol. 25, 329–334 (2003).

    CAS  Article  Google Scholar 

  50. 50.

    Trauth, J. A., Seidler, F. J. & Slotkin, T. A. An animal model of adolescent nicotine exposure: effects on gene expression and macromolecular constituents in rat brain regions. Brain Res. 867, 29–39 (2000).

    CAS  Article  Google Scholar 

  51. 51.

    Dwyer, J. B., McQuown, S. C. & Leslie, F. M. The Dynamic Effects of Nicotine on the Developing Brain. Pharmacol. Ther. 122, 125–139 (2009).

    CAS  Article  Google Scholar 

  52. 52.

    Addicott, M. A. et al. Nicotine and non-nicotine smoking factors differentially modulate craving, withdrawal and cerebral blood flow as measured with arterial spin labeling. Neuropsychopharmacology 39, 2750–2759 (2014).

    CAS  Article  Google Scholar 

  53. 53.

    Hillmer, A. T., Mason, G. F., Fucito, L. M., O’Malley, S. S. & Cosgrove, K. P. How imaging glutamate, GABA, and dopamine can inform the clinical treatment of alcohol dependence and withdrawal. Alcohol. Clin. Exp. Res. 39, 2268–2282 (2015).

    CAS  Article  Google Scholar 

  54. 54.

    Ravan, S., Martinez, D., Slifstein, M., Anissa, A. D. Chapter 18—Molecular imaging in alcohol dependence. in Handbook of Clinical Neurology Vol. 125 (eds Sullivan, E. V., Pfefferbaum, A.) 293–311 (Elsevier, 2014,

  55. 55.

    Weinberger, A. H. et al. Cigarette smoking is associated with increased risk of substance use disorder relapse: a nationally representative, prospective longitudinal investigation. J. Clin. Psychiatry 78, e152–e160 (2017).

    Article  Google Scholar 

Download references


The Coronary Artery Risk Development in Young Adults Study (CARDIA) is conducted and supported by the National Heart, Lung, and Blood Institute (NHLBI) in collaboration with the University of Alabama at Birmingham (HHSN268201300025C & HHSN268201300026C), Northwestern University (HHSN268201300027C), University of Minnesota (HHSN268201300028C), Kaiser Foundation Research Institute (HHSN268201300029C), and Johns Hopkins University School of Medicine (HHSN268200900041C). CARDIA is also partially supported by the Intramural Research Program of the National Institute on Aging (NIA) and an intra-agency agreement between NIA and NHLBI (AG0005). This manuscript has been reviewed by CARDIA for scientific content.

Author information



Corresponding author

Correspondence to Lenore J. Launer.

Ethics declarations

Conflict of interest

This work was initiated while D.C.G. was at the Colorado School of Public Health. The views expressed in this article are those of the authors and do not necessarily represent the views of the National Heart, Lung, and Blood Institute; National Institutes of Health; or the United States Department of Health and Human Services. The authors declare that they have no conflict of interest.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Elbejjani, M., Auer, R., Jacobs, D.R. et al. Cigarette smoking and gray matter brain volumes in middle age adults: the CARDIA Brain MRI sub-study. Transl Psychiatry 9, 78 (2019).

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • DOI:

Further reading


Quick links