Journal of Addictive Behaviors,Therapy & RehabilitationISSN: 2324-9005

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.

Research Article, J Addict Behav Ther Rehabil Vol: 2 Issue: 3

Cigarette Smoking, Dependence and Chronic Obstructive Pulmonary Disease: A Renewed Approach of an Old Relationship

Salameh Pascale1*, Khayat Georges2, Salame Joseph3 and Waked Mirna4
1Lebanese University - Faculties of Pharmacy & of Public Health, Beirut, Lebanon
2Hôtel Dieu de France Hospital - Beirut & Saint Joseph University- Faculty of Medicine, Beirut, Lebanon
3Lebanese University – Faculty of Medicine, Beirut, Lebanon
4Saint George Hospital - Beirut & Balamand University - Faculty of Medicine, Beirut, Lebanon
Corresponding author : Pascale Salameh, PharmD, MPH, PhD
Professor of Epidemiology, Jdeidet El Meten, Chalet Suisse street, Ramza Azzam bldg, 5th floor, Beirut, Lebanon
Tel: 009613385542; Fax: 009611696600
E-mail: [email protected], [email protected]
Received: July 04, 2013 Accepted: August 23, 2013 Published: August 26, 2013
Citation: Salameh P, Khayat G, Salamé J, Waked M (2013) Cigarette Smoking, Dependence and Chronic Obstructive Pulmonary Disease: A Renewed Approach of an Old Relationship. J Addict Behav Ther Rehabil 2:3. doi:10.4172/2324-9005.1000111


Cigarette Smoking, Dependence and Chronic Obstructive Pulmonary Disease: A Renewed Approach of an Old Relationship

Introduction: Cigarette dependence is common in cigarette smokers, but the evaluation of its relationship with Chronic Obstructive Pulmonary Disease (COPD) has been rarely performed, particularly for the dimensions of this dependence. The objective of this study is to assess such an association. Methods: We used data on current smokers from two samples: a cross-sectional national study and a case-control study. Results: We found a significant association between cigarette dependence and COPD in current smokers (OR between 2 and 4); the results were confirmed in samples of both studies, by multivariate analysis and dose-effect relationship. The results were in favor of an effect that is independent of cumulative smoking dose and of nicotine physiological dependence; moreover, there was a significant increase in the prevalence of COPD per LCD quintile and FTND dependence categories (p<0.001 for trends). Conclusion: Cigarette dependence is associated with COPD in current smokers, independently of the smoking dose and of nicotine dependence level.

Keywords: FTND, LCD, tobacco, chronic obstruction, Lebanon


FTND; LCD; Tobacco; Chronic obstruction; Lebanon


Tobacco consumption is one of the leading causes of deadly diseases in the world, and total tobacco-attributable deaths are projected to rise to 8.3 million in 2030, at which point they will represent almost 10% of all deaths globally, mainly in developing countries [1]. Some researchers draw a distinction between nicotine dependence and tobacco dependence [2], so that nicotine dependence is sometimes considered as one dimension of the more complex tobacco dependence [3]. In regular smokers, the Fagerström Test for Nicotine Dependence (FTND) is used as a screening tool for physiological nicotine dependence [4,5]; it is extensively used in various countries [6]. The Lebanon Cigarette Dependence (LCD) scale is another tool intended to evaluate the whole cigarette dependence concept, including nicotine dependence, reinforcement and craving intensity [7].
Smokers with nicotine dependence tend to have increased cigarette consumption, and this addiction may worsen the impact of smoking exposure by altering the frequency or depth of smoke inhalation, besides consuming a greater number of cigarettes; this was shown in patients with Chronic Obstructive Pulmonary Disease (COPD) [8]. The other way round, smokers with COPD exhibit higher levels of cigarette dependence, smoke more cigarettes a day, and have higher cotinine concentrations than smokers without COPD. However, they do not display higher motivation to quit than smokers without COPD [9]. As a result, it is reasonable to hypothesize that increased dependence to nicotine would facilitate the development and progression of COPD, a disease that is considered highly tobacco related and preventable [1,10].
Nevertheless, available studies have linked overall cigarette dependence to COPD, while very few of them looked at dimensions of tobacco dependence in this relationship. The objective of this analysis was to evaluate the association between overall cigarette dependence and its dimensions (nicotine dependence, psychological craving and reinforcement), smoking dose and COPD in current smokers.

Materials and Methods

Population and sampling
Sample 1 (epidemiological setting): A cross-sectional study was carried out between October 2009 and September 2010, using a multistage cluster sample all over Lebanon. Lebanese residents aged 40 years and above were enrolled in the study, with no exclusion criteria. From the list of circumscriptions, we randomly selected one hundred circumscriptions i.e. local communities. We then randomly chose individuals to be interviewed from a provided list of dwelling households through a local authority representative; all eligible individuals within a household were interviewed using a standardized questionnaire. Additional methodological details are available in a separate publication [11].
Sample 2 (clinical setting): A case-control study was performed between July 2009 and June 2010, comparing a group of patients with COPD from two tertiary care hospitals in Beirut with a control group. The COPD group was composed of incident outpatient cases of COPD. Cases were included if they were ≥ 40 years of age, free of other respiratory diseases, diagnosed as COPD by a chest physician, and had a post-bronchodilator FEV1/FVC <0.7 [12]. Outpatients consulting for various extra-pulmonary problems were included as controls if they were ≥ 40 years of age, free of any respiratory disease or symptom. Additional methodological details are available in a separate publication [13].
For both studies, the Lebanese University ethical review board waived the need for approval because they were observational studies that respected both confidentiality and autonomy of involved participants.


After informed consent, subjects had baseline spirometry (Micro Lab, Micro Medical Limited, England) by a trained technician and were interviewed to answer a questionnaire. Thirty minutes after inhaling 2 puffs of ipratropium bromide (18 μg/actuation) and albuterol sulfate (103 μg/actuation) (Combivent®) in a pressurized metered-dose aerosol unit, post bronchodilator spirometry was performed. The best of 3 trials was taken and used to classify patients as COPD. Carbon monoxide in exhaled air measurements were taken to verify smoking status of participants. COPD was defined and classified according to GOLD guidelines (FEV1/FVC <0.70 postbronchodilator) [12].

Questionnaire in Samples 1 and 2

The used questionnaire in both samples was composed of several parts, including the sociodemographic part, detailed smoking history, and detailed respiratory health using the American Thoracic Society questionnaire [14]. Current cigarette smoking was defined as “smoking more than one pack in a lifetime” and “smoking currently”; smokers could thus include irregular smokers. Cumulative smoking dose was evaluated by multiplying the mean numbers of packs smoked per day by the duration of smoking in years (pack*years).
For cigarette dependence evaluation in current smokers, we used the well known FTND test to measure nicotine dependence [4,15], and the LCD scale to measure cigarette dependence, including reinforcement, nicotine dependence and craving intensity subscales [7]. In fact, although other scales exist for cigarette dependence evaluations [16,17], they have never been used or validated in the Lebanese population. Oppositely, we chose the FTND scale because it has already been used in Lebanon [18,19], and the LCD scale since it was specifically developed and validated in the Lebanese population [7].
Whilst the point of the FTND is to measure physical dependence on tobacco, the LCD assesses a broader range of aspects relevant to smoking and to cessation difficulty (including items that measure reasons for smoking). Thus, the LCD aims to measure a broader, more multi-faceted concept of dependence, with the ability to cover the main components of DSM-IV and ICD-10 definitions of dependence. The LCD, with its three factors structure (reinforcement, craving intensity and nicotine dependence per se) shows a wider picture of cigarette dependence than the FTND that only reflects nicotine dependence [7]. During development, it had adequate psychometric properties (Chronbach alpha=0.78). During factor analysis, the KMO measure of sampling adequacy was 0.838, while the Bartlett’s test of sphericity gave significant results (p<0.001). Between factors correlation was low to moderate: r=0.336 between factors 2 and 3 (p<0.001), r=0.229 between factors 1 and 2 (p<0.001), and r=0.055 between factors 1 and 3 (p>0.05). Moreover, the first factor explained 27.52% of the total variance, while the second explained 19.24% and the third 9.41%, making a total of 56.16% of the data variance. The LCD score also had a good face validity and was inversely correlated with the number of times the participant tried to stop smoking without succeeding (r=-0.215; p<0.001) [7]. Afterwards, the LCD was validated in a second sample by running a confirmatory factor analysis, and showed adequate properties [7].

Statistical Analysis

Data entry was performed by independent lay persons that were unaware of the objectives of the study. Statistical analysis was performed using SPSS software, version 13.0. An alpha value of 0.05 will be considered significant. For sample 1, cluster sampling effect was taken into account according to Rumeau-Rouquette and collaborators [20]. Data weighting was performed according to the Central Administration of Statistics numbers, taking into account geographical dwelling, age and gender [21].
Afterwards, bivariate analyses were carried out on samples 1 and 2. We computed the OR of having COPD if the subject was considered dependent to cigarettes according to FTND test (if his or her score was superior to the cut-off of 5, which also corresponds to the median in our population) [22] and according to the LCD score (if his or her score was superior to the calculated median, in parallel to the median of FTND that corresponds to dependence cut-off).
In the latter case of categorical variables association, the Chi square test of independence was used in case all calculated cell counts were ≥ 5 within tables, while Fisher exact test was used in case the calculated cell count was <5. Moreover, a Pearson correlation coefficient was calculated to evaluate correlation between continuous variables (the FTND and LCD scores, and the cumulative smoking dose). A Somers’d test was also applied to test for ordinal variables and dose-effect relationships, using scales quintiles dependence categories for LCD, and the recommended using a five-category distribution including very low (scores 0-2), low (3-4), moderate (5), high (6-7) and very high (8-10) dependence [23].
Multivariate analysis were also performed to evaluate predictors of COPD as a dependent variable, taking into account potential confounding variables that had a p-value <0.20 in bivariate analysis: gender, residency, education, work status, and marital status were taken as nominal variables, while age, height, weight, and body mass index were considered continuous variables. Multivariate analyses were carried out using stepwise likelihood ratio logistic regressions on samples 1 and 2, after ensuring the model adequacy to data by Hosmer and Lemeshow test. Absence of interaction with potential confounding was checked and multi-colinearity absence was verified before accepting the final models.
Six separate models were generated: one using dichotomous FTND (nicotine dependence) as a major independent variable, the second using dichotomous LCD (cigarette dependence) as a major independent variable, the third using dichotomous FTND (nicotine dependence), LCD (cigarette dependence) and cumulative smoking dose (in pack-years) as major independent variables, the fourth using LCD and FTND as continuous variables, the fifth using LCD factors (as quantitative variables) and cumulative smoking (in pack-years) as major independent variables, and the sixth using LCD quintiles and FTND recommended dependence categories. Adjusted OR (aOR) were calculated.


Samples 1 and 2 description
In sample 1 (household survey), out of 2201 individuals, 637 were current smokers and answered to FTND and LCD test questions. The mean was 5.47 (SD=2.61); 315 (49.45%) had a FTND score higher than 5 and were thus considered dependent to nicotine, while 322 (50.55%) had a score lower than six. Moreover, in this epidemiological setting, participants without COPD had a mean FTND of 5.16 (SD=2.59), compared with 6.91(SD=2.32) for COPD ones (p<0.001); for the LCD, participants without COPD had a mean of 20.52 (SD=8.19), versus 26.58 (SD=8.08) for those with COPD (p<0.001).
In sample 2 (case-control study), a total of 738 individuals were included in the study: 527 healthy individuals and 211 cases of COPD. 95 cases were current smokers (45%); out of them, 65(68.4%) had a FTND score higher than 5 (mean=7.17; SD=2.19); in controls, 132 (25%) were current smokers, while 37 (28% of smokers) had an FTND higher than 5 (mean=4.32; SD=2.39). In this clinical sample, the results were respectively: 4.28 (SD=2.41) versus 7.17 (SD=2.19) for FTND, and 31.66 (SD=7.97) versus 37.95(9.04) for LCD (p<0.001 for both scales).
Bivariate analysis (sample 1 and sample 2)
According to the median score of 22, individuals with an LCD score threshold of 23 or more were considered cigarette dependent. Using this threshold, we were able to determine a dichotomous variable for cigarette dependence, and to link it with COPD, in comparison with the fixed threshold of 5 in case of using the FTND scale (which also corresponded to the median in our population). We found that dependence to cigarette, whether measured by FTND or by the LCD scale, gave a moderate to strong association with COPD. The association with the LCD was at least 50% higher than with the FTND scale (OR=2.93 versus 2.04, respectively; p<0.001 for both), demonstrating a better ability to predict COPD in sample 1 (Table 1). In sample 2, however, the FTND and LCD gave close results of a strong association (p<0.001 for both).
Table 1: Cigarette dependence in current smokers and COPD association.
Multivariate analysis
When both FTND and LCD were included within the same model (global model; sample 1), the only scale that maintained a significant association with COPD was the LCD (aOR=2.81; p<0.001). In this latter model, the cumulative dose of cigarette smoking also gave a significant result (aOR=1.38; p=0.023). In another model, the three LCD factors were included, along with the cumulative smoking dose: here the second factor (the “nicotine dependence” factor) was removed from the model due to non significance, while the cumulative smoking dose persisted (Table 1). Moreover, in bivariate analysis, for every increase of 1 point in LCD (range 1 to 40), the odds of COPD increased by 12% (OR=1.12[1.08-1.16]; p<0.001); for every increase of 1 point in FTND (range 1 to 10), the odds of COPD increased by 13% (OR=1.13[1.09-1.17]; p<0.001). In multivariate analysis, these OR became 1.06 and 1.20, respectively (p<0.05 for both).
In sample 2, the FTND units were better associated with COPD versus LCD units, both in bivariate and multivariate analyses; cumulative smoking was removed from the model in multivariate analysis. When the global model analysis is run on sample 2, the LCD dependence was independently and strongly associated with COPD: aOR=3.37 (p<0.001), while the FTND dependence was removed from the model. Interestingly, when the three LCD factors were included along with the cumulative smoking dose, the second factor “nicotine dependence” and the cumulative smoking dose were both removed from the model due to non significance; only factor 1 “reinforcement” (aOR=1.14; p=0.009) and factor 3 “craving intensity” (aOR=1.25; p<0.001) were retained in the model.
Dose-effect relationships
There was an inverse correlation between the LCD score and post-bronchodilator FEV1/FVC (r=-0.377; p<0.001), while the FTND score gave lower correlation with post-bronchodilator FEV1/ FVC (r=-0.131; p<0.001). In Figure 1, we show the dose-effect relationship of both FTND and LCD scales from sample 1. Although there was a significant dose-effect increase for all relationships (p<0.001 for Somers’d test), a steeper increase in the association with COPD is shown with the LCD quintiles compared to the FTND ones, particularly in the highest dependence category (Figure 1). For all of these results, substantially similar results were found using sample 2.
Figure 1: COPD prevalence by dependence categories of FTND and quintiles of LCD scale. Dependence categories for FTND are: 0-2 (very low dependence), 3-4 (low dependence), 5 (moderate dependence), 6-7 (high dependence), and 8-10 (very high dependence).
Moreover, additional bivariate and multivariate results are also shown in Table 1; they show a clear dose-effect relationship for LCD quintiles in both samples, while the FTND is only retained in the epidemiological sample; cumulative smoking dose was removed for both samples (Table 1).


In this study, cigarette dependence, measured by both the FTND (physiological nicotine dependence) and the LCD (physiological and psychological dependence) scores, showed a higher risk of COPD among current smokers. The association with nicotine dependence (FTND) has already been demonstrated in the literature, where smokers with COPD showed a higher dependence to nicotine than healthy smokers using Fagerström test score (p<0.001) [8,24], and the Heaviness of Smoking Index, a shorter version of the FTND [9].
Individual variations may give plausible explanations to this finding: correlations were found between some loci of smoking status, number of cigarettes per day as well as level of nicotine dependence and individual localizations in the genome [25-29]. Interestingly, it has been suggested that fast nicotine metabolizers may be less susceptible to nicotine harmful effects, but may need to use more tobacco to compensate for their rapid metabolism, so they are more exposed to other tobacco harmful effect [2,30-32].
The association with the LCD leads us to another finding: individuals who have higher cigarette dependence (whether due to psychological craving or reinforcement) have a higher risk of COPD, independent of smoking cumulative dose and of physiological nicotine dependence. The latter idea may be explained by the fact that meaningful individual differences in dependence are observed independent of differences in current daily cigarette consumption and duration of smoking [33]; to our knowledge, the association of these cigarette dependence dimensions with COPD is being reported for the first time. In fact, we found a better ability to predict COPD by the LCD score compared with the FTND, particularly for the “reinforcement” and “craving intensity” factors. These results were maintained in quantitative bivariate and multivariate analyses, in epidemiological (sample 1) and clinical settings (sample 2); the physiological nicotine dependence factor was steadily removed from the multivariate models due to non significance and the cumulative dose gave unstable results. Here also, one biological explanation for this result may be genetic, whereas the same genes may underlie psychological cigarette dependence and COPD susceptibility, regardless of nicotine physiological dependence: indeed, a recent study demonstrated that COPD related locus has an important yet unclear genetic influence on tobacco dependence [34].
There were some differences between the results of samples 1 and 2 that deserve our attention: First, the LCD was consistently better associated with the FTND in sample 1, while results were almost similar in sample 2 in bivariate analyses. This may be explained by the cross-sectional nature of sample 1 that is highly heterogeneous (diagnosed and undiagnosed cases of COPD and chronic bronchitis), while the sample 2 is more homogeneously divided between cases (diagnosed COPD cases and mostly heavy smokers) and controls (no respiratory disease and mainly light smokers). This would lead to heterogeneous distributions of dependence dimensions and disease presentations in sample 1, compared with parallelism of dependence (whichever scale is used) and disease in sample 2. Second, in multivariate analyses, the smoking dose significantly predicted COPD in sample 1, while it gave no significant association with the disease in sample 2, once LCD factors were introduced to the model. It was also removed when dependence scales were used as quantitative or ordinal variables in both samples. This could be explained by a better capture of COPD variance by dependence scales that include the smoking dose in addition to other items (LCD in particular), or by a low power to detect significant associations in some situations. Additional studies would be helpful to resolve this particular matter.
Furthermore, individual differences in nicotine metabolism and their relationship to susceptibility for regular tobacco use and tobacco related diseases are in the early stages of understanding [2]. More information is needed on the metabolic processes and the toxicokinetic variation of nicotine and other tobacco substrates, for a better appreciation of how these differences influence susceptibility to becoming nicotine or cigarette dependent [31,35], along with their relation to tobacco-related diseases [2]. We also expect that other exposures to toxics such as pesticides or other environmental factors may act as enzyme inducers or inhibitors and affect this complex relationship by interaction on oxidative stress [36,37]; this remains to be studied in specifically designed studies.
We are aware of the limitations of this work: there is a possibility of selection bias, which may hamper generalizability of the results. There is also a possibility of classification bias due to the methods we used in the cross-sectional study, such as the portable spirometer that may not be as sensitive as fixed ones and the standardized questionnaire that may involve a recall bias on several variables, including smoking related ones and potential confounding. However, the confirmatory analysis performed on the case-control study corroborates the findings of the cross-sectional study in a clinical setting; thus, we have no reason to believe that changing the methods would have changed the essential of our study results. Finally, longitudinal studies are needed to confirm the hypothesized causal link between dependence to cigarettes and the development of COPD.


In conclusion, cigarette and nicotine dependence are associated with COPD, independently of the smoking dose. Further studies are needed to account for the metabolic processes and the psychological effects differences of tobacco substrates, for a better appreciation of how these differences influence susceptibility to becoming nicotine or cigarette dependent, along with their relation to tobacco-related diseases predisposition.


All authors have no competing interest to declare. This work was funded by an unrestricted educational grant from Boehringer Ingleheim – Lebanon company.


Track Your Manuscript

Recommended Journals