Change in smoking status after spiral chest computed tomography scan screening
Cancer screening may provide a “teachable moment” for the reduction of high-risk behaviors. The current study evaluated smoking behavior changes in current and former smokers after low-dose, fast spiral chest computed tomography scan (CT) screening for lung carcinoma.
The study was comprised of 901 current smokers and 574 former smokers who participated in a low-dose, fast spiral chest CT scan screening study for lung carcinoma. Demographic, pulmonary function, screening recommendations, and smoking history variables were evaluated as predictors of self-reported point prevalence smoking abstinence 1 year after screening.
Of the current smokers at baseline, 14% reported smoking abstinence at follow-up. Older age and poorer lung function were associated with smoking abstinence. Ninety percent of former smokers reported smoking abstinence at a 1-year of follow-up. A longer duration of smoking abstinence at baseline was found to be predictive of abstinence in this group.
The 14% smoking abstinence rate was higher than would be expected for spontaneous rates of smoking cessation. Therefore, screening may provide a teachable moment for smokers. Low-dose, fast spiral chest CT scan screening recommendations were not found to be associated with smoking behavior change in either group. Further research is needed to evaluate the potential avenues through which lung carcinoma screening can be used as an opportunity for providing effective nicotine interventions. Cancer 2003. © 2003 American Cancer Society.
Cigarette smoking is the primary cause of lung carcinoma, accounting for nearly 90% of all lung carcinomas.1 One effective means of preventing lung carcinoma is abstinence from smoking. Even among individuals diagnosed with lung carcinoma, smoking cessation reduces the risk of development of a second primary tumor.2, 3 Research has also demonstrated that individuals who continue to smoke after a cancer diagnosis report lower quality of life compared with individuals who stop smoking.4 However, it has been estimated that 30–60% of smokers will continue to smoke after being diagnosed with cancer.5 For example, in a study of 317 smokers diagnosed with Mountain6 Stage I nonsmall cell lung carcinoma (NSCLC), researchers found a 12-month tobacco abstinence rate of 53% and a 24-month tobacco abstinence rate of 47%.7 Others have reported an 86% tobacco abstinence rate after thoracotomy for resection of a pulmonary malignancy.8
Despite the importance of providing smoking cessation interventions for individuals either diagnosed with lung carcinoma or for those considered to be at high risk for developing lung carcinoma, this is an underinvestigated area. To our knowledge, studies published to date are few in number and are confounded by methodologic shortcomings,5 including nonrandomization and small sample sizes. Currently, there are no clear treatment guidelines for intervening with tobacco users at high risk for developing lung carcinoma.
To our knowledge, few studies published to date have evaluated interventions for tobacco users who are undergoing cancer screening. Our literature search yielded only three studies. One study included 580 smokers who received a Papanicolaou (Pap) test for cervical carcinoma screening. It evaluated the efficacy of a telephone counseling intervention for stopping smoking.9 In that study, 3 telephone counseling calls did not appear to improve the smoking abstinence rates at a 6-month follow-up compared with usual care (10.9% vs. 10.5%). The limitations of that study were the timing of the telephone counseling interventions (delivered 4–6 weeks after subjects received their Pap test results) and a lack of information or instructions for the use of pharmacotherapy for smoking cessation. The second study involved low-dose chest computed tomography (CT) screening for the early detection of lung carcinoma, but no treatment for tobacco dependence.10 In that study of 134 current smokers, 23% reported abstinence from smoking and 26% reported decreased cigarettes smoked per day in a follow-up telephone survey. Those who stopped smoking were more likely to be younger (age 68 years or younger), report more cancer-related anxiety, or perceived more health benefits of stopping smoking. It is noteworthy that the majority (85%) of the participants stated they would have liked to receive smoking cessation treatment as part of the lung carcinoma screening program. The authors of that study concluded that lung carcinoma screening presents a “teachable moment” for smoking cessation. The third study, a pilot project, evaluated the change in smoking status for 55 women undergoing early detection screening for lung carcinoma that included sputum cytology, a chest X-ray, bronchoscopy, and a spiral CT scan.11 Participants were twice advised by the study physician to quit smoking: first, at the time they underwent screening and second, 2 weeks later when they received their screening results. At a 1-month follow up, 16% had stopped smoking. As a group, these participants showed high levels of nicotine addiction, low confidence in their ability to quit, and all perceived themselves at high risk for developing lung carcinoma. The authors of that study concluded that more needs to be learned regarding how to tailor nicotine interventions for cancer screening populations.
How to use personal health information to improve nicotine dependence treatment is not well understood. For example, in a study of 518 male spit tobacco users, researchers found that although advice from a dentist improved tobacco abstinence rates, providing feedback on clinically apparent oral lesions did not impact spit tobacco use.12 Similarly, in a sample of asymptomatic smokers, positive results from electron beam CT scans for coronary calcification impacted the subject's perceived cardiovascular risk, but did not appear to impact any smoking behavior change.13 Therefore, we need to learn more about how to use disease biomarkers to enhance the success of nicotine dependence interventions.14
In January 1999, a low-dose, fast spiral chest CT scan study was initiated at the Mayo Clinic (Rochester, MN) to determine whether the use of such screening for NSCLC has a reasonable likelihood of decreasing mortality by increasing the diagnosis of Stage IA/IB tumors according to the 1997 Mountain staging system.6 Participants in the low-dose, fast spiral chest CT scan study had at least a 20-pack-year smoking history and, therefore, were considered to be at risk for smoking-related morbidity and were motivated to engage in lung carcinoma screening. However, it is not known to what degree participation in low-dose, fast spiral chest CT scan screening may influence changes in smoking behavior. For example, it is possible that undergoing screening (i.e., acknowledging increased cancer risk, interacting with healthcare providers, and obtaining screening and screening results) might promote smoking abstinence. Laboratory studies have shown that feedback on health consequences of smoking enhances motivation to stop smoking.15 Therefore, it is plausible that receiving a positive screening result may have a particularly strong influence on abstinence. Alternately, there may be a potential for current or former smokers who receive a negative screening result to continue or restart smoking, possibly resulting from a perceived reduction in risk.
The primary goals of the current study were to evaluate the change in smoking behavior in a large sample of current and former smokers receiving low-dose, fast spiral chest CT scan screening and to identify factors associated with smoking abstinence. Because this group of individuals is at high risk for developing lung carcinoma, identifying predictors of smoking cessation is important for tailoring pharmacologic and behavioral interventions to the needs of individual smokers.
MATERIALS AND METHODS
Between January 20, 1999 and December 15, 1999, 1520 participants were enrolled and underwent a baseline prevalence low-dose, spiral chest CT scan. The study participants were comprised of 785 men and 735 women. Of these, 1508 (99%) were white and 12 were African American, Native American, or Hispanic. All were at least 50 years of age (mean age, 59 years; range, 50–85 years). Sixty-one percent were current smokers and 39% were former smokers. The median number of pack-years of smoking was 45 (range, 20–230 pack-years). Of the 1520 participants enrolled, 1475 (97%) returned and had smoking status information obtained at their first annual follow-up, which was performed within a 1-month window on either side of the 1-year anniversary of the baseline scan. These 1475 participants (901 current smokers and 574 former smokers at the time of the baseline CT scan) form the basis of the current study.
The community was informed about the low-dose, fast spiral chest CT scan screening study by local and regional television and newspaper coverage that carried information regarding the general outline of the study and eligibility requirements as well as funding by the National Cancer Institute. Participants were enrolled after written informed consent was obtained and included asymptomatic men and women 50 years of age or older who were current or former (quit fewer than 10 years previously) cigarette smokers with at least a 20 pack-year history of smoking. Ineligible participants had a history of any cancer within 5 years, other than nonmelanomatous skin carcinoma, cervical carcinoma in situ, or localized prostate carcinoma. Participants were required to be mentally competent and considered healthy enough to potentially undergo pulmonary resection (i.e., no congestive heart failure or disabling dyspnea at the time of enrollment). Individuals with a serious illness that decreased life expectancy to less than 5 years were excluded. The protocol was approved by the Mayo Foundation institutional review board.
A detailed description of the methods used in the low-dose, fast spiral CT scan screening study, including a summary of how the CT scan findings were used to guide recommendations for follow-up spiral CT scans or diagnostic evaluations, was provided in a previous publication.16 A brief description is presented with emphasis placed on the details most relevant to the current study.
The participants had to be either current cigarette smokers or had to have quit within the past 10 years. They also needed to have a pack-year history of smoking cigarettes of at least 20 years (examples: 1 pack per day for 20 years = 20 pack years; one-half pack per day for 40 years = 20 pack-years; 2 packs per day for 10 years = 20 pack-years). At the annual follow-up visit, the participants were asked, “Do you now smoke cigarettes?” If “yes,” they were asked, “How many cigarettes do you smoke per day?” If “no,” they were asked, “When was your quit date?” Of the current smokers at baseline who reported being abstinent from smoking at the 1-year follow-up visit, 33% did not report their quit date. Of those who reported a quit date, 95% reported being abstinent for 7 or more days. Smoking abstinence rates in the current study were based on participant self-report. Midway through the study, we began to collect expired air carbon monoxide (CO) levels to biochemically confirm point prevalence smoking abstinence. Of the 710 participants who completed the CO screening, 314 self-reported smoking abstinence. Of these, 98% were biochemically confirmed to be abstinent. These findings support the validity of the self-reported abstinence rates obtained from the total sample.
Spirometry (forced expiratory volume in 1 second [FEV1]) was performed on each participant using the Puritan Bennett Renaissance spirometer (Puritan Bennett, Pleasanton, CA). The technicians were certified pulmonary function technologists. Forced expiratory spirometry was performed in accordance with the Lung Health Study protocol and with the standards of the American Thoracic Society. All scans were performed on a multislice, spiral CT scanner. All CT scan images were viewed in cine-mode formats at a computer workstation by one of four investigator radiologists.
Logistic regression was employed to identify baseline characteristics predictive of abstinence from smoking at the 1-year follow-up visit. Separate analyses were performed for individuals who were current and former smokers at baseline. Characteristics considered in these analyses included continuous variables for age, average cigarettes per day, number of years smoked, percentage of predicted FEV1, categoric variables for gender, and the recommendation received as a result of the baseline low-dose, fast spiral chest CT scan screening findings (no follow-up, follow-up in 6 months, follow-up in 3 months, follow-up as soon as possible/biopsy, and other). Follow-up recommendations were based on the size of the largest noncalcified nodule detected on the CT scan. Nodules 3 mm or smaller, 4–7 mm, and 8 mm or larger were followed with a CT scan in 6 months, 3 months, and immediately, respectively. For participants who were former smokers at baseline, the duration of abstinence from smoking at the time of the baseline visit was also analyzed as a potential predictor of abstinence from smoking at the 1-year follow-up visit. Duration of abstinence was analyzed as both a continuous variable and also categorically using the following categories: 3.0 months or shorter duration, 3.1–6.0 months, 6.1–12.0 months, 12.1–24.0 months, and 24.1 months or longer duration. In all cases, two-tailed P values less than or equal to 0.05 were considered statistically significant.
Of the 1475 participants who returned and had smoking status information obtained at their first annual follow-up screening, 901 (448 men and 453 women) were current smokers at baseline. For these participants, the mean age ± standard deviation (SD) was 59.0 ± 6.5 years. After the baseline low-dose, fast spiral chest CT scan screening, the recommendation was no follow-up (i.e., no return before 12 months) for 386 participants(43%), follow-up in 6 months for 188 participants (21%), follow-up in 3 months for 230 participants(26%), biopsy or follow-up as soon as possible for 44 participants(5%), and other recommendations for 53 participants(6%). (This group was for another abnormality that was found that required “other” follow-up [e.g., a mammogram for a breast nodule, an ultrasound of a kidney for a cyst vs. mass, or an ultrasound for an abdominal aneurysm].) At the first annual follow-up, 129 (14%; 95% confidence interval [CI] = 12–17%) of these participants self-reported abstinence from smoking. Table 1 presents the findings of the univariate logistic regression analysis assessing potential predictors of smoking abstinence at 1 year for individuals who were current smokers at baseline. From this analysis, older age (odds ratio [OR] = 1.4 for every 10-year increase, P = 0.029) and lower percentage of predicted FEV1 (OR = 1.2 for every 10 percentage point decrease, P < 0.001) were univariately associated with abstinence at 1 year. From multivariate analysis, after adjusting for all other potential predictors presented in Table 1, a lower percentage of predicted FEV1 (OR = 1.2 for every 10 percentage point decrease, P = 0.002) was the only significant predictor of abstinence from smoking at 1 year.
Table 1. Characteristics Associated with Smoking Abstinence at 1 Year in Individuals Who Were Current Smokers at Baseline
|Age (yrs)b|| || ||0.029|
| 50–54||267||12|| |
| 55–59||258||14|| |
| 60–64||196||13|| |
| 65–69||109||17|| |
| ≥ 70||71||25|| |
|Gender|| || ||0.870|
| Male||448||15|| |
| Female||453||14|| |
|FEV1 (% predicted)b|| || ||< 0.001|
| ≤ 60||136||24|| |
| 61–70||124||19|| |
| 71–80||179||12|| |
| 81–90||197||13|| |
| ≥ 91||248||10|| |
|Recommendation|| || ||0.653|
| No follow-up||386||13|| |
| 6 mos||188||16|| |
| 3 mos||230||14|| |
| ASAP/biopsy||44||11|| |
| Other||53||19|| |
|Cigarettes per dayb|| || ||0.489|
| 10–19||65||11|| |
| 20||473||16|| |
| 21–39||223||13|| |
| 40||116||13|| |
| ≥ 41||24||13|| |
|Yrs smokedb|| || ||0.148|
| ≤ 25||29||10|| |
| 26–35||301||13|| |
| 36–45||396||13|| |
| ≥ 46||175||20|| |
For 574 participants (313 men and 261 women) who were former smokers at the baseline visit, the mean age ± SD age was 60.7 ± 6.8 years. The median duration of smoking abstinence was 3.8 years (interquartile range, 0.9–7.1 years). Based on the findings of the baseline low-dose, fast spiral chest CT scan screening, the recommendation was no follow-up for 250 participants (44%), follow-up in 6 months for 110 participants (19%), follow-up in 3 months for 153 participants (27%), biopsy or follow-up as soon as possible for 32 participants (6%), and other for 29 participants (5%). At the first annual follow-up visit, 516 (90%; 95% CI = 87–95%) of those participants who were former smokers at baseline reported that they were still abstinent from smoking and 10% of former smokers were now smoking again. Table 2 presents the findings of the univariate logistic regression analysis assessing potential predictors of abstinence at 1 year among individuals who were former smokers at baseline. For these former smokers, abstinence from smoking at the 1-year follow-up visit was not found to be significantly associated with age, gender, average number of cigarettes per day, percentage of predicted FEV1, or the recommendation received as a result of the baseline spiral CT scan findings. There was evidence (P = 0.012) that a longer smoking history was associated with a lower rate of abstinence from smoking at the 1-year follow-up. In addition, a strong association was detected between the duration of abstinence at the time of the baseline visit and abstinence from smoking at the 1-year follow-up visit (P < 0.001 with duration of abstinence treated as both a continuous and categoric variable). From visual inspection of the data, it appeared that for former smokers with more than 2 years of abstinence at the time of the baseline visit, there was no significant association between the duration of abstinence beyond 2 years and abstinence from smoking at the 1-year follow-up visit. This lack of association was confirmed in an analysis restricted to participants who had been abstinent from smoking for more than 2 years. Therefore, subsequent analyses were performed that treated duration of abstinence as a continuous variable with a value of 730 days assigned for participants who had been abstinent from smoking for more than 2 years at the time of the baseline visit. From a univariate analysis using this independent variable, longer duration of smoking abstinence at the time of the baseline visit was significantly associated with an increased likelihood of smoking abstinence at the 1-year follow-up visit (OR = 1.6 for each additional 3 months of abstinence up to 2 years; P < 0.001). From multivariate analysis, after adjusting for all other potential predictors presented in Table 2, longer duration of smoking abstinence at the time of the baseline visit (OR = 1.7 for each additional 3 months of smoking abstinence up to 2 years; P < 0.001) was the only significant predictor of abstinence from smoking at 1 year.
Table 2. Characteristics Associated with Smoking Abstinence at 1 Year in Individuals Who Were Former Smokers at Baseline
|Age (yrs)b|| || ||0.583|
| 50–54||117||88|| |
| 55–59||154||90|| |
| 60–64||135||89|| |
| 65–69||109||92|| |
| ≥ 70||59||92|| |
|Gender|| || ||0.221|
| Male||313||89|| |
| Female||261||92|| |
|FEV1 (% predicted)b|| || ||0.992|
| ≤ 60||107||91|| |
| 61–70||70||90|| |
| 71–80||92||85|| |
| 81–90||110||91|| |
| ≥ 91||180||91|| |
|Recommendation|| || ||0.163|
| No follow-up||250||89|| |
| 6 mos||110||85|| |
| 3 mos||153||94|| |
| ASAP/biopsy||32||94|| |
| Other||29||90|| |
|Cigarettes per dayb|| || ||0.298|
| 10–19||19||89|| |
| 20||250||87|| |
| 21–39||140||94|| |
| 40||126||90|| |
| ≥ 41||39||92|| |
|Yrs smokedb|| || ||0.012|
| ≤ 25||57||96|| |
| 26–35||192||91|| |
| 36–45||556||89|| |
| ≥ 46||99||85|| |
|Duration of abstinence (mos)|| || || |
| ≤ 3.0||55||58||< 0.001|
| 3.1 to 6.0||39||62|| |
| 6.1 to 12.0||48||79|| |
| 12.1 to 24.0||60||95|| |
| ≥ 24.1||368||98|| |
The current study evaluated smoking behavior change in current and former smokers after low-dose, fast spiral chest CT scan screening for lung carcinoma. Fourteen percent of current smokers at baseline reported smoking abstinence 1 year after CT scan screening. The design of the current study did not allow direct evaluation of the impact of the screening process itself on abstinence, as we did not have a comparison group of high-risk current and former smokers who did not undergo CT scan screening. However, the 14% abstinence rate is greater than the expected annual abstinence rate of 5–7% among self-quitters in the general population.17 The smoking cessation rate in the current study is similar to the abstinence rates reported by many clinical interventions.18 Therefore, screening may provide a teachable moment for smokers.
Lung carcinoma screening may provide an opportunity for nicotine dependence intervention. Previous investigators found that 86% of smokers undergoing CT scan screening stated they would have liked to receive assistance with smoking cessation as part of the lung carcinoma screening process.10 These findings support the use of cancer screening as a teachable moment in which to build on a patient's openness to learn new information and heightened motivation to promote health by providing tobacco dependence intervention.10, 19, 20 Furthermore, lung carcinoma screening provides an opportunity to implement relapse prevention efforts with former smokers, as the current findings show that 10% of former smokers resumed smoking in the year after screening. The current study findings suggest that clinicians promoting smoking abstinence with screening participants must not overlook relapse prevention in former smokers, particularly those in the early period of abstinence.
Among participants who were current smokers at baseline, pulmonary function results predicted smoking abstinence at the 1-year follow-up. Active smokers with poorer lung function were more likely to stop smoking after screening. This finding may suggest that undergoing spirometry and receiving abnormal results regarding pulmonary functioning may promote abstinence in smokers. The degree to which actual physical symptoms experienced by smokers with poorer lung function promote smoking abstinence versus the spirometry examination results is unknown. Some studies evaluating the impact of spirometry on smoking behavior change have found that smokers with abnormal spirometry findings are more likely to stop smoking,21, 22 whereas other studies have suggested that these smokers are less likely to change their smoking behavior.23, 24 Tobacco dependence interventions that have included spirometry feedback have demonstrated success,25, 26 although the independent impact of spirometry within these studies is unknown. As reduced FEV1 is an independent risk factor for lung carcinoma,27 feedback on abnormal findings within cancer screening may be useful in motivating smoking behavior change. Further investigation in this area is warranted.
Evaluation of predictors of smoking behavior change in current and former smokers demonstrated that low-dose, fast spiral chest CT scan screening recommendations were not associated with smoking abstinence 1 year after CT scan screening. Within the current study, we were able to evaluate the potential impact of screening outcomes on smoking behavior by comparing change in smoking behavior between individuals with normal CT scan findings and individuals with abnormal CT scan findings. It is noteworthy that the CT scan findings and subsequent follow-up recommendations given to current and former smokers were nearly identical. In both groups, on multivariate analysis, there was no evidence that screening recommendations predicted smoking behavior change. This finding is in contrast to results described by Ostroff et al.10 that current smokers who received normal screening results were less likely to stop smoking. These authors reported that such findings may suggest potential barriers to tobacco use treatment interventions with smokers who have normal CT scan screening results. Within the current study, we found no evidence that lung carcinoma screening provided a “green light” for continued smoking, as individuals with negative screening results were not less likely to change their smoking behavior.
One concern that has been raised about participation in a lung CT scan screening project is that screening results may promote a return to smoking. In the current study, 10% of subjects who were abstinent at baseline reported that they were smoking at the 1-year follow-up. The only predictor of recurrence was that a longer duration of smoking abstinence at baseline significantly predicted abstinence rate at the 1-year follow-up. Individuals abstinent at baseline for 1 year or less experienced a 30% relapse rate and individuals abstinent for 2 or more years experienced a 2% relapse rate. These relapse rates are similar to those demonstrated by clinical programs. For example, participants in an 8-week smoking cessation program, who were all abstinent at 1 year, experienced a 22% relapse rate at the 2-year follow-up.28 Therefore, our results provide further support for the inclusion of relapse prevention treatment for recently (past 12 months) abstinent smokers and indicate that participation in a lung carcinoma CT scan screening program does not promote a return to smoking in long-term abstainers.28
Several important factors were not evaluated in the current study. Individuals who adhere to medical recommendations may be more likely to quit smoking and the presence of noncancer-related medical diagnoses may impact an individual's motivation to quit. Future investigators should consider assessing these factors in individuals undergoing cancer screening. We also did not evaluate cognitive factors such as motivation for change, decision-making processes, or patient understanding or interpretation of screening (e.g., spiral chest CT scan screening and pulmonary function) feedback. Furthermore, we did not collect data concerning current smokers who made stop attempts but resumed smoking during the year after screening. Ostroff et al.10 found that of the 23% of current smokers who stopped smoking and 26% who reduced their smoking after lung carcinoma screening, the majority (87%) described the lung carcinoma screening process as a major influence on changing their smoking behavior. Perceived benefits of stopping smoking and anxiety about developing lung carcinoma were associated with smoking behavior change.10 Additional information on stop attempts, lapse and relapse, and cognitive processes related to smoking behavior change would provide greater understanding of the process of smoking behavior change after screening and could be used to tailor nicotine dependence treatment interventions for this group of high-risk smokers.
More research is needed to further evaluate the impact of the cancer screening process itself on smoking behavior change, the influence of specific screening results (e.g., spiral CT scans, spirometry) on promoting abstinence, the process of smoking behavior change in this high-risk group, and methods of effectively providing tobacco use treatment interventions to enhance smoking abstinence. The use of cancer screening feedback as a motivational component of a nicotine dependence intervention merits additional study.