Does menthol attenuate the effect of bupropion among African American smokers?
ABSTRACT
Background African Americans have higher tobacco-related morbidity and mortality and are more likely to smoke menthol cigarettes than their white counterparts. This study examined differences between African American menthol and non-menthol smokers in smoking characteristics and cessation.
Methods The study sample consisted of 600 African American smokers enrolled in a clinical trial that assessed the efficacy of sustained-release bupropion for smoking cessation. Menthol (n = 471) and non-menthol (n = 129) smokers were compared on smoking-related characteristics and abstinence rates at 6 weeks and 6 months.
Results Menthol smokers were younger (41.2 versus 52.9 years), more likely to be female (73.7% versus 56.6%) and more likely to smoke their first cigarette within 30 minutes of waking up (81.7% versus 69.8%) compared to non-menthol smokers (all P < 0.01). Cigarette taste (50% versus 40%, P = 0.054) was rated non-significantly higher by menthol smokers. Seven-day point-prevalence abstinence from smoking at 6 weeks were 28% and 42% (P = 0.006) and at 6 months were 21% and 27% (P = 0.21) for menthol and non-menthol smokers, respectively. At 6 weeks follow-up, stepwise logistic regression revealed that among those younger than 50 years, non-menthol smokers were more likely to quit smoking (odds ratio = 2.0; 95% CI = 1.03–3.95) as were those who received bupropion (odds ratio = 2.12; 95% CI = 1.32–3.39).
Conclusion African American menthol smokers had lower smoking cessation rates after 6 weeks of treatment with bupropion-SR, thereby putting menthol smokers at greater risk from the health effects of smoking. Lower overall cessation rates among African Americans menthol smokers may partially explain ethnic differences in smoking-related disease risks.
INTRODUCTION
One of the most striking differences in the smoking patterns between African American and white smokers is the preference for menthol cigarettes. Whereas approximately 80% of African American smokers smoke menthol cigarettes, the proportion among whites is only about 20% (Kabat et al. 1991). African Americans also smoke fewer cigarettes per day (Caraballo et al. 1998) and begin smoking later in life compared to whites (Kabat et al. 1991; Okuyemi et al. 2001). Despite smoking fewer cigarettes per day, African Americans experience disproportionately higher rates of smoking-related health consequences (Harris et al. 1993). African Americans have the highest incidence rates of all cancers combined, and the highest overall cancer mortality rates compared to other racial/ethnic groups (Harris et al. 1993). Because of their high preference for menthol cigarettes, it has been suggested that menthol cigarette smoking may contribute to the excess smoking-related morbidity experienced by African Americans.
Menthol is a naturally occurring flavoring element (Sidney et al. 1995) and one of thousands of chemicals that may be added to cigarettes during the manufacturing process. Menthol combustion produces carcinogenic compounds such as benzopyrenes (Schmeltz & Schlotzhauer 1968) which might contribute directly to higher lung cancer rates. However, research on the association between menthol cigarette use and lung cancer has produced mixed results. A case–control study consisting of 337 cases and 478 controls (Carpenter et al. 1999) did not find increased lung cancer rates among menthol smokers compared to non-menthol smokers. In contrast, a prospective study of over 11 000 patients enrolled in a health maintenance organization and followed for 6–12 years found increased risk of lung cancer among male menthol smokers (relative risk = 1.45 versus male non-menthol smokers; adjusted for age, race, education, number of cigarettes smoked per day and duration of smoking) but not in females (Sidney et al. 1995). While it remains unclear whether menthol increases the risk of lung cancer directly, it is possible that the relationship between menthol and tobacco-related morbidity is an indirect one, acting through other factors that might increase disease risk.
For example, due to its local anesthetic and cooling effects menthol may affect smoking topography in a number of ways, including puff volume and depth of smoke inhalation (Caskey et al. 1993). These factors probably increase exposure to tobacco smoke toxins and consequently disease risk. Another hypothesis is that menthol makes smoking more enjoyable. Menthol cigarette smokers may therefore be less able to quit smoking and continue smoking for longer periods of time. A longer duration of smoking would increase the risk of tobacco-related diseases.
Given the disproportionately high burden of tobacco-related health consequences experienced by African Americans and their high rates of smoking menthol cigarettes, there is a need to understand better the association between menthol and cessation in African Americans. Such an understanding is important in reducing health disparities in tobacco-related diseases between various racial/ethnic groups in the United States. To test the hypothesis that menthol smokers are less likely to quit, we conducted a secondary analysis of data from a clinical trial which tested bupropion-SR for smoking cessation in 600 African American smokers. Differences in smoking-related characteristics between menthol and non-menthol smokers were also examined.
METHODS
Study design
The parent study was a double-blind, placebo-controlled, randomized trial of 600 African American smokers (Ahluwalia et al. 2002). Participants were recruited at an inner-city community health center over a 16-month period and were assigned randomly to receive a 7-week supply of placebo or bupropion SR 150 twice daily. Medication was dispensed by study staff during clinic visits and participants were given instruction about usage. Participants in both groups received eight brief counseling sessions and a previously developed smoking cessation guide. Eligible people described themselves as ‘African American or black’, were at least 18 years of age, smoked at least 10 cigarettes per day, were interested in quitting in the next 30 days, spoke English and had a home address and working telephone. Participants were excluded if they had a contraindication for bupropion SR (predisposition to seizures, excessive alcohol use, bulimia or anorexia nervosa, current use of bupropion), were pregnant, currently used psychoactive medication, used other forms of tobacco or nicotine replacement in the past 30 days, were in drug treatment during the past 6 months or were being treated for depression.
At the baseline visit, participants completed a battery of assessments (e.g. socio-demographics, smoking behaviors, nicotine dependence, depression, perceived stress and withdrawal symptoms) and were assigned randomly to receive either bupropion SR (Zyban, Glaxo Wellcome, Research Triangle Park, NC, USA) 150 mg once a day or placebo once a day for 3 days, followed by twice a day for a total of 7 weeks. Participants set a target quit date 1 week after the start of medication, and returned on that day (quit day) for a second in-person visit (week 0). They returned for follow-up assessments and counseling the following week (week 1) and at weeks 3 and 6 and month 6. A detailed description of study procedures is provided elsewhere (Ahluwalia et al. 2002). Participants provided written informed consent during the first visit. The trial procedures were approved and monitored by the University of Kansas Medical Center's Human Subjects Committee.
Measures
The baseline assessment included measures of demographics, health information and smoking behaviors (see Table 1). Nicotine dependence was assessed with the modified six-item Fagerström Test for Nicotine Dependence (FTND) scale, a widely used measure of nicotine dependence with established reliability (Cronbach's alpha = 0.61) (Heatherton et al. 1991). Use of menthol cigarettes was assessed with the question, ‘Do you usually smoke menthol cigarettes?’ Outcomes reported in the current analysis are 7-day point prevalence smoking cessation at 6 weeks and 6 months. Self-reported abstinence was confirmed with expired carbon monoxide (CO) assessment (<10 p.p.m.) and only in the case of discrepancies was saliva collected and used to resolve the discrepancy with salivary cotinine analysis (<20 ng ml) (Ahluwalia et al. 2002; SRNT 2002).
| Characteristic | Menthol smokers n = 471 | Non-menthol smokers n = 129 | P-value |
|---|---|---|---|
| Age in years, mean (SD) | 41.8 (9.4) | 52.9 (12.3) | <0.001 |
| Female, % | 73.7 | 56.6 | <0.001 |
| Monthly income ≤ $1 800, % | 54.6 | 54.7 | 0.977 |
| ≤High school graduate, % | 50.7 | 47.3 | 0.487 |
| Employed, % | 78.3 | 65.9 | 0.004 |
| CPD, mean (SD) | 18.1 (5.1) | 18.3 (5.8) | 0.188 |
| FTND score, mean (SD) | 4.7 (1.9) | 4.4 (2.2) | 0.760 |
| Smoked 1st cigarette ≤ 30 minutes of waking up, % yes | 81.7 | 69.8 | 0.003 |
| How satisfying was most recent cigarette, % very/extremely | 57.3 | 50.4 | 0.160 |
| Taste of most recent cigarette, % very/extremely good | 49.9 | 40.3 | 0.054 |
| Salivary cotinine, mean (SD) | 292.9 (147.8) | 288.1 (123.8) | 0.712 |
| Expired CO, mean (SD) | 21.7 (14.7) | 22.7 (12.4) | 0.982 |
Statistical analysis
Surveys were double-data entered and analysis was performed with SAS version 8 (SAS Institute 1999). Subjects were classified into two groups: menthol smokers and non-menthol smokers. Categorical baseline variables were summarized by frequencies and percentages while quantitative baseline variables were summarized by means and standard deviations. The χ2 test was used to determine if any of the categorical baseline variables differed between the two groups. The two-sample t-test was used to assess if any of the quantitative baseline variables differed between the two groups. Logistic regression was used to assess the effect of smoking menthol cigarettes on 7-day point prevalence smoking cessation at 6 weeks while controlling for treatment. Stepwise logistic regression and best subsets logistic regression were used to explore the joint relationship of treatment, menthol cigarette smoking and the other baseline variables on 7-day point prevalence smoking cessation at 6 weeks. All statistical analyses were performed on an intention-to-treat basis as defined by Piantadosi (Piantadosi 1997) in that subjects were analyzed in the group into which they were randomized. We did not impute those lost to follow-up as smokers because we had shown previously that missing data were missing completely at random (Ahluwalia et al. 2002).
Given the known association between age and cessation, we examined a priori the distribution of menthol smoking by age and its potential interaction with menthol use on cessation. Age 50 years corresponds to the 3rd quartile of age distribution for our study sample. Menthol cigarette use was much more prevalent in those under the age of 50 years compared to those 50 and above. Furthermore, there was a significant interaction between age dichotomized at 50 years and menthol use on cessation. Logistic regression models were therefore performed separately by two age categories, <50 years and ≥50 years. Logistic regression for 7-day point prevalence abstinence is reported for 6 weeks only, because univariate analysis did not reveal significant differences in abstinence rates between menthol and non-menthol smokers at 6 months.
RESULTS
Of the 600 smokers enrolled in the study, 471 (78.5%) smoked menthol cigarettes whereas 129 (21.5%) smoked non-menthol cigarettes. Demographic and smoking characteristics of participants are shown in Table 1. Menthol cigarette smokers were younger (41.2 years versus 52.9 years for non-menthol smokers; P < 0.0001), more likely to be female (73.7% versus 56.6% for non-menthol smokers; P < 0.001) and more likely to be employed (78.3% versus 65.9% for non-menthol smokers; P = 0.0002).
Although menthol and non-menthol smokers had similar scores on the Fagerstrom Test for Nicotine Dependence (FTND), menthol smokers were more likely to smoke their first cigarette within 30 minutes of waking up (81.7% versus 69.8%; P = 0.003). Menthol smokers also rated the taste of their most recent cigarette (49.9% versus 40.3% for non-menthol; P = 0.054) non-significantly higher than non-menthol smokers. However, the rating for satisfaction with their most recent cigarette was not significantly different between the two groups (57.3% versus 50.4% for non-menthol; P = 0.160).
Overall 28.3% of menthol smokers were abstinent at 6 weeks (end of treatment, n = 535) compared to 41.5% of non-menthol smokers (P = 0.006). The abstinence rate at 6 months (n = 518) was also lower among menthol smokers but the difference was not statistically significant (21.4% versus 27.0% for non-menthol smokers; P = 0.21). When separated by treatment (Fig. 1; bupropion, n = 265; or placebo, n = 270), among those who received bupropion the 7-day point-prevalence abstinence rate at 6 weeks for non-menthol smokers (60.3%) was significantly higher than for menthol smokers (36.2%, P < 0.01). Abstinence rates did not differ by menthol status among those who received placebo (23.3% non-menthol versus 20.5% menthol; P = 0.63). Test for interaction between treatment and menthol status on cessation was not significant (P = 0.07)
Seven-day point-prevalence abstinence rates according to treatment received; *P < 0.01.
Further examination of the association between age and menthol smoking revealed that the ratio of menthol to non-menthol smokers was substantially higher among those younger than 50 years of age. Among those <50 years (n = 443), 88.7% smoked menthol cigarettes whereas only 49.7% of those ≥50 years (n = 157) smoked menthol cigarettes (P < 0.0001). There was also a significant interaction between categorized age (<50 versus ≤ 50) and menthol status on 7-day point-prevalence abstinence at 6 weeks (P = 0.02). Other variables in Table 1 that were associated with menthol status (gender, employment, cigarettes smoked per day, smoking ≤30 minutes of waking, satisfaction and taste of most recent cigarette) were also tested for potential interaction, but none showed significant interaction with menthol status on cessation. Because of significant interaction between age and menthol smoking on cessation, associations between menthol and cessation were further examined by two age groups, <50 versus ≥50 years. Figure 2 shows abstinence rates at 6 weeks for non-menthol and menthol smokers by age groups.
Seven-day point-prevalence abstinence rates by age groups; *P < 0.01.
At 6 weeks follow-up, among those <50 years, 44.4% of non-menthol smokers were abstinent compared to 24.9% for menthol smokers (P < 0.01). Abstinence rates among those ≥ 50 years did not differ significantly by menthol status (P = 0.57). Two stepwise logistic regression models were constructed to predict the probability of abstinence at 6 weeks; one for those aged <50 years and the other for those aged ≥ 50 years. Variables considered for inclusion in both models were those listed in Table 1 and whether they received bupropion or placebo. Among individuals <50 years, non-menthol smokers were twice as likely to quit smoking at the end of 6 weeks compared to menthol smokers (odds ratio = 2.02; 95% CI = 1.03–3.95). Those who received bupropion were more likely to quit (odds ratio = 2.12; 95% CI = 1.32–3.39), whereas those who smoked their first cigarette within 30 minutes of waking were less likely to quit (odds ratio = 0.37; 95% CI = 0.22–0.62). Among those ≥50 years, receiving bupropion was the only predictor of abstinence at 6 weeks (odds ratio = 4.43; 95% CI = 2.17–9.04).
DISCUSSION
In this analysis from a clinical trial of sustained-release bupropion for smoking cessation among African Americans, menthol cigarette smokers had lower cessation rates after 6 weeks of treatment with the study drug. The lower cessation rates among menthol cigarette smokers was found only in those younger than 50 years. This difference in cessation rates between menthol and non-menthol smokers seems to be driven by treatment with bupropion as there was no significant difference in cessation rates by menthol status in the placebo group. The magnitude of the difference in quit rates was also greater when broken down by treatment received. The cessation rates at 6 months were lower for menthol smokers, but the difference was not statistically significant. Lack of significance at 6 months could be due to a number of reasons, including delayed quitting among menthol smokers, higher relapse among non-menthol smokers or, in fact, a reflection of true finding (i.e. difference in rates at 6 weeks is a spurious finding). There was also a higher attrition rate at 6 months (14%) compared to 6 weeks (11%), which could reduce the power to detect a significant difference between menthol and non-menthol smokers even if one exists. However, there was no differential attrition at 6 months between menthol and non-menthol smokers.
To our knowledge, these are the first data from a clinical trial to show that African American menthol smokers have short-term lower cessation rates than non-menthol smokers. Because the vast majority of African American smokers smoke menthol cigarettes, lower cessation rates for menthol smokers could translate into an overall reduction in cessation rates for African Americans. A number of studies have shown that despite making more quit attempts than whites on average in a given year, African Americans are less successful in their cessation attempts compared to whites (CDC 1994, 1998). The greater difficulty with quitting experienced by African Americans have been attributed to a number of factors, including smoking high nicotine/tar cigarettes and higher levels of nicotine dependence (Royce et al. 1993). Findings from the present study suggest that smoking menthol cigarettes may be a contributory factor in the difficulty experienced by African Americans in smoking cessation. Although the literature on menthol cigarette smoking and the risks of lung cancer is inconclusive, lower cessation rates associated with smoking menthol cigarettes could explain in part the ethnic differences in smoking-related diseases.
Why would menthol smokers be less successful quitting smoking? First, menthol cigarette smokers may have greater difficulty in quitting because they are more nicotine-dependent than non-menthol smokers. Data from the current study showed that menthol smokers were more likely to smoke their first cigarette within 30 minutes of waking up, an indicator of dependence. This was similar to findings reported by Ahijevych & Parsley (1999). However, in the present study the effect of menthol on cessation persists after controlling for ‘smoking within 30 minutes of waking’, a measure of nicotine dependence. Secondly, menthol smokers in current study rated cigarette taste higher than their non-menthol counterparts. Due to the local anesthetic and cooling effects of menthol, menthol in cigarettes may allow smokers to take larger puffs, hold their breaths for longer periods of time and be able to inhale deeper compared to non-menthol smokers. The resultant effect will be increased exposure to nicotine and other tobacco smoke constituents. However, while some studies found no significant differences in the depth of inhalation or puff volume between menthol and non-menthol smokers (Sidney et al. 1989; Caskey et al. 1993; McCarthy et al. 1995; Ahijevych et al. 1996), other studies have reported that menthol cigarettes increased smokers’ carbon monoxide levels. One study of 20 male smokers (Jarvik et al. 1994) found that mentholated cigarettes decreased puff volumes but increased puff flow rates of inhaled smoke, which may increase smokers’ exposure to toxic effects of carbon monoxide. In a sample of 95 women, significantly higher puff volumes were identified in menthol compared to non-menthol smokers (45.8 versus 37.8 ml), with an equal representation of African American and white women in each menthol preference group (Ahijevych & Parsley 1999).
Thirdly, some data suggest that menthol may also affect metabolism of nicotine. Two studies showed that menthol smokers have higher cotinine/cigarette ratios (Clark et al. 1996; Ahijevych & Parsley 1999). Another study found that bupropion attenuates the effect of the enzyme CYP2B6, which has been shown to affect cessation (Lerman et al. 2002). However, it is not known if menthol affects the metabolism of bupropion. If the metabolism of other toxic/carcinogenic constituents of tobacco smoke is also slower in menthol smokers, this subgroup of smokers may potentially be exposed to relatively higher doses of toxins in tobacco smoke than non-menthol smokers with similar smoking levels. Fourthly, menthol status was determined at enrollment and we did not collect data about duration of use of menthol cigarettes or whether or not participants had switched the brand of cigarettes they smoked. If some non-menthol smokers were former menthol smokers who had switched brands in preparation for quitting, cessation rates could be biased in favor of the non-menthol group. This is a subject for further study. Furthermore, smoking of menthol cigarettes could be a marker for some other yet undetermined factor. This is all the more plausible given that the lower quit rates observed among menthol smokers in the current study were only found among those younger than 50 years of age. Menthol cigarette smoking could be a marker for some other behavior associated with young age. For example, if younger smokers, who are more likely to smoke menthol cigarettes, are less compliant with study medication, then age could be a contributing factor to the differential cessation rates found among menthol and non-menthol smokers. Finally, because cessation rates did not significantly differ between menthol and non-menthol smokers on placebo, the lower cessation rates for menthol smokers could be specific for bupropion. Additional support for this explanation is that differential cessation rates were found only during treatment (6 weeks) but not subsequently (6 months). The null finding among those who received placebo was consistent with findings from two recent studies (Hyland et al. 2002; Muscat et al. 2002) that found no association between use of mentholated cigarettes and cessation. However, one of the studies (Muscat et al. 2002) used a cross-sectional design and had a predominantly older sample (>80% aged>45 years) while the other study (Hyland et al. 2002) had a very low proportion of menthol (24%) and African American (<10%) smokers.
Our study has some limitations. Data were derived from a study that was not designed to test whether menthol smokers have lower quit rates than non-menthol smokers. The sample was predominantly menthol smokers, reflecting an expected prevalence of menthol smoking among African Americans. A sample with equal distribution of menthol versus non-menthol smokers may reveal other factors associated with menthol smoking not found in the current study. However, having a larger sample of non-menthol smokers in the study is unlikely to change factors that were found to be significant in this study. Also, information about menthol smoking was obtained by self-report but there was no reason to believe that participants would misrepresent their menthol smoking status, and the proportions of smokers in both groups were as expected. Furthermore, this study was limited to African Americans and therefore the study's findings may not generalize to smokers of other racial/ethnic groups in the United States and other parts of the world. Studies are needed to determine whether menthol smokers from other racial/ethnic groups also have lower cessation rates than their non-menthol smoking counterparts.
Clearly, there is a need to understand better the reason for the lower cessation rates among African American menthol smokers. Smoking cessation studies with a significant proportion of menthol smokers need to adjust for menthol status in their estimation of cessation rates. Also, because cigarette taste was rated higher by menthol smokers, studies are needed to examine whether some pharmacological interventions are more effective for menthol smokers. For example, the nicotine inhaler which contains menthol (although in small quantity), for taste and behavioral reasons, may prove to be more effective for smoking cessation among menthol smokers. Studies are also needed to assess if the lower cessation among menthol smokers is specific only to bupropion or if this pattern generalizes to other forms of pharmacotherapy. If a lower cessation rate among menthol smokers is specific for bupropion, then bupropion would not be the optimal treatment for menthol smokers.
In summary, our study shows that African American menthol smokers on sustained-release bupropion had lower smoking cessation rates after 6 weeks treatment. If these findings are confirmed in future studies, then lower overall cessation rates among African Americans may be explained partially by their predominantly smoking menthol cigarettes, which in turn may explain ethnic differences in smoking-related disease risks. Also, the higher success rate at 6 weeks in non-menthol smokers may provide a stronger basis from which to make the next quit attempt. Addressing factors contributing to lower cessation rates among menthol smokers is an important step towards eliminating health disparities due to tobacco use among various racial/ethnic groups.
ACKNOWLEDGEMENTS
An abstract with preliminary results was presented at the annual meeting of the Society for Research in Nicotine and Tobacco, Savannah, Georgia, 20–24 February 2002. This study was supported by grants from the National Cancer Institute (R01 CA77856, K07 CA90334, R24 CA95835-01). Glaxo-Welcome, Inc. provided study medication; however, they played no role in the design, conduct, interpretation and analysis of the study. The authors thank the staff at Swope Parkway Health Center for making this research possible.
