|Home | About | Journals | Submit | Contact Us | Français|
Smokers who switch to “lighter” cigarettes may be diverted from quitting smoking. We assessed factors associated with switching and the association between switching and (1) making a quit attempt, and (2) recent quitting, yielding a measure of net quitting (attempts × recent quitting).
In 2003, a total of 30 800 ever-smokers who smoked in the past year provided history of switching and 3 reasons for switching: harm reduction, quitting smoking and flavour. Among those who made a past-year quit attempt, recent quitting was defined as ≥90-day abstinence when surveyed. Multivariable logistic regression identified determinants of outcomes.
In all, 12 009 (38%) of ever-smokers switched. Among switchers, the most commonly cited reasons were flavour only (26%) and all 3 reasons (18%). Switchers (vs non-switchers) were more likely to make a quit attempt between 2002 and 2003 (51% vs 41%, p<0.001, adjusted odds ratio (AOR) 1.58, (95% confidence interval (CI) 1.48 to 1.69)), but less likely to have recently quit (9% vs 17%, p<0.001; AOR 0.40 (95% CI 0.35 to 0.45)), yielding lower overall net quitting (4.3% vs 7.0%, p<0.001; AOR 0.54, (95% CI 0.47 to 0.61)). The effects of switching on outcomes were most pronounced for reasons including quitting smoking, whereas switching for harm reduction alone had no association with outcomes.
Compared with no switching, a history of switching was associated with 46% lower odds of net quitting.
So-called low tar and nicotine cigarettes, often referred to as “lights”, now make up about 84% of the US market share1 and are smoked widely in other countries.2 “Lights” were introduced in the 1960s and targeted by the tobacco industry to health conscious smokers who might otherwise quit smoking.3 4 In an effort to categorise cigarettes according to the level of nicotine and tar that they contained, the US Federal Trade Commission (FTC) adopted a machine-based method of determining tar and nicotine content of cigarettes that established three groups: ≥15 mg tar, 6 to <15 mg tar and <6 mg tar.5 While the FTC never provided explicit guidance on cigarette brand descriptors such as “light” or “mild”, in the US the cigarette industry has used the “light” label to describe any brands with up to 10 mg of tar and 0.8 mg of nicotine, based on the FTC method.6 When smoked in real life, these so-called “light” cigarettes deliver amounts of tar, nicotine and other substances to the smoker that are comparable to “regular” cigarettes,7–10 and do not lead to a reduction in tobacco-related illness.11–13
The tobacco industry has been charged with using false descriptors such as “light” and “mild” to provide false reassurance to health-concerned smokers who may otherwise have quit smoking.3 4 14 Global misconceptions that “light” cigarettes are healthier have proved to be difficult to dispel, even after implementation of labelling bans and the Framework Convention on Tobacco Control (FCTC) labelling guidelines on such misleading descriptors.15–18 Furthermore, a history of ever-switching to a “lighter” or “lower yield” cigarette (hereafter referred to as “switching”) is common, with over a third of US adult smokers reporting having done so for the purpose of attempting to reduce health risks.19–21
Thoroughly understanding the potential consequences of switching is critical, particularly in countries such as the US where debate continues over how to end deceptive cigarette labelling and advertising. The 2006 landmark ruling against the tobacco industry by a US District Court refers to “light” and related labelling as “forbidden health descriptors”,22 and new federal legislation23 proposes to ban such descriptors. In spite of global efforts to expose the truth about false cigarette labels, the tobacco industry continues to introduce modified cigarettes with claims of reduced harm, possibly replicating the negative public health impact of “lights”.
A key concern about “lights” has been that smokers who switch to “lights” might be less likely to quit, reducing cessation and increasing the disease burden of tobacco. The total number of smokers who quit (“net quitting”) is a function of (1) the percentage making a quit attempt and (2) the percentage of individuals who succeed at quitting. The two may be influenced differentially by switching. In past studies, switchers were more aware of the health risks of smoking, demonstrated more interest in quitting and in some cases (but not all)24 made more quit attempts.19 25
The relationship between switching and successful quitting remains unclear. Two studies, one prospective but of relatively small sample size, failed to find an association between switching to reduce health risks and successful quitting.24 25 Three large, nationally representative retrospective studies found reduced successful quitting among smokers who had used or switched to a “lighter” cigarette in order to reduce health risks.19–21 Even beliefs about “lighter” cigarettes (as opposed to actually switching) may also influence smoking behaviour: in one study, smokers who believed switching would reduce their health risks were 25% less likely to succeed at quitting smoking in the subsequent year.16
Most prior studies have had little information on reasons for switching. For example, if switching to reduce health risks resolves the cognitive dissonance between health concerns and smoking,16 26 then quit attempts and successful quitting would be expected to be reduced. By contrast, quit attempts and successful quitting might be expected to be higher among those who switch specifically for the purpose of quitting smoking, as these smokers were already expressing an interest in quitting.
Prior research has not fully answered the question of whether switching helps, hinders, or has no bearing on quit attempts, successful quitting, or, perhaps most importantly, on net percentage quitting (attempts × quitting). Further, prior studies have not addressed whether the various reasons for switching might be associated with different effects on these outcomes. To address these questions, we determined the prevalence of and factors associated with switching, and assessed the association between switching and (1) making a quit attempt, (2) recent quitting (≥90-day abstinence among those making an attempt) and (3) their product, net quitting. We hypothesised these outcomes would vary by reason for switching.
We used data collected from the 2003 Tobacco Use Supplement to the Current Population Survey, a large Federal household tobacco survey of the US civilian non-institutionalised population cosponsored by National Cancer Institute and the Centers for Disease Control and Prevention27 28 (detailed methodology has been described elsewhere: http://riskfactor.cancer.gov/studies/tus-cps). The 2003 overall response rate was 83%; excluding proxies, it was 65%.
We focused on 30 800 respondents to the 2003 Tobacco Use Supplement who indicated that they were smoking in 2002 and provided complete information about their own smoking (proxy reports were excluded). Ever-smokers were defined as those who answered “yes” to the question, “Have you ever smoked at least 100 cigarettes in your lifetime?”. This population included current (n = 29 033) and recently quit smokers (n = 1767) (eg, those who had been abstinent for ≥90 days when surveyed). Ever-smokers were asked, “Do you currently smoke cigarettes everyday, some days, or not at all?”. There were 24 918 everyday and 4115 non-daily smokers.
Ever-smokers were asked, “Have you ever switched from a stronger cigarette to a lighter cigarette for at least 6 months?”. Respondents who answered “yes” (defined as “switchers”) were then asked to answer true/false follow-up questions about three possible reasons for switching: “you switched…” (1) “because you felt that a lighter cigarette would be less harmful to your health” (harm reduction); (2) “as a way to quit smoking” and (3) “to get a smoother or lighter flavour”. Respondents who answered affirmatively to two options were asked to rate the most important reason for switching. Because respondents were able to specify multiple reasons for switching, there were eight possible combinations (summarised in table 1) used to create a variable to probe reasons for switching.
Covariates studied included sociodemographic characteristics: age (15–29, 30–39, 40–49, 50–59, 60+), gender, race (white vs other) and highest level of education achieved (<12 years, 12 years (high school), 13+ years); smoking-related characteristics (level of nicotine dependence (smoke first cigarette within 30 min of awakening vs later)29 and cigarettes per day (<10, 10–19, 20–29, ≥30)) and type of cigarette smoked (“light/ultralight”, regular). Recent quitters reported the type of cigarette smoked in the year before quitting. Smokers who indicated “no usual type” or “some other type” of cigarette were excluded. Smokers indicated if they had made a “serious” quit attempt or a quit attempt for at least 24 h in the prior year (yes/no). Smokers were defined as having recently quit smoking if they reported they had not been smoking for ≥90 days at the time of the 2003 survey. The smoking relapse curve flattens after 90 days,30 and short-term success has been shown to predict long-term success.31
We first used descriptive statistics to characterise the overall sample and compare non-switchers with switchers, by reasons for switching (table 1). We further assessed differences among the reasons for switching by each covariate. We assessed the percentage of switchers (overall and by various reasons) versus non-switchers who (1) made a quit attempt (among all smokers, n = 30 800) and (2) achieved recent quitting (among those who made a quit attempt, n = 13 848) (table 2). Net quitting, expressed as a percentage, was the proportion of all 30 800 individuals who achieved recent quitting. To determine whether switching was independently associated with outcomes, we fitted separate multivariable logistic regression models in which the dependent variables were (1) making a quit attempt in the past year (n = 30 800), (2) recent quitting (among those who made a quit attempt, n = 13 848) and (3) net quitting (n = 30 800). We ran each of these models with the independent variable as (1) ever-switching (yes/no) and (2) all eight possible combinations of reasons for switching.
All analyses used SUDAAN software (V. 10; RTI International, Research Triangle Park, North Carolina, USA) to obtain variance estimates using a balanced repeated replication method that accounted for the complex sampling design.27 28 32 Results were weighted to reflect the US population.
The full sample of 30 800 (table 1) was almost half male and was predominantly white. Over half the sample reported smoking “light/ultralight” cigarettes (currently for current smokers, or in the year prior to quitting for recently quit smokers). There were 12 009 switchers (38%). Compared to non-switchers, switchers tended to be older, female, white, more educated and were more likely to smoke the first cigarette within 30 min of waking. Switchers were notably more likely than non-switchers to smoke “light/ultralight” cigarettes (either at the time of survey for current smokers, or in the year prior to quitting for recently quit smokers).
The most commonly reported reason for switching was flavour (26%), followed by a triad of reasons (harm reduction, quitting smoking and flavour) (18%), followed by the combination of harm reduction and flavour (15%). Among switchers, 41% cited a single reason, whereas 51% cited more than one reason (and 8% cited “none of these reasons”). Counting multiple reasons, 43% cited wanting to quit smoking, 51% cited harm reduction and 67% cited milder flavour as a reason for switching (totals add to over 100% because multiple reasons were cited).
Those who switched solely to quit smoking tended to be younger minority women with lower education and less dependence, while those who switched solely for harm reduction tended to be older, white, more highly educated males who smoked heavily. Smokers who had switched solely for flavour tended to be younger white males or females who smoked “light/ultralight” cigarettes. Those who switched for the combined reasons of quitting smoking and harm reduction tended to be middle aged, more educated males and those who switched for quitting smoking and flavour appeared to be younger, less educated females. Individuals who switched for combined harm reduction and flavour tended to be older, more highly dependent white men and women who smoked “light/ultralight” cigarettes. Those who switched for a combination of all three reasons tended to be evenly divided among age and gender categories and were more likely to be smoking “light/ultralight” cigarettes (all p values for differences among reasons for switching by covariates in table 1 <0.001).
Table 2 displays the analyses of quit attempts, recent quitting (≥90-day abstinence at survey) and net quitting by reason for switching. Compared to non-switchers, switchers were more likely to make a quit attempt in the past year (51% vs 41%), but, once they made an attempt, were less likely to have recently quit (9% vs 17%). After adjustment for all covariates in table 1, the odds of switchers making a quit attempt were 58% higher than those of non-switchers. However, switchers who made a quit attempt had a 60% lower odds of recently quitting smoking, compared to non-switchers.
Quit attempts and recent quitting are also shown by reason for switching. The percentage who attempted quitting among individuals who switched for any reason that included “to quit smoking” was higher than among non-switchers, regardless of other reasons cited. After adjustment for all variables in table 1, the odds of making a quit attempt among those who switched solely to quit smoking were 3.6 times greater than odds among non-switchers. Of those who switched solely to quit smoking, 71% made a quit attempt, compared to only 41% of non-switchers. There was no association between switching for harm reduction or flavour and making a quit attempt.
Among those who made a quit attempt, lower rates of achieving recent quitting were seen for switchers (all reasons except harm reduction). Whereas 17% of non-switchers who made a quit attempt recently quit smoking, only 9% of those who had switched for any reason recently quit. This difference held after adjusting for cofactors (adjusted odds ratio (AOR) 0.40, 95% confidence interval (CI) 0.35 to 0.45). Further, recent quitting varied by reason for switching: 6% of switchers who switched solely to quit smoking and 10% of switchers who switched solely for flavour, recently quit. After adjustment for important cofactors, odds of recent quitting among those who switched solely to quit smoking were 70% below those of non-switchers, and among those who switched solely for flavour, odds were 57% below those of non-switchers. The combination of quitting smoking and flavour was associated with 80% lower odds of recent quitting compared to non-switchers. There was no association between switching for harm reduction and recent quitting.
Thus, switching was generally associated with an increased likelihood of making a quit attempt, but a decreased likelihood of recently quitting smoking. How do these two opposing effects affect the net impact of switching on the likelihood of becoming a former smoker? Net quitting (table 2) was the percentage of all past-year smokers who recently quit (≥90-day abstinence at the time of survey). Overall, 51% of all switchers made a quit attempt, but only 9% of this group achieved recent quitting, for a net 4% who quit smoking. In contrast, among non-switchers, 41% attempted quitting and 17% of those succeeded, such that 7% of all non-switchers quit smoking. The adjusted odds of net quitting were 46% lower among switchers (for any reason) compared to non-switchers, and tended to be even lower for those who had switched for the combination of quitting smoking and flavour. Thus, switching was associated with lower net quitting that varied by reason for switching.
In this large, nationally representative sample of US adolescent and adult smokers, switching to a “lighter” cigarette for any reason was associated with 58% higher odds of making a quit attempt, but was also associated with 60% lower odds of achieving recent quitting in such an attempt. Our findings confirm and extend the findings from prior literature16 24 25 by delineating outcomes based on reasons for switching and by assessing net quitting (the product of making a quit attempt and recently quitting). Smokers who had switched for the purpose of quitting smoking (regardless of additional reasons) were the most likely to make a quit attempt, but also the least likely to have recently quit, with similar results for those who switched for any reason that included flavour. The net public health outcome associated with ever-switching was negative for all reasons except harm reduction, which appeared to be neutral.
Why would switching be associated with higher odds of making a quit attempt? The smokers who switched may have been the most motivated to change their smoking, whether by switching or by quitting. The Tobacco Use Supplement does not provide respondent data on their interest in quitting at the time they switched or at the time they made a quit attempt, so these data cannot address this hypothesis.
Given that some switchers may have been more interested in quitting, why were they less successful at quitting? As a result of switching, these smokers may have changed their smoking topography, possibly making smoking behaviour harder to change.33 This potential scenario suggests that having switched might actually undermine successful quitting. It is also possible that smokers who had the most trouble quitting preferentially switched: the smokers who wanted to quit and could, did so; the ones who could not sought an alternative, and switched. On this hypothesis, switching is not exercising a causal influence on cessation, but simply acting as a marker for smokers who were and are motivated to make a quit attempt and maintain abstinence, but who have trouble doing so. It is also possible that switchers began that process as highly motivated individuals who were able to attempt quitting, but whose motivation waned as they (falsely) came to see the “lighter” cigarette as less toxic and thus an acceptable alternative to outright cessation. It remains unclear why individuals who switched for harm reduction seemed less susceptible to these processes.
It is particularly concerning that smokers who switched as a means to quit smoking had the highest rates of quit attempts but also the lowest odds of recent quitting (57% to 80% lower than odds among non-switchers, depending on reasons for switching).
A lesser gap in odds of successful smoking cessation was seen for those who switched solely for better flavour (57% lower than the non-switchers). Of the 38% of smokers who reported switching (representing about 15 million US smokers), over a quarter cited doing so solely for flavour. Prior research suggests that smokers interpret descriptors such as “light” as an implicit indicator of reduced toxicity.34 Cigarette companies may further be exploiting this association with newer descriptors such as “smooth” and “fine”.35 Indeed, Philip Morris test marketed its purported reduced delivery Marlboro cigarettes under the name “Ultra Smooth”.
Thus smokers who switched for reasons that included flavour may have been diverted from making a quit attempt and quitting smoking by the daily reassurance from this sensory experience that their smoking was doing less damage to their health. In this sense, the apparent distinction between switching for flavour and switching for harm reduction begins to blur. However, smokers who switched only for flavour were about as likely as those who switched solely for harm reduction to make a quit attempt, but were less likely to have recently quit, suggesting that these motives reflect different dynamics.
Switching solely to smoke a less harmful cigarette did not appear to be associated with a change in odds of making a quit attempt or recent quitting: smokers who stated they switched for harm reduction were just as likely as non-switchers to attempt quitting and just as likely to succeed when they tried. This finding is in line with some prior research,24 25 but in contrast to other studies16 19–21 demonstrating less successful quitting among individuals who switched to lighter cigarettes for harm reduction. The survey question on switching asked in the current study differs slightly from most prior surveys, which asked only about switching for harm reduction,19 potentially leading to misclassification of reasons for switching.
If indeed smokers who switch in order to reduce their health risk are not thereby deterred from making quit attempts, or undermined in quitting success, then some of the concerns that have been expressed36 about the unintended consequences of introducing “safer” cigarettes might be misplaced. However, this study is not capable of definitively answering this question, which requires further study.
Our study suggests that switching for various reasons has a complex relationship to quitting, sometimes being associated with higher likelihood of making quit attempts, even while being associated with lower rates of success in those attempts. Overarching that complexity is the finding that, for most smokers, switching is associated with lower chances of abstinence and overall lower rates of smoking cessation.
Our study has several important strengths. The 2003 Tobacco Use Supplement is a large, nationally representative sample of US smokers, featuring a clear, two-part question on switching that assessed reasons for switching and ensured that the act of switching was not trivial or short lived. The methods of probing reasons for switching allowed for multiple combinations and showed that smokers tend to switch for combinations of reasons. This may have important implications when counselling smokers, including openly discussing people's intended reasons for switching and the potential consequences for smoking cessation.
The Tobacco Use Supplement lacked complete data on attitudes and beliefs toward “lighter” cigarettes (these questions were only asked of current, but not of former smokers). Finally, the design of the study, which lacked precise information on timing of switching, precludes definitive interpretation of the relationship between switching and subsequent smoking cessation as causal. It is even possible that the experience of quitting could have altered the participants' recall of reasons for switching. This study shares this limitation of retrospective data with almost all the studies on “lights”. The one study to prospectively examine switching and subsequent quitting was in turn limited by small sample size, constraining its ability to address the impact of switching on quitting.24
In conclusion, compared to non-switchers, switchers were more likely to make a quit attempt between 2002 and 2003, but were less likely to have recently quit smoking during this period. The magnitude of this association varied by reason for switching. Our analyses support the position that switching should be discouraged and that switching as a step towards quitting should be particularly discouraged. As neither “light” cigarettes nor any other kind of cigarettes have demonstrated convincing clinical evidence of reduced harm, smokers should be encouraged to stop smoking, rather than just change what cigarettes they smoke.
The authors gratefully acknowledge the critique provided by Mitchell Zeller on an earlier version of this manuscript, as well as the statistical expertise of Dr Charity Moore during revision. HAT, JEB and AMH had full access to all of the data in the study and take responsibility for its integrity and the accuracy of the analysis.
Funding: This publication was made possible by a grant from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research (KL2 RR024154-03 to HAT). Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NCRR or NIH. Information on the NCRR is available at http://www.ncrr.nih.gov/. Information on re-engineering the clinical research enterprise can be obtained from http://nihroadmap.nih.gov/clinicalresearch/overview-translational.asp.
Competing interests: None declared.
Ethics approval: These data were deemed exempt by the University of Pittsburgh Institutional Review Board (0610085).
Provenance and peer review: Not commissioned; externally peer reviewed.