Smokers of Menthol and Nonmenthol Cigarettes Exhibit Similar Levels of Biomarkers of Smoke Exposure

Cigarette smoking causes cancers of the respiratory tract and other sites, and is similarly associated with increased risks of cardiovascular disease, chronic obstructive pulmonary disease, and other detriments to health (1, 2). Although all conventional cigarettes are known to elevate chronic disease risks, there has long been considerable interest among regulatory authorities, public health bodies, and cigarette manufacturers in determining whether differences in cigarette design or composition may affect the magnitude of smoking-related exposures and disease risks (3, 4). This interest includes the development of knowledge with regard to the potential of cigarette product design elements to affect the manner in which cigarettes are smoked. An added cigarette ingredient or a product design feature that significantly affects human smoking behavior could conceivably affect (positively or negatively) the intensity or cumulative magnitude of exposures to smoke constituents and, consequently, associated chronic disease risks.

The development and refinement of biomarkers of smoke constituent exposure from blood, exhaled breath, and excreted urine offers opportunities to assess smokers' exposures in an approach that offers a promising potential to characterize the contribution of both the behavioral and the cigarette product-related elements of this familiar but complex exposure scenario (5). Whereas additional refinement and validation of such methods for this application are indicated, analyses of novel or existing smoke constituent biomarkers may also permit the evaluation of modified cigarette products that may have reduced risks to continuing smokers who are unwilling to quit (4-6). Cigarette smoke exposure biomarkers have also been applied to the task of comparing relative exposures that result from different types of commercial cigarettes designed to produce different smoke “tar” yields when assessed by standardized, validated mechanical smoking methods (7-9).

Menthol has a long history of uneventful, safe use in mint-flavored foods and confections, as well as in oral and topical consumer products. However, menthol is perhaps most familiar as a flavoring employed in manufactured cigarettes that are preferred by approximately 30% of smokers in the United States (10). Menthol is arguably the best-studied among flavoring materials that have traditionally been used in manufactured cigarettes around the world, having been the subject of dozens of smoke chemistry, toxicology, and epidemiology investigations. A 2003 conference organized by the National Cancer Institute and the Centers for Disease Control identified a number of additional research priorities to address gaps in knowledge with regard to the potential of cigarette mentholation to affect the manner in which cigarettes are smoked, with consequent potential to affect the exposure of smokers to cigarette smoke constituents and their attending risks to develop smoking-related illnesses (11).

Menthol is unique among tobacco flavoring ingredients by virtue of having been the subject of a number of epidemiologic studies intended to compare the risks for cancer and other smoking-related diseases between smokers of mentholated and nonmentholated cigarettes (12-22). The overwhelming weight of evidence provided by these studies indicates that menthol cigarettes do not convey additional risks to smokers relative to nonmentholated cigarettes (23, 24). Nevertheless, speculation persists that menthol employed as a cigarette flavoring may play some role in differential disease risks among smokers who prefer mentholated brands. In particular, speculation that menthol may somehow account for the high incidence of some smoking-related diseases experienced by black smokers, notably males, has arisen in light of the stronger preference for mentholated brands reported for this demographic group in the United States (25, 26).

The present study was undertaken to determine whether moderately heavy smokers of regular or mentholated cigarette brands of the popular “lights” product category [i.e. cigarettes yielding approximately 7-15 mg Federal Trade Commission (FTC) “tar”; ref. 27], exhibit differences in biomarkers of smoke constituent exposure. Subjects smoked specified brands yielding ∼9 to 10 mg FTC tar in their normal, ad libitum fashion throughout the course of the parallel-arm study, and provided blood and 24-hour urine samples on two occasions one week apart for determination of any evidence of differences in smoke exposure. The evaluated biomarkers of smoke exposure included blood carboxyhemoglobin, a measure of carbon monoxide exposure from smoking and other environmental exposures (28); six major free and glucuronide-conjugated urinary nicotine metabolites accounting for a substantial majority of total nicotine exposure (29); and total urinary 4-(N-nitrosomethylamino)-1-(3-pyridyl)-1-butanol (NNAL), a measure of exposure to the tobacco-specific nitrosamine 4-(methylnitrosamino)-I-(3-pyridyl)-1-butanone (NNK). NNK has been classified by the IARC as class 1, carcinogenic to humans (30), and total urinary NNAL, comprising the sum of free NNAL and its glucuronide conjugate (NNAL-Gluc), has been previously described as a sensitive and specific biomarker suitable for application in investigations of smoking exposure and carcinogenesis (6, 31).

This study of smoke exposure biomarkers complements available epidemiologic data on menthol cigarettes and contributes to a limited body of work that has sought to characterize such exposures of smokers in their natural daily environments rather than in the laboratory. We believe that the present study of 54 menthol and 58 nonmenthol cigarette smokers is the largest to date comparing biomarkers of exposure resulting from cigarettes of otherwise similar smoke yield, and the first to have included concurrent smoke chemistry analyses for those cigarettes.

The experiment was a parallel-arm study in which subjects identified as smokers of either menthol or nonmenthol cigarettes were supplied with specified brands of cigarettes and continued smoking ad libitum in their normal manner throughout the 1-wk study interval. The subjects' stated preference for menthol or nonmenthol cigarettes of the popular lights product category (∼9-10 mg FTC tar; Table 1) was a primary consideration in initial screening and no forced switching between cigarette types was requested or required. A 2-wk prestudy acclimation period preceded the study to allow the menthol cigarette–preferring smokers to become accustomed to a single specified menthol cigarette brand. Collections of 24-h urine and a blood sample were completed on the first day of the study. A second 24-h urine collection and blood sampling was done on the seventh day of the study to complete the experimental phase. Smoking rates were monitored by collection of cigarette butts during the two 24-h intervals preceding the blood and urine collections (study days 1 and 7).

One hundred twelve male or female subjects 24 to 70 y of age, having a minimum smoking history of 3 pack-years, and reporting consumption of ≥15 menthol or nonmenthol cigarettes daily for the past year, were recruited through newspaper advertisements and telephone solicitation (Piedmont Medical Research Laboratories). Study participants had stated no current plans for smoking cessation at preliminary interview, but all subjects were advised at study initiation and upon completion that the best thing for their health is to quit smoking. Research participants expressing a desire to quit smoking either during the study or following study completion were offered reimbursement for either nicotine replacement therapy or enrollment in a smoking cessation program. The menthol cigarette study group comprised 31 white and 23 black subjects, and the nonmenthol group comprised 53 white and 5 black subjects. Participants who continued through qualification, screening visits, a 2-wk designated cigarette acclimatization period, cigarette butt collections, and two clinical visits for blood and urinary biomarker determinations received monetary reimbursement. Demographic characteristics are further detailed in Table 2.

Health status was verified through medical screening. Exclusion criteria included use of any investigational drugs during the study period or within 1 mo of the screening visit; evidence of alcohol abuse or use of any illicit drugs; or a history of cancer, various respiratory or cardiovascular diseases, asthma, or diabetes. Pregnant or nursing subjects were also excluded, as were those reporting consumption of grapefruit or grapefruit juice.

The study protocol was reviewed and approved by an independent Institutional Review Board (Quorum Review). Written, informed participant consent was obtained, and applicable Good Clinical Practices Guidelines were followed.

A commercially available menthol cigarette brand and several nonmentholated brands of similar smoke yield were provided to participants during the week of the study and for a 2-wk acclimation period preceding the first clinical visit. Samples from the cigarette lots employed in the study were characterized for smoke yields of tar, nicotine, and carbon monoxide by standardized FTC protocols (32). Briefly, cigarettes were conditioned and mechanically smoked (35 mL puffs of 2-s duration drawn once per minute) for gravimetric and gas chromatographic determinations of nicotine and tar yield. Carbon monoxide yields were determined from nondispersive IR absorption spectrometry on collected mainstream smoke gas phase samples (Arista Laboratories). The same independent laboratory did analyses of smoke NNK yields by liquid chromatographic separation of collected smoke particulate and quantification by tandem quadrupole mass spectrometry. Menthol was determined in smoke particulate material by a standard method of analysis by the Lorillard Tobacco Company Analytical Chemistry group. Briefly, smoke particulate matter collected on a Cambridge glass fiber filter under FTC smoking conditions is extracted and analyzed for menthol with a gas chromatograph equipped with a flame ionization detector.

Study participants collected and returned their expended cigarette butts during the two 24-h urine collection intervals as a general confirmation of compliance with the study protocol and usage of designated cigarette brands. Four 4-mL tubes of blood were collected for carboxyhemoglobin determination after 2:30 pm on the days of the two clinic visits, as previous studies had indicated that smokers' carboxyhemoglobin levels typically rise throughout the day to approximate a steady-state condition in late afternoon (33). The two 24-h urine collections were retained on ice by study subjects and frozen for delivery of a 4-mL aliquot to an independent laboratory for smoke exposure biomarkers analysis.

Carboxyhemoglobin was determined with an Instrumentation Laboratory 682 CO-Oximeter, calibrated daily (Piedmont Medical Research Laboratories). Urinary nicotine, cotinine, and four other nicotine metabolites (nicotine-N'-glucuronide, cotinine-N'-glucuronide, trans-3′-hydroxycotinine, trans-3′-hydroxycotinine-O-glucuronide) were determined following solid-phase extraction of frozen (−70°C) aliquots of collected 24-h urine and liquid chromatography-tandem mass spectrometry with methods developed and validated by Covance Laboratories (34). Glucuronide conjugates were determined by liberation of the respective nicotine metabolite aglycones by enzymatic deconjugation with β−glucuronidase (from Helix pomatia). Free, total and glucuronide conjugates of urinary NNAL were determined by liquid chromatography-tandem mass spectrometry with liberation of conjugated NNAL with Escherichia coli β−glucuronidase (Covance Laboratories; ref. 35). Urinary metabolite levels were normalized to creatinine to account for daily and intersubject differences in urinary output volume. Urinary creatinine was measured in the Hitachi 917, using an enzymatic method based upon the established determination of sarcosine after the conversion of creatinine with the aid of creatinase and sarcosine oxidase. The Roche Creatinine Plus Reagent Kit was used. This method permits a precise and specific measurement of urine creatinine.

An independent third party statistical analysis of the study data was done for measurements taken at clinical visits 1 and 2. Combined findings from both visits are reported here, with primary comparisons of interest being the biomarker values for smokers of mentholated versus nonmentholated cigarettes, with discussion of additional analyses by sex and race.

All statistical analyses were done using SAS 9.1 (SAS Institute). Frequencies were cross-tabulated and Fisher's exact tests were used to calculate X² values for differences between variables by group (menthol and nonmenthol, race, gender, and visit). General linear model analyses of covariance were done examining the effects of gender and race, visit, and group on biomarkers. All biomarker response variables were natural log-transformed (ln) to reduce the effect of skewness and outliers. Group comparisons were based on differences between least squares means estimates from the mixed model. All P values were two-tailed, with P < 0.05 considered significant.

The yields of FTC tar, nicotine, and carbon monoxide, as well as mainstream smoke NNK and menthol delivery by the menthol test cigarette, are presented in Table 1. The menthol and nonmenthol cigarettes employed in the study were very similar in tar, nicotine, and carbon monoxide yield. NNK smoke yields were somewhat more variable, with the lowest value of 44.7 ng/cigarette determined for nonmenthol cigarette #5 and the highest value of 79.8 ng/cigarette NNK determined for nonmenthol cigarette # 2. These NNK values are all within the range of mainstream smoke NNK yields typically reported for contemporary cigarettes of the lights category. The majority of NNK found in cigarette smoke is believed to be formed during the tobacco curing process, so the differences in NNK yields seen here and elsewhere are likely primarily a reflection of differences in tobacco blends among brands. The menthol cigarette employed in this study yielded a mean of 0.34 mg menthol per cigarette under FTC laboratory smoking conditions, which is typical of such mentholated products and similar to previously reported mainstream smoke menthol delivery values.

Study participants' demographic and physiologic characteristics and cigarette consumption are presented in Table 2. Cigarette butt collections during the two 24-hour study intervals indicated generally similar, moderately heavy smoking rates (25-29 cigarettes per day) and similar adherence to cigarette brand assignments among the compared groups. Urine output volumes and creatinine excretion recorded during the two collection intervals were broadly similar, while displaying typical interindividual variation.

Balanced distribution by age, race, and sex for the mentholated and nonmentholated cigarette study groups was pursued through extensive recruitment, qualification, and screening efforts. However, recruitment of black subjects preferring nonmenthol cigarettes proved to be problematic, consistent with prior survey data indicating substantial race-associated differences in taste preferences for mentholated cigarettes between white and black smokers (26). A χ² test was done for each demographic variable to determine if there was statistical evidence of group differences. There was no statistical evidence that the smoker groups differed by sex (P = 0.81) or age (P = 0.86). However, there was significant evidence of a difference due to race (23 black subjects in the menthol group and 5 in the nonmenthol groups, P < 0.01).

No statistically significant differences in blood carboxyhemoglobin levels were seen in comparisons between smokers of mentholated and nonmentholated cigarettes (Table 3). The pooled menthol cigarette smoking subjects (both clinical visits, males and females; blacks and whites) exhibited a slightly lower blood carboxyhemoglobin than did the nonmenthol smokers (5.9% versus 6.5%; P = 0.13). Separate analyses by race (Table 4) indicated that white smokers exhibited small, statistically insignificant trends toward higher carboxyhemoglobin levels from nonmenthol cigarettes and lower levels from menthol cigarettes than were found among the respective black smoker groups.

No substantive differences in excreted nicotine metabolites per mL urine, normalized by urinary creatinine, were evident in comparisons between smokers of the mentholated and nonmentholated cigarettes (Table 3). Nor were significant differences in total urinary nicotine equivalents (the sum of six metabolites) evident in 24-hour totals for the two cigarette type groups. Separate analyses by race (Table 4) gave some indications of differences (black menthol smokers and white nonmenthol smokers having slightly higher levels of nicotine metabolites), but these differences were not statistically significant. These small differences may have been attributable to the modest differences in cigarettes per day consumed by these respective groups (Table 2) and do not suggest any independent effects of menthol on nicotine intake.

Urinary concentrations of NNAL and its glucuronide were somewhat lower (difference not statistically significant) in the pooled menthol cigarette–smoking subjects than in their nonmenthol cigarette group counterparts, with or without adjustment for urinary creatinine concentration (Table 3). Pooled subjects' median creatinine-corrected total NNAL (NNAL plus NNAL-glucuronide) was also lower for the menthol cigarette group than for the nonmenthol group (286 versus 437 pg total NNAL/mg creatinine, respectively; Table 3). The menthol cigarette smokers similarly had lower total daily urinary excretion of NNAL than did the nonmenthol cigarette group (300 versus 419 ng/24-hour urine; Table 3). Although not reaching statistical significance, the present observation of lower total NNAL excretion by smokers of mentholated cigarettes relative to that of nonmentholated cigarette smokers is consistent with prior work (36) that did not support the hypothesis that mentholated cigarette smoking results in greater absorption of tobacco smoke toxins. Indeed, the present findings suggest a trend toward lower exposures among menthol smokers, generally. Our observation of modestly, but consistently lower urinary NNAL, blood carboxyhemoglobin, and some urinary nicotine metabolite values for menthol cigarette smokers was somewhat surprising given the similarity of the machine-generated smoke yields of the study cigarettes. Separate biomarker analyses for black and white smokers (Table 4) indicated that values for the menthol cigarette groups tended to be lower for whites, significantly so for nicotine-glucuronide (P < 0.01), cotinine-glucuronide (P < 0.001), NNAL-glucuronide (P < 0.01), and total NNAL (P < 0.01).

An individual smoker's risks for the development of chronic detriments to health that arise from cigarette smoking are likely influenced by the interplay of the chemistry of the mainstream smoke aerosol and individual smoking behavior variables in addition to dietary, genetic, and collateral environmental factors (2, 5). To the extent that cigarette design features may affect smoke chemistry or smoking behaviors such as the individual smoker's puffing topography (inhalation depth and volume, percent of cigarette consumed, and frequency of puffing), different cigarettes may result in quantitatively or qualitatively different smoke constituent exposures, and potentially different risks for disease (9, 37).

The findings of prior smoking topography and biomarker studies comparing mentholated and nonmentholated cigarettes have been inconsistent; such studies are rendered difficult by smokers' taste preferences with regard to menthol and the limitations of current methods to measure smoking topography (37-40). Studies of smoke exposure biomarkers in the blood and urine of smokers under their normal, ad libitum smoking conditions reflect the combined effects of cigarette composition, smoking topography, and cigarette smoking rate without intrusive laboratory instrumentation or forced-puffing protocols that do not closely resemble real-world smoking behaviors. This study provided participants with commercial cigarettes closely matched for machine-measured yields of tar, nicotine, and carbon monoxide in an attempt to detect any independent effects of the presence or absence of menthol on smoking behavior or smoke constituent exposure as indicated by the measured blood and urine smoking biomarkers.

The estimate of 0.625 mg menthol delivery to the smoker per cigarette smoked that was developed by Benowitz and colleagues (36) is in fair agreement with an estimated systemic delivery that might be developed from the 0.34 mg menthol per cigarette smoke yield reported here from standardized machine smoking. The recent development of improved methods to measure the lung deposition and retention of individual cigarette smoke constituents in human smokers (41, 42) suggests that better estimates of both local and systemic exposures to menthol or to other smoke constituents of greater toxicologic interest may soon be attainable.

Total nicotine equivalents, calculated as the median of individual subjects' values for all six nicotine metabolites (both sexes, races, and clinical visits combined) yielded an estimate of moderately lower nicotine excretion by the menthol smokers than the nonmenthol cigarette smokers (Table 3). However, separate total nicotine equivalent calculations by race indicated slightly higher nicotine excretion by black smokers in the menthol group (Table 4). This difference was not statistically significant and may be at least partially attributable to the slightly higher daily smoking rate of 29 cigarettes per day for black menthol smokers relative to the 25 cigarettes per day smoked by the black nonmenthol smokers in this study. White smokers exhibited higher total urinary nicotine equivalents excretion by the nonmenthol cigarette smokers than the menthol smokers, despite having reported very similar daily cigarette smoking of 26 or 27 cigarettes per day.

Prior work has suggested that black smokers may experience higher nicotine exposures than whites at similar rates of daily smoking (43). This trend was not evident in the present study, possibly because the menthol and nonmenthol cigarettes employed in the study were closely matched in terms of mainstream smoke nicotine yield. This has not always been the case in some prior investigations. Whereas further research into black/white differences in smoking behavior, nicotine pharmacogenetics, and smoke constituent exposure may be warranted, the present findings suggest that such studies will be most informative if the smoke yields of the products are matched, or at least documented by concurrent smoke chemistry analysis for consideration in subsequent biomarkers determinations.

It has long been appreciated that smoke yield determinations by standardized mechanical smoking machine methods cannot account for the diversity of smoking behaviors manifested by individual smokers (44-46). However, chemical analysis of smoke constituents certainly has some value as a starting point in the comparative exposure assessment of different cigarettes (4, 9), at a minimum as a qualitative approach to identifying the smoke constituents likely to contribute most significantly to disease risks.

Because the study reported here was intended to explore the potential of cigarette mentholation to independently affect smoke exposure biomarkers by influencing smoking behavior, smoke constituent uptake, or other means, the commercial cigarettes employed in the study were closely matched in terms of machine-generated smoke yields (Table 1). Most different brands of cigarettes vary in tobacco blend, construction, filter efficiency, and other design parameters, but the presence of menthol in one of the cigarettes employed in this study is the major distinction between that test cigarette and the other nonmenthol cigarettes of similar smoke yield.

The assessment of the possible effects of cigarette mentholation on smokers' exposures independent of demographic and pharmacogenetic factors is hampered by two realities: The first of these is the previously reported and substantial difference in menthol cigarette taste preference between black and white smokers [approximately 80% of black smokers prefer menthol brands versus 25% of white smokers (26)]. Secondly, it seems that there may be measurable differences in metabolic pathways for key smoking-related biomarkers between black and white smokers (43). These realities have constrained prior efforts to assess menthol as an independent variable in smoking biomarker work (8). Indeed, significant difficulty in recruiting racially balanced study groups was also encountered in the present investigation.

The present findings do not indicate any substantive differences between relatively heavy smokers of menthol and nonmenthol cigarettes delivering ∼10 mg FTC tar in terms of biomarkers of exposure to the smoke constituents carbon monoxide, nicotine, and NNK. To the extent that these biomarkers of the gas/vapor and particulate phases of cigarette smoke may be indicative of concomitant exposure to these and other smoke constituents that may be involved in the etiology of smoking-related diseases, these findings do not suggest that menthol cigarettes convey risks to the smoker any greater than those of otherwise similar nonmenthol cigarettes. However, as the participants in this study were relatively heavy smokers having average consumption of 25 to 29 cigarettes daily (Table 2), there could be differences in biomarkers of exposure at lower daily rates of smoking that may not have been apparent among these subjects due to the saturation of processes affecting excreted biomarker levels. Such saturation could have particular significance for biomarker studies in black smokers, who typically smoke fewer cigarettes per day than do their white counterparts (10). However, a recently described model of lung cancer risk among black smokers considered both smoking intensity and duration and provided indications that the risk attending menthol cigarette smoking is similar or perhaps even marginally lower than is the risk from nonmentholated cigarettes (22). Further investigation into biomarker-based indices of exposures associated with cigarette mentholation could include exploration of the influence of subjects' daily smoking rate as well as consideration of cigarettes having very different machine-measured smoke yields. The present findings developed for relatively heavy smokers of menthol and nonmenthol cigarettes of similar and relatively low machine-generated smoke yields may not necessarily be predictive across the spectrum of commercial cigarette smoke yields and widely different cigarette smoking rates.

We believe that the present study is the largest to date reporting biomarkers of exposure in human subjects smoking menthol and nonmenthol cigarettes of similar machine-generated smoke yield (FTC smoking protocol) under realistic, ad libitum conditions. It also seems to be the first study to have reported concurrently generated smoke constituent yield data for the cigarettes consumed by the study participants.

In summary, the present blood (carboxyhemoglobin) and urine (total nicotine equivalents, total NNAL) biomarkers findings are not consistent with speculation that smokers of mentholated cigarettes may exhibit an increased smoking intensity or smoke constituent exposures relative to those smoking nonmentholated cigarettes of broadly similar design and FTC smoke yield. The present biomarkers findings are consistent with the weight of extant epidemiologic evidence (22, 24) and prior biomarkers studies (36, 40) suggesting that exposures and consequent risks attending the smoking of mentholated cigarettes are not substantially different from those of nonmentholated cigarettes.

J.D. Heck: Lorillard Tobacco Company employee.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

I am indebted to Drs. Melissa Hagan Hughes and Jack R. Reid for statistical audits, Dr. Kevin R. Reinert for insightful discussion of the study findings, and William Pruitt for data quality assurance. The statistical analyses provided by Dr. Scott Richter of the University of North Carolina at Greensboro are also greatly appreciated.