1000 Words Rogerian Essay

doi:10.1093/ntr/nts014 Advance Access Published on March 1, 2012 Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco 2012.

Nicotine & Tobacco Research, Volume 14, Number 10 (October 2012) 1140–1144Nicotine & Tobacco Research

1

doi: 10.1093/ntr/nts014 Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco 2012.

and most have found no increased risk of lung cancer among menthol smokers. In fact, some studies have suggested that menthol smokers have decreased lung cancer risk compared with nonmenthol smokers. A meta-analysis of eight epidemio- logic studies on the subject ( Lee, 2011 ) found that menthol cigarette use was associated with a statistically signifi cant lower lung cancer risk for female smokers, although not for male smokers or for smokers overall. The meta-analysis also found a statistically signifi cant lower lung cancer risk for menthol smokers in studies published since 2001, although not in earlier studies. Etzel, Kachroo, Liu, D’Amelio, and Dong (2008) found a non- signifi cant decreased risk of lung cancer for menthol smokers among current smokers in a case – control study of Blacks . Blot, Cohen, McLaughlin, Hargreaves, and Signorello (2011) recently published a study of Southern Community Cohort Study data that found lower lung cancer incidence and mortality among current menthol smokers compared with current nonmenthol smokers. Their study participants, however, came from 12 South- ern states and 2/3 of the study participants were Blacks . Study participants also had 1 – 8 years of follow-up. In this study, we analyze lung cancer mortality risk for menthol and nonmenthol smokers using nationally representative National Health Interview Survey (NHIS) data and 20 years of mortality follow-up.

Methods We conducted a survival analysis of participants in the 1987 NHIS Cancer Control Supplement who were followed for mortality through linkage with the National Death Index (NDI). The NHIS is a nationally representative household health survey of the U.S. civilian noninstitutionalized population that is conducted by the National Center for Health Statistics (NCHS) ( Schoenbom & Boyd, 1989 ). The NDI is maintained by NCHS and contains the death certifi cate information for U.S. decedents since 1979 ( MacMahon, 1983 ). Twenty- two thousand and forty- three NHIS participants aged 18 and over completed the 1987 Cancer Control Supplement, which asked participants about their cancer risk factors, including smoking, and collected information about menthol cigarette smoking ( National Center for Health Statistics, 2011 ).

Six thousand and seventy-three NHIS participants reported being current smokers, and these smokers were asked to present

Abstract Introduction: The U.S. Food and Drug Administration is currently assessing the public health impact of menthol cigarettes. Results from a recent U.S. cohort study, composed largely of Blacks and limited to 12 Southern states, found that menthol cigarette smokers had lower risks of lung cancer incidence and mortality than nonmenthol smokers.

Methods: We conducted a survival analysis of current smokers from the 1987 National Health Interview Survey Cancer Control Supplement ( n = 4,832), followed for mortality through linkage with the National Death Index. We estimated mortality hazard ratios ( HR s) for menthol smokers compared with nonmenthol smokers, adjusting for a full set of demographic and smoking characteristics.

Results: The overall HR for lung cancer mortality for menthol smokers was 0.69 (95% CI = 0.45 – 1.06). The HR for lung cancer mortality for menthol smokers at ages 50 and over was 0.59 (95% CI = 0.37 – 0.95). All-cause mortality net of lung cancer mortality did not differ for menthol and nonmenthol smokers.

Conclusion: We found evidence of lower lung cancer mortali- ty risk among menthol smokers compared with nonmenthol smokers at ages 50 and over in the U.S. population. It is not known, however, if these differences are due to the impact of menthol on cigarette smoking or long-term differences in cigarette design between menthol and nonmenthol cigarettes.

Introduction Menthol is a common additive in cigarettes in the United States, and menthol cigarettes have been commonly identifi ed as being popular among young, Black , and female smokers ( Lawrence et al., 2010 ). Numerous studies have been conducted on lung cancer risk among menthol smokers compared with nonmenthol smokers ( Brooks, Palmer, Strom, & Rosenberg, 2003 ; Carpenter, Jarvik, Morgenstern, McCarthy, & London, 1999 ; Kabat & Hebert, 1991 ; Murray, 2007 ; Sidney, Takawa, Friedman, Sadler, & Tashkin, 1995 ; Stellman, Chen, Muscat, Djordjevic, & Richie, 2003 ) ,

Original Investigation

Lung Cancer Mortality Risk for U.S. Menthol Cigarette Smokers Brian Rostron , Ph.D.

Center for Tobacco Products, Food and Drug Administration, Rockville, MD

Corresponding Author: Brian Rostron, Ph.D., Center for Tobacco Products, Food and Drug Administration, 9200 Corporate Boulevard, Room 300G, Rockville, MD 20850, USA. Telephone: 301-796-9360; Fax: 240-276-3761; E-mail: [email protected]

Received September 16 , 2011 ; accepted January 19 , 2012

2

Lung cancer mortality risk

a pack of the cigarettes that they smoked most often to the inter- viewer or report whether the brand of cigarettes that they smoked most often was a menthol or nonmenthol ( “ plain ” ) cigarette. One thousand four hundred and seventeen of the current smokers were identifi ed as being menthol cigarette smokers, 3 , 690 were identifi ed as being nonmenthol cigarette smokers, and 966 had unknown menthol cigarette preference. We analyzed the demographic and smoking characteristics of these smokers. All analyses were conducted using R version 2.10.1 ( R Development Core Team, 2009 ) and the survey package ( Lumley, 2010 ), using the appropriate NHIS person - weights and taking into account the NHIS complex survey design.

We then conducted a survival analysis of these NHIS par- ticipants using data from the public-use National Health Interview Survey — Linked Mortality File (NHIS-LMF ; National Center for Health Statistics, 2010 ). The NHIS-LMF provides mortality follow-up for NHIS participants through the end of 2006 through linkage with the NDI . Some cause-of-death and date- of-death data in the NHIS-LMF have been perturbed to prevent identifi cation of decedents, but NCHS has shown that this data perturbation has little effect on results from data and survival analysis ( National Center for Health Statistics Data Linkage Team, 2010 ). We restricted the survival analysis to smokers with reported menthol or nonmenthol preference in order to specifi - cally identify any association between menthol smoking and lung cancer risk, although the inclusion of smokers with unknown menthol preference did not appreciably affect the results. Five thousand and sixty- seven current smokers with reported menthol preference were eligible for mortality follow-up. We used Cox proportional hazard models to estimate mortality hazard ratios ( HR s) for menthol smokers compared with nonmenthol smokers, controlling for sex, race/ethnicity (non-Hispanic W hite or “ W hite, ” non-Hispanic Black or “ Black , ” non-Hispanic other race or “ other ” , and Hispanic), educational attainment (<12 years, 12 years, and >12 years), family income (<$20,000 and ≥ $20,000), and current use of other tobacco products (chewing tobacco and snuff, cigars, and pipes). All demographic and smoking charac- teristics were reported at baseline. We also controlled for cumu- lative pack years of smoking, which we calculated by dividing the number of cigarettes reported smoked per day at baseline by 20 cigarettes per pack and then multiplying the resulting quan- tity by the number of years since the age at which the smoker reported beginning to smoke regularly. We obtained similar results when we included number of cigarettes smoked per day and age at which the smoker started smoking regularly as separate control variables in alternative regression models. Follow-up time was expressed as age, as recommended in the research literature ( Korn, Graubard, & Midthune, 1997 ). Four thousand eight hun- dred and thirty two of the relevant smokers had information for all of the variables and could be included in the survival analysis. One thousand two hundred and twenty one of these individuals were identifi ed as deceased in mortality follow-up. Two hundred and one of these decedents had malignant neoplasm of the lung, bronchus, or trachea (International Classifi cation of Diseases, 10th revision [ICD-10], C33 – C34) identifi ed on the death certifi - cate as the underlying cause of death. One hundred and sixty seven of these decedents had reported smoking nonmenthol ciga- rettes at baseline and 34 had reported smoking menthol ciga- rettes. s were estimated for smokers at all ages and for smokers at ages 50 and above, given that approximately 95% of all lung can- cer mortality in the United States occurs at these ages ( Xu, Ko-

chanek, Murphy, & Tejada-Vera, 2010 ) and because 96% of lung cancer deaths observed among these NHIS participants occurred at these ages. This approach has been used previously in research on mortality risk factors ( Fuller, 2011 ). The consistency of the HR s over follow-up time was evaluated by calculating HR s for the fi rst 10 years of follow-up and comparing them with the HR s for the entire period. The HR s for the two periods were similar, with somewhat lower ratios for menthol smokers over the fi rst 10 years. The proportional hazard assumption was also evaluated by calculating chi-square d tests of the scaled Schoenfeld residuals from the models as functions of follow-up time, and the proportionality assumption was upheld.

We also calculated HR s for all-cause mortality and all-cause mortality net of lung cancer mortality for these smokers.

Results Table 1 presents demographic and smoking characteristics for the NHIS current smokers by menthol cigarette use. Menthol smokers were on average younger than nonmenthol smokers and smoked fewer cigarettes per day. Female, Black , and Hispanic smokers were more likely to smoke menthol cigarettes than male and non-Hispanic W hite smokers. The mean age and ciga- rettes smoked per day for smokers with unknown menthol pref- erence fell between those of menthol and nonmenthol smokers. Prevalence of unknown menthol preference was fairly consistent across groups, except for Blacks , who were more likely than other groups to report a menthol preference.

Table 2 presents HR s obtained from survival analysis of the NHIS current menthol and nonmenthol smokers. The HR for menthol cigarette use for these smokers was 0.69 (95% CI = 0.45 – 1.06). In analyses restricted to subgroups (results not shown), the estimated for menthol smoking for female smokers ( HR = 0.70, 95% CI = 0.39 – 1.26) was essentially the same as the estimated for male smokers ( HR = 0.71, 95% CI = 0.42 – 1.20). The estimated was lower for Black smokers ( HR = 0.41, 95% CI = 0.15 – 1.09) than for W hite smokers ( HR = 0.70, 95% CI = 0.42 – 1.17), although the difference was not statistically signifi cant. In analysis restricted to smokers at ages 50 and over, the was 0.59 (95% CI = 0.37 – 0.95). The for all ages for the fi rst 10 years of mortality follow-up was 0.57 (95% CI = 0.32 – 1.03). Inclusion of survey participants with unknown menthol fl avor preference in the analysis did not affect the overall for menthol smoking ( HR = 0.70, 95% CI = 0.45 – 1.07), and there was no difference in lung cancer mortality for smokers with unknown menthol preference and nonmenthol smokers ( HR = 0.98, 95% CI = 0.69 – 1.40).

A similar association for menthol smoking was not observed with other causes of mortality. The HR for menthol smoking for all-cause mortality for all ages was 0.92 (95% CI = 0.81 – 1.06). The for menthol smoking for all-cause mortality net of lung cancer mortality was 0.97 (95% CI = 0.84 – 1.12).

Discussion Using nationally representative NHIS data, we have found a lower risk of lung cancer mortality at ages 50 and over for menthol smokers compared with non-menthol smokers over 20 years

1140

Nicotine & Tobacco Research, Volume 14, Number 10 (October 2012)Nicotine & Tobacco Research

1

doi: 10.1093/ntr/nts014 Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco 2012.

and most have found no increased risk of lung cancer among menthol smokers. In fact, some studies have suggested that menthol smokers have decreased lung cancer risk compared with nonmenthol smokers. A meta-analysis of eight epidemio- logic studies on the subject ( Lee, 2011 ) found that menthol cigarette use was associated with a statistically signifi cant lower lung cancer risk for female smokers, although not for male smokers or for smokers overall. The meta-analysis also found a statistically signifi cant lower lung cancer risk for menthol smokers in studies published since 2001, although not in earlier studies. Etzel, Kachroo, Liu, D’Amelio, and Dong (2008) found a non- signifi cant decreased risk of lung cancer for menthol smokers among current smokers in a case – control study of Blacks . Blot, Cohen, McLaughlin, Hargreaves, and Signorello (2011) recently published a study of Southern Community Cohort Study data that found lower lung cancer incidence and mortality among current menthol smokers compared with current nonmenthol smokers. Their study participants, however, came from 12 South- ern states and 2/3 of the study participants were Blacks . Study participants also had 1 – 8 years of follow-up. In this study, we analyze lung cancer mortality risk for menthol and nonmenthol smokers using nationally representative National Health Interview Survey (NHIS) data and 20 years of mortality follow-up.

Methods We conducted a survival analysis of participants in the 1987 NHIS Cancer Control Supplement who were followed for mortality through linkage with the National Death Index (NDI). The NHIS is a nationally representative household health survey of the U.S. civilian noninstitutionalized population that is conducted by the National Center for Health Statistics (NCHS) ( Schoenbom & Boyd, 1989 ). The NDI is maintained by NCHS and contains the death certifi cate information for U.S. decedents since 1979 ( MacMahon, 1983 ). Twenty- two thousand and forty- three NHIS participants aged 18 and over completed the 1987 Cancer Control Supplement, which asked participants about their cancer risk factors, including smoking, and collected information about menthol cigarette smoking ( National Center for Health Statistics, 2011 ).

Six thousand and seventy-three NHIS participants reported being current smokers, and these smokers were asked to present

Abstract Introduction: The U.S. Food and Drug Administration is currently assessing the public health impact of menthol cigarettes. Results from a recent U.S. cohort study, composed largely of Blacks and limited to 12 Southern states, found that menthol cigarette smokers had lower risks of lung cancer incidence and mortality than nonmenthol smokers.

Methods: We conducted a survival analysis of current smokers from the 1987 National Health Interview Survey Cancer Control Supplement ( n = 4,832), followed for mortality through linkage with the National Death Index. We estimated mortality hazard ratios ( HR s) for menthol smokers compared with nonmenthol smokers, adjusting for a full set of demographic and smoking characteristics.

Results: The overall HR for lung cancer mortality for menthol smokers was 0.69 (95% CI = 0.45 – 1.06). The HR for lung cancer mortality for menthol smokers at ages 50 and over was 0.59 (95% CI = 0.37 – 0.95). All-cause mortality net of lung cancer mortality did not differ for menthol and nonmenthol smokers.

Conclusion: We found evidence of lower lung cancer mortali- ty risk among menthol smokers compared with nonmenthol smokers at ages 50 and over in the U.S. population. It is not known, however, if these differences are due to the impact of menthol on cigarette smoking or long-term differences in cigarette design between menthol and nonmenthol cigarettes.

Introduction Menthol is a common additive in cigarettes in the United States, and menthol cigarettes have been commonly identifi ed as being popular among young, Black , and female smokers ( Lawrence et al., 2010 ). Numerous studies have been conducted on lung cancer risk among menthol smokers compared with nonmenthol smokers ( Brooks, Palmer, Strom, & Rosenberg, 2003 ; Carpenter, Jarvik, Morgenstern, McCarthy, & London, 1999 ; Kabat & Hebert, 1991 ; Murray, 2007 ; Sidney, Takawa, Friedman, Sadler, & Tashkin, 1995 ; Stellman, Chen, Muscat, Djordjevic, & Richie, 2003 ) ,

Original Investigation

Lung Cancer Mortality Risk for U.S. Menthol Cigarette Smokers Brian Rostron , Ph.D.

Center for Tobacco Products, Food and Drug Administration, Rockville, MD

Corresponding Author: Brian Rostron, Ph.D., Center for Tobacco Products, Food and Drug Administration, 9200 Corporate Boulevard, Room 300G, Rockville, MD 20850, USA. Telephone: 301-796-9360; Fax: 240-276-3761; E-mail: [email protected]

Received September 16 , 2011 ; accepted January 19 , 2012

2

Lung cancer mortality risk

a pack of the cigarettes that they smoked most often to the inter- viewer or report whether the brand of cigarettes that they smoked most often was a menthol or nonmenthol ( “ plain ” ) cigarette. One thousand four hundred and seventeen of the current smokers were identifi ed as being menthol cigarette smokers, 3 , 690 were identifi ed as being nonmenthol cigarette smokers, and 966 had unknown menthol cigarette preference. We analyzed the demographic and smoking characteristics of these smokers. All analyses were conducted using R version 2.10.1 ( R Development Core Team, 2009 ) and the survey package ( Lumley, 2010 ), using the appropriate NHIS person - weights and taking into account the NHIS complex survey design.

We then conducted a survival analysis of these NHIS par- ticipants using data from the public-use National Health Interview Survey — Linked Mortality File (NHIS-LMF ; National Center for Health Statistics, 2010 ). The NHIS-LMF provides mortality follow-up for NHIS participants through the end of 2006 through linkage with the NDI . Some cause-of-death and date- of-death data in the NHIS-LMF have been perturbed to prevent identifi cation of decedents, but NCHS has shown that this data perturbation has little effect on results from data and survival analysis ( National Center for Health Statistics Data Linkage Team, 2010 ). We restricted the survival analysis to smokers with reported menthol or nonmenthol preference in order to specifi - cally identify any association between menthol smoking and lung cancer risk, although the inclusion of smokers with unknown menthol preference did not appreciably affect the results. Five thousand and sixty- seven current smokers with reported menthol preference were eligible for mortality follow-up. We used Cox proportional hazard models to estimate mortality hazard ratios ( HR s) for menthol smokers compared with nonmenthol smokers, controlling for sex, race/ethnicity (non-Hispanic W hite or “ W hite, ” non-Hispanic Black or “ Black , ” non-Hispanic other race or “ other ” , and Hispanic), educational attainment (<12 years, 12 years, and >12 years), family income (<$20,000 and ≥ $20,000), and current use of other tobacco products (chewing tobacco and snuff, cigars, and pipes). All demographic and smoking charac- teristics were reported at baseline. We also controlled for cumu- lative pack years of smoking, which we calculated by dividing the number of cigarettes reported smoked per day at baseline by 20 cigarettes per pack and then multiplying the resulting quan- tity by the number of years since the age at which the smoker reported beginning to smoke regularly. We obtained similar results when we included number of cigarettes smoked per day and age at which the smoker started smoking regularly as separate control variables in alternative regression models. Follow-up time was expressed as age, as recommended in the research literature ( Korn, Graubard, & Midthune, 1997 ). Four thousand eight hun- dred and thirty two of the relevant smokers had information for all of the variables and could be included in the survival analysis. One thousand two hundred and twenty one of these individuals were identifi ed as deceased in mortality follow-up. Two hundred and one of these decedents had malignant neoplasm of the lung, bronchus, or trachea (International Classifi cation of Diseases, 10th revision [ICD-10], C33 – C34) identifi ed on the death certifi - cate as the underlying cause of death. One hundred and sixty seven of these decedents had reported smoking nonmenthol ciga- rettes at baseline and 34 had reported smoking menthol ciga- rettes. s were estimated for smokers at all ages and for smokers at ages 50 and above, given that approximately 95% of all lung can- cer mortality in the United States occurs at these ages ( Xu, Ko-

chanek, Murphy, & Tejada-Vera, 2010 ) and because 96% of lung cancer deaths observed among these NHIS participants occurred at these ages. This approach has been used previously in research on mortality risk factors ( Fuller, 2011 ). The consistency of the HR s over follow-up time was evaluated by calculating HR s for the fi rst 10 years of follow-up and comparing them with the HR s for the entire period. The HR s for the two periods were similar, with somewhat lower ratios for menthol smokers over the fi rst 10 years. The proportional hazard assumption was also evaluated by calculating chi-square d tests of the scaled Schoenfeld residuals from the models as functions of follow-up time, and the proportionality assumption was upheld.

We also calculated HR s for all-cause mortality and all-cause mortality net of lung cancer mortality for these smokers.

Results Table 1 presents demographic and smoking characteristics for the NHIS current smokers by menthol cigarette use. Menthol smokers were on average younger than nonmenthol smokers and smoked fewer cigarettes per day. Female, Black , and Hispanic smokers were more likely to smoke menthol cigarettes than male and non-Hispanic W hite smokers. The mean age and ciga- rettes smoked per day for smokers with unknown menthol pref- erence fell between those of menthol and nonmenthol smokers. Prevalence of unknown menthol preference was fairly consistent across groups, except for Blacks , who were more likely than other groups to report a menthol preference.

Table 2 presents HR s obtained from survival analysis of the NHIS current menthol and nonmenthol smokers. The HR for menthol cigarette use for these smokers was 0.69 (95% CI = 0.45 – 1.06). In analyses restricted to subgroups (results not shown), the estimated for menthol smoking for female smokers ( HR = 0.70, 95% CI = 0.39 – 1.26) was essentially the same as the estimated for male smokers ( HR = 0.71, 95% CI = 0.42 – 1.20). The estimated was lower for Black smokers ( HR = 0.41, 95% CI = 0.15 – 1.09) than for W hite smokers ( HR = 0.70, 95% CI = 0.42 – 1.17), although the difference was not statistically signifi cant. In analysis restricted to smokers at ages 50 and over, the was 0.59 (95% CI = 0.37 – 0.95). The for all ages for the fi rst 10 years of mortality follow-up was 0.57 (95% CI = 0.32 – 1.03). Inclusion of survey participants with unknown menthol fl avor preference in the analysis did not affect the overall for menthol smoking ( HR = 0.70, 95% CI = 0.45 – 1.07), and there was no difference in lung cancer mortality for smokers with unknown menthol preference and nonmenthol smokers ( HR = 0.98, 95% CI = 0.69 – 1.40).

A similar association for menthol smoking was not observed with other causes of mortality. The HR for menthol smoking for all-cause mortality for all ages was 0.92 (95% CI = 0.81 – 1.06). The for menthol smoking for all-cause mortality net of lung cancer mortality was 0.97 (95% CI = 0.84 – 1.12).

Discussion Using nationally representative NHIS data, we have found a lower risk of lung cancer mortality at ages 50 and over for menthol smokers compared with non-menthol smokers over 20 years

1141

Lung cancer mortality risk

3

Nicotine & Tobacco Research

Table 1. Demographic and S moking C haracteristics for C urrent S mokers by M enthol S tatus and M enthol P revalence by D emographic C haracteristics: 1987 National Health Interview Survey (NHIS) Cancer Control Supplement P articipants

Characteristic

Current menthol smokers ( n = 1,417) mean ( CI )

Current nonmenthol smokers ( n = 3,690) mean ( CI )

Current smokers, menthol preference unknown ( n = 966) mean ( CI )

Age 38.6 (37.7, 39.5) 41.7 (41.1, 42.3) 39.8 (38.6, 41.0) Cigarettes per day 18.1 (17.4, 19.0) 20.7 (20.2, 21.2) 19.1 (18.2, 20.0) Deaths through December 31, 2006 All causes 306 993 241 Lung cancer 34 178 37 Lung cancer (ages 50+) 29 174 34

Group Flavor preference Menthol ( CI ) Nonmenthol ( CI ) Unknown ( CI ) Males ( n = 2,768), % 19.3 (17.1, 21.5) 63.3 (60.9, 65.6) 17.4 (15.5, 19.4) Females ( n = 3,305), % 24.6 (22.7, 26.5) 59.8 (57.8, 61.8) 15.5 (14.0, 17.0) Non-Hispanic Whites ( n = 4,712), % 17.0 (15.7, 18.3) 65.2 (63.6, 66.8) 17.8 (16.4, 19.2) Non-Hispanic Blacks ( n = 947), % 54.0 (49.5, 58.4) 38.6 (34.2, 42.9) 7.5 (5.1, 9.9) Non-Hispanic Other ( n = 94), % 16.1 (7.7, 24.5) 68.8 (57.3, 80.3) 15.1 (7.2, 23.0) Hispanic ( n = 320), % 24.3 (18.2, 30.4) 57.7 (51.8, 63.5) 18.0 (13.3, 22.7)

Note . All statistics were computed using the relevant NHIS person-weights.

of follow-up. These results agree with expectations that any association between lung cancer and menthol smoking would be greatest at ages in which smokers have smoked longer and accumulated more pack-years of smoking. The estimated HR s presented here are generally consistent overall and by subgroup with the estimated HR s reported previously by Blot et al. (2011) from a cohort study. Blot et al., for example, reported the same point estimate as this study for lung cancer mortality for current smokers in their analysis ( HR = 0.69, 95% CI = 0.45 – 0.95). Sim- ilar to this study, Blot et al. also found lower estimated odds ra- tios ( OR s) of lung cancer incidence associated with menthol smoking for Black smokers compared with W hite smokers

( OR = 0.52, 95% CI = 0.34 – 0.78 and OR = 0.84, 95% CI = 0.43 – 1.64, respectively). Blot et al. (2011) did fi nd lower estimated OR s for females smokers compared with male smokers ( OR = 0.43, 95% CI = 0.24 – 0.75 and OR = 0.77, 95% CI = 0.49 – 1.23, respec- tively), which was not similar to results found in this study.

These general results have also been suggested in previous research, as indicated by techniques such as meta-analysis. It is interesting to note that we and Blot et al. (2011) have found lower esti mated HR s and OR s for subgroups such as Blacks and, in the case of Blot et al., females, who tend to have higher menthol preva- lence among smokers than W hites and males generally. This trend

Table 2. Cox P roportional H azard M odel R esults for the A ssociation B etween S moking and Demographic Characteristics and Lung Cancer Mortality : 1987 National Health Interview Survey Cancer Control Supplement Current Smokers Followed for Mortality Through 2006

Predictor variable

All ages (18+), n = 4,832 Age 50+, n = 3,381

Hazard ratio 95% CI Hazard ratio 95% CI

Cigarette pack-years 1.01* (1.00, 1.01) 1.01* (1.00, 1.01) Menthol cigarette use (ref. = no) 0.69 (0.45, 1.06) 0.59* (0.37, 0.95) Sex (ref. = male) 0.80 (0.55, 1.17) 0.84 (0.58, 1.22) Race/ethnicity (ref. = non-Hispanic White) Black 1.28 (0.83, 1.97) 1.29 (0.84, 1.98) Other 1.00 (0.30, 3.41) 0.78 (0.21, 2.92) Hispanic 1.05 (0.45, 2.43) 1.05 (0.43, 2.56) Education (ref. = <12 years) 12 years 0.87 (0.58, 1.29) 0.86 (0.58, 1.28) >12 years 0.65 (0.41, 1.03) 0.65 (0.41, 1.05) Family income (ref. = <$20,000) 1.03 (0.72, 1.49) 1.08 (0.75, 1.57) Current other tobacco Use (ref. = no) 1.33 (0.79, 2.24) 1.26 (0.74, 2.14)

Note . Ref. = reference. Values for all predictor variables were reported at baseline in 1987. * p < .05.

4

Lung cancer mortality risk

may result from subgroups with higher menthol prevalence among smokers also generally having greater consistency in menthol smoking, which would tend to explain the estimates obtained for these groups. This explanation may also help to explain why similar results were not necessarily observed in earlier studies. Menthol smoking increased signifi cantly among smokers in the 1960s and 1970s ( Lee, 2011 ). Any association between menthol smoking and lower lung cancer risk may have therefore been attenuated in stud- ies conducted in a period such as the 1990s, given that smokers reporting menthol smoking at baseline would have likely smoked a higher proportion of nonmenthol cigarettes in their smoking his- tories compared with self-reported menthol smokers in a later period. Interestingly, Lee (2011) found no association in a meta- analysis between menthol smoking and lung cancer risk in studies published between 1991 and 2000 ( relative r isk [RR] = 1.00, 95% CI = 0.86 – 1.15) but a lower risk for menthol smokers in studies published from 2001 to 2008 (RR = 0.88, 95% CI = 0.77 – 0.99).

The reasons for this association are not clear at the present time. One possible cause in this particular study could be greater smoking cessation among menthol smokers compared with non- menthol smokers during the follow-up period. Some research has suggested that menthol smokers have lower cessation than nonmenthol smokers ( Levy et al., 2011 ), especially among certain racial and ethnic subgroups and in studies carried out more recently ( Foulds, Hooper, Pletcher, & Okuyemi, 2010 ), but defi nitive evidence of such an association is lacking at the present time . Blot et al. (2011) , for example, did not fi nd a difference in smoking cessation among menthol and nonmenthol smokers during follow-up in the Southern Community Cohort Study. The existence of similar fi ndings regarding lung cancer and men- thol in studies in which menthol smoking status was ascertained retroactively ( Etzel et al., 2008 ) or prospectively with a much shorter follow-up period ( Blot et al., 2011 ) would tend to suggest that differences in smoking cessation cannot explain all of the observed differences in lung cancer mortality risk between menthol and nonmenthol smokers. The observed lower HR for menthol smokers during the first 10 years of follow-up in this study, when there were generally fewer former smokers, compared with the entire study also tends to support this inference.

Differences in the design of menthol and nonmenthol ciga- rettes may also have an effect on lung cancer risk. Researchers have shown an association between changes in cigarette design features and increases in lung cancer rates for smokers over time, especially for adenocarcinoma ( Burns, Anderson, & Gray, 2011 ). Cigarette ventilation is one possible cause of differences in risk for menthol cigarettes, given that the delivery of carcinogenic con- stituents can be altered by ventilation ( Hoffmann & Hoffmann, 1997 ). Smokers of ventilated cigarettes often inhale more deeply in order to ensure consistent nicotine delivery ( Hoffmann & Hoffmann, 1997 ). Popular menthol brands in the U.S. market often have little or no fi lter tip ventilation, particularly for their “ full-fl avor ” subbrands ( Kozlowski, Mehta, & Sweeney, 1997 ). As a result, differences in lung cancer mortality for menthol and nonmenthol smokers may be due in part to differences in inhalation caused by product design and not necessarily to menthol itself.

The public health implications of any decreased lung cancer risk for menthol smoking compared with nonmenthol smoking, if ever conclusively demonstrated, are inevitably problematic. Smoking of any kind of cigarette is known to profoundly harm

individual and population health, and the Centers for Disease Control and Prevention ( CDC ) estimates that smoking is responsi- ble for 443,000 deaths in the United States each year ( CDC, 2008 ). Further study is needed into possible explanations for the observed association and the public health impact of potential reasons for it. This research could potentially identify ways to decrease the individual risk of cigarettes.

The chief limitation of this study is the limited number of lung cancer deaths in the sample. We do not believe that this is a signifi cant limitation of our analysis or results, given that our results are consistent in direction and magnitude with those obtained previously in the research literature, including results from meta-analysis ( Lee, 2011 ) . We note that caution should be taken in interpreting results in our analysis for subgroups such as Blacks , for whom data are particularly limited at this time. Additional lung cancer deaths and person-time of exposure for the NHIS participants should be available in the future when NCHS links NHIS data to more recent NDI data.

Other limitations of this study result from the information available in the 1987 NHIS Cancer Control Supplement. We do not, for example, have information about the number of years that menthol and nonmenthol smokers had regularly smoked menthol cigarettes prior to 1987, although it is not clear a priori in which direction any bias caused by mixed menthol and non- menthol smoking would have on the observed association.

Funding The author is employed by the Food and Drug Administration .

Declaration of Interests None declared .

Acknowledgments The opinions expressed in this paper are solely those of the author and do not necessarily refl ect those of the Food and Drug Administration.

References Blot , W. J. , Cohen , S. S. , McLaughlin , J. K. , Hargreaves , M. K. , & Signorello , L. B. ( 2011 ). Lung cancer risk among smokers of menthol cigarettes . Journal of the National Cancer Institute , 103 , 1 – 7 . doi:10.1093/jnci/djr102

Brooks , D. R. , Palmer , J. R. , Strom , B. L. , & Rosenberg , L. ( 2003 ). Menthol cigarettes and risk of lung cancer . American Journal of Epidemiology , 158 , 609 – 616 . doi:10.1093/aje/kwg182

Burns , D. M. , Anderson , C. M. , & Gray , N. ( 2011 ). Do changes in cigarette design infl uence the rise of adenocarcinoma of the lung? Cancer Causes and Control , 22 , 13 – 22 . doi:10.1007/s10552- 010-9660-0

Carpenter , C. L. , Jarvik , M. E. , Morgenstern , H. , McCarthy , W. J. , & London , S. J. ( 1999 ). Mentholated cigarette smoking and

3

Nicotine & Tobacco Research

Table 1. Demographic and S moking C haracteristics for C urrent S mokers by M enthol S tatus and M enthol P revalence by D emographic C haracteristics: 1987 National Health Interview Survey (NHIS) Cancer Control Supplement P articipants

Characteristic

Current menthol smokers ( n = 1,417) mean ( CI )

Current nonmenthol smokers ( n = 3,690) mean ( CI )

Current smokers, menthol preference unknown ( n = 966) mean ( CI )

Age 38.6 (37.7, 39.5) 41.7 (41.1, 42.3) 39.8 (38.6, 41.0) Cigarettes per day 18.1 (17.4, 19.0) 20.7 (20.2, 21.2) 19.1 (18.2, 20.0) Deaths through December 31, 2006 All causes 306 993 241 Lung cancer 34 178 37 Lung cancer (ages 50+) 29 174 34

Group Flavor preference Menthol ( CI ) Nonmenthol ( CI ) Unknown ( CI ) Males ( n = 2,768), % 19.3 (17.1, 21.5) 63.3 (60.9, 65.6) 17.4 (15.5, 19.4) Females ( n = 3,305), % 24.6 (22.7, 26.5) 59.8 (57.8, 61.8) 15.5 (14.0, 17.0) Non-Hispanic Whites ( n = 4,712), % 17.0 (15.7, 18.3) 65.2 (63.6, 66.8) 17.8 (16.4, 19.2) Non-Hispanic Blacks ( n = 947), % 54.0 (49.5, 58.4) 38.6 (34.2, 42.9) 7.5 (5.1, 9.9) Non-Hispanic Other ( n = 94), % 16.1 (7.7, 24.5) 68.8 (57.3, 80.3) 15.1 (7.2, 23.0) Hispanic ( n = 320), % 24.3 (18.2, 30.4) 57.7 (51.8, 63.5) 18.0 (13.3, 22.7)

Note . All statistics were computed using the relevant NHIS person-weights.

of follow-up. These results agree with expectations that any association between lung cancer and menthol smoking would be greatest at ages in which smokers have smoked longer and accumulated more pack-years of smoking. The estimated HR s presented here are generally consistent overall and by subgroup with the estimated HR s reported previously by Blot et al. (2011) from a cohort study. Blot et al., for example, reported the same point estimate as this study for lung cancer mortality for current smokers in their analysis ( HR = 0.69, 95% CI = 0.45 – 0.95). Sim- ilar to this study, Blot et al. also found lower estimated odds ra- tios ( OR s) of lung cancer incidence associated with menthol smoking for Black smokers compared with W hite smokers

( OR = 0.52, 95% CI = 0.34 – 0.78 and OR = 0.84, 95% CI = 0.43 – 1.64, respectively). Blot et al. (2011) did fi nd lower estimated OR s for females smokers compared with male smokers ( OR = 0.43, 95% CI = 0.24 – 0.75 and OR = 0.77, 95% CI = 0.49 – 1.23, respec- tively), which was not similar to results found in this study.

These general results have also been suggested in previous research, as indicated by techniques such as meta-analysis. It is interesting to note that we and Blot et al. (2011) have found lower esti mated HR s and OR s for subgroups such as Blacks and, in the case of Blot et al., females, who tend to have higher menthol preva- lence among smokers than W hites and males generally. This trend

Table 2. Cox P roportional H azard M odel R esults for the A ssociation B etween S moking and Demographic Characteristics and Lung Cancer Mortality : 1987 National Health Interview Survey Cancer Control Supplement Current Smokers Followed for Mortality Through 2006

Predictor variable

All ages (18+), n = 4,832 Age 50+, n = 3,381

Hazard ratio 95% CI Hazard ratio 95% CI

Cigarette pack-years 1.01* (1.00, 1.01) 1.01* (1.00, 1.01) Menthol cigarette use (ref. = no) 0.69 (0.45, 1.06) 0.59* (0.37, 0.95) Sex (ref. = male) 0.80 (0.55, 1.17) 0.84 (0.58, 1.22) Race/ethnicity (ref. = non-Hispanic White) Black 1.28 (0.83, 1.97) 1.29 (0.84, 1.98) Other 1.00 (0.30, 3.41) 0.78 (0.21, 2.92) Hispanic 1.05 (0.45, 2.43) 1.05 (0.43, 2.56) Education (ref. = <12 years) 12 years 0.87 (0.58, 1.29) 0.86 (0.58, 1.28) >12 years 0.65 (0.41, 1.03) 0.65 (0.41, 1.05) Family income (ref. = <$20,000) 1.03 (0.72, 1.49) 1.08 (0.75, 1.57) Current other tobacco Use (ref. = no) 1.33 (0.79, 2.24) 1.26 (0.74, 2.14)

Note . Ref. = reference. Values for all predictor variables were reported at baseline in 1987. * p < .05.

1142

Nicotine & Tobacco Research, Volume 14, Number 10 (October 2012)

3

Nicotine & Tobacco Research

Table 1. Demographic and S moking C haracteristics for C urrent S mokers by M enthol S tatus and M enthol P revalence by D emographic C haracteristics: 1987 National Health Interview Survey (NHIS) Cancer Control Supplement P articipants

Characteristic

Current menthol smokers ( n = 1,417) mean ( CI )

Current nonmenthol smokers ( n = 3,690) mean ( CI )

Current smokers, menthol preference unknown ( n = 966) mean ( CI )

Age 38.6 (37.7, 39.5) 41.7 (41.1, 42.3) 39.8 (38.6, 41.0) Cigarettes per day 18.1 (17.4, 19.0) 20.7 (20.2, 21.2) 19.1 (18.2, 20.0) Deaths through December 31, 2006 All causes 306 993 241 Lung cancer 34 178 37 Lung cancer (ages 50+) 29 174 34

Group Flavor preference Menthol ( CI ) Nonmenthol ( CI ) Unknown ( CI ) Males ( n = 2,768), % 19.3 (17.1, 21.5) 63.3 (60.9, 65.6) 17.4 (15.5, 19.4) Females ( n = 3,305), % 24.6 (22.7, 26.5) 59.8 (57.8, 61.8) 15.5 (14.0, 17.0) Non-Hispanic Whites ( n = 4,712), % 17.0 (15.7, 18.3) 65.2 (63.6, 66.8) 17.8 (16.4, 19.2) Non-Hispanic Blacks ( n = 947), % 54.0 (49.5, 58.4) 38.6 (34.2, 42.9) 7.5 (5.1, 9.9) Non-Hispanic Other ( n = 94), % 16.1 (7.7, 24.5) 68.8 (57.3, 80.3) 15.1 (7.2, 23.0) Hispanic ( n = 320), % 24.3 (18.2, 30.4) 57.7 (51.8, 63.5) 18.0 (13.3, 22.7)

Note . All statistics were computed using the relevant NHIS person-weights.

of follow-up. These results agree with expectations that any association between lung cancer and menthol smoking would be greatest at ages in which smokers have smoked longer and accumulated more pack-years of smoking. The estimated HR s presented here are generally consistent overall and by subgroup with the estimated HR s reported previously by Blot et al. (2011) from a cohort study. Blot et al., for example, reported the same point estimate as this study for lung cancer mortality for current smokers in their analysis ( HR = 0.69, 95% CI = 0.45 – 0.95). Sim- ilar to this study, Blot et al. also found lower estimated odds ra- tios ( OR s) of lung cancer incidence associated with menthol smoking for Black smokers compared with W hite smokers

( OR = 0.52, 95% CI = 0.34 – 0.78 and OR = 0.84, 95% CI = 0.43 – 1.64, respectively). Blot et al. (2011) did fi nd lower estimated OR s for females smokers compared with male smokers ( OR = 0.43, 95% CI = 0.24 – 0.75 and OR = 0.77, 95% CI = 0.49 – 1.23, respec- tively), which was not similar to results found in this study.

These general results have also been suggested in previous research, as indicated by techniques such as meta-analysis. It is interesting to note that we and Blot et al. (2011) have found lower esti mated HR s and OR s for subgroups such as Blacks and, in the case of Blot et al., females, who tend to have higher menthol preva- lence among smokers than W hites and males generally. This trend

Table 2. Cox P roportional H azard M odel R esults for the A ssociation B etween S moking and Demographic Characteristics and Lung Cancer Mortality : 1987 National Health Interview Survey Cancer Control Supplement Current Smokers Followed for Mortality Through 2006

Predictor variable

All ages (18+), n = 4,832 Age 50+, n = 3,381

Hazard ratio 95% CI Hazard ratio 95% CI

Cigarette pack-years 1.01* (1.00, 1.01) 1.01* (1.00, 1.01) Menthol cigarette use (ref. = no) 0.69 (0.45, 1.06) 0.59* (0.37, 0.95) Sex (ref. = male) 0.80 (0.55, 1.17) 0.84 (0.58, 1.22) Race/ethnicity (ref. = non-Hispanic White) Black 1.28 (0.83, 1.97) 1.29 (0.84, 1.98) Other 1.00 (0.30, 3.41) 0.78 (0.21, 2.92) Hispanic 1.05 (0.45, 2.43) 1.05 (0.43, 2.56) Education (ref. = <12 years) 12 years 0.87 (0.58, 1.29) 0.86 (0.58, 1.28) >12 years 0.65 (0.41, 1.03) 0.65 (0.41, 1.05) Family income (ref. = <$20,000) 1.03 (0.72, 1.49) 1.08 (0.75, 1.57) Current other tobacco Use (ref. = no) 1.33 (0.79, 2.24) 1.26 (0.74, 2.14)

Note . Ref. = reference. Values for all predictor variables were reported at baseline in 1987. * p < .05.

4

Lung cancer mortality risk

may result from subgroups with higher menthol prevalence among smokers also generally having greater consistency in menthol smoking, which would tend to explain the estimates obtained for these groups. This explanation may also help to explain why similar results were not necessarily observed in earlier studies. Menthol smoking increased signifi cantly among smokers in the 1960s and 1970s ( Lee, 2011 ). Any association between menthol smoking and lower lung cancer risk may have therefore been attenuated in stud- ies conducted in a period such as the 1990s, given that smokers reporting menthol smoking at baseline would have likely smoked a higher proportion of nonmenthol cigarettes in their smoking his- tories compared with self-reported menthol smokers in a later period. Interestingly, Lee (2011) found no association in a meta- analysis between menthol smoking and lung cancer risk in studies published between 1991 and 2000 ( relative r isk [RR] = 1.00, 95% CI = 0.86 – 1.15) but a lower risk for menthol smokers in studies published from 2001 to 2008 (RR = 0.88, 95% CI = 0.77 – 0.99).

The reasons for this association are not clear at the present time. One possible cause in this particular study could be greater smoking cessation among menthol smokers compared with non- menthol smokers during the follow-up period. Some research has suggested that menthol smokers have lower cessation than nonmenthol smokers ( Levy et al., 2011 ), especially among certain racial and ethnic subgroups and in studies carried out more recently ( Foulds, Hooper, Pletcher, & Okuyemi, 2010 ), but defi nitive evidence of such an association is lacking at the present time . Blot et al. (2011) , for example, did not fi nd a difference in smoking cessation among menthol and nonmenthol smokers during follow-up in the Southern Community Cohort Study. The existence of similar fi ndings regarding lung cancer and men- thol in studies in which menthol smoking status was ascertained retroactively ( Etzel et al., 2008 ) or prospectively with a much shorter follow-up period ( Blot et al., 2011 ) would tend to suggest that differences in smoking cessation cannot explain all of the observed differences in lung cancer mortality risk between menthol and nonmenthol smokers. The observed lower HR for menthol smokers during the first 10 years of follow-up in this study, when there were generally fewer former smokers, compared with the entire study also tends to support this inference.

Differences in the design of menthol and nonmenthol ciga- rettes may also have an effect on lung cancer risk. Researchers have shown an association between changes in cigarette design features and increases in lung cancer rates for smokers over time, especially for adenocarcinoma ( Burns, Anderson, & Gray, 2011 ). Cigarette ventilation is one possible cause of differences in risk for menthol cigarettes, given that the delivery of carcinogenic con- stituents can be altered by ventilation ( Hoffmann & Hoffmann, 1997 ). Smokers of ventilated cigarettes often inhale more deeply in order to ensure consistent nicotine delivery ( Hoffmann & Hoffmann, 1997 ). Popular menthol brands in the U.S. market often have little or no fi lter tip ventilation, particularly for their “ full-fl avor ” subbrands ( Kozlowski, Mehta, & Sweeney, 1997 ). As a result, differences in lung cancer mortality for menthol and nonmenthol smokers may be due in part to differences in inhalation caused by product design and not necessarily to menthol itself.

The public health implications of any decreased lung cancer risk for menthol smoking compared with nonmenthol smoking, if ever conclusively demonstrated, are inevitably problematic. Smoking of any kind of cigarette is known to profoundly harm

individual and population health, and the Centers for Disease Control and Prevention ( CDC ) estimates that smoking is responsi- ble for 443,000 deaths in the United States each year ( CDC, 2008 ). Further study is needed into possible explanations for the observed association and the public health impact of potential reasons for it. This research could potentially identify ways to decrease the individual risk of cigarettes.

The chief limitation of this study is the limited number of lung cancer deaths in the sample. We do not believe that this is a signifi cant limitation of our analysis or results, given that our results are consistent in direction and magnitude with those obtained previously in the research literature, including results from meta-analysis ( Lee, 2011 ) . We note that caution should be taken in interpreting results in our analysis for subgroups such as Blacks , for whom data are particularly limited at this time. Additional lung cancer deaths and person-time of exposure for the NHIS participants should be available in the future when NCHS links NHIS data to more recent NDI data.

Other limitations of this study result from the information available in the 1987 NHIS Cancer Control Supplement. We do not, for example, have information about the number of years that menthol and nonmenthol smokers had regularly smoked menthol cigarettes prior to 1987, although it is not clear a priori in which direction any bias caused by mixed menthol and non- menthol smoking would have on the observed association.

Funding The author is employed by the Food and Drug Administration .

Declaration of Interests None declared .

Acknowledgments The opinions expressed in this paper are solely those of the author and do not necessarily refl ect those of the Food and Drug Administration.

References Blot , W. J. , Cohen , S. S. , McLaughlin , J. K. , Hargreaves , M. K. , & Signorello , L. B. ( 2011 ). Lung cancer risk among smokers of menthol cigarettes . Journal of the National Cancer Institute , 103 , 1 – 7 . doi:10.1093/jnci/djr102

Brooks , D. R. , Palmer , J. R. , Strom , B. L. , & Rosenberg , L. ( 2003 ). Menthol cigarettes and risk of lung cancer . American Journal of Epidemiology , 158 , 609 – 616 . doi:10.1093/aje/kwg182

Burns , D. M. , Anderson , C. M. , & Gray , N. ( 2011 ). Do changes in cigarette design infl uence the rise of adenocarcinoma of the lung? Cancer Causes and Control , 22 , 13 – 22 . doi:10.1007/s10552- 010-9660-0

Carpenter , C. L. , Jarvik , M. E. , Morgenstern , H. , McCarthy , W. J. , & London , S. J. ( 1999 ). Mentholated cigarette smoking and

3

Nicotine & Tobacco Research

Table 1. Demographic and S moking C haracteristics for C urrent S mokers by M enthol S tatus and M enthol P revalence by D emographic C haracteristics: 1987 National Health Interview Survey (NHIS) Cancer Control Supplement P articipants

Characteristic

Current menthol smokers ( n = 1,417) mean ( CI )

Current nonmenthol smokers ( n = 3,690) mean ( CI )

Current smokers, menthol preference unknown ( n = 966) mean ( CI )

Age 38.6 (37.7, 39.5) 41.7 (41.1, 42.3) 39.8 (38.6, 41.0) Cigarettes per day 18.1 (17.4, 19.0) 20.7 (20.2, 21.2) 19.1 (18.2, 20.0) Deaths through December 31, 2006 All causes 306 993 241 Lung cancer 34 178 37 Lung cancer (ages 50+) 29 174 34

Group Flavor preference Menthol ( CI ) Nonmenthol ( CI ) Unknown ( CI ) Males ( n = 2,768), % 19.3 (17.1, 21.5) 63.3 (60.9, 65.6) 17.4 (15.5, 19.4) Females ( n = 3,305), % 24.6 (22.7, 26.5) 59.8 (57.8, 61.8) 15.5 (14.0, 17.0) Non-Hispanic Whites ( n = 4,712), % 17.0 (15.7, 18.3) 65.2 (63.6, 66.8) 17.8 (16.4, 19.2) Non-Hispanic Blacks ( n = 947), % 54.0 (49.5, 58.4) 38.6 (34.2, 42.9) 7.5 (5.1, 9.9) Non-Hispanic Other ( n = 94), % 16.1 (7.7, 24.5) 68.8 (57.3, 80.3) 15.1 (7.2, 23.0) Hispanic ( n = 320), % 24.3 (18.2, 30.4) 57.7 (51.8, 63.5) 18.0 (13.3, 22.7)

Note . All statistics were computed using the relevant NHIS person-weights.

of follow-up. These results agree with expectations that any association between lung cancer and menthol smoking would be greatest at ages in which smokers have smoked longer and accumulated more pack-years of smoking. The estimated HR s presented here are generally consistent overall and by subgroup with the estimated HR s reported previously by Blot et al. (2011) from a cohort study. Blot et al., for example, reported the same point estimate as this study for lung cancer mortality for current smokers in their analysis ( HR = 0.69, 95% CI = 0.45 – 0.95). Sim- ilar to this study, Blot et al. also found lower estimated odds ra- tios ( OR s) of lung cancer incidence associated with menthol smoking for Black smokers compared with W hite smokers

( OR = 0.52, 95% CI = 0.34 – 0.78 and OR = 0.84, 95% CI = 0.43 – 1.64, respectively). Blot et al. (2011) did fi nd lower estimated OR s for females smokers compared with male smokers ( OR = 0.43, 95% CI = 0.24 – 0.75 and OR = 0.77, 95% CI = 0.49 – 1.23, respec- tively), which was not similar to results found in this study.

These general results have also been suggested in previous research, as indicated by techniques such as meta-analysis. It is interesting to note that we and Blot et al. (2011) have found lower esti mated HR s and OR s for subgroups such as Blacks and, in the case of Blot et al., females, who tend to have higher menthol preva- lence among smokers than W hites and males generally. This trend

Table 2. Cox P roportional H azard M odel R esults for the A ssociation B etween S moking and Demographic Characteristics and Lung Cancer Mortality : 1987 National Health Interview Survey Cancer Control Supplement Current Smokers Followed for Mortality Through 2006

Predictor variable

All ages (18+), n = 4,832 Age 50+, n = 3,381

Hazard ratio 95% CI Hazard ratio 95% CI

Cigarette pack-years 1.01* (1.00, 1.01) 1.01* (1.00, 1.01) Menthol cigarette use (ref. = no) 0.69 (0.45, 1.06) 0.59* (0.37, 0.95) Sex (ref. = male) 0.80 (0.55, 1.17) 0.84 (0.58, 1.22) Race/ethnicity (ref. = non-Hispanic White) Black 1.28 (0.83, 1.97) 1.29 (0.84, 1.98) Other 1.00 (0.30, 3.41) 0.78 (0.21, 2.92) Hispanic 1.05 (0.45, 2.43) 1.05 (0.43, 2.56) Education (ref. = <12 years) 12 years 0.87 (0.58, 1.29) 0.86 (0.58, 1.28) >12 years 0.65 (0.41, 1.03) 0.65 (0.41, 1.05) Family income (ref. = <$20,000) 1.03 (0.72, 1.49) 1.08 (0.75, 1.57) Current other tobacco Use (ref. = no) 1.33 (0.79, 2.24) 1.26 (0.74, 2.14)

Note . Ref. = reference. Values for all predictor variables were reported at baseline in 1987. * p < .05.

1143

Lung cancer mortality risk

5

Nicotine & Tobacco Research

lung-cancer risk . Annals of Epidemiology , 9 , 114 – 120 . doi:10.1016/ S1047-2797(98)00042-8

Centers for Disease Control and Prevention . ( 2008 ). Smoking- attributable mortality, years of potential life lost, and productivity losses — United States, 2000 – 2004 . Morbidity & Mortality Weekly Report , 57 , 1226 – 1228 .

Etzel , C. J. , Kachroo , S. , Liu , M. , D’Amelio , A. , & Dong , Q. ( 2008 ). Development and validation of a lung cancer risk prediction model for African-Americans . Cancer Prevention Research , 1 , 255 – 265 . doi:10.1158/1940-6207.CAPR-08-0082

Foulds , J. , Hooper , M. W. , T. D. , Pletcher , M. J. , & Okuyemi , K. S. ( 2010 ) Do smokers of menthol cigarettes fi nd it harder to quit smoking? Nicotine & Tobacco Research , 12 , S102 – S109 . doi:10. 1093/ntr/ntq166

Fuller , T. D. ( 2011 ). Moderate alcohol consumption and the risk of mortality . Demography , 48 , 1105 – 1125 . doi:10.1007/ s13524-011-0035-2

Hoffmann , D. , & Hoffmann , I. ( 1997 ). The changing cigarette, 1950 – 1995 . Journal of Toxicology and Environmental Health , 50 , 307 – 364 . doi:10.1006/pmed.1997.0183

Kabat , G. C. , & Hebert , J. R. ( 1991 ). Use of mentholated cigarettes and lung cancer risk . Cancer Research , 51 , 6510 – 6513 .

Korn , E. L. , Graubard , B. I. , & Midthune , D. ( 1997 ). Time-to-event analysis of longitudinal follow-up of a survey: Choice of the time-scale . American Journal of Epidemiology , 145 , 72 – 80 .

Kozlowski , L. T. , Mehta , N. Y. , & Sweeney , C. T. ( 1997 ). Filter ventilation levels in selected U.S. cigarettes, 1997 . Morbidity & Mortality Weekly Report , 46 , 1047 .

Lawrence , D. , Rose , A. , Fagan , P. , Moolchan , E. T. , Gibson , J. T. , & Backinger , C. L. ( 2010 ). National patterns and correlates of mentholated cigarette use in the United States . Addiction , 105 , 13 – 31 . doi:10.1111/j.1360-0443.2010.03203.x

Lee , P. N. ( 2011 ). Systematic review of the epidemiological evidence comparing lung cancer risk in smokers of mentholated and unmentholated cigarettes . BMC Pulmonary Medicine , 11 , 18 . doi:10.1186/1471-2466-11-18

Levy , D. T. , Blackman , K. , Tauras , J. , Chaloupka , F. J. , Villanti , A. C. , Niaura , R. S. , et al. ( 2011 ). Quit attempts and quit rates

among menthol and nonmenthol smokers in the United States . American Journal of Public Health , 101 , 1241 – 1247 . doi:10.2105/ AJPH.2011.300178

Lumley , T. ( 2010 ). Survey: Analysis of complex survey samples (version 3.23) . Seattle, WA : Thomas Lumley .

MacMahon , B. ( 1983 ). The National Death Index . American Journal of Public Health , 73 , 1247 – 1248 . doi:10.2105/AJPH. 73.11.1247

Murray , R. P. ( 2007 ). Menthol cigarettes and health risks in Lung Health Study data . Nicotine & Tobacco Research , 9 , 101 – 107 . doi:10.1080/14622200601078418

National Center for Health Statistics . ( 2010 ). NHIS Linked Mortality Files . Retrieved from http :// www . cdc . gov / nchs / data_ access / data_linkage / mortality / nhis_linkage . htm

National Center for Health Statistics . ( 2011 ). National Health Interview Survey . Retrieved from http :// www . cdc . gov / nchs / nhis . htm

National Center for Health Statistics Data Linkage Team . ( 2010 ). Comparative Analysis of the NHIS Public-use and Restricted-use Linked Mortality Files: 2010 Data Release . Retrieved from http :// www . cdc . gov / NCHS / data / datalinkage / nhis_mort_compare_2010_fi nal . pdf

R Development Core Team . ( 2009 ). R: A language and environment for statistical computing (version 2.10.1) . Vienna, Austria : Author .

Schoenbom , C. R. , & Boyd , G. M. ( 1989 ). Smoking and other tobacco use: United States, 1987 . Vital and Health Statistics , 10 , 1 – 79 .

Sidney , S. , Takawa , I. S. , Friedman , G. D. , Sadler , M. C. , & Tashkin , D. P. ( 1995 ). Mentholated cigarette use and lung cancer . Archives of Internal Medicine , 155 , 727 – 732 . doi:10.1001/archinte.155. 7.727

Stellman , S. D. , Chen , Y. , Muscat , J. E. , Djordjevic , M. V. , & Richie , J. P. ( 2003 ). Lung cancer risk in white and black Americans . Annals of Epidemiology , 13 , 294 – 302 . doi:10.1016/S1047-2797 (02)00420-9

Xu , J. , Kochanek , K. D. , Murphy , S. L. , & Tejada-Vera , B. ( 2010 ). Deaths: Final data for 2007 . National Vital Statistics Report , 58 , 1 – 135 .

1144

Copyright of Nicotine & Tobacco Research is the property of Oxford University Press / USA and its content

may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express

written permission. However, users may print, download, or email articles for individual use.