Rostron B. Menthol Cigarette Use and Stroke Risk Among US SmokersA Critical Reappraisal. JAMA Intern Med. 2014;174(5):808-809. doi:10.1001/jamainternmed.2013.9600
Copyright 2014 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Vozoris1 recently reported that menthol cigarette use was associated with increased stroke risk compared with nonmenthol cigarette use among US smokers (odds ratio, 2.25; 95% CI, 1.33-2.78). These results, however, were not consistent across all demographic groups. For example, no increased risk was observed among African American smokers, a group with a high menthol smoking prevalence. I recently reported2 that menthol cigarette use was associated with lower lung cancer mortality among US smokers, although I found no difference in mortality risk for other causes. This study reexamines stroke risk among US menthol smokers using national health survey data and mortality follow-up.
I used data from the 1999 through 2010 National Health and Nutrition Examination Survey (NHANES),3 a nationally representative health survey of the US civilian noninstitutionalized population conducted by the National Center for Health Statistics. It includes a health interview as well as a physical examination and collection of biospecimens. Approximately 10 000 individuals participate in the NHANES every 2 years. I analyzed data for NHANES current smokers 20 years and older using logistic regression analysis to examine the association between having been diagnosed with a stroke by a health professional and menthol smoking. I included pack years of smoking, sex, age, race and ethnicity (using the NHANES race and ethnicity categories), educational attainment, ratio of family income to poverty threshold, use of other tobacco products, and body mass index as control variables in the analysis. Of the 7055 NHANES participants who reported that they were current smokers, 5745 had information for all regression variables, including menthol cigarette use (600 smokers did not have family income information and 294 were missing menthol information). In total, 1765 smokers were identified as menthol smokers and 3980 as nonmenthol smokers. Of these smokers, 195 reported having had a stroke. Stroke prevalence was 3.4% (95% CI, 2.9%-4.0%) among nonmenthol smokers and 3.3% (95% CI, 2.6%-4.3%) among menthol smokers. Stroke prevalence was similar among all NHANES smokers at 3.5% (95% CI, 3.0%-4.1%) for nonmenthol smokers and 3.8% (95% CI, 3.0%-4.7%) for menthol smokers. I conducted the analyses using the appropriate NHANES sample weights, taking into account the NHANES complex sample design.
I also analyzed stroke mortality risk using 1987 National Health Interview Survey (NHIS) Cancer Control Supplement data linked for mortality follow-up. The NHIS4 is an annual questionnaire-based health survey of the US civilian noninstitutionalized population. Approximately 22 000 NHIS participants 18 years and older completed the Cancer Control Supplement in 1987. These individuals were observed for mortality through the end of 2006 by linking their records to the National Death Index, which is maintained by the National Center for Health Statistics and contains death certificate information for all US decedents since 1979.5 I conducted a survival analysis of current smokers at baseline by menthol cigarette use, using a Cox proportional hazards regression model. I controlled for sex, age, race and ethnicity, educational attainment, family income, and use of other tobacco products. Of the 4832 smokers who had information for all variables, 1356 were identified as menthol smokers and 3476 as nonmenthol smokers. Among these individuals, 1221 deaths were ascertained through linkage with the National Death Index, of which 56 were caused by stroke (International Classification of Diseases, Tenth Revision, codes I60-I69).
All NHANES and NHIS data used in this analysis have been released for public use by the National Center for Health Statistics, so institutional review board approval was not needed or obtained.
The Table presents adjusted odds ratios for stroke for NHANES menthol smokers compared with nonmenthol smokers. No difference in risk was observed among smokers overall or among male or female smokers. Odds of stroke were lower for African American menthol smokers, but it is possible that this estimate was affected by a limited sample size and/or residual confounding. I also did not observe a difference in stroke mortality risk for menthol cigarette use among NHIS smokers (hazard ratio, 0.67; 95% CI, 0.34-1.33).
I found no evidence in national health survey data of a higher stroke risk for US menthol smokers compared with nonmenthol smokers. I examined an expanded set of NHANES data compared with that used by Vozoris1 (1999-2010 vs 2001-2008 participants) and controlled for a similar set of covariates. Results were consistent by sex, generally stable across race and ethnicity groups, and robust to alternative specifications for variables such as smoking exposure and income. Using NHIS data, I also found no evidence of higher stroke mortality associated with menthol cigarette use among smokers.
It is not clear to me how Vozoris1 obtained his findings, given that I cannot replicate his general results for stroke using the NHANES data and analyses that he specified. Moreover, the absence of observed differences in stroke prevalence among NHANES menthol smokers would suggest that methodological or analytical issues may have affected his results.
These findings are consistent with my research that found lower lung cancer mortality for US menthol smokers compared with nonmenthol smokers but no difference in mortality risk for other causes.2 Lower lung cancer risk for menthol smokers has been observed in previous studies6,7 and among certain groups in meta-analysis.8
Corresponding Author: Brian Rostron, PhD, 8000 Wildwood Dr, Takoma Park, MD 20912 (email@example.com).
Published Online: March 10, 2014. doi:10.1001/jamainternmed.2013.9600.
Conflict of Interest Disclosures: None reported.