![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Departments of
* Social and Preventive Medicine, School of Public Health and Health Professions and
Department of Medicine and the
** School of Social Work, University at Buffalo, State University of New York, Buffalo, NY, 14260;
Department of International Health, Johns Hopkins School of Public Health, Baltimore, MD, 21205;
Lombardi Cancer Center, Georgetown University, Washington, D.C. 20057-1421;
The Cancer Institute of New Jersey, UMDNJ-Robert Wood Johnson Medical School, New Brunswick, NJ 08903; and
# Prevention Research Center, Pacific Institute for Research and Evaluation, Berkeley, CA 94704
3To whom correspondence should be addressed. E-mail: jfreuden{at}buffalo.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT SUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
KEY WORDS: • alcohol consumption • alcohol dehydrogenase • epidemiology • humans • lung cancer
Lung cancer is the leading cause of death from cancer in the world for men and the second leading cause for women (1). Although there is clear evidence that the primary risk factor for lung cancer is tobacco exposure, there remains variation in disease risk that is not understood. One possible additional factor contributing to risk of this disease is alcohol consumption, particularly beer. Since 1984 when Potter and McMichael (2) first proposed that alcohol may affect the risk of lung cancer, there have been a number of studies examining this question with inconsistent results [reviewed in (3)]. However, because smoking is such a strong risk factor for lung cancer and because smoking and alcohol are correlated behaviors, it has been difficult to identify an independent contribution of alcohol, if any, to disease etiology. Further, few studies have examined both lifetime and more recent alcohol consumption in relation to risk. Additionally, there is evidence that other factors may affect the relationship between alcohol consumption and risk. In particular, there is some evidence that vitamin A intake may modify the alcohol-lung cancer relation (4–6).
One possible way to better understand etiologic associations is to examine genetic factors in relation to risk. If indeed alcohol were important etiologically, one might expect that there would be differences in the relation of alcohol and risk depending on genetic differences in alcohol metabolism. Further, understanding of genetic variation may be important in understanding etiologic relations in that heterogeneity within a study population may mask effects of an exposure.
Alcohol dehydrogenase (ADH), a rate-limiting enzyme in alcohol metabolism, catalyzes the oxidation of ethanol to acetaldehyde. There are six ADH genes designated as ADH1, ADH2, and so on; class I (ADH1 - 3) are those most involved in alcohol metabolism. There is in vitro evidence that a variant in the ADH3 gene has functional importance (7). There are several studies that have shown that risk associated with alcohol exposure may by modified by genotype of ADH3 for several chronic diseases (8–12) including cancers of the breast (8), mouth and pharynx (10,11); those with the variant that codes for a more efficient metabolism of alcohol to acetaldehyde may be at increased risk at similar levels of intake. The gene for cytochrome p4502E1 is also polymorphic and plays a role in alcohol metabolism. However, this pathway is limited to consumption of high intakes of alcohol.
We conducted a case-control study of lifetime alcohol consumption and lung cancer. We report here on the results from that study, i.e., associations of recent and lifetime consumption of alcoholic beverages with risk, and of the interaction of alcohol consumption with ADH3 genotype with risk of lung cancer in men and women in the western portion of New York State in the United States.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Controls between 35 and 65 y were randomly selected from a list of those holding a New York State driver’s license and residing in Erie and Niagara Counties; those 
65 y were randomly selected from the rolls of the Health Care Finance Administration. This study was conducted in conjunction with two other case-control studies, one of myocardial infarction and one of breast cancer, utilizing a common set of controls. Controls were selected to be frequency-matched for age, sex and race for the complete set of cases for all of the studies.
Interviews.
Interviews were conducted by trained interviewers and included both questionnaires filled out by participants and a computer-assisted personal interview. The interview lasted 
2.5 h and included information regarding medical history, diet, physical activity and social network.
Methodology for assessment of alcohol consumption was described previously (13). Participants reported on recent intake for the period 12 mo before diagnosis for cases and 12–24 mo before interview for controls. Participants defined their usual drink size for each beverage they drank using calibrated models and reported on frequency of consumption and amount drunk per drinking occasion. To obtain information on lifetime alcohol intake, we used the Cognitive Lifetime Drinking History (CLDH). Participants reported how old they were when they started drinking alcohol at least once a month for 6 mo and at what ages their drinking patterns changed. This information was used to define intervals during their lives when drinking patterns were fairly homogeneous. Starting from the time that they began drinking, for each interval, participants were asked about quantity and frequency of drinking alcoholic beverages. Beverage specific data on drink size and proportion of drinks as beer, wine or hard liquor, together with data on the number of drinks consumed in a typical 28-d period were used to calculate beverage-specific amounts of ethanol consumed in each interval. These estimates were summed across intervals to arrive at a lifetime total both for specific beverages and for all alcohol. Test-retest reliability of this questionnaire has been shown to be good. Pearson’s correlation were 0.84 for two administrations of the questionnaire for intake in the previous 12- to 24-mo period (14) and 0.74–0.85 for two administrations of the questionnaire regarding lifetime intake (13).
Lifetime smoking histories were also ascertained. Participants were asked if they had ever smoked at least 100 cigarettes, 20 cigars or 20 pipes in their lifetime. Those smoking less were considered never-smokers. Ever-smokers were asked when they had started smoking and for each decade of life when they smoked, information was obtained regarding the amount smoked per day and any periods during that decade when they stopped smoking. In addition, there was detailed information collected regarding passive smoke exposure. We calculated a variable of passive smoke exposure at home based on report of the number of people who smoked who lived with the participant and the number of years each smoker lived with them, summed for their lifetime. Lifetime total exposure to secondhand smoke at work was calculated for each job held on the basis of the hours per week of exposure to secondhand smoke and the number of years in that job. Other sources of passive smoke exposure were calculated by summing, for each decade, the frequency of 1) going to a restaurant/bar; 2) going to other social gatherings, such as dances and parties; and 3) going to other settings such as meetings, barber shop, indoor sporting events, where there was a lot of smoke. These occasions were then summed across the lifetime.
The diet history questionnaire was assessed using the food-frequency questionnaire developed by the National Cancer Institute (15,16), modified to include a few additional foods commonly consumed in the region. Diet questions referred to usual diet 12–24 mo before diagnosis for cases or before interview for controls. BMI was based on participant report of height and weight 1 y ago and calculated as weight (kg)/height2 (m2).
To assess bias related to nonparticipation, we conducted a short telephone interview with both those participating and those not participating (for whom we had permission to contact), at the time of contact. Questions included information on alcohol consumption.
Molecular genetic analyses.
All participants were asked to provide a blood sample. Because there was a larger group of controls than required and to preserve the resources of blood specimens, genotype was determined for a subset of controls. Two control series were generated, each with two controls matched to each case. One series was matched for gender, age (±5 y), race, smoking status (current, never, past smokers) and pack-years of smoking (±5 pack y). The second series was matched for gender, age (same year of birth) and race. (For the first group, because of the larger number of matching variables, the matching criteria for age were relaxed.) All genetic analyses were conducted at Georgetown University at the laboratory of Dr. Peter Shields. As previously described (10), using the method of Groppi et al. (17) in modified form, a 145-bp fragment including the polymorphism was amplified by PCR on extracted DNA. Before the PCR, highly homologous ADH1 and ADH2 genes were digested with the NlaIII restriction enzyme. This digestion mixture was then subjected to PCR and subsequent SspI enzymatic digestion to reveal ADH3 genotype (i.e., ADH31–1, ADH31–2 or ADH32–2). Positive and negative controls were included with each set of 14 samples. Cases and corresponding controls were assayed together; laboratory personnel had no knowledge of the case-control status of the samples. All results were scored independently by two readers, both unaware of identifying data including case-control status. For quality control, 20% of samples were repeated. In addition, duplicate samples were included among the study samples; these samples were not identifiable by laboratory personnel.
Statistical analyses.
Means and SD were calculated for continuous study variables. Two-sided comparisons of cases and controls were calculated within gender groups with Student’s t test. Categorical variables were compared with the 
2 test. Odds ratios (OR) and 95% CI were calculated with logistic regression, adjusting for age, years of education, race, BMI (kg/m2), vegetable intake (g/d), fruit intake (g/d), energy intake excluding energy from alcohol, and smoking and passive smoking.
Smoking adjustment was by packs per year smoked and total years smoked. In analyses of smoking and lung cancer that included different combinations of variables including pack years, years smoked, number of cigarettes smoked and smoking status, we found that this combination of variables was the one that explained the largest proportion of the variance in risk. We also controlled for lifetime passive smoke exposure at home, at work and in other settings. Analyses for the entire group were adjusted for gender. For analyses related to lifetime intake and to 12–24 mo ago, results for men and women were similar and CI overlapped almost completely; findings for the whole group are presented. Analyses for particular beverages were adjusted for intake of other alcohol (e.g., risk associated with beer intake was adjusted for wine and hard liquor consumption). Because of the exact matching of controls, the analyses of ADH3 were by conditional logistic regression. For these analyses, in addition to the matching variables, all analyses were adjusted for education, smoking, passive smoking and fruit and vegetable intake.
|
RESULTS |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
In addition to the expected differences in smoking for cases and controls, mean intakes of beer were higher for both male and female cases compared with controls, and wine intakes were higher for female controls (Table 1). Risks associated with lifetime alcohol consumption for three categories of consumption are shown (Table 2). Categories of consumption were: lifetime nondrinkers, and those below and above the median of intake for drinking controls, 3.5 drinks/wk. There was an indication of about a threefold increase in risk of lung cancer in the highest category of alcohol consumption compared with the lifetime nondrinkers in the crude analysis. When the confounders, particularly smoking, were accounted for, there was no indication that risk increased with increased total lifetime alcohol consumption (OR 1.13, 95% CI 0.47–2.72). For beverage-specific adjusted OR, there was no indication that risk was related to consumption of any of these beverages.
|
|
19 y did not differ from those drinking at an earlier age (OR 1.15, 95% CI 0.78–1.70) (data not shown).
|
Risks for lung cancer in categories of alcohol dehydrogenase genotype (ADH3) are shown (Table 4). Based on previous studies of this gene in relation to cancer of other sites, we had hypothesized that risk would be increased in the ADH31–1 group. However, we did not find any evidence of an elevation in risk for this group. For examination of both alcohol consumption and genotype (Table 5), there was no evidence at all of an increase in risk in high alcohol, ADH31–1 category for the set of controls matched for pack years of smoking. For the group matched only for gender, age and race in which smoking was controlled for in the analysis, there was some indication that risk increased in the high alcohol, ADH31–1 category; the CI was very wide and included the null (OR 2.57, 95% CI 0.82–8.00). The cut-off value for the two categories of alcohol was at the median for the controls, corresponding to very light drinking, a mean of <5 drinks/wk.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
70% in the highest category of beer consumption in the period 12–24 mo before the interview for controls and before diagnosis for cases. We also found some evidence of effect modification by vitamin A intake, i.e., higher levels of all alcohol and of beer were related to increased risk only among those who also had lower levels of vitamin A intake. There was no evidence that ADH31–1 genotype with alcohol consumption increased risk. It is important to note that in general, the intake of alcohol in this population was not heavy. The median intake of the controls in the upper category of lifetime total alcohol consumption was 3 drinks per drinking day or a mean of 1 drink/d. For beer consumption in the previous 12–24 mo, median intake for controls in the upper category was 1 beer/d. The generally low intakes in this population would preclude our finding an effect of heavier consumption on risk. The evidence for an association between alcohol consumption and risk of lung cancer was reviewed recently; there is some evidence that alcohol is related to risk, but results have been mixed (3). Although there are few available data, there is some epidemiologic and experimental evidence that an effect of alcohol on lung cancer, if any, would be in the later stages of tumor progression (18,19). We examined reports of alcohol consumption 12–24 mo before diagnosis/interview. Although we did not see unequivocal evidence of stronger associations with risk, evidence for an association with beer intake was stronger for more recent intake than for lifetime intake. However, this finding might be the result of multiple testing and should be interpreted carefully.
There have also been reports of differences in the effect of alcohol on risk depending on intake of vitamin A (4–6). We examined alcohol and lung cancer risk within strata of vitamin A; there was some evidence of increased risk, albeit characterized by an imprecise association, for total alcohol intake for men and beer intake in particular for women who were also in the lower half of vitamin A intakes. A possible mechanism for this finding could be that among those with a higher intake of vitamin A, the antioxidant properties of the vitamin can offset the oxidative effects of alcohol.
There are several possible mechanisms of an effect of alcohol on lung cancer risk. One of these would be effects of acetaldehyde on lung tissue. Acetaldehyde has been designated as a carcinogen (20). Evidence from in vitro systems indicates that there is a twofold difference in the Vmax of ADH31–1 compared with ADH32–2, with the ADH31–1 variant catalyzing alcohol to acetaldehyde more rapidly (7). Our finding of no association between ADH3 genotype and risk of lung cancer would tend to discount this as a possible mechanism of lung carcinogenesis.
A major concern in the examination of the literature regarding alcohol in relation to lung cancer has been confounding by smoking in that these behaviors tend to be strongly correlated, given the strong etiologic link between smoking and lung cancer. There are reports of alcohol and lung cancer among nonsmokers, which provide inconsistent evidence of an alcohol effect (21,22). In our study, we did not have a sufficient sample of cases who had never smoked to examine risk in this group alone. However, we did collect very detailed data regarding lifetime smoking practices including the number of cigarettes, pipes or cigars smoked during each decade of life when there was smoking and periods of attempting to quit. In addition, we assessed exposure to passive smoke at home and at work throughout the lifetime. There was considerable confounding by smoking. For example, for recent beer consumption with adjustment for smoking and other confounders, the OR went from 2.7 to 1.7. Although the remaining increase for women could be etiologic, concern remains that it can be explained by uncontrolled confounding. In a sensitivity analysis examining misclassification of both smoking status and drinking status, it was shown that the modest increase in risk seen in this and other studies could be accounted for by small amounts of misclassification (23). In our study, misclassification on smoking status and drinking status is less likely because of the intensive nature of the interview regarding lifetime practices. However, there may be differences in smoking practices (e.g., depth of inhalation, time of inhalation) that were not detected by our interview.
Additional sources of potential bias would be in the self-selection of participants for both the cases and controls, and recall bias, differential error for reports of past intake by cases and controls. There is some evidence, at least for breast cancer, that biased recall of alcohol consumption is not substantial (24,25). In this study, to minimize recall bias, both interviewers and participants were unaware of the hypotheses under study. There would not, of course, be recall bias in the genotype results. The ADH3 genotype does not affect the amount of alcohol consumed (8) and is not likely to be related to participation. Another source of error would be in measurement. Although we made considerable effort to probe into past exposures for alcohol, smoking and other relevant variables, error likely exists in our measurements of lifetime consumption. In addition to errors in estimation of total quantity of intake, there is also error resulting from differences among beers and hard liquors in their contents of compounds that may be important in relation to cancer risk. For the measurements of genotype, we included blind duplicates within each batch and there were no errors in those readings. Random measurement error would generally bias our findings toward the null value. A final important source of error would be related to the small sample size.
Our study provides some evidence that recent beer consumption may be associated with lung cancer, although the possibilities of chance, uncontrolled confounding and/or selection bias remain as important alternate explanations. We did not find evidence that those with the ADH3 genotype leading to more exposure to acetaldehyde from alcohol were at increased risk of lung cancer, but rather some indication of decreased risk. These null findings for gene-environment interactions would indicate that an acetaldehyde effect is not likely to explain any effect. The bulk of the recent evidence would indicate that there is a small effect of alcohol on risk, if any. Findings of increased risk related to alcohol consumption in nonsmokers remain provocative. Clearly, smoking remains the primary risk factor for lung cancer and the primary target of prevention for this disease with considerable bearing on public health.
|
FOOTNOTES |
|---|
2 Supported in part by grant no. AA 09802 from the National Institute on Alcohol Abuse and Alcoholism.
Manuscript received 16 April 2003. Initial review completed 19 May 2003. Revision accepted 6 August 2003.
|
LITERATURE CITED |
|---|
|
TOPABSTRACTSUBJECTS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
1. Ferlay, J., Bray, F., Pisani, P. & Parkin, D. M., GLOBOCAN (2000) Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0. IARC CancerBase No. 5. Lyon, IARCPress, 2001 2000 Limited version available from: URL: http://www.dep.iarc.fr/globocan/globocan.htm Last updated on 03/02/2001.
2. Potter, J. D. & McMichael, A. J. (1984) Alcohol, beer and lung cancer—a meaningful relationship?. Int. J. Epidemiol. 13:240-242.
3. Bandera, E. V., Freudenheim, J. L. & Vena, J. E. (2001) Alcohol consumption and lung cancer: A review of the epidemiologic evidence. Cancer Epidemiol. Biomark. Prev. 10:813-821.
4. Kvåle, G., Bjelke, E. & Gart, J. J. (1983) Dietary habits and lung cancer risk. Int. J. Cancer 31:397-405.[Medline]
5. Bandera, E. V., Freudenheim, J. L., Graham, S., Marshall, J. R., Haughey, B. P., Swanson, M., Brasure, J. & Wilkinson, G. (1992) Alcohol consumption and lung cancer in white males. Cancer Causes Control 3:85-93.
6. De Stefani, E., Correa, P., Fierro, L., Fontham, E.T.H., Chen, V. & Zavala, D. (1993) The effect of alcohol on the risk of lung cancer in Uruguay. Cancer Epidemiol. Biomark. Prev. 2:21-26.[Abstract]
7. Bosron, W. F. & Li, T. K. (1986) Genetic polymorphisms of human liver alcohol and aldehyde dehydrogenases and their relationship to alcohol metabolism and alcoholism. Hepatology 6:502-510.[Medline]
8. Freudenheim, J. L., Ambrosone, C. B., Moysich, K. B., Vena, J. E., Graham, S., Marshall, J. R., Muti, P., Laughlin, R., Nemoto, T., Harty, L., Crits, A. & Shields, P. G. (1999) Alcohol dehydrogenase 3 genotype modification of the association of alcohol consumption with breast cancer. Cancer Causes Control 10:369-377.[Medline]
9. Hines, L. M., Hankinson, S. E., Smith-Warner, S. A., Spiegelman, D., Kelsey, K. T., Colditz, G. A., Willett, W. C. & Hunter, D. J. (2000) A prospective study of the effect of alcohol consumption and ADH3 genotype on plasma steroid hormone levels and breast cancer risk. Cancer Epidemiol. Biomark. Prev. 9:1099-1105.
10. Harty, L. C., Caporaso, N. E., Hayes, R. B., Winn, D. M., Bravo-Otero, E., Blot, W. J., Kleinman, D. V., Brown, L. M., Armenian, , Fraumeni, J. F., Jr & Shields, P. G. (1997) Alcohol dehydrogenase 3 genotype and risk of oral cavity and pharyngeal cancers. J. Natl. Cancer Inst. 89:1698-1705.[Abstract]
11. Coutelle, C., Ward, P. J., Fleury, B., Quattrocchi, P., Chambrin, H., Iron, A., Couzigou, P. & Cassaigne, A. (1997) Laryngeal and oropharyngeal cancer, and alcohol dehydrogenase 3 and glutathione S-transferase M1 polymorphisms. Hum. Genet. 99:319-325.[Medline]
12. Day, C. P., Bashir, R., James, O.F.W., Bassendine, M. F., Crabb, D. W., Thomasson, H. R., Li, T. K. & Edenberg, H. J. (1991) Investigation of the role of polymorphisms at the alcohol and aldehyde dehydrogenase loci in genetic predisposition to alcohol-related end-organ damage. Hepatology 14:798-801.[Medline]
13. Russell, M., Marshall, J. R., Trevisan, M., Freudenheim, J. L., Chang, A.W.K., Markovic, N., Vana, J. E. & Priore, R. L. (1997) Test-retest reliability of the cognitive lifetime drinking history. Am. J. Epidemiol. 146:975-981.
14. McCann, S. E., Marshall, J. R., Trevisan, M., Russell, M., Muti, P., Markovic, N., Chan, A.W.K. & Freudenheim, J. L. (1999) Recent alcohol intake as estimated by the Health Habits and History Questionnaire and a more detailed alcohol intake questionnaire. Am. J. Epidemiol. 150:334-340.
15. Block, G., Hartman, A. M., Dresser, C. M., Carroll, M. D., Gannon, J. & Gardner, L. (1986) A data based approach to diet questionnaire design and testing. Am. J. Epidemiol. 124:453-469.
16. Block, G., Woods, M., Potosky, A. & Clifford, C. (1990) Validation of a self-administered diet history questionnaire using multiple diet records. J. Clin. Epidemiol. 43:1327-1335.[Medline]
17. Groppi, A., Begueret, J. & Iron, A. (1990) Improved methods for genotype determination of human alcohol dehydrogenase (ADH) at ADH 2 and ADH 3 loci by using polymerase chain reaction-directed mutagenesis. Clin. Chem. 36:1765-1768.
18. Potter, J. D., Sellers, T. A., Folsom, A. R. & McGovern, P. G. (1992) Alcohol, beer, and lung cancer in postmenopausal women: The Iowa Women’s Health Study. Ann. Epidemiol. 2:587-595.[Medline]
19. Nachiappan, V., Mufti, S. I., Chakravarti, A., Eskelson, C. D. & Rajasekharan, R. (1994) Lipid peroxidation and ethanol-related tumor promotion in Fischer-344 rats treated with tobacco-specific nitrosamines. Alcohol Alcohol. 29:565-574.
20. International Agency for Research on Cancer (1984) IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Humans, vol. 36, Allyl Compounds, Aldehydes, Epoxides and Peroxides 1984 IARC Lyon, France.
21. Kabat, K. C. & Wynder, E. L. (1984) Lung cancer in nonsmokers. Cancer 53:1214-1221.[Medline]
22. Koo, L. C. (1988) Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr. Cancer 11:155-171.[Medline]
23. Korte, J. E., Brennan, P., Henley, S. J. & Boffetta, P. (2002) Dose-specific meta-analysis and sensitivity analysis of the relation between alcohol consumption and lung cancer risk. Am. J. Epidemiol. 155:496-506.
24. Giovannucci, E., Stampfer, M. J., Colditz, G. A., Manson, J. E., Rosner, B. A., Longnecker, M. P., Speizer, F. E. & Willett, W. C. (1993) Recall and selection bias in reporting past alcohol consumption among breast cancer cases. Cancer Causes Control 4:441-448.[Medline]
25. Holmberg, L., Ohlander, E. M., Byers, T., Zack, M., Wolk, A., Bruce, A., Bergstrom, R., Bergkvist, L. & Adami, H. O. (1996) A search for recall bias in a case-control study of diet and breast cancer. Int. J. Epidemiol. 25:235-244.
This article has been cited by other articles:
|
|
 |
|
  C. Chao Associations between Beer, Wine, and Liquor Consumption and Lung Cancer Risk: A Meta-analysis Cancer Epidemiol. Biomarkers Prev., November 1, 2007; 16(11): 2436 - 2447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Rohrmann, J. Linseisen, H. C. Boshuizen, J. Whittaker, A. Agudo, P. Vineis, P. Boffetta, M. K. Jensen, A. Olsen, K. Overvad, et al. Ethanol Intake and Risk of Lung Cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC) Am. J. Epidemiol., December 1, 2006; 164(11): 1103 - 1114. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L Freudenheim, J. Ritz, S. A Smith-Warner, D. Albanes, E. V Bandera, P. A van den Brandt, G. Colditz, D. Feskanich, R A. Goldbohm, L. Harnack, et al. Alcohol consumption and risk of lung cancer: a pooled analysis of cohort studies Am. J. Clinical Nutrition, September 1, 2005; 82(3): 657 - 667. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A Ruano-Ravina, A Figueiras, and J M Barros-Dios Type of wine and risk of lung cancer: a case-control study in Spain Thorax, November 1, 2004; 59(11): 981 - 985. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
