Skip Navigation


American Journal of Epidemiology Advance Access originally published online on December 8, 2008
American Journal of Epidemiology 2009 169(2):132-142; doi:10.1093/aje/kwn306
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Web Only Tables
Right arrow All Versions of this Article:
169/2/132    most recent
kwn306v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Purdue, M. P.
Right arrow Articles by Hayes, R. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Purdue, M. P.
Right arrow Articles by Hayes, R. B.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology Published by the Johns Hopkins Bloomberg School of Public Health 2008.

Type of Alcoholic Beverage and Risk of Head and Neck Cancer—A Pooled Analysis Within the INHANCE Consortium

Mark P. Purdue, Mia Hashibe, Julien Berthiller, Carlo La Vecchia, Luigino Dal Maso, Rolando Herrero, Silvia Franceschi, Xavier Castellsague, Qingyi Wei, Erich M. Sturgis, Hal Morgenstern, Zuo-Feng Zhang, Fabio Levi, Renato Talamini, Elaine Smith, Joshua Muscat, Philip Lazarus, Stephen M. Schwartz, Chu Chen, Jose Eluf Neto, Victor Wünsch-Filho, David Zaridze, Sergio Koifman, Maria Paula Curado, Simone Benhamou, Elena Matos, Neonilia Szeszenia-Dabrowska, Andrew F. Olshan, Juan Lence, Ana Menezes, Alexander W. Daudt, Ioan Nicolae Mates, Agnieszka Pilarska, Eleonora Fabianova, Peter Rudnai, Debbie Winn, Gilles Ferro, Paul Brennan, Paolo Boffetta and Richard B. Hayes

Correspondence to Dr. Mark P. Purdue, National Cancer Institute, 6120 Executive Boulevard, EPS 8111, Rockville, MD 20852 (e-mail: purduem{at}mail.nih.gov).

Received for publication March 14, 2008. Accepted for publication September 8, 2008.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
The authors pooled data from 15 case-control studies of head and neck cancer (9,107 cases, 14,219 controls) to investigate the independent associations with consumption of beer, wine, and liquor. In particular, they calculated associations with different measures of beverage consumption separately for subjects who drank beer only (858 cases, 986 controls), for liquor-only drinkers (499 cases, 527 controls), and for wine-only drinkers (1,021 cases, 2,460 controls), with alcohol never drinkers (1,124 cases, 3,487 controls) used as a common reference group. The authors observed similar associations with ethanol-standardized consumption frequency for beer-only drinkers (odds ratios (ORs) = 1.6, 1.9, 2.2, and 5.4 for ≤5, 6–15, 16–30, and >30 drinks per week, respectively; Ptrend < 0.0001) and liquor-only drinkers (ORs = 1.6, 1.5, 2.3, and 3.6; P < 0.0001). Among wine-only drinkers, the odds ratios for moderate levels of consumption frequency approached the null, whereas those for higher consumption levels were comparable to those of drinkers of other beverage types (ORs = 1.1, 1.2, 1.9, and 6.3; P < 0.0001). Study findings suggest that the relative risks of head and neck cancer for beer and liquor are comparable. The authors observed weaker associations with moderate wine consumption, although they cannot rule out confounding from diet and other lifestyle factors as an explanation for this finding. Given the presence of heterogeneity in study-specific results, their findings should be interpreted with caution.

alcohol drinking; alcoholic beverages; beer; case-control studies; head and neck neoplasms; meta-analysis; wine

Abbreviations: CI, confidence interval; HNC, head and neck cancer; INHANCE, International Head and Neck Cancer Epidemiology; OR, odds ratio


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 
Alcohol consumption is a major risk factor for head and neck cancer (HNC), that is, cancers of the oral cavity, pharynx, and larynx (1, 2). Although the causal mechanism is not fully understood, ethanol may influence cancer risk directly through topical carcinogenic effects and/or indirectly by enhancing the effects of other carcinogens (e.g., tobacco) (1–3). There is speculation that other ingredients of alcoholic beverages besides ethanol may additionally influence cancer risk. Asbestos filtration products, tannins, N-nitroso compounds, urethane, and other possible carcinogens have been found in some alcoholic beverages (1, 3). Conversely, antioxidants, such as resveratrol, found in red wine may be anticarcinogenic (4).

In light of this uncertainty surrounding the mechanisms underlying alcohol-induced carcinogenesis, there has been interest as to whether different alcoholic beverages are differentially associated with HNC risk. There are several reports of differences in relative risk among beer, wine, and liquor consumption (Web Table 1). (This information is described in the first of 6 supplementary tables; each is referred to as "Web table" in the text and is posted on the Journal’s website (http://aje.oxfordjournals.org/).) However, the observed differences in risk are inconsistent across studies (5–17). A common limitation of these studies is the inability to adequately isolate the effects of beer, wine, and liquor consumption, given their intercorrelated patterns of consumption in many populations. One method of addressing this issue is to identify drinkers of only beer, wine, or liquor and to estimate their respective associations with HNC risk relative to never drinkers. However, limitations in sample size among individual studies have prevented the use of this analytical approach across all 3 beverage types.

Larger studies are needed to adequately investigate the effects of different alcoholic beverages; with this in mind, we conducted a pooled analysis of studies participating in the International Head and Neck Cancer Epidemiology (INHANCE) Consortium to investigate the independent associations with HNC risk for consumption of beer, wine, and liquor. In particular, we estimated these associations among persons who consumed 1 beverage type exclusively (referred to hereafter as "pure drinkers").


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 APPENDIX
 References
 
A detailed description of the INHANCE Consortium has been previously published (18). For the purposes of this study, HNC cases were restricted to invasive tumors of the oral cavity, oropharynx, hypopharynx, oral cavity or pharynx not otherwise specified, larynx, and HNC unspecified (18). Cancers of the salivary gland (International Classification of Diseases for Oncology, Second Edition, codes C07–C08) were excluded from our analysis because of the different etiologic pattern from that of other HNCs. The INHANCE pooled data set (version 1.1) used for this analysis included 15 individual case-control studies, comprising 10,301 HNC cases and 15,329 controls (19–32). Individuals with missing data on age, sex, or race/ethnicity were excluded (23 cases, 102 controls), as were subjects with missing or conflicting information regarding drinking status (620 cases, 390 controls). Finally, participants from the India and Sudan centers of the International Agency for Research on Cancer Multicenter Study (32) were excluded (551 cases, 618 controls) because of the extremely low prevalence of alcohol consumption in these populations. After these exclusions, the data set for this analysis included 9,107 HNC cases and 14,219 controls.

Characteristics of the studies included in the pooled data are shown in Table 1. Most were hospital based, and all included controls frequency matched to cases on age, sex, and additional factors. Study interviews were conducted face-to-face, except in Iowa where subjects completed self-administered questionnaires. Written, informed consent was obtained from all study subjects, and the investigations were approved by institutional review boards at each of the institutes involved. All study questionnaires were reviewed to assess the comparability of the data and wording of interview questions. The questionnaires were conceptually similar across studies; subjects were asked if they were alcohol drinkers (definitions by study are included in the Appendix) and then asked subsequent questions on frequency of drinking, duration of drinking, and different types of alcoholic beverages consumed. Data from individual studies were received by the pooling center with personal identifiers removed. Each data item was checked for illogical or missing values. Queries were sent to investigators, and inconsistencies were resolved.


View this table:
[in this window]
[in a new window]

  		Table 1. Summary of Individual Studies in the INHANCE Consortium Pooled Analysis of Alcoholic Beverages

 
To facilitate comparisons of consumption frequency across alcoholic beverage types and to adjust for differences between studies in specified beverage volumes, we standardized the calculated average beverage-specific weekly number of drinks on the basis of ethanol volume. We calculated this measure for each alcoholic beverage type by converting the beverage volume specified in the questionnaire to milliliters, multiplying this by the average number of drinks per week of that beverage, applying estimates of the beverage-specific volume percentage of pure ethanol (5% for beer, 12% for wine, 40% for liquor) (1), and dividing the weekly consumption of pure ethanol by the average per-drink volume of pure ethanol across alcoholic beverage types for the 15 studies (15.6 mL). The standardized estimates of beverage consumption frequency can thus be assumed to have, on average, identical volumes of ethanol per drink.

The cutpoints for categorizing the average weekly number of ethanol-standardized drinks for a beverage (never drinker, ≤5, 6–15, 16–30, >30 drinks per week) were selected to maximize the number of categories of consumption level while minimizing the occurrence of sparse data within the upper categories for some studies. This latter issue arose because of considerable variation across populations in the popularity of different beverages, as is summarized in Table 1. Participants in European studies consumed larger quantities of wine than did subjects from North and Latin America. Less extreme differences were observed for liquor consumption (generally higher in North and Latin American studies), while beer consumption was fairly uniform across studies. For each alcoholic beverage, we also calculated the duration (never drinker, ≤10, 11–20, 21–40, >40 years) of consumption reported by subjects and the lifetime cumulative number of standardized drinks consumed (never drinker, ≤10, 11–20, 21–40, 41–80, >80 drink-years; calculation previously described (18)).

From the data on alcohol consumption, we were able to identify mutually exclusive subgroups across studies: never drinkers (1,124 cases, 3,487 controls), drinkers of beer only (858 cases, 986 controls), drinkers of liquor only (499 cases, 527 controls), drinkers of wine only (1,021 cases, 2,460 controls), and drinkers of multiple beverage types (5,605 cases, 6,759 controls). Our analyses focused on investigations of alcoholic beverage consumption (ethanol-standardized consumption frequency, duration of consumption, lifetime cumulative consumption) among each of the groups of pure drinkers, with never drinkers as a common reference group. However, we also examined associations with beverage-specific, ethanol-standardized consumption frequency among all study subjects (i.e., including drinkers of multiple beverage types).

We used a 2-stage random effects modeling approach (33) for our pooled analysis of alcoholic beverage types. At the first stage, associations between HNC and measures of alcoholic beverage consumption were assessed by estimating odds ratios and 95% confidence intervals based on unconditional logistic regression models for each case-control study. To adjust for matching factors and potential confounding, we included age (categories shown in Web Table 2), sex, education, race/ethnicity, study center, pack-years of cigarette smoking, years of cigar smoking (continuous), and years of pipe smoking (continuous) in all models. Analyses conducted among all study subjects (i.e., including drinkers of multiple beverage types) were also adjusted for consumption frequency of other beverages. For subjects missing information on educational level (535 cases, 438 controls), we applied multiple imputation using the PROC MI procedure (SAS, version 9, software; SAS Institute, Inc., Cary, North Carolina) under the assumption that the education data were missing at random (34). Within each geographic region, we predicted educational level using logistic regression models with parameters for age, sex, race/ethnicity, study, and case/control status (35). The logistic regression results to assess summary estimates for alcoholic beverage consumption for the 5 imputations were combined by using the PROC MIANALYZE procedure (SAS Institute, Inc.).

At the second stage of analysis, summary effect estimates were calculated by using a maximum-likelihood–based random effects model. We tested for evidence of heterogeneity between the study-specific odds ratios for categories of alcoholic beverage use by calculating likelihood ratio tests comparing the deviance statistics from main-effects models with those from models specifying interaction between alcoholic beverage consumption and study.

Among pure drinkers, we conducted a test of heterogeneity in odds ratio estimates for levels of ethanol-standardized consumption frequency across the 3 beverage types. This was performed by constructing a likelihood ratio test comparing the deviance statistic of a model containing beverage-specific parameters for consumption frequency with that of a constrained model with parameters for different levels of consumption frequency specified as being identical across beverage types.

We also conducted additional analyses stratifying by geographic region (Europe, North America, Latin America), study design (hospital based, population based), cancer site (oral, pharynx, larynx), smoking history (never smoked, ever smoked), sex, and age (<55, ≥55 years). We also conducted influence analyses, where each study was excluded one at a time to assess the impact upon the magnitude of the overall summary estimate and its statistical significance.

All reported P values are 2 sided.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 APPENDIX
 References
 
Of the 9,107 cases and 14,219 controls included in the analysis, approximately 88% and 75%, respectively, reported alcohol consumption. The majority of drinkers reported consuming multiple beverage types (70% of case drinkers, 63% of control drinkers). Among pure drinkers, the most common beverage reported was wine (43% of case pure drinkers, 62% of control pure drinkers), followed by beer (36% of case pure drinkers, 25% of control pure drinkers) and liquor (21% of case pure drinkers, 13% of control pure drinkers). Selected characteristics of cases and controls are shown in Web Table 2. As expected, the distribution of pure beverage consumption differed substantially by geographic region; the majority of wine-only drinkers were European (83% of cases, 87% of controls), and most beer-only drinkers were North American (71% of cases, 53% of controls). Pure liquor consumption varied by geographic region to a lesser extent than for beer and wine; a plurality of liquor-only drinkers were from North America (45% of cases, 46% of controls). Pure drinkers and multiple-beverage drinkers were similar in their distributions by age, sex, ethnicity, educational level, and pack-years of cigarette use. Women and never smokers were less likely to consume alcohol.

Analyses of pure drinkers of each beverage type
Figure 1 shows study-specific odds ratios for beer, liquor, and wine consumption (vs. alcohol never drinkers) from analyses restricted to alcohol never drinkers and pure drinkers. The study-specific odds ratios were fairly homogeneous for beer but not for liquor or wine; tests of between-study heterogeneity for the latter 2 beverage types were statistically significant. In particular, the heterogeneity for wine consumption was greatest among the European studies. The random effects summary odds ratios for ever drinking of beer, liquor, and wine were 2.1 (95% confidence interval (CI): 1.6, 2.7), 2.2 (95% CI: 1.4, 3.4), and 1.6 (95% CI: 1.0, 2.6), respectively.


Figure 1
View larger version (16K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. Study-specific head and neck cancer odds ratios for consumption of separate alcoholic beverages—pure drinkers versus never drinkers. Diamond symbols represent odds ratios; symbol size is proportional to the number of cases among never drinkers and pure drinkers of that beverage. Horizontal lines represent 95% confidence intervals. The P value for the test of heterogeneity across individual studies is shown at the bottom of each graph. OR, odds ratio; Int., International.

 
The results from pooled analyses of ethanol-standardized beverage consumption frequency and duration among pure drinkers of each beverage type are summarized in Table 2. All summary odds ratio estimates demonstrated between-study heterogeneity. We observed similar associations with HNC risk among beer-only drinkers (odds ratios (ORs) = 1.6, 1.9, 2.2, and 5.4 for ≤5, 6–15, 16–30, and >30 drinks per week, respectively) and liquor-only drinkers (ORs = 1.6, 1.5, 2.3, and 3.6) relative to never drinkers. Associations with years of consumption and lifetime cumulative consumption were also comparable between beer-only and liquor-only drinkers. Among wine-only drinkers, odds ratios for moderate levels of consumption were close to null (ORs = 1.1 and 1.2 for ≤5 and 6–15 drinks per week, respectively); increases in risk were observed only for higher consumption levels (ORs = 1.9 and 6.3 for 16–30 and >30 drinks per week, respectively). A test of heterogeneity in odds ratio estimates for levels of consumption frequency across all 3 beverage types was statistically significant (P < 0.01). When we conducted pairwise tests of heterogeneity between beverage types, it became apparent that the odds ratio estimates for wine-only drinkers were the source of the heterogeneity (beer vs. liquor, P = 0.60; beer vs. wine, P < 0.01; liquor vs. wine, P = 0.02).


View this table:
[in this window]
[in a new window]

  		Table 2. Head and Neck Cancer Risk and Pure Consumption of Beer, Wine, and Liquor, With Never Drinkers as the Referent Group, in the INHANCE Consortium

 
Study-specific results (Web Tables 3–5) were inspected, and influence analyses were performed to attempt to identify sources of between-study heterogeneity in the pooled estimates of ethanol-standardized beverage consumption frequency among pure drinkers and never drinkers. The summary odds ratios for beer-only consumption were influenced slightly by exclusion of the Iowa study (OR = 1.6, 1.7, 1.7, and 4.2 for ≤5, 6–15, 16–30, and >30 drinks per week, respectively), although statistically significant between-study heterogeneity remained (P < 0.0001). Exclusion of the Tampa study influenced the pooled results for consumption of liquor only (ORs = 1.5, 1.8, 1.9, and 3.2); no residual between-study heterogeneity was detected (P = 0.15). The associations with pure wine consumption became considerably weaker in magnitude upon exclusion of the Switzerland study (ORs = 1.0, 1.0, 1.4, and 3.7) and increased with exclusion of the Milan study (ORs = 1.2, 1.4, 2.4, and 8.7); in each case, however, substantial heterogeneity was still present among the remaining pooled studies.

The results of stratified analyses of pure ethanol-standardized beverage consumption level (never drinker, ≤15, >15 drinks per week) are summarized in Tables 3–5. We observed some differences between geographic regions in the magnitude of risk for the highest consumption level. For beer, summary odds ratio estimates for North American studies (ORs = 2.0 and 4.1 for ≤15 and >15 standard drinks per week, respectively) were slightly stronger than those for studies from Europe (ORs = 1.4 and 3.5) and Latin America (ORs = 1.5 and 2.1). For liquor, the summary odds ratio estimates for North American studies (ORs = 1.9 and 3.2) and Latin American studies (ORs = 2.1 and 2.6) were stronger than those of studies from Europe (ORs = 1.2 and 1.7). For wine, the odds ratio estimates for European studies (ORs = 1.5 and 4.0) were stronger than those for studies from North America (ORs = 1.1 and 2.8) and Latin America (ORs = 1.1 and 3.7). Odds ratios were stronger for oral and pharyngeal cancer than for laryngeal cancer and among smokers compared with never smokers. The pattern of associations across beverage types did not change with anatomic site or smoking status. Additional analyses investigating the joint effects of smoking and pure beverage consumption (ever vs. never drinker) also did not meaningfully differ by beverage type (results not shown). Beverage associations did not differ by study design, sex, or age group.


View this table:
[in this window]
[in a new window]

  		Table 3. Head and Neck Cancer Risk and Pure Ethanol-standardized Consumption of Beer, With Analysis Stratified by Selected Study and Subject Characteristics, in the INHANCE Consortium

 

View this table:
[in this window]
[in a new window]

  		Table 5. Head and Neck Cancer Risk and Pure Ethanol-standardized Consumption of Wine, With Analysis Stratified by Selected Study and Subject Characteristics, in the INHANCE Consortium

 
Analyses of all study subjects
In our pooled analysis of all study subjects, the pattern of findings across alcoholic beverage types was comparable to that observed in the analysis of pure drinkers (Web Table 6).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
APPENDIX
 References
 
Overall, we found estimates of the relative risks of HNC associated with alcohol consumption to be very similar for beer and liquor. In light of our use of ethanol-standardized estimates of beverage consumption frequency, our finding of similar associations with risk for beer and liquor consumption argues in favor of ethanol and its metabolites as the principal carcinogenic agents in these alcoholic beverages, rather than beverage-specific constituents. We cannot rule out the possibility that another byproduct of alcohol production—created at levels highly correlated with ethanol concentration—is in fact the actual HNC carcinogen in alcoholic beverages. To our knowledge, no such candidate compound has been identified.

Our summary odds ratio estimates for wine consumption, however, were somewhat different from those for the other 2 beverage types. Whereas the dose-response relation was generally linear for beer and liquor, no such trend was observed for wine. At moderate levels of consumption, the odds ratio estimates for wine were much weaker than the corresponding estimates for beer or liquor. At a high consumption level (>30 standard drinks per week), the odds ratio estimate for wine was generally comparable to those of the other beverage types. This pattern in risk across beverage types was consistent across most subgroup strata with nonsparse numbers, and it was also apparent in analyses of all subjects (i.e., pure and mixed-beverage drinkers), suggesting that these findings are not attributable to issues concerning the study of pure drinkers.

One possible reason for the observed weaker association with moderate wine consumption is that it reflects residual confounding. Wine consumption has been associated with higher intake of a healthy diet, higher education, and lower smoking levels in studies conducted in the United States and Northern Europe (36–38). Although we performed detailed adjustment for education and smoking in our modeling, we did not adjust for diet or other lifestyle factors that may influence the risk of HNC, and we may not have completely accounted for the effects of smoking. Had we been able to do so, the associations among wine drinkers may have been more similar to those among beer and liquor drinkers. A second possible explanation involves the fact that wine is more frequently consumed during meals than other alcoholic beverages (39). The possibility of an "alcohol washing effect" by the chewing and swallowing of foods has been proposed, whereby ingestion of foods could "wash" alcohol drinking and reduce the effect of ethanol and its carcinogenic metabolites on the oral, pharyngeal, and esophageal mucosa (40, 41). Finally, it is possible that the carcinogenic effects of ethanol in these beverages are offset by other anticarcinogenic compounds in wine such as resveratrol, a phenolic compound with antioxidant properties present in red wine that has been shown to inhibit tumor initiation and progression in experimental studies (4). We did not have information on the type of wine consumed (red vs. white) for our pooled analysis; consequently, we could not investigate this hypothesis further. We caution that this last explanation is speculative, given the paucity of supporting evidence.

Some differences between regions in associations with beer (strongest in North America), liquor (strongest in North and Latin America), and wine (strongest in Europe) were observed in the analysis of pure drinkers. This may reflect differences within each broad category of beverage consumption level used in the subgroup analyses (<15, ≥15 standard drinks per week) in the average number of drinks per week from each region, given that the consumption levels of beer, liquor, and wine were highest among studies from North America, North and Latin America, and Europe, respectively. Another possible explanation is that these differences may indicate greater accuracy in assessing lifetime consumption patterns (and thus less attenuation of the effect due to misclassification) for the beverage predominant in that region (42).

We did not observe any clear differences between beverage types in analyses stratified by smoking status. However, the odds ratio estimates for each beverage type were stronger and more stable among smokers compared with those among never smokers. These differences are consistent with effect modification between alcohol and tobacco use, evidence of which has been previously reported by several studies (43). It is biologically plausible that alcohol consumption could amplify the carcinogenic effects of smoking, as ethanol has been shown to increase the permeability of oral mucosa to tobacco combustion products (44, 45).

We also did not observe meaningful differences across beverage types with regard to their associations with HNC at different anatomic sites. However, the odds ratio estimates for alcohol consumption were stronger for cancers of the oral cavity and pharynx than for laryngeal cancers, irrespective of beverage type. This pattern has been previously reported in several studies and is biologically plausible given that the oral cavity and pharynx are directly exposed to alcoholic beverages, whereas only parts of the larynx (supraglottis and epilarynx) are exposed.

An important strength of our study is its large size. The INHANCE Consortium, with detailed information on lifetime alcohol consumption for more than 9,000 cases and 14,000 controls from 15 case-control studies, is a unique resource for investigating whether beer, liquor, and wine consumptions are differentially associated with HNC risk. In particular, the large size of this study enabled us to estimate associations with alcohol consumption among individuals who reported consuming only beer, wine, or liquor. As such, it is, by far, the largest study to investigate beverage effects among pure drinkers. Comparisons across pure-drinker subgroups arguably represent the most informative means of investigating whether the relative risk of HNC from alcohol consumption differs by alcoholic beverage type.

This study also has limitations. Foremost among these is the presence of heterogeneity in study-specific results for each measure of alcohol consumption examined, which limits the conclusions that can be made from our pooled analysis. Our inspection of study-specific findings, influence analyses, and subgroup analyses stratifying by study characteristics did not generally point to any clear sources of heterogeneity, although exclusion of the Tampa study from the pooled analysis of ethanol-adjusted frequency of liquor-only consumption resulted in a heterogeneity test statistic that was no longer statistically significant. Differences across studies in the wording and design of their questions assessing lifetime alcohol consumption are a potential source of heterogeneity. A second study limitation, which may contribute to the observed heterogeneity, is the absence of more detailed information regarding alcoholic beverages (e.g., red vs. white wine consumption, beverage-specific percentage of ethanol content for a given country, the use of nonalcoholic "mixers" with liquor). Such unmeasured differences among the underlying study populations in their alcoholic beverage consumption patterns could contribute to between-study heterogeneity. For example, the heterogeneity for liquor consumption may be due, in part, to possible differences among study populations in the type of liquor consumed and/or the frequency with which liquor is consumed "straight" versus "mixed." There is some evidence that ethanol concentration, independent of ethanol amount, may influence HNC risk (16). A third limitation is our inability to adjust for dietary factors or other lifestyle characteristics that could potentially confound our findings. Finally, we acknowledge the inherent limitation of the retrospective case-control design for investigating the etiologic relevance of past exposures such as lifetime alcohol consumption.

In conclusion, we observed comparable estimates of HNC relative risk for consumption of beer, liquor and, at high consumption levels, wine in our pooled analysis within the INHANCE Consortium. We observed, however, a comparatively weaker risk at low consumption levels for wine than for the other beverage types. Given the presence of heterogeneity in study-specific results and the possible existence of confounding from diet and other lifestyle factors, our findings should be interpreted with caution.


   APPENDIX
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
References
 
The definitions of ever alcohol drinking were as follows: "ever" consumed alcohol (Central Europe; Aviano, Milan, and Italy Multicenter; France; and Switzerland studies); >4 drinks in a year (Seattle study); ≥1 drink per month for ≥6 months in a lifetime (Los Angeles study); ≥12 drinks of a kind of alcohol in a lifetime (Puerto Rico study); once or more per month (Multicenter, Latin America studies); average ≥1 drink per week for ≥1 year (Iowa study), once or more per week for ≥1 year (Tampa and Houston studies); ≥4 times per month of beer, wine, or liquor (North Carolina study).


View this table:
[in this window]
[in a new window]

  		Table 4. Head and Neck Cancer Risk and Pure Ethanol-standardized Consumption of Liquor, With Analysis Stratified by Selected Study and Subject Characteristics, in the INHANCE Consortium

 


   ACKNOWLEDGMENTS
 
Author affiliations: National Cancer Institute, Bethesda, Maryland (Mark P. Purdue, Debbie Winn, Richard B. Hayes); International Agency for Research on Cancer, Lyon, France (Mia Hashibe, Julien Berthiller, Silvia Franceschi, Maria Paula Curado, Gilles Ferro, Paul Brennan, Paolo Boffetta); Istituto di Ricerche Farmacologiche Mario Negri and University of Milan, Milan, Italy (Carlo La Vecchia); Aviano Cancer Centre, Aviano, Italy (Luigino Dal Maso, Renato Talamini); Instituto de Investigación Epidemiológica, San José, Costa Rica (Rolando Herrero); Institut Català d'Oncologia, Barcelona, Spain (Xavier Castellsague); University of Texas M. D. Anderson Cancer Center, Houston, Texas (Qingyi Wei, Erich M. Sturgis); School of Public Health, University of Michigan, Ann Arbor, Michigan (Hal Morgenstern); School of Public Health, University of California, Los Angeles, California (Zuo-Feng Zhang); Institut universitaire de médecine sociale et préventive, Lausanne, Switzerland (Fabio Levi); College of Public Health, University of Iowa, Iowa City, Iowa (Elaine Smith); Penn State College of Medicine, Hershey, Pennsylvania (Joshua Muscat, Philip Lazarus); Fred Hutchinson Cancer Research Center, Seattle, Washington (Stephen M. Schwartz, Chu Chen); Universidade de Sao Paulo, Sao Paulo, Brazil (Jose Eluf Neto, Victor Wünsch-Filho); Cancer Research Centre, Moscow, Russia (David Zaridze); Escola Nacional de Suade Publica, Rio de Janeiro, Brazil (Sergio Koifman); INSERM U794, Evry, France (Simone Benhamou); Institute of Oncology Angel H. Roffo, University of Buenos Aires, Buenos Aires, Argentina (Elena Matos); Institute of Occupational Medicine, Lodz, Poland (Neonilia Szeszenia-Dabrowska); University of North Carolina School of Public Health, Chapel Hill, North Carolina (Andrew F. Olshan); Institute of Oncology and Radiobiology, Havana, Cuba (Juan Lence); Universidade Federal de Pelotas, Pelotas, Brazil (Ana Menezes); Hospital de Clinicas de Porto Alegre, Porto Alegre, Brazil (Alexander W. Daudt); Faculty of Medicine, University Carol Davila, Bucharest, Romania (Ioan Nicolae Mates); Second Maxillofacial Surgery Clinic, Medical Academy, Warsaw, Poland (Agnieszka Pilarska); Specialized State Health Institute, Banská Bystrica, Slovakia (Eleonora Fabianova); and National Institute of Environmental Health, Budapest, Hungary (Peter Rudnai).

The individual studies were funded by the following sources: 1) Milan study (Italian Association for Research on Cancer (AIRC)); 2) Aviano and Italy multicenter studies (AIRC, Italian League Against Cancer, and Italian Ministry of Research); 3) French study (Swiss Cancer League, Switzerland (KFS1069-09-2000); League Against Cancer of Fribourg, Switzerland (FOR381.88); Cancer Research, Switzerland (AKT 617); and Fund for Clinical Research Against Cancer, Gustave-Roussy Institute, Villejuif, France (88D28)); 4) Swiss study (Swiss League Against Cancer and the Swiss Research Against Cancer/Oncosuisse (KFS-700 and OCS-1633)); 5) Central Europe study (World Cancer Research Fund and the European Commission's INCOCOPERNICUS Program (contract no. IC15-CT98-0332)); 6) Seattle study (National Institutes of Health (NIH) (R01CA048896 and R01DE012609)); 7) Iowa study (NIH NIDCR R01DE11979, NIDCR R01DE13110, NIH FIRCA TW01500, and Veterans Affairs Merit Review Funds); 8) North Carolina study (NIH grant R01CA61188 and in part by a grant from the National Institute of Environmental Health Sciences (P30ES010126)); 9) Tampa study (NIH P01CA068384 and K07CA104231); 10) Los Angeles study (NIH grants P50CA90388, R01DA11386, R03CA77954, T32CA09142, U01CA96134, and R21ES011667, as well as the Alper Research Program for Environmental Genomics of the UCLA Jonsson Comprehensive Cancer Center); 11) Houston study (NIH R01ES11740 and R01CA100264); 12) Puerto Rico study (funded by the Intramural Research Program of the US National Institutes of Health (National Cancer Institute)); 13) Latin America study (Fondo para la Investigacion Cientifi ca y Tecnologica (Argentina), Institut Municipal d'Invesigacio Medica (Barcelona), Fundação de Amparo à Pesquisa no Estado de São Paulo (no. 01/01768-2), and European Commission (IC18-CT97-0222)); and 14) International Agency for Research on Cancer (IARC) Multicenter Study (Fondo de Investigaciones Sanitarias of the Spanish government (FIS 97/0024, FIS 97/0662, and BAE 01/5013), International Union Against Cancer, and Yamagiwa-Yoshida Memorial International Cancer Study grant).

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 APPENDIX
 References
 

  1. Alcohol drinking. IARC Working Group, Lyon, 13–20 October 1987. IARC Monogr Eval Carcinog Risks Hum (1988) 44:1–378.[Medline]
  2. Blot WJ. Invited commentary: More evidence of increased risks of cancer among alcohol drinkers. Am J Epidemiol (1999) 150(11):1138–1140.[Free Full Text]
  3. Ogden GR, Wight AJ. Aetiology of oral cancer: alcohol. Br J Oral Maxillofac Surg (1998) 36(4):247–251.[Medline]
  4. Aggarwal BB, Bhardwaj A, Aggarwal RS, et al. Role of resveratrol in prevention and therapy of cancer: preclinical and clinical studies. Anticancer Res. (2004) 24(5A):2783–2840.[Abstract/Free Full Text]
  5. Blot WJ, McLaughlin JK, Winn DM, et al. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res. (1988) 48(11):3282–3287.[Abstract/Free Full Text]
  6. Kabat GC, Wynder EL. Type of alcoholic beverage and oral cancer. Int J Cancer (1989) 43(2):190–194.[Web of Science][Medline]
  7. Merletti F, Boffetta P, Ciccone G, et al. Role of tobacco and alcoholic beverages in the etiology of cancer of the oral cavity/oropharynx in Torino, Italy. Cancer Res. (1989) 49(17):4919–4924.[Abstract/Free Full Text]
  8. Barra S, Franceschi S, Negri E, et al. Type of alcoholic beverage and cancer of the oral cavity, pharynx and oesophagus in an Italian area with high wine consumption. Int J Cancer (1990) 46(6):1017–1020.[Web of Science][Medline]
  9. Mashberg A, Boffetta P, Winkelman R, et al. Tobacco smoking, alcohol drinking, and cancer of the oral cavity and oropharynx among U.S. veterans. Cancer (1993) 72(4):1369–1375.[Medline]
  10. Bundgaard T, Wildt J, Frydenberg M, et al. Case-control study of squamous cell cancer of the oral cavity in Denmark. Cancer Causes Control (1995) 6(1):57–67.[CrossRef][Web of Science][Medline]
  11. Grønbaek M, Becker U, Johansen D, et al. Population based cohort study of the association between alcohol intake and cancer of the upper digestive tract. BMJ (1998) 317(7162):844–847.[Abstract/Free Full Text]
  12. Kjaerheim K, Gaard M, Andersen A. The role of alcohol, tobacco, and dietary factors in upper aerogastric tract cancers: a prospective study of 10,900 Norwegian men. Cancer Causes Control (1998) 9(1):99–108.[CrossRef][Web of Science][Medline]
  13. Zambon P, Talamini R, La Vecchia C, et al. Smoking, type of alcoholic beverage and squamous-cell oesophageal cancer in northern Italy. Int J Cancer (2000) 86(1):144–149.[CrossRef][Web of Science][Medline]
  14. Garrote LF, Herrero R, Reyes RM, et al. Risk factors for cancer of the oral cavity and oro-pharynx in Cuba. Br J Cancer (2001) 85(1):46–54.[CrossRef][Web of Science][Medline]
  15. Schlecht NF, Pintos J, Kowalski LP, et al. Effect of type of alcoholic beverage on the risks of upper aerodigestive tract cancers in Brazil. Cancer Causes Control (2001) 12(7):579–587.[CrossRef][Web of Science][Medline]
  16. Huang WY, Winn DM, Brown LM, et al. Alcohol concentration and risk of oral cancer in Puerto Rico. Am J Epidemiol (2003) 157(10):881–887.[Abstract/Free Full Text]
  17. Castellsague X, Quintana MJ, Martinez MC, et al. The role of type of tobacco and type of alcoholic beverage in oral carcinogenesis. Int J Cancer (2004) 108(5):741–749.[CrossRef][Web of Science][Medline]
  18. Hashibe M, Brennan P, Benhamou S, et al. Alcohol drinking in never users of tobacco, cigarette smoking in never drinkers, and the risk of head and neck cancer: pooled analysis in the International Head and Neck Cancer Epidemiology Consortium. J Natl Cancer Inst (2007) 99(10):777–789.[Abstract/Free Full Text]
  19. Franceschi S, Talamini R, Barra S, et al. Smoking and drinking in relation to cancers of the oral cavity, pharynx, larynx, and esophagus in northern Italy. Cancer Res. (1990) 50(20):6502–6507.[Abstract/Free Full Text]
  20. Baron AE, Franceschi S, Barra S, et al. A comparison of the joint effects of alcohol and smoking on the risk of cancer across sites in the upper aerodigestive tract. Cancer Epidemiol Biomarkers Prev (1993) 2(6):519–523.[Abstract/Free Full Text]
  21. Benhamou S, Tuimala J, Bouchardy C, et al. DNA repair gene XRCC2 and XRCC3 polymorphisms and susceptibility to cancers of the upper aerodigestive tract. Int J Cancer (2004) 112(5):901–904.[CrossRef][Web of Science][Medline]
  22. Altieri A, Bosetti C, Gallus S, et al. Wine, beer and spirits and risk of oral and pharyngeal cancer: a case-control study from Italy and Switzerland. Oral Oncol (2004) 40(9):904–909.[Medline]
  23. Levi F, Pasche C, La Vecchia C, et al. Food groups and risk of oral and pharyngeal cancer. Int J Cancer (1998) 77(5):705–709.[CrossRef][Web of Science][Medline]
  24. Hashibe M, Boffetta P, Zaridze D, et al. Evidence for an important role of alcohol- and aldehyde-metabolizing genes in cancers of the upper aerodigestive tract. Cancer Epidemiol Biomarkers Prev (2006) 15(4):696–703.[Abstract/Free Full Text]
  25. Rosenblatt KA, Daling JR, Chen C, et al. Marijuana use and risk of oral squamous cell carcinoma. Cancer Res. (2004) 64(11):4049–4054.[Abstract/Free Full Text]
  26. Wang D, Ritchie JM, Smith EM, et al. Alcohol dehydrogenase 3 and risk of squamous cell carcinomas of the head and neck. Cancer Epidemiol Biomarkers Prev (2005) 14(3):626–632.[Abstract/Free Full Text]
  27. Olshan AF, Weissler MC, Watson MA, et al. GSTM1, GSTT1, GSTP1, CYP1A1, and NAT1 polymorphisms, tobacco use, and the risk of head and neck cancer. Cancer Epidemiol Biomarkers Prev (2000) 9(2):185–191.[Abstract/Free Full Text]
  28. Elahi A, Zheng Z, Park J, et al. The human OGG1 DNA repair enzyme and its association with orolaryngeal cancer risk. Carcinogenesis (2002) 23(7):1229–1234.[Abstract/Free Full Text]
  29. Cui Y, Morgenstern H, Greenland S, et al. Polymorphism of xeroderma pigmentosum group G and the risk of lung cancer and squamous cell carcinomas of the oropharynx, larynx and esophagus. Int J Cancer (2006) 118(3):714–720.[CrossRef][Web of Science][Medline]
  30. Zhang Z, Shi Q, Liu Z, et al. Polymorphisms of methionine synthase and methionine synthase reductase and risk of squamous cell carcinoma of the head and neck: a case-control analysis. Cancer Epidemiol Biomarkers Prev (2005) 14(5):1188–1193.[Abstract/Free Full Text]
  31. Hayes RB, Bravo-Otero E, Kleinman DV, et al. Tobacco and alcohol use and oral cancer in Puerto Rico. Cancer Causes Control (1999) 10(1):27–33.[CrossRef][Web of Science][Medline]
  32. Herrero R, Castellsague X, Pawlita M, et al. Human papillomavirus and oral cancer: the International Agency for Research on Cancer Multicenter Study. J Natl Cancer Inst (2003) 95(23):1772–1783.[Abstract/Free Full Text]
  33. Stukel TA, Demidenko E, Dykes J, et al. Two-stage methods for the analysis of pooled data. Stat Med (2001) 20(14):2115–2130.[CrossRef][Web of Science][Medline]
  34. Greenland S, Finkle WD. A critical look at methods for handling missing covariates in epidemiologic regression analyses. Am J Epidemiol (1995) 142(12):1255–1264.[Abstract/Free Full Text]
  35. Rubin DB. Multiple Imputation for Nonresponse in Surveys (1987) New York, NY: John Wiley & Sons.
  36. Johansen D, Friis K, Skovenborg E, et al. Food buying habits of people who buy wine or beer: cross sectional study. BMJ (2006) 332(7540):519–522.[Abstract/Free Full Text]
  37. Tjonneland A, Gronbaek M, Stripp C, et al. Wine intake and diet in a random sample of 48763 Danish men and women. Am J Clin Nutr (1999) 69(1):49–54.[Abstract/Free Full Text]
  38. Klatsky AL, Armstrong MA, Kipp H. Correlates of alcoholic beverage preference: traits of persons who choose wine, liquor or beer. Br J Addict (1990) 85(10):1279–1289.[CrossRef][Web of Science][Medline]
  39. Hupkens CL, Knibbe RA, Drop MJ. Alcohol consumption in the European community: uniformity and diversity in drinking patterns. Addiction (1993) 88(10):1391–1404.[CrossRef][Web of Science][Medline]
  40. Schlecht NF, Franco EL, Pintos J, et al. Interaction between tobacco and alcohol consumption and the risk of cancers of the upper aero-digestive tract in Brazil. Am J Epidemiol (1999) 150(11):1129–1137.[Abstract/Free Full Text]
  41. Dal Maso L, La Vecchia C, Polesel J, et al. Alcohol drinking outside meals and cancers of the upper aero-digestive tract. Int J Cancer (2002) 102(4):435–437.[CrossRef][Medline]
  42. Ferraroni M, Decarli A, Franceschi S, et al. Validity and reproducibility of alcohol consumption in Italy. Int J Epidemiol (1996) 25(4):775–782.[Abstract/Free Full Text]
  43. Tobacco smoke and involuntary smoking. IARC Working Group, Lyon, 11–18 June 2002. IARC Monogr Eval Carcinog Risks Hum (2004) 83:1–1438.[Medline]
  44. Squier CA, Cox P, Hall BK. Enhanced penetration of nitrosonornicotine across oral mucosa in the presence of ethanol. J Oral Pathol (1986) 15(5):276–279.[CrossRef][Web of Science][Medline]
  45. Du X, Squier CA, Kremer MJ, et al. Penetration of N-nitrosonornicotine (NNN) across oral mucosa in the presence of ethanol and nicotine. J Oral Pathol Med (2000) 29(2):80–85.[CrossRef][Web of Science][Medline]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Am J EpidemiolHome page
J. H. Lubin, M. Purdue, K. Kelsey, Z.-F. Zhang, D. Winn, Q. Wei, R. Talamini, N. Szeszenia-Dabrowska, E. M. Sturgis, E. Smith, et al.
Total Exposure and Exposure Rate Effects for Alcohol and Smoking and Risk of Head and Neck Cancer: A Pooled Analysis of Case-Control Studies
Am. J. Epidemiol., October 15, 2009; 170(8): 937 - 947.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow Web Only Tables
Right arrow All Versions of this Article:
169/2/132    most recent
kwn306v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Purdue, M. P.
Right arrow Articles by Hayes, R. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Purdue, M. P.
Right arrow Articles by Hayes, R. B.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?