Diet and body mass, and oral and oropharyngeal squamous cell carcinomas: Analysis from the IARC multinational case–control study
Tobacco and alcohol use are the main risk factors for oral and oropharyngeal cancers, yet, dietary habits may also be of importance. Data from a series of case–control studies conducted in 9 countries worldwide (1,670 cases and 1,732 controls) were used to investigate the role of several food groups and body mass index (BMI). Low BMI significantly increased the odds ratio (OR) of cancer more than 2-fold among ever- and never-tobacco users and ever- and never-alcohol drinkers. After adjustment for potential confounders, high intake of fruits and vegetables significantly reduced the OR of cancer compared to low intake among ever-tobacco users (OR 0.4, 95% confidence interval [CI] 0.3–0.6), although not among never-tobacco users (OR 1.1, 95% CI 0.6–2.0). Similarly, the protective effect of high fruit and vegetable consumption was present among ever-drinkers (OR 0.4, 95% CI 0.3–0.6), but not among never-drinkers (OR 1.0, 95% CI 0.6–1.6). In conclusion, low BMI increases the risk of oral cancer, and vegetables and fruits may modulate the carcinogenic effects of tobacco and alcohol. © 2005 Wiley-Liss, Inc.
Tobacco smoking and alcohol drinking are the major risk factors for oral and oropharyngeal squamous cell carcinomas (SCCs) and have been shown to account for over 90% of these cancers in several populations.1, 2 Paan chewing and use of betel-quid are also associated with these cancers, particularly in India.3 Beyond these established risk factors, poor diet, characterised by low fruit and vegetable intake and high meat and fat consumption, has also been related to increased oral SCCs.4, 5, 6, 7, 8, 9, 10, 11 In a summary analysis of all oral cancer case–control studies that investigated fruit and/or vegetable consumption, the International Agency for Research on Cancer (IARC) reported a significantly reduced risk of oral cancer among individuals with high intake of fruits and vegetables compared to individuals with low intake.12 While these findings were consistent across studies, doubt remains as to whether the confounding effects of the main risk factors, or even other important socioeconomic factors, had been adequately controlled for. It is known that individuals who smoke or drink commonly consume a diet low in fruits and vegetables.13 Because of the inability to rule out the potential for the confounding effects of tobacco on diet, high fruit and vegetable intake were characterised as “possibly protective against oral cancer”.12
Obesity or high body mass index (BMI), commonly associated with increased risk of cancer at several sites,14 was similarly investigated as a risk factor for oral cancer by the IARC. However, the data were deemed “inadequate” to reach conclusions regarding the cancer risk associated with BMI because of the paucity of available studies. However, some studies showed that high BMI was inversely associated with the risk of oral and oropharyngeal SCCs15 and one study suggested that leanness long antedates cancer diagnosis.16 Importantly, the few studies that investigated the effects of BMI on oral cancer by smoking habits showed that the inverse association was diminished among non-smokers,15, 16, 17, 18 again suggesting the potential for residual confounding effects of tobacco and alcohol consumption on diet within the strata of smoking and drinking.
It is difficult to determine whether diet and/or BMI act independently from established risk factors, or whether previous studies have been unable to adequately control for the confounding effects of tobacco and alcohol exposure. Therefore, the aim of this study was to investigate several aetiologic factors related to oral and oropharyngeal SCCs, including fruit and vegetable intake, and BMI, while controlling for the most important risk factors, tobacco and alcohol use.
Material and methods
Oral and oropharyngeal cancer cases and cancer-free controls from the IARC Multicenter Oral Cancer Study were selected for this analysis.19 Briefly, incident cases of oral (including lip [excluding external lip], base of tongue, other parts of the tongue, gum, floor of mouth, palate and other parts of the mouth) and oropharyngeal (oropharynx and tonsil) SCC, which henceforth will be referred to as ‘oral cancer’, were recruited from referral centres, surgery clinics and hospitals in 9 countries (Italy, Spain, Poland, Northern Ireland, India, Cuba, Canada, Australia and Sudan) from 1996 to 1999. Only histologically confirmed cases of SCC were included. Cancer-free controls were recruited from hospital settings in Italy, Spain, Poland, Cuba, Canada, Australia, Sudan and a community setting in Northern Ireland, and were either presenting at hospitals for cancer screening or were visitors of cancer patients in India; controls with alcohol and/or tobacco related diseases, such as chronic lung disease, cancer of the lung or liver or coronary heart disease, were excluded. Also diseases potentially related to long-term modifications of dietary habits (i.e., diabetes mellitus, cirrhosis, etc.) were not eligible. Major eligible admission diagnoses among controls included non-alcohol related fractures and sprains, orthopaedic diseases, benign neoplasms and acute surgical conditions (e.g., appendicitis). Controls were frequency-matched by age (in 5-year categories) and gender within each study site. A total of 1,670 oral cancer cases and 1,732 controls were included in this analysis. One-third of the study population was from India and most others were from Europe. The overall participation rates for cases and controls were 88.7 and 87.3%, respectively, and were similar when stratified by country.
Risk factor assessment
Information was collected via an interviewer-administered questionnaire on demographics, tobacco smoking and chewing histories (for India only20) and alcohol consumption patterns. Height and weight 2 years prior to diagnosis were queried at the time of interview. Diet was assessed for the period 2 years prior to disease onset for cases and 2 years prior to study enrolment for controls. A food frequency questionnaire was used to examine weekly consumption of several foods, including, milk, yoghurt, bread, pasta or rice, maize dishes, meat, fish, ham and salami and sausage, egg, cheese, potatoes, raw green vegetables and salads, cruciferae (broccoli, cabbage and brussel sprouts), carrots, fresh tomatoes, pulses, fresh fruit juice, apples or pears, citrus fruit, bananas, cakes and desserts. Dietary findings from the largest participating centres have been reported (Italy,21 India,22 Cuba,23 Spain24 and Poland25).
BMI was computed as the reported weight (in kilograms) divided by the reported height (in meters) squared, based on weight that was queried for the period 2 years prior to questionnaire administration. Within each country, individuals (cases and controls combined) were ranked on the basis of their BMI and divided into tertiles. Individuals within each tertile from each country were combined.
Dietary habits were assessed analysing the total weekly intake frequency of individual food items. Three summary measures were defined on the basis of the individual reported consumption of (i) any kind of fresh fruit (including apples, pears and citrus fruit), (ii) any kind of vegetable (raw green vegetables, cruciferae, carrots and tomatoes) and (iii) the total fresh fruit and vegetable intake. To define country-specific quartiles for individual food items and summary variables, individuals (cases and control combined) within each country were ranked on the basis of their reported weekly frequency of total consumption of each food item. Individuals within each quartile of each country were combined.
Unconditional logistic regression models were used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for BMI, each reported food item and the fruit and vegetable summary measures among oral cancer cases compared to cancer-free controls. All models included terms for the design variables (age, gender and study country), 3 categories of education (as a proxy for socioeconomic status): (i) none, (ii) 1–6 years, (iii) 7 or more years; 5 categories of tobacco smoking: (i) never, (ii) former, (iii) current <20 cigarettes per day, (iv) current 20–29 cigarettes per day and (v) current >29 cigarettes per day; 5 categories of tobacco chewing (including different combinations of tobacco, betel nut, areca nut and calcium hydroxide): (i) never, (ii) former, (iii) current <20 wads per day, (iv) current 20–29 wads per day and (v) current >29 wads per day; 5 categories of alcohol use: (i) never, (ii) former, (iii) current <7 drinks per week, (iv) current 7–20 drinks per week and (v) current >20 drinks per week; and finally, the total number of portions per week of all food items (as a categorical variable by country-specific quartile). This last variable was used to control for possible systematic differential reports by cases of diet [i.e., as a proxy measure to allow for total diet (energy) intake].26 In the categorical variable of tobacco use, any use of cigars or pipes was included in the highest category of tobacco use. As the effects of tobacco use were strong, an additional continuous variable of ‘number of cigarettes per day’ among current and former smokers was included to adjust for potential residual confounding. BMI was also controlled for in the analyses of food items.
For the fruit and vegetable summary measures, stratified analyses were performed among categories of tobacco smoking (never/former/current), alcohol drinking (never/former/current) and BMI (dichotomised at the country-specific median value into low versus high). Similar results were obtained for the summary measures among former and current smokers, as well, and among former and current drinkers, therefore, these strata were collapsed into ‘ever-users’. A sub-analysis to assess the impact of tobacco chewing, which was only assessed in India, was conducted. Similar results with regard to exposure variables were reported; therefore, the never-chewers who did not smoke were included in the ‘never-users’ tobacco category, while the ‘ever-users’ category represents those who smoked or chewed tobacco, or reported use of both forms of tobacco. Dose-response trends were evaluated for all ordinal variables, including quartiles of food summary measures and BMI; all p values for the linear trend tests are given in the tables. The effects of fruit and vegetable intake by tobacco use and alcohol drinking were evaluated by chi-square tests for heterogeneity.
Only 235 (14.1%) cases compared to 824 (47.6%) controls had never smoked or chewed tobacco. In total, 679 (40.8%) cases and 852 (50.7%) controls had never drunk alcoholic beverages. Current tobacco smoking (OR 4.0, 95% CI 3.3–4.9), paan chewing (OR 12.1, 95% CI 9.1–16.1) and former or current alcohol drinking (OR 2.2, 95% CI 1.8–2.7) significantly increased the odds of oral cancer. Compared to controls, cases were significantly less likely to have 7 or more years of education (43.4% of cases versus 57.1% of controls).
Cases were more commonly in the lowest category of BMI (43.7% versus 21.4% of controls, p < 0.001). Compared to individuals in the high BMI category, individuals in the low BMI category were almost 3-fold (95% CI 2.3–3.5) more likely to have oral SCCs (Table I). When stratified by tobacco use, significantly increased odds associated with low BMI compared to high BMI were consistently present among never- (OR 2.5, 95% CI 1.6–4.0) and ever- (OR 2.9, 95% CI 2.3–3.8) tobacco users, as well as never- (OR 2.6, 95% CI 2.0–3.5) and ever- (OR 3.2, 95% CI 2.2–4.6) drinkers (Table I). Significant trends were noted by decreasing tertiles of BMI for all strata (p ≤ 0.001, Table I).
Table I. Distribution of Cases and Controls1 and Corresponding ORs2 with 95% CIs for Cancer of the Oral Cavity and Oropharynx According to Approximate Country-Specific Tertiles of BMI
|Medium||454:513||1.5 (1.2–1.9)||1.7 (1.1–2.6)||1.5 (1.2–1.9)||1.5 (1.1–1.9)||1.6 (1.1–2.2)|
|Low||626:320||2.8 (2.3–3.5)||2.5 (1.6–4.0)||2.9 (2.3–3.8)||2.6 (2.0–3.5)||3.2 (2.2–4.6)|
|pfor trend|| ||<0.0001||0.0001||<0.0001||<0.0001||<0.0001|
Table II gives the distribution of oral cancer cases and controls according to selected foods and the corresponding ORs. Food items that significantly decreased the odds of oral cancer included fish, carrots, pulses, raw green vegetables, apples and pears, citrus fruit and fruit juices (Table II). Consumption of cruciferae reduced the odds of cancer in each of the quartiles of intake; however, the trend was not significant. Subjects in the medium-low and medium-high quartiles of ham and salami consumption had significantly elevated odds of oral cancer compared to individuals in the lowest category of consumption; a weaker association was present comparing the high to the low quartiles of intake (Table II). No effect was noted with intake of other food items, including yoghurt, bread, pasta or rice, maize, meat, eggs, cheese, potatoes or bananas (data not shown).
Table II. Distribution of Cases and Controls and Corresponding ORs1 with 95% CIs for Cancer of the Oral Cavity and Oropharynx by Approximate Country-Specific Quartile of intake of 12 Selected Food items and Summary Measures
| Milk||546:452||1.0||417:436||0.9 (0.7–1.2)||449:524||0.9 (0.7–1.2)||248:264||0.9 (0.7–1.2)||0.4223|
| Cheese||775:685||1.0||397:438||0.7 (0.5–0.9)||235:273||0.8 (0.6–1.1)||237:260||0.8 (0.6–1.1)||0.2644|
| Ham and salami||731:884||1.0||361:304||1.6 (1.2–2.0)||282:184||1.9 (1.4–2.5)||255:280||1.2 (0.9–1.6)||0.0331|
| Fish||507:403||1.0||517:567||0.7 (0.5–0.8)||314:382||0.6 (0.4–0.7)||322:326||0.8 (0.6–1.1)||0.0315|
| Carrots||576:381||1.0||483:526||0.7 (0.6–0.9)||370:380||0.8 (0.6–1.0)||227:391||0.6 (0.4–0.7)||<0.0001|
| Cruciferae||610:486||1.0||465:506||0.8 (0.6–1.0)||358:380||1.0 (0.8–1.2)||224:302||0.7 (0.5–0.9)||0.0616|
| Pulses||568:447||1.0||470:513||0.8 (0.6–1.0)||368:395||0.7 (0.5–0.9)||251:320||0.6 (0.5–0.8)||0.0003|
| Raw green vegetables||493:332||1.0||417:393||0.8 (0.6–1.0)||365:444||0.6 (0.5–0.8)||383:509||0.7 (0.5–0.9)||0.0004|
| Tomatoes||519:424||1.0||338:319||1.0 (0.7–1.2)||317:345||0.9 (0.7–1.2)||485:589||0.9 (0.7–1.2)||0.4596|
| Apples and pears||857:544||1.0||409:425||0.6 (0.5–0.8)||226:419||0.3 (0.3–0.4)||150:272||0.4 (0.3–0.5)||<0.0001|
| Citrus fruits||479:272||1.0||617:520||0.8 (0.6–1.1)||316:528||0.4 (0.3–0.6)||244:357||0.6 (0.4–0.7)||<0.0001|
| Fruit juices||903:783||1.0||429:466||0.9 (0.7–1.1)||147:195||0.7 (0.6–1.0)||163:224||0.8 (0.6–1.1)||0.0478|
| Fruit summary3||549:355||1.0||442:389||0.9 (0.7–1.1)||365:463||0.7 (0.5–0.8)||285:457||0.7 (0.5–0.9)||<0.0001|
| Vegetable summary4||509:410||1.0||451:350||1.1 (0.9–1.4)||395:445||0.9 (0.7–1.2)||298:468||0.7 (0.6–1.0)||<0.0001|
| Total fruit and vegetable||505:323||1.0||472:386||1.0 (0.7–1.2)||373:446||0.7 (0.6–0.9)||285:504||0.6 (0.4–0.8)||<0.0001|
Each of the 3 summary measures for fruit or vegetable intake was associated with significantly reduced odds of oral cancer and corresponding dose-response trends (Table II). Specifically, there was a 30% reduction in the odds of oral cancer among individuals in the high-intake quartile for both the fruit summary (OR 0.7, 95% CI 0.5–0.9) and the vegetable summary (OR 0.7, 95% CI 0.6–1.0). Total fruit and vegetable consumption also reduced the odds of oral cancer among those in the high-intake quartile compared to the low-intake quartile (OR 0.6, 95% CI 0.4–0.8, Table II). The effects of the fruit summary measure and the vegetable summary measure were independent of one another (correlation between 2 summary measures = 0.39, data not shown).
No strong correlations were found among different food items, as all of the correlation coefficients were <0.63. When stratified by case or control status, high consumption of fruits was chiefly due to consumption of apple and pears and citrus fruit. They were positively correlated among both cases and controls. Consumption of fruits was positively correlated with consumption of bread and cheese among cases and with ham and salami consumption among control women only. Vegetable consumption was not correlated with any of the food items considered; only among control women did we find a positive correlation of 0.52 with fruit consumption.
The protective effects of fruit and vegetable consumption were investigated in stratified analyses among categories of never- and ever-tobacco users, never- and ever-drinkers, and low and high BMI (dichotomised at the country-specific median value, Table III). Among never-tobacco users, high total fruit and vegetable consumption was not associated with oral cancer (OR 1.1, 95% CI 0.6–2.0). However, among ever-tobacco users (including former and current), being in the high-intake quartile of total fruit and vegetable intake significantly decreased the odds of oral cancer compared to individuals in the low-intake quartile of consumption (OR 0.4, 95% CI 0.3–0.6); a significant trend of decreasing odds of cancer with increasing total fruit and vegetable consumption was present (p < 0.001, Table III). The effect of fruit and vegetable intake by smoking status was significantly heterogeneous. Similar results were observed in the stratified analysis of never- and ever-drinkers. Among never-drinkers, there was no association with high fruit and vegetable intake and oral cancer (OR 1.0, 95% CI 0.6–1.6). Yet, among ever-drinkers, high intake compared to low intake of total fruits and vegetables seemed protective against oral cancer (OR 0.4, 95% CI 0.3–0.6, Table III). Heterogeneity by drinking status was significant for vegetables but not for fruits. Among both the low and high BMI strata, high intake of total fruit and vegetables significantly decreased the odds of oral cancer compared to low intake of total fruit and vegetables (low BMI: OR 0.5, 95% CI 0.3–0.7; high BMI: OR 0.5, 95% CI 0.3–0.8); significant trends of decreasing odds of cancer with increasing total fruit and vegetable consumption were present in both the low and high BMI strata (p < 0.001, Table III). This analysis was repeated stratified by tumour site; similar patterns of the protective effects of high intake compared to low intake of fruits and vegetables were found among both oral cavity and oropharynx cases (data not shown).
Table III. ORs1 and Corresponding 95% CIs for Cancer of the Oral Cavity and Oropharynx by Approximate Country-Specific Quartile of the Calculated Summary Measures for Fruit and Vegetable Intake, Stratified by Tobacco use, Alcohol Drinking and BMI2
| Medium- low||1.6 (0.9–2.7)||0.8 (0.6–1.0)||0.8 (0.5–1.2)||1.0 (0.7–1.3)||0.7 (0.5–0.9)||1.2 (0.8–1.7)|
| Medium-high||0.9 (0.5–1.5)||0.6 (0.5–0.8)||0.7 (0.5–1.0)||0.6 (0.5–0.9)||0.7 (0.5–1.0)||0.7 (0.5–1.0)|
| Highest||1.2 (0.7–2.1)||0.5 (0.4–0.7)5||0.8 (0.5–1.2)||0.6 (0.5–0.9)6||0.7 (0.5–1.1)||0.6 (0.4–0.8)|
|p for trend||0.9982||<0.0001||0.2199||0.0017||0.1360||0.0002|
| Medium-low||1.3 (0.8–2.2)||1.0 (0.8–1.4)||1.7 (1.1–2.5)||0.8 (0.6–1.1)||1.3 (0.9–1.8)||1.0 (0.7–1.4)|
| Medium-high||0.9 (0.6–1.6)||0.9 (0.7–13)||1.3 (0.9–2.0)||0.8 (0.6–1.1)||0.9 (0.7–1.4)||0.8 (0.6–1.2)|
| Highest||1.4 (0.8–2.5)||0.6 (0.4–0.8)5||1.2 (0.7–1.9)||0.6 (0.4–0.9)6||0.6 (0.4–1.0)||0.7 (0.5–1.1)|
|p for trend||0.5144||0.00021||0.6627||0.0072||0.0240||0.0611|
|Total fruit and vegetable intake|
| Medium-low||1.1 (0.6–1.9)||0.9 (0.6–1.1)||1.4 (1.0–2.1)||0.8 (0.5–1.0)||0.9 (0.7–1.3)||1.1 (0.7–1.5)|
| Medium-high||0.9 (0.5–1.6)||0.7 (0.5–0.9)||1.2 (0.8–1.7)||0.6 (0.4–0.8)||0.8 (0.5–1.1)||0.6 (0.4–0.9)|
| Highest||1.1 (0.6–2.0)||0.4 (0.3–0.6)5||1.0 (0.6–1.6)||0.4 (0.3–0.6)6||0.5 (0.3–0.7)||0.5 (0.3–0.8)|
|p for trend||0.8131||<0.0001||0.6941||<0.0001||0.0009||<0.0001|
While the majority of oral SCCs are explained by tobacco use and alcohol consumption, nutrition and diet may also play an important role in oral cancer risk. In this large multinational case–control study, we confirmed prior reports where, contrary to cancers at other anatomic sites where obesity is a major risk factor,27 low BMI was associated with increased oral cancer risk.15, 16, 28 This inverse association was present in the stratified analyses, including among never- and ever-tobacco users and among never- and ever-drinkers. Importantly, the questionnaires included information on height and weight in the 2-year period prior to cancer diagnosis (or control enrolment), thereby reducing the likelihood that low BMI was the result of cancer development. It is therefore more likely that this association is real, and may possibly be the result of long-term nutritional deficiency among oral cancer cases.16
Our study confirmed the protective effect of fruit and vegetables against oral cancer.5, 6, 7, 8 A large number of potentially anticarcinogenic agents are found in fruits and vegetables, including carotenoids, dithiolthiones, glucosinolates, indoles, isothiocyanates, protease inhibitors, plant sterols, allium compounds, limonenes and also vitamins C and E, selenium and dietary fibre.29 It is still not clear, however, which constituent(s) of fruits and vegetables are potentially responsible for reducing oral cancer risk. Folate is crucial for normal DNA synthesis and repair,29 and its deficiency may increase the risk of oral cancer by inducing an imbalance in DNA precursors.30 In particular, the decreased absorption of folate in the body can also be due to high intake of alcohol.30 Vegetables and fruits, especially citrus fruits, are also rich in flavonoids and polyphenols having antioxidant, antimutagenic and antiproliferative properties.31, 32
Fish also appeared as an indicator of decreased risk of oral cancer in the present study, and a favourable effect has also been reported at other cancer sites.33 Fish intake supplies n-3 polyunsaturated fatty acids that are incorporated into cell membranes and influence several biological responses,34 such as suppression of neoplastic transformation, cell growth inhibition, immune system and inflammation,35, 36 apoptosis and anti-angiogenesis.37
With respect to meat consumption, positive associations have been reported with different types of meat, in particular red meat and cured meat.7, 8, 9 We also found increased risk of oral cancer associated with consumption of ham and salami that may be related to both high salt and nitrite content or unfavourable effect of animal fats and cholesterol.7 However, the apparent favourable effect of fish and the direct association with ham and salami intake may, at least in part, represent an unspecific indicator of dietary patterns on oral cancer.
In the stratified analysis, we found that fruit and vegetable consumption was inversely related to oral cancer only among ever-tobacco users or ever-drinkers, but not among never-tobacco users or never-drinkers. This observation may be interpreted in 2 ways. First, the association is real, and the intake of foods (i.e., fruits and vegetables) known to be rich in nutrients and micronutrients protective against oral cancer modulate the carcinogenic effects of tobacco and alcohol. An alternative explanation may be that the protective effect is due to residual confounding of smoking and drinking among people who have these habits, since heavy smokers and drinkers tend to consume fruits and vegetables less commonly compared to non-smokers and non-drinkers.13 However, in our study, as in similar studies, oral cancer was very rare in people who had never consumed alcohol or, most notably, who had never used tobacco. Indeed the carefully alcohol- and tobacco-adjusted ORs we presented for fruits and vegetables are not materially different from the unadjusted ones (high intake of fruits and vegetables versus low intake OR 0.5, 95% CI 0.4–0.6).
This large study enabled detailed assessment of tobacco and alcohol history, as well as dietary assessment across a broad distribution of intake of several food items. We calculated country-specific quartiles to systematically combine individuals with high, moderate (in 2 levels) and low intake of each food item within each region. This was important to address the significant heterogeneity of consumption across cultures (test for heterogeneity of fruit and vegetable consumption by study country, chi-squared 23.52, p < 0.001). While using the same questionnaire to assess food intake from a variety of cultures (as opposed to a culture-specific questionnaire) has the potential to miss culture-specific food items, such as tropical fruits, employing homogenous methods to ascertain diet has advantages as well, including comparability of results. We were unable to adjust for total energy intake,26 as only selected foods were queried based on usual consumption, but we used total number of portions per week of all food items as a surrogate measure of individual total dietary intake.
In this large study, several strong associations between nutrition, diet and oral cancer risk were found, across a number of heterogeneous populations from developing and developed countries. This confirms that, after tobacco and alcohol use, poor diet is one of the most important risk factors in oral cancer worldwide.
The authors thank Ms. Trudy Perdrix-Thoma for editing the manuscript and preparing it for submission.
In addition to the aforementioned authors, the following investigators contributed to the study: Nubia Muñoz, International Agency for Research on Cancer, Lyon, France; Javier Pintos, McGill University, Montreal, Canada; Peter J.F. Snijders and Chris J.L.M. Meijer, Vrije Universiteit Medical Center, Amsterdam, The Netherlands; Raphael Viscidi, The Johns Hopkins University School of Medicine, Baltimore, Maryland; Frank Kee, Queen's University Belfast, Northern Ireland, United Kingdom; Leticia Fernández, Instituto Nacional de Oncología y Radiobiología, Havana, Cuba; Ali Idris, Toombak and Smoking Research Center, Khartoum, Sudan; María José Sánchez, Escuela Andaluza de Salud Pública, Granada, Spain; Adoración Nieto, Facultad de Medicina, Sevilla, Spain; F. Xavier Bosch, Institut Català d'Oncologia, Barcelona, Spain; Renato Talamini, Centro di Riferimento Oncologico di Aviano, Aviano, Italy; Alessandra Tavani, Ricerche Farmacologiche ‘Mario Negri’, Milan, Italy; Michael Pawlita, Ulrich Reidel, Deutsches Krebsforschungszentrum, Heidelberg, Germany; Jolanta Lissowska, Cancer Center, Warsaw, Poland; Barbara Rose, Sydney Head and Neck Cancer Institute, Royal Prince Alfred Hospital, Sydney, Australia; Hema Sridhar, Kidwai Memorial Institute of Oncology, Bangalore, India; Prabda Balaram, Regional Cancer Center, Trivandrum, India; Thangarajan Rajkumar, Cancer Institute (WIA), Chennai, India.
Supported by grant S06 96202489 05F02 from Europe Against Cancer (to the International Agency for Research on Cancer, Lyon, France [N. Muñoz and R. Herrero]) and grants FIS 97/0662, FIS 01/1236 and BAE 01/5013 (to the Institut Catala' d'Oncologia, Barcelona, Spain [X. Castellsagué and F. Xavier Bosch]) from “Fondo de Investigaciones Sanitarias” (FIS), Madrid, Spain. Funding was also provided by the International Union Against Cancer (UICC) Yamagiwa-Yoshida Memorial International Cancer Study (to X. Castellsagué), the National Cancer Institute of Canada (to McGill University, Montreal, Quebec, Canada [E. Franco]), the Italian Association for Research on Cancer (to the Centro di Riferimento Oncologico di Aviano, Aviano, Italy [S. Franceschi]) and the Pan American Health Organization (to the Instituto Nacional de Oncología y Radiobiología, Havana, Cuba [L. Fernández]).