Design: Two hospital-based case-control studies were conducted in central and eastern Europe and Latin America. CASE / CONTROL SELECTION: Cases and controls were recruited in Moscow (Russia), Bucharest (Romania) and Lodz and Warsaw (Poland) from 1998 to 2002, and from 1998 to 2003 in Buenos Aires (Argentina), La Habana (Cuba) and the Brazilian towns Rio de Janeiro, São Paulo, Pelotas, Porto Alegre and Goiânia. Incident cases of squamous cell carcinoma of the head and neck (oral cavity, pharynx, larynx) and oesophagus, as well as age- (in quinquennia) and sex frequency-matched controls, were enrolled from 1998 to 2003. Control subjects included residents of the study areas for at least 1 year who were admitted to the same hospitals as the cases or in a comparable catchment area (controls in São Paulo were not recruited from oncology hospitals, and population-based controls were enrolled in Warsaw). Controls were hospitalised for prespecified conditions thought to be unrelated to tobacco use or alcohol consumption. Both studies were coordinated by the International Agency for Research on Cancer according to an identical protocol for both case and control recruitment.
Data analysis: Data from the central European and Latin American studies were analysed separately. Multivariable logistic regression models, including terms for oral health indicators, age, sex, education, country (central Europe) or centre (Latin America), tobacco pack-years and cumulative alcohol consumption, were used to obtain odds ratios (OR) and 95% confidence intervals (CI). All oral hygiene indicators and covariates were analysed as categorical variables, except for age, cumulative alcohol consumption, tobacco pack-years, and age when full-time education was completed, which were analysed as continuous variables. Case-control comparisons were made using t tests for continuous variables and chi-square tests for categorical variables; two-sided P values were obtained. Effect modification was assessed by stratifying on smoking (never/ former/ current) and drinking (never/ ever) habit. Heterogeneity between centres was tested.
Results: Risk factors for head and neck cancer, independent of tobacco use and alcohol consumption, were as follows: poor condition of the mouth [central Europe OR, 2.89 (95% CI, 1.74-4.81); Latin America OR, 1.89 (95% CI, 1.47-2.42)]; lack of toothbrush use [Latin America OR, 2.36 (95% CI, 1.28-4.36)], and daily mouthwash use [Latin America OR, 3.40 (95% CI,1.96-5.89]. Missing six to 15 teeth was an independent risk factor for oesophageal cancer [central Europe OR, 2.84 (95% CI, 1.26-6.41); Latin America OR, 2.18 (95% CI, 1.04-4.59)].
Conclusions: These results indicate that periodontal disease (as indicated by poor condition of the mouth and missing teeth) and daily mouthwash use may be independent causes of cancers of the head, neck and oesophagus.