The aim of this study was to determine risk factors for pharyngeal cancer and to propose 10 result-based preventive measures. It was a case-control study conducted in Madrid, Spain, with 232 consecutive patients diagnosed between January 1 1990 and December 31, 1995, sex- and age-matched with 232 control individuals with no oncological disease or history. By means of an interviewer-administered questionnaire, seven different epidemiological areas were surveyed, namely: (1) sociodemographic variables, (2) familial all-site cancer history, (3) medical history, (4) lifestyle (habits), (5) diet, (6) occupational exposure, and (7) non-occupational exposure. Of the great number of factors within each epidemiological area, the following were found to be risk factors after adjustment for tobacco smoking and alcoholic beverage drinking: (1) tobacco smoking, (2) alcoholic beverage drinking, (3) low and low-middle socioeconomic background, (4) low educational level, (5) rural milieu, (6) working, or having worked, as a manual worker in agriculture, (7) working, or having worked as a manual worker in building industry, (8) having an upper aerodigestive tract cancer familial history, (9) having a medical history of alcholism, low weight/malnutrition, gastroesophageal reflux or chronic obstructive bronchopneumonia, (10) low dietary intake of fruit, fruit juice, uncooked vegetables, dietary fibre-containing foods, fish and milk and dairy products, (11) high dietary intake of meat and fried foods, (12) deficient oral and dental hygiene, (13) abuse of black coffee, (14) abuse of 'carajillo' (a typical Spanish drink composed of black coffee and flambéed brandy), (15) occupational exposure to pesticides, solvents and dust of different origins. On the basis of our results and those reported by other authors, we put forward 10 measures for the prevention of pharyngeal cancer. However, due to the small size of the nasopharyngeal cancer subsample (n = 35, 15.08 per cent), our results as well as the preventive measures are to considered as referring uniquely to oropharyngeal and hypopharyngeal cancers. In addition, from descriptive statistical data inspection one can conclude that nasopharyngeal cancer is likely to bear risk factors different from those for oropharyngeal and hypopharyngeal cancers, thus nasopharyngeal cancer warrants specific epidemiological investigation with a sufficiently large patient sample.