In order to explore multiple risk factors of nasopharyngeal carcinoma (NPC). a total of 9775 male adult residents in six townships were studied. Sociodemographic characterlsties. history of cigarettes smoking and alcoho1 drinking, dietary habits, personal history of sinusitis and family history of NPC were collected by well-trained interviewers with structured questionnaires at the recruitment. Blood samples were also obtained and tested for immunoelobulin A antibodies against viral capsid antigen (anti-EBV VCA/lgA) of Epstein-Barr virus (EBV) and antibodies to EBVspecific DNase (anti-EBV DNase). A total of 13 new NPC cases occurred during an average follow-up period of 8.1 years. Cox's propoitional hazard models and multipie logistic regression models were used to analyze multiple risk factors for NPC. The multivariate-adjusted relative risk of developing NPC for anti-EBV VCA/lpA was 40.6 (95%CI 7.5-220.1) for a titer equal to 1:10. and 726.7 (95%CI 177.9-2967.7) for a titergreater than or equal to 1:40. The multivariate-adjusted relative risk for anti-EBV DNase was 3.4 (95%CI 0.4-32.9) for a level of 2.0-3.9 units. 34.1 (95%CI 6.8-171.1) for 4.0-5.9 units, and 113.6 (95%CI 27.7-465.9) for 6.0 or more units. Stratification analysis of NPC cases diagnosed within and after first two years of follow-up, the significantly high relative risks associated with elevated EBV antibodies remained. After adjustment for age and anti-EBV antibody level, consumption of fermented soy bean products. family history of NPC, and personal history of sinusitis were also associated with the development of NPC. Interactions were found between seropositivity of anti-EBV DNase and cigarette smoking and fermented soy bean products consumption.