Esophageal cancer is a leading cause of cancer death, especially in developing countries. In high-risk regions, squamous cell carcinoma is the most common type of esophageal cancer, and its etiology remains poorly understood. The purpose of this study was to evaluate the association between human papillomavirus (HPV) infection and esophageal squamous cell carcinoma (ESCC) and related precursor lesions in a high-risk area of China. We conducted a cross-sectional study among adult inhabitants of Linxian, China. All subjects were interviewed about potential risk factors, had the length of their esophagus sampled by a balloon cytology examination and underwent endoscopy with mucosal iodine staining and biopsy of all unstained lesions. A multivalent HPV hybridization probe, Digene Hybrid Capture II (Gaithersburg, MD), which recognizes high-risk types 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59 and 68, was used to determine the HPV infection status of the cytologic specimens, and the endoscopic biopsies were used to classify each subject's esophageal disease. 740 subjects completed the cytologic and endoscopic exams, and 702 had adequate cytologic and biopsy specimens. Using a cutpoint of > or =3.0 pg/ml of HPV DNA to define a positive test, HPV positivity was identified in 13% (61/475) of subjects without squamous dysplasia, 8% (8/102) with mild dysplasia, 7% (6/83) with moderate dysplasia, 16% (6/38) with severe dysplasia and zero (0/4) with invasive ESCC. Changing the cutpoint defining a positive test did not change the association of HPV infection and dysplasia grade. In this high-risk population, infection of esophageal cells with high-risk HPV types occurs in 13% of asymptomatic adults with no evidence of squamous dysplasia and a similar proportion of individuals with mild, moderate or severe dysplasia. This suggests that HPV infection is not a major risk factor for ESCC in this high-risk Chinese population. Further studies are warranted to determine if infection with this agent is associated with neoplastic progression in a subset of cases.