ObjectivesTo classify the high risk population for primary liver cancer (PLC) in rural China.MethodsBetween June 2011 and June 2013, hepatitis B surface antigen (HBsAg) carriers were identified in clinical laboratory of Qidong People’s Hospital and surveyed by questionnaires. Cox proportional hazard regression model was introduced to demonstrate independent risk factors associated with PLC occurrence. Moreover, receiver characteristic operating (ROC) curve was utilized to evaluate discrimination power of risk factor panel for PLC risk classification.ResultsA total of 1 296 HBsAg carriers were enrolled, among which 686 participants were male with an average age of 45.73±11.58 years, and 610 participants were female with an average age of 45.67±12.33 years. After a mean follow up period of 5.5 years, 43 incident PLC cases were confirmed, which generated a PLC incidence of 60.5 millions person years. Multi-univariate Cox model showed that increase of age (HR=1.055, 95%CI 1.029 to 1.083, P<0.000 1), male (HR=3.263, 95%CI 1.567 to 6.796,P=0.001 6), having family history of PLC (HR=2.315, 95%CI 1.260 to 4.252, P=0.006 8), HBeAg positivity (HR=2.367, 95%CI 1.267 to 4.419, P=0.006 9) and GGT abnormality (HR=2.721, 95%CI 1.457 to 5.083, P=0.001 7) were the independent risk factors of PLC.ConclusionRoutine host, viral and liver biochemical parameters which are readily accessible in daily clinical practice can be utilized in identification of the targeted population for prevention of PLC in rural China.