OBJECTIVE: To assess the 4-year anti-HBs persistence after revaccination with 3-dose of hepatitis B vaccine (HepB) among low-responsive adults. METHODS: A total of 24 237 healthy adults who had no history of hepatitis B infection and hepatitis B vaccination, resided in the local area for more than six months and were aged 18-49 years were selected from 79 villages of Zhangqiu county, Shandong province, China in 2009. Blood samples were obtained and hepatitis B surface antigen (HBsAg), antibody against hepatitis B surface antigen (anti-HBs) and antibody against hepatitis B core antigen (anti-HBc) were detected using ELISA method. A total of 11 590 persons who were negative for all of these indicators were divided into four groups by cluster sampling method. Each group was vaccinated with one of the following four types of HepB at 0-1-6 months schedule: 20 μg HepB derived in Saccharomyces cerevisiae (HepB-SC), 20 μg HepB derived in Chinese hamster ovary cell (HepB-CHO), 10 μg HepB-SC and 10 μg HepB derived in Hansenula polymorpha (HepB-HP). Blood samples were collected one month after the third dose of primary immunization and tested for anti-HBs using chemiluminescence microparticle immunoassay (CMIA). The 892 low-responders were revaccinated with three doses of HepB at 0-1-6 months schedule and the type of HepB was the same as which was used for primary immunization. During the follow-up to low-responders, the following informations were collected: the demographic characteristics (including age, gender), histories of hepatitis B infection, hepatitis B vaccination, smoking, drinking and chronic diseases. Blood samples were collected one month (T1) and four years after revaccination and anti-HBs, anti-HBc and HBsAg (if anti-HBs <10 mU/ml) were detected by CMIA. The risk factors associated with positive rate of anti-HBs and GMC of anti-HBs were identified by multiple logistic regression analysis and multifactor linear regression model analysis respectively. Anti-HBs titer at T1 was grouped according to the level and was considered as the independent variable in the model analysis. RESULTS: A total of 529 participants were identified from 892 low-responders. Among 529 participants, 276 (52.2%) were males and 253 (47.8%) were females. The positive rate was 82.6% (437/529) at T1 and it decreased to 28.2% (149/529) four years after revaccination. The corresponding GMC decreased from 542.06 (95% CI: 466.72-629.56) mU/ml to 27.69 (95% CI: 23.08-33.23) mU/ml. Multivariable analysis showed the positive rate of anti-HBs 4 years after revaccination was independently associated with anti-HBs titer at T1. The positive rate among those whose anti-HBs titer more than 1 000 mU/ml at T1 was significantly higher than those whose anti-HBs titer less than 100 mU/ml. The OR (95%CI) was 39.67 (13.81-114.01). The GMC was associated with HepB type for revaccination and anti-HBs titer at T1. The GMC among those revaccinated 20 μg HepB was significantly higher than those revaccinated 20 μg HepB-CHO, 10 μg HepB-SC and 10 μg HepB-HP. The b (95% CI) was -0.40 (-0.78--0.02), -0.57 (-1.01- -0.15) and -0.63 (-1.03- -0.23), respectively. The GMC among those whose anti-HBs titer 100-999 mU/ml and those whose anti-HBs titer ≥1 000 mU/ml at T1 were higher than those whose anti-HBs titer <100 mU/ml. The b (95% CI) was 0.93 (0.53-1.33) and 3.31 (2.88-3.73) respectively. CONCLUSION: Anti-HBs GMC decreased rapidly 4 years after revaccination among low-responsive adults, but still kept good protecion. The anti-HBs persistence after revaccination was associated with HepB type for revaccination and anti-HBs level of titer one month after revaccination.