Sa1483 - Real-World Rates of Hepatitis B Surface Antigen (HBSAG) Seroclearance in Patients with Chronic Hepatitis B (CHB): A Systematic Review, Conventional Aggregated Data Meta-Analysis (ADMA) and Individual Patient Data Meta-Analysis (IPDMA)
Yee Hui Yeo,Hsiu J. Ho,Hwai-I Yang,Tai-Chung Tseng,Min-Sun Kwak,Young Min Park,James Y. Fung,Maria Buti,Manuel Rodriguez,Sombat Treeprasertsuk,Carmen M. Preda,Teerapat Ungtrakul,Phunchai Charatcharoenwitthaya,Xiangyong Li,Michael H. Le,Bin Wei,Biyao Zou,An K. Le,Donghak Jeong,Nicholas Chien,Leslie Kam,Tetsuya Hosaka,Mar Riveiro,Doina Proca,Fumitaka Suzuki,Mariko Kobayashi,Tassanee Sriprayoon,Yutian Chong,Tawesak Tanwandee,Man-Fung Yuen,Hyo-Suk Lee,Jia-Horng Kao,Anna Lok,Chun-Ying Wu,Mindie H. Nguyen
DOI: https://doi.org/10.1016/s0016-5085(18)33750-8
IF: 29.4
2018-01-01
Gastroenterology
Abstract:Background: HBsAg seroclearance is a rare but important marker of CHB functional control.However, there is considerable variability in HBsAg clearance rates across published datasets.Our goal was to perform ADMA and IPDMA to estimate pooled annual and cumulative HBsAg seroclearance rates and to identify their predictors.Method: Two researchers independently searched PubMed and Embase from inception to July 5, 2017 for relevant studies and extracted data.Authors of all included studies were contacted for original individual participant level data.We estimated pooled rates of HBsAg seroclearance using random effects model in ADMA and Cox proportional-hazard models with a cluster statement in IPDMA.This study with pre-planned subgroup analyses has been registered in PROSPERO.Results: Of the 4469 abstracts and articles reviewed, 26 full-article studies (n=39,661) were included.ADMA showed a pooled annual HBsAg seroclearance rate of 1.11% (95% CI 0.84-1.37%).Individual data from ten cohorts (total: 12,950 patients, 8,065 males and 4,885 females) from the US, Europe and Asia were obtained and harmonized.Two cohorts (n=6,285) included untreated patients only and four cohorts (n=1,697) recruited treated patients only.There were three cohorts (n=4,968) that included both treated and untreated patients.There were 1,205 patients who showed HBsAg seroclearance over 107,466.4person-years of followup, yielding an annual incidence rate of 1.12% (95%CI: 1.06-1.19%).The 5-, 10-, 15-, and 20-year pooled cumulative rates were 3.95%, 9.82%, 17.32%, 22.44%, respectively (Figure 1A).Multivariate Cox regression showed that males (adjusted hazard ratio [aHR]:1.17,95%CI:1.02-1.34)and age >55 years (vs.<35 years) (aHR:1.77,95%CI:1.45-2.15)were significant independent predictors for higher rates of HBsAg seroclearance, while cirrhosis at entry (aHR:0.59,95%CI:0.35-0.99),positive HBeAg (aHR:0.33,95%CI:0.26-0.42),HBV DNA>20,000 IU/mL (aHR:0.31,95%CI:0.25-0.39)and quantitative HBsAg (qHBsAg)>1200 IU/mL (aHR:0.2,95%CI:0.17-0.24)were independently associated with lack of HBsAg seroclearance (Table 1).However, Asian (vs.non-Asian) and population-based community (vs.clinical center) study setting were not.Conclusion: The pooled annual HBsAg seroclearance rate in this multi-national study of 12,950 persons with CHB was approximately 1.1% from both ADMA and IPDMA.Male sex, older age, HBeAg negativity, non-cirrhosis, lower HBV DNA and qHBsAg were independent predictors of HBsAg seroclearance, but not race or study setting (population-based community vs. clinical cohorts).