Hepatocellular carcinoma risk reduction associated with antiviral treatment in patients with noncirrhotic chronic hepatitis B: a systematic literature review and comparative meta-analysis Authors: Arpan Mohanty, Alice E. Stead, Laura E. Telep, Amanda W. Singer, Tristan Curteis, Anna Zolotor, Catherine Frenette, Dona Khoshabafard, Kyle Hammond, Tom McQuaid, Robin K. Kelley Published: 2026 ABSTRACT BackgroundChronic hepatitis B (CHB) affects nearly 300 million people worldwide and is a leading cause of hepatocellular carcinoma (HCC). While cirrhosis is a risk factor for HCC, hepatitis B virus (HBV) DNA integration can lead to HCC in non-cirrhotic individuals. This study synthesized existing data to compare HCC risk among non-cirrhotic patients treated with antivirals versus those untreated.MethodsA systematic literature review was conducted. Databases were searched in February 2023 and handsearches were performed to identify studies reporting HCC incidence in non-cirrhotic CHB patients, both treated with antivirals for a mean or median duration of at least three years and untreated. A meta-analysis compared HCC incidence rates between treated and untreated populations.ResultsTwo thousand, one hundred sixty-four records were screened and 61 publications on 50 studies were included. HCC cumulative incidence rates increased with greater follow-up, ranging from 0 to 3% at one year and 0 to 30% at 10 years. Incidence rates ranged from 0 per 100 to 2.43 per 100 person-years. The comparative meta-analysis of four studies indicated that treated populations had a 40% reduction in HCC incidence compared to untreated groups (incidence rate ratio [IRR]: 0.59; 95% CI: 0.50, 0.69), with incidence rates expressed per 100 person-years. Sensitivity analyses confirmed the robustness of these findings.ConclusionsAntiviral therapy significantly reduces HCC risk in non-cirrhotic CHB patients. These results, with other literature, support treatment guideline expansion to include non-cirrhotic individuals and prevent HCC development. Further research is necessary to evaluate the impact of early versus late treatment initiation and explore potential differences between antivirals.Systematic review registrationPROSPERO ID: CRD42023396260. Date of registration: 11 April 2023.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12876-026-04996-y. Introduction Chronic hepatitis B (CHB), defined as the presence of hepatitis B surface antigen (HBsAg) for at least 6 months after hepatitis B virus (HBV) infection, is a significant global health challenge that affects nearly 300 million people worldwide [1]. CHB is assessed through key markers such as HBsAg, HBV DNA levels, alanine aminotransferase (ALT), and liver disease progression. Hepatitis B e-antigen (HBeAg), also used in CHB assessment, is a marker of HBV replication, while the presence of antibodies against HBeAg is an equally important marker of partial immune control [2]. Long-term complications of CHB infection include liver failure, advanced liver disease, and hepatocellular carcinoma (HCC), all of which are associated with high morbidity and mortality [3]. CHB accounts for over half of HCC cases globally [4]. Cirrhosis is a well-established risk factor for HCC across all major causes of liver diseases, including chronic viral hepatitis [5, 6]. However, HBV is unique in that early host cell chromosomal integration of viral DNA contributes to HCC development even in non-cirrhotic individuals, with HBV DNA integrations found in approximately 80% of HCC tumor tissue [7]. HBV is designated as a group 1 carcinogen by the International Agency for Research on Cancer (IARC) [8] and is carcinogenic and mutagenic to host cell DNA [7]. HBV can drive oncogenesis through the production of viral proteins that disrupt normal cell growth, through DNA integrations which facilitate chromosomal translocations, and by promoting inflammation that results in cell turnover and clonal expansion [7]. These processes contribute to genomic instability and malignant transformation [7, 9]. There is evidence that viral DNA integration and concomitant chromosomal translocations occur within hours of infection, and initial studies evaluating the impact of antiviral treatment have shown reduction of integrations and reduced expansion of mutant clone populations [10, 11]. This likely explains in part why long-term antiviral therapy with nucleos(t)ide analogues (NAs) has demonstrated the ability to mitigate HCC risk, particularly when initiated early in the disease course and maintained over a significant duration of time [12, 13]. The critical role of DNA damage in malignant transformation and tumor progression underscores the need for early interventions to counteract HBV’s mutagenic potential and mitigate HCC risk. Current international treatment guidelines from the European Association for the Study of the Liver (EASL) [2], the American Association for the Study of Liver Diseases (AASLD) [14], and the Asian Pacific Association for the Study of the Liver (APASL) [15] recommend initiating antiviral therapy in cirrhotic individuals with CHB. The EASL 2025 guidelines recognize the potential benefits of earlier treatment in all HBV DNA-positive individuals to prevent viral replication and subsequent liver damage, while acknowledging there is less conclusive evidence supporting a significant reduction in HCC risk with NA therapy in non-cirrhotic populations due to limited long-term follow-up data and low event incidence [2]. Overall, treatment guidelines continue to focus primarily on treating cirrhotic individuals with CHB, while guidelines for management of non-cirrhotic individuals are more complex and take into consideration factors such as viral load and ALT values [7]. High-quality synthesis of evidence of HCC risk reduction attributable to antiviral therapy may elucidate the benefits of treatment expansion for individuals with CHB, specifically in non-cirrhotic, immune-tolerant, or younger individuals earlier in disease. Conducting randomized controlled trials (RCTs) to evaluate the efficacy of antiviral therapy in reducing HCC incidence is potentially unethical and impractical due to the long latency period for HCC development, which requires extended follow-up, incurs high costs, and faces challenges due to heterogeneous and evolving clinical practices. Thus, this de novo, clinically focused systematic literature review (SLR) considers both RCTs and observational studies to identify the incidence rates of HCC in individuals with non-cirrhotic CHB, stratified by those treated and untreated with antivirals. A meta-analysis was performed to directly compare the incidence of HCC between individuals who were treated with antiviral therapies and individuals who were untreated within the same study to evaluate the hypothesis that antiviral treatment reduces the long-term risk of HCC in non-cirrhotic populations. This study is intended to help inform clinical decision-making for individuals with non-cirrhotic CHB by examining the benefits of initiating antiviral treatment prior to progression to advanced liver disease. Methods Search strategy and selection criteria The SLR was conducted according to a prespecified protocol which was registered in PROSPERO (ID: CRD42023396260). Searches were performed in February 2023 across electronic databases including MEDLINE, MEDLINE In-Process, MEDLINE Epub Ahead of Print, Embase, The Cochrane Database of Systematic Reviews (CDSR) and the Cochrane Central Register of Controlled Trials (CENTRAL), restricting results to literature published from 2002 onwards, to identify primary research relating to the study objective. The search strategies used are shown in Tables S1-S4. Following the initial searches, Ovid, which was used to search MEDLINE, MEDLINE In-Process, MEDLINE Epub Ahead of Print, and Embase, announced that some Embase records had been missing. An additional search was performed in March 2023 to identify relevant records which were missing from the Ovid platform during the initial searches (Table S5). The bibliographies of all relevant SLRs and network meta-analyses (NMas) included from the database searches were hand-searched, as were select conference proceedings from the previous 2 years, to identify any additional relevant studies (Table S6). Each title and abstract, and subsequently each full-text article, was assessed for inclusion by two independent reviewers. Discrepancies in inclusion were resolved through discussion until consensus was reached, with a third independent reviewer making the final decision if necessary. Studies eligible for inclusion in the SLR were peer-reviewed, published in English, reported HCC incidence in children or adults with non-cirrhotic CHB based on any combination of clinical assessment, non-invasive testing/imaging, or biopsy, and included populations untreated with antivirals or treated with antivirals for at least 3 years, reported as either mean or median treatment duration (excluding mixed cohorts), with any of the following interventions: tenofovir alafenamide (TAF)/Vemlidy®, tenofovir disoproxil fumarate (TDF)/Viread®, entecavir (ETV)/Baraclude®, telbivudine (LdT)/Tyzeka®, adefovir dipivoxil/Hepsera®, clevudine/Levovir®, or lamivudine (LAM)/Epivir-HBV/Zeffix/Heptodin®. Included records were examined for overlapping populations prior to data extraction. Source data analysis Data extraction for included studies was performed in line with guidelines from the York University Centre for Reviews and Dissemination (CRD) [16]. Study characteristics, baseline characteristics of the patient population, and relevant efficacy, safety, and patient-reported outcomes were extracted into a pre-specified Microsoft Excel extraction grid by one reviewer, and verified by a second, independent reviewer. Any discrepancies were discussed and, if necessary, a third individual was enlisted to arbitrate the final decision. The quality of included studies was assessed by one individual and verified by a second using the assessment tool developed by the Joanna Briggs Institute (JBI) for prevalence data, modified to account for the diagnosis of HBV infection and HCC in the patient population (Table S8) [17]. Based on studies identified in the SLR, the feasibility of performing meta-analyses to evaluate the impact of longer-term antiviral treatment on the risk of HCC in individuals with non-cirrhotic CHB, as well as simply the reduction in HCC risk among treated versus untreated populations, was assessed. A thorough assessment of heterogeneity investigated differences in study, patient, and treatment characteristics among the included studies. Additionally, comprehensive investigations of available data on the incidence of HCC were carried out, with both cumulative incidence and incidence rates explored. Based on the results of the feasibility assessment, an unadjusted meta-analysis was performed on a subset of studies reporting on both treated and untreated populations of patients with CHB to estimate the incidence rate ratio (IRR) and incidence rate difference (IRD) of HCC between patients treated with NAs (who were treatment-naïve at baseline) and untreated patients. Reported (and, where appropriate, derived) study-level values for person-time and number of HCC events were used as input data for the meta-analysis. Methods for derivation of meta-analysis inputs are shown in Supplementary File 1. The meta package (version 6.5-0) was used in R (version 4.3.0) [18, 19]. The metainc function was used for this analysis with both fixed effect and random effects estimates produced for the IRR and the IRD. The default settings of the meta package were used. The Mantel-Haenszel method was used for pooling [20], and 95% confidence intervals (CIs) were generated for each estimate. I2 was calculated to assess statistical heterogeneity across studies. In order to identify potential outliers, a Baujat plot was generated using the baujat function in the meta package [21]. A sensitivity analysis was then conducted to investigate the results of the direct comparison meta-analysis excluding an identified outlier. Results SLR study characteristics A total of 2,803 records were retrieved by the electronic database searches. After de-duplication of results, 2,164 were suitable for review. After title and abstract review, 319 records were selected for full text review. Of these, 56 fulfilled the eligibility criteria for inclusion. Supplementary searches of congresses and SLR bibliographies identified five additional records. In total, 61 publications reporting on 50 unique studies were included in the SLR, shown in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) diagram (Fig. S1). A summary of study characteristics of included studies is shown in Table 1. The majority were retrospective cohort studies (n = 33), while 11 were prospective cohort studies, three were national database/registry studies, three were RCT open-label extension studies, and one was an RCT. Tseng 2017 included both retrospective and prospective cohort designs [64]. Most studies (n = 39) were conducted in Asia, with South Korea being the most represented country (n = 16 studies). Japan (n = 9) and Taiwan (n = 8) were the second and third most frequently represented countries. The remaining studies were multinational (n = 7) or conducted in China (n = 6), Italy (n = 3), or France (n = 1). Sample sizes for included studies ranged from 39 to 25,940 participants, with 14 studies including 1,000 or more participants and only six studies reporting fewer than 100 participants. Most studies demonstrated appropriate methodology, with an adequate sample size in 49/50 studies and 44/50 studies using valid methods to identify HCC development (Table S8). Table 1Summary of study characteristicsStudyStudy designCountryStudy periodSource of dataTreatments evaluatedSample sizeaDefinition of cirrhosis statusAndo 2018 [22]Retrospective cohortJapanJan 2000–Jan 2013bNagoya University HospitalLAM; LAM + ADV; ETV; ETV + ADV; TDFN = 105Cirrhosis was diagnosed by histological analysis and/or appearance of cirrhotic liver on ultrasonography, for example, superficial nodularity, coarse parenchymal echo pattern, signs of portal hypertension (splenomegaly > 120 mm, diameter of portal vein > 12 mm, patent collateral veins, or ascites), and/or the presence of gastroesophageal varices on endoscopyChiu 2020 [23]Retrospective cohortTaiwanJan 2007–Jan 2016bChang Gung Memorial HospitalNo treatment (ETV or TDF cessation)N = 460Noncirrhosis was diagnosed based on biopsy, Fibrosis-4 or liver stiffness measurement at baseline, combined repeated ultrasound findings without clinical features, such as splenomegaly, ascites, or gastroesophageal varices at baselineCho 2014 [24]Retrospective cohortSouth KoreaJan 2007–Jan 2012bSamsung Medical CenterETV; LAM; CLV; sequential or combination NA therapy; no treatmentN = 1,817Cirrhosis was determined by liver biopsy or imaging modalities with findings of irregular surface nodularity, parenchymal or morphological abnormality and manifestations of portal hypertension designated as liver cirrhosisChoi 2020 [25]Retrospective cohortSouth KoreaJan 2007–Dec 2016cAsan Medical CenterETV; TDFN = 2,612Cirrhosis was defined by the presence of any of the following findings: coarse liver echotexture and nodular liver surface on ultrasonography, clinical features of portal hypertension (e.g., ascites, splenomegaly, or varices), or thrombocytopenia (< 150,000/mm3)Choi 2022a [26]Retrospective cohortSouth KoreaJan 2000–Apr 2020cAsan Medical Center; Seoul National University Hospital; Severance Hospital; Samsung Medical CenterETV; TDF; no treatmentN = 2,073Cirrhosis was defined as the presence of any of the following characteristics: coarse liver echotexture and nodular liver surface on ultrasonography, clinical features of portal hypertension (e.g., ascites, splenomegaly, and varices), and platelet count below 100,000/µLChoi 2022b [27]Retrospective cohortSouth Korea; Hong Kong; TaiwanNRNRETV; TDFN = 7,545NRChun 2021 [28]National registry/ databaseSouth KoreaJan 2012–Dec 2018cNRETV; TDFN = 7,602Cirrhosis was defined as ICD-10 code K74Dezanet 2021 [29]Prospective cohortFranceJan 2002–Jan 2018cFour centers in FranceTDF-based ART; TAF-based ARTeN = 155Liver fibrosis was assessed using the FibroTest® calculated from standard battery of biochemical markers. Scores ranged from 0 to 1. METAVIR equivalents of this measure, as established in the HIV-HBV coinfected population, were used to grade liver fibrosis (F2: 0.48–0.58, F3: 0.59–0.73, F4: ≥0.74)Hosaka 2010 [30]Retrospective cohortJapanSep 1995–Sep 2007bToranomon HospitalLAM + ADVN = 186Cirrhosis was diagnosed by laparoscopy or liver biopsy, and in the patients who had not received them, by clinical data, imaging modalities, and portal hypertensionHosaka 2013 [31]Retrospective cohortJapanJan 1973–Dec 2011cToranomon HospitalETV; LAM; no treatmentN = 601Cirrhosis was determined based on laparoscopy, liver biopsy, imaging modalities, or portal hypertensionHosaka 2022 [32]Retrospective cohortJapanNR–Dec 2020dToranomon HospitalETVN = 132Cirrhosis was determined by laparoscopy, liver biopsy, imaging modalities, or portal hypertensionJi 2022 [33]RCT open-label extensionChinaNov 2013–NR9 first-class hospitals (not specified)ETV; ETV+Biejia-RuanganN = 500NRKim 2015 [34]RCT open-label extensionMultinationalJun 2005–Jan 2016NRTDF; ADV followed by TDFN = 482Cirrhosis was defined as Ishak fibrosis scores of 5 or 6Kim 2017a [35]Retrospective cohortSouth KoreaJan 2007–Dec 2015cAsan Medical CenterETVN = 850Cirrhosis was diagnosed based on histological and/or radiographical findings with evidence of portal hypertension (e.g., splenomegaly, ascites, or varices) in the presence of significant thrombocytopenia (platelet count < 150,000/mm3)Kim 2017b [36]Retrospective cohortSouth KoreaJan 2007–Dec 2015cSamsung Medical CenterETVN = 432Cirrhosis was clinically defined when patients showed cirrhotic configuration of the liver on imaging studies (nodular liver surface or caudate lobe hypertrophy) with thrombocytopenia (< 150 × 103/L) or splenomegaly (by imaging) and/or by the presence of varices (by upper endoscopy or imaging studies)Kim 2018 [37]Retrospective cohortSouth KoreaJan 2007–Apr 2017dUlsan University HospitalETV; TDFN = 198Cirrhosis was determined via liver biopsy or clinically defined in the absence of liver biopsy based on repeated liver imagingstudies (nodular liver surface or caudate lobe hypertrophy) with thrombocytopenia (< 150 × 1000/mm3) or splenomegaly (by imaging), and/or by the presence of varices (by upper endoscopy or imaging studies)Ko 2019 [38]Prospective cohortHong KongJan 2002–Dec 2015dQueen Mary HospitalETVN = 1,033Cirrhosis was defined by the following combined parameters: (a) ultrasonographic evidence of small-sized liver with or without development of splenomegaly/ascites/varices; and (b) APRI = (AST/upper limit of normal)/platelet count (x109L) x 100) greater than 2Kobashi 2011 [39]Prospective cohortJapanNov 2000–Nov 2010dNRETV; LAMN = 194Cirrhosis diagnosis was confirmed by liver histology, clinical signs (encephalopathy or ascites), evidence of esophagogastric varices by endoscopy, or imaging modalities such as ultrasonography, CT, or MRIKozuka 2019 [40]Retrospective cohortJapanSep 2006–June 2017cOsaka City University HospitalETVN = 105Liver cirrhosis was diagnosed by histological examination according to the METAVIR scoring system using imaging modalities such as ultrasonography, CT, or MRI and by signs of portal hypertensionLampertico 2007 [41]RCT open-label extensionItalyJan 2001–NRHospital (not specified)LAM + ADVN = 39Cirrhosis was defined on clinical grounds, that is, albumin level < 3.5 g/dL, platelet count < 100,000 mm3, clinical decompensation, and/or ultrasound demonstrationof surface nodularity, splenomegaly, and > 15-mm portal vein diameterLee 2016 [42]Retrospective cohortSouth KoreaJan 2009–Dec 2012bBundang Jesaeng HospitalETVN = 66Cirrhosis was diagnosed based on ultrasonographic findingsLee 2019 [43]RCTSouth KoreaDec 2008–Apr 201518 medical centers (not specified)ETV; LAM; ETV + ADV;e ETV+tenofovir;e LAM + ADV;e LAM+tenofovireN = 72Advanced fibrosis and cirrhosis were confirmed according to grades F3 and F4, or clinical evidence of overt cirrhosis (gastroesophageal varices identified on endoscopy, orplatelet count < 100,000/mm3 and splenomegaly [> 12 cm]) with a high viral load (HBV DNA 2,000 IU/mL)Lee 2021 [44]National registry/ databaseSouth KoreaJan 2012–Dec 2014dHospital (not specified)TDF; TDF-based combination therapyN = 2,030Cirrhosis was defined as patients with ICD-10 K74 codeLee 2022 [45]Retrospective cohortSouth KoreaJan 2007–Dec 2020bKorea University Guro Hospital; Korea University Ansan Hospital; Gacheon University Gil Medical Center; Wonju Severance Christian HospitalETV; TDFN = 787Cirrhosis was diagnosed based on ultrasonographic findings including undulation of liver surfaces, blunt liver edge, severe coarse liver echotexture, liver shrinkage, splenomegaly, or ascitesLi 2017 [46]Prospective cohortChinaMar 2012–May 2015First Affiliated Hospital of Zhengzhou UniversityETV; LAM; ADV; LAM + ADVN = 965NRLim 2014 [47]Retrospective cohortSouth KoreaNov 1999–Mar 2013cAsan Medical CenterLAM; ETVN = 878Cirrhosis was diagnosed when the liver echotexture was coarse and nodular and/or when features of portal hypertension (e.g., ascites, splenomegaly, and varices) were notedLin 2019 [48]Retrospective cohortTaiwanJan 2009–Dec 2010bMackay Memorial HospitalETV; LAM; LdTN = 270Patients underwent abdominal ultrasonography to determine the changes of cirrhosisMoon 2015 [49]Retrospective cohortSouth KoreaJan 2003–Dec 2012dChonbuk National University HospitalLAM; LdT; CLV; ADV;e ETV;e ETV + ADV;e TDFeN = 330Cirrhosis diagnosis was based on liver biopsies or radiologic findings with or without varices at upper gastrointestinal endoscopy. Radiologic findings of cirrhosis were defined as coarse liver echotexture with parenchymal nodularity, irregular surface nodularity, and the features of portal hypertensionNagata 2016 [50]Retrospective cohortJapanDec 1999–Jun 2010bTokai University HospitalLAM; ETV; no treatment (LAM or ETV cessation)N = 94NRNiro 2013 [51]Prospective cohortItaly1998–2007bCasa Sollievo Sofferenza HospitalLAM; ETV; ADV; LdT; tenofovir; LAM + ADVN = 74To define grading and staging of liver disease, a liver biopsy was advised before commencing antiviral treatment. Histological activity was evaluated by the modified Scheuer’ score, which grades the inflammatory activity on a scale from 0 to 3 and fibrosis from 0 to 4. In patients who did not consent to liver biopsy, cirrhosis was diagnosed with indirect criteria, such as platelet count ≥ 100,000 mm3, ultrasound features of surface nodularity, splenomegaly and portal vein diameter > 13 mm, and/or the endoscopic detection of esophageal varicesOh 2020 [52]Retrospective cohortSouth KoreaJan 2011–Dec 2015b9 academic hospitals (not specified)ETV; TDFN = 935The severity of fibrosis was calculated using the fibrosis-4 score (e.g., no fibrosis if the fibrosis-4 scores were < 1.45; possible fibrosis if the fibrosis-4 scores were between 1.45–3.25; and advanced fibrosis if the fibrosis-4 scores were > 3.25). Cirrhosis was defined based on the radiologic features of cirrhosis (e.g., undulating liver surface, compensatory left lobe hypertrophy, splenomegaly, and portosystemic shunts, and so forth)Orito 2015 [53]Retrospective cohortJapanNR15 hospitals in the Japanese Red Cross Liver NetworkETV; LAM; LAM+ADVeN = 492NRPAGE-B Cohort [54]Prospective cohortGreece; Spain; Germany; Italy; The Netherlands; TurkeyNR–Dec 2012b10 treatment centers (not specified)ETV; TDF; ETV + ADV; ETV + TDF; TDF + LAM; TDF + LdTN = 1,023Elastographic cirrhosis was considered to be present in cases with reliable liver stiffness measurements of ≥ 12 kPa, which represents the lowest LSM cut-off for the diagnosis of cirrhosis according to the international guidelinesPapatheodoridis 2015 [55]Retrospective cohortGreece; Spain; Italy; The Netherlands; TurkeyNR–Oct 2013d7 liver clinics (not specified)ETV; TDF; ETV + ADV; ETV + TDF; TDF + LAM; TDF + LdTN = 1,085Patients were classified according to the severity of their liver disease into (a) patients with CHB only if they had a pretreatment liver biopsy without lesions of cirrhosis, (b) patients with compensated cirrhosis if they had histological findings and/or ultrasonographic findings (nodules in the hepatic parenchyma, spleen > 12 cm, portal vein > 16 mm) and/or endoscopic findings of cirrhosis (varices, portal gastropathy) and (c) patients with decompensated cirrhosis if they had developed ascites, variceal bleeding, encephalopathy and/or non-obstructive jaundice (bilirubin >3 mg/dl). Patients without a pretreatment liver biopsy and without any other sign of cirrhosis were considered as cases with unclassified disease severityPapatheodoridis 2017 [56]Prospective cohortGreece; Spain; Germany; Italy; The Netherlands; TurkeyNR–Apr 2016d10 liver clinics (not specified)ETV; TDF; ETV + ADV; ETV + TDF; TDF + LAM; TDF + LdTN = 1,379The severity of liver disease was classified into CHB only (without cirrhosis) according to findings from liver biopsies and CHB with compensated cirrhosis according to histological, elastographic (reliable liver stiffness measurement > 14 kPa), ultrasonographic, and/or endoscopic findings. Patients without a pretreatment liver biopsy and without any other sign of cirrhosis were considered as cases with unclassified disease severityPapatheodoridi 2022 [57]Retrospective cohortGreece; TaiwanJan 2011–Jan 2017bNRPatients who discontinued NA therapy (not specified); ETV; LAM; TDFN = 490Absence of cirrhosis was defined as Ishak staging score of ≤ 4 or liver stiffness ≤ 10 kPa at transient elastography [58]Pellicelli 2014 [59]Retrospective cohortItalyJan 1998–Dec 2012cLiver units or gastroenterology treatment centers that were part of the CLEO groupLAM; ETV; ADV; LdT; TDF; LAM + ADVN = 193Cirrhosis diagnosis was based on histological criteria (Ishak stages 5 and 6). In patients that did not undergo liver biopsy, cirrhosis was diagnosed on the basis of other criteria, such as ultrasound signs (spleen size > 12 cm, coarse nodular eco-pattern in the hepatic parenchyma), and/or transient elastometry value (FibroScan®) > 12.5 kPa and/or endoscopic findings compatible with cirrhosis (esophageal varices, portal gastropathy), and/or platelet count < 100,000 mm3RETRACT-B [60]Prospective cohortBelgium; Germany; Greece; The Netherlands; Spain; Hong Kong; Taiwan; CanadaJan 2001–Jan 2020b13 treatment centers (not specified)Patients who discontinued NA therapy (not specified)N = 1,368Cirrhosis was diagnosed based on histologic findings or ultrasonographic evidence with or without splenomegalyShim 2017 [61]Retrospective cohortSouth KoreaJan 2007–June 2016cKyung Hee University HospitalETVN = 133Cirrhosis was diagnosed with imaging, endoscopic, or histologic examinations. Typical imaging findings were an irregular nodular surface, coarse parenchymal echogenicity reflecting underlying nodularity, and/or caudate lobe hypertrophy. The presence of esophageal or gastric varices and ascites was also included as a diagnostic criterion for cirrhosisShin 2018 [62]Retrospective cohortSouth KoreaJan 2007–May 2017dUlsan University Hospital in KoreaETVN = 454Cirrhosis was determined via liver biopsy, or clinically defined in the absence of liver biopsy based on repeated liver imaging studies (nodular liver surface or caudate lobe hypertrophy) with thrombocytopenia (< 150 K/µL) or splenomegaly (by imaging), and/or by the presence of varices (by upper endoscopy or imaging studies)Sou 2020 [63]Retrospective cohortTaiwanJan 2007–Oct 2012bKaohsiung Chang Gung Memorial Hospital or China Medical University HospitalETVN = 890Cirrhosis was diagnosed according to either (1) histology or (2) repeated ultrasounds with consistent findings suggestive of cirrhosis in addition to clinical features such as splenomegaly, thrombocytopenia, ascites, or gastroesophageal varicesTseng 2017 [64]Retrospective cohort and prospective cohortTaiwanJan 1985–Dec 2015cERADICATE-B study; National Taiwan University Hospital; Taipei Tzuchi HospitalNA treatment (not specified); no treatmentN = 2,607Cirrhosis at baseline was diagnosed either by histology, APRI value > 2.0, or ultrasonographic findings together with clinical features such as thrombocytopenia, gastroesophageal varices, or ascitesWang 2016 [65]Retrospective cohortTaiwanOct 2011–May 2013dChang Gung Memorial HospitalTDFN = 144Cirrhosis was diagnosed on the basis of liver stiffness measurement and/or if repeated ultrasound and clinical findings were suggestive of cirrhotic conditions, such as splenomegaly, thrombocytopenia, varices, ascites, and hepatic encephalopathyWong 2013 [66]Prospective cohortHong KongDec 2005–Aug 2012bHepatitis clinics at Prince of Wales HospitalETVN = 1,199Cirrhosis was defined as a shrunken small liver with a nodular surface noted on imaging oft he liver and clinical features of portal hypertension (e.g., ascites, splenomegaly, and varices)Wu 2017 [67]Retrospective cohortTaiwanJan 2007–Jan 2014dChang Gung Memorial HospitalTDF; ETVN = 296Cirrhosis was diagnosed according to repeated ultrasounds suggestive of cirrhosis in addition to clinical features such as splenomegaly, thrombocytopenia, ascites or varicesYamada 2015 [68]Retrospective cohortJapanJul 2004–Jul 2012bOsaka University Hospital; Osaka Liver ForumETVN = 404Cirrhosis was defined by a shrunken, small liver with nodular surface as noted on liver imaging and by clinical features of portal hypertensionYang 2013 [69]Retrospective cohortTaiwanAug 2006–May 2010dChang Gung Memorial HospitalETVN = 314Cirrhosis was diagnosed by: (i) liver biopsy histology or (ii) combined repeated ultrasound findings suggestive of cirrhosis from at least two examinations performed a minimum of 6 months apart with clinical features such as splenomegaly, thrombocytopenia, esophageal varices, ascites or hepatic encephalopathyYang 2014 [70]Retrospective cohortTaiwanJan 1990–Jan 2009bNational Taiwan University HospitalLAMN = 54NRYip 2020 [71]National registry/databaseHong KongJan 2008–Dec 2018cCDARSETV; TDFN = 25,940Cirrhosis was identified by ICD-9-CM diagnosis codes for cirrhosis and its related complicationsYuen 2007 [72]Prospective cohortHong KongJun 1994–Aug 1997bQueen Mary HospitalLAM; famciclovir followed by LAM; no treatmentN = 266Cirrhosis development was defined as ultrasonographic evidence of small-sized liver with or without splenomegaly and ascitesAbbreviations: ADV adefovir, ALT alanine transaminase, APRI AST to platelet ratio, ART antiretroviral therapy, AST aspartate aminotransferase, CLEO Club Epatologi Ospedalieri, CLV clevudine, CT computed tomography, ETV entecavir, LAM lamivudine, LdT telbivudine, MRI magnetic resonance imaging, NA nucleo(s/t)ide analog, NR not reported, RCT randomized control trial, TAF tenofovir alafenamide, TDF tenofovir disoproxil fumarateaThe sample size represents the number of patients included in each study who were relevant to this SLR, which may be a sub-population of the overall number of patients included in a given studybThese dates represent that study’s start and end of patient enrollment, as study start and end date were not reportedcThese dates represented the start of patient enrollment, and the end of data collection, as the study start date and end date were not reporteddThese dates represent the data collection start and end date, as study start date and end date were not reported eTreatments were received during follow-up but not at baseline Table 1Summary of study characteristicsStudyStudy designCountryStudy periodSource of dataTreatments evaluatedSample sizeaDefinition of cirrhosis statusAndo 2018 [22]Retrospective cohortJapanJan 2000–Jan 2013bNagoya University HospitalLAM; LAM + ADV; ETV; ETV + ADV; TDFN = 105Cirrhosis was diagnosed by histological analysis and/or appearance of cirrhotic liver on ultrasonography, for example, superficial nodularity, coarse parenchymal echo pattern, signs of portal hypertension (splenomegaly > 120 mm, diameter of portal vein > 12 mm, patent collateral veins, or ascites), and/or the presence of gastroesophageal varices on endoscopyChiu 2020 [23]Retrospective cohortTaiwanJan 2007–Jan 2016bChang Gung Memorial HospitalNo treatment (ETV or TDF cessation)N = 460Noncirrhosis was diagnosed based on biopsy, Fibrosis-4 or liver stiffness measurement at baseline, combined repeated ultrasound findings without clinical features, such as splenomegaly, ascites, or gastroesophageal varices at baselineCho 2014 [24]Retrospective cohortSouth KoreaJan 2007–Jan 2012bSamsung Medical CenterETV; LAM; CLV; sequential or combination NA therapy; no treatmentN = 1,817Cirrhosis was determined by liver biopsy or imaging modalities with findings of irregular surface nodularity, parenchymal or morphological abnormality and manifestations of portal hypertension designated as liver cirrhosisChoi 2020 [25]Retrospective cohortSouth KoreaJan 2007–Dec 2016cAsan Medical CenterETV; TDFN = 2,612Cirrhosis was defined by the presence of any of the following findings: coarse liver echotexture and nodular liver surface on ultrasonography, clinical features of portal hypertension (e.g., ascites, splenomegaly, or varices), or thrombocytopenia (< 150,000/mm3)Choi 2022a [26]Retrospective cohortSouth KoreaJan 2000–Apr 2020cAsan Medical Center; Seoul National University Hospital; Severance Hospital; Samsung Medical CenterETV; TDF; no treatmentN = 2,073Cirrhosis was defined as the presence of any of the following characteristics: coarse liver echotexture and nodular liver surface on ultrasonography, clinical features of portal hypertension (e.g., ascites, splenomegaly, and varices), and platelet count below 100,000/µLChoi 2022b [27]Retrospective cohortSouth Korea; Hong Kong; TaiwanNRNRETV; TDFN = 7,545NRChun 2021 [28]National registry/ databaseSouth KoreaJan 2012–Dec 2018cNRETV; TDFN = 7,602Cirrhosis was defined as ICD-10 code K74Dezanet 2021 [29]Prospective cohortFranceJan 2002–Jan 2018cFour centers in FranceTDF-based ART; TAF-based ARTeN = 155Liver fibrosis was assessed using the FibroTest® calculated from standard battery of biochemical markers. Scores ranged from 0 to 1. METAVIR equivalents of this measure, as established in the HIV-HBV coinfected population, were used to grade liver fibrosis (F2: 0.48–0.58, F3: 0.59–0.73, F4: ≥0.74)Hosaka 2010 [30]Retrospective cohortJapanSep 1995–Sep 2007bToranomon HospitalLAM + ADVN = 186Cirrhosis was diagnosed by laparoscopy or liver biopsy, and in the patients who had not received them, by clinical data, imaging modalities, and portal hypertensionHosaka 2013 [31]Retrospective cohortJapanJan 1973–Dec 2011cToranomon HospitalETV; LAM; no treatmentN = 601Cirrhosis was determined based on laparoscopy, liver biopsy, imaging modalities, or portal hypertensionHosaka 2022 [32]Retrospective cohortJapanNR–Dec 2020dToranomon HospitalETVN = 132Cirrhosis was determined by laparoscopy, liver biopsy, imaging modalities, or portal hypertensionJi 2022 [33]RCT open-label extensionChinaNov 2013–NR9 first-class hospitals (not specified)ETV; ETV+Biejia-RuanganN = 500NRKim 2015 [34]RCT open-label extensionMultinationalJun 2005–Jan 2016NRTDF; ADV followed by TDFN = 482Cirrhosis was defined as Ishak fibrosis scores of 5 or 6Kim 2017a [35]Retrospective cohortSouth KoreaJan 2007–Dec 2015cAsan Medical CenterETVN = 850Cirrhosis was diagnosed based on histological and/or radiographical findings with evidence of portal hypertension (e.g., splenomegaly, ascites, or varices) in the presence of significant thrombocytopenia (platelet count < 150,000/mm3)Kim 2017b [36]Retrospective cohortSouth KoreaJan 2007–Dec 2015cSamsung Medical CenterETVN = 432Cirrhosis was clinically defined when patients showed cirrhotic configuration of the liver on imaging studies (nodular liver surface or caudate lobe hypertrophy) with thrombocytopenia (< 150 × 103/L) or splenomegaly (by imaging) and/or by the presence of varices (by upper endoscopy or imaging studies)Kim 2018 [37]Retrospective cohortSouth KoreaJan 2007–Apr 2017dUlsan University HospitalETV; TDFN = 198Cirrhosis was determined via liver biopsy or clinically defined in the absence of liver biopsy based on repeated liver imagingstudies (nodular liver surface or caudate lobe hypertrophy) with thrombocytopenia (< 150 × 1000/mm3) or splenomegaly (by imaging), and/or by the presence of varices (by upper endoscopy or imaging studies)Ko 2019 [38]Prospective cohortHong KongJan 2002–Dec 2015dQueen Mary HospitalETVN = 1,033Cirrhosis was defined by the following combined parameters: (a) ultrasonographic evidence of small-sized liver with or without development of splenomegaly/ascites/varices; and (b) APRI = (AST/upper limit of normal)/platelet count (x109L) x 100) greater than 2Kobashi 2011 [39]Prospective cohortJapanNov 2000–Nov 2010dNRETV; LAMN = 194Cirrhosis diagnosis was confirmed by liver histology, clinical signs (encephalopathy or ascites), evidence of esophagogastric varices by endoscopy, or imaging modalities such as ultrasonography, CT, or MRIKozuka 2019 [40]Retrospective cohortJapanSep 2006–June 2017cOsaka City University HospitalETVN = 105Liver cirrhosis was diagnosed by histological examination according to the METAVIR scoring system using imaging modalities such as ultrasonography, CT, or MRI and by signs of portal hypertensionLampertico 2007 [41]RCT open-label extensionItalyJan 2001–NRHospital (not specified)LAM + ADVN = 39Cirrhosis was defined on clinical grounds, that is, albumin level < 3.5 g/dL, platelet count < 100,000 mm3, clinical decompensation, and/or ultrasound demonstrationof surface nodularity, splenomegaly, and > 15-mm portal vein diameterLee 2016 [42]Retrospective cohortSouth KoreaJan 2009–Dec 2012bBundang Jesaeng HospitalETVN = 66Cirrhosis was diagnosed based on ultrasonographic findingsLee 2019 [43]RCTSouth KoreaDec 2008–Apr 201518 medical centers (not specified)ETV; LAM; ETV + ADV;e ETV+tenofovir;e LAM + ADV;e LAM+tenofovireN = 72Advanced fibrosis and cirrhosis were confirmed according to grades F3 and F4, or clinical evidence of overt cirrhosis (gastroesophageal varices identified on endoscopy, orplatelet count < 100,000/mm3 and splenomegaly [> 12 cm]) with a high viral load (HBV DNA 2,000 IU/mL)Lee 2021 [44]National registry/ databaseSouth KoreaJan 2012–Dec 2014dHospital (not specified)TDF; TDF-based combination therapyN = 2,030Cirrhosis was defined as patients with ICD-10 K74 codeLee 2022 [45]Retrospective cohortSouth KoreaJan 2007–Dec 2020bKorea University Guro Hospital; Korea University Ansan Hospital; Gacheon University Gil Medical Center; Wonju Severance Christian HospitalETV; TDFN = 787Cirrhosis was diagnosed based on ultrasonographic findings including undulation of liver surfaces, blunt liver edge, severe coarse liver echotexture, liver shrinkage, splenomegaly, or ascitesLi 2017 [46]Prospective cohortChinaMar 2012–May 2015First Affiliated Hospital of Zhengzhou UniversityETV; LAM; ADV; LAM + ADVN = 965NRLim 2014 [47]Retrospective cohortSouth KoreaNov 1999–Mar 2013cAsan Medical CenterLAM; ETVN = 878Cirrhosis was diagnosed when the liver echotexture was coarse and nodular and/or when features of portal hypertension (e.g., ascites, splenomegaly, and varices) were notedLin 2019 [48]Retrospective cohortTaiwanJan 2009–Dec 2010bMackay Memorial HospitalETV; LAM; LdTN = 270Patients underwent abdominal ultrasonography to determine the changes of cirrhosisMoon 2015 [49]Retrospective cohortSouth KoreaJan 2003–Dec 2012dChonbuk National University HospitalLAM; LdT; CLV; ADV;e ETV;e ETV + ADV;e TDFeN = 330Cirrhosis diagnosis was based on liver biopsies or radiologic findings with or without varices at upper gastrointestinal endoscopy. Radiologic findings of cirrhosis were defined as coarse liver echotexture with parenchymal nodularity, irregular surface nodularity, and the features of portal hypertensionNagata 2016 [50]Retrospective cohortJapanDec 1999–Jun 2010bTokai University HospitalLAM; ETV; no treatment (LAM or ETV cessation)N = 94NRNiro 2013 [51]Prospective cohortItaly1998–2007bCasa Sollievo Sofferenza HospitalLAM; ETV; ADV; LdT; tenofovir; LAM + ADVN = 74To define grading and staging of liver disease, a liver biopsy was advised before commencing antiviral treatment. Histological activity was evaluated by the modified Scheuer’ score, which grades the inflammatory activity on a scale from 0 to 3 and fibrosis from 0 to 4. In patients who did not consent to liver biopsy, cirrhosis was diagnosed with indirect criteria, such as platelet count ≥ 100,000 mm3, ultrasound features of surface nodularity, splenomegaly and portal vein diameter > 13 mm, and/or the endoscopic detection of esophageal varicesOh 2020 [52]Retrospective cohortSouth KoreaJan 2011–Dec 2015b9 academic hospitals (not specified)ETV; TDFN = 935The severity of fibrosis was calculated using the fibrosis-4 score (e.g., no fibrosis if the fibrosis-4 scores were < 1.45; possible fibrosis if the fibrosis-4 scores were between 1.45–3.25; and advanced fibrosis if the fibrosis-4 scores were > 3.25). Cirrhosis was defined based on the radiologic features of cirrhosis (e.g., undulating liver surface, compensatory left lobe hypertrophy, splenomegaly, and portosystemic shunts, and so forth)Orito 2015 [53]Retrospective cohortJapanNR15 hospitals in the Japanese Red Cross Liver NetworkETV; LAM; LAM+ADVeN = 492NRPAGE-B Cohort [54]Prospective cohortGreece; Spain; Germany; Italy; The Netherlands; TurkeyNR–Dec 2012b10 treatment centers (not specified)ETV; TDF; ETV + ADV; ETV + TDF; TDF + LAM; TDF + LdTN = 1,023Elastographic cirrhosis was considered to be present in cases with reliable liver stiffness measurements of ≥ 12 kPa, which represents the lowest LSM cut-off for the diagnosis of cirrhosis according to the international guidelinesPapatheodoridis 2015 [55]Retrospective cohortGreece; Spain; Italy; The Netherlands; TurkeyNR–Oct 2013d7 liver clinics (not specified)ETV; TDF; ETV + ADV; ETV + TDF; TDF + LAM; TDF + LdTN = 1,085Patients were classified according to the severity of their liver disease into (a) patients with CHB only if they had a pretreatment liver biopsy without lesions of cirrhosis, (b) patients with compensated cirrhosis if they had histological findings and/or ultrasonographic findings (nodules in the hepatic parenchyma, spleen > 12 cm, portal vein > 16 mm) and/or endoscopic findings of cirrhosis (varices, portal gastropathy) and (c) patients with decompensated cirrhosis if they had developed ascites, variceal bleeding, encephalopathy and/or non-obstructive jaundice (bilirubin >3 mg/dl). Patients without a pretreatment liver biopsy and without any other sign of cirrhosis were considered as cases with unclassified disease severityPapatheodoridis 2017 [56]Prospective cohortGreece; Spain; Germany; Italy; The Netherlands; TurkeyNR–Apr 2016d10 liver clinics (not specified)ETV; TDF; ETV + ADV; ETV + TDF; TDF + LAM; TDF + LdTN = 1,379The severity of liver disease was classified into CHB only (without cirrhosis) according to findings from liver biopsies and CHB with compensated cirrhosis according to histological, elastographic (reliable liver stiffness measurement > 14 kPa), ultrasonographic, and/or endoscopic findings. Patients without a pretreatment liver biopsy and without any other sign of cirrhosis were considered as cases with unclassified disease severityPapatheodoridi 2022 [57]Retrospective cohortGreece; TaiwanJan 2011–Jan 2017bNRPatients who discontinued NA therapy (not specified); ETV; LAM; TDFN = 490Absence of cirrhosis was defined as Ishak staging score of ≤ 4 or liver stiffness ≤ 10 kPa at transient elastography [58]Pellicelli 2014 [59]Retrospective cohortItalyJan 1998–Dec 2012cLiver units or gastroenterology treatment centers that were part of the CLEO groupLAM; ETV; ADV; LdT; TDF; LAM + ADVN = 193Cirrhosis diagnosis was based on histological criteria (Ishak stages 5 and 6). In patients that did not undergo liver biopsy, cirrhosis was diagnosed on the basis of other criteria, such as ultrasound signs (spleen size > 12 cm, coarse nodular eco-pattern in the hepatic parenchyma), and/or transient elastometry value (FibroScan®) > 12.5 kPa and/or endoscopic findings compatible with cirrhosis (esophageal varices, portal gastropathy), and/or platelet count < 100,000 mm3RETRACT-B [60]Prospective cohortBelgium; Germany; Greece; The Netherlands; Spain; Hong Kong; Taiwan; CanadaJan 2001–Jan 2020b13 treatment centers (not specified)Patients who discontinued NA therapy (not specified)N = 1,368Cirrhosis was diagnosed based on histologic findings or ultrasonographic evidence with or without splenomegalyShim 2017 [61]Retrospective cohortSouth KoreaJan 2007–June 2016cKyung Hee University HospitalETVN = 133Cirrhosis was diagnosed with imaging, endoscopic, or histologic examinations. Typical imaging findings were an irregular nodular surface, coarse parenchymal echogenicity reflecting underlying nodularity, and/or caudate lobe hypertrophy. The presence of esophageal or gastric varices and ascites was also included as a diagnostic criterion for cirrhosisShin 2018 [62]Retrospective cohortSouth KoreaJan 2007–May 2017dUlsan University Hospital in KoreaETVN = 454Cirrhosis was determined via liver biopsy, or clinically defined in the absence of liver biopsy based on repeated liver imaging studies (nodular liver surface or caudate lobe hypertrophy) with thrombocytopenia (< 150 K/µL) or splenomegaly (by imaging), and/or by the presence of varices (by upper endoscopy or imaging studies)Sou 2020 [63]Retrospective cohortTaiwanJan 2007–Oct 2012bKaohsiung Chang Gung Memorial Hospital or China Medical University HospitalETVN = 890Cirrhosis was diagnosed according to either (1) histology or (2) repeated ultrasounds with consistent findings suggestive of cirrhosis in addition to clinical features such as splenomegaly, thrombocytopenia, ascites, or gastroesophageal varicesTseng 2017 [64]Retrospective cohort and prospective cohortTaiwanJan 1985–Dec 2015cERADICATE-B study; National Taiwan University Hospital; Taipei Tzuchi HospitalNA treatment (not specified); no treatmentN = 2,607Cirrhosis at baseline was diagnosed either by histology, APRI value > 2.0, or ultrasonographic findings together with clinical features such as thrombocytopenia, gastroesophageal varices, or ascitesWang 2016 [65]Retrospective cohortTaiwanOct 2011–May 2013dChang Gung Memorial HospitalTDFN = 144Cirrhosis was diagnosed on the basis of liver stiffness measurement and/or if repeated ultrasound and clinical findings were suggestive of cirrhotic conditions, such as splenomegaly, thrombocytopenia, varices, ascites, and hepatic encephalopathyWong 2013 [66]Prospective cohortHong KongDec 2005–Aug 2012bHepatitis clinics at Prince of Wales HospitalETVN = 1,199Cirrhosis was defined as a shrunken small liver with a nodular surface noted on imaging oft he liver and clinical features of portal hypertension (e.g., ascites, splenomegaly, and varices)Wu 2017 [67]Retrospective cohortTaiwanJan 2007–Jan 2014dChang Gung Memorial HospitalTDF; ETVN = 296Cirrhosis was diagnosed according to repeated ultrasounds suggestive of cirrhosis in addition to clinical features such as splenomegaly, thrombocytopenia, ascites or varicesYamada 2015 [68]Retrospective cohortJapanJul 2004–Jul 2012bOsaka University Hospital; Osaka Liver ForumETVN = 404Cirrhosis was defined by a shrunken, small liver with nodular surface as noted on liver imaging and by clinical features of portal hypertensionYang 2013 [69]Retrospective cohortTaiwanAug 2006–May 2010dChang Gung Memorial HospitalETVN = 314Cirrhosis was diagnosed by: (i) liver biopsy histology or (ii) combined repeated ultrasound findings suggestive of cirrhosis from at least two examinations performed a minimum of 6 months apart with clinical features such as splenomegaly, thrombocytopenia, esophageal varices, ascites or hepatic encephalopathyYang 2014 [70]Retrospective cohortTaiwanJan 1990–Jan 2009bNational Taiwan University HospitalLAMN = 54NRYip 2020 [71]National registry/databaseHong KongJan 2008–Dec 2018cCDARSETV; TDFN = 25,940Cirrhosis was identified by ICD-9-CM diagnosis codes for cirrhosis and its related complicationsYuen 2007 [72]Prospective cohortHong KongJun 1994–Aug 1997bQueen Mary HospitalLAM; famciclovir followed by LAM; no treatmentN = 266Cirrhosis development was defined as ultrasonographic evidence of small-sized liver with or without splenomegaly and ascitesAbbreviations: ADV adefovir, ALT alanine transaminase, APRI AST to platelet ratio, ART antiretroviral therapy, AST aspartate aminotransferase, CLEO Club Epatologi Ospedalieri, CLV clevudine, CT computed tomography, ETV entecavir, LAM lamivudine, LdT telbivudine, MRI magnetic resonance imaging, NA nucleo(s/t)ide analog, NR not reported, RCT randomized control trial, TAF tenofovir alafenamide, TDF tenofovir disoproxil fumarateaThe sample size represents the number of patients included in each study who were relevant to this SLR, which may be a sub-population of the overall number of patients included in a given studybThese dates represent that study’s start and end of patient enrollment, as study start and end date were not reportedcThese dates represented the start of patient enrollment, and the end of data collection, as the study start date and end date were not reporteddThese dates represent the data collection start and end date, as study start date and end date were not reported eTreatments were received during follow-up but not at baseline Summary of study characteristics Cirrhosis was defined on clinical grounds, that is, albumin level < 3.5 g/dL, platelet count < 100,000 mm3, clinical decompensation, and/or ultrasound demonstration of surface nodularity, splenomegaly, and > 15-mm portal vein diameter Advanced fibrosis and cirrhosis were confirmed according to grades F3 and F4, or clinical evidence of overt cirrhosis (gastroesophageal varices identified on endoscopy, or platelet count < 100,000/mm3 and splenomegaly [> 12 cm]) with a high viral load (HBV DNA 2,000 IU/mL) Abbreviations: ADV adefovir, ALT alanine transaminase, APRI AST to platelet ratio, ART antiretroviral therapy, AST aspartate aminotransferase, CLEO Club Epatologi Ospedalieri, CLV clevudine, CT computed tomography, ETV entecavir, LAM lamivudine, LdT telbivudine, MRI magnetic resonance imaging, NA nucleo(s/t)ide analog, NR not reported, RCT randomized control trial, TAF tenofovir alafenamide, TDF tenofovir disoproxil fumarate aThe sample size represents the number of patients included in each study who were relevant to this SLR, which may be a sub-population of the overall number of patients included in a given study bThese dates represent that study’s start and end of patient enrollment, as study start and end date were not reported cThese dates represented the start of patient enrollment, and the end of data collection, as the study start date and end date were not reported dThese dates represent the data collection start and end date, as study start date and end date were not reported eTreatments were received during follow-up but not at baseline SLR baseline characteristics The baseline characteristics extracted across studies included age, sex, race or ethnicity, prior and current treatments, ALT levels, HBeAg status, and coinfections (Table 2). Apart from Ji 2022 [33], all studies reported age at baseline, with mean age across reporting study cohorts ranging from 38.4 to 59.0 years. Sex was reported by 49 studies, revealing a predominantly male demographic ranging from 45.1% to 95.5% of study cohorts. Race or ethnicity was reported by 21 studies; five studies focused exclusively on white populations and eleven on Asian populations. Table 2Summary of baseline characteristicsStudyCohortAgeSexRace or ethnicityPrior treatmentsCurrent treatmentsALTHBV DNAHBeAg statusCoinfectionsAndo 2018 [22]Full study cohort (N = 133)Median: 51 years (range 20–79)Male: 79/133Female: 54/133NRIFN: 0/133NA: 133/133 (100%)Median: 23 IU/L (range 8-107)NRPositive: 41/133Negative: 92/133No coinfectionsChiu 2020 [23]ETV group (n = 303)Mean: 51.3 years (SD 10.5)Male: 244/303Female: 59/303Asian: 100%ETV: 100%No current treatmentsMean: 322.5 U/L (SD 487.6)Mean: 5.52 log IU/ml (SD 1.52)Negative: 100%No coinfectionsTDF group (n = 157)Mean: 52.1 years (SD 10.4)Male: 126/157Female: 31/157Asian: 100%TDF: 100%No current treatmentsMean: 316.4 U/L(SD 443.9)Mean: 6.03 log IU/mL(SD 1.60)Negative: 100%No coinfectionsCho 2014 [24]NA-treated (n = 1,378)Mean: 47.5 years (SD 10.5)Male: 908 (65.9%)NRNo prior treatmentsCLV: 49 (3.5%); ETV: 1,002 (73.1%); LAM: 69 (5.0%); combination or sequential: 258 (18.7)%Mean: 155 U/L (SD 276.8)Mean: 6.38 log IU/mL (SD 1.50)Positive: 54.9%No coinfectionsUntreated inactive CHB (n = 1,014)Mean: 51.7 years (SD 10.2)Male: 608 (60.0%)NRNo prior treatmentsNo current treatmentsMean: 21.4 U/L (SD 7.6)Mean: 1.86 log IU/mL (SD 0.72)Positive: 0%Negative: 100%No coinfectionsChoi 2020 [25]Full study cohort (N = 4,639)Mean: 47.0 years (SD 10.9)Male: 2,990 (64.5%)Korean: 100%No prior treatmentsTDF: 1,574 (33.9%); ETV: 3,065 (66.1%)Median: 93 U/mL (IQR 47–184)Median: 6.7 log IU/mL (IQR 5.5–7.9)Positive: 2,732 (64.2%)No coinfectionsChoi 2022a [26]All treated patients (N = 2,073)Mean: 42.1 years (SD 11.8)Male: 1,306 (63.0%)Female: 767 (37.0%)NRNo prior treatmentsTDF: 792 (38.2%); ETV: 1,281 (61.8%)Median: 131 IU/L (IQR 84–262)Median: 8.0 log IU/mL (IQR 7.2–8.3)Positive: 2,073 (100%)No coinfectionsChoi 2022b [27]Full study cohort (N = 7,545)Median: 44 yearsMale: 61.60%NRNo prior treatmentsETV or TDF: 100%NRNRPositive: 100%NRChun 2021 [28]Individuals without cirrhosis at baseline (n = 7,602)Mean: 48 years (SD 10)Male: 4,665 (61.4%)NRNo prior treatmentsETV or TDF: 100%Median: 69 IU/L (IQR 40–125)NRNRNo coinfectionsDezanet 2021 [29]Full study cohort (N = 169)Median: 41.5 years (IQR 36.0-48.3)Male: 83.4%NRLAM: 149 (88.2%)TDF-based ART/TAF-based ART: 100%NRNRPositive: 99 (58.6%)HIV: 100%Hosaka 2010 [30]Individuals who developed HCCa (n = 18)Median: 52 years (range 35–75)Male: 15 (83%)NRLAM: 100%ADV + LAM: 100%Median: 151 IU/L (range 61–576)Median: 7.1 log copies/mL (range 4.4->7.6)Positive: 8 (44%)NRIndividuals who did not develop HCCa (n = 229)Median: 45 years (range 26–75)Male: 183 (80%)NRLAM: 100%ADV + LAM: 100%Median: 104 IU/L (range 13 − 1,563)Median: 7.1 log copies/mL (range < 2.6->7.6)Positive: 132 (58%)NRHosaka 2013 [31]ETV-treated individualsa (n = 316)Mean: 46 years (SD 12.1)Male: 210/316Female: 106/316Japanese: 100%No prior treatmentsETV: 100%Median: 61 IU/L (IQR 39–109)Median: 6.3 log copies/mL (IQR 5.2–7.9)Positive: 135 (43%)No coinfectionsLAM-treated individualsa (n = 182)Mean: 45 years (SD 10.7)Male: 136/182Female: 46/182Japanese: 100%No prior treatmentsLAM: 100%Median: 73 IU/L (IQR 42–121)Median: 6.4 log copies/mL (IQR 5.0-7.4)Positive: 71 (39%)No coinfectionsUntreated individualsa (n = 316)Mean: 46 years (SD 13.5)Male: 210/316Female: 106/316Japanese: 100%No prior treatmentsNo current treatmentsMedian: 60 IU/L (IQR 28–144)Median: 6.6 log copies/mL (IQR 4.5–7.8)Positive: 133 (42%)No coinfectionsHosaka 2022 [32]Full study cohort (N = 180)Mean: 51 years (SD 9.9)Male: 111 (61.7%)NRNRETV: 100%Median: 51 IU/L (IQR 31–98)Median: 4.8 log IU/mL (IQR 3.5–5.6)Negative: 100%No coinfectionsJi 2022 [33]ETV-treated (n = 500)NRNRNRNRETV: 100%NRNRPositive: 100%No coinfectionsKim 2015 [34]Individuals without cirrhosis at baseline (n = 482)Mean: 38.4 years (SD 11.8)Male: 345 (72%)White: 283 (59%)Asian: 148 (31%)Other: 51 (11%)NRTDF: 100%Mean: 143 U/L (SD 113.1)Mean: 7.7 log copies/mL (SD 1.5)Positive: 199 (41%)No coinfectionsKim 2017a [35]Individuals without cirrhosis at baseline (n = 1,185)Mean: 44 years (SD 11)Male: 779 (65.7)NRNo prior treatmentsETV: 100%Median: 131 IU/mL (IQR 84–265)NRNRNo coinfectionsKim 2017b [36]Full study cohorta (N = 875)Mean: 47.7 years (SD 10.7)Male: 564 (64.5%)Korean: 100%No prior treatmentsETV: 100%Median: 89 U/L (IQR 51–145)Mean: 6.5 log IU/mL (SD 1.3)Positive: 483 (55.2%)No coinfectionsKim 2018 [37]Non-cirrhotic, propensity-matched individuals receiving TDF treatment (n = 198)Mean: 48 years (SD 11)Male: 132 (66.7%)NRNo prior treatmentsTDF: 100%Mean: 188.2 IU/mL (SD 280.9)Mean: 6.4 log IU/mL (SD 1.6)Positive: 152 (77.3%)No coinfectionsKo 2019 [38]Individuals who developeda HCC (n = 66)Median: 57.0 years (IQR Q3 − Q1 10.4)Male: 54 (81.8%)Asian: 100%NA: 2 (3%)ETV: 100%Median: 58.5 U/L (IQR Q3 − Q1 48)Median: 6.1 log IU/mL (IQR Q3 − Q1 2.8)Positive: 14 (21.2%)NRIndividuals who did not develop HCCa (n = 1,159)Median: 49.4 years (IQR Q3 − Q1 16.1)Male: 741 (63.9%)Asian: 100%NA: 76 (6.6%)ETV: 100%Median: 69 U/L (IQR Q3 − Q1 119)Median: 6.5 log IU/mL (IQR Q3 − Q1 2.4)Positive: 293 (25.3%)NRKobashi 2011 [39]Full study cohort (N = 256)Median: 50 years (range 22–88)Male: 179 (69.9%)Female: 77 (30.1%)NRNo prior treatmentsETV: 129 (50.4%)LAM: 127 (49.6%)Median: 87 IU/L (IQR 13 − 1,660)Median: 7 log copies/mL (range 2.6–8.8)Positive: 132 (51.6%)Negative: 124 (48.4%)No coinfectionsKozuka 2019 [40]Non-cirrhotic individuals receiving ETV treatment (n = 105)Median: 43 years (range 23–78)Male: 67 (63.8%)Asian: 100%IFN: 22 (21%)ETV: 100%Median: 119 U/L (range 22 − 2,605)Median: 7.4 log copies/mL (range 3.1->9.1)Positive: 52 (49.5%)No coinfectionsLampertico 2007 [41]Full study cohorta (N = 145)Median: 56 years (range 19–77)Male: 122 (84%)White: 100%LAM: 100%ADV + LAM: 100%Median: 58 IU/L (range 13 − 2,870)Median: 6 log copies/mL (range 3.5–9.3)Negative: 124 (86%)NRLee 2016 [42]Full study cohorta (N = 102)Mean: 46.4 years (SD 11.2)Male: 67 (66%)Female: 35 (34%)NRNo prior treatmentsETV: 100%Mean: 224.9 U/L (SD 440.8)Median: 7,340,040.5 IU/mL (IQR 474,098.3–138,000,000)Positive: 71 (70%)NRLee 2019 [43]Full study cohorta(N = 462)Mean: 51.1 years (SD 9.4)Male: 250 (54.1%)NRNo prior treatmentsETV: 231 (50.0%)LAM: 231 (50.0%)Mean: 40.0 IU/L (SD 30.3)Mean: 5.5 log IU/mL (SD 1.3)Positive: 222 (48.1%)No coinfectionsLee 2021 [44]Propensity-score matched with TDF combination-therapy (n = 2,499)Mean: 51.5 years (SD 8.7)Male: 1,765 (70.6%)Almost all patients had Korean ethnicityADV: 159 (6.4%); CLV: 202 (8.1%); ETV: 192 (7.7%); LAM: 1,900 (76.0%); telbivudine: 46 (1.8%)TDF-based combination therapy: 100%Median: 28 IU/mL (IQR 21–40)NRNRNo coinfectionsPropensity-score matched with TDF monotherapy (n = 2,499)Mean: 50.7 years (SD 8.8)Male: 1,712 (68.5%)Almost all patients had Korean ethnicityADVr: 127 (5.1%); CLV: 308 (12.3%); ETV: 562 (22.5%); LAM: 1,426 (57.1%); LdT: 76 (3.0%)TDF: 100%Median: 27 IU/mL (IQR 20–41)NRNRNo coinfectionsLee 2022 [45]Derivation cohorta (n = 1,239)Mean: 47.72 years (SD 10.95)Male: 736 (59.4%)NRNRTDF: 471 (38.3%); ETV: 760 (61.7%)Median: 87 U/L (IQR 44–180)Median: 1.8 × 106 IU/mL (IQR 7.5 × 104-3.26 × 107)Positive: 625 (51.6%)NRLi 2017 [46]Full study cohorta (N = 1,200)Mean/median NRMale: 816/1,200Female: 384/1,200NRNo prior treatmentsADV: 455 (37.92%); ETV: 425 (35.42%); LAM: 153 (12.75%); LAM + ADV: 167 (13.92%)Mean/median NRMean/median NRPositive: 565 (47.08%)No coinfectionsLim 2014 [47]Non-cirrhotic LAM-treated individuals, propensity score matched to ETV-treated cohort (n = 878)Mean: 42.2 years (SD 10.6)Male: 578 (65.8%)NRNo prior treatmentsLAM: 100%Median: 172 IU/mL (IQR 96–321)Mean: 7.68 log IU/mL (SD 1.08)Positive: 656 (74.7%)No coinfectionsLin 2019 [48]Individuals who developeda HCC (n = 18)Mean: 52.8 years (SD 6.1)Male: 15 (83.3%)NRNRETV: 5/18; LAM: 8/18; LdT: 5/18Mean: 161.4 IU/L (SD 177.3)Mean: 6.7 × 106 IU/mL (SD 130.0)Positive: 7/18Negative: 11/18No coinfectionsIndividuals who did not develop HCCa (n = 378)Mean: 47.1 (SD 12.6)Male: 285 (75.4%)NRNRETV: 173/378; LAM: 160/378; LdT: 45/378Mean: 361.7 IU/L (SD 496.3)Mean: 150.0 × 106 IU/mL (SD 160.0)Positive: 164/378Negative: 214/378No coinfectionsMoon 2015 [49]Full study cohorta (N = 524)Mean: 46.2 years (SD 12.2)Male: 340 (64.9%)NRNo prior treatmentsCLV: 70 (13.4%); ETV: 161 (30.7%); LAM: 293 (55.9%); multi-drug treatmentbMean: 157.6 IU/L (SD 195.0)Mean: 6.4 log IU/mL (SD 1.2)Positive: 360 (68.7%)HCV: 4 (0.9%)Nagata 2016 [50]Non-cirrhotic HBeAg- individuals (n = 39)Mean: 50.0 years (SD 11.8)Male: 28 (72%)Female: 11 (28%)NRNo prior treatmentsETV: 14 (36%); LAM: 25 (64%); multi-drug treatmentcMean/median NRMean/median NRPositive: 0%Negative: 100%No coinfectionsNon-cirrhotic HBeAg+ individuals (n = 55)Mean: 44.3 years (SD 14.5)Male: 37 (67%)Female: 18 (33%)NRNo prior treatmentsETV: 17 (31%); LAM: 38 (69%); multi-drug treatmentdMean/median NRMean/median NRPositive: 100%Negative: 0%No coinfectionsNiro 2013 [51]Individuals without cirrhosis at baseline (n = 74)Mean: 58 years (SD 11)Male: 55 (74%)All but 1 patient was ItalianNA: 38 (51%); IFN monotherapy: 14 (19%); IFN + NA therapy: 24 (32%)NA: 100%Mean: 243 [units NR] (SD 245)Mean/median NRNRNo coinfectionsOh 2020 [52]Propensity-matched individuals receiving ETVa (n = 516)Mean: 49.2 years (SD 12.6)Male: 319 (61.8%)NRNo prior treatmentsETV: 100%NRMedian: 6.4 log IU/mL (IQR 5.4–7.5)Positive: 314 (60.9%)No coinfectionsPropensity-matched individuals receiving TDFa (n = 516)Mean: 49.0 years (SD 9.4)Male: 325 (63.0%)NRNo prior treatmentsTDF: 100%NRMedian: 6.4 log IU/mL (IQR 5.4–7.5)Positive: 311 (60.3%)No coinfectionsOrito 2015 [53]Full study cohorta (N = 602)Median: 52 years (range 21–79)Male: 381/602Female: 221/602NRNRETV: 405/602; LAM: 56/602; multi-drug combination therapyeMedian: 69 IU/L (range 9 − 2,821)Median: 6.8 log copies/mL (range 2.3–9.1)Positive: 295/600Negative: 305/600No coinfectionsPAGE-B Cohort [54]Full study cohorta (N = 1,427)Mean: 52.1 years (SD 13.1)Male: 992 (69.5%)White: 100%NA: 642 (45.0%); IFN: 358 (25.1%)TDF: 609 (42.7%); ETV: 515 (36.1%); ETV + ADV: 10 (0.7%); ETV + TDF: 20 (1.4%); TDF + LAM: 268 (18.8%); TDF + LdT: 5 (0.3%)Mean/median NRMean/median NRPositive: 261 (18.4%)No coinfectionsPapatheodoridis 2015 [55]Full study cohorta (N = 1,666)Mean: 53.7 years (SD 13.5)Male: 1,194 (71.7%)White: 100%NA: 659 (39.6%); IFN: 418 (25.1%)TDF: 606 (36.4%); ETV: 715 (42.9%); ETV + ADV: 14 (0.8%); ETV + TDF: 24 (1.4%); TDF + LAM: 305/1,666; TDF + LdT: 2/1,666Mean/median NRMean/median NRPositive: 257 (15.4%)No coinfectionsPapatheodoridis 2017 [56]Non-cirrhotic individuals receiving ETV or TDF (n = 1,379)Mean: 51 years (SD 14)Male: 932 (68%)White: 100%NA: 524 (38%); IFN: 324 (24%)TDF: 626 (45%); ETV: 577 (42%); ETV + ADV: 11 (1%); ETV + TDF: 27 (2%); TDF + LAM: 131/1,379; TDF + LdT: 7/1,379Mean/median NRMean/median NRPositive: 266 (19%)No coinfectionsPapatheodoridi 2022 [57]Propensity score-matched individuals who discontinued NA therapy (n = 245)Mean: 57 years (SD 12)Male: 178 (73%)White: 143 (58%)Asian: 102 (42%)TDF: 108 (44.1%); ETV: 111 (45.3%); LAM: 21 (8.6%)No current treatmentsMedian: 111 IU/L (IQR Q3 − Q1 67)Median: 5.8 log IU/mL (IQR Q3 − Q1 2.2)Negative: 100%No coinfectionsPropensity score-matched individuals who continued NA therapy > 5 years (n = 245)Mean: 58 years (SD 12)Male: 178 (73%)White: 143 (58%)Asian: 102 (42%)NRNA: 100%; TDF: 92 (37.6%); ETV: 133 (54.3%); LAM: 15 (6.1%)Median: 144 IU/L (IQR Q3 − Q1 415)Median: 5.6 log IU/mL (IQR Q3 − Q1 2.7)Negative: 100%No coinfectionsPellicelli 2014 [59]Non-cirrhotic individuals receiving NA therapy ≥ 18 months (n = 193)Mean: 52.8 years (SD 13)Male: 135 (70%)White: 100%IFN: 83 (43%)TDF: 14/193; ADV: 6/193; ETV: 93/193; LAM: 57/193; LdT: 4/193; LAM + ADV: 19/193; NA rescue therapy: 35 (18%)Mean: 86 IU/L (SD 40)Mean: 5.2 log IU/mL (SD 1.7)Negative: 100%No coinfectionsRETRACT-B [60]Full study cohorta (N = 1,552)Mean: 52.9 years (SD 11.2)Male: 1,122 (72.3%)White: 175 (11.3%)Black: 13 (0.8%)Asian: 1,359 (87.6%)Other: 5 (0.3%)NA: 100%; TDF: 421 (27.1%); ETV: 981 (63.2%); IFN: 113 (8.6%)No current treatmentsMedian: 0.6 xULN (range 0.4–0.8)NRNegative: 100%No coinfectionsShim 2017 [61]Non-cirrhotic individuals treated with ETV (n = 133)Median: 49 years (IQR 44–55)Male: 75 (56.4%)Korean: 100%No prior treatmentsETV: 100%Median: 105 U/L (IQR 68–213)Median: 6.81 log IU/mL (IQR 5.55–7.97)Positive: 75 (56.4%)No coinfectionsShin 2018 [62]Full study cohorta (N = 894)Mean: 52.0 years (SD 11.3)Male: 597 (66.8%)NRNo prior treatmentsETV: 100%Median: 89 IU/mL (IQR 46–153)Mean: 6.39 log IU/mL (SD 1.41)Positive: 537 (60.1%)No coinfectionsSou 2020 [63]Full study cohorta (N = 1,397)Median: 50 years (IQR Q3 − Q1 17)Male: 1,018 (72.9%)Asian: 100%No prior treatmentsETV: 100%Median: 104 U/L (IQR Q3 − Q1 205)Median: 5.96 log IU/mL (IQR Q3 − Q1 2.35)Positive: 491 (35.1%)No coinfectionsTseng 2017 [64]Untreated ERADICATE-B study cohort (n = 2,075)Mean/median NRMale: 1,222 (58.89%)Female: 853 (41.11%)Asian: 100%No prior treatmentsNo current treatmentsMean/median NRMean/median NRPositive: 372 (17.93%)Negative: 1,703 (82.07%)No coinfectionsNA-treated validation cohort (n = 532)NRNRAsian: 100%NRNA: 100%NRMean/median NRNRNRWang 2016 [65]NA-experienced individuals (n = 105)Mean: 48.9 years (SD 13.4)Male: 83 (79%)NRNA: 100%TDF: 100%Mean: 208.2 IU/L (SD 336.6)Mean: 6.2 log copies/mL (SD 1.9)Positive: 51 (48.6%)No coinfectionsNA-naïve individuals (n = 131)Mean: 48.9 years (SD 12.9)Male: 100 (76.3%)NRNo prior treatmentsTDF: 100%Mean: 238.7 IU/L (SD 419.8)Mean: 6.9 log copies/mL (SD 1.5)Positive: 42 (32.1%)No coinfectionsWong 2013 [66]Non-cirrhotic individuals (n = 1,199)Mean: 50 years (SD 12)Male: 863 (72%)NRNA: 358 (30%); IFN: 28 (2%)ETV: 100%Mean: 156 IU/L (SD 339)Mean: 5.0 log IU/mL (SD 2.2)Positive: 358 (32%)Negative: 813 (68%)No coinfectionsWu 2017 [67]Individuals receiving TDF (n = 106)Mean: 47.1 years (SD 12.1)Male: 74 (70%)NRNo prior treatmentsTDF: 100%Mean: 303 U/L (SD 372)Mean: 7.35 log IU/mL (SD 0.70)Positive: 50 (47.1%)No coinfectionsIndividuals receiving ETV (n = 313)Mean: 47 years (SD 12.3)Male: 230 (73%)NRNo prior treatmentsETV: 100%Mean: 316 U/L (SD 440)Mean: 7.18 log IU/mL (SD 0.74)Positive: 172 (55.0%)No coinfectionsYamada 2015 [68]Non-cirrhotic individuals receiving ETV (n = 404)Mean: 51.3 years (SD 12.1)Male: 233 (58%)Female: 171/404NRIFN: 44 (11%)ETV: 100%Mean: 156.1 IU/L (SD 210.8)Median: 6.9 log copies/mLPositive: 181 (45%)Negative: 219/404No coinfectionsYang 2013 [69]NA-naïve individualsa (n = 323)Mean: 48.0 years (SD 12.2)Male: 256 (79.3%)NRIFN: 20 (6.2%)ETV: 100%Mean: 324.5 U/L (SD 476.4)Mean: 6.4 log copies/mL (SD 1.5)Positive: 118 (36.5%)No coinfectionsNA-experienced individualsa (n = 164)Mean: 48.2 years (SD 11.2)Male: 144 (87.8%)NRADV: 45 (27.4%); LAM: 155 (94.5%); IFN: 15 (9.1%)ETV: 100%Mean: 277.9 U/L (SD 460.3)Mean: 6.3 log copies/mL (SD 1.7)Positive: 69 (42.1%)No coinfectionsYang 2014 [70]Individuals who received LAM prophylaxis for transplant (n = 26)Mean: 41 years (SD 14.2)Male: 16/26Female: 10/26NRNRLAM+immunosuppressive therapy: 100%NRMean: 3.1 log IU/mL (SD 2.8)Negative: 100%No coinfectionsIndividuals who received LAM post-transplant (n = 28)Mean: 42.7 years (SD 10.5)Male: 18/28Female: 10/28NRNRLAM+immunosuppressive therapy: 100%NRMean: 2.6 log IU/mL (SD 2.1)Negative: 100%No coinfectionsYip 2020 [71]ETV-treated individualsa (n = 28,041)Mean: 53.4 years (SD 13.0)Male: 18.094 (64.5%)Asian: 100%No prior treatmentsETV: 100%Median: 62.0 U/L (IQR 33.0-137.0)Mean: 5.3 log IU/mL (SD 2.2)Positive: 8,317 (29.7%)No coinfectionsTDF-treated individualsa (n = 1,309)Mean: 43.2 years (SD 13.1)Male: 591 (45.1%)Asian: 100%No prior treatmentsTDF: 100%Median: 43.0 U/L (IQR 25.0-103.0)Mean: 4.9 log IU/mL (SD 2.7)Positive: 721 (55.1%)No coinfectionsYuen 2007 [72]Non-cirrhotic LAM-treated individuals (n = 142)Median: 33.9 years (range 20.2–54.4)Male: 106/142Female: 36/142Asian: 100%NA: 99/142LAM: 100%Median: 65 U/L (range 14–838)Median: 8.7 log copies/mL (range 4.2–11.7)Positive: 100%No coinfectionsNon-cirrhotic untreated individuals (n = 124)Median: 33.4 years (range 20.8–59)Male: 90/124Female: 34/124Asian: 100%NRNo current treatmentsMedian: 56.5 U/L (range 9-850)Median: 6.8 log copies/mL (range 1.5–9.6)Positive: 100%No coinfectionsAbbreviations: ADV adefovir, ALT alanine aminotransferase, ART antiretroviral therapy, CHB chronic hepatitis B, CLV clevudine, ETV entecavir, HBeAg hepatitis B e-antigen, HBV hepatitis B virus, HCC hepatocellular carcinoma, HCV hepatitis C virus, HIV human immunodeficiency virus, IFN interferon, IQR interquartile range, IU international unit, LAM lamivudine, LdT telbivudine, NA nucleos(t)ide analogue, NR not reported, SD standard deviation, TAF tenofovir alafenamide, TDF tenofovir disoproxil fumarate, ULN upper limit of normalaBaseline characteristics were not reported separately for individuals without cirrhosis at baseline (study reports a sub analysis of HCC risk in non-cirrhotic individuals and therefore is eligible for SLR inclusion)b251 out of 363 patients initially on initial LAM/LdT/CLV switched to ADV (n = 149), ETV (n = 90), and other NAs (n = 12). Of the 161 patients initially started on ETV, 2 patients switched to ETV + ADV and 2 patients switched to TDFc17 out of 25 patients receiving LAM discontinued treatment. Of these 17 patients, 4 switched to ETV. 6 out of 14 patients receiving ETV discontinued treatment, of these, 2 patients restarted ETV after relapsingd14 out of 38 patients receiving LAM discontinued treatment, of these, 2 patients started ETV after relapsingen=74 switched from LAM monotherapy to LAM + ADV; n = 67 switched from LAM to ETV Table 2Summary of baseline characteristicsStudyCohortAgeSexRace or ethnicityPrior treatmentsCurrent treatmentsALTHBV DNAHBeAg statusCoinfectionsAndo 2018 [22]Full study cohort (N = 133)Median: 51 years (range 20–79)Male: 79/133Female: 54/133NRIFN: 0/133NA: 133/133 (100%)Median: 23 IU/L (range 8-107)NRPositive: 41/133Negative: 92/133No coinfectionsChiu 2020 [23]ETV group (n = 303)Mean: 51.3 years (SD 10.5)Male: 244/303Female: 59/303Asian: 100%ETV: 100%No current treatmentsMean: 322.5 U/L (SD 487.6)Mean: 5.52 log IU/ml (SD 1.52)Negative: 100%No coinfectionsTDF group (n = 157)Mean: 52.1 years (SD 10.4)Male: 126/157Female: 31/157Asian: 100%TDF: 100%No current treatmentsMean: 316.4 U/L(SD 443.9)Mean: 6.03 log IU/mL(SD 1.60)Negative: 100%No coinfectionsCho 2014 [24]NA-treated (n = 1,378)Mean: 47.5 years (SD 10.5)Male: 908 (65.9%)NRNo prior treatmentsCLV: 49 (3.5%); ETV: 1,002 (73.1%); LAM: 69 (5.0%); combination or sequential: 258 (18.7)%Mean: 155 U/L (SD 276.8)Mean: 6.38 log IU/mL (SD 1.50)Positive: 54.9%No coinfectionsUntreated inactive CHB (n = 1,014)Mean: 51.7 years (SD 10.2)Male: 608 (60.0%)NRNo prior treatmentsNo current treatmentsMean: 21.4 U/L (SD 7.6)Mean: 1.86 log IU/mL (SD 0.72)Positive: 0%Negative: 100%No coinfectionsChoi 2020 [25]Full study cohort (N = 4,639)Mean: 47.0 years (SD 10.9)Male: 2,990 (64.5%)Korean: 100%No prior treatmentsTDF: 1,574 (33.9%); ETV: 3,065 (66.1%)Median: 93 U/mL (IQR 47–184)Median: 6.7 log IU/mL (IQR 5.5–7.9)Positive: 2,732 (64.2%)No coinfectionsChoi 2022a [26]All treated patients (N = 2,073)Mean: 42.1 years (SD 11.8)Male: 1,306 (63.0%)Female: 767 (37.0%)NRNo prior treatmentsTDF: 792 (38.2%); ETV: 1,281 (61.8%)Median: 131 IU/L (IQR 84–262)Median: 8.0 log IU/mL (IQR 7.2–8.3)Positive: 2,073 (100%)No coinfectionsChoi 2022b [27]Full study cohort (N = 7,545)Median: 44 yearsMale: 61.60%NRNo prior treatmentsETV or TDF: 100%NRNRPositive: 100%NRChun 2021 [28]Individuals without cirrhosis at baseline (n = 7,602)Mean: 48 years (SD 10)Male: 4,665 (61.4%)NRNo prior treatmentsETV or TDF: 100%Median: 69 IU/L (IQR 40–125)NRNRNo coinfectionsDezanet 2021 [29]Full study cohort (N = 169)Median: 41.5 years (IQR 36.0-48.3)Male: 83.4%NRLAM: 149 (88.2%)TDF-based ART/TAF-based ART: 100%NRNRPositive: 99 (58.6%)HIV: 100%Hosaka 2010 [30]Individuals who developed HCCa (n = 18)Median: 52 years (range 35–75)Male: 15 (83%)NRLAM: 100%ADV + LAM: 100%Median: 151 IU/L (range 61–576)Median: 7.1 log copies/mL (range 4.4->7.6)Positive: 8 (44%)NRIndividuals who did not develop HCCa (n = 229)Median: 45 years (range 26–75)Male: 183 (80%)NRLAM: 100%ADV + LAM: 100%Median: 104 IU/L (range 13 − 1,563)Median: 7.1 log copies/mL (range < 2.6->7.6)Positive: 132 (58%)NRHosaka 2013 [31]ETV-treated individualsa (n = 316)Mean: 46 years (SD 12.1)Male: 210/316Female: 106/316Japanese: 100%No prior treatmentsETV: 100%Median: 61 IU/L (IQR 39–109)Median: 6.3 log copies/mL (IQR 5.2–7.9)Positive: 135 (43%)No coinfectionsLAM-treated individualsa (n = 182)Mean: 45 years (SD 10.7)Male: 136/182Female: 46/182Japanese: 100%No prior treatmentsLAM: 100%Median: 73 IU/L (IQR 42–121)Median: 6.4 log copies/mL (IQR 5.0-7.4)Positive: 71 (39%)No coinfectionsUntreated individualsa (n = 316)Mean: 46 years (SD 13.5)Male: 210/316Female: 106/316Japanese: 100%No prior treatmentsNo current treatmentsMedian: 60 IU/L (IQR 28–144)Median: 6.6 log copies/mL (IQR 4.5–7.8)Positive: 133 (42%)No coinfectionsHosaka 2022 [32]Full study cohort (N = 180)Mean: 51 years (SD 9.9)Male: 111 (61.7%)NRNRETV: 100%Median: 51 IU/L (IQR 31–98)Median: 4.8 log IU/mL (IQR 3.5–5.6)Negative: 100%No coinfectionsJi 2022 [33]ETV-treated (n = 500)NRNRNRNRETV: 100%NRNRPositive: 100%No coinfectionsKim 2015 [34]Individuals without cirrhosis at baseline (n = 482)Mean: 38.4 years (SD 11.8)Male: 345 (72%)White: 283 (59%)Asian: 148 (31%)Other: 51 (11%)NRTDF: 100%Mean: 143 U/L (SD 113.1)Mean: 7.7 log copies/mL (SD 1.5)Positive: 199 (41%)No coinfectionsKim 2017a [35]Individuals without cirrhosis at baseline (n = 1,185)Mean: 44 years (SD 11)Male: 779 (65.7)NRNo prior treatmentsETV: 100%Median: 131 IU/mL (IQR 84–265)NRNRNo coinfectionsKim 2017b [36]Full study cohorta (N = 875)Mean: 47.7 years (SD 10.7)Male: 564 (64.5%)Korean: 100%No prior treatmentsETV: 100%Median: 89 U/L (IQR 51–145)Mean: 6.5 log IU/mL (SD 1.3)Positive: 483 (55.2%)No coinfectionsKim 2018 [37]Non-cirrhotic, propensity-matched individuals receiving TDF treatment (n = 198)Mean: 48 years (SD 11)Male: 132 (66.7%)NRNo prior treatmentsTDF: 100%Mean: 188.2 IU/mL (SD 280.9)Mean: 6.4 log IU/mL (SD 1.6)Positive: 152 (77.3%)No coinfectionsKo 2019 [38]Individuals who developeda HCC (n = 66)Median: 57.0 years (IQR Q3 − Q1 10.4)Male: 54 (81.8%)Asian: 100%NA: 2 (3%)ETV: 100%Median: 58.5 U/L (IQR Q3 − Q1 48)Median: 6.1 log IU/mL (IQR Q3 − Q1 2.8)Positive: 14 (21.2%)NRIndividuals who did not develop HCCa (n = 1,159)Median: 49.4 years (IQR Q3 − Q1 16.1)Male: 741 (63.9%)Asian: 100%NA: 76 (6.6%)ETV: 100%Median: 69 U/L (IQR Q3 − Q1 119)Median: 6.5 log IU/mL (IQR Q3 − Q1 2.4)Positive: 293 (25.3%)NRKobashi 2011 [39]Full study cohort (N = 256)Median: 50 years (range 22–88)Male: 179 (69.9%)Female: 77 (30.1%)NRNo prior treatmentsETV: 129 (50.4%)LAM: 127 (49.6%)Median: 87 IU/L (IQR 13 − 1,660)Median: 7 log copies/mL (range 2.6–8.8)Positive: 132 (51.6%)Negative: 124 (48.4%)No coinfectionsKozuka 2019 [40]Non-cirrhotic individuals receiving ETV treatment (n = 105)Median: 43 years (range 23–78)Male: 67 (63.8%)Asian: 100%IFN: 22 (21%)ETV: 100%Median: 119 U/L (range 22 − 2,605)Median: 7.4 log copies/mL (range 3.1->9.1)Positive: 52 (49.5%)No coinfectionsLampertico 2007 [41]Full study cohorta (N = 145)Median: 56 years (range 19–77)Male: 122 (84%)White: 100%LAM: 100%ADV + LAM: 100%Median: 58 IU/L (range 13 − 2,870)Median: 6 log copies/mL (range 3.5–9.3)Negative: 124 (86%)NRLee 2016 [42]Full study cohorta (N = 102)Mean: 46.4 years (SD 11.2)Male: 67 (66%)Female: 35 (34%)NRNo prior treatmentsETV: 100%Mean: 224.9 U/L (SD 440.8)Median: 7,340,040.5 IU/mL (IQR 474,098.3–138,000,000)Positive: 71 (70%)NRLee 2019 [43]Full study cohorta(N = 462)Mean: 51.1 years (SD 9.4)Male: 250 (54.1%)NRNo prior treatmentsETV: 231 (50.0%)LAM: 231 (50.0%)Mean: 40.0 IU/L (SD 30.3)Mean: 5.5 log IU/mL (SD 1.3)Positive: 222 (48.1%)No coinfectionsLee 2021 [44]Propensity-score matched with TDF combination-therapy (n = 2,499)Mean: 51.5 years (SD 8.7)Male: 1,765 (70.6%)Almost all patients had Korean ethnicityADV: 159 (6.4%); CLV: 202 (8.1%); ETV: 192 (7.7%); LAM: 1,900 (76.0%); telbivudine: 46 (1.8%)TDF-based combination therapy: 100%Median: 28 IU/mL (IQR 21–40)NRNRNo coinfectionsPropensity-score matched with TDF monotherapy (n = 2,499)Mean: 50.7 years (SD 8.8)Male: 1,712 (68.5%)Almost all patients had Korean ethnicityADVr: 127 (5.1%); CLV: 308 (12.3%); ETV: 562 (22.5%); LAM: 1,426 (57.1%); LdT: 76 (3.0%)TDF: 100%Median: 27 IU/mL (IQR 20–41)NRNRNo coinfectionsLee 2022 [45]Derivation cohorta (n = 1,239)Mean: 47.72 years (SD 10.95)Male: 736 (59.4%)NRNRTDF: 471 (38.3%); ETV: 760 (61.7%)Median: 87 U/L (IQR 44–180)Median: 1.8 × 106 IU/mL (IQR 7.5 × 104-3.26 × 107)Positive: 625 (51.6%)NRLi 2017 [46]Full study cohorta (N = 1,200)Mean/median NRMale: 816/1,200Female: 384/1,200NRNo prior treatmentsADV: 455 (37.92%); ETV: 425 (35.42%); LAM: 153 (12.75%); LAM + ADV: 167 (13.92%)Mean/median NRMean/median NRPositive: 565 (47.08%)No coinfectionsLim 2014 [47]Non-cirrhotic LAM-treated individuals, propensity score matched to ETV-treated cohort (n = 878)Mean: 42.2 years (SD 10.6)Male: 578 (65.8%)NRNo prior treatmentsLAM: 100%Median: 172 IU/mL (IQR 96–321)Mean: 7.68 log IU/mL (SD 1.08)Positive: 656 (74.7%)No coinfectionsLin 2019 [48]Individuals who developeda HCC (n = 18)Mean: 52.8 years (SD 6.1)Male: 15 (83.3%)NRNRETV: 5/18; LAM: 8/18; LdT: 5/18Mean: 161.4 IU/L (SD 177.3)Mean: 6.7 × 106 IU/mL (SD 130.0)Positive: 7/18Negative: 11/18No coinfectionsIndividuals who did not develop HCCa (n = 378)Mean: 47.1 (SD 12.6)Male: 285 (75.4%)NRNRETV: 173/378; LAM: 160/378; LdT: 45/378Mean: 361.7 IU/L (SD 496.3)Mean: 150.0 × 106 IU/mL (SD 160.0)Positive: 164/378Negative: 214/378No coinfectionsMoon 2015 [49]Full study cohorta (N = 524)Mean: 46.2 years (SD 12.2)Male: 340 (64.9%)NRNo prior treatmentsCLV: 70 (13.4%); ETV: 161 (30.7%); LAM: 293 (55.9%); multi-drug treatmentbMean: 157.6 IU/L (SD 195.0)Mean: 6.4 log IU/mL (SD 1.2)Positive: 360 (68.7%)HCV: 4 (0.9%)Nagata 2016 [50]Non-cirrhotic HBeAg- individuals (n = 39)Mean: 50.0 years (SD 11.8)Male: 28 (72%)Female: 11 (28%)NRNo prior treatmentsETV: 14 (36%); LAM: 25 (64%); multi-drug treatmentcMean/median NRMean/median NRPositive: 0%Negative: 100%No coinfectionsNon-cirrhotic HBeAg+ individuals (n = 55)Mean: 44.3 years (SD 14.5)Male: 37 (67%)Female: 18 (33%)NRNo prior treatmentsETV: 17 (31%); LAM: 38 (69%); multi-drug treatmentdMean/median NRMean/median NRPositive: 100%Negative: 0%No coinfectionsNiro 2013 [51]Individuals without cirrhosis at baseline (n = 74)Mean: 58 years (SD 11)Male: 55 (74%)All but 1 patient was ItalianNA: 38 (51%); IFN monotherapy: 14 (19%); IFN + NA therapy: 24 (32%)NA: 100%Mean: 243 [units NR] (SD 245)Mean/median NRNRNo coinfectionsOh 2020 [52]Propensity-matched individuals receiving ETVa (n = 516)Mean: 49.2 years (SD 12.6)Male: 319 (61.8%)NRNo prior treatmentsETV: 100%NRMedian: 6.4 log IU/mL (IQR 5.4–7.5)Positive: 314 (60.9%)No coinfectionsPropensity-matched individuals receiving TDFa (n = 516)Mean: 49.0 years (SD 9.4)Male: 325 (63.0%)NRNo prior treatmentsTDF: 100%NRMedian: 6.4 log IU/mL (IQR 5.4–7.5)Positive: 311 (60.3%)No coinfectionsOrito 2015 [53]Full study cohorta (N = 602)Median: 52 years (range 21–79)Male: 381/602Female: 221/602NRNRETV: 405/602; LAM: 56/602; multi-drug combination therapyeMedian: 69 IU/L (range 9 − 2,821)Median: 6.8 log copies/mL (range 2.3–9.1)Positive: 295/600Negative: 305/600No coinfectionsPAGE-B Cohort [54]Full study cohorta (N = 1,427)Mean: 52.1 years (SD 13.1)Male: 992 (69.5%)White: 100%NA: 642 (45.0%); IFN: 358 (25.1%)TDF: 609 (42.7%); ETV: 515 (36.1%); ETV + ADV: 10 (0.7%); ETV + TDF: 20 (1.4%); TDF + LAM: 268 (18.8%); TDF + LdT: 5 (0.3%)Mean/median NRMean/median NRPositive: 261 (18.4%)No coinfectionsPapatheodoridis 2015 [55]Full study cohorta (N = 1,666)Mean: 53.7 years (SD 13.5)Male: 1,194 (71.7%)White: 100%NA: 659 (39.6%); IFN: 418 (25.1%)TDF: 606 (36.4%); ETV: 715 (42.9%); ETV + ADV: 14 (0.8%); ETV + TDF: 24 (1.4%); TDF + LAM: 305/1,666; TDF + LdT: 2/1,666Mean/median NRMean/median NRPositive: 257 (15.4%)No coinfectionsPapatheodoridis 2017 [56]Non-cirrhotic individuals receiving ETV or TDF (n = 1,379)Mean: 51 years (SD 14)Male: 932 (68%)White: 100%NA: 524 (38%); IFN: 324 (24%)TDF: 626 (45%); ETV: 577 (42%); ETV + ADV: 11 (1%); ETV + TDF: 27 (2%); TDF + LAM: 131/1,379; TDF + LdT: 7/1,379Mean/median NRMean/median NRPositive: 266 (19%)No coinfectionsPapatheodoridi 2022 [57]Propensity score-matched individuals who discontinued NA therapy (n = 245)Mean: 57 years (SD 12)Male: 178 (73%)White: 143 (58%)Asian: 102 (42%)TDF: 108 (44.1%); ETV: 111 (45.3%); LAM: 21 (8.6%)No current treatmentsMedian: 111 IU/L (IQR Q3 − Q1 67)Median: 5.8 log IU/mL (IQR Q3 − Q1 2.2)Negative: 100%No coinfectionsPropensity score-matched individuals who continued NA therapy > 5 years (n = 245)Mean: 58 years (SD 12)Male: 178 (73%)White: 143 (58%)Asian: 102 (42%)NRNA: 100%; TDF: 92 (37.6%); ETV: 133 (54.3%); LAM: 15 (6.1%)Median: 144 IU/L (IQR Q3 − Q1 415)Median: 5.6 log IU/mL (IQR Q3 − Q1 2.7)Negative: 100%No coinfectionsPellicelli 2014 [59]Non-cirrhotic individuals receiving NA therapy ≥ 18 months (n = 193)Mean: 52.8 years (SD 13)Male: 135 (70%)White: 100%IFN: 83 (43%)TDF: 14/193; ADV: 6/193; ETV: 93/193; LAM: 57/193; LdT: 4/193; LAM + ADV: 19/193; NA rescue therapy: 35 (18%)Mean: 86 IU/L (SD 40)Mean: 5.2 log IU/mL (SD 1.7)Negative: 100%No coinfectionsRETRACT-B [60]Full study cohorta (N = 1,552)Mean: 52.9 years (SD 11.2)Male: 1,122 (72.3%)White: 175 (11.3%)Black: 13 (0.8%)Asian: 1,359 (87.6%)Other: 5 (0.3%)NA: 100%; TDF: 421 (27.1%); ETV: 981 (63.2%); IFN: 113 (8.6%)No current treatmentsMedian: 0.6 xULN (range 0.4–0.8)NRNegative: 100%No coinfectionsShim 2017 [61]Non-cirrhotic individuals treated with ETV (n = 133)Median: 49 years (IQR 44–55)Male: 75 (56.4%)Korean: 100%No prior treatmentsETV: 100%Median: 105 U/L (IQR 68–213)Median: 6.81 log IU/mL (IQR 5.55–7.97)Positive: 75 (56.4%)No coinfectionsShin 2018 [62]Full study cohorta (N = 894)Mean: 52.0 years (SD 11.3)Male: 597 (66.8%)NRNo prior treatmentsETV: 100%Median: 89 IU/mL (IQR 46–153)Mean: 6.39 log IU/mL (SD 1.41)Positive: 537 (60.1%)No coinfectionsSou 2020 [63]Full study cohorta (N = 1,397)Median: 50 years (IQR Q3 − Q1 17)Male: 1,018 (72.9%)Asian: 100%No prior treatmentsETV: 100%Median: 104 U/L (IQR Q3 − Q1 205)Median: 5.96 log IU/mL (IQR Q3 − Q1 2.35)Positive: 491 (35.1%)No coinfectionsTseng 2017 [64]Untreated ERADICATE-B study cohort (n = 2,075)Mean/median NRMale: 1,222 (58.89%)Female: 853 (41.11%)Asian: 100%No prior treatmentsNo current treatmentsMean/median NRMean/median NRPositive: 372 (17.93%)Negative: 1,703 (82.07%)No coinfectionsNA-treated validation cohort (n = 532)NRNRAsian: 100%NRNA: 100%NRMean/median NRNRNRWang 2016 [65]NA-experienced individuals (n = 105)Mean: 48.9 years (SD 13.4)Male: 83 (79%)NRNA: 100%TDF: 100%Mean: 208.2 IU/L (SD 336.6)Mean: 6.2 log copies/mL (SD 1.9)Positive: 51 (48.6%)No coinfectionsNA-naïve individuals (n = 131)Mean: 48.9 years (SD 12.9)Male: 100 (76.3%)NRNo prior treatmentsTDF: 100%Mean: 238.7 IU/L (SD 419.8)Mean: 6.9 log copies/mL (SD 1.5)Positive: 42 (32.1%)No coinfectionsWong 2013 [66]Non-cirrhotic individuals (n = 1,199)Mean: 50 years (SD 12)Male: 863 (72%)NRNA: 358 (30%); IFN: 28 (2%)ETV: 100%Mean: 156 IU/L (SD 339)Mean: 5.0 log IU/mL (SD 2.2)Positive: 358 (32%)Negative: 813 (68%)No coinfectionsWu 2017 [67]Individuals receiving TDF (n = 106)Mean: 47.1 years (SD 12.1)Male: 74 (70%)NRNo prior treatmentsTDF: 100%Mean: 303 U/L (SD 372)Mean: 7.35 log IU/mL (SD 0.70)Positive: 50 (47.1%)No coinfectionsIndividuals receiving ETV (n = 313)Mean: 47 years (SD 12.3)Male: 230 (73%)NRNo prior treatmentsETV: 100%Mean: 316 U/L (SD 440)Mean: 7.18 log IU/mL (SD 0.74)Positive: 172 (55.0%)No coinfectionsYamada 2015 [68]Non-cirrhotic individuals receiving ETV (n = 404)Mean: 51.3 years (SD 12.1)Male: 233 (58%)Female: 171/404NRIFN: 44 (11%)ETV: 100%Mean: 156.1 IU/L (SD 210.8)Median: 6.9 log copies/mLPositive: 181 (45%)Negative: 219/404No coinfectionsYang 2013 [69]NA-naïve individualsa (n = 323)Mean: 48.0 years (SD 12.2)Male: 256 (79.3%)NRIFN: 20 (6.2%)ETV: 100%Mean: 324.5 U/L (SD 476.4)Mean: 6.4 log copies/mL (SD 1.5)Positive: 118 (36.5%)No coinfectionsNA-experienced individualsa (n = 164)Mean: 48.2 years (SD 11.2)Male: 144 (87.8%)NRADV: 45 (27.4%); LAM: 155 (94.5%); IFN: 15 (9.1%)ETV: 100%Mean: 277.9 U/L (SD 460.3)Mean: 6.3 log copies/mL (SD 1.7)Positive: 69 (42.1%)No coinfectionsYang 2014 [70]Individuals who received LAM prophylaxis for transplant (n = 26)Mean: 41 years (SD 14.2)Male: 16/26Female: 10/26NRNRLAM+immunosuppressive therapy: 100%NRMean: 3.1 log IU/mL (SD 2.8)Negative: 100%No coinfectionsIndividuals who received LAM post-transplant (n = 28)Mean: 42.7 years (SD 10.5)Male: 18/28Female: 10/28NRNRLAM+immunosuppressive therapy: 100%NRMean: 2.6 log IU/mL (SD 2.1)Negative: 100%No coinfectionsYip 2020 [71]ETV-treated individualsa (n = 28,041)Mean: 53.4 years (SD 13.0)Male: 18.094 (64.5%)Asian: 100%No prior treatmentsETV: 100%Median: 62.0 U/L (IQR 33.0-137.0)Mean: 5.3 log IU/mL (SD 2.2)Positive: 8,317 (29.7%)No coinfectionsTDF-treated individualsa (n = 1,309)Mean: 43.2 years (SD 13.1)Male: 591 (45.1%)Asian: 100%No prior treatmentsTDF: 100%Median: 43.0 U/L (IQR 25.0-103.0)Mean: 4.9 log IU/mL (SD 2.7)Positive: 721 (55.1%)No coinfectionsYuen 2007 [72]Non-cirrhotic LAM-treated individuals (n = 142)Median: 33.9 years (range 20.2–54.4)Male: 106/142Female: 36/142Asian: 100%NA: 99/142LAM: 100%Median: 65 U/L (range 14–838)Median: 8.7 log copies/mL (range 4.2–11.7)Positive: 100%No coinfectionsNon-cirrhotic untreated individuals (n = 124)Median: 33.4 years (range 20.8–59)Male: 90/124Female: 34/124Asian: 100%NRNo current treatmentsMedian: 56.5 U/L (range 9-850)Median: 6.8 log copies/mL (range 1.5–9.6)Positive: 100%No coinfectionsAbbreviations: ADV adefovir, ALT alanine aminotransferase, ART antiretroviral therapy, CHB chronic hepatitis B, CLV clevudine, ETV entecavir, HBeAg hepatitis B e-antigen, HBV hepatitis B virus, HCC hepatocellular carcinoma, HCV hepatitis C virus, HIV human immunodeficiency virus, IFN interferon, IQR interquartile range, IU international unit, LAM lamivudine, LdT telbivudine, NA nucleos(t)ide analogue, NR not reported, SD standard deviation, TAF tenofovir alafenamide, TDF tenofovir disoproxil fumarate, ULN upper limit of normalaBaseline characteristics were not reported separately for individuals without cirrhosis at baseline (study reports a sub analysis of HCC risk in non-cirrhotic individuals and therefore is eligible for SLR inclusion)b251 out of 363 patients initially on initial LAM/LdT/CLV switched to ADV (n = 149), ETV (n = 90), and other NAs (n = 12). Of the 161 patients initially started on ETV, 2 patients switched to ETV + ADV and 2 patients switched to TDFc17 out of 25 patients receiving LAM discontinued treatment. Of these 17 patients, 4 switched to ETV. 6 out of 14 patients receiving ETV discontinued treatment, of these, 2 patients restarted ETV after relapsingd14 out of 38 patients receiving LAM discontinued treatment, of these, 2 patients started ETV after relapsingen=74 switched from LAM monotherapy to LAM + ADV; n = 67 switched from LAM to ETV Summary of baseline characteristics Male: 79/133 Female: 54/133 Positive: 41/133 Negative: 92/133 Male: 244/303 Female: 59/303 Male: 126/157 Female: 31/157 Mean: 316.4 U/L (SD 443.9) Mean: 6.03 log IU/mL (SD 1.60) Positive: 0% Negative: 100% Male: 1,306 (63.0%) Female: 767 (37.0%) Male: 210/316 Female: 106/316 Male: 136/182 Female: 46/182 Male: 210/316 Female: 106/316 White: 283 (59%) Asian: 148 (31%) Other: 51 (11%) Male: 179 (69.9%) Female: 77 (30.1%) ETV: 129 (50.4%) LAM: 127 (49.6%) Positive: 132 (51.6%) Negative: 124 (48.4%) Male: 67 (66%) Female: 35 (34%) Full study cohorta (N = 462) ETV: 231 (50.0%) LAM: 231 (50.0%) Male: 816/1,200 Female: 384/1,200 Positive: 7/18 Negative: 11/18 Positive: 164/378 Negative: 214/378 Male: 28 (72%) Female: 11 (28%) Positive: 0% Negative: 100% Male: 37 (67%) Female: 18 (33%) Positive: 100% Negative: 0% Male: 381/602 Female: 221/602 Positive: 295/600 Negative: 305/600 White: 143 (58%) Asian: 102 (42%) White: 143 (58%) Asian: 102 (42%) White: 175 (11.3%) Black: 13 (0.8%) Asian: 1,359 (87.6%) Other: 5 (0.3%) Male: 1,222 (58.89%) Female: 853 (41.11%) Positive: 372 (17.93%) Negative: 1,703 (82.07%) Positive: 358 (32%) Negative: 813 (68%) Male: 233 (58%) Female: 171/404 Positive: 181 (45%) Negative: 219/404 Male: 16/26 Female: 10/26 Male: 18/28 Female: 10/28 Male: 106/142 Female: 36/142 Male: 90/124 Female: 34/124 Abbreviations: ADV adefovir, ALT alanine aminotransferase, ART antiretroviral therapy, CHB chronic hepatitis B, CLV clevudine, ETV entecavir, HBeAg hepatitis B e-antigen, HBV hepatitis B virus, HCC hepatocellular carcinoma, HCV hepatitis C virus, HIV human immunodeficiency virus, IFN interferon, IQR interquartile range, IU international unit, LAM lamivudine, LdT telbivudine, NA nucleos(t)ide analogue, NR not reported, SD standard deviation, TAF tenofovir alafenamide, TDF tenofovir disoproxil fumarate, ULN upper limit of normal aBaseline characteristics were not reported separately for individuals without cirrhosis at baseline (study reports a sub analysis of HCC risk in non-cirrhotic individuals and therefore is eligible for SLR inclusion) b251 out of 363 patients initially on initial LAM/LdT/CLV switched to ADV (n = 149), ETV (n = 90), and other NAs (n = 12). Of the 161 patients initially started on ETV, 2 patients switched to ETV + ADV and 2 patients switched to TDF c17 out of 25 patients receiving LAM discontinued treatment. Of these 17 patients, 4 switched to ETV. 6 out of 14 patients receiving ETV discontinued treatment, of these, 2 patients restarted ETV after relapsing d14 out of 38 patients receiving LAM discontinued treatment, of these, 2 patients started ETV after relapsing en=74 switched from LAM monotherapy to LAM + ADV; n = 67 switched from LAM to ETV Regarding treatment status, five studies reported on untreated patients with CHB. The most common treatments received during follow-up were ETV (n = 39 studies), NA combinations (n = 26 studies), TDF (n = 19 studies), and LAM (n = 16 studies). HBeAg status varied across studies, with eight studies including only HBeAg-negative individuals and five including only HBeAg-positive individuals. In mixed populations (n = 37 studies), HBeAg positivity ranged from 15.4% to 77.3%. ALT levels were reported in 45 studies, with 15 studies reporting mean values ranging from 21.4 to 361.7 IU/L and 24 studies reporting median values ranging from 23.0 to 172.0 IU/L. HBV DNA was reported by 46 studies, showing mean levels ranging from 1.86 log IU/mL in untreated, HBeAg-negative individuals with inactive CHB [24], to 7.7 log IU/mL in individuals who were HBeAg-positive for at least 6 months prior to study enrollment [34]. Additionally, Moon 2015 included participants with hepatitis C virus (HCV) co-infection [49] and Dezanet 2021 included those with human immunodeficiency virus (HIV) co-infection [29]. SLR outcomes A summary of outcomes reported across included studies is shown in Table 3. Most studies assessed specific time points or mean/median follow-up durations of at least five years, and 24 studies assessed incidence beyond five years of follow-up. Incidence at specific timepoints less than five years was assessed by 17 studies. Incidence at specific timepoints at five years of follow-up or beyond was assessed by 24 studies. Table 3Summary of study outcomesStudyCohortCohort Description N Follow-up DurationAssessment TimepointCumulative IncidenceIncidence RateMeasures of Association n %Ando 2018 [22]NA-treated virological respondersPatients with sustained disappearance of viremia, defined as continuous HBV DNA levels < 2.1 log copies/mL105Median: 4.8years (range 1.0–13.1)a1 yearNR0NRNR3 yearsNR2.2NRNR5 years45.3NRHR: 5.563 (95% CI: 1.438, 21.519), p = 0.013 (cirrhotic vs. non-cirrhotic)Chiu 2020 [23]All PSM TDF or ETV discontinuation patientsAll PSM patients, all of whom were HBeAg- and discontinued ETV or TDF treatment per 2012 APASL guidelines; follow-up for HCC development may include re-treatment period for some patients460TDF: mean: 158.7 weeks (SD 10.8)ETV: mean: 159 weeks (SD 29.1)60 months133.5NRNRCho 2014 [24]NA complete respondersComplete responders to NA therapy, defined as continuous virological response (HBV DNA < 2,000 IU/mL)754Mean: 37.6 months (SD 21.6)a1 yearNR2.0NRNR3 yearsNR4.7NRNR5 yearsNR7.2NRNRNA incomplete respondersIncomplete responders to NA therapy, defined as HBV DNA fluctuation ≥ 2,000 IU/mL during follow-up179Mean: 51.7 months (SD 19.8)a1 yearNR3.3NRNR3 yearsNR6.4NRNR5 yearsNR11.1NRNRHBeAg- untreated inactive CHBHBeAg- patients with normal ALT and serum HBV DNA < 2,000 IU/mL throughout follow-up884Mean: 42.7 months (SD 22.4)a5 yearsNR0.8NRNRHBeAg- NA complete respondersHBeAg- complete responders to NA therapy, defined as continuous virological response (HBV DNA < 2,000 IU/mL)316Mean: 37.9 months (SD 21.5)a5 years156.9NRNRChoi 2020 [25]ETV- or TDF-treatedNR2,612Median: 5.6 years (range 1.0–10.9)aAnnual (entire follow-up)NR0.6NRHR: 2.30 (95% CI: 1.77, 2.99), p < 0.001 (cirrhotic vs. non-cirrhotic)Choi 2022a [26]All ETV- or TDF-treated patientsAll treated patients, all of whom were HBeAg+2,073Median: 5.7 years (IQR 3.6–8.1)Entire follow-up47NR0.39 per 100 person-years (95% CI NR)NRETV- or TDF-treated, HBV DNA ≥ 8.00 log IU/mLHBeAg+ treated patients with HBV DNA ≥ 8.00 log IU/mL1,108Median: 5.8 years (IQR 3.5–8.2)Entire follow-up10NR0.15 per 100 person-years (95% CI: 0.08, 0.27)HR: 1 (reference)ETV- or TDF-treated, HBV DNA 7.00–7.99 log IU/mLHBeAg+ treated patients with HBV DNA 7.00–7.99 log IU/mL521Median: 5.5 years (IQR 3.7–8.3)Entire follow-up14NR0.46 per 100 person-years (95% CI: 0.26, 0.75)HR: 2.48 (95% CI: 1.10, 5.62), p = 0.03 (treated patients with HBV DNA 7.00–7.99 log IU/mL vs. treated patients with HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA 6.00–6.99 log IU/mLHBeAg+ treated patients with HBV DNA 6.00–6.99 log IU/mL274Median: 5.5 years (IQR 3.5–7.6)Entire follow-up13NR0.84 per 100 person-years (95% CI: 0.47, 1.40)HR: 3.69 (95% CI: 1.58, 8.59), p = 0.002 (treated patients with HBV DNA 6.00–6.99 log IU/mL vs. treated patients with HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA 5.00–5.99 log IU/mLHBeAg+ treated patients with HBV DNA 5.00–5.99 log IU/mL170Median: 5.0 years (IQR 3.5–7.0)Entire follow-up10NR1.14 per 100 person-years (95% CI: 0.58, 2.03)HR: 6.10 (95% CI: 2.46, 15.11), p < 0.001 (treated patients with HBV DNA 5.00–5.99 log IU/mL vs. treated patients with HBV DNA ≥ 8.00 log IU/mL)PSM high viral loadPSM cohort of high viral load patients, ≥ 8.00 log IU/mL, PSM to medium viral load patients, 5.00-7.99 log IU/mL930Median: 6.0 years (IQR 3.7–8.3)Entire follow-up10NR0.18 per 100 person-years (95% CI: 0.09, 0.31)HR: 1 (reference)PSM medium viral loadPSM cohort of medium viral load patients, 5.00-7.99 log IU/mL, PSM to high viral load patients, ≥ 8.00 log IU/mL930Median: 5.4 years (IQR 3.6–7.9)Entire follow-up35NR0.66 per 100 person-years (95% CI: 0.47, 0.91)HR: 3.76 (95% CI: 1.86, 7.59), p < 0.001 (PSM medium viral load patients vs. PSM high viral load patients)PSM untreated, moderate viral loadUntreated patients with moderate viral load, 5.00-7.99 log IU/mL, PSM to treated patients with moderate viral load947Median: 7.8 years (IQR 4.9–11.9)aEntire follow-upNRNR1.52 per 100 person-years (95% CI: 1.28, 1.81)HR: 1 (reference)PSM treated, moderate viral loadHBeAg+ treated patients with moderate viral load, 5.00-7.99 log IU/mL, PSM to untreated patients with moderate viral load947Median: 5.6 years (IQR 3.5–8.0)aEntire follow-upNRNR0.67 per 100 person-years (95% CI: 0.47, 0.91)HR: 0.44 (95% CI: 0.30, 0.64), p < 0.001 (PSM treated patients with moderate viral load vs. PSM untreated patients with moderate viral load)PSM untreated, high viral loadUntreated patients with high viral load, ≥ 8.00 log IU/mL, PSM to treated patients with high viral load969Median: 7.8 years (IQR 4.9–11.9)aEntire follow-upNRNR0.26 per 100 person-years (95% CI: 0.16, 0.39)HR: 1 (reference)PSM treated, high viral loadHBeAg+ treated patients with high viral load, ≥ 8.00 log IU/mL, PSM to untreated patients with high viral load969Median: 5.6 years (IQR 3.5–8.0)aEntire follow-upNRNR0.10 per 100 person-years (95% CI: 0.04, 0.22)HR: 0.48 (95% CI: 0.19, 1.21), p = 0.12 (PSM treated patients with high viral load vs. PSM untreated patients with high viral load)Choi 2022b [27]ETV- or TDF-treated, HBV DNA ≥ 8.00 log IU/mLAll included patients were HBeAg+2,929Median: 4.28 yearsaEntire follow-upNR0.580.10 per 100 person-years (95% CI: 0.05, 0.17)HR: 1 (reference)ETV- or TDF-treated, HBV DNA ≥ 7.00 and < 8.00 log IU/mLAll included patients were HBeAg+1,984Median: 4.28 yearsaEntire follow-upNR3.080.48 per 100 person-years (95% CI: 0.35, 0.65)HR: 4.04 (95% CI: 1.68, 9.73), p = 0.0020 (HBV DNA ≥ 7.00 and < 8.00 log IU/mL vs. HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA ≥ 6.00 and < 7.00 log IU/mLAll included patients were HBeAg+1,561Median: 4.28 yearsaEntire follow-upNR8.041.38 per 100 person-years (95% CI: 1.11, 1.69)HR: 8.05 (95% CI: 3.34, 19.35), p < 0.0001 (HBV DNA ≥ 6.00 and < 7.00 log IU/mL vs. HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA ≥ 5.00 and < 6.00 log IU/mLAll included patients were HBeAg+1,071Median: 4.28 yearsaEntire follow-upNR7.511.17 per 100 person-years (95% CI: 0.88, 1.53)HR: 6.23 (95% CI: 3.10, 12.51), p < 0.0001 (HBV DNA ≥ 5.00 and < 6.00 log IU/mL vs. HBV DNA ≥ 8.00 log IU/mL)Chun 2021 [28]ETV- or TDF-treated, totally sedentaryPatients with no physical activity1,437bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR2.43 per 100 patient-years (95% CI NR)HR: 1 (reference)ETV- or TDF-treated, < 500 MET-min/weekPatients with a low level of physical activity as quantified using metabolic equivalent (MET)2,425bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.84 per 100 patient-years (95% CI NR)HR: 0.83 (95% CI: 0.65, 1.06), p = 0.13 (< 500 MET-min/week vs. totally sedentary)ETV- or TDF-treated, 500–1,000 MET-min/weekPatients with a moderate level of physical activity as quantified using metabolic equivalent (MET)2,326bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.67 per 100 patient-years (95% CI NR)HR: 0.72 (95% CI: 0.56, 0.92), p = 0.009 (500–1,000 MET-min/week vs. totally sedentary)ETV- or TDF-treated, 1,000–1,499 MET-min/weekPatients with a moderate level of physical activity as quantified using metabolic equivalent (MET)988bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.56 per 100 patient-years (95% CI NR)HR: 0.66 (95% CI: 0.47, 0.92), p = 0.02 (1,000–1,499 MET-min/week vs. totally sedentary)ETV- or TDF-treated, ≥ 1,500 MET-min/weekPatients with a high level of physical activity as quantified using metabolic equivalent (MET)433bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.61 per 100 patient-years (95% CI NR)HR: 0.68 (95% CI: 0.44, 1.06), p = 0.09 (≥ 1,500 MET-min/week vs. totally sedentary)Dezanet 2021 [29]TDF-/TAF-based ART-treated, low fibrosis without progressionHIV+ patients receiving TDF-/TAF-based ART with sustained no-to-mild fibrosis (F0-F1)50Median: 7.4 years (IQR 3.1–12.9)Entire follow-up00.0NRNRTDF-/TAF-based ART-treated, low fibrosis with progressionHIV+ patients receiving TDF-/TAF-based ART with initially low level of fibrosis, but a slow, linear increase towards advanced fibrosis during treatment38Median: 8.5 years (IQR 3.1–13.0)Entire follow-up00.0NRNRTDF-/TAF-based ART-treated, moderate fibrosis with high fluctuationHIV+ patients receiving TDF-/TAF-based ART with a heterogenous distribution of initial fibrosis levels (ranging from F0-F1 to F4) and equally heterogenous fluctuations in FibroTest scores during treatment67Median: 6.6 years (IQR 3.2–13.2)Entire follow-up23.0NRNRHosaka 2010 [30]LAM + ADV-treatedPatients with LAM-resistant mutations186Median: 115 weeks (range 25–282)1 yearNR2NRNR3 yearsNR11NRNR5 yearsNR16NRHR: 1.42 (95% CI: 1.04, 1.96), p = 0.03 (cirrhotic vs. non-cirrhotic)Hosaka 2013 [31]PSM ETV-treatedPatients who received ETV, PSM to the control group237bMedian: 3.3 years (IQR 2.3–4.3)a2 yearsNR0NRNR3 yearsNR0NRNR4 yearsNR0.8NRNR5 yearsNR2.5NRHR: 0.37 (95% CI: 0.15, 0.91), p = 0.030 (PSM ETV vs. PSM untreated)PSM untreatedUntreated historical control group, PSM to the ETV group231bMedian: 7.6 years (IQR 3.4–13.7)a2 yearsNR1.0NRNR3 yearsNR1.6NRNR4 yearsNR2.2NRNR5 yearsNR3.6NRHR: 1 (reference)PSM LAM-treatedPatients who received LAM, PSM to the ETV group133bMedian: 6.8 years (IQR 5.0–9.9)a2 yearsNR3.2NRNR3 yearsNR3.2NRNR4 yearsNR3.2NRNR5 yearsNR4.9NRNRHosaka 2022 [32]ETV-treatedAll included patients were HBeAg-132bMedian: 11 yearsa5 years9NRNRHR: 4.24 (95% CI: 1.81, 9.91), p = 0.001 (preexisting cirrhosis vs. no preexisting cirrhosis)Ji 2022 [33]ETV-treatedPatients who received ETV + placebo500Median: 89 monthsaEntire follow-up438.6NRNRKim 2015 [34]TDF-treatedPatients receiving open-label TDF482Median: 7.36 years (range 0.00–7.55)aCumulative incidence: annual (entire follow-up)Incidence rate: entire follow-up80.2828.4 per 10,000 person-years (95% CI NR)NRKim 2017a [35]ETV-treated virological respondersPatients with a virological response to ETV, defined as serum HBV DNA < 15 IU/mL at 1 year of therapy850Median: 5.7 years (IQR 4.8–7.0)Entire follow-up37NR0.74 per 100 patient-years (95% CI: 0.52, 1.02)NRKim 2017b [36]ETV-treated maintained virologic responsePatients with complete virologic response to ETV, defined as HBV DNA < 12 IU/mL (limit of detection), and persistently undetectable HBV DNA throughout follow-up224Median: 4.5 years (range 1.0–8.7)a5 yearsNR4.0NRHR: 1 (reference)ETV-treated low-level viremiaPatients with HBV DNA ≥ 12 IU/mL but < 2,000 IU/mL persistently or intermittently during follow-up, whether or not they previously achieved complete virologic response on ETV208Median: 4.5 years (range 1.0–8.7)a5 yearsNR6.9NRHR: 1.65 (95% CI: 0.65, 4.17), p = 0.29 (low-level viremia vs. maintained virologic response)All ETV-treated patientsAll relevant patients, including the maintained virologic response and low-level viremia cohorts432Median: 4.5 years (range 1.0–8.7)aEntire follow-upNRNRNRHR: 3.53 (95% CI: 2.11, 5.94), p < 0.001 (cirrhotic vs. non-cirrhotic)Kim 2018 [37]PSM TDF-treatedPatients were PSM to an ETV-treated group198Median: 33 months (IQR 21–46)aEntire follow-up000 per 100 person-years (95% CI NR)NRKo 2019 [38]ETV-treatedNR1,033bMedian: 6.6 years (IQR 2.8)aEntire follow-up383.7NRNR2 years7NRNRNR3 years14NRNRNR4 years24NRNRNR5 years29NRNRNR6 years32NRNRNR7 years36NRNRNR8 years36NRNRNR9 years38NRNRNR10 years38NRNRNRKobashi 2011 [39]ETV- or LAM-treatedNR194Median: 4.25 years (range 0.41–10.0)a1 yearNR1.6NRNR3 yearsNR3.5NRNR5 yearsNR3.5NRNR7 yearsNR7.1NRNR10 years1429.6NRNRKozuka 2019 [40]ETV-treatedNR105Median: 4.4 years (range 1.0–10.7)a3 yearsNR3.5NRNR5 yearsNR5.4NRNR7 yearsNR5.4NRHR: 4.789 (95% CI: 1.296, 17.689), p = 0.019 (cirrhotic vs. non-cirrhotic)Lampertico 2007 [41]ADV- and LAM-treatedAll patients had LAM-resistant HBV39Median: 42 months (range 12–74)aEntire follow-up0NRNRNRLee 2016 [42]ETV-treatedNR66Median: 45.2 months (IQR 36.0–58.3)a1 yearNR0NRNR3 yearsNR1.90NRNR5 yearsNR1.90NRHR: 9.07 (95% CI: 1.09, 75.69), p = 0.042 (cirrhotic vs. non-cirrhotic)Lee 2019 [43]ETV- or LAM-treatedPatients had a baseline METAVIR score of F372NR5 years58.9NRNRLee 2021 [44]PSM TDF combination therapyPSM patients who received TDF-containing combination therapy after switching from other NA treatments2,030Median: 45.3 months (IQR 39.0–48.6)aEntire follow-up58NR0.75 per 100 patient-years (95% CI NR)HR: 1 (reference)PSM TDF monotherapyPSM patients who received TDF monotherapy for at least 6 months after switching from other NA treatments2,030Median: 43.7 months (IQR 34.8–49.0)aEntire follow-up67NR0.87 per 100 patient-years (95% CI NR)HR: 1.21 (95% CI: 0.85, 1.71), p = 0.3 (PSM TDF monotherapy vs. PSM TDF combination therapy)Lee 2022 [45]ETV- or TDF-treatedNR787Median: 63 months (IQR 37.6–98.2)aEntire follow-up29NRNRHR: 2.974 (95% CI: 1.865, 4.743), p < 0.0001 (cirrhotic vs. non-cirrhotic)Li 2017 [46]NA-treatedPatients with < 9.7 kPa LSM were considered to be patients without cirrhosis965NR1 yearNR0.83NRNR2 yearsNR1.83NRNR3 years822.75NRNRLim 2014 [47]PSM LAM-treatedPatients treated with LAM 100 mg/day, PSM to an ETV-treated cohort (which was not extracted due to overlap with other studies included in this SLR)878Median: 8.7 years (IQR 6.5–11.5)aEntire follow-up40NR0.80 per 100 patient-years (95% CI NR)HR: 1.26 (95% CI: 0.68, 2.34), p = 0.46 (PSM ETV vs. PSM LAM)Lin 2019 [48]NA-treatedPatients treated with LAM, ETV, or LdT who were followed up for more than 2 years270Mean: 5 yearsEntire follow-up5NRNRNR2 years2NRNRHR: 4.107 (95% CI: 1.424, 11.843), p = 0.009 (cirrhotic vs. non-cirrhotic)Moon 2015 [49]NA-treatedNR330Median: 48 months (range 8–60)a1 yearNR0NRNR2 yearsNR0.30NRNR3 yearsNR0.30NRNR4 yearsNR0.30NRNR5 yearsNR2.10NRHR: 17.16 (95% CI: 4.67, 63.12), p < 0.001 (cirrhotic vs. non-cirrhotic)Nagata 2016 [50]All ETV- or LAM-treated patientsAll patients who started treatment with LAM or ETV, including several who discontinued treatment if they satisfied APASL stopping criteria94HBeAg-: mean: 108 months (range 18–182)HBeAg+: mean: 111 months (range 16–178)Entire follow-up55.3NRNRETV- or LAM-treated, HBeAg-HBeAg- patients who started treatment with LAM or ETV, including several who discontinued treatment if they satisfied APASL stopping criteria39Mean: 108 months (range 18–182)Annual over entire follow-upNR0.6NRNRETV- or LAM-treated, HBeAg+HBeAg+ patients who started treatment with LAM or ETV, including several who discontinued treatment if they satisfied APASL stopping criteria55Mean: 111 months (range 16–178)Annual over entire follow-upNR0.6NRNRNiro 2013 [51]Continuous NA therapyNR74Mean: 6 years (SD 3)aEntire follow-up11NRNROh 2020 [52]PSM ETV-treatedNR278Median: 4.7 years (IQR 4.3–5.3)aEntire follow-up3NRNRHR: 1 (reference)PSM TDF-treatedNR278Median: 4.8 years (IQR 3.9–5.4)aEntire follow-up4NRNRHR: 1.41 (95% CI: 0.32, 6.32), p = 0.651 (PSM TDF vs. PSM ETV)Orito 2015 [53]All NA-treated patientsAll relevant patients492Median: 90 months (range 12–204)Annual over entire follow-up130.34NRNRNA-treated, HBV DNA+Patients who did not lose HBV DNA positivity during follow-up188Median: 90 months (range 12–204)aEntire follow-up158.0NRNRNA-treated, HBV DNA-Patients who became HBV DNA- during follow-up230Median: 90 months (range 12–204)aEntire follow-up135.7NRNRNA-treated, HBeAg+ during therapyPatients who were HBeAg+ during follow-up99Median: 90 months (range 12–204)aEntire follow-up44.0NRNRNA-treated, HBeAg seroconversionPatients who experienced HBeAg seroconversion during follow-up39Median: 90 months (range 12–204)aEntire follow-up25.1NRNRNA-treated, HBsAg+ during therapyPatients who were HBsAg+ during follow-up479Median: 90 months (range 12–204)aEntire follow-up132.7NRNRNA-treated, HBsAg seroclearancePatients who experienced HBsAg seroclearance during follow-up13Median: 90 months (range 12–204)aEntire follow-up00.0NRNRPAGE-B Cohort [54]All NA-treated, baseline non-cirrhotic patientsPatients who did not develop HCC within the first 5 years of treatment1,023Median: 8.4 yearsaYears 5–12121.2NRHR: 4.81 (95% CI: 2.36, 9.79), p < 0.001 (cirrhotic vs. non-cirrhotic)6 yearsNRNR0.5 per 100 patient-years (95% CI NR)NR8 yearsNRNR1.2 per 100 patient-years (95% CI NR)NR10 yearsNRNR1.7 per 100 patient-years (95% CI NR)NRNA-treated, non-cirrhotic at baseline and at year 5Patients who did not develop HCC within the first 5 years of treatment and had LSM < 12 kPa at year 5870Median: 8.4 yearsaYears 5–129NRNRNR6 yearsNRNR0.5 per 100 patient-years (95% CI NR)NR8 yearsNRNR1.1 per 100 patient-years (95% CI NR)NR10 yearsNRNR1.3 per 100 patient-years (95% CI NR)NRNA-treated, non-cirrhotic at baseline, cirrhotic at year 5Patients who did not develop HCC within the first 5 years of treatment and had LSM ≥ 12 kPa at year 53Median: 8.4 yearsaYears 5–121NRNRNRPapatheodoridis 2015 [55]ETV- and/or TDF-treatedNR1,085Median: 38.8 months (IQR 23.4–49.9)a1 yearNR0.6NRNR3 yearsNR1.4NRNR5 yearsNR3.7NRHR: 4.50 (95% CI: 2.55, 7.96), p < 0.001 (compensated cirrhotic vs. non-cirrhotic)Papatheodoridis 2017 [56]ETV- and/or TDF-treatedNR1,379Median: 72 months (IQR 35)a5 yearsNRNR0.49 per 100 patient-years (95% CI NR)NRETV- and/or TDF-treatedPatients analyzed after excluding person-time accrued and HCC cases diagnosed within the first year of follow-upNRMedian: 72 months (IQR 35)a5 yearsNRNR0.51 per 100 patient-years (95% CI NR)NRPapatheodoridi 2022 [57]PSM NA discontinuationHBeAg- patients from a Caucasian cohort who discontinued NA treatment within 5 years and an Asian cohort who discontinued treatment within 3 years without developing HCC245Median: 47 months (IQR 34)a1 year70.9NRNR3 years71.9NRNR5 years75.1NRHR: 1.49 (95% CI: 0.44, 5.12), p = 0.523 (PSM NA discontinuation vs. PSM NA continuation)PSM NA continuationHBeAg- patients from a Caucasian cohort who continued NA treatment beyond 5 years and an Asian cohort who continued treatment beyond 3 years without developing HCC245Median: 40 months (IQR 32)a1 year70.4NRNR3 years72.5NRNR5 years74.9NRHR: 1 (reference)NA discontinuation, with low virological relapseNon-PSM HBeAg- patients with low virological relapse, defined as HBV DNA > 2,000 IU/mL228Median: 47 months (IQR 34)aEntire follow-up52.2NRHR: 1.64 (95% CI: 0.18, 14.78), p = 0.659 (NA discontinuation with vs. without low virological relapse)NA discontinuation, without low virological relapseNon-PSM HBeAg- patients without low virological relapse, defined as HBV DNA > 2,000 IU/mL97Median: 47 months (IQR 34)aEntire follow-up22.1NRHR: 1 (reference)NA discontinuation, with high virological relapseNon-PSM HBeAg- patients with high virological relapse, defined as HBV DNA > 20,000 IU/mL172Median: 47 months (IQR 34)aEntire follow-up42.3NRHR: 1.45 (95% CI: 0.24, 8.71), p = 0.688 (NA discontinuation with vs. without high virological relapse)NA discontinuation, without high virological relapseNon-PSM HBeAg- patients without high virological relapse, defined as HBV DNA > 20,000 IU/mL153Median: 47 months (IQR 34)aEntire follow-up32.0NRHR: 1 (reference)NA discontinuation, with HBsAg seroclearanceNon-PSM HBeAg- patients with HBsAg seroclearance, which was assessed prior to HCC diagnosis60Median: 47 months (IQR 34)aEntire follow-up00.0NRHR: 2.61 (95% CI: 0.43, 15.63), p = 0.295 (NA discontinuation with vs. without HBsAg seroclearance)NA discontinuation, without HBsAg seroclearanceNon-PSM HBeAg- patients without HBsAg seroclearance, which was assessed prior to HCC diagnosis265Median: 47 months (IQR 34)aEntire follow-up72.6NRHR: 1 (reference)Pellicelli 2014 [59]LAM-treatedAll patients were HBeAg- and had HBV genotype D57Mean: 69 months (SD 30)aEntire follow-upNRNR0.001 per 100 person-years (95% CI NR)OR: 7.9 (95% CI: 0.74, 85.00), p = 0.08 (cirrhotic vs. non-cirrhotic)NA-treatedAll patients were HBeAg- and had HBV genotype D193Mean: 69 months (SD 30)Entire follow-up21NROR: 7.6 (95% CI: 1.37, 42.20), p = 0.02 (cirrhotic vs. non-cirrhotic)10 years20.5NRNRRETRACT-B [60]NA discontinuationAll patients were HBeAg-, virally suppressed, and met their study center’s criteria for stopping NA therapy1,368Median: 18.4 years (range 7.9–39.4)aEntire follow-up100.7NRNRShim 2017 [61]ETV-treatedNR133Median: 3.6 years (IQR 2.0–5.1)aEntire follow-up2NR4.0 per 1,000 patient-years (95% CI NR)HR: 6.42 (95% CI: 1.42, 28.95), p = 0.016 (stage 1 cirrhotic vs. non-cirrhotic)HR: 24.89 (95% CI: 5.77, 107.30), p < 0.001 (stage 2 cirrhotic vs. non-cirrhotic)HR: 17.01 (95% CI: 3.86, 74.90), p < 0.001 (stage 3 cirrhotic vs. non-cirrhotic)Shin 2018 [62]ETV-treatedNR454Median: 5.4 years (IQR 3.0–7.3)aEntire follow-up71.5NRNR5 yearsNR1.5NRNR7 years72.5NRNRAnnual (7 years)NR0.4NRNRETV-treated, with good adherencePatients with ETV adherence ≥ 90%320Median: 5.4 years (IQR 3.0–7.3)aEntire follow-up51.6NRHR: 1.002 (95% CI: 0.978, 1.027), p = 0.836 (good vs. poor adherence)5 yearsNR1.7NRNR7 years51.7NRNRAnnual (7 years)NR0.2NRNRETV-treated, with poor adherencePatients with ETV adherence < 90%134Median: 5.4 years (IQR 3.0–7.3)aEntire follow-up21.5NRHR: 1 (reference)5 yearsNR1.5NRNR7 years24.9NRNRAnnual (7 years)NR0.7NRNRSou 2020 [63]ETV-treatedNR890NR5 yearsNR2.5NRNRAnnual (0 to < 5 years)NR0.63NRNR10 yearsNR3.2NRNR340NRAnnual (5–10 years)NR0.14NRNRTseng 2017 [64]All ERADICATE-B untreated patientsAll patients in the retrospective, untreated ERADICATE-B cohort2,075Median: 15.38 years (range 2.51–27.72)Entire follow-up137NR0.41 per 100 person-years (95% CI NR)NRERADICATE-B untreated patients with 15-year follow-upRetrospective, untreated ERADICATE-B cohort patients who were followed-up for 15 years1,241Median: 15.38 years (range 2.51–27.72)a15 years105NRNRNRERADICATE-B untreated inactive carriersRetrospective, untreated ERADICATE-B cohort patients who met the following criteria: HBeAg-, HBV DNA < 2,000 IU/mL, HBsAg < 1,000 IU/mL, and ALT < 40 U/L408Median: 15.38 years (range 2.51–27.72)aAnnual (15 years)NR0.06NRNRNA-treated validation cohortPatients in the validation cohort received long-term NA therapy532Median: 6.58 years (range 1.28–11.88)Entire follow-up10NR0.27 per 100 person-years (95% CI NR)NRWang 2016 [65]TDF-treatedNR141Median: 154 weeks (range 52–196)36 months00NRNRWong 2013 [66]ETV-treatedNR1,199Mean: 43 months (SD 13)3 yearsNR1.8NRNR5 years212.1NRHR: 3.2 (95% CI: 1.5, 6.4), p = 0.002 (cirrhotic vs. non-cirrhotic)Wu 2017 [67]ETV- or TDF-treatedNR296ETV-treated: mean: 49 months (SD 19.1)aTDF-treated: mean: 37.9 months (SD 7.2)aEntire follow-up3NRNRHR: 6.22 (95% CI: 1.81, 21.37), p = 0.004 (cirrhotic vs. non-cirrhotic)Yamada 2015 [68]ETV-treatedNR404Mean: 49.2 months (range 14–109)3 years162.7NRNR5 years165.3NRHR: 4.765 (95% CI: 2.415, 9.404), p < 0.001 (cirrhotic vs. non-cirrhotic)ETV-treated, with virological response at 24 weeksPatients with a virological response, defined by serum HBV DNA levels that were continuously < 2.6 log copies/mL275bMean: 49.2 months (range 14–109)a3 yearsNR3.5NRNR5 yearsNR7.2NRNRETV-treated, without virological response at 24 weeksPatients without a virological response, defined by serum HBV DNA levels that were continuously ≥ 2.6 log copies/mL129bMean: 49.2 months (range 14–109)a3 yearsNR1.0NRNR5 yearsNR1.0NRNRYang 2013 [69]ETV-treatedNR314Median: 36 months (range 12–73)a1 yearNR0NRNR2 yearsNR0.7NRNR3 yearsNR1.9NRNR4 yearsNR3.80NRHR: 7.10 (95% CI: 2.68, 18.83), p < 0.001 (cirrhotic vs. non-cirrhotic)Yang 2014 [70]LAM prophylaxisRenal transplant recipients who were HBeAg- and received LAM prophylaxis26Mean: 6.7 years (SD 2.4)5 years00NRNR10 years00NRNRLAM post-transplantRenal transplant recipients who were HBeAg- and received LAM later in the post-transplant period for HBV reactivation28Mean: 9.4 years (SD 4.7)5 years00NRNR10 years14.2NRNRYip 2020 [71]TDF-treatedNR1,271Median: 2.8 years (IQR 1.4–4.5)aEntire follow-up20.2NRNRETV-treatedNR24,219Median: 3.7 years (IQR 1.7–5.0)aEntire follow-up6462.7NRNRYuen 2007 [72]LAM-treatedAll patients were HBeAg+142Median: 89.9 months (range 26.5–128.3)Entire follow-up1NRNRNRUntreatedAll patients were HBeAg+124Median: 107.8 months (range 30.9–127.3)Entire follow-up3NRNRNRAbbreviations: ADV adefovir, ALT alanine transaminase, APASL Asian-Pacific Association for the Study of the Liver, ART antiretroviral therapy, CI confidence interval, DNA deoxyribonucleic acid, ETV entecavir, HBeAg hepatitis B e-antigen, HR hazard ratio, HBsAg hepatitis B surface antigen, HBV hepatitis B virus, HCC hepatocellular carcinoma, HIV human immunodeficiency virus, IQR interquartile range, LAM lamivudine, LdT telbivudine, LSM liver stiffness measurement, MET metabolic equivalent, N cohort size (denominator), n cases of HCC (numerator), NA nucleo(s/t)ide analog, NR not reported, OR odds ratio, PSM propensity score-matched, SD standard deviation, TAF tenofovir alafenamide, TDF tenofovir disoproxilaFollow-up durations were not reported separately for the cohort of interest and may include patients for whom outcomes were not extractedbCohort sample size calculated with information provided Table 3Summary of study outcomesStudyCohortCohort Description N Follow-up DurationAssessment TimepointCumulative IncidenceIncidence RateMeasures of Association n %Ando 2018 [22]NA-treated virological respondersPatients with sustained disappearance of viremia, defined as continuous HBV DNA levels < 2.1 log copies/mL105Median: 4.8years (range 1.0–13.1)a1 yearNR0NRNR3 yearsNR2.2NRNR5 years45.3NRHR: 5.563 (95% CI: 1.438, 21.519), p = 0.013 (cirrhotic vs. non-cirrhotic)Chiu 2020 [23]All PSM TDF or ETV discontinuation patientsAll PSM patients, all of whom were HBeAg- and discontinued ETV or TDF treatment per 2012 APASL guidelines; follow-up for HCC development may include re-treatment period for some patients460TDF: mean: 158.7 weeks (SD 10.8)ETV: mean: 159 weeks (SD 29.1)60 months133.5NRNRCho 2014 [24]NA complete respondersComplete responders to NA therapy, defined as continuous virological response (HBV DNA < 2,000 IU/mL)754Mean: 37.6 months (SD 21.6)a1 yearNR2.0NRNR3 yearsNR4.7NRNR5 yearsNR7.2NRNRNA incomplete respondersIncomplete responders to NA therapy, defined as HBV DNA fluctuation ≥ 2,000 IU/mL during follow-up179Mean: 51.7 months (SD 19.8)a1 yearNR3.3NRNR3 yearsNR6.4NRNR5 yearsNR11.1NRNRHBeAg- untreated inactive CHBHBeAg- patients with normal ALT and serum HBV DNA < 2,000 IU/mL throughout follow-up884Mean: 42.7 months (SD 22.4)a5 yearsNR0.8NRNRHBeAg- NA complete respondersHBeAg- complete responders to NA therapy, defined as continuous virological response (HBV DNA < 2,000 IU/mL)316Mean: 37.9 months (SD 21.5)a5 years156.9NRNRChoi 2020 [25]ETV- or TDF-treatedNR2,612Median: 5.6 years (range 1.0–10.9)aAnnual (entire follow-up)NR0.6NRHR: 2.30 (95% CI: 1.77, 2.99), p < 0.001 (cirrhotic vs. non-cirrhotic)Choi 2022a [26]All ETV- or TDF-treated patientsAll treated patients, all of whom were HBeAg+2,073Median: 5.7 years (IQR 3.6–8.1)Entire follow-up47NR0.39 per 100 person-years (95% CI NR)NRETV- or TDF-treated, HBV DNA ≥ 8.00 log IU/mLHBeAg+ treated patients with HBV DNA ≥ 8.00 log IU/mL1,108Median: 5.8 years (IQR 3.5–8.2)Entire follow-up10NR0.15 per 100 person-years (95% CI: 0.08, 0.27)HR: 1 (reference)ETV- or TDF-treated, HBV DNA 7.00–7.99 log IU/mLHBeAg+ treated patients with HBV DNA 7.00–7.99 log IU/mL521Median: 5.5 years (IQR 3.7–8.3)Entire follow-up14NR0.46 per 100 person-years (95% CI: 0.26, 0.75)HR: 2.48 (95% CI: 1.10, 5.62), p = 0.03 (treated patients with HBV DNA 7.00–7.99 log IU/mL vs. treated patients with HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA 6.00–6.99 log IU/mLHBeAg+ treated patients with HBV DNA 6.00–6.99 log IU/mL274Median: 5.5 years (IQR 3.5–7.6)Entire follow-up13NR0.84 per 100 person-years (95% CI: 0.47, 1.40)HR: 3.69 (95% CI: 1.58, 8.59), p = 0.002 (treated patients with HBV DNA 6.00–6.99 log IU/mL vs. treated patients with HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA 5.00–5.99 log IU/mLHBeAg+ treated patients with HBV DNA 5.00–5.99 log IU/mL170Median: 5.0 years (IQR 3.5–7.0)Entire follow-up10NR1.14 per 100 person-years (95% CI: 0.58, 2.03)HR: 6.10 (95% CI: 2.46, 15.11), p < 0.001 (treated patients with HBV DNA 5.00–5.99 log IU/mL vs. treated patients with HBV DNA ≥ 8.00 log IU/mL)PSM high viral loadPSM cohort of high viral load patients, ≥ 8.00 log IU/mL, PSM to medium viral load patients, 5.00-7.99 log IU/mL930Median: 6.0 years (IQR 3.7–8.3)Entire follow-up10NR0.18 per 100 person-years (95% CI: 0.09, 0.31)HR: 1 (reference)PSM medium viral loadPSM cohort of medium viral load patients, 5.00-7.99 log IU/mL, PSM to high viral load patients, ≥ 8.00 log IU/mL930Median: 5.4 years (IQR 3.6–7.9)Entire follow-up35NR0.66 per 100 person-years (95% CI: 0.47, 0.91)HR: 3.76 (95% CI: 1.86, 7.59), p < 0.001 (PSM medium viral load patients vs. PSM high viral load patients)PSM untreated, moderate viral loadUntreated patients with moderate viral load, 5.00-7.99 log IU/mL, PSM to treated patients with moderate viral load947Median: 7.8 years (IQR 4.9–11.9)aEntire follow-upNRNR1.52 per 100 person-years (95% CI: 1.28, 1.81)HR: 1 (reference)PSM treated, moderate viral loadHBeAg+ treated patients with moderate viral load, 5.00-7.99 log IU/mL, PSM to untreated patients with moderate viral load947Median: 5.6 years (IQR 3.5–8.0)aEntire follow-upNRNR0.67 per 100 person-years (95% CI: 0.47, 0.91)HR: 0.44 (95% CI: 0.30, 0.64), p < 0.001 (PSM treated patients with moderate viral load vs. PSM untreated patients with moderate viral load)PSM untreated, high viral loadUntreated patients with high viral load, ≥ 8.00 log IU/mL, PSM to treated patients with high viral load969Median: 7.8 years (IQR 4.9–11.9)aEntire follow-upNRNR0.26 per 100 person-years (95% CI: 0.16, 0.39)HR: 1 (reference)PSM treated, high viral loadHBeAg+ treated patients with high viral load, ≥ 8.00 log IU/mL, PSM to untreated patients with high viral load969Median: 5.6 years (IQR 3.5–8.0)aEntire follow-upNRNR0.10 per 100 person-years (95% CI: 0.04, 0.22)HR: 0.48 (95% CI: 0.19, 1.21), p = 0.12 (PSM treated patients with high viral load vs. PSM untreated patients with high viral load)Choi 2022b [27]ETV- or TDF-treated, HBV DNA ≥ 8.00 log IU/mLAll included patients were HBeAg+2,929Median: 4.28 yearsaEntire follow-upNR0.580.10 per 100 person-years (95% CI: 0.05, 0.17)HR: 1 (reference)ETV- or TDF-treated, HBV DNA ≥ 7.00 and < 8.00 log IU/mLAll included patients were HBeAg+1,984Median: 4.28 yearsaEntire follow-upNR3.080.48 per 100 person-years (95% CI: 0.35, 0.65)HR: 4.04 (95% CI: 1.68, 9.73), p = 0.0020 (HBV DNA ≥ 7.00 and < 8.00 log IU/mL vs. HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA ≥ 6.00 and < 7.00 log IU/mLAll included patients were HBeAg+1,561Median: 4.28 yearsaEntire follow-upNR8.041.38 per 100 person-years (95% CI: 1.11, 1.69)HR: 8.05 (95% CI: 3.34, 19.35), p < 0.0001 (HBV DNA ≥ 6.00 and < 7.00 log IU/mL vs. HBV DNA ≥ 8.00 log IU/mL)ETV- or TDF-treated, HBV DNA ≥ 5.00 and < 6.00 log IU/mLAll included patients were HBeAg+1,071Median: 4.28 yearsaEntire follow-upNR7.511.17 per 100 person-years (95% CI: 0.88, 1.53)HR: 6.23 (95% CI: 3.10, 12.51), p < 0.0001 (HBV DNA ≥ 5.00 and < 6.00 log IU/mL vs. HBV DNA ≥ 8.00 log IU/mL)Chun 2021 [28]ETV- or TDF-treated, totally sedentaryPatients with no physical activity1,437bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR2.43 per 100 patient-years (95% CI NR)HR: 1 (reference)ETV- or TDF-treated, < 500 MET-min/weekPatients with a low level of physical activity as quantified using metabolic equivalent (MET)2,425bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.84 per 100 patient-years (95% CI NR)HR: 0.83 (95% CI: 0.65, 1.06), p = 0.13 (< 500 MET-min/week vs. totally sedentary)ETV- or TDF-treated, 500–1,000 MET-min/weekPatients with a moderate level of physical activity as quantified using metabolic equivalent (MET)2,326bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.67 per 100 patient-years (95% CI NR)HR: 0.72 (95% CI: 0.56, 0.92), p = 0.009 (500–1,000 MET-min/week vs. totally sedentary)ETV- or TDF-treated, 1,000–1,499 MET-min/weekPatients with a moderate level of physical activity as quantified using metabolic equivalent (MET)988bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.56 per 100 patient-years (95% CI NR)HR: 0.66 (95% CI: 0.47, 0.92), p = 0.02 (1,000–1,499 MET-min/week vs. totally sedentary)ETV- or TDF-treated, ≥ 1,500 MET-min/weekPatients with a high level of physical activity as quantified using metabolic equivalent (MET)433bMedian: 3.1 years (IQR 2.0–4.6)aEntire follow-upNRNR1.61 per 100 patient-years (95% CI NR)HR: 0.68 (95% CI: 0.44, 1.06), p = 0.09 (≥ 1,500 MET-min/week vs. totally sedentary)Dezanet 2021 [29]TDF-/TAF-based ART-treated, low fibrosis without progressionHIV+ patients receiving TDF-/TAF-based ART with sustained no-to-mild fibrosis (F0-F1)50Median: 7.4 years (IQR 3.1–12.9)Entire follow-up00.0NRNRTDF-/TAF-based ART-treated, low fibrosis with progressionHIV+ patients receiving TDF-/TAF-based ART with initially low level of fibrosis, but a slow, linear increase towards advanced fibrosis during treatment38Median: 8.5 years (IQR 3.1–13.0)Entire follow-up00.0NRNRTDF-/TAF-based ART-treated, moderate fibrosis with high fluctuationHIV+ patients receiving TDF-/TAF-based ART with a heterogenous distribution of initial fibrosis levels (ranging from F0-F1 to F4) and equally heterogenous fluctuations in FibroTest scores during treatment67Median: 6.6 years (IQR 3.2–13.2)Entire follow-up23.0NRNRHosaka 2010 [30]LAM + ADV-treatedPatients with LAM-resistant mutations186Median: 115 weeks (range 25–282)1 yearNR2NRNR3 yearsNR11NRNR5 yearsNR16NRHR: 1.42 (95% CI: 1.04, 1.96), p = 0.03 (cirrhotic vs. non-cirrhotic)Hosaka 2013 [31]PSM ETV-treatedPatients who received ETV, PSM to the control group237bMedian: 3.3 years (IQR 2.3–4.3)a2 yearsNR0NRNR3 yearsNR0NRNR4 yearsNR0.8NRNR5 yearsNR2.5NRHR: 0.37 (95% CI: 0.15, 0.91), p = 0.030 (PSM ETV vs. PSM untreated)PSM untreatedUntreated historical control group, PSM to the ETV group231bMedian: 7.6 years (IQR 3.4–13.7)a2 yearsNR1.0NRNR3 yearsNR1.6NRNR4 yearsNR2.2NRNR5 yearsNR3.6NRHR: 1 (reference)PSM LAM-treatedPatients who received LAM, PSM to the ETV group133bMedian: 6.8 years (IQR 5.0–9.9)a2 yearsNR3.2NRNR3 yearsNR3.2NRNR4 yearsNR3.2NRNR5 yearsNR4.9NRNRHosaka 2022 [32]ETV-treatedAll included patients were HBeAg-132bMedian: 11 yearsa5 years9NRNRHR: 4.24 (95% CI: 1.81, 9.91), p = 0.001 (preexisting cirrhosis vs. no preexisting cirrhosis)Ji 2022 [33]ETV-treatedPatients who received ETV + placebo500Median: 89 monthsaEntire follow-up438.6NRNRKim 2015 [34]TDF-treatedPatients receiving open-label TDF482Median: 7.36 years (range 0.00–7.55)aCumulative incidence: annual (entire follow-up)Incidence rate: entire follow-up80.2828.4 per 10,000 person-years (95% CI NR)NRKim 2017a [35]ETV-treated virological respondersPatients with a virological response to ETV, defined as serum HBV DNA < 15 IU/mL at 1 year of therapy850Median: 5.7 years (IQR 4.8–7.0)Entire follow-up37NR0.74 per 100 patient-years (95% CI: 0.52, 1.02)NRKim 2017b [36]ETV-treated maintained virologic responsePatients with complete virologic response to ETV, defined as HBV DNA < 12 IU/mL (limit of detection), and persistently undetectable HBV DNA throughout follow-up224Median: 4.5 years (range 1.0–8.7)a5 yearsNR4.0NRHR: 1 (reference)ETV-treated low-level viremiaPatients with HBV DNA ≥ 12 IU/mL but < 2,000 IU/mL persistently or intermittently during follow-up, whether or not they previously achieved complete virologic response on ETV208Median: 4.5 years (range 1.0–8.7)a5 yearsNR6.9NRHR: 1.65 (95% CI: 0.65, 4.17), p = 0.29 (low-level viremia vs. maintained virologic response)All ETV-treated patientsAll relevant patients, including the maintained virologic response and low-level viremia cohorts432Median: 4.5 years (range 1.0–8.7)aEntire follow-upNRNRNRHR: 3.53 (95% CI: 2.11, 5.94), p < 0.001 (cirrhotic vs. non-cirrhotic)Kim 2018 [37]PSM TDF-treatedPatients were PSM to an ETV-treated group198Median: 33 months (IQR 21–46)aEntire follow-up000 per 100 person-years (95% CI NR)NRKo 2019 [38]ETV-treatedNR1,033bMedian: 6.6 years (IQR 2.8)aEntire follow-up383.7NRNR2 years7NRNRNR3 years14NRNRNR4 years24NRNRNR5 years29NRNRNR6 years32NRNRNR7 years36NRNRNR8 years36NRNRNR9 years38NRNRNR10 years38NRNRNRKobashi 2011 [39]ETV- or LAM-treatedNR194Median: 4.25 years (range 0.41–10.0)a1 yearNR1.6NRNR3 yearsNR3.5NRNR5 yearsNR3.5NRNR7 yearsNR7.1NRNR10 years1429.6NRNRKozuka 2019 [40]ETV-treatedNR105Median: 4.4 years (range 1.0–10.7)a3 yearsNR3.5NRNR5 yearsNR5.4NRNR7 yearsNR5.4NRHR: 4.789 (95% CI: 1.296, 17.689), p = 0.019 (cirrhotic vs. non-cirrhotic)Lampertico 2007 [41]ADV- and LAM-treatedAll patients had LAM-resistant HBV39Median: 42 months (range 12–74)aEntire follow-up0NRNRNRLee 2016 [42]ETV-treatedNR66Median: 45.2 months (IQR 36.0–58.3)a1 yearNR0NRNR3 yearsNR1.90NRNR5 yearsNR1.90NRHR: 9.07 (95% CI: 1.09, 75.69), p = 0.042 (cirrhotic vs. non-cirrhotic)Lee 2019 [43]ETV- or LAM-treatedPatients had a baseline METAVIR score of F372NR5 years58.9NRNRLee 2021 [44]PSM TDF combination therapyPSM patients who received TDF-containing combination therapy after switching from other NA treatments2,030Median: 45.3 months (IQR 39.0–48.6)aEntire follow-up58NR0.75 per 100 patient-years (95% CI NR)HR: 1 (reference)PSM TDF monotherapyPSM patients who received TDF monotherapy for at least 6 months after switching from other NA treatments2,030Median: 43.7 months (IQR 34.8–49.0)aEntire follow-up67NR0.87 per 100 patient-years (95% CI NR)HR: 1.21 (95% CI: 0.85, 1.71), p = 0.3 (PSM TDF monotherapy vs. PSM TDF combination therapy)Lee 2022 [45]ETV- or TDF-treatedNR787Median: 63 months (IQR 37.6–98.2)aEntire follow-up29NRNRHR: 2.974 (95% CI: 1.865, 4.743), p < 0.0001 (cirrhotic vs. non-cirrhotic)Li 2017 [46]NA-treatedPatients with < 9.7 kPa LSM were considered to be patients without cirrhosis965NR1 yearNR0.83NRNR2 yearsNR1.83NRNR3 years822.75NRNRLim 2014 [47]PSM LAM-treatedPatients treated with LAM 100 mg/day, PSM to an ETV-treated cohort (which was not extracted due to overlap with other studies included in this SLR)878Median: 8.7 years (IQR 6.5–11.5)aEntire follow-up40NR0.80 per 100 patient-years (95% CI NR)HR: 1.26 (95% CI: 0.68, 2.34), p = 0.46 (PSM ETV vs. PSM LAM)Lin 2019 [48]NA-treatedPatients treated with LAM, ETV, or LdT who were followed up for more than 2 years270Mean: 5 yearsEntire follow-up5NRNRNR2 years2NRNRHR: 4.107 (95% CI: 1.424, 11.843), p = 0.009 (cirrhotic vs. non-cirrhotic)Moon 2015 [49]NA-treatedNR330Median: 48 months (range 8–60)a1 yearNR0NRNR2 yearsNR0.30NRNR3 yearsNR0.30NRNR4 yearsNR0.30NRNR5 yearsNR2.10NRHR: 17.16 (95% CI: 4.67, 63.12), p < 0.001 (cirrhotic vs. non-cirrhotic)Nagata 2016 [50]All ETV- or LAM-treated patientsAll patients who started treatment with LAM or ETV, including several who discontinued treatment if they satisfied APASL stopping criteria94HBeAg-: mean: 108 months (range 18–182)HBeAg+: mean: 111 months (range 16–178)Entire follow-up55.3NRNRETV- or LAM-treated, HBeAg-HBeAg- patients who started treatment with LAM or ETV, including several who discontinued treatment if they satisfied APASL stopping criteria39Mean: 108 months (range 18–182)Annual over entire follow-upNR0.6NRNRETV- or LAM-treated, HBeAg+HBeAg+ patients who started treatment with LAM or ETV, including several who discontinued treatment if they satisfied APASL stopping criteria55Mean: 111 months (range 16–178)Annual over entire follow-upNR0.6NRNRNiro 2013 [51]Continuous NA therapyNR74Mean: 6 years (SD 3)aEntire follow-up11NRNROh 2020 [52]PSM ETV-treatedNR278Median: 4.7 years (IQR 4.3–5.3)aEntire follow-up3NRNRHR: 1 (reference)PSM TDF-treatedNR278Median: 4.8 years (IQR 3.9–5.4)aEntire follow-up4NRNRHR: 1.41 (95% CI: 0.32, 6.32), p = 0.651 (PSM TDF vs. PSM ETV)Orito 2015 [53]All NA-treated patientsAll relevant patients492Median: 90 months (range 12–204)Annual over entire follow-up130.34NRNRNA-treated, HBV DNA+Patients who did not lose HBV DNA positivity during follow-up188Median: 90 months (range 12–204)aEntire follow-up158.0NRNRNA-treated, HBV DNA-Patients who became HBV DNA- during follow-up230Median: 90 months (range 12–204)aEntire follow-up135.7NRNRNA-treated, HBeAg+ during therapyPatients who were HBeAg+ during follow-up99Median: 90 months (range 12–204)aEntire follow-up44.0NRNRNA-treated, HBeAg seroconversionPatients who experienced HBeAg seroconversion during follow-up39Median: 90 months (range 12–204)aEntire follow-up25.1NRNRNA-treated, HBsAg+ during therapyPatients who were HBsAg+ during follow-up479Median: 90 months (range 12–204)aEntire follow-up132.7NRNRNA-treated, HBsAg seroclearancePatients who experienced HBsAg seroclearance during follow-up13Median: 90 months (range 12–204)aEntire follow-up00.0NRNRPAGE-B Cohort [54]All NA-treated, baseline non-cirrhotic patientsPatients who did not develop HCC within the first 5 years of treatment1,023Median: 8.4 yearsaYears 5–12121.2NRHR: 4.81 (95% CI: 2.36, 9.79), p < 0.001 (cirrhotic vs. non-cirrhotic)6 yearsNRNR0.5 per 100 patient-years (95% CI NR)NR8 yearsNRNR1.2 per 100 patient-years (95% CI NR)NR10 yearsNRNR1.7 per 100 patient-years (95% CI NR)NRNA-treated, non-cirrhotic at baseline and at year 5Patients who did not develop HCC within the first 5 years of treatment and had LSM < 12 kPa at year 5870Median: 8.4 yearsaYears 5–129NRNRNR6 yearsNRNR0.5 per 100 patient-years (95% CI NR)NR8 yearsNRNR1.1 per 100 patient-years (95% CI NR)NR10 yearsNRNR1.3 per 100 patient-years (95% CI NR)NRNA-treated, non-cirrhotic at baseline, cirrhotic at year 5Patients who did not develop HCC within the first 5 years of treatment and had LSM ≥ 12 kPa at year 53Median: 8.4 yearsaYears 5–121NRNRNRPapatheodoridis 2015 [55]ETV- and/or TDF-treatedNR1,085Median: 38.8 months (IQR 23.4–49.9)a1 yearNR0.6NRNR3 yearsNR1.4NRNR5 yearsNR3.7NRHR: 4.50 (95% CI: 2.55, 7.96), p < 0.001 (compensated cirrhotic vs. non-cirrhotic)Papatheodoridis 2017 [56]ETV- and/or TDF-treatedNR1,379Median: 72 months (IQR 35)a5 yearsNRNR0.49 per 100 patient-years (95% CI NR)NRETV- and/or TDF-treatedPatients analyzed after excluding person-time accrued and HCC cases diagnosed within the first year of follow-upNRMedian: 72 months (IQR 35)a5 yearsNRNR0.51 per 100 patient-years (95% CI NR)NRPapatheodoridi 2022 [57]PSM NA discontinuationHBeAg- patients from a Caucasian cohort who discontinued NA treatment within 5 years and an Asian cohort who discontinued treatment within 3 years without developing HCC245Median: 47 months (IQR 34)a1 year70.9NRNR3 years71.9NRNR5 years75.1NRHR: 1.49 (95% CI: 0.44, 5.12), p = 0.523 (PSM NA discontinuation vs. PSM NA continuation)PSM NA continuationHBeAg- patients from a Caucasian cohort who continued NA treatment beyond 5 years and an Asian cohort who continued treatment beyond 3 years without developing HCC245Median: 40 months (IQR 32)a1 year70.4NRNR3 years72.5NRNR5 years74.9NRHR: 1 (reference)NA discontinuation, with low virological relapseNon-PSM HBeAg- patients with low virological relapse, defined as HBV DNA > 2,000 IU/mL228Median: 47 months (IQR 34)aEntire follow-up52.2NRHR: 1.64 (95% CI: 0.18, 14.78), p = 0.659 (NA discontinuation with vs. without low virological relapse)NA discontinuation, without low virological relapseNon-PSM HBeAg- patients without low virological relapse, defined as HBV DNA > 2,000 IU/mL97Median: 47 months (IQR 34)aEntire follow-up22.1NRHR: 1 (reference)NA discontinuation, with high virological relapseNon-PSM HBeAg- patients with high virological relapse, defined as HBV DNA > 20,000 IU/mL172Median: 47 months (IQR 34)aEntire follow-up42.3NRHR: 1.45 (95% CI: 0.24, 8.71), p = 0.688 (NA discontinuation with vs. without high virological relapse)NA discontinuation, without high virological relapseNon-PSM HBeAg- patients without high virological relapse, defined as HBV DNA > 20,000 IU/mL153Median: 47 months (IQR 34)aEntire follow-up32.0NRHR: 1 (reference)NA discontinuation, with HBsAg seroclearanceNon-PSM HBeAg- patients with HBsAg seroclearance, which was assessed prior to HCC diagnosis60Median: 47 months (IQR 34)aEntire follow-up00.0NRHR: 2.61 (95% CI: 0.43, 15.63), p = 0.295 (NA discontinuation with vs. without HBsAg seroclearance)NA discontinuation, without HBsAg seroclearanceNon-PSM HBeAg- patients without HBsAg seroclearance, which was assessed prior to HCC diagnosis265Median: 47 months (IQR 34)aEntire follow-up72.6NRHR: 1 (reference)Pellicelli 2014 [59]LAM-treatedAll patients were HBeAg- and had HBV genotype D57Mean: 69 months (SD 30)aEntire follow-upNRNR0.001 per 100 person-years (95% CI NR)OR: 7.9 (95% CI: 0.74, 85.00), p = 0.08 (cirrhotic vs. non-cirrhotic)NA-treatedAll patients were HBeAg- and had HBV genotype D193Mean: 69 months (SD 30)Entire follow-up21NROR: 7.6 (95% CI: 1.37, 42.20), p = 0.02 (cirrhotic vs. non-cirrhotic)10 years20.5NRNRRETRACT-B [60]NA discontinuationAll patients were HBeAg-, virally suppressed, and met their study center’s criteria for stopping NA therapy1,368Median: 18.4 years (range 7.9–39.4)aEntire follow-up100.7NRNRShim 2017 [61]ETV-treatedNR133Median: 3.6 years (IQR 2.0–5.1)aEntire follow-up2NR4.0 per 1,000 patient-years (95% CI NR)HR: 6.42 (95% CI: 1.42, 28.95), p = 0.016 (stage 1 cirrhotic vs. non-cirrhotic)HR: 24.89 (95% CI: 5.77, 107.30), p < 0.001 (stage 2 cirrhotic vs. non-cirrhotic)HR: 17.01 (95% CI: 3.86, 74.90), p < 0.001 (stage 3 cirrhotic vs. non-cirrhotic)Shin 2018 [62]ETV-treatedNR454Median: 5.4 years (IQR 3.0–7.3)aEntire follow-up71.5NRNR5 yearsNR1.5NRNR7 years72.5NRNRAnnual (7 years)NR0.4NRNRETV-treated, with good adherencePatients with ETV adherence ≥ 90%320Median: 5.4 years (IQR 3.0–7.3)aEntire follow-up51.6NRHR: 1.002 (95% CI: 0.978, 1.027), p = 0.836 (good vs. poor adherence)5 yearsNR1.7NRNR7 years51.7NRNRAnnual (7 years)NR0.2NRNRETV-treated, with poor adherencePatients with ETV adherence < 90%134Median: 5.4 years (IQR 3.0–7.3)aEntire follow-up21.5NRHR: 1 (reference)5 yearsNR1.5NRNR7 years24.9NRNRAnnual (7 years)NR0.7NRNRSou 2020 [63]ETV-treatedNR890NR5 yearsNR2.5NRNRAnnual (0 to < 5 years)NR0.63NRNR10 yearsNR3.2NRNR340NRAnnual (5–10 years)NR0.14NRNRTseng 2017 [64]All ERADICATE-B untreated patientsAll patients in the retrospective, untreated ERADICATE-B cohort2,075Median: 15.38 years (range 2.51–27.72)Entire follow-up137NR0.41 per 100 person-years (95% CI NR)NRERADICATE-B untreated patients with 15-year follow-upRetrospective, untreated ERADICATE-B cohort patients who were followed-up for 15 years1,241Median: 15.38 years (range 2.51–27.72)a15 years105NRNRNRERADICATE-B untreated inactive carriersRetrospective, untreated ERADICATE-B cohort patients who met the following criteria: HBeAg-, HBV DNA < 2,000 IU/mL, HBsAg < 1,000 IU/mL, and ALT < 40 U/L408Median: 15.38 years (range 2.51–27.72)aAnnual (15 years)NR0.06NRNRNA-treated validation cohortPatients in the validation cohort received long-term NA therapy532Median: 6.58 years (range 1.28–11.88)Entire follow-up10NR0.27 per 100 person-years (95% CI NR)NRWang 2016 [65]TDF-treatedNR141Median: 154 weeks (range 52–196)36 months00NRNRWong 2013 [66]ETV-treatedNR1,199Mean: 43 months (SD 13)3 yearsNR1.8NRNR5 years212.1NRHR: 3.2 (95% CI: 1.5, 6.4), p = 0.002 (cirrhotic vs. non-cirrhotic)Wu 2017 [67]ETV- or TDF-treatedNR296ETV-treated: mean: 49 months (SD 19.1)aTDF-treated: mean: 37.9 months (SD 7.2)aEntire follow-up3NRNRHR: 6.22 (95% CI: 1.81, 21.37), p = 0.004 (cirrhotic vs. non-cirrhotic)Yamada 2015 [68]ETV-treatedNR404Mean: 49.2 months (range 14–109)3 years162.7NRNR5 years165.3NRHR: 4.765 (95% CI: 2.415, 9.404), p < 0.001 (cirrhotic vs. non-cirrhotic)ETV-treated, with virological response at 24 weeksPatients with a virological response, defined by serum HBV DNA levels that were continuously < 2.6 log copies/mL275bMean: 49.2 months (range 14–109)a3 yearsNR3.5NRNR5 yearsNR7.2NRNRETV-treated, without virological response at 24 weeksPatients without a virological response, defined by serum HBV DNA levels that were continuously ≥ 2.6 log copies/mL129bMean: 49.2 months (range 14–109)a3 yearsNR1.0NRNR5 yearsNR1.0NRNRYang 2013 [69]ETV-treatedNR314Median: 36 months (range 12–73)a1 yearNR0NRNR2 yearsNR0.7NRNR3 yearsNR1.9NRNR4 yearsNR3.80NRHR: 7.10 (95% CI: 2.68, 18.83), p < 0.001 (cirrhotic vs. non-cirrhotic)Yang 2014 [70]LAM prophylaxisRenal transplant recipients who were HBeAg- and received LAM prophylaxis26Mean: 6.7 years (SD 2.4)5 years00NRNR10 years00NRNRLAM post-transplantRenal transplant recipients who were HBeAg- and received LAM later in the post-transplant period for HBV reactivation28Mean: 9.4 years (SD 4.7)5 years00NRNR10 years14.2NRNRYip 2020 [71]TDF-treatedNR1,271Median: 2.8 years (IQR 1.4–4.5)aEntire follow-up20.2NRNRETV-treatedNR24,219Median: 3.7 years (IQR 1.7–5.0)aEntire follow-up6462.7NRNRYuen 2007 [72]LAM-treatedAll patients were HBeAg+142Median: 89.9 months (range 26.5–128.3)Entire follow-up1NRNRNRUntreatedAll patients were HBeAg+124Median: 107.8 months (range 30.9–127.3)Entire follow-up3NRNRNRAbbreviations: ADV adefovir, ALT alanine transaminase, APASL Asian-Pacific Association for the Study of the Liver, ART antiretroviral therapy, CI confidence interval, DNA deoxyribonucleic acid, ETV entecavir, HBeAg hepatitis B e-antigen, HR hazard ratio, HBsAg hepatitis B surface antigen, HBV hepatitis B virus, HCC hepatocellular carcinoma, HIV human immunodeficiency virus, IQR interquartile range, LAM lamivudine, LdT telbivudine, LSM liver stiffness measurement, MET metabolic equivalent, N cohort size (denominator), n cases of HCC (numerator), NA nucleo(s/t)ide analog, NR not reported, OR odds ratio, PSM propensity score-matched, SD standard deviation, TAF tenofovir alafenamide, TDF tenofovir disoproxilaFollow-up durations were not reported separately for the cohort of interest and may include patients for whom outcomes were not extractedbCohort sample size calculated with information provided Summary of study outcomes Median: 4.8 years (range 1.0–13.1)a TDF: mean: 158.7 weeks (SD 10.8) ETV: mean: 159 weeks (SD 29.1) Cumulative incidence: annual (entire follow-up) Incidence rate: entire follow-up HBeAg-: mean: 108 months (range 18–182) HBeAg+: mean: 111 months (range 16–178) HR: 6.42 (95% CI: 1.42, 28.95), p = 0.016 (stage 1 cirrhotic vs. non-cirrhotic) HR: 24.89 (95% CI: 5.77, 107.30), p < 0.001 (stage 2 cirrhotic vs. non-cirrhotic) HR: 17.01 (95% CI: 3.86, 74.90), p < 0.001 (stage 3 cirrhotic vs. non-cirrhotic) ETV-treated: mean: 49 months (SD 19.1)a TDF-treated: mean: 37.9 months (SD 7.2)a Abbreviations: ADV adefovir, ALT alanine transaminase, APASL Asian-Pacific Association for the Study of the Liver, ART antiretroviral therapy, CI confidence interval, DNA deoxyribonucleic acid, ETV entecavir, HBeAg hepatitis B e-antigen, HR hazard ratio, HBsAg hepatitis B surface antigen, HBV hepatitis B virus, HCC hepatocellular carcinoma, HIV human immunodeficiency virus, IQR interquartile range, LAM lamivudine, LdT telbivudine, LSM liver stiffness measurement, MET metabolic equivalent, N cohort size (denominator), n cases of HCC (numerator), NA nucleo(s/t)ide analog, NR not reported, OR odds ratio, PSM propensity score-matched, SD standard deviation, TAF tenofovir alafenamide, TDF tenofovir disoproxil aFollow-up durations were not reported separately for the cohort of interest and may include patients for whom outcomes were not extracted bCohort sample size calculated with information provided Cumulative incidence was reported by all but two studies, seven of which reported the annual cumulative incidence ranging from 0.06% in untreated, HBeAg negative inactive carriers [64] to 0.72% in ETV-treated individuals with medication adherence rates of < 90% [62]. The cumulative incidence of HCC in both treated and untreated cohorts was also reported across specific follow-up periods. Over a one-year follow-up (reported by 10 studies), the incidence ranged from 0 to 3.3%. This range increased over three years (reported by 16 studies), reaching 0 to 11.0%, and remained similar at five years (reported by 21 studies), with values from 0 to 11.1%. By ten years (reported by six studies), the cumulative incidence expanded further, ranging from 0% to 29.6%. The 13 studies reporting incidence rates assessed a total of 38 cohorts or timepoints. The sum of person-years across studies ranged from 500.6 to 33,249.7. Reported incidence rates ranged from 0 per 100 person-years in TDF-treated individuals propensity score matched to an ETV-treated group [37], to 2.43 per 100 person-years in ETV- or TDF-treated individuals who were completely sedentary in a study examining the association between physical activity and risk of HCC in patients with CHB [28]. Overall, of the 38 cohorts or timepoints assessed, 15 of the reported incidence rates were ≤ 0.5 per 100 person-years, 10 were > 0.5 and ≤ 1 per 100 person-years, and 13 were > 1 per 100 person-years. Hazard ratios (HR) assessing the risk of HCC development were reported by 26 studies, and odds ratios (OR) were reported by one study. Among these, 18 studies compared individuals with baseline cirrhosis to those without baseline cirrhosis, while seven studies compared treatment groups, such as the type of NA treatment received, treatment versus no treatment, treatment discontinuation, and treatment adherence. Additionally, four studies compared cohorts with varying HBV DNA levels or virological responses to treatment. Surveillance and diagnostic methodologies were reported by all studies, summarized in Table S9. Hosaka 2010 reported two sets of methodologies across publications [30, 73], which are considered separately in this section. Therefore, of 51 study methodologies, 37 reported HCC surveillance frequency, 33 reported surveillance details, 21 reported diagnostic guidelines, and 35 reported diagnostic details. Generally, patients were screened for HCC approximately every six months; 14 study methodologies screened patients every six months, 12 every three to six months, and three every six to 12 months. Two study methodologies reported more frequent screening, with one cohort of Hosaka 2010 screening patients monthly [30], and Ji 2022 screening patients at least every 12 weeks [33]. Four study methodologies reported less frequent screening, with the PAGE-B cohort and Papatheodoridis 2017 reporting screening at least every 12 months [54, 56], Papatheodoridis 2015 reporting approximately annual screening for all patients [55], and Wong 2013 reporting screening every one to two years [66]. All studies that reported surveillance methodology also utilized imaging, most frequently ultrasonography and/or alpha-fetoprotein (AFP) measurement, along with other tumor markers in some cases. Diagnostic guidelines reported were AASLD (n = 11), EASL (n = 5), AASLD and EASL (n = 2), Korean Liver Cancer Study Group (n = 2), and Liver Cancer Study Group of Japan (n = 1). Three additional study methodologies reported following certain diagnostic guidelines, although the specific guidelines were not reported. HCC was diagnosed by imaging, biopsy/histology, and/or AFP measurement, as reported by 35 studies. Two additional database studies did not report diagnostic criteria but reported that International Classification of Diseases (ICD) codes were used to identify patients who developed HCC. Of the study methodologies that did report specific diagnostic criteria, nine diagnosed HCC based on imaging alone, while others utilized a combination of diagnostic techniques. Feasibility assessment and meta-analysis Across studies, substantial differences were observed in both study and patient characteristics. The feasibility assessment determined that analyzing all studies identified in the SLR to assess rates of HCC incidence across populations with average covariate values indicating earlier treatment initiation (i.e., younger age or higher baseline viral load) would be of limited robustness due to the extent of heterogeneity and aggregate nature of the data. Additionally, study populations differed notably in terms of treatment status, in that some were only treated before study follow-up. Four studies identified in the SLR (Choi 2022a [26], Yuen 2007 [72], Tseng 2017 [64], and Hosaka 2013[31]) included both treated and untreated populations and reported HCC incidence for each. These studies were therefore deemed feasible for inclusion in a direct comparison meta-analysis which allowed for evaluation of HCC risk among populations within the same study, ensuring that differences in outcomes can be attributed more reliably to antiviral treatment effects. Following the conclusion of the SLR, a targeted literature review update performed in December 2024 identified a more recent publication by Choi et al. in a cohort which overlapped with the original study selected for the meta-analysis [74]. This new publication included a larger sample size, including both HBeAg-positive and -negative individuals, whereas the originally identified study was limited to those who were HBeAg-positive. Therefore, data from the newer publication were used in place of the prior one for the meta-analysis. A summary of this study is provided in Table 4. Table 4Summary of Choi 2024 included in the meta-analysisStudyStudy DesignStudy CountrySample SizeTreatments AssessedCohort DescriptionDefinition of cirrhosis statusOutcomesChoi 2024 [74]Retrospective cohortSouth Korea4,693ETV or TDFHBeAg-positive and -negative individuals who were treatment naïve at baselineCirrhosis was defined as the presence of any of the following characteristics: coarse liver echotexture and nodular liver surface on ultrasonography; clinical features of portal hypertension (e.g., ascites, splenomegaly and varices); and platelet count < 100 × 103/µL193 patients developed HCC (0.53 per 100 person-years; 95% CI NR)CI Confidence interval, ETV Entecavir, HBeAg Hepatitis B e-antigen, HCC Hepatocellular carcinoma, NR Notreported, TDF tenofovir disoproxil Table 4Summary of Choi 2024 included in the meta-analysisStudyStudy DesignStudy CountrySample SizeTreatments AssessedCohort DescriptionDefinition of cirrhosis statusOutcomesChoi 2024 [74]Retrospective cohortSouth Korea4,693ETV or TDFHBeAg-positive and -negative individuals who were treatment naïve at baselineCirrhosis was defined as the presence of any of the following characteristics: coarse liver echotexture and nodular liver surface on ultrasonography; clinical features of portal hypertension (e.g., ascites, splenomegaly and varices); and platelet count < 100 × 103/µL193 patients developed HCC (0.53 per 100 person-years; 95% CI NR)CI Confidence interval, ETV Entecavir, HBeAg Hepatitis B e-antigen, HCC Hepatocellular carcinoma, NR Notreported, TDF tenofovir disoproxil Summary of Choi 2024 included in the meta-analysis CI Confidence interval, ETV Entecavir, HBeAg Hepatitis B e-antigen, HCC Hepatocellular carcinoma, NR Notreported, TDF tenofovir disoproxil All studies included in the direct comparison meta-analysis were retrospective or prospective cohorts conducted in Asia. Average duration of follow-up ranged from 3.3 to 15.4 years. Notable differences were observed in patient characteristics between the treated and untreated populations of Tseng 2017 and Yuen 2007, which did not use propensity score matching or similar methods. From the random effects model, an IRR of 0.59 (95% CI: 0.50, 0.69) was found, indicating that treated populations had 0.59 times the rate of HCC compared to untreated populations, shown in Panel A of Fig. 1. An IRD of -0.27 (95% CI: -0.58, 0.04) was found, indicating 27 (-4, 58) excess cases of HCC per 100 person-years of follow-up among untreated individuals attributable to treatment status (shown in Panel B of Fig. 1). Fig. 1Unadjusted comparative meta-analysis of HCC incidence in treated versus untreated populations, IRR (Panel A) and IRD (Panel B). Abbreviations: CI, confidence interval; IRD, incidence rate difference; IRR, incidence rate ratio Unadjusted comparative meta-analysis of HCC incidence in treated versus untreated populations, IRR (Panel A) and IRD (Panel B). Abbreviations: CI, confidence interval; IRD, incidence rate difference; IRR, incidence rate ratio Sensitivity analysis A Baujat plot constructed to explore sources of heterogeneity in the meta-analysis identified Hosaka 2013 as a potential outlier based on its contribution to the pooled treatment effect and to overall heterogeneity, shown in Fig. 2. The exclusion of Hosaka 2013 in a sensitivity analysis led to a similar estimate of risk reduction, with an IRR of 0.58 (95% CI: 0.49, 0.68). Compared to untreated populations, treated populations had 0.58 times the rate of HCC, shown in Panel A of Fig. 3. An IRD of -0.31 (95% CI: -0.62, 0.00) was found, indicating that among the untreated populations there were 31 excess cases of HCC per 100 person-years of follow up that could be attributed to their treatment status, shown in Panel B of Fig. 3. Despite removing Hosaka 2013, which appeared to contribute to heterogeneity in the Baujat plot, the I² values remained reasonably consistent between the primary and sensitivity analyses (IRR I²: 0% in both analyses; IRD I²: 79% before, 85% after). Fig. 2Sensitivity analysis: Baujat plot of the unadjusted comparative meta-analysis Sensitivity analysis: Baujat plot of the unadjusted comparative meta-analysis Fig. 3Sensitivity analysis of unadjusted comparative meta-analysis of HCC incidence in treated versus untreated populations, IRR (Panel A) and IRD (Panel B). Abbreviations: CI, confidence interval; IRD, incidence rate difference; IRR, incidence rate ratio Sensitivity analysis of unadjusted comparative meta-analysis of HCC incidence in treated versus untreated populations, IRR (Panel A) and IRD (Panel B). Abbreviations: CI, confidence interval; IRD, incidence rate difference; IRR, incidence rate ratio Discussion Our rigorous meta-analysis highlights the potential benefit of initiating antiviral therapy earlier in CHB disease course. Drawing on four studies suitable for direct comparison and focusing on treatment-naïve individuals as a proxy for earlier-stage CHB, we found an approximate 40% reduction in risk of HCC in treated populations as compared to untreated populations. This meta-analysis addresses a key evidence gap in the absence of long-term RCTs, providing timely insights into the benefits of early antiviral therapy in non-cirrhotic CHB while results from extended follow-up trials such as ATTENTION are awaited. In addition, it also reinforces growing evidence linking prolonged HBV DNA exposure to increased HCC risk and supports existing literature on the benefits of antiviral therapy for individuals with CHB, across the CHB spectrum [75–78]. We also demonstrate that the cumulative incidence rates of HCC increased with greater follow-up time, ranging from 0% to 3% at one year and 0% to 30% at 10 years. Of the 50 studies included in the SLR, 44 reported an average follow-up of at least three years and 23 reported at least five years. Our findings suggest a correlation between cumulative exposure to replicating HBV and increased HCC risk. These findings align with evidence from studies published after the initial SLR which would have otherwise met our eligibility criteria [74, 79–90]. The clinical importance of preventing progression to HCC is highlighted by the ongoing global burden of primary liver cancer, with national registry data from China demonstrating that liver cancer ranked fourth in incidence and second in cancer mortality in 2022 [91]. Current clinical guidelines recommend antiviral therapy for individuals with non-cirrhotic CHB without risk factors for HCC, primarily based on elevated ALT levels or significant liver fibrosis with high HBV DNA [7]. The 2024 World Health Organization (WHO) HBV guidelines expand indications for CHB treatment but still do not recommend HBV treatment specifically for HCC risk reduction [1]. Several studies have linked HBV viremia to increased HCC risk; however, few have directly compared treated versus untreated individuals to assess the impact of antiviral therapy on HCC risk. In a study by Choi et al. non-cirrhotic patients with high HBV viral load had the lowest HCC risk when treated promptly, while those with moderate viral load had the highest risk, which was not completely mitigated by long-term therapy—suggesting that delayed treatment may lead to irreversible oncogenic changes from prolonged viral replication [76]. Notably, a 4-year interim analysis in the 12-year ATTENTION trial demonstrated that TAF initiation in non-cirrhotic patients with moderate-to-high HBV viral load and normal or mildly elevated ALT concentrations resulted in reduced risk of liver-related serious adverse events, including HCC, compared to those who were untreated [75]. However, the difference was insufficient to stop the trial early. Final results are expected in 2032. Given the established role of DNA damage, mutagenesis, and cancer development [9], these findings collectively underscore the importance of addressing viral DNA integration as part of treatment strategies early in the disease course. This supports evolving treatment guidelines, such as those from EASL, which acknowledge the potential benefit of earlier antiviral therapy in non-cirrhotic populations [2]. Earlier initiation of antiviral therapy may mitigate risk, particularly for younger patients with CHB with low ALT and viral load who may lack consistent clinical follow-up and later develop active disease. A recent review [77] found a similar HCC risk reduction associated with treatment, with individuals receiving antiviral therapy having a 40–60% lower risk of HCC than untreated individuals, although this study was not limited to NAs or non-cirrhotic individuals and is therefore not directly comparable. Prior to HBV-controlling HIV regimens, HIV co-infection was associated with accelerated progression of HBV-related liver disease compared to mono-infected individuals, including HCC and cirrhosis [92]. The recent study by Lui et al. supports HCC risk reduction associated with antiviral treatment in co-infected individuals, a population often treated earlier in the HBV disease course. However, it is important to note that the participants in this study with HIV co-infection were significantly younger and had a lower percentage of cirrhosis than the HIV-uninfected group [78]. This review’s strengths include the use of a predefined protocol, searches of multiple databases and grey literature sources, and a dual-review approach to minimize bias and errors. A particular strength of the evidence base is that hepatic conditions that could be considered confounding factors (such as cirrhosis and other advanced liver conditions) were excluded; this is in contrast with existing meta-analyses which did not exclude studies reporting only on cirrhotic patients or on mixed populations [93, 94]. Additionally, modification of the JBI prevalence checklist to account for the diagnosis of both CHB and HCC conditions ensured a more rigorous evaluation of the diagnosis and reporting quality among studies, enhancing the reliability and validity of our findings. A major strength of the direct comparison meta-analysis was its approach to comparing cohorts within the same studies, allowing for more reliable attribution of HCC risk to treatment effects. Although there were too few studies to adjust for characteristics in a meta-regression, two of the studies [31, 81] employed propensity score matching techniques to balance baseline characteristics between their treated and untreated cohorts. While the inclusion of methodologically heterogeneous studies in the analysis could be considered a limitation in that it may complicate interpretation of the risk reduction associated with treatment, the fact that half of the studies employed propensity score matching is a strength in that it likely led to a greater balance of patient characteristics between the study cohorts than would otherwise be expected in incidence rates derived without propensity score adjustment. Limitations of the review include the restriction to studies with an abstract or full text in English and the limitation of the evidence base due to the strict eligibility criteria of the review. As most included studies were conducted in Asia, relevant data published in non-English languages may have been missed, potentially influencing effect estimates and limiting regional generalizability. While these criteria were intentional choices given the aim of gathering evidence specific to expanding NA usage in patients with early stages of HBV infection, they limited the number of studies for inclusion. Notably, the requirement of at least three years of NA treatment resulted in the exclusion of some studies focusing only on untreated patients with CHB, limiting the potential of these analyses to compare treated and untreated patients except in the few studies reporting on both treated and untreated cohorts. Consequently, this prevented an adjusted meta-analysis due to the insufficient number of studies. Additionally, there is potential for ascertainment bias in registry data, as treated patients may be more likely to return for follow-up and report symptoms that prompt imaging compared to untreated cohorts. A limitation of the meta-analysis is that only four studies were suitable for inclusion due to heterogeneity. Most studies did not explicitly report person-time at risk, one of two inputs to the meta-analysis of HCC incidence. For studies that did not report person-time, this value was approximated using the number-at-risk table accompanying the cumulative incidence curve for HCC and multiplying the number of patients at risk by average duration of follow-up, potentially resulting in overestimates of total person-time at risk. Heterogeneity in the direct comparison meta-analysis likely arose from differences in patient inclusion criteria, antivirals tested, or other methodological differences across the studies. Hosaka 2013 [31] and Yuen 2007 [72] both included individuals treated with LAM, an antiviral known to be characterized by high rates of drug resistance [95]. Notably, Hosaka 2013 reported a higher IRR of HCC than the other studies, potentially due to differences in patient and treatment characteristics. The exclusion of Hosaka 2013 in the sensitivity analysis did not alter the estimated risk reduction, indicating a result robust to outliers. However, the narrower confidence intervals suggest increased precision in the estimate potentially due to reduced heterogeneity in the remaining studies. A key limitation of both the SLR and meta-analysis is the relatively limited follow-up time of included studies, as it can take much longer than the several years assessed in many of the included studies for HCC to develop [96]. Additionally, most of the studies are located in Asia, which may restrict the applicability of the findings across diverse populations. Notably, aggressive HCC development has been observed in populations in Sub-Saharan Africa, a group that is largely underrepresented in this review [97]. This underrepresentation emphasized the necessity for further research in diverse regions to better understand the long-term risk and progression of HCC across different populations, particularly since differences in genotype, screening, and access to antiviral therapy that may modify treatment effects have been noted in this population [98–100]. In addition, future research should explore differences in HCC risk reduction when antiviral treatment is initiated in early versus late HBV disease stages. Utilizing longitudinal individual patient data to adjust for time since onset of HBV infection and treatment duration would robustly assess the impact of early antiviral treatment on long-term HCC risk; alternatively, it could be used to adjust for repeated measures of viral load, age, and HBeAg status as proxy variables for disease stage. Research should also focus on newer generation antivirals, such as TAF-based regimens. Given the required duration of treatment for inclusion, studies on newer agents were not captured in this SLR. Finally, future research may also investigate the potential benefit of antiviral treatment in reducing the risk of intrahepatic cholangiocarcinoma, which is also associated with CHB [101]. This meta-analysis supports the benefits of antiviral therapy in reducing HCC risk among non-cirrhotic individuals with CHB. The analysis did not directly assess HDV DNA integration; however, the observed risk reduction is consistent with broader evidence that viral suppression can lower hepatocarcinogenic risk. Combined with existing literature, these findings further support treatment expansion for non-cirrhotic individuals with CHB to reduce HBV DNA integration early in the disease course and prevent HCC development. Supplementary Information Supplementary Material 1. Supplementary Material 1.