1 4 1 3 4 8 9 16 Methods Literature search and study design Two independent researchers conducted the literature search, study selection and data extraction, with any disagreements resolved by consensus among them. The researchers conducted a systemic literature search using the electronic databases MEDLINE (1966 to January 2006), EMBASE (1980 to June 2006), OVID (1966 to January, week 3, 2006) and the Cochrane library clinical trials registry (issue 1, 2007). The following keywords were used: ‘Hepatitis B’, ‘Interferon’, ‘Lamivudine’ and ‘combination therapy’. In addition, a manual search using citations in previous publications was preformed. The following inclusion criteria were used: (i) study design: randomized controlled trials; (ii) study population: HBeAg-positive patients; (iii) intervention: interferon vs. interferon and lamivudine therapy. Our search was not restricted by language. The following exclusion criteria were used: (i) examining the nonadult population; (ii) not reporting any of the primary efficacy measures as defined by the authors. When several publications pertaining to a single study were identified, the most recent and complete publication was used. The included studies were divided into two groups according to their use of conventional (CON) or pegylated (PEG) interferon-α, with patients within each group given interferon monotherapy, or interferon and lamivudine combination therapy. Data were extracted for study methodology and for the defined efficacy measures. Only data pertaining to the regimens in question were extracted, while data concerning other regimens were reviewed, and if found to be of significance to our study, were noted and discussed. Separate meta-analyses examining the defined efficacy measures were preformed. In addition, we compared the rates of sustained responses across groups aiming at the identification of a preferable regimen. Intention to treat analysis was used throughout this study, excluding histological response analysis, because of its low outcomes reporting rates. Efficacy measures and definitions 2 Study quality and homogeneity 15 Statistical analysis comprehensive meta analysis Q P Results Study selection and characteristics n n n 16 21 n 22 23 16 19 17 18 20 23 n 16 23 n 17 18 n 17 19 23 n 16 18 16 19 22 23 20 21 18 16 17 19 21 23 20 Tables 1 2 Table 2 Patient selection criteria of studies included in the meta-analysis Study Inclusion criteria Exclusion criteria Ayaz (2006) 1. HBsAg positive for >6 m and anti-HBeAg and HBsAg negative 2. Presence of HBV DNA 3. Evidence of inflammation on biopsy within 6 m of enrolment and ALT>1.5 NL 1. Previous treatment with INF, antiviral or immunosuppressive agents 2. HIV, hepatitis C or D 3. Other aetiologies of liver disease, alcohol intake >40 g/day, decompensated liver disease or cancer 4. No informed consent 5. Pregnancy 6. Any contraindications for INF use 7. Leucocytes, neutrophil or platelet count of <2500, <1000 and <100 000/mL, respectively, or haemoglobin <10 g/dL Song (2004) 1. 19–65 years old 2. HBsAg positive for >6 m and HBeAg positive 3. HBV DNA>500 000 copies/mL 4. Evidence of inflammation by 2 NL <ALT <500 Not reported Deng (2003) 1. 15–60 years old 2. HBeAg and HBV DNA positive for >6 m 3. HBV DNA>103 000 copies/mL 4. Evidence of inflammation by ALT>2 NL 1. Immunosuppressive or antiviral therapy within 6 m 2. Hepatitis of other aetiologies 3. Decompensated liver disease 4. Pregnancy or breast feeding Yalcin (2002) 1. 16–80 years old 2. HBeAg and HBsAg positive 3. HBV DNA positive 4. Evidence of inflammation by histology and 1.5<ALT<10 NL, on three occasions within 6 m 3 3 3 Cindoruk (2002) 1. Adults 2. HBeAg positive 3. HBV DNA positive 4. Evidence of inflammation by histology and by abnormal ALT levels for >6 m 1. Previous INF therapy 2. HIV, hepatitis C or D 3. Decompensated liver disease 4. Diabetes, autoimmune, or other psychiatric or serious medical illness 5. High alcohol intake or current drug abuse 6. Pregnancy Schalm (2000) 1. 16–70 years old 2. HBsAg and HBeAg positive at screening and at >6 and >3 m prior respectively 3. HBV DNA>500 000 copies/mL 4. Evidence of inflammation by histology or persistently elevated ALT for >3 m 1. Contraindication to or previous INF therapy, or antiviral therapy within 6 m 2. HIV, hepatitis C or D 3. Decompensated liver disease 4. Liver disease of other aetiology Lau (2005) 1. Adults 2. HBsAg positive for >6 m and HBeAg positive 3. HBV DNA>500 000 copies/mL 4. Evidence of inflammation on biopsy and 1<ALT<10 NL 3 Janssen (2005) 1. >16 years old 2. HBsAg positive for >6 m and HBeAg positive on two occasions within 8 w of randomization 3. Evidence of inflammation by two measurements of ALT>2 NL within 8 w of randomization 3 3 3 ALT, alanine aminotransferase; HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; INF, interferon; m, months; NEG, HBeAg negative; NL, upper limit of normal; POS, HBeAg positive; w, weeks; y, years. Table 1 Characteristics of studies included in the meta-analysis Study n Study design Jadad score Therapy period Follow-up period Therapy regimen Conventional interferon-α  Ayaz (2006)  68 RCT 2 12 m  6 m INF-α-2a 9 MU × 3/w with or without LMV 100 mg/day  Song (2004)  90 RCT 2 12 m  6 m INF-α 3 MU × 3/w with or without LMV 100 mg/day  Deng (2003)  62 RCT 2 48 w 24 w INF-α-1b 5 MU × 3/w with or without LMV 100 mg/day  Yalcin (2002)  49 RCT 2 52 w 52 w INF-α-2b 10 MU × 3/w with or without LMV 100 mg/day  Cindoruk (2002) 100 RCT 2 6 m  6 m INF-α 9 MU × 3/w with or without LMV 100 mg/day  Schalm (2000) 144 RCT, DB 4 24 w 40 w INF-α 10 MU × 3/w with or without LMV 100 mg/day Pegylated interferon-α  Lau (2005) 542 RCT, DB 5 48 w 24 w PegINF-α-2a 180 μg × 1/w with or without LMV 100 mg/day  Janssen (2005) 266 RCT, DB 4 52 w 26 w PegINF-α-2b 100 μg × 1/w with or without LMV 100 mg/day DB, double blind; INF, conventional interferon; LMV, lamivudine; m, months; PegINF, pegylated interferon; RCT, randomized controlled; w, weeks. Sustained virological response P P P Q P Fig. 1 Fig. 1 *  Sustained biochemical response P P P Q P Fig. 2 Fig. 2 Sustained biochemical response. CON, conventional interferon monotherapy vs. its combination with lamivudine; PEG, pegylated interferon monotherapy vs. its combination with lamivudine.  Sustained hepatitis B e antigen clearance P P P Q P Fig. 3 Fig. 3 *  Sustained seroconversion P P P Q P Fig. 4 Fig. 4 *  Histological response n P n P n P n P n P n P All-cause and liver-related mortality There were no reported deaths of any aetiology for either group (0% for both). Safety P n P n P Q P YMDD mutation emergence P n P P n Q P Conventional interferon combination therapy vs. pegylated interferon monotherapy P P P P P Discussion 2 2 P 24 P P 20 P 25 P P 26 P P P 16 17 19 P 27 1 3 28 29 27 30 31 22 30 Our study contains several limitations. Firstly, our use of intention to treat analysis, the methodological heterogeneity of the included studies, and the heterogeneity of their treatment and follow-up protocols, may have introduced some inaccuracies in our analysis. Notably, while the PEG group comprised large, carefully planned, well-executed studies, the CON group involved smaller, lower quality ones, thus weakening our conclusions in its regard. Secondly, the absence of adequate controls precluded the authors from studying the subsets of treatment-naïve and previously treated populations. Those concerns, however, were alleviated by the low patient lost for follow-up rates, the lack of statistically significant heterogeneity across studies, the beneficial effects of the combination with conventional interferon across the measured indicators and the agreement between our conclusions and those of other studies. 32 33 34 35 36 40 Conclusion Pegylated interferon-α monotherapy is the treatment of choice for HBeAg-positive chronic hepatitis B, with no added benefit with lamivudine addition. However, when conventional interferon therapy is considered, its combination with lamivudine should be entertained.