| Greenup AJ (2014) | ☆ At least somewhat representative of the average HBV-infected pregnant woman | ☆ Drawn from the same community (same inclusion and exclusion criteria also) | ☆ Reporting on adherence within the paper, reduction of viral load used to assess women’s response to treatment | ☆ Always the case | ☆☆ Comparable for HBV DNA level and comparable HBeAg positive. Same regimen for infant immunoprophylaxis and confirmation that all infants received it | No details given on laboratory methods for infants, and no details of which assay was used for testing HBsAg | ☆ Yes | > 20% LFU in control group, although <20% LFU in two treatment groups | 7 (low) |
| Zhang H (2014) | ☆ At least somewhat representative of the average HBV-infected pregnant woman | ☆ Drawn from the same community (same inclusion and exclusion criteria also) | ☆ Monthly HBV DNA level testing was done to check maternal adherence | ☆ Always the case | ☆☆ Comparable for HBV DNA level and comparable HBeAg positive. Same regimen for infant immunoprophylaxis | ☆ Describes testing done and refers to a central laboratory employed for this study | ☆ Yes | ☆ LFU reported and <20% LFU in all treatment and control groups | 9 (low) |
| Jackson V (2015) | ☆ At least somewhat representative of the average HBV-infected pregnant woman | ☆ Drawn from the same community (same inclusion and exclusion criteria also) | ☆ Mentions good treatment compliance in all but one patient, and measures decrease in viral load in 35/36 women taking treatment just prior to delivery and saw a significant decrease in most patients (also show these results in a figure in the paper) | ☆ Always the case | HBV DNA level and HBeAg not described in control group. Mentions that all infants received the same regimen for infant immunoprophylaxis; however, in the control group, many women defaulted from care/moved to other maternities, so this does not seem well verified | ☆ Laboratory assays described, with indication of record linkage (results viewed retrospectively in medical records) | ☆ Yes | <80% retention in both treated and control groups | 6 (high) |
| Liu CP (2015) | ☆ At least somewhat representative of the average HBV-infected pregnant woman | Many more women included in the control group (highly disproportionate, which could indicate non-similarity with the treated) | Some limited data presented on decrease of maternal viral load, but no mention of linking this with compliance/adherence/time on treatment, and no detailed results provided | ☆ Always the case | ☆ HBV DNA level and/or HBeAg not described for both treated and control groups. Similar infant prophylaxis between treated and control groups | ☆ Laboratory assays described, with indication of record linkage (results viewed retrospectively in medical records) | ☆ Yes | No loss to follow up described because it was a retrospective cohort study (or listed as such) where the infants needed to have had test results at the testing time-point (this is therefore misclassified as a cohort study, and has a high risk of bias for loss to follow up) | 5 (high) |
| Pan CQ (2017) | ☆ At least somewhat representative of the average HBV-infected pregnant woman | Same population and criteria, however, no indication of how this group was chosen (usually says “unwillingness”, for example) | Some data presented on decrease of maternal viral load, but no mention of linking this with compliance/adherence/time on treatment. Additionally, because of study design (retrospective) there is low/no chance of adherence monitoring | ☆ Always the case | ☆☆ Comparable for HBV DNA level and comparable HBeAg positive. Same regimen for infant immunoprophylaxis | ☆ Reference to the hospital’s centralized laboratory and linkage to medical records for assessing infant outcome | ☆ Yes | No loss to follow up described because it was a retrospective cohort study (or listed as such) where the infants needed to have had test results at the testing time-point (this is therefore misclassified as a cohort study, and has a high risk of bias for loss to follow up) | 6 (high) |
| He T (2018) | ☆ At least somewhat representative of the average HBV infected pregnant woman | ☆ Drawn from the same community with same inclusion and exclusion criteria | ☆ Detailed information on reduction of viral load given, including specific data for each woman (every one had a −6 to −8 log reduction) | ☆ Always the case | ☆☆ Comparable for HBV DNA level and comparable HBeAg positive. Same regimen for infant immunoprophylaxis | ☆ Linkage to medical records | ☆ Yes | Retrospective cohort mentioned but no loss to follow up described, no mention of how there was perfect retention | 8 (low) |
| Wakano Y (2018) | Not representative of the general population (women who have had a child infected previously) | ☆ Drawn from the same community with same inclusion and exclusion criteria | ☆ >2 log reduction of HBV DNA levels in all treated women | ☆ Always the case | ☆ Comparable for HBV DNA level and comparable HBeAg positive. Different immunoprophylaxis regimens mixed among the groups of treated and non-treated | Laboratory assays not well described | ☆ Yes | ☆ 100% retention | 6 (high) |
| Foaud HM (2019) | ☆ Truly representative of the average HBV-infected pregnant woman | Control group comprised women who were not candidates for lamivudine (likely to be quite different from those who received it) | ☆ States that women were given lamivudine monthly and were questioned regarding compliance at each visit | ☆ Always the case | ☆ HBeAg proportion not comparable, and HBV DNA at baseline not given. Same regimen for infant immunoprophylaxis | ☆ Lab testing done centrally as part of the study, laboratory assays for defining infant outcome described | ☆ Yes | <80% follow up at 6–12 months in control group, though ~86% follow up in treated group at that time-point. (Note: at later time-point that study defined, there was >80% follow up) | 6 (high) |
