Our systematic reviews shows that hepatitis B vaccine, hepatitis B immunoglobulin, or the combination of vaccine plus immunoglobulin given to the newborn infants of mothers positive for hepatitis B surface antigen prevents the occurrence of hepatitis B. Furthermore, the combination of vaccine plus immunoglobulin was superior to vaccine alone. These benefits were not significantly associated with the methodological quality of the trials, the mother's hepatitis B e antigen status, time of injection, or number of infants dropping out of the study.
Our review has several potential limitations. Firstly, some analyses include few trials and a small number of newborn infants. Secondly, most trials were of low methodological quality. We did not, however, find a strong association between methodological quality and results. This supports the robustness of our results, but does not exclude the possibility of bias.61-63
Thirdly, although we did not find asymmetries in funnel plots, we cannot exclude publication bias. Fourthly, only a few investigators responded to our request for further information and often that the details were lost. Fifthly, most trials reported only surrogate outcomes (hepatitis B surface antigen status or antibody levels to hepatitis B surface antigen) and not long term clinical outcomes. One trial with long term follow-up did find more patients with chronic hepatitis in the plasma derived vaccine plus hepatitis B immunoglobulin group compared with the plasma derived vaccine group.32
Our results show that hepatitis B vaccination prevents the occurrence of hepatitis B in the newborn infants of mothers positive for hepatitis B surface antigen. We found no significant difference between recombinant vaccine and plasma derived vaccine on hepatitis B infections (relative risk 1.00, 95% CI 0.70 to 1.42). However, more infants who received recombinant vaccine achieved antibody levels to hepatitis surface antigen > 10 IU/l (1.96, 1.39 to 2.78). The advantage of recombinant vaccine might be due to the difference in chemical and physical characteristics of the antigen components of the vaccines.64
Recombinant vaccine is used in high income countries owing to the fear of acquiring human immunodeficiency virus and other infections, including transmissible spongiform encephalopathies.65
Our finding seems to support the introduction of recombinant vaccines in clinical practice.
The recommended schedules for immune prophylaxis against hepatitis B varies among countries.66,67
In general we were unable to show significant differences among different doses, different schedules, and different forms of plasma derived vaccine and recombinant vaccine on hepatitis B occurrence. Furthermore, our subgroup analyses did not show a strong association between timing of injection (within 12, 24, or 48 hours) and magnitude of effects. The number of infants evaluated in these comparisons was small. Therefore larger trials are needed before equivalence or non-inferiority can be claimed.
Our meta-analyses found that hepatitis B immunoglobulin alone or when added to hepatitis B vaccine decreased the risk of hepatitis B infection (0.52, 0.44 to 0.63). A recent non-randomised study reported no benefit of adding hepatitis B immunoglobulin to vaccine in mothers negative for hepatitis B e antigen.68
In our analysis, only one small trial out of 11 trials included infants of such mothers.55
Our subgroup analysis did not find any statistically significant difference between infants of mothers negative or positive for hepatitis B e antigen. More randomised trials seem warranted on the addition of hepatitis B immunoglobulin to vaccine for infants of mothers negative for hepatitis B e antigen. It should be noted that hepatitis B immunoglobulin, as with plasma derived vaccine, has the potential for transmission of bloodborne infections.69
Few trials reported sufficiently on adverse events. According to what was reported, hepatitis B vaccine and hepatitis B immunoglobulin seem safe. These results are in accordance with two Cochrane reviews on hepatitis B vaccination of healthcare workers and dialysis patients.70,71
Furthermore, cohort studies found that hepatitis B vaccination is well tolerated and that severe adverse events are rare.72-79
One cohort study did find, however, that hepatitis B vaccine increased the risk of chronic arthritis and acute ear infections.80
We are unable to determine if the reliability of this finding owing to the methodological weaknesses of cohort studies.66
Randomised clinical trials may overlook adverse events because of the relatively low numbers of participants or poor reporting of adverse events.81-83
Further trials ought to focus on adverse events after the International Conference on Harmonisation's guidelines for clinical trials.79
What is already known on this topic
Mother to child transmission accounts for up to 50% of hepatitis B carriers
Repeated vaccination over months is required to mount an effective antibody response
Immunoglobulin is immediately effective and seems protective for several months, after which it wanes
What this study adds
Vaccine decreased the risk of hepatitis B infection among infants of mothers positive for hepatitis B surface antigen
Immunoglobulin alone or added to vaccine decreased the risk of hepatitis B infection among infants of mothers positive for hepatitis B surface antigen
Evidence on immunisation for infants of mothers positive for hepatitis B surface antigen but negative for hepatitis B e antigen is weak
In general, the risk of perinatal transmission from mothers negative for hepatitis B e antigen is considered much lower than that from mothers who are positive for the antigen.5-9
Further, the infants of hepatitis B e antigen negative mothers often clear an asymptomatic infection.15
Our findings are mainly based on immune prophylaxis for infants of mothers positive for hepatitis B surface antigen and hepatitis B e antigen. Evidence from randomised clinical trials is insufficient to either support or refute immune prophylaxis for infants of mothers negative for hepatitis B e antigen. The applicability of our findings to mothers negative for hepatitis B e antigen, which are of high proportions in, for example, the United States and northern Europe, is therefore limited.84
Cost effectiveness studies indicate that hepatitis B vaccination for infants of mothers positive for hepatitis B surface antigen are cost effective in countries with low,85-88
intermediate, and high prevalence.89-92
We identified no cost effectiveness studies assessing the effects of adding hepatitis B immunoglobulin to vaccine. As hepatitis B immunoglobulin may reduce the risk of hepatitis B infection, the need to carry out cost effectiveness studies based on randomised trials seems justified.
Two trials that discussed a new way to potentially prevent vertical transmission of hepatitis B did not fulfil our inclusion criteria.14
The two trials randomised pregnant women positive for hepatitis B surface antigen to hepatitis B immunoglobulin versus no intervention before delivery.93,94
In the group receiving immunoglobulin, fewer infants were positive for hepatitis B surface antigen at follow-up. The methodological quality of those trials was low. Furthermore, the mothers are at risk of developing immune complex disease due to hepatitis B immunoglobulin reacting with their own circulating hepatitis B surface antigens. More trials are therefore needed before this intervention should be adopted.