Selection of studies
We identified 3,215 titles in electronic databases. 165 additional titles were found by manual search. After removal of duplicates, abstract and full-text screening, 41 publications were identified which contained relevant data on HBV VC in Germany. After removal of two non-informative conference abstracts, 39 publications were included in the systematic review (2, 16-52). Additionally, 29 publications matching the inclusion criteria were added from the scoping review (53-81), giving a total of 68 publications in the final review. For details see figure 1.
Figure 1: Study flow
Study characteristics
Sixty-five of the included publications were articles in scientific journals, two were dissertations and one was a conference abstract.
The included publications reported on data from 67 different studies. The nationwide population-based survey DEGS1 assessed VC using different study tools and the results were therefore published in two separate publications (2, 41).
All 67 included studies had a cross-sectional design, while six of them additionally contained a cohort element (28, 35, 47, 70, 75, 82). However, longitudinal data was not considered for this review.
In 27 studies, data was collected up to 2010. In three studies the data collection time frame was not reported, these studies were published in 2012 (57), 2018 (47) and 2019 (29).
Twenty-nine studies contained nationwide data. For three of these studies (reported in four publications), the sample was drawn from the German residential register (2, 39-41), for two from other national registers (38, 48) and for two from the health insurance refund claims of Associations of Statutory Health Insurance Physicians (ASHIP) (16, 44).
In 29 studies, vaccination status was taken from vaccination cards or from medical records, were vaccination was documented. In 13 studies, blood samples were tested for isolated positive anti-HBs as an indicator of vaccination or medical records were screened for serological results. In 20 studies, the vaccination status was self-reported and in two studies medical records were checked without clearly reporting for which parameter (53, 63). In two studies, health insurance data was analysed (16, 44). For DEGS1, participants presented vaccination cards and anti-HBs was investigated in a sub-group.
In 35 studies the coverage of the complete VC schedule (called complete VC hereafter) was reported. In 12 of them, coverage of incomplete vaccination schedule (called incomplete VC hereafter). For the remaining 32 studies the reported VC was assessed as “not specified”. For 11 studies VC of current protection was calculated.
In two studies, only VC among the susceptible study population was available (26, 57). In two further studies VC for susceptible part study population was calculated separately (47, 49).
The main study characteristics of the included publications are shown in additional file 3.
Hepatitis B vaccination coverage by population group
In seven studies, HBV VC was investigated in different study populations, either as a single entity or as a sub-population (e.g. HIV-positive MSM) (18, 19, 28, 37, 49, 60, 62). Where applicable, these studies were therefore allocated to more than one population group below (28, 37, 49).
Twenty-four studies reported data on HBV VC among children and adolescents. Among these, thirteen studies contained data from the yearly mandatory national primary school entry medical and developmental check-up, showing a complete HBV VC between 86% and 90.5% (2005 and 2017) (36, 43, 45, 46, 64-69, 71-73). Two studies contained data from the national population-based health survey of children and adolescents (German Health Interview and Examination Survey for Children and Adolescents, KiGGS and KiGGS Wave 2), reporting a complete VC among three to 17 year old children of 65.8% (2003-2006) and 84.4% (2014-2017) (39, 40), respectively.
Three studies estimated HBV VC in the adult general population. Using random samples of individuals aged 18 to 79 years drawn between 2008 and 2011, the nationwide health survey DEGS1 reported an unspecified VC of 32.9% (by check of vaccination cards and self-report) versus 22.9% (by serological marker) (2, 41). A study of patients from an emergency room in Berlin performed in 2010/2011 showed an unspecified VC of 25.2% (self-reported) (56) and in another study complete VC of patients visiting occupational physicians was 52.1%.
Regarding travel-related indication for HBV vaccination, one study was identified reporting an unspecified VC of 89% in a web-based survey (59).
Thirteen studies were related to immunocompromised populations including three in PLWH (two of them only including MSM) (28, 49, 70), seven in patients with autoimmune diseases (21, 30, 33, 34, 50, 52, 58), two in patients with liver cirrhosis resp. after liver transplantation (26, 51) and one in alcohol-dependant people during detoxication therapy (75). The unspecified VC in PLWH was between 47.1% and 47.7% in HIV-positive MSM and 11.5% in HIV-positive patients non-restricted to MSM. In other immunosuppressed populations, the complete/not specified VC was between 11.8% and 89.0% For three of 13 studies relating to immunosuppressed populations, the VC with current protection was reported (33, 34, 51).
Sixteen studies investigated VC among populations at occupational risk of HBV exposure. Six studies looked at students of different health occupations (e.g. medical students, nursing students) and reported a complete/unspecified VC between 63.9% and 93.1% (54, 61, 74, 78-80). In nine studies the HBV VC was measured among hospital personnel including medical doctors, nurses, paramedics and other medical staff resulting in a complete/not specified VC between 63.6% and 96.5% (29, 38, 42, 53, 55, 76, 77, 81, 83). One study estimated the not specified VC among educational personnel in schools for handicapped persons at 80.1% (22). For eight studies, VC with current protection was reported (54, 61, 76-81). In one of them, a study among health care students, current protection was higher than 80% (94%) (61).
Sixteen studies considered people at increased risk of non-occupational HBV exposure. Two studies were conducted among household contacts of PLWVH reporting a not specified VC of 54.0% and 55.2% in family and partners (57, 60) and 61.7% in children and siblings of PLWVH (60). For one study, the VC with current protection was reported (57) (17% of family members with known anti-HBs titre >10 IU/L).
Seven studies reported outcomes among individuals who were part of a migrant population from HBV endemic countries. One of these studies was a sub analysis of data from the school entry examination in Bavaria (37). Two studies measured a not specified VC of 4.4% resp. 14.9% in unaccompanied refugee minors (UAM) (31, 35) and three a not specified VC between 9.1% and 18.6% in adult refugees (23, 27, 32). For patients with a direct or indirect migration background of a general medicine practice the not specified VC was 17.4% (25). The complete VC in pre-school children with direct/indirect migration background was 84.5% (37).
Three studies were related to MSM. One of them reported on an internet-based survey showing complete VC of 52.3% (20). The two other studies were conducted among HIV-positive MSM (see above) with a VC between 47.1% and 47.7% (28, 49).
In four studies data was collected among PWID. The not specified VC was between 10.5% and 52.5% in drug consumption facilities and in consumption places on the street (24, 47, 62), and between 19.0% and 49.0% in opioid substitution treatment centres (62, 63).
For other populations at non-occupational risk for HBV exposure (e.g. prisoners, sex workers), no studies were identified.
The VC by population group including main study characteristics is illustrated in figures 2-4.
Figure 2: vaccination coverage in Germany among the general population and travellers to HBV endemic countries, %, 2005-2019, (*publication allocated to two population groups, $ HBV VC only in HBV susceptible people, $$ HBV VC also in HBV susceptible people available, +current: additional information on % of currently protected available)
Figure 3: vaccination coverage in Germany among populations at risk for severe HBV infection due to immunosuppression, %, 2005-2019, (*publication allocated to two population groups, $ HBV VC only in HBV susceptible people, $$ HBV VC also in HBV susceptible people available, +current: additional information on % of currently protected available)
Figure 4: vaccination coverage in Germany among populations at risk for occupational HBV exposure, %, 2005-2019, (*publication allocated to two population groups, $ HBV VC only in HBV susceptible people, $$ HBV VC also in HBV susceptible people available, +current: additional information on % of currently protected available)
Figure 5: vaccination coverage in Germany among populations at risk for non-occupational HBV exposure, %, 2005-2019, (*publication allocated to two population groups, $ HBV VC only in HBV susceptible people, $$ HBV VC also in HBV susceptible people available, +current: additional information on % of currently protected available)
Risk of Bias
Risk of bias was low in 51 of 68 publications. In the remaining 17 publications risk of bias was assessed to be high. All publications among children/adolescents and adults in the GP were at low risk of bias. Six of 16 publications on populations at occupational risk for HBV were at high risk of bias, whereas the proportion was 8/16 for publications on populations at non-occupational risk for HBV. Important weaknesses of publications at high risk of bias were the use of self-reporting (n=7), serology (n=8) or medical records with unclear parameters (n=2) as study instruments, and a missing proper case definition for complete/incomplete HBV immunisation (n=12) as well as weak methods used to represent the target population (no national representativeness n=12, inadequate sampling frame n=13, inadequate sampling n=8). For details see additional file 4.