Chronic infection with the hepatitis B (HBV) or hepatitis C virus (HCV) is a significant cause of liver disease-related morbidity and mortality in the European Union/European Economic Area (EU/EEA) [1]. Both viruses are transmitted through contact with infected blood, blood products and other bodily fluids. HBV is vaccine preventable which, along with other primary prevention measures including health care infection control and antenatal screening, have led to a decrease in acute and chronic hepatitis B (CHB) incidence in many EU/EEA countries [2]. Health care infection control together with harm reduction programmes among people who inject drugs (PWID) have also led to some decrease in the HCV incidence in many countries [3]. Many EU/EEA countries now face a dichotomy: a declining incidence of new HBV/HCV infections in the general population due to the success of primary prevention [2, 4] alongside a projected increase in liver disease-related morbidity and mortality due to ageing of the chronically infected population [5, 6]. With the availability of antiviral treatment that can effectively halt disease progression in CHB, including progression to cirrhosis and hepatocellular carcinoma, and new direct acting antivirals for chronic hepatitis C (CHC) that report cure rates in more than 90% of cases [7, 8], elimination of chronic viral hepatitis is a possibility. Elimination requires expanded access to screening, efficient linkage to care and retention in treatment among risk populations. Timely, reliable prevalence data are needed to understand which populations are most affected to better target screening and treatment programmes, and to monitor the performance and impact of these activities at a strategic level. Indeed, for screening to have a more favourable cost-effectiveness ratio and lead to an overall net gain in population health, current evidence indicates that it should be targeted to higher prevalence populations including PWID and other risk populations, where the expected case yield would be highest. [9, 10] However, the prevalence threshold above which a favourable cost effectiveness ratio varies considerably between EU/EEA countries.

In terms of key at-risk populations, men who have sex with men (MSM) are considered a high risk population for viral hepatitis due to the efficacy of sexual contact in transmitting HBV and the high prevalence of other sexually transmitted infections especially Human Immuno-Deficiency Virus (HIV). Whilst sexual contact was historically considered an ineffective route of HCV transmission, an increased HCV incidence among MSM who have not/do not inject drugs has been reported since the early 2000s. There is increasing evidence of permucosal transmission of HCV, especially among HIV positive MSM, although sexually acquired HCV infection remains rare in HIV negative, non-injecting MSM. [11, 12] Hahné et al. reported Hepatitis B surface Antigen (HBsAg) and anti-HCV (measures of evidence of chronic HBV and chronic or resolved HCV infection respectively) prevalence among MSM in the EU/EEA ranging from < 1% to 4% and from > 1% to 2.9%, respectively [13].

People detained in prison settings are considered a high risk population for blood-borne virus infection due the criminalisation of high transmission risk behaviour such as injecting drug use and sex work, coupled with pre-detention social vulnerability (such as experience of domestic abuse, poverty and homelessness) among many people detained and convicted. Prison-acquired blood-borne virus infections may also occur due to the continuation of transmission risk behaviour, the limited availability of harm reduction services and the lack of adequate infection control practices [14, 15]. Dolan et al. meta-analysed data in Global Burden of Disease regions: in Western Europe, HBsAg and anti-HCV prevalence among people in prison was reported to be 2.4% and 15.5%, respectively, while in Eastern Europe it was 10.4% for HBsAg and 20.2% for anti-HCV [16]. HBsAg and anti-HCV estimates are also available for nine and 13 EU/EEA countries, respectively, although no study quality assessment nor country-level meta-analysis/pooling were performed.

Of the three at-risk populations included in this study, PWID are considered at highest risk due to the efficacy of unsafe injecting behaviour in transmitting HBV and HCV. This together with clustering of social and environmental risk factors in this marginalised population such as a history of incarceration, poverty, homelessness and multi-morbidity compound their vulnerability [17]. Nelson et al. conducted a global review of HBsAg and anti-HCV prevalence among PWID in 2010, and reported prevalence data for 30 EU/EEA countries for anti-HCV and for 26 EU/EEA countries for HBsAg. The prevalence of anti-HCV ranged from 21.1% in Finland to 90.5% in Latvia, whereas HBsAg prevalence ranged from 0.0% in Ireland and Cyprus to 21.3% in Estonia [17]. Wiessing et al. performed a systematic review of various epidemiological measures of the HCV epidemic among PWID in Europe [18]. Although anti-HCV prevalence was not an included outcome, their findings across the cascade of care show that 72% of anti-HCV infected PWID are viraemic; that 49% are unaware of their infection; and that 9.5% of diagnosed cases are reported to be on treatment. A review focused on the EU/EEA in 2009, Hahné et al. reported HBsAg prevalence among PWID to be between 0.0% and 21.3% and anti-HCV prevalence to be between 5.3% and 90% [13]. An updated synthesis of the prevalence in this priority population is required.

Our study is part of a larger project funded by the European Centre for Disease Prevention and Control (ECDC) that seeks to provide a timely update on available estimates across a number of low risk populations (the general population, pregnant women and first-time blood donors) and as a comparator/contrast to collate prevalence estimates in high risk populations. We describe the results of the study into chronic viral hepatitis low risk populations and among migrants elsewhere [19, 20]. In the study reported here, we seek to update and expand the work of the previous ECDC systematic review (from 2009) by Hahné et al. [21] of prevalence estimates for markers of hepatitis B (HBsAg) and C (anti-HCV) in three key risk groups: MSM, people who inject drugs and people incarcerated in prison. Our study seeks to contribute to the elimination of viral hepatitis in Europe by providing information to support the design and management of primary and secondary prevention strategies.

This is the first review to collate, assess and compare prevalence estimates across these three key at-risk groups in the EU/EEA. Although gaps in evidence exist, this study reports 68 HBsAg/anti-HCV single study/pooled prevalence estimates from 23 of 31 EU/EEA countries, 42 of which are considered as intermediate/high prevalence using the WHO endemicity threshold for HBV/HCV (≥2%) [58]. This includes 20 of the 23 estimates among PWID, 20 of the 28 high quality estimates among people in prison, and four of the 17 estimates among HIV negative/unknown HIV sero-status MSM. Geographical trends are difficult to determine due to heterogeneity of, and gaps in, evidence, although the reported data here are suggestive of higher prevalence among MSM (for anti-HCV) and among PWID (for both viruses) among countries in eastern and southern Europe.

Limitations in the estimates reported for people in prison and MSM relate to geographical and population coverage, study quality and heterogeneity of the included estimates. To retrieve estimates for people in prison and for MSM, we conducted a very broad search of the published literature with no language or population restrictions, and validated retrievals directly with countries, yet found many geographical gaps in the data. Indeed, only a third of EU/EEA countries are represented among the studies that met the inclusion criteria for people in prison and only seven countries reported estimates among MSM. It is unlikely we failed to identify and include all existing high quality data, and consider it most likely that the data just do not exist or are not published. In the absence of larger, more robust studies from which prevalence can be derived, we consider the data reported here are the best available although there may have been more recent estimates published since the date of our search (March 2015). Significant heterogeneity in study design within and between risk groups hamper the statistical comparison and pooling of prevalence across countries and populations. To control for strong sources of bias in studies among people in prison when pooling data, we developed and applied a study quality assessment. The five domains were considered equally important sources of bias and it is possible that estimates included in the analysis have residual selection biases. Further, our study quality assessment did not consider sample size and there is clearly more uncertainty in the estimates derived from smaller studies.

For pragmatic reasons, we extracted prevalence estimates for PWID from the data repository coordinated by EMCDDA. With limited methodological information accompanying the EMCDDA data sets, it is possible that this data set is not exhaustive. However, EMCDDA were recently identified by another wide-ranging systematic review as the source of the most routinely collected, European-level data on the viral hepatitis prevalence among PWID [18]. We adopted an algorithmic approach favouring the most recent national level data to select estimates, and found estimates meeting this criteria for just seven MS for HBsAg and 16 MS for anti-HCV. As with the retrievals from the systematic search, there is considerable heterogeneity in study design (intervention-related and observational), sampling method (single and multi-centre sampling methods) and sample size (from < 50 to > 6000 participants) across these 23 estimates. Beyond favouring the geographical and time-frame parameters, we did not systematically assess the quality of these studies and selection biases relating to study setting, population and sample size are likely to exist.

Using 2% prevalence as the endemicity threshold set out in the 2017 WHO HBV and HCV testing guidelines, [58] by comparing risk group prevalence with the general population prevalence (from previously published reviews of the literature [59–64]) and by comparing across risk groups, our findings generally support the continued classification of PWID and people in prison as the key populations for both chronic hepatitis B and C infection. Whilst this study does not seem to support the continued classification of HIV negative/unknown HIV sero-status MSM as a high (> 2%) prevalence population for chronic hepatitis B infection, we are cautious in this conclusion given the wide confidence interval (that sometimes includes the 2% threshold) around the MSM-derived HBV estimates. We therefore still consider MSM a key population for targeted action, given that anti-HCV prevalence in MSM is higher than in the general population, the evidence of the ongoing transmission of viral hepatitis among MSM populations and the complex interaction of viral hepatitis and HIV [11, 12]. Both infection with, and antiretroviral treatment for, HIV are suspected to increase progression to chronicity as well as to accelerate fibrosis [11, 12]. Global anti-HCV prevalence among HIV positive MSM has been estimated as high as 6.4% [65], and end-stage liver disease is a leading key cause of death among co-infected HIV positive patients in some high income countries [66]. A cohort study found an unexpectedly high proportion of MSM (23% compared to the 5–10% chronicity rate expected in the general adult population [67]) develop a chronic hepatitis B infection following HBV exposure regardless of HIV status, with a younger (adult) age at infection significantly associated with an increased risk of developing CHB [68]. It would therefore seem an effective use of health resources to (continue to) offer HCV screening to HIV positive men in addition to MSM-wide HBV vaccination [69]. Specific cost-effectiveness analyses of offering individual or combined blood-borne virus screening in MSM would also greatly aid public health decision making and be a useful addition to the evidence.

Differences in the prevalence among people in prison between countries are related to the differential distribution of risk factors among the prison population together with differences in prison conditions, such as the availability of harm reduction interventions and infection control practices and infrastructure across the EU/EEA countries represented in this study [16]. The high prevalence of HBsAg in the prison population in some countries could be attributable to the incarceration of people born in intermediate or high prevalence countries and consequent over-representation of migrants in the incarcerated population. Recent estimates suggest that the proportion of the prison population that is foreign-born ranges from < 5% in Bulgaria and Hungary, to more than 15% in France, Germany, Italy, Portugal and Spain, and up to 72% in Luxembourg [70]. It is possible that the incarceration of foreign-born migrants is a driver of the high prevalence of chronic hepatitis B and C infection in prisons although as there is no systematic EU-wide data collection on the demographic profile of the incarcerated population, our understanding of the dynamics of migration, incarceration and chronic infection is limited [70].

Our study seeks to contribute to the elimination of viral hepatitis in Europe by providing information to support the design and management of primary and secondary prevention strategies. However, expected prevalence is just one of a number of factors that affects the cost-effectiveness of testing strategies. Programme-related factors such as ease of reaching the target population, uptake of screening, actual (viraemic) prevalence, linkage to care and treatment initiation also playing a key role [9, 10, 3].

We see four key public health implications emerging from our experience in this study. The first is the indication for systematic screening and linkage to care of people in prison given the high prevalence, the overlap with the PWID population, and the possible continuation of risk behaviour. Secondly, we see significant public health benefits of providing treatment as prevention, especially for CHC, among populations that share risk behaviour, in line with national and international clinical and public health guidelines [58]. Analyses have shown that treatment among high-risk dynamically interactive populations such as PWIDs, is cost-effective, especially given the shorter and more tolerable treatment regimens [8]. Thirdly, the need for diagnostic testing and treatment is particularly important for PWID where the prevalence, and therefore risk of intra-population transmission, of hepatitis C is very high. Targeted PWID screening in accessible locations as part of broader harm reduction measures may help break down barriers of stigma and among this vulnerable and high risk population [16, 71]. The criminalisation of drug use has been suggested to be as the single most important determinant of the high blood-borne virus prevalence among people in prison [16]. Finally, the lower prevalence of CHB we found among some risk populations in some countries is likely a direct result of the adoption and implementation of primary prevention measures, especially childhood immunisation, in the general population [72]. This highlights the importance of adequately resourcing primary prevention measures as well as continuing to offer HBV vaccination to risk groups to protect public health.

The limitations of this study also provide ideas for future research, specifically the improvement in the design of studies and greater geographical representation to fill the gaps in evidence. The development and consistent application of an EU/EEA or international standard for the design and quality assessment of seroprevalence studies to inform pooling and/or statistical comparison of data across studies and populations would also greatly improve understanding of prevalence across countries and populations. Finally, and probably most importantly, there is clear and urgent need for more implementation studies to determine the features of screening programmes and strategies among risk populations that effectively reach, diagnose and link to care people infected with chronic viral hepatitis.

Our study highlights the heterogeneity in prevalence across risk groups across Europe. Prevalence generally increases in an Eastern and Southern direction. There are also many countries, especially in the Eastern and Southern part of Europe, that are not represented in our results, highlighting the need to build capacity for and resource the development of robust epidemiological studies among key risk groups. Step One of elimination action planning is to know your epidemic, the ‘who’ and the ‘where’, [73] and both the evidence and the data gaps contained in this review should contribute to this strategic aim.