Abstract
To investigate the prevalence of hepatitis B virus (HBV) infection among alcohol users.
A systematic search of articles in the PubMed, Web of Science and EMBASE databases was conducted. The methodological quality of each study was scored, and a meta-analysis was performed taking into account the heterogeneity expected among the studies. Publication bias was assessed using Begg’s and Egger’s tests.
The search identified 998 reports that yielded 18 eligible studies. The studies comprised 12,204 alcohol users, who were mostly men. The mean score on the quality evaluation was 6.9, and 11 studies were classified as having a low risk of bias. The estimated worldwide prevalence of HBV was 20.0% (95%CI: 19.0–20.0). The heterogeneity among the studies was substantial (I2 = 96.7%). In subgroup analyses, it was observed that among alcohol user dependents with no description of liver damage, alcohol users with different stages of chronic liver disease and alcohol users who all had cirrhosis, the estimated prevalence was 10.0% (95%CI: 8.0–14.0), 14.0% (95%CI: 13.0–15.0) and 32.0% (95%CI: 29.0–35.0), respectively. The meta-regression analysis showed that the study quality score had an influence on the investigated prevalence (P = 0.005). Nevertheless, the funnel plot showed asymmetry, and there was evidence of publication bias according to Egger’s test (P = 0.003) but not Begg’s test (P = 0.869).
The prevalence of HBV among alcohol users was high. HBV infection and alcohol consumption are factors affecting the development and worsening of liver disease; therefore, we suggest that adult alcohol users should be carefully monitored.
INTRODUCTION
The hepatitis B virus (HBV) is a hepatotropic virus that belongs to the family Hepadnaviridae, and its transmission can occur vertically (from mother to child), via blood to blood contact or via sexual contact. HBV can cause acute and chronic liver diseases (CLD) such as cirrhosis or even a common type of primary liver cancer, hepatocellular carcinoma (HCC) (Busch and Thimme, 2015; WHO, 2019). Chronic HBV infection is characterized by persistent hepatitis B surface antigen (HBsAg) seropositivity for ≥6 months (Schweitzer et al., 2015). Approximately 257 million people worldwide have chronic HBV infection. Moreover, this infection is considered a public health issue, and the World Health Organization (WHO) estimates that its global prevalence ranges from 0.7% in the Americas to 6.2% in the Western Pacific (WHO, 2018a); however, it is possible that many people are unaware of being infected. In 2015, Schweitzer and collaborators carried out a systematic review analyzing HBsAg seroprevalence and excluding high-risk groups, and the overall prevalence was 3.61% (Schweitzer et al., 2015). There has been a decline in HBV transmission over the last few decades due to vaccines mostly administered to infants, in addition to regional differences in coverage, but deaths still occur in adults if they are not diagnosed and treated (WHO, 2018a). Furthermore, the WHO recommends the vaccination of adult individuals at high risk of HBV infection (WHO, 2019).
Alcohol consumption is a well-known causal factor of liver disease worldwide and has become increasingly relevant (Cholankeril and Ahmed, 2018; Kim et al., 2019). Briefly, alcohol consumption induces inflammation and increases oxidative stress, possibly leading to the development of alcoholic hepatitis. The excessive alcohol consumption mainly promotes the translocation of gut bacteria and lipopolysaccharides to the portal system. Then, interactions occur with toll-like receptors, resulting in the release of immunogenic and inflammatory mediators (Jampana and Khan, 2011). Frequent and excessive alcohol consumption can lead to several histopathological alterations, such as steatosis, fibrosis, cirrhosis and HCC (Hosseini et al., 2019). There are factors that can contribute to the acceleration of these alterations, and these factors are the type of alcohol and the dose and duration of alcohol consumption. Genetic factors, iron overload, obesity and viral hepatitis infections also contribute significantly to the evolution of liver disease (O’Shea et al., 2010). More than 50.0% of the individuals in the Americas, Western Pacific and Europe consume alcohol. According to the last ‘Global status report on alcohol and health’, approximately 3.3 million deaths or 5.3% of all deaths worldwide were attributed to excessive alcohol consumption (WHO, 2018b).
A relevant aspect that was presumed by Gitto and collaborators is that the effect of alcohol consumption in combination with HBV infection enhances liver damage through mechanisms such as a weak immune response, enhanced oxidative stress, and the activation of pro-inflammatory cytokine pathways (Gitto et al., 2009). In addition, alcohol consumption may increase the chances of acquiring sexually transmitted infections (STIs) by increasing risky behavior, which could partially explain an increase in the prevalence of viral hepatitis among alcohol users compared to the general population (Cortes et al., 2017). However, according to the WHO, the contribution of alcohol consumption to the burden of infectious diseases such as HBV has not yet been sufficiently reported (WHO, 2018b).
In view of these factors that worsen liver disease, namely, alcohol consumption and viral hepatitis, the present systematic review and meta-analysis aimed to investigate the prevalence of HBV infection among alcohol users.
METHODS
Study design
A systematic search of articles reporting the prevalence of HBV infection among alcohol users was conducted in three health and biomedical databases: Medline via PubMed, Web of Science and EMBASE. This systematic review and meta-analysis were reported following the ‘Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA)’ guidelines (Moher et al., 2009). It was prospectively registered in the PROSPERO database (International Prospective Register of Systematic Reviews), registration number CRD42018106535.
Search strategy and selection criteria
The PubMed, Web of Science and EMBASE databases were searched from 1 January 2000 to 26 March 2019 without language restrictions to identify the maximum number of original research articles about HBV infection among alcohol users worldwide over those two decades. PubMed was searched with the use of Medical Subject Headings (MeSH) terms and free text words (keywords) with ‘OR’ and ‘AND’ Boolean operators as follows: (‘Hepatitis B virus’[Mesh] OR HBV OR ‘Hepatitis B’[Mesh]) AND (‘Alcoholics’[Mesh] OR ‘Alcoholism’[Mesh] OR ‘Liver Diseases, Alcoholic’[Mesh]). Web of Science was searched as follows: ((‘Hepatitis B virus’:ti,ab,kw OR ‘HBV’:ti,ab,kw) AND ‘alcoholism’:ti,ab,kw OR ‘alcohol liver disease’:ti,ab,kw) NOT ‘clinical trials’:ti,ab,kw AND [2000–2019]/py. EMBASE was searched using database-specific subject headings and keywords as follows: ((‘Hepatitis B virus’ OR ‘HBV’) AND ‘alcoholism’ OR ‘alcohol liver disease’) NOT ‘clinical trials’ AND [2000–2019]. Reference lists from the selected studies as well as from review articles were manually searched to increase the chance of identifying relevant literature (Suppl. material 1).
The inclusion criteria were as follows: (a) positivity for HBV defined by the presence of the surface antigen HBsAg, (b) alcohol users (considered as individuals with some alcohol abuse or dependence), (c) age ≥ 18 years and (d) studies providing data that facilitated the HBsAg prevalence calculation.
The exclusion criteria were: (a) studies not reporting enough data to estimate the HBsAg prevalence; (b) reviews, case series, case reports, conference abstracts or clinical trials; (c) studies reporting overlapping/updated data and (d) sample size < 50.
Study selection and data collection
Initially, duplicate publications were excluded. Then, records not relevant to the objective were excluded after the titles/abstracts were read by two independent researchers, with > 95% agreement, and a list of possibly suitable articles was generated. Disagreements regarding the identification of appropriated studies were read and debated by two other researchers until they reached a consensus. In the next step, full-text versions of the articles were evaluated for eligibility after having been read and discussed by the researchers. One article might include one or more studies (prevalence measures in different populations).
Following full-text reviews, data from the relevant studies were extracted by the researchers, of which 25% were double extracted by one investigator to ensure consistency. A specific form was designed, and the following variables were extracted: surname of the first author, publication year, country, type of population, sample size, cut-off value for alcohol consumption, positivity for HBsAg among alcohol users and the mean age and sex of the alcohol users.
Quality assessment
The methodological quality was assessed in each study included in the current systematic review and meta-analysis according to the ‘Joanna Briggs Institute Critical Appraisal Checklist for Studies Reporting Prevalence Data’ (Munn et al., 2015). In this scale, the quality is scored from 1 to 9, with 9 being the highest quality. We classified scores as having high (0–3), moderate (4–6) and low (7–9) risks of bias. The quality score was considered a variable and was analyzed in the meta-regression. This score was not considered a study selection criterion.
Quantitative analysis
The meta-analysis was performed using Stata software v. 13 (Stat Corp LP, Texas USA). The following commands were used: metan for the random effects meta-analysis (considering the heterogeneity expected among the studies); metareg for the meta-regression and metabias to evaluate the publication bias (Begg and Mazumdar, 1994; Egger et al., 1997; Higgins et al., 2008).
For the evaluation of the possible heterogeneity among the studies, the I2 statistic was used, which reports the variation among the studies (percentage) that is due to heterogeneity and not chance (Higgins et al., 2008). The I2 statistic was considered evidence of mild, moderate and high heterogeneity when the values were 25.0–50.0, 51.0–75.0% and >75%, respectively. Low heterogeneity indicates that the variability among estimates is compatible with random variation.
RESULTS
Studies included
The database search identified 998 reports: 487 through Pubmed, 149 through Web of Science and 362 through EMBASE. Of these, 92 were excluded because they were duplicates. After the screening of titles and abstract, another 783 were excluded because they did not meet the selection criteria. The full texts of 123 possibly suitable records were retrieved for screening, and 111 of these were excluded for the reasons summarized in Fig. 1. Therefore, 12 eligible articles identified in databases were included in this systematic review. In addition, 4 articles were identified by the manual search. Finally, 16 articles (12 plus 4) contributed 18 studies with HBV prevalence measures.
Process of search and selection of studies in the systematic review of the HBV prevalence among alcohol users.
General characteristics of the studies
Only studies in which the diagnosis of HBV infection was performed by HBsAg detection were included. Of these 18 studies, nine reported the diagnosis by enzyme immunoassay, three by microparticle enzyme immunoassay and one by radioimmunoassay, and the others did not contain this information. Regarding alcohol users, more than half of the studies (n = 11) described the cut-off value of the amount of alcohol consumption per day and that ranged from ≥12 grams per day (g/day) to > 80 g/day. In most studies, the prevalence of HBV among alcohol users was not the primary goal, but it was possible to extract this information. The minimum duration of alcohol consumption was mentioned in nine studies and ranged from 2 weeks to > 10 years. However, only three studies evaluated the mean duration of alcohol consumption and it ranged from 19 to 27 years.
Of the 18 studies included, 14 contained data from hospitals. Among those who mentioned the Units/Departments were: Units of Liver, Hepatology, Gastroenterology, Internal Medicine, or access to hospital records. One study included a community-based cohort, and three were from Center for Addictions or Mental Health Unit.
The studies comprised a total of 12,204 alcohol users, and sample sizes ranged from 58 to 2720 (mean = 678, median = 317). It was possible to collect the mean age of the alcohol users in only six studies, and it was 44.6 years (33.5 to 50.9 years). Table 1 shows that 11 studies described sex information, and 88.6% (72.6–100.0%) were males. In relation to the location of the individuals included in the studies, three studies included individuals from the American continent (16.7%), one from Africa (5.6%), six from Europe (33.3%) and eight from Asia (44.4%). The mean score of the quality assessment was 6.9 (median = 7, range from 5 to 8); seven studies were classified as having a moderate risk of bias and 11 as having a low risk of bias.
Characteristics of the selected studies reporting the seroprevalence of HBV infection among alcohol users
| Publication | Country | Continent | Type of population | % of male | Sample size | Positive HBsAg N (%) | Alcohol consumption cut-off (g/day) | Quality score |
|---|---|---|---|---|---|---|---|---|
| Sagnelli et al., 2018 | Italy | Europe | CLD | 75.8 | 2720 | 244 (9.0) | ≥40 M, ≥30 F | 6 |
| Mukherjee et al., 2017 | India | Asia | CLD | NA | 2429 | 477 (19.6) | NA | 6 |
| Cortes et al., 2017 | Brazil | America | Dependent | 100 | 90 | 0 (0.0) | NA | 7 |
| Tekin et al., 2015 | Turkey | Europe | Cirrhotic | NA | 300 | 49 (16.3) | ≥40 | 6 |
| Wang et al., 2014 | China | Asia | Cirrhotic | 98.9 | 859 | 400 (46.6) | ≥40 M, ≥20 F or > 80 within past 2 weeks | 5 |
| Mbaawuaga et al., 2014 | Nigeria | Africa | Various populationsa | NA | 533 | 72 (13.5) | NA | 7 |
| Gonçalves et al., 2013 | Brazil | America | Cirrhotic | 93.4 | 716 | 114 (15.9) | >80 M, > 40 F | 5 |
| Lin et al., 2013 | Taiwan | Asia | Cirrhotic | 82.9 | 334 | 132 (39.5) | >80 | 6 |
| Pazeto et al., 2012 | Brazil | America | Dependent | 92.6 | 68 | 2 (2.9) | NA | 8 |
| Stroffolini et al., 2010 | Italy | Europe | CLD | NA | 613 | 50 (8.2) | ≥12 | 8 |
| Stroffolini et al., 2010 | Italy | Europe | Cirrhotic | NA | 136 | 12 (8.8) | ≥12 | 8 |
| Fung et al., 2007 | China | Asia | CLD | 94.3 | 105 | 5 (4.8) | ≥30 M, ≥20 F | 7 |
| Addolorato et al., 2007 | Italy | Europe | Cirrhotic | 72.6 | 84 | 13 (15.5) | ≥60 M, ≥48 F at least 2 days/week | 6 |
| Krupitsky et al., 2006 | Russia | Europe | Dependent | 87.9 | 503 | 55 (11.0) | NA | 8 |
| Wang et al., 2003 | Taiwan | Asia | General population | 100 | 2429 | 472 (19.5) | ≥12 at least 4 days/week | 7 |
| Lin et al., 2002 | Taiwan | Asia | CLD | NA | 65 | 13 (20.0) | NA | 8 |
| Lin et al., 2002 | Taiwan | Asia | Gastric disordersb | NA | 58 | 4 (6.9) | NA | 8 |
| Kwon et al., 2000 | Korea | Asia | Cirrhotic | 98.8 | 162 | 47 (29.0) | ≥80 | 8 |
| Publication | Country | Continent | Type of population | % of male | Sample size | Positive HBsAg N (%) | Alcohol consumption cut-off (g/day) | Quality score |
|---|---|---|---|---|---|---|---|---|
| Sagnelli et al., 2018 | Italy | Europe | CLD | 75.8 | 2720 | 244 (9.0) | ≥40 M, ≥30 F | 6 |
| Mukherjee et al., 2017 | India | Asia | CLD | NA | 2429 | 477 (19.6) | NA | 6 |
| Cortes et al., 2017 | Brazil | America | Dependent | 100 | 90 | 0 (0.0) | NA | 7 |
| Tekin et al., 2015 | Turkey | Europe | Cirrhotic | NA | 300 | 49 (16.3) | ≥40 | 6 |
| Wang et al., 2014 | China | Asia | Cirrhotic | 98.9 | 859 | 400 (46.6) | ≥40 M, ≥20 F or > 80 within past 2 weeks | 5 |
| Mbaawuaga et al., 2014 | Nigeria | Africa | Various populationsa | NA | 533 | 72 (13.5) | NA | 7 |
| Gonçalves et al., 2013 | Brazil | America | Cirrhotic | 93.4 | 716 | 114 (15.9) | >80 M, > 40 F | 5 |
| Lin et al., 2013 | Taiwan | Asia | Cirrhotic | 82.9 | 334 | 132 (39.5) | >80 | 6 |
| Pazeto et al., 2012 | Brazil | America | Dependent | 92.6 | 68 | 2 (2.9) | NA | 8 |
| Stroffolini et al., 2010 | Italy | Europe | CLD | NA | 613 | 50 (8.2) | ≥12 | 8 |
| Stroffolini et al., 2010 | Italy | Europe | Cirrhotic | NA | 136 | 12 (8.8) | ≥12 | 8 |
| Fung et al., 2007 | China | Asia | CLD | 94.3 | 105 | 5 (4.8) | ≥30 M, ≥20 F | 7 |
| Addolorato et al., 2007 | Italy | Europe | Cirrhotic | 72.6 | 84 | 13 (15.5) | ≥60 M, ≥48 F at least 2 days/week | 6 |
| Krupitsky et al., 2006 | Russia | Europe | Dependent | 87.9 | 503 | 55 (11.0) | NA | 8 |
| Wang et al., 2003 | Taiwan | Asia | General population | 100 | 2429 | 472 (19.5) | ≥12 at least 4 days/week | 7 |
| Lin et al., 2002 | Taiwan | Asia | CLD | NA | 65 | 13 (20.0) | NA | 8 |
| Lin et al., 2002 | Taiwan | Asia | Gastric disordersb | NA | 58 | 4 (6.9) | NA | 8 |
| Kwon et al., 2000 | Korea | Asia | Cirrhotic | 98.8 | 162 | 47 (29.0) | ≥80 | 8 |
CLD: chronic liver disease; F: female; HBsAg: HBV surface antigen; M: male; NA: not available.
aVarious populations: include groups such as pregnant women, automobile accident victims, blood donors, health care workers from hospitals and sex workers.
bGastric disorders: include patients with normal liver function, and the main diagnoses were diseases of the esophagus, stomach or duodenum, gallbladder or bile ducts and pancreas, noninfectious enterocolitis and other disorders of the intestines or peritoneum.
Characteristics of the selected studies reporting the seroprevalence of HBV infection among alcohol users
| Publication | Country | Continent | Type of population | % of male | Sample size | Positive HBsAg N (%) | Alcohol consumption cut-off (g/day) | Quality score |
|---|---|---|---|---|---|---|---|---|
| Sagnelli et al., 2018 | Italy | Europe | CLD | 75.8 | 2720 | 244 (9.0) | ≥40 M, ≥30 F | 6 |
| Mukherjee et al., 2017 | India | Asia | CLD | NA | 2429 | 477 (19.6) | NA | 6 |
| Cortes et al., 2017 | Brazil | America | Dependent | 100 | 90 | 0 (0.0) | NA | 7 |
| Tekin et al., 2015 | Turkey | Europe | Cirrhotic | NA | 300 | 49 (16.3) | ≥40 | 6 |
| Wang et al., 2014 | China | Asia | Cirrhotic | 98.9 | 859 | 400 (46.6) | ≥40 M, ≥20 F or > 80 within past 2 weeks | 5 |
| Mbaawuaga et al., 2014 | Nigeria | Africa | Various populationsa | NA | 533 | 72 (13.5) | NA | 7 |
| Gonçalves et al., 2013 | Brazil | America | Cirrhotic | 93.4 | 716 | 114 (15.9) | >80 M, > 40 F | 5 |
| Lin et al., 2013 | Taiwan | Asia | Cirrhotic | 82.9 | 334 | 132 (39.5) | >80 | 6 |
| Pazeto et al., 2012 | Brazil | America | Dependent | 92.6 | 68 | 2 (2.9) | NA | 8 |
| Stroffolini et al., 2010 | Italy | Europe | CLD | NA | 613 | 50 (8.2) | ≥12 | 8 |
| Stroffolini et al., 2010 | Italy | Europe | Cirrhotic | NA | 136 | 12 (8.8) | ≥12 | 8 |
| Fung et al., 2007 | China | Asia | CLD | 94.3 | 105 | 5 (4.8) | ≥30 M, ≥20 F | 7 |
| Addolorato et al., 2007 | Italy | Europe | Cirrhotic | 72.6 | 84 | 13 (15.5) | ≥60 M, ≥48 F at least 2 days/week | 6 |
| Krupitsky et al., 2006 | Russia | Europe | Dependent | 87.9 | 503 | 55 (11.0) | NA | 8 |
| Wang et al., 2003 | Taiwan | Asia | General population | 100 | 2429 | 472 (19.5) | ≥12 at least 4 days/week | 7 |
| Lin et al., 2002 | Taiwan | Asia | CLD | NA | 65 | 13 (20.0) | NA | 8 |
| Lin et al., 2002 | Taiwan | Asia | Gastric disordersb | NA | 58 | 4 (6.9) | NA | 8 |
| Kwon et al., 2000 | Korea | Asia | Cirrhotic | 98.8 | 162 | 47 (29.0) | ≥80 | 8 |
| Publication | Country | Continent | Type of population | % of male | Sample size | Positive HBsAg N (%) | Alcohol consumption cut-off (g/day) | Quality score |
|---|---|---|---|---|---|---|---|---|
| Sagnelli et al., 2018 | Italy | Europe | CLD | 75.8 | 2720 | 244 (9.0) | ≥40 M, ≥30 F | 6 |
| Mukherjee et al., 2017 | India | Asia | CLD | NA | 2429 | 477 (19.6) | NA | 6 |
| Cortes et al., 2017 | Brazil | America | Dependent | 100 | 90 | 0 (0.0) | NA | 7 |
| Tekin et al., 2015 | Turkey | Europe | Cirrhotic | NA | 300 | 49 (16.3) | ≥40 | 6 |
| Wang et al., 2014 | China | Asia | Cirrhotic | 98.9 | 859 | 400 (46.6) | ≥40 M, ≥20 F or > 80 within past 2 weeks | 5 |
| Mbaawuaga et al., 2014 | Nigeria | Africa | Various populationsa | NA | 533 | 72 (13.5) | NA | 7 |
| Gonçalves et al., 2013 | Brazil | America | Cirrhotic | 93.4 | 716 | 114 (15.9) | >80 M, > 40 F | 5 |
| Lin et al., 2013 | Taiwan | Asia | Cirrhotic | 82.9 | 334 | 132 (39.5) | >80 | 6 |
| Pazeto et al., 2012 | Brazil | America | Dependent | 92.6 | 68 | 2 (2.9) | NA | 8 |
| Stroffolini et al., 2010 | Italy | Europe | CLD | NA | 613 | 50 (8.2) | ≥12 | 8 |
| Stroffolini et al., 2010 | Italy | Europe | Cirrhotic | NA | 136 | 12 (8.8) | ≥12 | 8 |
| Fung et al., 2007 | China | Asia | CLD | 94.3 | 105 | 5 (4.8) | ≥30 M, ≥20 F | 7 |
| Addolorato et al., 2007 | Italy | Europe | Cirrhotic | 72.6 | 84 | 13 (15.5) | ≥60 M, ≥48 F at least 2 days/week | 6 |
| Krupitsky et al., 2006 | Russia | Europe | Dependent | 87.9 | 503 | 55 (11.0) | NA | 8 |
| Wang et al., 2003 | Taiwan | Asia | General population | 100 | 2429 | 472 (19.5) | ≥12 at least 4 days/week | 7 |
| Lin et al., 2002 | Taiwan | Asia | CLD | NA | 65 | 13 (20.0) | NA | 8 |
| Lin et al., 2002 | Taiwan | Asia | Gastric disordersb | NA | 58 | 4 (6.9) | NA | 8 |
| Kwon et al., 2000 | Korea | Asia | Cirrhotic | 98.8 | 162 | 47 (29.0) | ≥80 | 8 |
CLD: chronic liver disease; F: female; HBsAg: HBV surface antigen; M: male; NA: not available.
aVarious populations: include groups such as pregnant women, automobile accident victims, blood donors, health care workers from hospitals and sex workers.
bGastric disorders: include patients with normal liver function, and the main diagnoses were diseases of the esophagus, stomach or duodenum, gallbladder or bile ducts and pancreas, noninfectious enterocolitis and other disorders of the intestines or peritoneum.
Worldwide prevalence
The meta-analyses were performed to estimate the prevalence of HBV infection among alcohol users in the 18 selected studies. The estimated worldwide prevalence of HBV was 20.0% (95% CI: 19.0–20.0) and ranged from 3.0% (95% CI: 1.0–12.0) to 47.0% (95% CI: 41.0–52.0) (Fig. 2). Based on the meta-analyses, we also evaluated possible heterogeneity among the studies, and the heterogeneity found was substantial (I2 = 96.7%).
Overall worldwide prevalence of HBV infection among alcohol users.
Subgroup analyses
The estimated prevalence of HBV was also analyzed by meta-analyses in subgroups according to the type of population. The subgroups were divided as follows: Dependent group: dependent alcohol users interned/attended at Center for Addictions or Mental Health Unit with no description of liver damage; CLD group: alcohol users with different etiologies and stages of CLD and Cirrhotic group: alcohol users who all had cirrhosis as the stage of CLD. It was observed that among Dependent group (n = 661), the seroprevalence was 10.0% (95% CI: 8.0–14.0) and ranged from 3.0 (95% CI: 1.0–12.0) to 11.0% (95% CI: 8.0–14.0). Among CLD group (n = 5932), the prevalence was 14.0% (95% CI: 13.0–15.0) and ranged from 5.0% (95% CI: 2.0–12.0) to 20.0% (95% CI: 11.0–36.0). Finally, among Cirrhotic group (n = 2591), the estimated mean prevalence of HBV was 32.0% (95% CI: 29.0–35.0) and ranged from 9.0% (95% CI: 5.0–16.0) to 47.0% (95% CI: 41.0–52.0) (Fig. 3).
Prevalence of HBV infection among alcohol users in the Dependent group (A), CLD group (B) and Cirrhotic group (C).
Meta-regression analysis
In the meta-regression analysis, it was possible to evaluate the influence of the study quality score with regard to the article objectives, the study publication year and the sample size on the investigated prevalence (HBsAg). However, only the study quality score had a statistically significant effect (P = 0.005) (Suppl. material 2). Nevertheless, in relation to the assessment of publication bias, it was possible to observe that the funnel plot showed asymmetry, and there was evidence of publication bias according to Egger’s test (P = 0.003) but not Begg’s test (P = 0.869) (Suppl. material 3).
When stratifying studies with different risks in relation to the quality assessment, there was a slight difference in the estimated prevalence of HBV in the meta-analysis. In the studies with quality scores of 5–6, the prevalence was 21.0% (95% CI: 20.0–23.0) and ranged from 9.0% (95% CI: 8.0–10.0) to 47.0% (95% CI: 41.0–52.0), and in the studies with quality scores of 7–8, the prevalence was 17.0% (95% CI: 15.0–18.0) and ranged from 3.0% (95% CI: 1.0–12.0) to 29.0% (95% CI: 21.0–40.0).
Additional studies
The other selected studies among alcohol users (three) were from the general population, gastric disorder patients and various populations that included groups such as pregnant women, automobile accident victims, blood donors, health care workers from hospitals and sex workers. The prevalence of HBV found in the general population was 19.5%; in gastric disorder patients, it was 6.9% and in various populations, it was 13.5% (Table 1).
DISCUSSION
The total HBV infection prevalence among alcohol users was 20.0%, which is higher than that described in the general population, which is between 0.7 and 6.2% worldwide (WHO, 2018a). In addition, a recent systematic review described low HBsAg prevalence in the European Union. There were data available from 13 countries, and the prevalence was estimated to be 0.9% (95% CI: 0.7–1.2), ranging from 0.1 to 4.4% in the general population, corresponding to approximately 4.7 million cases (Hofstraat et al., 2017). A possible explanation for the prevalence found in the present meta-analysis is that, in addition to alcohol consumption, more than half of the studies included patients with CLD or cirrhosis. A study conducted in India found that HBV was the main cause of CLD without cirrhosis, which was found in 40.8% of 8163 cases. Regarding cirrhotic patients, alcohol emerged as the predominant cause, which was present in 34.3% of 4413 cases (Mukherjee et al., 2017). Another study investigated the etiologies of liver cirrhosis in 1516 patients, and alcohol alone was the main cause, and it was present in 39.7% of the patients (Gonçalves et al., 2013). It is also relevant to highlight that alcohol use disorders are becoming more common across all sociodemographic groups, and alcoholic liver disease (ALD) is already replacing HCV infection as the leading indication for liver transplantation (Cholankeril and Ahmed, 2018).
The prevalence of HBV among alcohol users was stratified by subgroup according to the type of population. In alcohol users, the lowest prevalence (10.0%) was found; however, the prevalence was higher than that in the general population. Alcohol consumption could leave some people more vulnerable to exposure to situations resulting in an increased risk for HBV acquisition, such as unsafe sex or drug use. Rehm and collaborators presented data showing that alcohol consumption is an independent risk factor for unsafe sex, and the more alcohol consumed, the greater the amount by which the intention to use condoms decreases. Subsequently, this plays an important role in the acquisition of STIs (Rehm et al., 2012), such as HBV. In young adults, alcohol consumed in conjunction with the use of marijuana increases the chance of unsafe sex more than when both factors are evaluated separately (Hayaki et al., 2018). In addition to this possible risky sexual behavior linked to drug use, exposure to HBV may also occur when people inject drugs and share needles. Alcohol consumption and drug use can eventually increase an individual’s susceptibility to viral infections such as HBV and can have an impact at the onset of the infection; likewise, HBV could affect recovery in individuals who are alcohol and drug dependents (Pöschl and Seitz, 2004; Krupitsky et al., 2006). It is important to emphasize that the prevalence of drug use disorders in adults varies widely among countries, and the highest prevalences are found in the United States, Australia and New Zealand (Degenhardt et al., 2019).
It seems logical that those who already have established CLD have a higher prevalence of HBV infection (14.0%) and that the prevalence is even higher in individuals who have cirrhosis (32.0%). It is likely that an individual who has had cirrhosis throughout his life has already been exposed to more risk factors for liver damage, such as contact with HBV, alcohol consumption and exposure to HCV. We did not obtain sufficient data to allow us to assess the prevalence of HCV coinfection in alcohol users infected with HBV. It has been described that all these factors could interact and lead to molecular alterations in the liver, resulting in synergistic effects such as hepatocyte apoptosis, oxidative stress, miR-122 and immune dysregulation, all of which would promote the progression of liver disease and unfavorable outcomes (Punzalan et al., 2015).
A recent population-based study in the United States (25,379,768 deaths among adults ≥20 years) presented data that the mortality due to cirrhosis caused by ALD had an annual percentage increase of 4.5% (95% CI: 3.8–5.3) in the period from 2007 to 2016 (Kim et al., 2019). A study conducted by Lee and collaborators defined alcohol consumption as ≥40 g/day for ≥5 years, and showed that this consumption in patients with HBV infection from Taiwan increased the risk of mortality related to HCC in 30% compared to HBV infection without alcohol consumption (Lee et al., 2013). In addition, another study described that the cumulative incidence of HCC (in 10 years) was increased in patients with cirrhosis who were infected with HBV and consumed alcohol (52.8%) compared with those who only consumed alcohol (25.6%) or HBV infection (39.8%) (P < 0.001) (Lin et al., 2013). Excessive alcohol consumption increases the risk for HCC even in individuals already under nucleoside or nucleotide therapy for HBV who have low viral load (Iida-Ueno et al., 2017). Tekin and collaborators even recommend that alcohol users with cirrhosis who are infected with HBV should be screened for HCC (Tekin et al., 2015). These data show excessive alcohol consumption to be a cause of the development of liver disease and the worsening of liver damage that is increasing in importance, especially when associated with hepatitis B infection.
CONCLUSIONS
The prevalence of HBV (HBsAg) found in alcohol users was high, and this should be a public health warning. Because HBV infection and alcohol consumption are known to be factors affecting the development and worsening of liver disease, we suggest that adult alcohol users should be carefully monitored with serological screening. In addition, preventive interventions could be carried out as educational measures.
Conflict of interest statement
The authors declare that they have no conflict of interest.
