Background
Viral hepatitis is an emerging global health problem. In 2015, an estimated 1.34 million deaths occurred due to viral hepatitis globally. This number is equal to deaths caused by tuberculosis and higher than those deaths caused by human immunodeficiency virus. In the same year, hepatitis B and C viruses (HBV&HCV) alone were responsible for 96% of hepatitis mortality [
1]. Untreated hepatitis B and C viral infections can lead to life treating long-term complications such as liver cirrhosis and cancer [
2].
Women of childbearing age can potentially transmit HBV to their babies. They transmit an infection to newborn usually during birth or soon after birth following close contact. Newborns who exposed to HBV will have almost 85–90% risk of developing chronic liver diseases [
3].
In Ethiopia, the rate of HBV transmission from infected mother to the newborn is not well studied. However, one study revealed that 75% of newborns born from HBV infected women were positive with hepatitis B surface antigen(HBsAg) in 2012 [
4].
Hepatitis is becoming an emerging public health concern in Ethiopia. Recent systematic review of all types of viral hepatitis in Ethiopia concluded that the prevalence of HBV among the population is 7.4% [
5]. Several HBV epidemiological studies among pregnant women are available in Ethiopia [
6‐
14]. However, the results of these studies showed a wide variation of prevalence ranging from 2.3 to 7.8% [
5,
7,
8,
10‐
14] over time and across geographical areas. Despite the availability of results from each study, there are no nationwide data on the prevalence of HBV infection among pregnant women in Ethiopia which clearly shows the presence of research gaps.
Furthermore, a well organized and synthesized data on viral hepatitis are limited and there was a recommendation to conduct a systematic review. A single systematic review was conducted in Ethiopia [
5]. However, in this systematic review, the prevalence of HBV among pregnant women was not estimated.
From the policy perspective, the health burden due to viral hepatitis, in general, are still given less attention in the country’s health system [
5,
15]. A recent report showed that not only the general population but also healthcare professional’s awareness of the epidemiology of hepatitis virus is low [
10,
15].
Several studies in Ethiopia recommended incorporation of routine antenatal care (ANC) screening program for Hepatitis B [
11,
12,
14,
16]. However, regular antenatal screening of pregnant women is not common and compulsory in Ethiopia [
17]. The absence of regular HBV screening program could be partly explained by lack of awareness on the overall burden of hepatitis B among pregnant women in Ethiopia by health professionals and policymakers. Therefore, this systematic review was conducted to give a quantitative estimate of the burden of HBV infection among pregnant women as a step to use for a better understanding of its epidemiology in Ethiopia and inform policymakers to take practical action at the policy level.
Discussion
The present systematic review and Meta-analysis aimed to estimate the prevalence of HBV infection among pregnant women in Ethiopia. The information obtained from this systematic review and Meta-analysis may improve knowledge on the epidemiology of HBV infection among pregnant women in Ethiopia.
The overall meta-analysis showed that the pooled prevalence of HBV infection in Ethiopia among pregnant women was 4.7%(95% CI 4.0–5.4%). This finding is close to WHO endemic definition of HBV infection(5–7%) [
41] and prevalence of chronic hepatitis B in Ethiopia estimated by Aparna Schweitzer et al.
.. (6.03%) [
42]. However, this finding is lower than the prevalence of HBV infection among general populations in Ethiopia (7.4%) [
5]. This may indicate the risk of HBV infection among pregnant women is lower than the risk of the general population. Alternatively, the high prevalence of HBV infection in the general population may be because of high mean age among the general population than the mean age of pregnant women. Moghaddasifar I et al also postulated that, the low prevalence of HBV infection among Iranian pregnant women(1.2%) is due to their lower mean age [
43].
The available systematic reviews and Meta-analysis in Africa for instance in Nigeria, Cameron and Ghana showed a high prevalence of HBV infection among pregnant women ranging from 9.8 to 14.1% [
34,
44,
45]. The findings from Nigeria, Cameron and Ghana are higher than that of the estimates from this Meta analysis. This difference could be attributed to the real difference in the prevalence of HBV infection among the general population. The prevalence of HBV infection among the general population in Nigeria, Cameron and Ghana were 13.6, 11.8 and 12.3% respectively that is higher than Ethiopian estimate (7.4%) [
5].
Furthermore, the difference in the risk of contracting HBV infection, socio-economic, environmental and behavioral risk factors may contribute to this difference.
In the contrary, the prevalence of HBV infection among pregnant women in Ethiopia is higher than estimation from several studies in European and American countries such as Denmark (0.26%) [
46], Brazil(1%) [
47], Spain (0.1%) [
48], Guatemala (0.22) [
49], Southeastern Turkey(1.74%) [
50] and Northern Turkey (2.1%) [
51]. The low prevalence of HBV infection among pregnant women in Europe and America may show the effect of higher socio-economic status, level of hygiene and vaccination coverage.
In this Meta-analysis, we couldn’t detect publication bias using eggers regression test. Lack of significant publication bias in the regression test might be the absence of significant dispersion in the sample among the included studies. It is well documented that lack of significant dispersion and reasonable sample size may cause non-significant eggers regression test [
52]. Borenstein M et al. [
52] argued even in the presence of significant dispersion and reasonable numbers of studies, regression test has lower power. Therefore, the failure to find evidence of publication bias using regression test should not lead to a false sense of assurance.
The intermediate level of HBV infection among pregnant women in this systematic review may indicate policymakers and programmers to implement routine and universal HBV screening program for all pregnant women.
Several other studies in Ethiopia [
8,
10‐
14,
16], a meta-analysis in Iran [
43] and studies in other countries [
50,
53‐
55] also recommended the establishment of antenatal HBV screening program. Universal screening program enables women to access HBV treatment. Furthermore, it may increase child vaccination coverage. For instance, in Denmark after implementation of universal screenings for hepatitis B, vaccination coverage raise to 96% [
46].
The finding from this systematic review needs to be interpreted cautiously because all studies included in this systematic review and Meta-analysis were institutional based studies which covered a portion of pregnant women in Ethiopia. These women might be relatively healthy, educated and economically empowered so that their risk of HBV infection is minimal. Furthermore, the included studies used different HBsAg screening tools. Some of the screening tools (Linear Chemicals, Joaquim Costa, and Barcelona, Spain) cannot detect less than 1 ng/mL of HBsAg in specimens. As result, false negative results may occur if the quantity of HBsAg present in the specimen < 1 ng/mL. In rare cases, HBsAg tests do not detect certain HBV mutant strains. Mutated HBsAg may not be detected by the available HBsAg screening tools (Linear chemicals. Joaquim Costa, Barcelona, Spain) as a result false negative results could be occurred. All these limitations may underestimate the true pooled prevalence of HBV infection among pregnant women in Ethiopia.
Limitations and strengths of the study
Most of the studies included were conducted in more recent years. As a result, the prevalence estimates in this Meta-analysis is likely to show the current situation of HBV infection among pregnant women in Ethiopia.
The full text of one article was not accessed. Efforts were made to access this article from the primary and coauthors through email and phone. However, positive response was not acquired. As a result, to some extent, this review might not be free from full publication bias.
Finally, in this review, we included articles published only in the English language. Thus, the introduction of language bias is expected. However, publication in other languages in Ethiopia is uncommon. For this reason, the introduction of language bias is minimal.