Prevalence of and risk factors for Plasmodium spp. co-infection with hepatitis B virus: a systematic review and meta-analysis

Background Plasmodium spp. and hepatitis B virus (HBV) are among the most common infectious diseases in underdeveloped countries. This study aimed to determine the prevalence of Plasmodium spp. and HBV co-infection in people living in endemic areas of both diseases and to assess the risk factors related to this co-infection. Methods The PubMed, Web of Science, and Scopus databases were searched. Observational cross-sectional studies and retrospective studies assessing the prevalence of Plasmodium species and HBV co-infection were examined. The methodological quality of the included studies was assessed with the Newcastle-Ottawa Scale (NOS), a tool for assessing the quality of nonrandomized studies in meta-analyses, and heterogeneity among the included studies was assessed with Cochran's Q test and the I2 (inconsistency) statistic. The pooled prevalence of the co-infection and its 95% confidence interval (CI) were estimated using the random-effects model, depending on the amount of heterogeneity there was among the included studies. The pooled odds ratio (OR) represented the difference in qualitative variables, whereas the pooled mean difference (MD) represented the difference in quantitative variables. Meta-analyses of the potential risk factors for Plasmodium spp. and HBV co-infection, including patient age and gender, were identified and represented as pooled odds ratios (OR) and 95% CIs. Publication bias among the included studies was assessed by visual inspection of a funnel plot to search for asymmetry. Results Twenty-two studies were included in the present systematic review and meta-analysis. Overall, the pooled prevalence estimate of Plasmodium spp. and HBV co-infection was 6% (95% CI 4–7%, Cochran's Q statistic < 0.001, I2: 95.8%), with prevalences of 10% in Gambia (95% CI: 8–12%, weight: 4.95%), 8% in Italy (95% CI 5–12%, weight: 3.8%), 7% in Nigeria (95% CI 4–10%, weight: 53.5%), and 4% in Brazil (95% CI 2–5%, weight: 19.9%). The pooled prevalence estimate of Plasmodium spp. and HBV co-infection was higher in studies published before 2015 (7%, 95% CI 4–9%, Cochran's Q statistic < 0.001, I2: 96%) than in those published since 2015 (3%, 95% CI 1–5%, Cochran's Q statistic < 0.001, I2: 81.3%). No difference in age and risk of Plasmodium spp. and HBV co-infection group was found between the Plasmodium spp. and HBV co-infection and the Plasmodium monoinfection group (p: 0.48, OR: 1.33, 95% CI 0.60–2.96). No difference in gender and risk of Plasmodium spp. and HBV co-infection group was found between the Plasmodium spp. and HBV co-infection and HBV co-infection group and the Plasmodium monoinfection group (p: 0.09, OR: 2.79, 95% CI 0.86–9.10). No differences in mean aspartate aminotransferase (AST), mean alanine aminotransferase (ALT), or mean total bilirubin levels were found (p > 0.05) between the Plasmodium spp. and HBV co-infection group and the Plasmodium monoinfection group. Conclusions The present study revealed the prevalence of Plasmodium spp. and HBV co-infection, which will help in understanding co-infection and designing treatment strategies. Future studies assessing the interaction between Plasmodium spp. and HBV are recommended.


Background
Malaria in humans is caused by the infection of at least one of the five Plasmodium species, including Plasmodium falciparum, Plasmodium vivax, Plasmodium ovale, Plasmodium malariae, and Plasmodium knowlesi [1]. Malaria is transmitted by the bite of an infected female Anopheles mosquito [2]. Malaria remains endemic among people around the world, especially people in the World Health Organization (WHO) African Region and the WHO South-East Asia Region [3]. According to the latest World Malaria Report 2019, there were an estimated 228 million cases of malaria with an estimated number of malaria deaths at 405,000, mostly children under 5 years of age (67%) [4]. Most malaria cases were reported from six countries in the African region including Nigeria (25%), the Democratic Republic of the Congo (12%), Uganda (5%), Côte d'Ivoire (4%), Mozambique (4%) and Niger (4%) [4]. Hepatitis B infection is caused by hepatitis B virus (HBV), a double-stranded DNA virus belonging to the Hepadenaviridae family, and its infection leads to a wide range of clinical spectra from acute to chronic hepatitis, cirrhosis, and hepatocellular carcinoma [5]. The infection of HBV in individuals was acquired through exposure to potentially infectious blood or blood products or through percutaneous exposure to sharp contamination [6]. Globally, an estimated 257 million people were living with chronic HBV infection in 2015, and most people in the WHO African Region and the Western Pacific Region were affected [7]. The reduction in HBV prevalence among children (1.3%) was seen after the introduction of the hepatitis B vaccine in 2015 [7].
Plasmodium spp. and HBV infections are endemic among people residing in the same regions and are prone to co-infect individuals because of their geographical coincidence [8,9]. In individuals with Plasmodium spp. and HBV co-infection, the two pathogens use the liver as their host during their developmental stages, which may result in liver injuries and lead to increased mortality and morbidity [10]. A previous study demonstrated that HBV infections could reduce the ability of the liver to kill Plasmodium parasites [10,11]. A study in patients with Plasmodium spp. and HBV co-infection demonstrated that P. falciparum modulates HBV viremia in patients with chronic HBV infection [12]. A study conducted in Asia proposed that chronic HBV infection may lead to the synergistic multiplication of P. falciparum malaria, but the overall risk of death was not significantly higher in coinfected patients [13]. Another study proposed that HBV infection might reduce the density of Plasmodium parasites in malaria patients with no organ dysfunction [9], while a previous study demonstrated that Plasmodium and HBV co-infection significantly increased the density of malaria parasites [14].
Currently, studies on the risk factors for Plasmodium spp. and HBV co-infection and the differences in liver function tests (LFTs) are sparse and inconsistent. Demographic profiles of patients, including age and gender, might explain the possible risks for co-infection, as suggested by previous studies indicating that most cases of Plasmodium spp. and HBV co-infection occured in patients aged 20-50 years [15,16], while another study demonstrated a higher proportion of Plasmodium spp. and HBV co-infection cases among patients who are ≥ 50 years of age [17]. A better understanding of the overall prevalence of Plasmodium spp. and HBV co-infection, the potential risk factors, and the LFTs in infected individuals could help endemic countries to diagnose, prevent, and control Plasmodium spp. and HBV co-infection. Therefore, the present study aimed to synthesize evidence regarding Plasmodium spp. and HBV co-infection using a meta-analytic approach and to assess the effects of demographic profiles, including age and gender, on the risks of Plasmodium spp. and HBV coinfection. In addition, the differences in the levels of LFTs between individuals with Plasmodium spp. and HBV coinfection and those with Plasmodium spp. monoinfection were also investigated. those methods. HBV infections were diagnosed by RDT or PCR, enzyme-linked immunosorbent assay (ELISA), or a combination of those methods.

Search strategies
The PubMed, Web of Science, and Scopus databases were searched for studies on co-infection with Plasmodium spp. and HBV that were published prior to April 16, 2020. To maximize the search, only three Medical Subject Headings (MeSH) were used: (malaria OR plasmodium) AND "hepatitis B" AND "co-infection". To maximize the number of included studies, studies on humans published in any language were considered. The references of the included studies and relevant review articles were also reviewed to identify additional relevant studies.

Eligibility criteria and selection of studies
Studies were included in this study if they were (1) observational studies performed on participants coinfected with Plasmodium spp. and HBV and (2) observational studies reporting on the number of Plasmodium spp. and HBV coinfected and monoinfected participants. Studies that did not meet the inclusion criteria were excluded, including animal studies, books and book chapters, experimental studies, studies on malaria co-infection with other pathogens, studies that did not examine coinfection, case reports/case series, randomized control trials, studies with a similar group of participants, and studies with no full text available. The relevant review articles were stored and reviewed as potentially eligible articles that met the inclusion criteria only, but they were not included in the meta-analysis.

Data extraction
Two authors (MK and KUK) independently screened and selected the relevant studies related to the eligibility criteria. Data from the potentially eligible studies were retrieved and extracted to a standardized data extraction form (Microsoft Excel, Microsoft Corporation, USA). Any discrepancies in study selection and extractions were resolved by consensus or discussion. Information on authors, year of publication, study area, years of the study conducted, study design, age range, gender, number of Plasmodium spp. monoinfections, number of HBV monoinfections, number of Plasmodium spp. and HBV co-infection, detection methods for Plasmodium spp. and HBV, and laboratory data (LFTs) were extracted.

Quality of the included studies
The quality of the included studies was determined following the Newcastle-Ottawa Scale (NOS), a tool for assessing the quality of nonrandomized studies in meta-analyses that consists of three domains: selection, comparability, and outcomes [19]. A maximum of six stars (the highest quality) for the cross-sectional and retrospective studies was modified from the NOS to assessing the quality of the included studies. Any studies rated ≥ 5 stars were considered high-quality studies, while studies rated < 5 stars were considered low-quality studies ( Table 2).

Data synthesis
The primary outcome of the present study was the pooled prevalence of Plasmodium spp. and HBV co-infection, which was estimated using the Freeman-Tukey double arcsine transformation method with the DerSimonian and Laird random-effects model. These analyses were run using the "metaprop" command provided in Stata (StataCorp, USA) [20,21]. In 2015, the WHO released the 'WHO Guidelines for the treatment of malaria' [22], which consisted of recommendations on the diagnosis and treatment of severe and uncomplicated malaria among at-risk populations [22]. The subgroup analysis of the pooled prevalence estimate between years of studies conducted (year 2014 or less, year 2015 and beyond) was performed to identify any differences in the pooled prevalence estimate between subgroups. The secondary outcome of the present study was the pooled effect estimates (odds ratio, OR) for the age group, gender, aspartate aminotransferase (AST) levels, alanine aminotransferase (ALT) levels, and total bilirubin levels. Differences in the effect estimates between the groups were examined. The results of the pooled effect estimates were presented as a pooled OR and its 95% confidence interval (CI) for dichotomous variables, whereas the results of the pooled effect estimates were presented as pooled mean difference (MD) and its 95% CI for continuous variables. The heterogeneity across the included studies was assessed with Cochran's Q test and Higgins I 2 (inconsistency) statistic. A significant Cochran's Q test (p < 0.05) with an I 2 value greater than 50% indicated substantial heterogeneity. The fixed-effects model was used in cases of no significant heterogeneity, while the random-effects model was used in cases of significant heterogeneity across the included studies.

Publication bias assessment
The publication bias among the included studies was assessed by a visual inspection of the funnel plot to search for asymmetry (the asymmetrical distribution of the included studies in the graph between the OR and SE (logOR)). The funnel plot was generated with data on the age of participants and status of Plasmodium spp. and HBV co-infection.

The prevalence of Plasmodium spp. and HBV co-infection
All 22 studies included in the present systematic review and meta-analysis provided data on the prevalence of Plasmodium spp. and HBV co-infection. The results showed that the prevalence estimate of Plasmodium spp. and HBV co-infection among 22 studies varied widely, ranging between 1 and 41%. The highest proportion of Plasmodium spp. and HBV co-infection was found in Nigerian blood donors (41%) as reported by Aernan et al. [10]. Overall, the pooled prevalence estimate of Plasmodium spp. and HBV co-infection was 6% (95% CI 4-7%, Cochran's Q statistic < 0.001, I 2 : 95.8%), and there was highly significant heterogeneity (p < 0.001, I 2 : 95.8%) among the included studies (Fig. 2).

Subgroup analysis on the prevalence of Plasmodium spp. and HBV co-infection
A subgroup analysis of the pooled prevalence estimate between studies conducted before 2015 and studies conducted since 2015 revealed that the pooled prevalence estimate of Plasmodium spp. and HBV co-infection in 13 studies published before 2015 was 7% (95% CI 4-9%, Cochran's Q statistic < 0.001, I 2 : 96%), while the pooled prevalence estimate of Plasmodium spp. and HBV coinfection in three studies published since 2015 was 3% (95% CI 1-5%, Cochran's Q statistic < 0.001, I 2 : 81%). The pooled prevalence estimate of Plasmodium spp. and HBV co-infection in studies conducted during 2013-2015 and studies that did not specify the year of studies conducted are demonstrated in Fig. 2.
An analysis of gender and risk of Plasmodium spp. and HBV co-infection was performed. The results showed that among the five included studies, there was no significant difference in gender and risk of Plasmodium spp. and HBV co-infection (p: 0.09, OR: 2.79, 95% CI 0.86-9.10, I 2 : 84%) (Fig. 5). The highest proportion of males (OR: 16.1, 95% CI 7.5-34.7) was found in blood donors as reported by Aernan et al. [10].

Quality of the included studies
Overall, the quality of the 22 included studies was assessed using the NOS with some modifications for cross-sectional and retrospective studies. All included studies were high-quality and had ratings ranging from six stars to seven stars ( Table 2).

Publication bias
The funnel plot revealed an asymmetrical distribution of the included studies in the graph between the OR and SE (logOR) (Fig. 7), indicating publication bias due to the small study effects.

Discussion
The present systematic review and meta-analysis provided information on the overall prevalence of Plasmodium spp. and HBV co-infection. The results demonstrated that there was a high prevalence estimate of Plasmodium spp. and HBV co-infection (6%). Nevertheless, the pooled prevalence estimate of Plasmodium spp. and HBV co-infection was not precisely estimated, as indicated by the highly significant amount of heterogeneity among the included studies (I 2 = 95.76, p < 0.001).
When assessing the individual studies, Plasmodium spp. and HBV co-infection was prevalent in most tropical and sub-Saharan African countries, particularly Nigeria [10,15,16,26,[28][29][30][31][33][34][35][36], Ghana [8,23,25], Central African Republic [24], and The Gambia [11]. The pooled prevalence estimate of Plasmodium spp. and HBV coinfection was highest (6%) among the included studies conducted in Nigeria. However, due to the significant amount of heterogeneity (p < 0.001, I 2 : 95.41%) across these included studies, the pooled prevalence estimate in Nigeria might be confounded by at least one study conducted in Nigerian blood donors, which reported the highest proportion of Plasmodium spp. and HBV coinfection (41%) [10]. Compared with studies conducted in Nigeria, the pooled prevalence estimate of Plasmodium spp. and HBV co-infection in Ghana was lower in the included studies (2%). There was also a significant amount of heterogeneity in the pooled subgroup analysis (p < 0.001, I 2 : 94.32%), which might be due to at least one study reporting a prevalence of 16% among transfusion recipients [8]. The pooled prevalence estimate of Plasmodium spp. and HBV co-infection in studies conducted in Brazil showed the lowest amount of heterogeneity (p < 0.001, I 2 : 79.02%) compared with studies conducted in Nigeria and Ghana. The study conducted  among African immigrants in Italy reported a high prevalence of Plasmodium spp. and HBV co-infection and suggested that this finding was due to the participants having acquired semi-immunity with submicroscopic levels of parasitaemia, resulting in no sign or symptoms of malaria before detection [27]. The subgroup analysis of the years of studies revealed that the pooled prevalence estimate of Plasmodium spp. and HBV co-infection was higher among studies conducted before 2015 (7%) than those conducted since 2015 (3%). However, the pooled prevalence of Plasmodium spp. and HBV co-infection among studies conducted before 2015 was not precisely estimated, as indicated by the significant amount of heterogeneity among the 13 included studies [9-11, 14, 16, 17, 24, 26, 28, 31, 32, 34, 35] (p < 0.001, I 2 : 96%). The pooled prevalence of Plasmodium spp. and HBV co-infection among studies conducted since 2015 (3%) was lower than the  prevalence among studied conducted before 2015. In addition, a significantly lower amount of heterogeneity among the three studies [23,29,30] (p < 0.001, I 2 : 81.3%) was found, indicating a more reliable pooled prevalence estimate of Plasmodium spp. and HBV co-infection than the prevalence among the studies conducted before 2015. In light of these results, the reduction in the pooled prevalence between studies conducted in and after 2015 compared with studies conducted before 2015 might be due to the new WHO guidelines implemented in 2015, which updated the recommendations for the treatment of malaria, including the dosing of drugs in children and the use of drugs for preventing malaria in high-risk groups (young children, pregnant women, tuberculosis or HIV/ AIDS patients, non-immune travellers) [22].
Little is known about the effect of Plasmodium spp. and HBV co-infection on disease severity. The interaction between two pathogens in hosts has been studied. The interactions between Plasmodium spp. and HBV in the same individuals in Papua New Guinea demonstrated that patients with severe malaria had the lowest prevalence of HBV infection [37]. The study of the immune response among patients with malaria and HBV coinfection demonstrated a robust pro-inflammatory Type 1 immune response (Th1), which is important for Plasmodium spp. clearance induced by HBV; nevertheless, it caused disease severity from co-infection [38]. HBV replication in liver cells also enhanced the production of interferon (IFN)-γ and IFN-α/β. A study in experimental mice demonstrated that intrahepatic HBV replication was inhibited by the Plasmodium yoelii 17X NL [39]. They also reported that Plasmodium spp. and HBV coinfected individuals demonstrated elevated concentrations of IL-10 (eight-fold) and C-C Motif Chemokine Ligand 2 (CCL2) in comparison to those with P. vivax infection [14]. CCL2 has been reported to be produced by hepatocytes during HBV infection [40] and is related to P. vivax infection [41].
A meta-analysis was conducted to determine whether age was a potential risk factor for Plasmodium spp. and HBV co-infection. The meta-analysis of age included seven studies [15-17, 26, 28-30] and revealed that no significant difference in age was found between the Plasmodium spp. and HBV co-infection group and the Plasmodium spp. monoinfection group. There was a significant heterogeneity among the included studies (p < 0.0001, I 2 : 69%). To explore the source of heterogeneity among the included studies that might affect the pooled OR estimate, a subgroup analysis of age (< 20, 20-50, and > 50 years of age) was performed. The subgroup analysis demonstrated that no significant difference in age or risk of co-infection was found among the three age groups. The pooled OR estimates were reliable in individuals aged < 20 (p < 0.49, I 2 : 0%) and those  Table 2 Quality of the included studies A star system developed by the Newcastle-Ottawa Scale (NOS) for assessing the quality of non-randomized studies in meta-analyses, which consists of three domains including selection, comparability, and outcomes.
A maximum of six stars (the highest quality) for the cross-sectional and retrospective studies was modified from the NOS for assessing the quality of the included studies. Any study rated  1 Abah and Udoidang 2019 [30] aged > 50 years (p: 0.06, I 2 : 60%) with no significant amount of heterogeneity among the included studies. This finding implies that an age < 20 and an age > 50 years were not potential risk factors for Plasmodium spp. and HBV co-infection, while the pooled OR estimates of age were not reliable in individuals aged between 20-50 years, as indicated by the significant amount of heterogeneity among the included studies (p < 0.0001, I 2 : 82%). This means that an age between 20 and 50 years old might be a potential risk factor for Plasmodium spp. and HBV co-infection, but this finding could not be confirmed in the present meta-analysis and needs to be evaluated by further longitudinal studies. In light of these results, research conducted in individuals aged between 20-50 years by two studies, Afolabi et al. [15] and Dabo et al. [16], demonstrated the highest prevalences of Plasmodium spp. and HBV co-infection in their studies. The higher prevalence of Plasmodium spp. and HBV co-infection among the youthful age group could be attributed to their higher-risk behaviours for infection by both Plasmodium spp. and HBV, including unprotected sexual activities, drug abuse, fashionable tattooing and skin piercing, the lack of mosquito nets for beds, and failure to complete the recommended drug doses for treatments [10,30]. Another possible explanation for the high proportion of Plasmodium spp. and HBV co-infection among the youthful age group could be explained by females in this age group giving birth to children. These females (pregnant) visited clinics or hospitals during their pregnancy health check-up, which resulted in a diagnosis and contributed to the high prevalence of Plasmodium spp. and HBV co-infection. This rationale was supported by the high proportion of pregnant women coinfected with Plasmodium spp. and HBV in four of the included studies [23,26,29,31].
The meta-analysis of gender among five included studies [10,16,26,28,30] demonstrated that male and female participants were comparable in risk of Plasmodium spp. and HBV co-infection (p: 0.09) with a significant amount of heterogeneity across the included studies (p < 0.0001, I 2 : 84%). This means that the pooled OR of gender in those five studies was not precisely estimated due to the high heterogeneity. In light of this result, the association between gender and risk of Plasmodium spp. and HBV co-infection should be further investigated, as the high proportion of males coinfected with Plasmodium spp. and HBV than females was reported in the study by Aernan et al. [30].
The differences in LFTs between Plasmodium spp. and HBV co-infection were performed using two studies [14,28]. The results demonstrated that no significance in any of the three LFTs, including AST, ALT, and total bilirubin levels, was found (p > 0.05); there was a significant amount of heterogeneity among these studies (p < 0.0001, I 2 ≥ 99%). Considering individual studies in the meta-analysis, one study demonstrated lower ALT and AST levels in the Plasmodium spp. and HBV co-infection group than in the Plasmodium spp. monoinfection group [14], while another study demonstrated no difference in the ALT and AST levels between groups [28]. The included 2018 study by Kolawole and Kana indicated high ALT (54.5%, 6/11 cases) and total bilirubin levels (72.7%, 8/11 cases) among patients with Plasmodium spp. and HBV co-infection [33].
The present study had several limitations. First, there was a limited number of included studies reporting on the prevalence of, potential risk factors for, and LFTs among individuals with Plasmodium spp. and HBV co-infection and in comparison with Plasmodium spp. monoinfection. Therefore, the meta-analyses of the pooled prevalences, pooled ORs, and pooled MDs were limited by these data. Second, there was a significant amount of heterogeneity among the included studies, resulting in imprecise estimates of the pooled prevalence estimate of Plasmodium spp. and HBV coinfection. Therefore, the pooled prevalence estimates should be interpreted with caution. Third, the present study did not analyse the other potential risk factors for Plasmodium spp. and HBV co-infection since most of the included studies did not examine or report for these factors.

Conclusion
The present study demonstrated the status of Plasmodium spp. and HBV co-infection. The results of this study can be used to support health care communities, helping them to recognize the double burden of these two