Background

Hepatitis C virus (HCV) infection is a bloodborne disease which is globally distributed that mainly affects developing countries. In 2019, an estimated 58 million chronic cases were reported in the world [1], 75% of which occurred in low- and middle-income countries (LMIC) [2]. Worldwide, new HCV infections occur in high-risk populations, including people who inject drugs (PWID) and men who have sex with men (MSM) [3]. In sub-Saharan Africa, HCV is mainly transmitted via unsafe medical practices and contaminated blood transfusion [4]. Other transmission routes include needlestick injury in healthcare workers, mother-to-child transmission (MTCT), and social practices such as piercing and tattooing [3].

HCV infections are usually asymptomatic and silently progress over time. Up to 25% of new infections spontaneously resolve [5], and the remaining evolve to chronicity with complications such as liver cirrhosis and hepatocellular carcinoma (HCC) [3]. Since no vaccine is available to prevent HCV infection and its chronic consequences, screening and treatment of cases using effective direct-acting antivirals (DAA) are required [6]. HCV antibody (anti-HCV) assays are typically positive within 4–10 weeks after the initial infection, persist lifelong, and indicate a current or past exposure to the virus [7]. Thus, the WHO requires a confirmatory test by detecting HCV ribonucleic acid (HCV RNA) or core antigen (HCVcAg) [8].

In Burkina Faso, a West African country, the epidemiologic patterns of HCV infection are poorly documented. The only nationwide study conducted in 2010 reported a seroprevalence of 3.6% (95% CI: 3.3–3.8) in the general population [9]. In addition, various seroprevalence data classify the country among those of intermediate or high seroprevalence (seroprevalence ≥ 2% or ≥ 5%, respectively) [10,11,12]. However, detailed characteristics of the infection in specific populations and geographic areas are scarce, representing a limitation for implementing effective control strategies. Thus, this study aimed to provide detailed knowledge on the epidemiologic profile of HCV infection in Burkina Faso by synthesizing data on HCV seroprevalence in various settings and populations at-risk in Burkina Faso. Secondary objectives included the prevalence of HCV RNA and the distribution of genotypes in the country.

Methods

Study design and guidelines

We conducted a systematic review with meta-analysis following the Joanna Briggs Institute guidelines for systematic reviews of studies reporting prevalence data [13]. This manuscript is reported according to the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) recommendations [14].

Study setting

Burkina Faso is a landlocked country in West Africa with an overall population of 20,487,979 people in 2019 [15]. The country is divided into 13 regions, with the central region accounting for 14.8% of the total population. About one in four inhabitants live in rural areas. This developing country is characterized by the persistence of tropical infections and the epidemiological transition to non-transmissible diseases.

Data source and study selection

We searched the following databases to identify records published between 1990 and 2020: PubMed, Web of Science, Scopus, and African Index Medicus. A complete list of the search queries is shown in Additional file 1: Table S1. Besides, a manual search was conducted on African Journals Online (AJOL), Google Scholar, the reference lists of eligible reports, and the electronic library of the University Joseph Ki-Zerbo (www.biblio-ujkz.com), the largest university in the country. The last search was conducted on July 1, 2020, and reports in English or French were eligible. After removing duplicates using the EndNote reference manager, two independent investigators (SO and JCRPO) screened the records based on their titles and abstracts. Those deemed relevant were retained for full-text review. Any report (journal article, conference abstract, government report) of HCV antibody testing among people living in Burkina Faso was considered for inclusion. Reports that did not mention sample size and number of cases (or prevalence) were excluded. Disagreements between the two investigators were solved by discussion. When there was no consensus, the final decision was made by a third investigator (ML).

Data extraction

A pre-piloted standardized electronic data extraction form was used to extract relevant data: authors names, year of publication, study setting, study design, sampling method, study period, characteristics of study subjects (age, sex, location, co-morbidities), sample size, number of participants positive for anti-HCV, and type of biological assay. We also extracted HCV RNA and genotyping data when available. In this paper, a “report” refers to any document mentioning HCV seroprevalence data, and a “study” relates to the measurement of anti-HCV antibodies in a given population or setting (rural/urban). Therefore, one report may describe multiple studies, and in this case, we extracted data for each study separately. When data were suspected to originate from the same source, only one publication was included.

Quality appraisal

Joanna Briggs Institute checklist for prevalence studies was used to appraise the methodological quality of the studies [13]. Before independently using the tool, the reviewers agreed on the minimum acceptable information for each item. Sample size was considered adequate when more than 411 subjects were included. This was based on an assumed seroprevalence of 3.6% [9], a precision of 1.8%, and a confidence level of 95%. For the sampling procedure, only probabilistic sampling was considered representative of the target population. Anti-HCV seroprevalence data was considered valid when the diagnostic was based on biological testing.

Data synthesis and analysis

We categorized the study populations into two groups based on the risk of HCV transmission. The low-risk group included the general population, blood donors, pregnant women, and children. People living with Human Immunodeficiency Virus (HIV), sex workers, MSM, and healthcare workers represented the high-risk group.

A meta-analysis of proportion was performed using the 'meta' and 'metafor' packages in the statistical program R. Data were transformed using the Freeman-Tukey double arcsine method to account for small proportions. The Dersimonian and Laird method, based on the random-effects model, was used to perform the meta-analysis and summarize data in a forest plot. Confidence-interval (CI) for individual studies proportions were calculated using the Clopper-Pearson method. Heterogeneity was assessed based on the Cochran Q test and quantified by the I2 index. Heterogeneity was considered significant when the p-value of the Cochran Q test was less than 0.05. The level of heterogeneity was rated as high, medium, or low when the I2 index was 75%, 50%, and 25%, respectively [16]. Publication bias was evaluated graphically by a funnel plot of the transformed proportion against the sample size [17]. We used the Egger test to assess the symmetry of the plot (p < 0.1).

Results

Search results

A total of 377 records were identified from the database search. After duplicates exclusion and titles and abstracts screening, 32 reports were selected for full-text review. Of these, 25 met the inclusion criteria, and six reports were further included, resulting in 31 reports. Finally, the qualitative review and the meta-analysis included 37 studies covering a population of 169,635, as some reports provided data on multiple populations or settings. The selection process is summarized in Fig. 1.

Fig. 1
figure 1

Flowchart of records identification and study selection. A “report” refers to any document mentioning HCV seroprevalence data, and a “study” relates to the measurement of anti-HCV antibodies in a specific population or setting. Of the 31 reports that met the review criteria, four reported seroprevalence data on separate study subjects or settings (as shown in Table 1), and each seroprevalence data was extracted as an individual study. Therefore, 31 reports met the criteria for this review, and 37 prevalence data, referred to as “studies”, were extracted

Full size image

Study characteristics

The details about the characteristics of each study are reported in Table 1. Studies were conducted between 1994 and 2018. The central region was the most represented (22/37 studies), and seven studies addressed the rural area. The cross-sectional design (33/37) was the main method used, with only eight studies (21.62%) using a probabilistic sampling. Sample size was considered adequate (> 411) in 46.0% of the studies, and the presence of anti-HCV antibodies was ascertained by biological assays in 34 studies (92.0%) (Additional file 1: Table S2). Enzyme immunoassays (ELISA, CLEIA) were the most used (24/34 studies), followed by rapid tests (7/34) and immunoblotting (3/34).

Table 1 Characteristics of included studies reporting HCV seroprevalence in Burkina Faso between 1990 and 2020
Full size table

Seroprevalence in the low-risk group

Low-risk populations were addressed in 31 studies and involved a total of 168,151 subjects, of whom 8330 were positive for HCV antibodies. Ten studies focused on the general population, eleven on blood donors, eight on pregnant women, and two on children. The seroprevalence ranged between 0 and 8.69%, and the pooled seroprevalence was 3.72% (95% CI: 3.20–4.28). However, significant heterogeneity was found in this group (I2 = 94.42%, p < 0.001) (Fig. 2).

Fig. 2
figure 2

Meta-analysis of HCV seroprevalence in Burkina Faso by risk group. Pooled seroprevalence based on random-effects model. Low-risk group includes the general population, blood donors, pregnant women, and children. High-risk group is constituted of people living with Human Immunodeficiency Virus (HIV), sex workers, men having sex with men (MSM) and healthcare workers

Full size image

Seroprevalence in the high-risk group

Subjects in the high-risk group were investigated in six studies: two on healthcare professionals, two on men having sex with men (MSM), one on people living with HIV, and one on sex workers. A total of 1,484 high-risk persons were studied, and 96 had antibodies to HCV. Seroprevalence varied from 1.28% to 10.94% (Fig. 2), and the combined seroprevalence was 4.75% (95% CI: 1.79–8.94). Heterogeneity was also significant in this group (I2 = 90.26%, p < 0.001).

Seroprevalence by study subjects

Among populations with a low risk of infection, blood donors had the highest seroprevalence (4.51%, 95% CI: 3.73–5.35, I2 = 96.51%, p < 0.001), followed by pregnant women (3.95%, 95% CI: 2.67–5.46, I2 = 71.86%, p < 0.001). The seroprevalence in the general population was 3.11% (95% CI: 2.05–4.38, I2 = 92.14%, p < 0.001), and 2.11% (95% CI: 0.68–4.16, I2 = 0%, p = 0.59) in children. HCV seroprevalence was highest among sex workers (9.85%, 95% CI: 6.82–13.34) and HIV patients (7.76%, 95% CI: 4.63–11.59). Healthcare workers had the lowest seroprevalence (1.57%, 95% CI: 0.28–3.68, p = 0.18). The summary of the seroprevalence by study subjects is shown in Table 2.

Table 2 HCV seroprevalence in subgroups
Full size table

Seroprevalence by study setting

The analysis of the seroprevalence by study setting was performed considering only low-risk groups (Table 2). Seven studies were conducted in rural areas and covered a population of 12,046 people. The pooled seroprevalence was 3.27% (95% CI: 2.35–4.33, I2 = 59.24%, p = 0.02). Urban populations were addressed in 24 studies and yielded a combined seroprevalence of 4.02% (95% CI: 3.40–4.68, I2 = 95.44%, p < 0.001).

Seroprevalence by the decade of data collection

Only low-risk populations were considered for the analysis of HCV seroprevalence by the period of data collection. Two studies were conducted between 1990 and 2000, resulting in a pooled seroprevalence of 5.33% (95% CI: 1.42–11.45, I2 = 91.06%, p < 0.001). The seroprevalence decreased to 3.95% (95% CI: 2.85–5.21, I2 = 90.37%, p < 0.001, 13 studies) during 2000–2010 and to 3.78% (95% CI: 3.19–4.41, I2 = 92.71%, p < 0.001, 13 studies) between 2010 and 2020 (Table 2). However, this decrease was not significant (p for difference = 0.78).

HCV RNA prevalence

Eight studies assessed the presence of HCV RNA, including two in rural settings and only one in the high-risk group. HCV RNA prevalence ranged from 0.47 to 7.03%. The highest prevalence estimates were found in HIV-positive patients (6.47%) and urban populations surveyed in 1994–1995 (7.03%). The seven studies that tested HCV RNA in low-risk populations included a total of 4,759 individuals, of whom 81 were positive. The meta-analysis yielded a pooled HCV RNA prevalence of 1.65% (95% CI: 0.74–2.89, I2 = 85.04%, p < 0.001) in the low-risk group. Figure 3 shows the forest plot of HCV RNA prevalence.

Fig. 3
figure 3

Meta-analysis of HCV RNA prevalence in Burkina Faso by risk group. Pooled prevalence based on random-effects model. Low-risk group includes the general population, blood donors, and pregnant women. Only one study evaluated HCV RNA in high-risk groups (people living with HIV)

Full size image

HCV genotypes

The genotyping of the virus was performed in six studies [18,19,20,21,22], and genotype 2 was the most predominant (60.3%), followed by genotype 1 (25.0%) and genotype 3 (7.4%). Mixed infections accounted for 5.9%, and genotype 4 was found in only one subject (1.5%) (Fig. 4).

Fig. 4
figure 4

Distribution of HCV genotypes in Burkina Faso. Mixed infections consisted of genotypes 2/3 and 2/4

Full size image

Publication bias

No publication bias was observed through the graphical assessment of the funnel plot (Fig. 5) and the Egger test for funnel plot asymmetry (p-value = 0.23).

Fig. 5
figure 5

Funnel plot of HCV seroprevalence meta-analysis. Double arcsine transformed proportion of individual studies is plotted against the sample size. The symmetrical distribution of studies in the funnel plot suggests an absence of publication bias

Full size image

Discussion

We evaluated the seroprevalence of HCV infection in Burkina Faso between 1990 and 2020 in several settings and populations. The reported seroprevalence in low-risk groups was 3.72% (95% CI: 3.20–4.28) and may reflect that of the general population, classifying the country among those with intermediate seroprevalence [8]. Previous analyses reported a seroprevalence between 4.9 and 6.1% in Burkina Faso, ranking the country among those with the highest seroprevalence in West Africa [10, 12]. However, these reports included high-risk populations in their estimations. In addition, as HCV seroprevalence decreased over time, the inclusion of more recent studies in our review may explain our lower rate. The rate reported in this review is similar to that of the only nationwide survey conducted in 2010, which found a weighted seroprevalence of 3.6% (95% CI: 3.3–3.8) in the general population [9].

Although non-significant, a decreasing trend was observed over the past three decades and can be attributed to several factors, including the improvement of transfusion and injection safety. Since 2000, blood transfusion in Burkina Faso is managed by the national center for blood transfusion (CNTS), which has the capacity for infection screening. However, in 2009, its production capacity covered only half of the needs, with the remaining directly collected in the health facilities not supplied by the CNTS. A survey of 42 of these health centers showed that 14.3% were not routinely performing HCV testing [23]. In addition, a residual risk of HCV transmission persists, estimated at one in 213 donations, as blood screening for HCV is only based on the detection of anti-HCV antibodies [24, 25]. Therefore, nationwide coverage of the blood supply by the CNTS and nucleic acid testing for HCV is recommended to reduce the risk of transfusion-transmitted hepatitis C.

Improved injection safety is the other factor that could explain the downward trend of infection rates. Indeed, in 1996, it was estimated that 11% of rural and 80% of urban health centers were using new and sterile syringes and needles for injections [26]. In 2000, a survey of 52 nationally representative health facilities found that this proportion increased to 96%, and no shortage of syringes or needles was reported in those health centers [27].

The pooled seroprevalence of HCV infection among pregnant women indicates a risk for MTCT, estimated at 4.2–7.8% among viremic women [28]. Nevertheless, the WHO does not recommend routine testing of pregnant women for HCV infection, as currently there is no treatment to prevent MTCT, and DAAs are not indicated during pregnancy [2]. Therefore, adequate detection and treatment of childbearing age women during the preconception period could be recommended.

Rural communities were understudied, although they accounted for 73.7% of the latest national population census [15]. Therefore, the reported seroprevalence may underestimate the magnitude of the actual situation in rural settings. It is thus essential to assess the burden of HCV infection on rural populations, as limited care access, low literacy, and low socioeconomic status are known factors of HCV infection [4, 29].

Few studies were conducted among key populations for HCV infection. Interestingly, two studies included MSM and one involved sex workers, two hard-to-reach groups. However, no study was conducted among PWID. Evidence exists about injecting drug usage in Burkina Faso, but no data are available on PWID numbers [30]. Monitoring the extent of injecting drug use and HCV transmission among PWID and other key populations (e.g., MSM and sex workers) should be implemented.

The pooled HCV RNA prevalence among low-risk populations was 1.65% (95% CI: 0.74–2.89%) and may be indicative of the prevalence in the general population. By applying this rate to the total population of the country in 2019 [15], approximately 301,174 people are estimated to be active HCV carriers in Burkina Faso. The Polaris observatory estimated the total number of active carriers in Burkina Faso at 247,000 in 2015, corresponding to a viremic prevalence of 1.3% (95% CI: 1.0–1.4%), and higher than the global prevalence of 1% [31]. As HCV infection can be treated with highly effective DAAs, public effort should be undertaken to identify active carriers and link them to care.

Only six studies evaluated HCV genotype distribution, and genotypes 2 and 1 were the most prevalent, as reported by previous publications [4, 12, 32]. Since the advent of pangenotypic DAAs, genotype determination is considered of little interest in deciding the adequate treatment for chronic HCV [2, 6]. However, recent reports of treatment failures with pangenotypic DAAs in patients of African descent infected with genotypes 1 and 4 [33, 34] are alarms for monitoring circulating genotypes and evaluating the effectiveness of HCV treatment.

Our study is limited by the small sample size in rural and high-risk populations and the diversity of anti-HCV antibody measurement methods. Also, as expected from a meta-analysis of prevalence data, significant heterogeneity was observed and could be attributed to the various populations and geographic areas [13]. These factors may have underestimated or overestimated the actual seroprevalence. Despite these limitations, our results are valuable in guiding public health response in a setting where data on HCV infection in specific populations and settings are scarce.

Conclusions

In conclusion, the seroprevalence of HCV infection is intermediate in Burkina Faso, with a decreasing trend over the past 30 years, due to improved blood transfusion and injection safety. There is a paucity of data at the national level and for rural and high-risk populations. The fight against hepatitis C infection requires high-quality and nationally representative data to guide public health response. Although general population testing is recommended, special attention should be paid to rural and high-risk populations, with screening and linkage to care.