Molecular characteristics of Hepatitis B and chronic liver disease in a cohort of HB carriers from Bamako, Mali

Chronic carriage of Hepatitis B Virus (HBV) reaches endemic levels (≥8% of the general populations) in many parts of Sub-Saharan Africa, South East Asia and Latin America. HB chronic infection is the main attributable risk factor worldwide for severe chronic liver disease and, significantly, hepatocellular carcinoma (HCC) [1,2]. In West Africa, where rates of chronic HB carriage in the general, non-vaccinated populations are 10-18%, HCC is the leading cancer in males (standardized incidence rates (ASRW) of 15-30/10⁵ person years) and the third in females (ASRW of 7-15/10⁵ person years) [3]. Most HCC patients are aged 35–50 and present with advanced tumors with an average time since first symptoms of only 2–3 months. Very few patients are diagnosed with cirrhosis before HCC detection. A clinical study in The Gambia has shown that, at HCC diagnosis, ultrasonography reveals concomitant cirrhosis in about 65% of the cases, corresponding in many cases to reactive, rather than precursor cirrhosis [4].

In West Africa, HBV infection mostly occurs perinatally through horizontal or vertical transmission and the rate of carriage by 15 years in non-HB vaccinated subjects is almost equal to adult rates [5,6]. In The Gambia, HB carriage is detected in over 90% of patients with HCC before age 50 [7]. Alcohol is not a significant risk factor at population level. In association with HBV, the other main risk factor for HCC is dietary exposure to aflatoxin-B1 (AFB1), a mycotoxins produced by the fungus Aspergilus sp. which contaminates several staple foods in semi-arid and tropical regions. Exposure to AFB1 induces a “signature” mutation in the tumour suppressor gene TP53 (R249S, AGG to AGT at codon 249, arginine to serine substitution). This mutation has been detected in about 50% of HCC cases in The Gambia [8]. Using a sensitive and quantitative mass spectrometry method, Short-Oligonucleotide Fragment Analysis (SOMA), R249S is also detectable as a marker of current exposure in subjects exposed to AFB1 in the general population [9].

In Mali, studies in specific population groups (pregnant women, students, blood donors) reported a variable rate of HB carriage in adults ranging between 10 and 18% [10-17]. In a cohort of 91 patients receiving treatment for chronic hepatic diseases in hospital Gabriel Touré (Bamako), 54% were found to be HB carriers, compared to 4.3% in a matched group of blood donors used as controls [18]. In the same center a year-long longitudinal study of 57 patients with severe chronic liver disease, jaundice and ascites were observed to be the main clinical signs. In these patients mortality over 1 year was 82.5%, with the three main causes being HCC, abdominal bleeding and hepatic encephalopathy [19]. In this study, we have analyzed serological and molecular parameters of HB carriage in a group of 152 adults unselected for pathological conditions and recruited in outpatients clinics and in several occupational groups. The 152 participants were confirmed positive for HBsAg at Center for infectiology Charles Merieux (CICM) in Bamako, Mali. Carriers were tested for fibrosis status and hepatitis activity by Fibrotest-Actitest and the presence of mutated TP53 R249S DNA, a marker of mutagenesis by aflatoxin, in the serum or plasma.

We have studied a group of volunteers recruited from various groups of the population in the district of Bamako, Mali, for HB carriage (as defined by persistent HBsAg + status) and we have further characterized several molecular, serological and biochemical parameters in the sub-set of those who were confirmed HB carriers. The results show that the rate of HB chronic carriage is high in this study population (18.8%), and that at least 1/3 of the carriers have significant signs of liver fibrosis (moderate to severe Fibrotest scores, ≥F2, 35.4%) or active hepatitis (viral load ≥ 10⁴ copies/mL, 34.2%). Thus, a substantial proportion of these HB carriers with clear signs of active hepatitis might benefit from anti HB-treatment for reducing the risk of progression towards severe liver disease or liver cancer, in particular in a context where HB carriers have detectable mutagenic exposure to aflatoxin [28-30].

The population recruited for this study includes volunteer hospital visitors and members of several occupational groups, unselected for pathology or condition. The focus on such defined groups may cause a selection bias. It should be noted that previous studies in Mali have been conducted on particular groups of high-risk (e.g. patients with liver disease, HIV-positive subjects, [11,18,19]) or low-risk (e.g. blood donors, [10,12,17]) subjects. Several studies have analyzed seroprevalence in pregnant women and in women of childbearing age [13-16]. Studies in blood donors and in healthy women reported a prevalence of HB carriage ranging between 10 and 18%. The proportion of carriers in the present study group (18.8%) is therefore in the high range of these previous estimates and is also among the highest reported in studies in West Africa. Rates between 11 and 17% have been consistently observed in studies performed in The Gambia or in Senegal [4,7,31]. However, most of these studies have been carried out on selected “control” groups and data from The Gambia are for an essentially rural population [4,7].

Of note, childhood HB vaccination has been introduced in the Expanded Program of Immunization (EPI) in Mali in 2002. It is expected that, provided adequate vaccination coverage and efficacy, the rate of HB carriage in the general population will decrease in the next two-three decades [32].

The volunteers recruited in this study were categorized in different sub-groups according to their social/professional status and the context of their recruitment. Rates of HB carriage of 16.8% and 17% were detected in two professional groups, respectively, military staff and staff recruited in a hospital setting (students; nurses and other hospital staff). In contrast, high rates (23%) were observed in non-patient volunteers recruited at outpatient clinics in major hospitals in Bamako city (Hospital Gabriel Touré and Point G), and the difference between the three recruitment groups was found to be significant (P = 0.0021). The social, demographic and health status of subjects in these “non-patient” subjects recruited in a hospital setting is not precisely known. We have noted that this group is composed of a mixed population with generally lower socio-economic status than the two professional groups. Therefore, further studies on the rate of HB carriage in different strata of the Malian population are needed. It is indeed possible that the rate of carriage might be even higher than the average reported here in less-economically favored groups of the population.

The structure of the study was such that subjects who were tested as HBsAg + at initial recruitment (276 positive subjects) were recalled for subsequent blood samples for other analyses at different time point in the study. Only part of the initial HBsAg + group was available for these subsequent analyses, which were thus carried out among subjects from a smaller series of 195 confirmed HB carriers (Figure 1). Thus, we cannot rule out a selection bias due to compliance with recall. However, there is no indication that a particular recruitment sub-group was more represented in the recall group than in the initial recruitment population.

Based on alignment of HBS sequences with reference database sequences, the most common HBV genotype in HB carriers was genotype E, present in 91.1% subjects. A small number of subjects were carriers of either D or A genotypes, and 5 subjects had sequences that matched both genotype E and D. The dominance of genotype E in this population is in agreement with studies in other West African populations. The only subject with A genotype had the A3 variant, a sub-genotype documented in West African populations [33]. The presence of both genotype E and D sequences in about 5.6% of the subjects is suggestive of either co-infection or recombination between these genotypes, an event that has been shown in a small number of infected subjects in different populations of West Africa [33].

Monitoring of HBV viral load is the most widely used method in assessing liver disease severity, predicting cirrhosis and HCC, deciding initiation of antiviral therapy, assessing treatment response and detecting the emergence of drug resistance. In this cohort, 89.5% of HBsAg + subjects (136/152) had detectable HBV viral load, with an average at 3.2 ± 1.7 log₁₀ IU/mL and 32.2% (49/152) had loads over 4 log₁₀ IU/mL, considered as cutoff value for high viral loads. Although there was an overall high correlation between HBe + status and viral load, over 75% of the subjects with viral load ≥4.0 log₁₀ UI/mL were negative for HBe, an observation compatible with previous studies on HB carriers in The Gambia [34]. No previous data is available for chronic HB carriers in Mali. In Senegal, a study on 225 HB carriers recruited in a hospital setting in Senegal (Dakar) has detected an average viral load of 3.6 log₁₀ IU/mL and 22% had viral loads ≥4.2 log₁₀ UI/mL [35].

The use of Fibrotest and Actitest as surrogate endpoint to assess inactive HB carrier status and to predict liver disease-free survival has been evaluated in a cohort of 1,074 patients that includes 47% of patients of African origin [20]. This study showed that subjects with F scores ≤ 2.9 (corresponding to ≤ F1 status) and A scores ≤ 2.7 (corresponding to ≤ A1) had 100% of HB-complication-free survival after 4 years of follow-up, and that F scores outperformed combined classical predictors such as HBe antibody status, low AST activity, low viral load and absence of co-infections with HIV, HCV or HDV. Thus, normal Fibrotest and Actitest values have been proposed as criteria for a new definition of inactive HB carrier status. In the present study, of 100 subjects with complete data for viral load, Fibrotest and Actitest, 35 (35%) matched these criteria of inactive carrier status. Of these subjects, 8 (22.9%) had a viral load ≥ 4 log₁₀ IU/ml. On the other hand, of 147 subjects tested by Fibrotest, 33 and 19 were scored as having moderate or severe fibrosis, respectively, representing together over 1/3 of the study participants (35.4%). There was no significant correlation between F scores and viral load. Regarding Actitest, 15 subjects had A status ≥ 2 (10.2%) and this score was highly correlated with a viral load ≥ 4 log₁₀ IU/mL. Overall, 44% of the total number of subjects tested for viral load, Fibrotest or Actitest was scored as “positive” for at least one of the three tests. These observations suggest that a substantial proportion of HB carriers in the population may benefit from treatment with nucleoside analogs such as Adefovir or Tenofovir [36]. However, there is controversy on whether Fibrotest may represent an accurate method for the diagnosis of fibrosis in West-African HB carriers. A study in Dakar on HB carriers with normal AST levels and high viral load, found discordant results for the prediction of fibrosis using Fibrotest, Fibrometer (another biochemical test) or Fibroscan (liver stiffness measurements) [35]. It should be noted that markers such as Haptoglobin, included in both Fibrotest and Actitest, are strongly expressed in subjects infected by chronic malaria and sickle cell anaemia [37], two prevalent conditions in Bamako. Therefore, further studies taking into account the specific epidemiological context are needed to determine an appropriate algorithm for decision of initiating anti-viral treatment in West African populations.

Levels of TP53 mutation at codon 249 (R249S) in circulating-free DNA from plasma have been proposed as a direct biomarker for the mutagenic effect of aflatoxin in normal subjects exposed through the diet. This mutation is present at high levels (>2500 copies/mL plasma) in liver cancer patients and its presence at such levels is concordant with mutation status in HCC tissues [38]. However, in asymptomatic subjects from The Gambia, low levels of R249S (70–2500 copies/mL) have been reported in subjects from the general rural population, with seasonal variations that may reflect variations in dietary intake of contaminated food [22]. In the present study, we have observed that about 60% of HB carriers from Bamako had R249S levels above threshold values (40 copies/mL), a proportion that is similar to HB carriers in rural Gambia. However, in subjects from Bamako, plasma concentrations (311 ± 340) were lower than in carriers from rural Gambia, in whom mean R249S levels detected at different periods of the year varied between 480 and 5,690 copies/mL serum. Overall, these results suggest that carriers form Bamako may have less exposure to aflatoxin than carriers from rural Gambia. This difference may reflect the combined effect of differences in dietary patterns in urban Bamako and rural Gambia, as well as the effect of interventions aimed at reducing crop contamination by aflatoxin, which are taking place in Mali since the mid 1990’s.

In conclusion, this study provides a detailed assessment of HB chronic infection status in a group of subjects, unselected for clinical symptoms, in Bamako-Mali. Previous studies in Mali and in West Africa have mostly addressed HB status in specific groups such as blood donors [10,12,17], pregnant women [13,16], or patients diagnosed with chronic active hepatitis [35]. The results show that the rate of HB carriage in the population of Bamako is in the higher range of estimates for West Africa, and that about 30% of adult carriers have HB and/or liver biomarkers that would qualify them for treatment with HB anti-viral therapies. On the other hand, mean levels of mutant R249S DNA, a marker of exposure to aflatoxin, are significantly lower than those measured in rural Gambia, where aflatoxin-contaminated crops are an important component of the diet. Future studies are needed to assess importance of these factors for the risk of fatal liver disease and liver cancer in the general population of Bamako/Mali, in a context of epidemiological transition.