Final Report Summary - PROLIFICA (Prevention of Liver Fibrosis and Cancer in Africa)
Executive Summary:
Primary liver cancer (hepatocellular carcinoma: HCC) is the most common malignancy in adult men and the second most common malignancy in adult women in West Africa. Since the survival from diagnosis is less than a year this makes HCC a major cause of premature adult mortality in West Africa. Hepatitis B virus (HBV) is a small, hepatotropic, double-stranded DNA virus causing acute or chronic hepatitis infection. Chronic HBV infection may lead to cirrhosis, liver failure or HCC. Recent estimates suggest that there are over 250 million people worldwide who have chronic HBV infection with high prevalence of infection in sub-Saharan Africa and Asia.
HBV can be prevented with a vaccine based on the envelope protein of the virus and can be treated using nucleoside analogues. Since 1990, under guidance from the World Health Organisation (WHO), the majority of countries have implemented infant vaccination programmes which integrated HBV vaccine with expanded programmes of immunisation. Treatment of HBV with nucleoside analogues suppresses viral replication, prevents disease progression, reverses liver fibrosis and cirrhosis and reduces the risk of HCC. However, there are currently no national treatment programmes for HBV in sub-Saharan Africa.
The principal aims of Prolifica were:
• to identify the main risk factors for HCC in a sub-Saharan African population using a case-control design
• to demonstrate that treatment of chronic HBV infection is feasible and an effective method of reducing the incidence of HCC in resource poor settings.
• to demonstrate that HCC can be detected early and treated effectively
PROLIFICA established two main research platforms:
1. The hepatocellular carcinoma case-control study (HC4) in Gambia, Senegal and Nigeria
2. The West African Treatment Cohort study (WATCH) to identify and treat patients with chronic hepatitis B infection in Gambia and Senegal.
In the HC4 study we collected data and biological samples from over 500 patients with HCC and controls with chronic HBV infection and normal population controls. Using this resource we have identified proteomic and metabonomics profiles which help to identify patients with HCC. These novel biomarkers will be developed into diagnostic assays for future clinical evaluation. We developed a point-of-care test for SCCA-IgM for use in near patient testing for HCC using a lateral flow assay system. The test proved reliable in serum samples wheter the SCCA-IgM was > 200Au/ml.
WATCH is the first study to evaluate the feasibility, effectiveness and cost-effectiveness of screening for HBV in the community and treating those eligible for treatment. We found a high uptake of screening (68.9%) and good linkage into care (81.3%). Amongst patients identified in the community only about 5% met eligibility criteria. The cost of screening was €7.50 per person and the cost effectiveness €500 per Disability Life Year averted.
Data from the WATCH study was used to populate a simulation model of the global HBV epidemic and to estimate the potential impact of various public health interventions. The model demonstrates that mortality due to HBV infection cannot be controlled in the foreseeable future without implementing a screen and treat strategy as illustrated in the WATCH study. In addition mother-to-child transmission of HBV is already becoming the dominant mode of transmission and to prevent new chronic HBV infections in the future will require implementation of methods to prevent this route of infection.
Project Context and Objectives:
Primary liver cancer (hepatocellular carcinoma: HCC) is the most common malignancy in adult men and the second most common malignancy in adult women in West Africa1. Since the survival from diagnosis is less than a year this makes HCC the major cause of premature adult mortality in West Africa. Worldwide HCC is one of the ten most common malignancies and in Europe the incidence of this type of cancer is rising in virtually all member states2. Although any cause of chronic liver disease and cirrhosis may lead to HCC globally the most important causes are hepatitis B virus (HBV) and hepatitis C virus (HCV) infections3. According to WHO estimates 250 million people worldwide are persistently infected with HBV and 180 million with HCV. Together these viruses are responsible for approximately one million deaths annually from HCC or cirrhosis with liver failure.
Fortunately only a minority of people with chronic HBV or HCV infection will develop HCC or liver failure 4. The main risk factor for progression to HCC in both infections appears to be the development of advanced liver fibrosis and cirrhosis 5. Other, established, common risk factors for HCC include, male gender, advanced age, exposure to aflatoxin and having a first degree relative with the disease6 7. In Asia and more recently in Africa high levels of virus in serum and high alanine transaminase (ALT) values have been correlated with progression of chronic HBV infection to HCC8 9. Most of our understanding about the natural history of HBV infection arises from European and Asian studies and a recent review pointed out the paucity of information on the natural history of persistent hepatitis B infection in Africa 10. Although the prevalence of chronic infection in adults is similar in Asia and Africa there are several difference which are likely to lead to variation in natural history of disease. The age at infection, which influences the rate of viral persistence differs, with childhood infection around the age of 2-3 years predominating in Africa whilst some 40% of HBV carriers in Asia result from perinatal infection 11. The virus differs with genotype E predominating in the Gambian population whilst B and C are the most common HBV genotypes in Asia. Co-factors also differ with little or no alcohol consumption in rural Gambia but high levels of aflatoxin exposure which we have recently suggested could play a role in cirrhosis 12. In addition there is of course a different human genetic background.
In West Africa over 60% of HCC cases have chronic HBV infection13. The WHO campaign to include HBV immunisation in the Expanded Programme on Immunisation schedules of all countries in the world is a major success. National surveys have confirmed that this intervention has reduced the prevalence of chronic HBV and, in Taiwan, the incidence of HCC amongst children 14. However, implementing HBV vaccine programmes for children still leaves a legacy of chronically infected people, who acquired the infection prior to the implementation of the programme, which will persist for the next half century. Furthermore, the changing demographics of people in sub-Saharan Africa, with an ageing population means that whilst the incidence of HCC may fall as a result of vaccination the number of cases will actually double by 2050 unless steps are taken to provide treatment for HBV and other aetiological factors.
Treatment is now available for the management of chronic HBV infection in developed countries. Until the 1990s the only treatment option for chronic HBV was interferon alpha which is not appropriate for widespread use in sub-Saharan Africa because serious side effects are common, the cost is high, there is need for intensive monitoring and therapy is contra-indicated in patients with advanced liver fibrosis or cirrhosis. However, treatment may now be delivered using oral nucleoside or nucleotide analogues which inhibit the HBV polymerase/reverse transcriptase enzyme 15. These drugs are administered once daily and are relatively free from side effects. However, as with the treatment of human immunodeficiency virus (HIV), the majority of patients will require indefinite treatment with nucleos(t)ide analogues in order to maintain suppression of viral replication. In a placebo-controlled trial in Asian patients treated with Lamivudine it was shown that suppression of viral replication prevented progression of liver fibrosis and reduced the risk of developing HCC 16. Two of the drugs (Lamivudine and Tenofovir) used to treat HBV infection are also used to treat HIV infection and are available in Africa at affordable prices.
Hepatitis B vaccine coverage in SSA is imperfect 1 and a large number of people born before the introduction of the vaccine continue to carry the virus with a risk of cirrhosis and HCC. 2 In SSA screening and treatment for hepatitis B are rarely accessible3,4 and HCC remain a major cause of death in young individuals.
The main objectives of PROLIFICA were 1) to evaluate whether screening and treatment for HBV infection is feasible and effective at reducing the risk of HCC in West Africa (Gambia & Senegal), 2) to identify biomarkers of cirrhosis and HCC (Gambia, Senegal, Nigeria) and 3) to develop West African research platforms on HBV and HCC and transfer capacity building on HBV infection and liver cancer.
The project was launched in 2011 and ended in January 2016. It has been conducted with the support of eight partners (summarized in Table 1) from UK, France, Gambia, Nigeria and Senegal.
Table 1: Main Partners in the PROLIFICA Project
Partner Institution Lead Contact Key Work Area
Imperial College London, UK Prof. Mark Thursz (MT) Lead project, co-ordinator, metabonomics
International Agency for Research on Cancer (IARC), France Dr Maimuna Mendy (MM) Cancer, virology, biomarkers, proteomics
Institute National de la Sante et de la Recherché Medicale (INSERM), France Prof Fabien Zoulim
Dr Isabelle Chemin (IC) Infectious diseases, HBV virology
Xeptagen Inc. Italy Dr Giorgio Fassina Diagnostic development, biomarkers
MRC Gambia, Gambia Prof. Umberto D’Alesssandro (UDA) Clinical studies
Le Dantec Hospital, Senegal Prof Soulemayne Mboup (SMB) Clinical studies
Royal Victoria Teaching Hospital Dr Ignatius Baldeh (IB) Clinical studies
Jos University Teaching Hospital (JUTH), Jos, Nigeria Prof Edith Okeke (EO) Clinical studies
The project has been conducted within 7 workpackages (WP) summarized in Table 2:
Table 2: PROLIFICA Workpackages
Workpackage Lead unit/person Description
WP1 Imperial College (MT) Coordination & management
WP2
HCC case study Imperial College (MM) HCC control study dataset
WP3
Anti- viral treatment cohort (WATCH) study Imperial College (MT), MRC (UDA) and Senegal (SMB) Treatment study
WP4
Biomarker discovery and mechanistic investigations through proteomics
Imperial College (STR)
IARC (MM) Novel markers for liver cancer validation
WP5
Development of metabonomic markers for the diagnosis of HCC Imperial College (STR)
IARC (MM) Confirmed list of biomarkers
WP6 INSERM, Lyon, France (IC) HBV molecular and serodiagnostics
WP7 All partners Capacity building
As lead Partner, the main role of Imperial College has been to act as the overall administrative co-ordinator of the project to ensure efficient collaboration, exchange of information and knowledge and fulfilment of ethical, regulatory and bio-safety rules. In addition, Imperial monitors the progress of the project, anticipating possible hurdles and working with the partners to find solutions. Specifically this includes detailing the overall work programme and budget, co-ordinating and monitoring scientific and financial reports, collecting and compiling results and organising project meetings.
In addition, Imperial has been also involved in several scientific aspects of the project, in collaboration with other partners. These include establishing whether the treatment of HBV is an effective intervention to control HBV replication, liver progression and prevent development of liver cancer (work package 3, training radiographers in the use of ultrasound for the early diagnosis of liver cancer and Fibroscan to assess the stage of HBV-related liver disease (WP3), the development of metabonomic biomarkers for the diagnosis of HCC and capacity building (WP7).
The three African partners were primarily responsible for the clinical studies: HCC patients (WP2) and screening and treatment cohort study (WATCH study WP3). They also had key roles in the HBV and hepatitis C (HCV) infection studies (WP6) that was primarily the responsibility of INSERM. C
The role of hepatitis virus infection in the pathogenesis of HCC is complicated by viral diversity, different levels of viral replication and the presence of HBV mutants that are capable of modulating the carcinogenesis process. The main goal of WP6 was to provide information on the contribution of HBV genotype and HBV diversity in the development of HCC in The Gambia, Senegal and Nigeria and generate full genome sequences of the different genotypes circulating in the three countries.
IARC, Xeptagen and Imperial College have been the main partners responsible for investigating the possibility of identifying biomarkers for both HCC and HBV in biological samples taken from patients in both the case control and treatment studies. In addition they will performe extensive analysis of the generated sequence alignment to identify viral genetics patterns and investigate their association with liver disease. They will also investigate infection with hepatitis C (HCV) as a possible risk factor of HCC. IARC and Imperial used proteomics and metabonomics respectively to identify, evaluate and validate biomarkers for HCC.
All Partners have been involved in capacity building and dissemination of the project-related technology and knowledge on Hepatitis B, liver cancer, screening, prevention, detection and treatment (WP7) throughout the Partnership, but especially for West Africa. In particular, the aim was to create reference centres for hepatology and liver cancer in the three West African countries involved in the project
Project Results:
3.1 Coordination and management of the consortium (WP1)
In order to achieve its objectives, the PROLIFICA programme has developed two platforms of clinical and scientific research on viral hepatitis and liver cancer in The Gambia, Senegal and Nigeria: the West African Treatment Cohort for Hepatitis B (WATCH) and the Hepatocellular Carcinoma Case-Control Study (HC4). Consortium management coordinated the transfer of data and samples from these platforms to the other workpackage teams.
3.2 HCC case-control study (HC4) in The Gambia, Senegal and Nigeria (WP2)
A total of 630 patients with suspicion of end-stage liver disease (cirrhosis, decompensated cirrhosis, and/or HCC) have been enrolled in Prolifica: 164 patients including 61 patients with HCC have been enrolled in Senegal; 466 patients including 50% with HCC in Gambia.
The analysis of the epidemiological, clinical and virological characteritics of the complete study population is under progress.
However, preliminary analysis of 69 HCC patients shows that 49 (72%) are males with a median age at 43.5 years (22-80). At time of diagnosis, the median number of tumoral lesions is 3 (1-10) and the median size of the tumor size is 12.25 cm (1.5-32): 61 (88.4%) have a mean size ≥ 5 cm at time of diagnosis and 21 (34%) have ascites. HBV is the main cause of HCC (66%) with a median viral load at 43,760 IU/mL (<50-3.109). HDV coinfection is observed in 6%, HIV infection in 9% and 6% of the HCC cases is attributable to HCV infection.
Median AFP levels was 909 (1-35,350) and 19 (27.5%) had AFP levels < 200 ng/mL (Table below)
Table 3: Characteristics of HCC patients in the Gambia
Median length of survival in those who died was 61 days (4-789 days).Cause of death was determined in only 25 patients with haematemesis being the main important reported cause (44%) (see figure 3 below).
Figure 1: Kaplan-Meier survival estimates of Gambian patients with HCC
We also analysed the survival of compensated and decompensated cirrhotic patients. (Shimakawa Y et al. APT 2015 5 ) see figure 4 below. The cumulative mortality rate at 28-day was 70.0% (21/30) in the cases and 13.1% (8/61) in the controls (P < 0.001). The cumulative mortality at 90-day was 92.9% (26/28) in the cases and 28.1% (16/57) in the control group (P < 0.001). Positive anti-HEV IgG was not associated with higher 28-day mortality in cirrhotic patients. After adjusting for age, sex, and case/control status, none of the potential precipitating events of ACLF were associated with higher 28-day mortality. The causes of death in patients who died within 90 days after enrolment were as follows: end-stage liver disease (n=18, 69.2%), heart failure (n=1, 3.9%) and unknown (n=7, 26.9%) in the ACLF cases and end-stage liver disease (n=10, 62.5%), breast cancer (n=1, 6.2%) and unknown (n=5, 31.3%) in the control group.
Figure 2: Survival of patients with compensated or decompensated cirrhosis
As part as the PROLIFICA programme we also analysed whether hepatitis E infection, which is common infection in resource-poor setting, is prevalent in Gambia and is a cause of hepatic decompensation in this country. In 204 healthy volunteers, 13.7% (95% CI: 9.6-19.2) were tested positive for anti-HEV IgG, and none had positive HEV viremia. After adjusting for age and sex, the following were associated with positive anti-HEV IgG: being a Christian, a farmer, drinking water from wells, handling pigs and eating pork. In 40 cases (median age: 45 years, 72.5% male) and 71 controls (39 years, 74.6% male), ≥70% were infected with hepatitis B virus. Although hepatitis B flare and sepsis were important precipitating events of ACLF, none had marker of acute HEV. These results have published in 2016 in Alimentary Pharmacoly and Therapeutics as follows
As part of the PROLIFICA programme, using the WATCH and HC4 studies, we analysed whether birth order is a risk factor of HCC as it has been suggested in Asian studies. Indeed, early age at infection with HBV increases the risk of chronic infection. Moreover, early HBV infection may further independently increase the risk of hepatocellular carcinoma (HCC) beyond its effect on chronicity.The distribution of birth order, a proxy for mode and timing of HBV transmission, was compared in The Gambia between hepatitis B surface antigen (HBsAg)-positive HCC cases recruited from hospitals (n =72) and two HBsAg positive control groups without HCC: population-based controls from a community HBV screening (n = 392) and hospital based controls (n = 63). Results: HCC risk decreased with increasing birth order in the population-based case–control analysis. Using first birth order as the reference, the odds ratios were 0.52 (95% CI: 0.20–1.36) 0.52 (0.17–1.56) 0.57 (0.16–2.05) and 0.14 (0.03–0.64) for second, third, fourth and greater than fourth birth order respectively (P = 0.01). A similarinverse association was observed in the hospital-based case–control comparison (P = 0.04). Conclusions: Compared to controls, HCC cases had earlier birth order, a proxy for young maternal age and maternal HBV viraemia at birth. This finding suggests that in chronic HBV carriers perinatal mother-to-infant transmission may increase HCC risk more than horizontal transmission. Providing HBV vaccine within 24 hours of birth to interrupt perinatal transmission might reduce the incidence of HCC in The Gambia. These results have been published in Liver international, 2015.This work was part of the PhD of Yusuke Shimakawa who has been associated to the PROLIFICA project.
3.3 Anti-viral treatment - The West African Treatment Cohort (WATCH) study (WP3)
This “screen and treat” intervention was set up in The Gambia and Senegal and has been successfully completed. Using a rapid point-of-care test for HBs antigen, a total of 16,310 subjects have been screened over both countries (in domestic places, working places, blood banks and facilities). Out of them 2,132 individuals were tested positive giving a total prevalence at 13%. After excluding, the facility screening which might overestimate the HBV prevalence, the proportion of HBV-infected subjects remained high at 10.3% (1,569/15,199).
Each HBV-infected individual was invited to see a clinician trained in Hepatology for standardised comprehensive liver assessment: physical examination, abdominal ultrasound (Portable MyLab25Gold, Esaote, Cambridge, UK), fasting liver stiffness measurement (LSM) using hepatic transient elastography (Fibroscan®, FS402, Echosens®, France), and routine serum hematology and biochemistry tests. Blood samples were tested for hepatitis B e antigen (HBeAg) (ELISA-ETI-EBK Plus, Diasorin, Italy), antibodies to hepatitis C (AxSYM, anti-HCV, Abbott, USA) and hepatitis D viruses (ETI-AB-DELTAK-2, Diasorin, Italy). Antibodies to HIV-1 and HIV-2 were detected using enzyme immunoassay (EIA, Genscreen ULTRA HIV Ag-Ab, Bio-Rad, USA).
HBV DNA levels were measured using an in-house quantitative real-time polymerase chain reaction (q-PCR) (detection limit of 50 IU/ml) which was validated against a commercial HBV qPCR (Abbott, USA, excellent correlation with the commercial assay (r 2= 0.90)) (publication under revision). All samples were tested at the MRC Unit in Fajara. Quality control and HBV genotyping were performed by a reference laboratory in France (INSERM, Lyon). Detailed results per country are provided below.
3.3.1 The “screen and treat” intervention in The Gambia (lead: MRC The Gambia Unit, Prof. UDA)
Screening was conducted in Western Gambia where 750,000 people live in 1,450 census enumeration areas (EAs), defined by the Gambia Bureau of Statistics. We used EA as a sampling unit, and one EA consists of an entire village, a part of large village/town, or a cluster of small hamlets. Because HBV prevalence may differ between urban and rural populations, we first stratified 1,450 EAs into urban (n=1,197) and rural (n=253) communities. Then, from each stratum 27 EAs were selected by simple random sampling using a random number generator. In the selected EAs, all inhabitants aged ≥30 years were eligible for screening. We excluded those aged <30 years because the national hepatitis B vaccination programme started in 1990.6 We organised a meeting in each EA, with the help of the village or city head. Following community approval, a team of fieldworkers conducted a census by visiting all households to register the name, age and sex of all eligible people and invited them for screening. After pre-test counseling and written consent, finger-prick whole blood was tested for HBsAg using a POC test (Determine®, Alere, USA) whose performance was validated in the field (sensitivity 88.5%, specificity 100%).7 Results were provided on site to the participants with post-test counseling, and those who tested positive were referred to the closest liver clinic.
This work has been done with the support of the MRC unit but also the NPHL, Ministry of Health Gambia (including EFSTH, formerly RVTH) who provided two excellent fieldworkers: Alagie Sanneh and Debo Jallow have been consistently involved in the field work activity of the project. Their participation to each event has been smoothly facilitated by management from NPHL. They have had a significant impact on the sensitisation, screening and recruitment of patients from communities and the clinics. We also received a strong support from Dr Makie Taal (ministry of health) who facilitated the relation with EFSTH and its main blood bank.
Eligibility for treatment was determined according to the 2012 European Association for the Study of the Liver (EASL) guidelines 8. In the absence of contra-indications, tenofovir was provided free of charge (provided by Gilead Company). Adherence to treatment was assessed using the Morisky adherence scale.
Between December 2011 and January 2014, all selected EAs agreed to participate to the study and 5,980 out of 8,170 eligible people (68.9%, 95% CI: 65.0-72.4%) participated in the community screening. Between January and December 2013, 5,559 potential blood donors were screened (99.3% men) for HBs Ag in the main blood bank in The Gambia (Banjul).The prevalence of HBsAg was 8.8%, (95% CI: 7.9-9.7%) and 13.0% (95% CI: 12.1-13.9%) in the Gambian community and the main Gambian blood bank, respectively.
Linkage to care, which was defined as first attendance to the liver clinic was high, with 402 of 495 (81.3%) HBsAg-positive people from the community attending clinic; however, 300 out of 721 (41.6%) of HBsAg positive blood donors linked into care. Treatment indication was met by 18 of 402 (4.4%, 95% CI: 2.5-7.7%) patients from the communities and 29 (9.7%, 95% CI: 6.8-13.6%) from the blood donors. Male gender was strongly associated with treatment eligibility (OR: 4.35 95% CI: 1.50-12.58 P=0.007).
Figure 3: Flow chart of the screen and treat intervention in The Gambia
The majority (87·9%, 617/702, 95% CI: 85.3-90.1%) of the HBsAg-positive individuals who attended the clinic after screening was classified as inactive chronic carriers (Table 1 below); 3.3% (13/395) and 7.9% (23/291) were HBeAg-positive from the community and blood bank, respectively. From the community screening, 48 (12.2%) individuals had elevated ALT ≥40 IU/mL and 41 (10.7%) had HBV DNA ≥ 2,000 IU/ml. After excluding eleven participants without valid LSM, 10 (2.6%) had extensive fibrosis (F3) and 11 (2.9%) had cirrhosis (F4). Co-infection with HIV, HCV or HDV was observed in 3.3%, 1.0% and 2.0% of the participants, respectively. Among the HBV-infected blood donors, 55 (18.8%) had ALT ≥40 IU/mL, 38 (14.4%) had HBV DNA≥ 2,000 IU/ml and 50 (17.5%) had extensive fibrosis/cirrhosis (≥F3). The Table below summarizes the clinical and biological characteristics of the HBV-infected study population in The Gambia.
Table 4: Clinical and Biological characteristics of the HBV infected study population in The Gambia
These results have been submitted to Lancet Global Health and are under revision for publication.
3.3.2 The Keneba-Manduar longitudinal cohort of HBV-infected subjects in The Gambia
The Medical Research Council (MRC), the International Agency for Research on Cancer (IARC/WHO) and the Gambia Government have been supporting studies on HBV infection in The Gambia since the 1980’s, and have established a population-based open cohort of treatment-naïve chronic HBV carriers. As part of the PROLIFICA project, we analysed this Gambian cohort to describe i) the sero-clearance rates of hepatitis B e antigen (HBeAg) and surface antigen (HBsAg) in West Africa; ii) the incidence of HCC, end-stage liver disease (ESLD) and all-cause mortality; iii) longitudinal changes in serum HBV DNA and alanine transaminase (ALT) levels ; and iv) the prevalence of significant liver fibrosis and chronic liver disease requiring antiviral therapy according to the European Association for the Study of the Liver (EASL) criteria. We also estimated the HBV-related disease burden attributable to the mother-to-infant transmission in SSA by examining the associations between these outcomes and maternal HBsAg, a proxy for mother-to-infant HBV transmission. 405 chronic carriers (95% genotype E), recruited at a median age of 10.8 years, were followed for a median length of 28.4 years. Annually, 7.4% (95% CI: 6.3-8.8%) cleared HBeAg and 1.0% (0.8-1.2%) cleared HBsAg. The incidence of HCC was 55.5/100,000 carrier-years (95% CI: 24.9-123.5). In the 2012-2013 survey (n=301), 5.6% (95% CI: 3.4-9.1%) had significant liver fibrosis. HBV genotype A (versus E), chronic aflatoxin B1 exposure, and an HBsAg-positive mother, a proxy for mother-to-infant transmission, were risk factors for of liver fibrosis. Interestingly, 16.0% of chronic carriers were infected via mother-to-infant transmission, and 63.0% (95% CI: 47.0-74.1%) of the cases requiring antiviral therapy were attributable to this mode of transmission. These data are unique in Sub-Saharan Africa and have been published in Gut in 2015 9.
3.3.3 Development and validation of diagnostic tools for HBV-infection adapted to the African setting.
HBsAg rapid point-of-care tests
In order to ensure the quality of our screening strategy, we assessed the performances of different rapid tests in the real life African field conditions in The Gambia within the PROLIFICA programme 7. We confirmed very good performances of the tests (sensitivity around 90% and specificity 100%) in the field. Importantly, as part of this validation study, we found that participants with false negative results were all inactive chronic carriers (i.e not in need of treatment) with low levels of HBsAg. This means that we might have slightly underestimated the prevalence of HBV in our study population, BUT it is unlikely that we missed any patient in need of treatment since inactive chronic carriers do not fulfil the treatment eligibility criteria.
Non-invasive markers of liver fibrosis in West African HBV-infected subjects
The lack of non-invasive and inexpensive markers for the assessment of chronic hepatitis B infection is one of the main barriers for offering care and treatment of HBV chronically infected subjects in Africa. Only two studies assessed the diagnostic performances of non-invasive markers of fibrosis in sub-Saharan Africa with discrepant results. New simple and inexpensive markers of fibrosis for African resource-poor countries are highly needed. Therefore, within the PROLIFICA project, and since we used this device to assess the liver disease in our study population, we first confirmed the excellent diagnostic accuracy of transient elastography (Fibroscan®) for predicting severe fibrosis and cirrhosis in HBV-infected patients in a West African setting (AUROCs around to 0.90). However, the Fibroscan device remains expensive, thus we developed and validated a simple and inexpensive biomarker of liver fibrosis: the Gamma-glutamyl transpeptidase to Platelet ratio: GPR (GGT/Platelets ratio) index which might be a better routine laboratory test than APRI (currently recommend by the World Health Organization) and Fib-4 tests.
In-house HBV qPCR (WP3 & WP6)
Before starting the PROLIFICA intervention programme, we validated the performance of an in house technique which was assessed using coefficient of variation for repeatability and reproducibility with 2 levels of concentration, accuracy (A), uncertainty (U) and contamination. Plasma from HBV infected patient were then quantified by the in-house technique and the results obtained were compared. The in-house results were also anlaysed using a commercial method (m2000 sp/rt, Abbott Laboratories) for comparison. The correlation (r2) and concordance (biais) of in-house method and m2000 one were determinated. This in-house HBV qPCR against the Abbott real time commercial assay and we found an excellent correlation coefficient at r^2=0.90. This in-house assay had a lower limit of sensitivity of 50 IU/ml which is comparable with commercially available HBV DNA assays based on real-time PCR analysis. We also found that the repeatability for both concentration tested, coefficient of variation (CV) respectively 2.74 % for the intermediate concentration and 1.28% for the high concentration. For the reproducibility (evaluated after 10 days), CV showed 2.44% for the intermediate concentration and 1.37 % for the high concentration. The accuracy and uncertainty were A=2.79 %; U = ± 0. 25 and A= 0.86 %; U = ± 0, 27 for the high concentration. Contamination rate was 1.28 %. The comparison with the m2000sp/rt for 68 samples covering the clinical interest area, showed r2=0.904 for correlation and a biais = -0,259 for concordance. Since our study, therefore, showed a relative good performance for viral load quantification using SYBRGreen dye, the monitoring of patient was done using this method.
This work has been submitted to Journal of Viral Hepatitis (under revision).
3.3.3 The “screen and treat” intervention in Senegal (WP3, lead: Hopital Le Dantec, Dakar, Senegal)
In Senegal, the screening intervention was performed in domestic communities (urban and rural areas) around Thies the second most important Senegalese city with the support of the Medical school of Thies (Prof. Mourtalla Ka). Screening for HBsAg was also performed in work places (Mines) and in facilities (see Table 2 below).
A total of 4,774 persons have been screened, including 3,660 persons from the community (domestic and work places).
Table 5: Screening results in Senegal
To date, the clinical and virological assessment of more than 700 HBV-infected subjects has been performed. Data are under analysis and will be presented as an oral communication in the next Francophone conference of HIV and Viral hepatitis in resource-limited countries (AFRAVIH, Brussels); the publication of the Senegalese “screen & treat” intervention is under preparation.
In the observation cohort (those who are HBsAg + and who did not meet EASL criteria for treatment), preliminary results (n= 138) showed a prevalence of HBeAg of less than 1% (1/138) and 99% of anti-HBe Ab. AgHBs quantification mean value was 3 times less (1465 UI [ 0.050-14 936]) than those under treatment. Delta co-infection in this group was 2.8 % (4/145).
Figure 4: Proportion of HBV infected subjects with low, medium or high viral load
3.3.5 Treated cohort in Gambia and Senegal (lead MRC Gambia and Le Dantec)
The treated cohort is derived from the community screening (WATCH) and the end-stage liver disease study (HC4 study). A cohort of 242 patients under tenofovir has been set up in both countries.
In Gambia, 127 patients are currently under tenfovir.
None of the eligible patients refused antiviral therapy. Twelve months after starting tenofovir therapy, 38 patients (80.9%) had a high adherence score, 7 (14.9%) a medium adherence and 2 (4.3%) a low adherence. At 12 month of treatment, 43/47 (91.5%) achieved a virological response defined by undetectable HBV viral load; 38 (79.7%) had normal transaminases (< 40 IU/mL), and 9 (19%) had ALT≥40 IU/mL (median 46 (43-62) IU/mL) with a baseline median (IQR) ALT at 49 (34-104) IU/mL. No clinical or biological adverse events were observed after 12 months of treatment.
In Senegal, 115 ? patients are under treatment, the first patient under TDF was recruited in April 2013. We monitored HBV viral load replication during TDF treatment using in-house qPCR. Our first results showed at J0 a median viral load at 6.58 log copies/ml. After only 3 months under treatment, the initial viral load decreased to 3.85 log copies/ml. After 1 year under TDF, 89 % (39/44) of patients tested had undetectable viral load. For those who have detectable viral load 11 % (5/44), 60% over them (3/5) have a viral load of less than 300 copies/ml or 2.47 log copies/mL (≈ 80 UI) and the most elevated value was 1200 copies/ml or 3.07 log copies/ml (≈ 300 UI).
Figure 5: Viral kinetic of HBV DNA in the Senegalese patients on tenofovir
3.4 Biomarker discovery and mechanistic investigation through proteomics (WP4)
3.4.1 Point of Care testing device for monitoring patients at risk of HCC based on a pre-existing dataset of candidate protein markers (WP4, lead: Xeptagen company)
Within the PROLIFICA consortium, XEPTAGEN has developed a new Point of Care Testing (POCT) device for enabling the surveillance of patients at risk of hepatocellular carcinoma (HCC) development based on the assessment of circulating levels of Squamous Cell Carcinoma Antigen (SCCA)-IgM immune complexes.
SCCA-IgM has been proposed as a novel biomarker for surveillance of patients at risk of HCC growth since the increase above 200 AU/mL of serum levels of SCCA-IgM may indicate a worsening of the liver disease in patients with cirrhosis, with a higher risk of developing HCC. Occurrence of SCCA-IgM at high levels (>500 AU/mL) enables the identification of subjects at higher risk of HCC development at least one year before the clinical evidence of tumor thus allowing a serological HCC surveillance.
The POCT has been developed as a lateral flow assay (LFA) for the detection of SCCA-IgM by gold nanoparticles- conjugated antibodies. The LFA has been made up of four membranes: sample pad, conjugate pad, nitrocellulose membrane and absorbent pad. All membranes are partially overlapped and assembled in a LFA plastic cassette. Cassettes have a hole above the sample pad for sample of serum or blood introduction. The prototype LFA has been validated with more than 300 samples demonstrating its diagnostic accuracy in detecting patients with SCCA-IgM > 200 AU/mL.
The adoption of the assessment of SCCA-IgM immune complexes using a low cost disposable LFA POCT for surveillance purposes may represent a significant step forward in the management of patients suffering from chronic liver diseases in low income countries.
3.4.2 Discovery & Validation Strategies (WP4 lead IARC)
Following the initial discovery strategies conducted at the Fred Hutchinson Cancer Research Centre and the University of Oxford in collaboration with other sites; a select group of well performing proteins have been entered into various stages of validation by measurement in liver disease cases from numerous populations.
Two independent biomarker discovery strategies were conducted, both of which are presented in full in the previous report. To summarize, both strategies involved a comprehensive deep proteome analysis of pooled plasma samples put through a pipeline of immunodepletion, reverse-phase chromatography and gel electrophoresis in one strategy and isoelectric focussing in the other with final analysis performed by tandem mass spectrometry. Robust data mining strategies were then employed to generate a priority list of shortlisted biomarkers based on the comparison of controls and case groupings, probability of association with a specific pathological status, the nature and abundance of target proteins, suspected biological relevance and finally the availability of affordable and robust methods for further assessment .
Following this multi-dimentional discovery approach, the proposed markers have been amalgamated and prioritized for measurement in distinct populations mainly through antibody detection methods (ELISA). Robust statistical methods have also been implemented, allowing the use of individual subject measurements to assess and compare expression trends through various disease stages and populations.
Extensive validation
Two independent validation strategies were executed on proteins identified from the biomarker discovery pipelines previously discussed. The primary strategy was aimed at comparing differential protein expression trends with established liver disease markers to assess significance and reproducibility of discriminatory expression patterns suggested by mass spectrometry.
Previous reports have detailed the validation of these markers in populations from the Case-Control Gambia Liver Cancer Study (GLCS) and Jos University Teaching Hospital (JUTH), neither of which were part of PROLIFICA. Comprehensive assays have also been run and reported on PROLIFICA Gambia patients as presented below.
Independent population validation
The proteins proposed from the GLCS discovery track were primarily validated in an independent cohort of African subjects collected from the Jos University Teaching Hospital (JUTH) in Nigeria. In total, plasma aliquots for 55 subjects classified as either Nigerian healthy controls (NN), Nigerian Asymptomatic Carriers (NASC), Nigerian cirrhotics (NCirr) or Nigerian HCCs (NHCC). The categorizations were based on tests for HBV, HCV and HIV virus status, AFP, biochemical liver indices, US, or Computed Tomography (CT), endoscopy and or liver biopsy. Table 6 summarizes the key results from the Jos participants; a full evaluation of these findings and those from the GLCS subjects is made in the published manuscript 10.
Table 6: Summary of sample numbers tested per protein along with AUC’s, confidence intervals and associated p-values for JUTH subjects
Protein & Sample Numbers Tested NN vs. NHCC NN vs. NCirr NCirr vs. NHCC
ApoA1 NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.7667 (95% CI: 0.6020 to0.9313)
p=0.01800 AUC: 0.6167 (95% CI: 0.3154 to 0.9179)
p=0.4477 AUC: 0.6587 (95% CI: 0.3987 to 0.9187) p=0.2435
AAT NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.7857 (95% CI: 0.6039 to 0.9676) p=0.01126 AUC: 0.5833 (CI: 0.2647 to 0.9019)
p=0.5876 AUC: 0.8016 (95% CI: 0.6360 to 0.9672) p=0.02672
CC3 NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.6238 (95% CI: 0.4200 to 0.8277) p=0.2720 AUC: 0.8167 (95% CI: 0.5921 to 1.042) p=0.03937 AUC: 0.8571 (95% CI: 0.6985 to 1.016) p=0.008705
HPX NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.8048 (95% CI: 0.6455 to 0.9640) p=0.006864 AUC: 1.000 (95% CI: 1.000 to 1.000) p=0.001146 AUC: 0.7937 (95% CI: 0.5594 to 1.028) p=0.03098
With this knowledge in hand, two of the best performing markers proposed by each of the groups involved in the discovery, validation and statistical combinations were selected for direct measurement by ELISA in the large, multi-country samples being collected under PROLIFICA. The four most discriminant proteins selected for measurement were HPX, Apo A1, OPN & LTBP2. The latter two proteins were identified within a sister study performed on Gambian, Thai and French participants 11.
The diagnostic criteria for various stages of liver disease utilized were as follows:
1. Controls
• Healthy Controls: HBV & HIV negative
2. Cases: Taken from HC4 Case-Control arm with cases of the following profiles:
• Fibrosis: Liver stiffness measurement (LSM) ≥7 kPa by Fibroscan
• Chronic HBV: HBV positive, no signs of progressive liver disease by US/biochemistry
• LC: Presence of either histological confirmation, LSM ≥11 by Fibroscan, or ultrasonographic diagnosis of liver cirrhosis
HCC
• Clinical, No Histology: Focal liver lesion ≥2cm which is consistent with HCC by US and AFP levels ≥ 400 ng/ml in patients with liver cirrhosis
• Histologically Confirmed: Combination of Ultrasound, AFP, Clinical Indices & a liver biopsy confirmation.
ALL HIV positive cases were excluded.
Samples from Nigeria (JUTH)
At the current phase of our proteomics studies, a total of 238 case and control samples were selected from the HC4 arm recruited in Nigeria. A summary of the clinical parameters for these patients is presented in Table 7. A total of 238 samples have so far been measured for OPN and LTBP2 with assays for ApoA1 and HPX are to be performed shortly. The validation experiments were conducted by ELISA based antibody detection methods.
Table 7: Summary of key clinical parameters for JUTH PROLIFICA subjects
In brief, antibodies from the protein of interest were adsorbed to the base of a microtitre well plate, reacting with and binding specific antigens in the diluted plasma. Any unbound protein was removed by four consecutive wash steps followed by incubation with Horseradish Peroxidase (HRP) conjugated antibodies to the protein of interest. This was followed by a final quadruple wash step, after which the chromomeric substrate 3,3’5,5’-tetramethylbenzidine (TMB) was added and allowed to develop to optimal intensity, for approximately 30 minutes and the reaction then stopped by addition of an acid solution. In their specific measurements, samples for OPN were diluted 1:100 and LTBP2 1:5.
Extrapolation and data analyses were conducted using the GraphPad Prism software. Key conclusions from our results thus far show that LTBP2 may be differentially expressed between various stages of CLD; however it appears to not be a robust enough marker to withstand heterogeneity in case definitions and aetiology of liver disease (Figure 6 & Figure 7).
Figure 6: Dot plot with accompanying table showing summary of LTBP2 expression levels across control and case groups in PROLIFICA Gambia participants.
Figure 7: Dot plot with accompanying table showing summary of LTBP2 expression levels across control and case groups in PROLIFICA Gambia participants.
Osteopontin (figure 2a, 2b) has been consistent in all its measurements in numerous populations as being over-expressed in HCC tumorigenesis. This is a marker which warrants further study into the specific mechanisms by which it is elevated and how specific these processes are to HCC.
Figure 8: Dot plot showing summary of OPN expression levels across control and case groups in PROLIFICA Gambia participants.
Figure 9: Dot plot with accompanying table showing summary of OPN expression levels across control and case groups in PROLIFICA Nigeria participants.
Compilation of Serological, Biochemical and Results
To enable the selection of well described and clinically distinct cases and controls in functional biomarker studies, it is important that key assay results are reported for selected patient populations. This has been a challenging area and one where the team within this
Conclusion
IARC have successfully set up the main project biobank, which will be available to the Partners post the current project completion. In addition, they have identified a selection of well described and clinically distinct cases potential functional biomarkers from their studies. In particular OPN and Osteopontin merit further investigation.
3.5 Development of metabonomic markers for the diagnosis of HCC (WP5)
This work has been conducted in collaboration with IARC, France.
Urine and blood (serum) samples taken from patients in the West African clinical studies are periodically transferred from source to either a specific PROLIFICA Biobank at IARC or to Imperial College for metabonomic biomarker studies. To date, samples from both Gambia and Nigeria have been sent to Imperial College, where they have been processed for the identification of liver cancer biomarkers.
The ultimate goal of this study is the development of a urinary-based test for use at the village level in Africa for screening of at-risk populations.
Urine samples from Nigeria and Gambia have been sent to Imperial College for metabonomic studies, which enabled the detection and identification of metabolites responsible for mediating the phenotypic expression of altered metabolism. Metabolic profiling involves the combination of spectroscopic techniques of analysis, nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS), with advance multivariate statistical tools to interrogate the spectral data produced (Figures 5 & 6).
Figure 10: Metabonomic study workflow
As explained in the previous report, 1H NMR was used as a first stage metabolic screening tool to metabolically phenotype 290 subjects (comprising patients with HCC, cirrhosis (Cir), non-cirrhotic liver disease (DC) and healthy controls (NC)). Urine samples from a further cohort of 463 subjects were analysed, the results of which validated the initial cohort. These analyses were completed by a Nigerian medical PhD student, Dr Nimzing Ladep, studying at Imperial.
Figure 11: Training and validation sets of urine samples analysed by 1H-NMR
The urinary metabotype of patients with HCC was distinct from those with DC and NC. The lower performance of HCC versus Cir was not unexpected, because most HCC cases occur in the background of cirrhosis. The intensities (relative concentrations) of several metabolites were significantly increased in urine of participants with HCC, compared to controls (Fig.3).
Figure 12: Urinary metabolites in HCC shown in a pair-wise comparison of A) HCC vs. Healthy, B) HCC vs. Liver disease; and C) HCC vs. Cirrhosis in both training and validation sets of experiments. Metabolites in bold texts were significantly up regulated in HCC while those in italics were significantly down regulated in HCC. Those represented by regular texts did not attain any significant association.
The diagnostic performance of a panel for HCC versus cirrhosis, comprising four metabolites (inosine, indole-3-acetate, N-acetylated amino acid, and galactose) was significantly higher than serum AFP, a finding that was corroborated in a validation cohort (Table 6).
Table 8: Diagnostic accuracies of urinary models compared to serum AFP in the discrimination of HCC from three control groups in the training and validation sets of subjects recruited in Nigeria and Gambia
Metabolites that contributed most strongly in discriminating between groups were subjected to further analyses. Specifically, liver disease controls were categorized to one (1), intermediate HCC (BCLC A-C) to two (2), and advanced HCC (BCLC D) to three (3), and the scores were regressed against the spectral profiles. The relative concentration of methionine, acetylcarnitine, indole-3-acetate, N-acetylated amino acid, dimethylglycine, 1-methylnicotinamide, and creatine were significantly positively correlated to category and clinical stage of HCC and are shown in figure 4. In comparison, there was poor correlation with stage of HCC by serum AFP (P50.7).
Figure 13: Correlation of relative concentration of urinary metabolites by disease category and BCLC stage of HCC
This study provides some evidence that a battery of discriminatory metabolites may be a future diagnostic accuracy that would be simple to perform and a paper was published last year on these data: Ladep NG, Dona AC, Lewis MR, Crossey MM, Lemoine M, Okeke E, Shimakawa Y, Duguru M, Njai HF, Fye HK, Taal M, Chetwood J, Kasstan B, Khan SA, Garside DA, Wijeyesekera A, Thillainayagam AV, Banwat E, Thursz MR, Nicholson JK, Njie R, Holmes E, Taylor-Robinson SD. “Discovery and validation of urinary metabotypes for the diagnosis of hepatocellular carcinoma (HCC) in West Africans”. Hepatology. 2014. doi: 10.1002/hep.27264
Complementary and more sensitive mass spectrometry (MS) methods followed up NMR analysis to increase the metabolome coverage and for practical reasons: NMR is a preserve of developed world research institutes whereas installing and running MS is cheaper and panels of validated metabolites could be translated into the clinical arena in the developing world. In this regard, the PROLIFICA project has obtained a triple quadrupole (TQ) mass spectrometer for targeted analyses, based at the MRC, Gambia, for future analysis of samples across West Africa.
Choosing which metabolites to analyse in the targeted MS assay is a process driven by hypotheses formed from previous experimental evidence. Untargeted metabolomics experiments are performed using high resolution accurate instrument such as quadrupole time of flight (Q-ToF) mass spectrometers. Due to the complex nature of the biological samples, separation by ultra-performance liquid chromatography (UPLC) is performed before the MS analysis.
Subsets of both urine sample sets were selected to be analysed by UPLC-QTOF-MS at Imperial College (Figure 10). MS was performed in both positive and negative ion electrospray (ESI+ and ESI-) modes.
Figure 14: Training and validation sets of urine samples analysed by UPLC-ESI-MS
A quality control (QC) sample was prepared by pooling identical aliquots of all the samples. The QC sample was used to condition the column prior to the analysis and then re-injected at regular intervals throughout the run (every 10 samples) to ensure system suitability and stability.
Further more comprehensive data processing and statistical analysis than the one explained in the previous report was performed. Principal components analysis (PCA) showed tight clustering of the QC samples in both the training and validation sets in ESI+ and ESI- modes, indicating good quality of all the data obtained and small analytical variation. Supervised models were subsequently generated to maximize the separation between classes and identify biomarkers associated with specific classes corresponding to the nature and stage of disease (Figure 11).
Figure 15: Multivariate statistical analysis of the validation set run by UPLC-(-)ESI-MS. A) Scores plot of the unsupervised analysis by principal component analysis (PCA). B) Scores and S-plots of the supervised analysis by orthogonal partial least square regression discriminant analysis (OPLS-DA).
Using the S-plot from the supervised analysis, the variable influence of the models was examined. The up-regulated metabolites in HCC were putatively identified using accurate mass (m/z), isotopic pattern, fragmentation pattern and retention time information. m/z values were searched against online databases including HMDB (http://www.hmdb.ca/) and METLIN (http://metlin.scripps.edu/index.php). When possible, authentic standards were analysed for definitive identification.
The urinary metabotype of patients with HCC was distinct from those with Cir, DC and NC with areas under the receiver operating characteristic (ROC) curves of 0.86(0.78-0.94) 0.93(0.89-0.97) and 0.89(0.80-0.98) in the training set; and 0.81(0.73-0.89) 0.96(0.94-0.99) and 0.90(0.85-0.96) respectively in the validation cohort, showing significantly higher diagnostic accuracy of liver cancer than serum AFP.
From the initial list of discriminant metabolites tentatively identified, 11 were confirmed with standards (Table 7) and used to develop the targeted method that will be run in the TQ instrument in the MRC Gambia. Method development was done by a Gambian postdoctoral researcher, Dr. Haddy Fye, trained at Imperial College.
Table 9: Up-regulated metabolites in HCC used for the development of the targeted UPLC-MS method. Metabolite IDs confirmed with standards
ID RT (min) m/z Ion Formula
1-Methylnicotinamide 0.57 137.071 [M+H]+ C7H9N2O
Acetylcarnitine 0.97 204.124 [M+H]+ C9H17NO4
Propionylcarnitine 1.58 218.139 [M+H]+ C10H19O4
Hydroxyphenyllactic acid 2.65 181.05 [M-H]- C9H10O4
Homovanillic acid 2.78 261.006 [M-H]- C9H10O7S
Kynurenic acid 2.85 190.051 [M+H]+ C10H7NO3
4-Hydroxybenzaldehyde 3.59 121.028 [M-H]- C7H6O2
Indole lactic acid 4.38 204.065 [M-H]- C11H11NO3
L-Octanoylcarnitine 5.59 288.217 [M+H]+ C15H29NO4
Glycocholic acid 7.38 464.301 [M-H]- C26H43NO6
Glycochenodeoxycholic acid-3-sulfate 7.85 528.263 [M-H]- C26H43NO8S
The metabolic significance of the presence of some of these low molecular compounds being higher in the urine of patients with HCC requires further studies. For example, propionylcarnitine is known to be in abundance in the urine of individuals with methylmalonyl-CoA mutase deficiency. Mutation of hepatocytes during the process of carcinogenesis may thus lead to defective isomerization of methylmalonyl-CoA to succinyl-CoA, which could impact the citric acid cycle. Indolelactic acid, a product of tryptophan metabolism is common in urine of humans. Its presence in higher concentration in the urine of patients with HCC suggests a differential metabolism of this amino acid in this group of patients. Glycocholic acid has been described to be perturbed in chronic liver disease as well as HCC patients and in whom there is significant trend of increase in concentration relative to severity of liver damage. The foregoing thus opens up a world of further basic science studies regarding the role of these urinary metabolites in the process and survival of hepatocarcinogenesis. For example, increases in urine of these metabolites may not necessarily be representative of the fact that they are increased in concentration in blood of HCC patients. Differential renal function may alter excretion of metabolites in the urine. Also, the tumour microenvironment of HCC may not necessarily be replete with these metabolites. Further studies involving the use of sera as well as liver cancer tissues could throw more light regarding role of these compounds in HCC.
Figure 16 shows the extracted ion chromatograms (EIC) of the discriminant metabolites up-regulated in HCC in a QC urine sample run by the profiling method in both ion modes (A) as well as the EIC of the corresponding standards run by the developed targeted assay (B).
Figure 16:Extracted ion chromatograms (EIC) of the discriminant metabolites up-regulated in HCC. A) EIC in QC urine sample run by the profiling method in positive (A1) and negative (A2) ion mode. B) EIC in Stadard mix run by the developed targeted assay 1: 1-Methylnicotinamide. 2: Acetylcarnitine. 3:Propionylcarnitine. 4:Hydroxyphenyllactic acid. 5: Homovanillic acid. 6:Kynurenic acid. 7:4-Hydroxybenzaldehyde. 8:Indole lactic acid. 9: L-Octanoylcarnitine. 10: Glycocholic acid. 11: Glycochenodeoxycholic acid-3-sulfate
The main differences in both UPLC methods are highlighted in the Table 8 below. The use of a TQ instrument in the targeted method allows the detection of compounds in negative and positive in a single chromatographic run, reducing the run time by half. By changing the column dimensions, the flow rate was decreased 3 times, as well as the run time per sample. Reduction in run time and in flow rate means a noticeable decrease of the consumption of solvent necessary to analyse a sample, which is critical in the developing world.
Table 10: Comparison of the profiling and chromatographic methods
UPLC method Profiling method Targeted method
Run time 25 min per sample
(1 chromatographic run of 12.5 min per polarity) 10 min per sample
(1 single run, both polarities)
Column type HSS T3, 2.1 X 150 mm HSS T3, 1 X 100 mm
Column temperature 45 °C 40°C
Flow rate 0.6 ml/min 0.2 ml/min
The developed method has been transferred to the instrument in the MRC Gambia, and Dr. Haddy Fye is currently validating and assessing its performance by analysing new urine samples collected in the Gambia.
Imperial College will continue to investigate the performance of the panel of identified biomarkers as an early diagnostic tool for liver cancer.
The work to date has been published in Hepatology, and further papers will ensue:
3.6 HBV & HCV infection: virological factors for HCC development (WP6)
The first objective of this workpackage was to implement standardised protocols for HBV molecular and serodiagnostics and evaluation/quality control of standardized protocol in all sites and this has been achieved. The second workpackage objective was to identify and characterize HBV genotypes and variants from the HC4 and WATCH platforms
3.6.1 Establish an in-house HBV DNA assay and transfer skills to Gambia & Senegal
Through training within INSERM and with subsequent help from partners, we have been able to transfer the in-house qPCR techniques to the African labs. We crosschecked all the samples for viral load qPCR and found comparable results (Figure 12) in the assay performed in INSERM and also in Gambia and Senegal, as part of our quality control programme. The technology transfer of the HBV qPCR has been achieved successfully. The assay design, validation and implementation of in-house qPCR in the PROLIFICA programme is currently being written as a publishable document.
Figure 17: The correlation curve showing the significantly similar results obtained with the same samples tested in MRC Gambia and INSERM as well as Senegal and INSERM
3.6.2 Viral genotype, sequence and variant analysis
Establishment of full Surface gene amplification and HBV full genome amplification:
We have performed full length amplification HBV genome as well as full HBV surface gene amplification from the test batch samples successfully and sequenced (from commercial source) to determine the genotypes. The common procedures were followed in all the subsequent screening and surveillance programme (Figure 13) The genotyping was primarily performed with the help of online NCBI genotyping tools by using the sequenced data (Figure 14) and further analysed using HBV Data base (https://hbvdb.ibcp.fr/HBVdb/) and other bioinformatics software (e.g. Clustal, MEGA)
Figure 18: Diagrammatic representation of the strategy and result HBV DNA amplification and sequencing
Figure 19: Representative picture showing HBV genotyping data and the results from NCBI viral genotyping tools.
Samples for the study
To date, we have received 3 types of cohort samples from Gambia through the IARC biobank. The WATCH cohort consists of 495 HBsAg positive samples. The Blood Donor Cohort includes 300 HBsAg samples and the HC4 cohort, more than 200 HBsAg positive samples. The WATCH cohort is a community based screening cohort, whereas the blood donor cohort is a facility based cohort. Patients are screened for HBsAg in the blood bank before blood donation. Those that are positive are recruited into the blood donor cohort. On the other hand, the HC4 cohort is a hospital based patient cohort who are screened for HCC. They are diagnosed as having HCC, liver cirrhosis or chronic HBV with fibrosis depending on their clinical, biochemical, pathological and radiological findings. For our mutation screening and genotyping studies, we have screened for HBV by DNA extraction and PCR on full surface gene and subsequent sequencing on 195 WATCH samples (based on >150 IU/mL HBV DNA as the cut off), 121 HC4 HBsAg+ samples and 64 HBsAg- HC4 samples (to check for occult HBV), and 98 HBsAg+ blood donor samples.
HBV full surface sequencing by direct PCR:
We have PCR amplified and sequenced the full surface gene (PreS1/PreS2/S overlapping a major part of polymerase including the RT region) as described previously. The results are as follows: - 122 WATCH samples (122/195, 62% success), 95 HBsAg+ HC4 (95/121, 78% success), 22 HBsAg- HC4 (Occult HBV, 22/64, 34% success) and 90 HBsAg+ blood donor (90/98, 90% success). We are in the process of submitting this sequence data to the NCBI database. The majority of the sequences generated in this PROLIFICA study are either of HBV/Genotype E or HBV/Gen A (see the genotype data below). This will be the largest collection of HBV/Gen E data from Sub Saharan Africa (SSA).
HBV full surface sequence by cloning of PCR products:
Due to the presence of PreS in frame deletion mutant (see the deletion mutant data) in a large number of samples, we have had a problem with the PCR based sequencing. To overcome this problem, we cloned the PCR products in a TA cloning vector (TOPO TA Cloning kit, Life Technologies) and then sequenced. We have selected 5-10 clones from each individual and thus generated over 100 cloned full surface sequences from 21 individuals (WATCH and HC4 cohort). This has helped us to understand and confirm the deletion mutant and their quasispecies complexity (see below). The sequence data will be submitted to the NCBI database very soon.
Generation of Full Length HBV sequence by cloning:
To further explore the full genome HBV genetic variability, we have amplified the full genome HBV (as described by Ghosh S et al., 2010) and subsequently cloned into PCR XL TOPO vector (TOPO XL cloning kit, Life technologies) and sequenced by multiple primers. The sequence fragments were joined by seqscape software (Applied Biosystem). Thus we generated 25 full genome HBV clone from 8 samples.
Generation of sequence of PreC/C region:
In order to explore the prevalence of Basal core promoter (BCP – 1762, 1764) and PreCore promoter (CP - 1896, 1899 stop codon) mutations, we have amplified and sequenced the region between nt1680-nt2460 of the HBV genome. This is very important, as >95% of the HBsAg+ individuals have HBeAg negative infection with very low viremia (>2000 IU/mL). We have successfully sequenced 60 individuals (47 HBeAg negative and 13 HBeAg positive) from the WATCH and HC4 cohort. However, the success of PCR amplification is low (>55%) in this region, due to the very low viral load and complexity of the HBV genome in this region. The following data table (Table 9) suggests that BCP/CP mutation is very high (80-90%) and this may be one of the major reasons for the high incidence of HBeAg negative infection (90 - 95%) in SSA.
Table 11: BCP/CP mutation profile
HBsAg quantification:
In order to understand and explore the role of HBV surface gene mutation on the expression of HBsAg in blood, we have quantified HBsAg from the serum of all HBs sequenced samples (217 HBsAg + samples from the HC4 and WATCH cohort). The quantification was performed using Cobas e 411 immunoassay analyser (Roche). We have also tried to see if there is any correlation between HBV viral load and HBsAg quantification (Figure 15). Interestingly, we found some individuals who had a very high viral load and low HBsAg or vice versa. This phenomenon may be related to genetic variability of the HBs gene.
Figure 20: Correlation between HBV viral Load (IU/mL) and HBsAg (IU/mL)
HBV Genotyping:
To determine the genotype of the strains and their phylogenetic relation, a tree was built on the 221 Surface gene sequence (only in-del free region) obtained in this study, along with the reference sequence of all known HBV genotype/subtypes obtained from the NCBI database. A ‘Maximum-Likelihood’ tree was built using the Jukes-Cantor model and 1000 bootstrap events in MEGA6 software. The final tree was edited in Figtree software (Figure 16). This showed that the majority of the individuals had Genotype E (>82%, fluorescence blue) and the rest had Genotype A (15.4%, fluorescence green), except for 4, which may have a recombinant of Genotype E and A (red). Further full length cloning based sequencing studies are needed to confirm (Figure 17). We have been unable to amplify the full length from the 4 samples, due to a limited number of biological specimens and a very low viral load.
Figure 21: Phylogenetic tree reconstruction of full length surface (PreS1/PreS2/S) sequences.
Figure 22: Probable site of recombination in the 2 strains were shown in the picture bellow using web-based NCBI genotyping tools of the full surface gene.
Role of HBV PreS deletion in the development of HCC in SSA:
We have observed a high incidence of PreS deletion (>34%) among the sequence generated in the PROLIFICA samples. To further explore its role in HCC in SSA, we have analysed 217 full surface sequences from the HC4 and WATCH cohorts with genotype E or A in relation to their clinical phenotype. We excluded 4 individuals who may carry recombinant strains, due to its very low prevalence. We divided the 217 individuals into 3 groups: - Chronic HBV (No LC/HCC), Liver Cirrhosis (LC) and Hepatocellular Carcinoma (HCC). We have 42 LC and 36 HCC. Comparing the groups, we have observed PreS2 deletion, Genotype A individuals are associated with the development of HCC in SSA. It is also possible that the lower incidence of HBsAg among the HCC group, compared to the No LC/HCC group (Table 10), may be related to the PreS2 deletion. We are currently exploring these findings.
Table 12: Demographic, Biochemical and HBV profiles
Type of PreS deletion in PROLIFICA samples:
We have mainly identified 3 types of PreS deletion strains among the strains isolated from different individuals (Figure 18). PreS region of HBV is frequently deleted by immune selection pressure, as it contains several T and B cell epitopes and is often found to be associated with liver cirrhosis or HCC (Pollicino et al., 2014). In our study, the most prevalent was the PreS2 deletion (>32%), followed by truncated S (9%) and PreS1 deletion (3%). These deletions were present in individuals with both the genotype (E and A). The most important observation of this sequencing study is the high incidence of PreS2 deletion in a specific region (aa5-20 of preS2) and its strong association with the HCC (>55%). This type of mutant was previously described by Pollicino et al (2014) and Su et al (2008 and 2014) and implicated in ground glass hepatocyte (GGH) formation and liver pathogenesis by oxidative stress.
Figure 23: Representative amino acid sequence alignment to show the different deletion strains
Presence of wild type and mutant PreS2 deletion strains in the same individuals:
To explore the quasispecies structure (since we have had difficulties in direct PCR product sequencing) within the same individuals, we sequenced the PCR products after cloning. The cloning and sequencing revealed that almost all the individuals carrying the PreS2 mutant have wildtype strains. Also, the in frame deletions are of different lengths within the same individual (Figure 19). To further confirm and explore the complex wildtype and mutant ratio, we are currently exploring selected samples for ultra-deep pyrosequencing (UDPS).
Figure 24: Wild type and mutant PreS2 quasispecies within the same individual
Immunohistochemistry for anti-HBs on liver biopsy:
It has previously been described (Su et al., 2008 and 2014) that the PreS1 and PreS2 deletion mutant forms GGH type 1 and type 2 are an important indication of liver pathogenesis. To explore this phenomenon, we also did immunohistochemistry (IHC) on the liver biopsy samples from PROLIFICA study patients and observed the same staining profile (Figure 20). Currently, we are investigating a large series of biopsy samples matched with wild type and mutant strains, to further explore/establish this hypothesis.
Figure 25: Ground Glass hepatocytes and PreS deletion
In-vitro studies on PreS2 deletion mutant:
To further explore the impact of PreS2 deletion mutant on the liver pathology, we designed some in-vitro studies (Figure 21). We cloned both full length Wild type and Mutant strains and amplified them with a primer containing SapI RE site. After SapI digestion and the purification of PCR products, we transfected into the Huh7 cells. Flowing transfection, we collected the supernatant and checked for HBsAg secretion by quantitative ELISA. We found that the mutants have less secretion of HBsAgcompared to the wild type. However, when we co-transfect the wild type and mutant, the secretion of HBsAg was restored to near normal. We also performed RNA blots to confirm normal replication and transcription capacity of both wild type and mutant strains within the cells.
Figure 26: In-vitro characterization of PreS2 deletion mutant
Immunoflourescence studies on in-vitro transfected cells:
In order to explore the intracellular HBsAg present in the wild type and mutant containing cells, we undertook immunofluorescence on the transfected cells after fixation. The confocal visualization of the transfected cells confirmed that PreS2 mutants store a high amount of HBsAg within the cells (Figure 22) and a high probability of having oxidative stress, DNA damage and ER stress. We have unsuccessfully tried to explore this hypothesis by western blot and qPCR (mRNA expression) using several ER stress markers. It is possible that due to very low transfection efficiency (5-15%), we are not able to see the difference on the whole cell extract. However, we are currently performing co-localization of HBV infected cells (with or without mutant) and different oxidative stress and DNA damage markers via IF, to further explore the possibility.
Figure 27: Immuno-fluorescence of transfected Huh7 cell lines
3.6.3 Occult HBV:
We have found 22 patients (22/64, 34%) in the HC4 cohort who are negative for HBsAg and positive for the HBV DNA test. We have also checked for the quantitative HBsAg in these patient’s samples and found all were below the detection limit (>0.025 IU/ml) for confirmation of the sero-negative status. We are currently exploring the HBeAg, HBcAg and Anti-HBs among these samples. All 22 samples were sequenced for the full surface gene and we are currently looking for any genetic variability of the strains..
3.6.4 Current and future virology work:
• We are currently analyzing all the sequences for point mutation.
• Undertaking UDPS to establish the wild type mutant quasispecies ratio within an individual.
• Using IHC on a series of biopsy samples to further explore the role of GGH1 and GGH2.
• Investigating the co-localization of oxidative stress markers on the transfected cells containing wildtype or mutant HBs
• Submitting all the generated sequence data into the INSERM database,
• Exploration of occult HBV and its molecular basis (if any).
• Submission of 2 manuscripts on HBV viral load, HBV virology and PreS deletion from the PROLIFICA study by November 2015.
INSERM has successfully completed their objectives for this period: the implementation of standardized protocols and the technology transfer of HBV qPCR to PROLIFICA partners. They are well placed to complete their project objectives.
3.7 Training & capacity building (WP7)
Objective: to build capacity by creating reference centres
All the partners (except Xeptagen) have been involved in capacity building
3.7.1 Clinical training and capacity building
In The Gambia five clinicians have been trained in hepatology: Dr Gibril Ndow, Dr Louise Saar, Dr Abu Kambi, Dr Momodou-Lamin Tekanyi, and Dr Sheikh-Omar Bittaye and one nurse (Saydiba Tamba).
The training consisted in lectures, supervision of clinical work in the liver clinic. All have been trained on the use of Fibroscan and Liver ultrasound and liver biopsy. They have also been trained in the liver diagnosis, management of chronic hepatitis B and end-stage liver disease.
The nurse has been train to manage HBV patients under treatment.
Two medical officers employed at the EFSTH (previously named RVTH) were recruited to join PROLIFICA as trainee hepatologists.
The trainees dedicated two days of the week entirely for this training. They worked closely with and under supervision of PI and consultant hepatologist Dr Ramou Njie, as well as Dr Maud Lemoine. During the past year they have seen and managed a number of patients with chronic liver disease such as hepatocellular carcinoma, cirrhosis and chronic hepatitis. Consequently, most of the patients seen at the EFSTH (previously named RVTH) with chronic liver disease seen on Wednesday afternoons are referred to the liver clinic at MRC Fajara for specialist review. Their skills in how to properly assess a patient with liver disease, carry out a liver biopsy, Fibroscan and ultrasonography, have been greatly enhanced.
In Senegal, 4 hepatologists (Prof. Ka. Dr Toure, Dr Madoky and Dr Da Veiga)and one nurse have been trained to use the Fibroscan.
Two Fibroscan devices and two Ulrasoound machines have been bought and given to the local partner (MRC in Gambia and University de Thies in Senegal).
As part as the project we also built a liver Unit at the MRC unit in Gambia and in the main teaching hospital in Banjul (Gambia)
Within the PROLIFICA programme, we have been able to settle an endoscopy clinic in The Gambia (this was already present in Senegal) with the support of the World Gastroenterology Society (Dr Des Leddin). Dr Ramou Njie and Maud Lemoine participated to the teaching and training in January 2014 with the support of the MRC clinical services (Dr Suzanne Anderson and Prof. Umberto D’Alessandro).
3.7.2 Lab Training
Local lab technicians and senior virologists have been trained to run qPCR and set up a HBV in-house PCR under the supervision of INSERM, Lyon France (Prof. Zoulim’s team) and Dr Maimuna Mendy (IARC, France). Training has been done locally but also at INSERM (Lyon, France) to train in particular Dr Amina Sow from Senegal and Dr Harr Njai from Gambia. Penda Suso junior Lab tec has been trained in Lyon from Feb to march 2016 to run and interpret serologies and qPCR.
In addition, Imperial has facilitated the setting up of a Mass Spec Laboratory at MRC Gambia in conjunction with Waters Inc and also provided training to MRC staff in the use of metabonomics to identify liver cancer biomarkers. This involved comprehensive training by specialised researchers at Imperial, followed by training on site, once the mass spectroscopy laboratory had been established at the MRC. Dr Njie has been involved with the on-site establishment of the laboratory and MRC staff training.
There is one dedicated trained scientist, Dr Haddy Fye, who managed the facility and was responsible for all local metabonomic research. She will eventually assist in training local staff in the technology. The Imperial metabonomic staff has supervised Dr Haddy Fye’s work and help with any issues, while also providing updates to her training. Imperial is currently investigating funding possibilities to support the future cost of running the facility.
Histopathology: liver biopsy samples went through a typical histopathology laboratory protocol which Penda Suso has documented as an SOP. Histopathological analysis has been done by Prof. Khalil. Initially, only one block was being cut but this was later revised to two so that one was used locally for diagnosis and the other sent to IARC (for archiving). Locally, eight slides were sectioned (2 for diagnosis; 3 sent to UK and 3 were archived).Penda Suso also successfully attended a one month’s training (20th August – 20th September 2013) on histological techniques at Imperial College London, St. Mary’s Hospital, London.
3.7.3 Academic training
The PROLIFICA has been able to train 16 students and junior doctors (please see below) in clinical and basic science research.
Degree Year Name
Bsc in Medicine (n=6) 2013
2014 Yuki Takao
Lauren Berg
Harry Posner
William Stranger
Edmund Donnelly
Mario Lapore
Master in public health (n=1) 2013 Liliane Mpanzani
PhD (n=6)
Epidemiology 2011-2014 Yusuke Shimakawa
Public health 2013-2016 Shevanthi Nayagam
Immunology 2013-2014 Nick
Medicine 2010-2013 Nimzing Ladep
Medicine 2011-2015 Torben Kimhofer
Medicine 2015-2018 Gibril Ndow
Science 2014-2017 Nick Easom
Post Doc (n=3)
2011-2013 Harr Njai
2013- Haddy Fye
2014-2015 Jess Howell
Fellowship (n=1) 2011-2013 Maud Lemoine
Potential Impact:
PROLIFICA has demonstrated that screening and treatment for hepatitis B in sub-Saharan Africa is feasible, cost-effective and necessary in order to control morbidity and mortality related to this infection. Furthermore the work demonstrates that screen and treat will be an essential component of the public health interventions required to eliminate this infection.
4.1 Local impact in Gambia and Senegal
On a number of occasions during the PROLIFICA programme the investigators have held meetings with members of the health ministries in Gambia and Senegal. These meetings have helped to raise the levels of awareness about HBV infection and its role in the development of cirrhosis, liver failure and liver cancer. In Gambia the Minister of Health attended the WHO World Hepatitis Day conference in Glasgow in September 2015 and was a signatory to the Glasgow Declaration on Hepatitis (http://www.who.int/hepatitis/glasgow-declaration-on-viral-hepatitis/en/). He is committed to increasing the public health response to viral hepatitis in Gambia and is currently developing a national plan to counter HBV in the country. It is expected that this plan will encompass the screen and treat strategy that was pioneered in the PROLIFICA project.
The HBV viral load assay developed for use by PROLIFICA is now being used effectively in Gambia to monitor patients on treatment. The skills and laboratory facilities generated by the project will now be available to support the more widespread application of screening, assessment and treatment monitoring anticipated in the Ministry of Health’s plan. During the PROLIFICA project four locally trained doctors and about 20 fieldworkers were trained in the skills required for screening and assessment of patients with HBV infection. These medical staff will prove invaluable once the Government backed plan is implemented.
The Senegalese ministry of health are also now developing a national plan for viral hepatitis. This includes a commitment to purchase Tenofovir for use in HBV mono-infected patients. The clinicians and laboratories participating in the PROLIFICA project are advising the Ministry about strategy for a national programme as well as providing the necessary molecular virology support.
4.2 Natural history of HBV
Using the clinical and laboratory facilities established as part of the PROLIFICA project in Gambia we were able to define the natural history of chronic HBV infection in West Africa. Cohorts of patients recruited over 25 years ago in the villages of Keneba and Manduar had been followed at 4-5 year intervals using liver biochemistry and viral serology. Full clinical assessments were performed on 301 patients. The following key findings were reported:
• Annually, 7.4% (95% CI 6.3% to 8.8%) cleared HBeAg and 1.0% (0.8% to 1.2%) cleared HBsAg.
• The incidence of HCC was 55.5/100 000 carrier-years (95% CI 24.9 to 123.5).
• 5.5% (95% CI 3.4% to 9.0%) had significant liver fibrosis.
• HBV genotype A (versus E), chronic aflatoxin B1 exposure and an HBsAg-positive mother, a proxy for mother-to-infant transmission, were risk factors for liver fibrosis.
• A small proportion (16.0%) of chronic carriers were infected via mother-to-infant transmission; however, this population represented a large proportion (63.0%) of the cases requiring antiviral therapy.
Most importantly this study highlighted the importance of mother to child transmission in sub-Saharan Africa which had previously been ignored due to the low proportion of transmission via this route. Future public health interventions will therefore need to target this mode of infection.
4.3 Epidemiological Modelling
As part of the PROLIFICA project we developed a simulation model of the global HBV epidemic, incorporating data on the natural history of HBV, prevalence, mortality, vaccine coverage, treatment dynamics and demography. We used the model to estimate the impact of current interventions and the impact of scaling up of existing interventions for prevention of infection and introducing population screening and treatment as a novel intervention. The model was populated with data taken directly from the PROLFICA WATCH study.
The model demonstrates the enormous impact that infant and neonatal vaccination programmes have already made in some global regions. Specifically it is estimated that 210 million new chronic HBV infections were averted between 1990 and 2015. The impact of vaccination on HBV related deaths is more modest with approximately 1.3 million deaths averted by 2030. However the model shows clearly that there is a need to scale up existing interventions as well as to introduce novel interventions in order to control mortality. For example if infant vaccination was scaled up from the current average of 70% to a target of 90%, and if birth dose vaccination was scaled up to 80% of children born to mothers with high level viraemia and if population wide testing and treatment was able to access 80% of the eligible population then there would be a global reduction of 90% of new chronic HBV infections and 62% reduction in mortality. Furthermore, this scale up of public health interventions would lead to achievement of a realistic HBV elimination threshold by 2090 which could not otherwise be reached.
The Global Epidemiological model has been shared with the WHO’s Global Hepatitis Programme in order to respond to the World Health Assembly’s request to evaluate what is required and whether it is feasible to eliminate HBV (http://www.who.int/mediacentre/news/releases/2014/WHA-20140522/en/) .
4.4 Return on investment
PROLIFICA data and the Global Epidemiological Model were used to estimate the costs and the return on investment for a comprehensive package of possible public health interventions for Senegal. The package of interventions include infant vaccination, measures to prevent mother-to-child-transmission and screening and treatment for HBV. We assumed an overall societal perspective by considering both direct and indirect costs associated with HBV interventions. The former were mainly the costs associated with seeking and obtaining prevention or treatment interventions (medical costs) while the latter were the costs to society due to the premature death of people infected with HBV. Direct costs were represented by the medical fees charged to patients at either public or private health facilities. In reality, fees are just a representation of the real costs as they depend on several factors including whether the patient is covered by health insurance, whether the patient seeks treatment at a private or a public health facility and on the baseline availability of certain diagnostic tests or treatments
Between 2015 and 2030, investing 265 billion CFA (€400 million) into this comprehensive package of interventions will give an economic return of 631 billion CFA (€963 million). This would amount to a return on investment of 2.38 CFA per CFA invested & could avert 21 000 deaths by 2030.
4.5 Awareness of limitations of vaccination only strategy
Until the World Health Assembly resolution on viral hepatitis in 2010 the WHO had only one strategy to tackle viral hepatitis which was the global implementation of the infant vaccination integrated with the expanded programme of immunisation in all member states. The WHO had also recommended birth dose vaccination for children born to mothers with high levels of viraemia but this policy has been poorly observed. The data emerging from PROLFICA have highlighted the need for additional public health interventions to control new chronic infections and to reduce the mortality associated with HBV
4.6 WHO guidelines
Prof. Mark Thursz and Dr Maud Lemoine are members of the HBV & HCV Guidelines Development groups for the World Health Organisation Global Hepatitis Programme. Dr Nayagam is a member of the WHO guideline development group for hepatitis screening.
4.6 Laboratory enhancement in Gambia
PROLIFICA’s Metabonomics workpackage results have stimulated further research on biomarker development for the early diagnosis of HCC. The Waters Corporation kindly donated a QToF mass spectrometry system for use at the MRC Laboratories facility in Gambia. This allowed Imperial researchers (Prof Taylor-Robinson and Prof Elaine Holmes) to set up a metabonomics laboratory run by Dr Haddy Fye, a post doctoral scientist trained at Imperial. The lab is now analysing urinary metabolic profiles in suspected HCC patients.
4.8 Cohorts of HBV infected patients to study long term outcomes
The WATCH platform generated two cohorts of patients that will be used to answer key questions about the management of HBV in the future. 1. The treated patient cohort will be used to monitor levels of adherence and the ability of treatment to prevent advanced liver disease over a prolonged duration of treatment. 2. The observational cohort of patients who did not meet treatment eligibility criteria will be used to determine the rate of spontaneous viral clearance, the rate of transfer from ineligible to eligible for treatment and the liver related outcomes in this group. This data will inform policy on the need for a frequency of follow-up in this patient group.
List of Websites:
www.prolifica.eu
Primary liver cancer (hepatocellular carcinoma: HCC) is the most common malignancy in adult men and the second most common malignancy in adult women in West Africa. Since the survival from diagnosis is less than a year this makes HCC a major cause of premature adult mortality in West Africa. Hepatitis B virus (HBV) is a small, hepatotropic, double-stranded DNA virus causing acute or chronic hepatitis infection. Chronic HBV infection may lead to cirrhosis, liver failure or HCC. Recent estimates suggest that there are over 250 million people worldwide who have chronic HBV infection with high prevalence of infection in sub-Saharan Africa and Asia.
HBV can be prevented with a vaccine based on the envelope protein of the virus and can be treated using nucleoside analogues. Since 1990, under guidance from the World Health Organisation (WHO), the majority of countries have implemented infant vaccination programmes which integrated HBV vaccine with expanded programmes of immunisation. Treatment of HBV with nucleoside analogues suppresses viral replication, prevents disease progression, reverses liver fibrosis and cirrhosis and reduces the risk of HCC. However, there are currently no national treatment programmes for HBV in sub-Saharan Africa.
The principal aims of Prolifica were:
• to identify the main risk factors for HCC in a sub-Saharan African population using a case-control design
• to demonstrate that treatment of chronic HBV infection is feasible and an effective method of reducing the incidence of HCC in resource poor settings.
• to demonstrate that HCC can be detected early and treated effectively
PROLIFICA established two main research platforms:
1. The hepatocellular carcinoma case-control study (HC4) in Gambia, Senegal and Nigeria
2. The West African Treatment Cohort study (WATCH) to identify and treat patients with chronic hepatitis B infection in Gambia and Senegal.
In the HC4 study we collected data and biological samples from over 500 patients with HCC and controls with chronic HBV infection and normal population controls. Using this resource we have identified proteomic and metabonomics profiles which help to identify patients with HCC. These novel biomarkers will be developed into diagnostic assays for future clinical evaluation. We developed a point-of-care test for SCCA-IgM for use in near patient testing for HCC using a lateral flow assay system. The test proved reliable in serum samples wheter the SCCA-IgM was > 200Au/ml.
WATCH is the first study to evaluate the feasibility, effectiveness and cost-effectiveness of screening for HBV in the community and treating those eligible for treatment. We found a high uptake of screening (68.9%) and good linkage into care (81.3%). Amongst patients identified in the community only about 5% met eligibility criteria. The cost of screening was €7.50 per person and the cost effectiveness €500 per Disability Life Year averted.
Data from the WATCH study was used to populate a simulation model of the global HBV epidemic and to estimate the potential impact of various public health interventions. The model demonstrates that mortality due to HBV infection cannot be controlled in the foreseeable future without implementing a screen and treat strategy as illustrated in the WATCH study. In addition mother-to-child transmission of HBV is already becoming the dominant mode of transmission and to prevent new chronic HBV infections in the future will require implementation of methods to prevent this route of infection.
Project Context and Objectives:
Primary liver cancer (hepatocellular carcinoma: HCC) is the most common malignancy in adult men and the second most common malignancy in adult women in West Africa1. Since the survival from diagnosis is less than a year this makes HCC the major cause of premature adult mortality in West Africa. Worldwide HCC is one of the ten most common malignancies and in Europe the incidence of this type of cancer is rising in virtually all member states2. Although any cause of chronic liver disease and cirrhosis may lead to HCC globally the most important causes are hepatitis B virus (HBV) and hepatitis C virus (HCV) infections3. According to WHO estimates 250 million people worldwide are persistently infected with HBV and 180 million with HCV. Together these viruses are responsible for approximately one million deaths annually from HCC or cirrhosis with liver failure.
Fortunately only a minority of people with chronic HBV or HCV infection will develop HCC or liver failure 4. The main risk factor for progression to HCC in both infections appears to be the development of advanced liver fibrosis and cirrhosis 5. Other, established, common risk factors for HCC include, male gender, advanced age, exposure to aflatoxin and having a first degree relative with the disease6 7. In Asia and more recently in Africa high levels of virus in serum and high alanine transaminase (ALT) values have been correlated with progression of chronic HBV infection to HCC8 9. Most of our understanding about the natural history of HBV infection arises from European and Asian studies and a recent review pointed out the paucity of information on the natural history of persistent hepatitis B infection in Africa 10. Although the prevalence of chronic infection in adults is similar in Asia and Africa there are several difference which are likely to lead to variation in natural history of disease. The age at infection, which influences the rate of viral persistence differs, with childhood infection around the age of 2-3 years predominating in Africa whilst some 40% of HBV carriers in Asia result from perinatal infection 11. The virus differs with genotype E predominating in the Gambian population whilst B and C are the most common HBV genotypes in Asia. Co-factors also differ with little or no alcohol consumption in rural Gambia but high levels of aflatoxin exposure which we have recently suggested could play a role in cirrhosis 12. In addition there is of course a different human genetic background.
In West Africa over 60% of HCC cases have chronic HBV infection13. The WHO campaign to include HBV immunisation in the Expanded Programme on Immunisation schedules of all countries in the world is a major success. National surveys have confirmed that this intervention has reduced the prevalence of chronic HBV and, in Taiwan, the incidence of HCC amongst children 14. However, implementing HBV vaccine programmes for children still leaves a legacy of chronically infected people, who acquired the infection prior to the implementation of the programme, which will persist for the next half century. Furthermore, the changing demographics of people in sub-Saharan Africa, with an ageing population means that whilst the incidence of HCC may fall as a result of vaccination the number of cases will actually double by 2050 unless steps are taken to provide treatment for HBV and other aetiological factors.
Treatment is now available for the management of chronic HBV infection in developed countries. Until the 1990s the only treatment option for chronic HBV was interferon alpha which is not appropriate for widespread use in sub-Saharan Africa because serious side effects are common, the cost is high, there is need for intensive monitoring and therapy is contra-indicated in patients with advanced liver fibrosis or cirrhosis. However, treatment may now be delivered using oral nucleoside or nucleotide analogues which inhibit the HBV polymerase/reverse transcriptase enzyme 15. These drugs are administered once daily and are relatively free from side effects. However, as with the treatment of human immunodeficiency virus (HIV), the majority of patients will require indefinite treatment with nucleos(t)ide analogues in order to maintain suppression of viral replication. In a placebo-controlled trial in Asian patients treated with Lamivudine it was shown that suppression of viral replication prevented progression of liver fibrosis and reduced the risk of developing HCC 16. Two of the drugs (Lamivudine and Tenofovir) used to treat HBV infection are also used to treat HIV infection and are available in Africa at affordable prices.
Hepatitis B vaccine coverage in SSA is imperfect 1 and a large number of people born before the introduction of the vaccine continue to carry the virus with a risk of cirrhosis and HCC. 2 In SSA screening and treatment for hepatitis B are rarely accessible3,4 and HCC remain a major cause of death in young individuals.
The main objectives of PROLIFICA were 1) to evaluate whether screening and treatment for HBV infection is feasible and effective at reducing the risk of HCC in West Africa (Gambia & Senegal), 2) to identify biomarkers of cirrhosis and HCC (Gambia, Senegal, Nigeria) and 3) to develop West African research platforms on HBV and HCC and transfer capacity building on HBV infection and liver cancer.
The project was launched in 2011 and ended in January 2016. It has been conducted with the support of eight partners (summarized in Table 1) from UK, France, Gambia, Nigeria and Senegal.
Table 1: Main Partners in the PROLIFICA Project
Partner Institution Lead Contact Key Work Area
Imperial College London, UK Prof. Mark Thursz (MT) Lead project, co-ordinator, metabonomics
International Agency for Research on Cancer (IARC), France Dr Maimuna Mendy (MM) Cancer, virology, biomarkers, proteomics
Institute National de la Sante et de la Recherché Medicale (INSERM), France Prof Fabien Zoulim
Dr Isabelle Chemin (IC) Infectious diseases, HBV virology
Xeptagen Inc. Italy Dr Giorgio Fassina Diagnostic development, biomarkers
MRC Gambia, Gambia Prof. Umberto D’Alesssandro (UDA) Clinical studies
Le Dantec Hospital, Senegal Prof Soulemayne Mboup (SMB) Clinical studies
Royal Victoria Teaching Hospital Dr Ignatius Baldeh (IB) Clinical studies
Jos University Teaching Hospital (JUTH), Jos, Nigeria Prof Edith Okeke (EO) Clinical studies
The project has been conducted within 7 workpackages (WP) summarized in Table 2:
Table 2: PROLIFICA Workpackages
Workpackage Lead unit/person Description
WP1 Imperial College (MT) Coordination & management
WP2
HCC case study Imperial College (MM) HCC control study dataset
WP3
Anti- viral treatment cohort (WATCH) study Imperial College (MT), MRC (UDA) and Senegal (SMB) Treatment study
WP4
Biomarker discovery and mechanistic investigations through proteomics
Imperial College (STR)
IARC (MM) Novel markers for liver cancer validation
WP5
Development of metabonomic markers for the diagnosis of HCC Imperial College (STR)
IARC (MM) Confirmed list of biomarkers
WP6 INSERM, Lyon, France (IC) HBV molecular and serodiagnostics
WP7 All partners Capacity building
As lead Partner, the main role of Imperial College has been to act as the overall administrative co-ordinator of the project to ensure efficient collaboration, exchange of information and knowledge and fulfilment of ethical, regulatory and bio-safety rules. In addition, Imperial monitors the progress of the project, anticipating possible hurdles and working with the partners to find solutions. Specifically this includes detailing the overall work programme and budget, co-ordinating and monitoring scientific and financial reports, collecting and compiling results and organising project meetings.
In addition, Imperial has been also involved in several scientific aspects of the project, in collaboration with other partners. These include establishing whether the treatment of HBV is an effective intervention to control HBV replication, liver progression and prevent development of liver cancer (work package 3, training radiographers in the use of ultrasound for the early diagnosis of liver cancer and Fibroscan to assess the stage of HBV-related liver disease (WP3), the development of metabonomic biomarkers for the diagnosis of HCC and capacity building (WP7).
The three African partners were primarily responsible for the clinical studies: HCC patients (WP2) and screening and treatment cohort study (WATCH study WP3). They also had key roles in the HBV and hepatitis C (HCV) infection studies (WP6) that was primarily the responsibility of INSERM. C
The role of hepatitis virus infection in the pathogenesis of HCC is complicated by viral diversity, different levels of viral replication and the presence of HBV mutants that are capable of modulating the carcinogenesis process. The main goal of WP6 was to provide information on the contribution of HBV genotype and HBV diversity in the development of HCC in The Gambia, Senegal and Nigeria and generate full genome sequences of the different genotypes circulating in the three countries.
IARC, Xeptagen and Imperial College have been the main partners responsible for investigating the possibility of identifying biomarkers for both HCC and HBV in biological samples taken from patients in both the case control and treatment studies. In addition they will performe extensive analysis of the generated sequence alignment to identify viral genetics patterns and investigate their association with liver disease. They will also investigate infection with hepatitis C (HCV) as a possible risk factor of HCC. IARC and Imperial used proteomics and metabonomics respectively to identify, evaluate and validate biomarkers for HCC.
All Partners have been involved in capacity building and dissemination of the project-related technology and knowledge on Hepatitis B, liver cancer, screening, prevention, detection and treatment (WP7) throughout the Partnership, but especially for West Africa. In particular, the aim was to create reference centres for hepatology and liver cancer in the three West African countries involved in the project
Project Results:
3.1 Coordination and management of the consortium (WP1)
In order to achieve its objectives, the PROLIFICA programme has developed two platforms of clinical and scientific research on viral hepatitis and liver cancer in The Gambia, Senegal and Nigeria: the West African Treatment Cohort for Hepatitis B (WATCH) and the Hepatocellular Carcinoma Case-Control Study (HC4). Consortium management coordinated the transfer of data and samples from these platforms to the other workpackage teams.
3.2 HCC case-control study (HC4) in The Gambia, Senegal and Nigeria (WP2)
A total of 630 patients with suspicion of end-stage liver disease (cirrhosis, decompensated cirrhosis, and/or HCC) have been enrolled in Prolifica: 164 patients including 61 patients with HCC have been enrolled in Senegal; 466 patients including 50% with HCC in Gambia.
The analysis of the epidemiological, clinical and virological characteritics of the complete study population is under progress.
However, preliminary analysis of 69 HCC patients shows that 49 (72%) are males with a median age at 43.5 years (22-80). At time of diagnosis, the median number of tumoral lesions is 3 (1-10) and the median size of the tumor size is 12.25 cm (1.5-32): 61 (88.4%) have a mean size ≥ 5 cm at time of diagnosis and 21 (34%) have ascites. HBV is the main cause of HCC (66%) with a median viral load at 43,760 IU/mL (<50-3.109). HDV coinfection is observed in 6%, HIV infection in 9% and 6% of the HCC cases is attributable to HCV infection.
Median AFP levels was 909 (1-35,350) and 19 (27.5%) had AFP levels < 200 ng/mL (Table below)
Table 3: Characteristics of HCC patients in the Gambia
Median length of survival in those who died was 61 days (4-789 days).Cause of death was determined in only 25 patients with haematemesis being the main important reported cause (44%) (see figure 3 below).
Figure 1: Kaplan-Meier survival estimates of Gambian patients with HCC
We also analysed the survival of compensated and decompensated cirrhotic patients. (Shimakawa Y et al. APT 2015 5 ) see figure 4 below. The cumulative mortality rate at 28-day was 70.0% (21/30) in the cases and 13.1% (8/61) in the controls (P < 0.001). The cumulative mortality at 90-day was 92.9% (26/28) in the cases and 28.1% (16/57) in the control group (P < 0.001). Positive anti-HEV IgG was not associated with higher 28-day mortality in cirrhotic patients. After adjusting for age, sex, and case/control status, none of the potential precipitating events of ACLF were associated with higher 28-day mortality. The causes of death in patients who died within 90 days after enrolment were as follows: end-stage liver disease (n=18, 69.2%), heart failure (n=1, 3.9%) and unknown (n=7, 26.9%) in the ACLF cases and end-stage liver disease (n=10, 62.5%), breast cancer (n=1, 6.2%) and unknown (n=5, 31.3%) in the control group.
Figure 2: Survival of patients with compensated or decompensated cirrhosis
As part as the PROLIFICA programme we also analysed whether hepatitis E infection, which is common infection in resource-poor setting, is prevalent in Gambia and is a cause of hepatic decompensation in this country. In 204 healthy volunteers, 13.7% (95% CI: 9.6-19.2) were tested positive for anti-HEV IgG, and none had positive HEV viremia. After adjusting for age and sex, the following were associated with positive anti-HEV IgG: being a Christian, a farmer, drinking water from wells, handling pigs and eating pork. In 40 cases (median age: 45 years, 72.5% male) and 71 controls (39 years, 74.6% male), ≥70% were infected with hepatitis B virus. Although hepatitis B flare and sepsis were important precipitating events of ACLF, none had marker of acute HEV. These results have published in 2016 in Alimentary Pharmacoly and Therapeutics as follows
As part of the PROLIFICA programme, using the WATCH and HC4 studies, we analysed whether birth order is a risk factor of HCC as it has been suggested in Asian studies. Indeed, early age at infection with HBV increases the risk of chronic infection. Moreover, early HBV infection may further independently increase the risk of hepatocellular carcinoma (HCC) beyond its effect on chronicity.The distribution of birth order, a proxy for mode and timing of HBV transmission, was compared in The Gambia between hepatitis B surface antigen (HBsAg)-positive HCC cases recruited from hospitals (n =72) and two HBsAg positive control groups without HCC: population-based controls from a community HBV screening (n = 392) and hospital based controls (n = 63). Results: HCC risk decreased with increasing birth order in the population-based case–control analysis. Using first birth order as the reference, the odds ratios were 0.52 (95% CI: 0.20–1.36) 0.52 (0.17–1.56) 0.57 (0.16–2.05) and 0.14 (0.03–0.64) for second, third, fourth and greater than fourth birth order respectively (P = 0.01). A similarinverse association was observed in the hospital-based case–control comparison (P = 0.04). Conclusions: Compared to controls, HCC cases had earlier birth order, a proxy for young maternal age and maternal HBV viraemia at birth. This finding suggests that in chronic HBV carriers perinatal mother-to-infant transmission may increase HCC risk more than horizontal transmission. Providing HBV vaccine within 24 hours of birth to interrupt perinatal transmission might reduce the incidence of HCC in The Gambia. These results have been published in Liver international, 2015.This work was part of the PhD of Yusuke Shimakawa who has been associated to the PROLIFICA project.
3.3 Anti-viral treatment - The West African Treatment Cohort (WATCH) study (WP3)
This “screen and treat” intervention was set up in The Gambia and Senegal and has been successfully completed. Using a rapid point-of-care test for HBs antigen, a total of 16,310 subjects have been screened over both countries (in domestic places, working places, blood banks and facilities). Out of them 2,132 individuals were tested positive giving a total prevalence at 13%. After excluding, the facility screening which might overestimate the HBV prevalence, the proportion of HBV-infected subjects remained high at 10.3% (1,569/15,199).
Each HBV-infected individual was invited to see a clinician trained in Hepatology for standardised comprehensive liver assessment: physical examination, abdominal ultrasound (Portable MyLab25Gold, Esaote, Cambridge, UK), fasting liver stiffness measurement (LSM) using hepatic transient elastography (Fibroscan®, FS402, Echosens®, France), and routine serum hematology and biochemistry tests. Blood samples were tested for hepatitis B e antigen (HBeAg) (ELISA-ETI-EBK Plus, Diasorin, Italy), antibodies to hepatitis C (AxSYM, anti-HCV, Abbott, USA) and hepatitis D viruses (ETI-AB-DELTAK-2, Diasorin, Italy). Antibodies to HIV-1 and HIV-2 were detected using enzyme immunoassay (EIA, Genscreen ULTRA HIV Ag-Ab, Bio-Rad, USA).
HBV DNA levels were measured using an in-house quantitative real-time polymerase chain reaction (q-PCR) (detection limit of 50 IU/ml) which was validated against a commercial HBV qPCR (Abbott, USA, excellent correlation with the commercial assay (r 2= 0.90)) (publication under revision). All samples were tested at the MRC Unit in Fajara. Quality control and HBV genotyping were performed by a reference laboratory in France (INSERM, Lyon). Detailed results per country are provided below.
3.3.1 The “screen and treat” intervention in The Gambia (lead: MRC The Gambia Unit, Prof. UDA)
Screening was conducted in Western Gambia where 750,000 people live in 1,450 census enumeration areas (EAs), defined by the Gambia Bureau of Statistics. We used EA as a sampling unit, and one EA consists of an entire village, a part of large village/town, or a cluster of small hamlets. Because HBV prevalence may differ between urban and rural populations, we first stratified 1,450 EAs into urban (n=1,197) and rural (n=253) communities. Then, from each stratum 27 EAs were selected by simple random sampling using a random number generator. In the selected EAs, all inhabitants aged ≥30 years were eligible for screening. We excluded those aged <30 years because the national hepatitis B vaccination programme started in 1990.6 We organised a meeting in each EA, with the help of the village or city head. Following community approval, a team of fieldworkers conducted a census by visiting all households to register the name, age and sex of all eligible people and invited them for screening. After pre-test counseling and written consent, finger-prick whole blood was tested for HBsAg using a POC test (Determine®, Alere, USA) whose performance was validated in the field (sensitivity 88.5%, specificity 100%).7 Results were provided on site to the participants with post-test counseling, and those who tested positive were referred to the closest liver clinic.
This work has been done with the support of the MRC unit but also the NPHL, Ministry of Health Gambia (including EFSTH, formerly RVTH) who provided two excellent fieldworkers: Alagie Sanneh and Debo Jallow have been consistently involved in the field work activity of the project. Their participation to each event has been smoothly facilitated by management from NPHL. They have had a significant impact on the sensitisation, screening and recruitment of patients from communities and the clinics. We also received a strong support from Dr Makie Taal (ministry of health) who facilitated the relation with EFSTH and its main blood bank.
Eligibility for treatment was determined according to the 2012 European Association for the Study of the Liver (EASL) guidelines 8. In the absence of contra-indications, tenofovir was provided free of charge (provided by Gilead Company). Adherence to treatment was assessed using the Morisky adherence scale.
Between December 2011 and January 2014, all selected EAs agreed to participate to the study and 5,980 out of 8,170 eligible people (68.9%, 95% CI: 65.0-72.4%) participated in the community screening. Between January and December 2013, 5,559 potential blood donors were screened (99.3% men) for HBs Ag in the main blood bank in The Gambia (Banjul).The prevalence of HBsAg was 8.8%, (95% CI: 7.9-9.7%) and 13.0% (95% CI: 12.1-13.9%) in the Gambian community and the main Gambian blood bank, respectively.
Linkage to care, which was defined as first attendance to the liver clinic was high, with 402 of 495 (81.3%) HBsAg-positive people from the community attending clinic; however, 300 out of 721 (41.6%) of HBsAg positive blood donors linked into care. Treatment indication was met by 18 of 402 (4.4%, 95% CI: 2.5-7.7%) patients from the communities and 29 (9.7%, 95% CI: 6.8-13.6%) from the blood donors. Male gender was strongly associated with treatment eligibility (OR: 4.35 95% CI: 1.50-12.58 P=0.007).
Figure 3: Flow chart of the screen and treat intervention in The Gambia
The majority (87·9%, 617/702, 95% CI: 85.3-90.1%) of the HBsAg-positive individuals who attended the clinic after screening was classified as inactive chronic carriers (Table 1 below); 3.3% (13/395) and 7.9% (23/291) were HBeAg-positive from the community and blood bank, respectively. From the community screening, 48 (12.2%) individuals had elevated ALT ≥40 IU/mL and 41 (10.7%) had HBV DNA ≥ 2,000 IU/ml. After excluding eleven participants without valid LSM, 10 (2.6%) had extensive fibrosis (F3) and 11 (2.9%) had cirrhosis (F4). Co-infection with HIV, HCV or HDV was observed in 3.3%, 1.0% and 2.0% of the participants, respectively. Among the HBV-infected blood donors, 55 (18.8%) had ALT ≥40 IU/mL, 38 (14.4%) had HBV DNA≥ 2,000 IU/ml and 50 (17.5%) had extensive fibrosis/cirrhosis (≥F3). The Table below summarizes the clinical and biological characteristics of the HBV-infected study population in The Gambia.
Table 4: Clinical and Biological characteristics of the HBV infected study population in The Gambia
These results have been submitted to Lancet Global Health and are under revision for publication.
3.3.2 The Keneba-Manduar longitudinal cohort of HBV-infected subjects in The Gambia
The Medical Research Council (MRC), the International Agency for Research on Cancer (IARC/WHO) and the Gambia Government have been supporting studies on HBV infection in The Gambia since the 1980’s, and have established a population-based open cohort of treatment-naïve chronic HBV carriers. As part of the PROLIFICA project, we analysed this Gambian cohort to describe i) the sero-clearance rates of hepatitis B e antigen (HBeAg) and surface antigen (HBsAg) in West Africa; ii) the incidence of HCC, end-stage liver disease (ESLD) and all-cause mortality; iii) longitudinal changes in serum HBV DNA and alanine transaminase (ALT) levels ; and iv) the prevalence of significant liver fibrosis and chronic liver disease requiring antiviral therapy according to the European Association for the Study of the Liver (EASL) criteria. We also estimated the HBV-related disease burden attributable to the mother-to-infant transmission in SSA by examining the associations between these outcomes and maternal HBsAg, a proxy for mother-to-infant HBV transmission. 405 chronic carriers (95% genotype E), recruited at a median age of 10.8 years, were followed for a median length of 28.4 years. Annually, 7.4% (95% CI: 6.3-8.8%) cleared HBeAg and 1.0% (0.8-1.2%) cleared HBsAg. The incidence of HCC was 55.5/100,000 carrier-years (95% CI: 24.9-123.5). In the 2012-2013 survey (n=301), 5.6% (95% CI: 3.4-9.1%) had significant liver fibrosis. HBV genotype A (versus E), chronic aflatoxin B1 exposure, and an HBsAg-positive mother, a proxy for mother-to-infant transmission, were risk factors for of liver fibrosis. Interestingly, 16.0% of chronic carriers were infected via mother-to-infant transmission, and 63.0% (95% CI: 47.0-74.1%) of the cases requiring antiviral therapy were attributable to this mode of transmission. These data are unique in Sub-Saharan Africa and have been published in Gut in 2015 9.
3.3.3 Development and validation of diagnostic tools for HBV-infection adapted to the African setting.
HBsAg rapid point-of-care tests
In order to ensure the quality of our screening strategy, we assessed the performances of different rapid tests in the real life African field conditions in The Gambia within the PROLIFICA programme 7. We confirmed very good performances of the tests (sensitivity around 90% and specificity 100%) in the field. Importantly, as part of this validation study, we found that participants with false negative results were all inactive chronic carriers (i.e not in need of treatment) with low levels of HBsAg. This means that we might have slightly underestimated the prevalence of HBV in our study population, BUT it is unlikely that we missed any patient in need of treatment since inactive chronic carriers do not fulfil the treatment eligibility criteria.
Non-invasive markers of liver fibrosis in West African HBV-infected subjects
The lack of non-invasive and inexpensive markers for the assessment of chronic hepatitis B infection is one of the main barriers for offering care and treatment of HBV chronically infected subjects in Africa. Only two studies assessed the diagnostic performances of non-invasive markers of fibrosis in sub-Saharan Africa with discrepant results. New simple and inexpensive markers of fibrosis for African resource-poor countries are highly needed. Therefore, within the PROLIFICA project, and since we used this device to assess the liver disease in our study population, we first confirmed the excellent diagnostic accuracy of transient elastography (Fibroscan®) for predicting severe fibrosis and cirrhosis in HBV-infected patients in a West African setting (AUROCs around to 0.90). However, the Fibroscan device remains expensive, thus we developed and validated a simple and inexpensive biomarker of liver fibrosis: the Gamma-glutamyl transpeptidase to Platelet ratio: GPR (GGT/Platelets ratio) index which might be a better routine laboratory test than APRI (currently recommend by the World Health Organization) and Fib-4 tests.
In-house HBV qPCR (WP3 & WP6)
Before starting the PROLIFICA intervention programme, we validated the performance of an in house technique which was assessed using coefficient of variation for repeatability and reproducibility with 2 levels of concentration, accuracy (A), uncertainty (U) and contamination. Plasma from HBV infected patient were then quantified by the in-house technique and the results obtained were compared. The in-house results were also anlaysed using a commercial method (m2000 sp/rt, Abbott Laboratories) for comparison. The correlation (r2) and concordance (biais) of in-house method and m2000 one were determinated. This in-house HBV qPCR against the Abbott real time commercial assay and we found an excellent correlation coefficient at r^2=0.90. This in-house assay had a lower limit of sensitivity of 50 IU/ml which is comparable with commercially available HBV DNA assays based on real-time PCR analysis. We also found that the repeatability for both concentration tested, coefficient of variation (CV) respectively 2.74 % for the intermediate concentration and 1.28% for the high concentration. For the reproducibility (evaluated after 10 days), CV showed 2.44% for the intermediate concentration and 1.37 % for the high concentration. The accuracy and uncertainty were A=2.79 %; U = ± 0. 25 and A= 0.86 %; U = ± 0, 27 for the high concentration. Contamination rate was 1.28 %. The comparison with the m2000sp/rt for 68 samples covering the clinical interest area, showed r2=0.904 for correlation and a biais = -0,259 for concordance. Since our study, therefore, showed a relative good performance for viral load quantification using SYBRGreen dye, the monitoring of patient was done using this method.
This work has been submitted to Journal of Viral Hepatitis (under revision).
3.3.3 The “screen and treat” intervention in Senegal (WP3, lead: Hopital Le Dantec, Dakar, Senegal)
In Senegal, the screening intervention was performed in domestic communities (urban and rural areas) around Thies the second most important Senegalese city with the support of the Medical school of Thies (Prof. Mourtalla Ka). Screening for HBsAg was also performed in work places (Mines) and in facilities (see Table 2 below).
A total of 4,774 persons have been screened, including 3,660 persons from the community (domestic and work places).
Table 5: Screening results in Senegal
To date, the clinical and virological assessment of more than 700 HBV-infected subjects has been performed. Data are under analysis and will be presented as an oral communication in the next Francophone conference of HIV and Viral hepatitis in resource-limited countries (AFRAVIH, Brussels); the publication of the Senegalese “screen & treat” intervention is under preparation.
In the observation cohort (those who are HBsAg + and who did not meet EASL criteria for treatment), preliminary results (n= 138) showed a prevalence of HBeAg of less than 1% (1/138) and 99% of anti-HBe Ab. AgHBs quantification mean value was 3 times less (1465 UI [ 0.050-14 936]) than those under treatment. Delta co-infection in this group was 2.8 % (4/145).
Figure 4: Proportion of HBV infected subjects with low, medium or high viral load
3.3.5 Treated cohort in Gambia and Senegal (lead MRC Gambia and Le Dantec)
The treated cohort is derived from the community screening (WATCH) and the end-stage liver disease study (HC4 study). A cohort of 242 patients under tenofovir has been set up in both countries.
In Gambia, 127 patients are currently under tenfovir.
None of the eligible patients refused antiviral therapy. Twelve months after starting tenofovir therapy, 38 patients (80.9%) had a high adherence score, 7 (14.9%) a medium adherence and 2 (4.3%) a low adherence. At 12 month of treatment, 43/47 (91.5%) achieved a virological response defined by undetectable HBV viral load; 38 (79.7%) had normal transaminases (< 40 IU/mL), and 9 (19%) had ALT≥40 IU/mL (median 46 (43-62) IU/mL) with a baseline median (IQR) ALT at 49 (34-104) IU/mL. No clinical or biological adverse events were observed after 12 months of treatment.
In Senegal, 115 ? patients are under treatment, the first patient under TDF was recruited in April 2013. We monitored HBV viral load replication during TDF treatment using in-house qPCR. Our first results showed at J0 a median viral load at 6.58 log copies/ml. After only 3 months under treatment, the initial viral load decreased to 3.85 log copies/ml. After 1 year under TDF, 89 % (39/44) of patients tested had undetectable viral load. For those who have detectable viral load 11 % (5/44), 60% over them (3/5) have a viral load of less than 300 copies/ml or 2.47 log copies/mL (≈ 80 UI) and the most elevated value was 1200 copies/ml or 3.07 log copies/ml (≈ 300 UI).
Figure 5: Viral kinetic of HBV DNA in the Senegalese patients on tenofovir
3.4 Biomarker discovery and mechanistic investigation through proteomics (WP4)
3.4.1 Point of Care testing device for monitoring patients at risk of HCC based on a pre-existing dataset of candidate protein markers (WP4, lead: Xeptagen company)
Within the PROLIFICA consortium, XEPTAGEN has developed a new Point of Care Testing (POCT) device for enabling the surveillance of patients at risk of hepatocellular carcinoma (HCC) development based on the assessment of circulating levels of Squamous Cell Carcinoma Antigen (SCCA)-IgM immune complexes.
SCCA-IgM has been proposed as a novel biomarker for surveillance of patients at risk of HCC growth since the increase above 200 AU/mL of serum levels of SCCA-IgM may indicate a worsening of the liver disease in patients with cirrhosis, with a higher risk of developing HCC. Occurrence of SCCA-IgM at high levels (>500 AU/mL) enables the identification of subjects at higher risk of HCC development at least one year before the clinical evidence of tumor thus allowing a serological HCC surveillance.
The POCT has been developed as a lateral flow assay (LFA) for the detection of SCCA-IgM by gold nanoparticles- conjugated antibodies. The LFA has been made up of four membranes: sample pad, conjugate pad, nitrocellulose membrane and absorbent pad. All membranes are partially overlapped and assembled in a LFA plastic cassette. Cassettes have a hole above the sample pad for sample of serum or blood introduction. The prototype LFA has been validated with more than 300 samples demonstrating its diagnostic accuracy in detecting patients with SCCA-IgM > 200 AU/mL.
The adoption of the assessment of SCCA-IgM immune complexes using a low cost disposable LFA POCT for surveillance purposes may represent a significant step forward in the management of patients suffering from chronic liver diseases in low income countries.
3.4.2 Discovery & Validation Strategies (WP4 lead IARC)
Following the initial discovery strategies conducted at the Fred Hutchinson Cancer Research Centre and the University of Oxford in collaboration with other sites; a select group of well performing proteins have been entered into various stages of validation by measurement in liver disease cases from numerous populations.
Two independent biomarker discovery strategies were conducted, both of which are presented in full in the previous report. To summarize, both strategies involved a comprehensive deep proteome analysis of pooled plasma samples put through a pipeline of immunodepletion, reverse-phase chromatography and gel electrophoresis in one strategy and isoelectric focussing in the other with final analysis performed by tandem mass spectrometry. Robust data mining strategies were then employed to generate a priority list of shortlisted biomarkers based on the comparison of controls and case groupings, probability of association with a specific pathological status, the nature and abundance of target proteins, suspected biological relevance and finally the availability of affordable and robust methods for further assessment .
Following this multi-dimentional discovery approach, the proposed markers have been amalgamated and prioritized for measurement in distinct populations mainly through antibody detection methods (ELISA). Robust statistical methods have also been implemented, allowing the use of individual subject measurements to assess and compare expression trends through various disease stages and populations.
Extensive validation
Two independent validation strategies were executed on proteins identified from the biomarker discovery pipelines previously discussed. The primary strategy was aimed at comparing differential protein expression trends with established liver disease markers to assess significance and reproducibility of discriminatory expression patterns suggested by mass spectrometry.
Previous reports have detailed the validation of these markers in populations from the Case-Control Gambia Liver Cancer Study (GLCS) and Jos University Teaching Hospital (JUTH), neither of which were part of PROLIFICA. Comprehensive assays have also been run and reported on PROLIFICA Gambia patients as presented below.
Independent population validation
The proteins proposed from the GLCS discovery track were primarily validated in an independent cohort of African subjects collected from the Jos University Teaching Hospital (JUTH) in Nigeria. In total, plasma aliquots for 55 subjects classified as either Nigerian healthy controls (NN), Nigerian Asymptomatic Carriers (NASC), Nigerian cirrhotics (NCirr) or Nigerian HCCs (NHCC). The categorizations were based on tests for HBV, HCV and HIV virus status, AFP, biochemical liver indices, US, or Computed Tomography (CT), endoscopy and or liver biopsy. Table 6 summarizes the key results from the Jos participants; a full evaluation of these findings and those from the GLCS subjects is made in the published manuscript 10.
Table 6: Summary of sample numbers tested per protein along with AUC’s, confidence intervals and associated p-values for JUTH subjects
Protein & Sample Numbers Tested NN vs. NHCC NN vs. NCirr NCirr vs. NHCC
ApoA1 NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.7667 (95% CI: 0.6020 to0.9313)
p=0.01800 AUC: 0.6167 (95% CI: 0.3154 to 0.9179)
p=0.4477 AUC: 0.6587 (95% CI: 0.3987 to 0.9187) p=0.2435
AAT NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.7857 (95% CI: 0.6039 to 0.9676) p=0.01126 AUC: 0.5833 (CI: 0.2647 to 0.9019)
p=0.5876 AUC: 0.8016 (95% CI: 0.6360 to 0.9672) p=0.02672
CC3 NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.6238 (95% CI: 0.4200 to 0.8277) p=0.2720 AUC: 0.8167 (95% CI: 0.5921 to 1.042) p=0.03937 AUC: 0.8571 (95% CI: 0.6985 to 1.016) p=0.008705
HPX NN: 10, NASC: 18, NCirr: 6 HCC: 21 AUC: 0.8048 (95% CI: 0.6455 to 0.9640) p=0.006864 AUC: 1.000 (95% CI: 1.000 to 1.000) p=0.001146 AUC: 0.7937 (95% CI: 0.5594 to 1.028) p=0.03098
With this knowledge in hand, two of the best performing markers proposed by each of the groups involved in the discovery, validation and statistical combinations were selected for direct measurement by ELISA in the large, multi-country samples being collected under PROLIFICA. The four most discriminant proteins selected for measurement were HPX, Apo A1, OPN & LTBP2. The latter two proteins were identified within a sister study performed on Gambian, Thai and French participants 11.
The diagnostic criteria for various stages of liver disease utilized were as follows:
1. Controls
• Healthy Controls: HBV & HIV negative
2. Cases: Taken from HC4 Case-Control arm with cases of the following profiles:
• Fibrosis: Liver stiffness measurement (LSM) ≥7 kPa by Fibroscan
• Chronic HBV: HBV positive, no signs of progressive liver disease by US/biochemistry
• LC: Presence of either histological confirmation, LSM ≥11 by Fibroscan, or ultrasonographic diagnosis of liver cirrhosis
HCC
• Clinical, No Histology: Focal liver lesion ≥2cm which is consistent with HCC by US and AFP levels ≥ 400 ng/ml in patients with liver cirrhosis
• Histologically Confirmed: Combination of Ultrasound, AFP, Clinical Indices & a liver biopsy confirmation.
ALL HIV positive cases were excluded.
Samples from Nigeria (JUTH)
At the current phase of our proteomics studies, a total of 238 case and control samples were selected from the HC4 arm recruited in Nigeria. A summary of the clinical parameters for these patients is presented in Table 7. A total of 238 samples have so far been measured for OPN and LTBP2 with assays for ApoA1 and HPX are to be performed shortly. The validation experiments were conducted by ELISA based antibody detection methods.
Table 7: Summary of key clinical parameters for JUTH PROLIFICA subjects
In brief, antibodies from the protein of interest were adsorbed to the base of a microtitre well plate, reacting with and binding specific antigens in the diluted plasma. Any unbound protein was removed by four consecutive wash steps followed by incubation with Horseradish Peroxidase (HRP) conjugated antibodies to the protein of interest. This was followed by a final quadruple wash step, after which the chromomeric substrate 3,3’5,5’-tetramethylbenzidine (TMB) was added and allowed to develop to optimal intensity, for approximately 30 minutes and the reaction then stopped by addition of an acid solution. In their specific measurements, samples for OPN were diluted 1:100 and LTBP2 1:5.
Extrapolation and data analyses were conducted using the GraphPad Prism software. Key conclusions from our results thus far show that LTBP2 may be differentially expressed between various stages of CLD; however it appears to not be a robust enough marker to withstand heterogeneity in case definitions and aetiology of liver disease (Figure 6 & Figure 7).
Figure 6: Dot plot with accompanying table showing summary of LTBP2 expression levels across control and case groups in PROLIFICA Gambia participants.
Figure 7: Dot plot with accompanying table showing summary of LTBP2 expression levels across control and case groups in PROLIFICA Gambia participants.
Osteopontin (figure 2a, 2b) has been consistent in all its measurements in numerous populations as being over-expressed in HCC tumorigenesis. This is a marker which warrants further study into the specific mechanisms by which it is elevated and how specific these processes are to HCC.
Figure 8: Dot plot showing summary of OPN expression levels across control and case groups in PROLIFICA Gambia participants.
Figure 9: Dot plot with accompanying table showing summary of OPN expression levels across control and case groups in PROLIFICA Nigeria participants.
Compilation of Serological, Biochemical and Results
To enable the selection of well described and clinically distinct cases and controls in functional biomarker studies, it is important that key assay results are reported for selected patient populations. This has been a challenging area and one where the team within this
Conclusion
IARC have successfully set up the main project biobank, which will be available to the Partners post the current project completion. In addition, they have identified a selection of well described and clinically distinct cases potential functional biomarkers from their studies. In particular OPN and Osteopontin merit further investigation.
3.5 Development of metabonomic markers for the diagnosis of HCC (WP5)
This work has been conducted in collaboration with IARC, France.
Urine and blood (serum) samples taken from patients in the West African clinical studies are periodically transferred from source to either a specific PROLIFICA Biobank at IARC or to Imperial College for metabonomic biomarker studies. To date, samples from both Gambia and Nigeria have been sent to Imperial College, where they have been processed for the identification of liver cancer biomarkers.
The ultimate goal of this study is the development of a urinary-based test for use at the village level in Africa for screening of at-risk populations.
Urine samples from Nigeria and Gambia have been sent to Imperial College for metabonomic studies, which enabled the detection and identification of metabolites responsible for mediating the phenotypic expression of altered metabolism. Metabolic profiling involves the combination of spectroscopic techniques of analysis, nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS), with advance multivariate statistical tools to interrogate the spectral data produced (Figures 5 & 6).
Figure 10: Metabonomic study workflow
As explained in the previous report, 1H NMR was used as a first stage metabolic screening tool to metabolically phenotype 290 subjects (comprising patients with HCC, cirrhosis (Cir), non-cirrhotic liver disease (DC) and healthy controls (NC)). Urine samples from a further cohort of 463 subjects were analysed, the results of which validated the initial cohort. These analyses were completed by a Nigerian medical PhD student, Dr Nimzing Ladep, studying at Imperial.
Figure 11: Training and validation sets of urine samples analysed by 1H-NMR
The urinary metabotype of patients with HCC was distinct from those with DC and NC. The lower performance of HCC versus Cir was not unexpected, because most HCC cases occur in the background of cirrhosis. The intensities (relative concentrations) of several metabolites were significantly increased in urine of participants with HCC, compared to controls (Fig.3).
Figure 12: Urinary metabolites in HCC shown in a pair-wise comparison of A) HCC vs. Healthy, B) HCC vs. Liver disease; and C) HCC vs. Cirrhosis in both training and validation sets of experiments. Metabolites in bold texts were significantly up regulated in HCC while those in italics were significantly down regulated in HCC. Those represented by regular texts did not attain any significant association.
The diagnostic performance of a panel for HCC versus cirrhosis, comprising four metabolites (inosine, indole-3-acetate, N-acetylated amino acid, and galactose) was significantly higher than serum AFP, a finding that was corroborated in a validation cohort (Table 6).
Table 8: Diagnostic accuracies of urinary models compared to serum AFP in the discrimination of HCC from three control groups in the training and validation sets of subjects recruited in Nigeria and Gambia
Metabolites that contributed most strongly in discriminating between groups were subjected to further analyses. Specifically, liver disease controls were categorized to one (1), intermediate HCC (BCLC A-C) to two (2), and advanced HCC (BCLC D) to three (3), and the scores were regressed against the spectral profiles. The relative concentration of methionine, acetylcarnitine, indole-3-acetate, N-acetylated amino acid, dimethylglycine, 1-methylnicotinamide, and creatine were significantly positively correlated to category and clinical stage of HCC and are shown in figure 4. In comparison, there was poor correlation with stage of HCC by serum AFP (P50.7).
Figure 13: Correlation of relative concentration of urinary metabolites by disease category and BCLC stage of HCC
This study provides some evidence that a battery of discriminatory metabolites may be a future diagnostic accuracy that would be simple to perform and a paper was published last year on these data: Ladep NG, Dona AC, Lewis MR, Crossey MM, Lemoine M, Okeke E, Shimakawa Y, Duguru M, Njai HF, Fye HK, Taal M, Chetwood J, Kasstan B, Khan SA, Garside DA, Wijeyesekera A, Thillainayagam AV, Banwat E, Thursz MR, Nicholson JK, Njie R, Holmes E, Taylor-Robinson SD. “Discovery and validation of urinary metabotypes for the diagnosis of hepatocellular carcinoma (HCC) in West Africans”. Hepatology. 2014. doi: 10.1002/hep.27264
Complementary and more sensitive mass spectrometry (MS) methods followed up NMR analysis to increase the metabolome coverage and for practical reasons: NMR is a preserve of developed world research institutes whereas installing and running MS is cheaper and panels of validated metabolites could be translated into the clinical arena in the developing world. In this regard, the PROLIFICA project has obtained a triple quadrupole (TQ) mass spectrometer for targeted analyses, based at the MRC, Gambia, for future analysis of samples across West Africa.
Choosing which metabolites to analyse in the targeted MS assay is a process driven by hypotheses formed from previous experimental evidence. Untargeted metabolomics experiments are performed using high resolution accurate instrument such as quadrupole time of flight (Q-ToF) mass spectrometers. Due to the complex nature of the biological samples, separation by ultra-performance liquid chromatography (UPLC) is performed before the MS analysis.
Subsets of both urine sample sets were selected to be analysed by UPLC-QTOF-MS at Imperial College (Figure 10). MS was performed in both positive and negative ion electrospray (ESI+ and ESI-) modes.
Figure 14: Training and validation sets of urine samples analysed by UPLC-ESI-MS
A quality control (QC) sample was prepared by pooling identical aliquots of all the samples. The QC sample was used to condition the column prior to the analysis and then re-injected at regular intervals throughout the run (every 10 samples) to ensure system suitability and stability.
Further more comprehensive data processing and statistical analysis than the one explained in the previous report was performed. Principal components analysis (PCA) showed tight clustering of the QC samples in both the training and validation sets in ESI+ and ESI- modes, indicating good quality of all the data obtained and small analytical variation. Supervised models were subsequently generated to maximize the separation between classes and identify biomarkers associated with specific classes corresponding to the nature and stage of disease (Figure 11).
Figure 15: Multivariate statistical analysis of the validation set run by UPLC-(-)ESI-MS. A) Scores plot of the unsupervised analysis by principal component analysis (PCA). B) Scores and S-plots of the supervised analysis by orthogonal partial least square regression discriminant analysis (OPLS-DA).
Using the S-plot from the supervised analysis, the variable influence of the models was examined. The up-regulated metabolites in HCC were putatively identified using accurate mass (m/z), isotopic pattern, fragmentation pattern and retention time information. m/z values were searched against online databases including HMDB (http://www.hmdb.ca/) and METLIN (http://metlin.scripps.edu/index.php). When possible, authentic standards were analysed for definitive identification.
The urinary metabotype of patients with HCC was distinct from those with Cir, DC and NC with areas under the receiver operating characteristic (ROC) curves of 0.86(0.78-0.94) 0.93(0.89-0.97) and 0.89(0.80-0.98) in the training set; and 0.81(0.73-0.89) 0.96(0.94-0.99) and 0.90(0.85-0.96) respectively in the validation cohort, showing significantly higher diagnostic accuracy of liver cancer than serum AFP.
From the initial list of discriminant metabolites tentatively identified, 11 were confirmed with standards (Table 7) and used to develop the targeted method that will be run in the TQ instrument in the MRC Gambia. Method development was done by a Gambian postdoctoral researcher, Dr. Haddy Fye, trained at Imperial College.
Table 9: Up-regulated metabolites in HCC used for the development of the targeted UPLC-MS method. Metabolite IDs confirmed with standards
ID RT (min) m/z Ion Formula
1-Methylnicotinamide 0.57 137.071 [M+H]+ C7H9N2O
Acetylcarnitine 0.97 204.124 [M+H]+ C9H17NO4
Propionylcarnitine 1.58 218.139 [M+H]+ C10H19O4
Hydroxyphenyllactic acid 2.65 181.05 [M-H]- C9H10O4
Homovanillic acid 2.78 261.006 [M-H]- C9H10O7S
Kynurenic acid 2.85 190.051 [M+H]+ C10H7NO3
4-Hydroxybenzaldehyde 3.59 121.028 [M-H]- C7H6O2
Indole lactic acid 4.38 204.065 [M-H]- C11H11NO3
L-Octanoylcarnitine 5.59 288.217 [M+H]+ C15H29NO4
Glycocholic acid 7.38 464.301 [M-H]- C26H43NO6
Glycochenodeoxycholic acid-3-sulfate 7.85 528.263 [M-H]- C26H43NO8S
The metabolic significance of the presence of some of these low molecular compounds being higher in the urine of patients with HCC requires further studies. For example, propionylcarnitine is known to be in abundance in the urine of individuals with methylmalonyl-CoA mutase deficiency. Mutation of hepatocytes during the process of carcinogenesis may thus lead to defective isomerization of methylmalonyl-CoA to succinyl-CoA, which could impact the citric acid cycle. Indolelactic acid, a product of tryptophan metabolism is common in urine of humans. Its presence in higher concentration in the urine of patients with HCC suggests a differential metabolism of this amino acid in this group of patients. Glycocholic acid has been described to be perturbed in chronic liver disease as well as HCC patients and in whom there is significant trend of increase in concentration relative to severity of liver damage. The foregoing thus opens up a world of further basic science studies regarding the role of these urinary metabolites in the process and survival of hepatocarcinogenesis. For example, increases in urine of these metabolites may not necessarily be representative of the fact that they are increased in concentration in blood of HCC patients. Differential renal function may alter excretion of metabolites in the urine. Also, the tumour microenvironment of HCC may not necessarily be replete with these metabolites. Further studies involving the use of sera as well as liver cancer tissues could throw more light regarding role of these compounds in HCC.
Figure 16 shows the extracted ion chromatograms (EIC) of the discriminant metabolites up-regulated in HCC in a QC urine sample run by the profiling method in both ion modes (A) as well as the EIC of the corresponding standards run by the developed targeted assay (B).
Figure 16:Extracted ion chromatograms (EIC) of the discriminant metabolites up-regulated in HCC. A) EIC in QC urine sample run by the profiling method in positive (A1) and negative (A2) ion mode. B) EIC in Stadard mix run by the developed targeted assay 1: 1-Methylnicotinamide. 2: Acetylcarnitine. 3:Propionylcarnitine. 4:Hydroxyphenyllactic acid. 5: Homovanillic acid. 6:Kynurenic acid. 7:4-Hydroxybenzaldehyde. 8:Indole lactic acid. 9: L-Octanoylcarnitine. 10: Glycocholic acid. 11: Glycochenodeoxycholic acid-3-sulfate
The main differences in both UPLC methods are highlighted in the Table 8 below. The use of a TQ instrument in the targeted method allows the detection of compounds in negative and positive in a single chromatographic run, reducing the run time by half. By changing the column dimensions, the flow rate was decreased 3 times, as well as the run time per sample. Reduction in run time and in flow rate means a noticeable decrease of the consumption of solvent necessary to analyse a sample, which is critical in the developing world.
Table 10: Comparison of the profiling and chromatographic methods
UPLC method Profiling method Targeted method
Run time 25 min per sample
(1 chromatographic run of 12.5 min per polarity) 10 min per sample
(1 single run, both polarities)
Column type HSS T3, 2.1 X 150 mm HSS T3, 1 X 100 mm
Column temperature 45 °C 40°C
Flow rate 0.6 ml/min 0.2 ml/min
The developed method has been transferred to the instrument in the MRC Gambia, and Dr. Haddy Fye is currently validating and assessing its performance by analysing new urine samples collected in the Gambia.
Imperial College will continue to investigate the performance of the panel of identified biomarkers as an early diagnostic tool for liver cancer.
The work to date has been published in Hepatology, and further papers will ensue:
3.6 HBV & HCV infection: virological factors for HCC development (WP6)
The first objective of this workpackage was to implement standardised protocols for HBV molecular and serodiagnostics and evaluation/quality control of standardized protocol in all sites and this has been achieved. The second workpackage objective was to identify and characterize HBV genotypes and variants from the HC4 and WATCH platforms
3.6.1 Establish an in-house HBV DNA assay and transfer skills to Gambia & Senegal
Through training within INSERM and with subsequent help from partners, we have been able to transfer the in-house qPCR techniques to the African labs. We crosschecked all the samples for viral load qPCR and found comparable results (Figure 12) in the assay performed in INSERM and also in Gambia and Senegal, as part of our quality control programme. The technology transfer of the HBV qPCR has been achieved successfully. The assay design, validation and implementation of in-house qPCR in the PROLIFICA programme is currently being written as a publishable document.
Figure 17: The correlation curve showing the significantly similar results obtained with the same samples tested in MRC Gambia and INSERM as well as Senegal and INSERM
3.6.2 Viral genotype, sequence and variant analysis
Establishment of full Surface gene amplification and HBV full genome amplification:
We have performed full length amplification HBV genome as well as full HBV surface gene amplification from the test batch samples successfully and sequenced (from commercial source) to determine the genotypes. The common procedures were followed in all the subsequent screening and surveillance programme (Figure 13) The genotyping was primarily performed with the help of online NCBI genotyping tools by using the sequenced data (Figure 14) and further analysed using HBV Data base (https://hbvdb.ibcp.fr/HBVdb/) and other bioinformatics software (e.g. Clustal, MEGA)
Figure 18: Diagrammatic representation of the strategy and result HBV DNA amplification and sequencing
Figure 19: Representative picture showing HBV genotyping data and the results from NCBI viral genotyping tools.
Samples for the study
To date, we have received 3 types of cohort samples from Gambia through the IARC biobank. The WATCH cohort consists of 495 HBsAg positive samples. The Blood Donor Cohort includes 300 HBsAg samples and the HC4 cohort, more than 200 HBsAg positive samples. The WATCH cohort is a community based screening cohort, whereas the blood donor cohort is a facility based cohort. Patients are screened for HBsAg in the blood bank before blood donation. Those that are positive are recruited into the blood donor cohort. On the other hand, the HC4 cohort is a hospital based patient cohort who are screened for HCC. They are diagnosed as having HCC, liver cirrhosis or chronic HBV with fibrosis depending on their clinical, biochemical, pathological and radiological findings. For our mutation screening and genotyping studies, we have screened for HBV by DNA extraction and PCR on full surface gene and subsequent sequencing on 195 WATCH samples (based on >150 IU/mL HBV DNA as the cut off), 121 HC4 HBsAg+ samples and 64 HBsAg- HC4 samples (to check for occult HBV), and 98 HBsAg+ blood donor samples.
HBV full surface sequencing by direct PCR:
We have PCR amplified and sequenced the full surface gene (PreS1/PreS2/S overlapping a major part of polymerase including the RT region) as described previously. The results are as follows: - 122 WATCH samples (122/195, 62% success), 95 HBsAg+ HC4 (95/121, 78% success), 22 HBsAg- HC4 (Occult HBV, 22/64, 34% success) and 90 HBsAg+ blood donor (90/98, 90% success). We are in the process of submitting this sequence data to the NCBI database. The majority of the sequences generated in this PROLIFICA study are either of HBV/Genotype E or HBV/Gen A (see the genotype data below). This will be the largest collection of HBV/Gen E data from Sub Saharan Africa (SSA).
HBV full surface sequence by cloning of PCR products:
Due to the presence of PreS in frame deletion mutant (see the deletion mutant data) in a large number of samples, we have had a problem with the PCR based sequencing. To overcome this problem, we cloned the PCR products in a TA cloning vector (TOPO TA Cloning kit, Life Technologies) and then sequenced. We have selected 5-10 clones from each individual and thus generated over 100 cloned full surface sequences from 21 individuals (WATCH and HC4 cohort). This has helped us to understand and confirm the deletion mutant and their quasispecies complexity (see below). The sequence data will be submitted to the NCBI database very soon.
Generation of Full Length HBV sequence by cloning:
To further explore the full genome HBV genetic variability, we have amplified the full genome HBV (as described by Ghosh S et al., 2010) and subsequently cloned into PCR XL TOPO vector (TOPO XL cloning kit, Life technologies) and sequenced by multiple primers. The sequence fragments were joined by seqscape software (Applied Biosystem). Thus we generated 25 full genome HBV clone from 8 samples.
Generation of sequence of PreC/C region:
In order to explore the prevalence of Basal core promoter (BCP – 1762, 1764) and PreCore promoter (CP - 1896, 1899 stop codon) mutations, we have amplified and sequenced the region between nt1680-nt2460 of the HBV genome. This is very important, as >95% of the HBsAg+ individuals have HBeAg negative infection with very low viremia (>2000 IU/mL). We have successfully sequenced 60 individuals (47 HBeAg negative and 13 HBeAg positive) from the WATCH and HC4 cohort. However, the success of PCR amplification is low (>55%) in this region, due to the very low viral load and complexity of the HBV genome in this region. The following data table (Table 9) suggests that BCP/CP mutation is very high (80-90%) and this may be one of the major reasons for the high incidence of HBeAg negative infection (90 - 95%) in SSA.
Table 11: BCP/CP mutation profile
HBsAg quantification:
In order to understand and explore the role of HBV surface gene mutation on the expression of HBsAg in blood, we have quantified HBsAg from the serum of all HBs sequenced samples (217 HBsAg + samples from the HC4 and WATCH cohort). The quantification was performed using Cobas e 411 immunoassay analyser (Roche). We have also tried to see if there is any correlation between HBV viral load and HBsAg quantification (Figure 15). Interestingly, we found some individuals who had a very high viral load and low HBsAg or vice versa. This phenomenon may be related to genetic variability of the HBs gene.
Figure 20: Correlation between HBV viral Load (IU/mL) and HBsAg (IU/mL)
HBV Genotyping:
To determine the genotype of the strains and their phylogenetic relation, a tree was built on the 221 Surface gene sequence (only in-del free region) obtained in this study, along with the reference sequence of all known HBV genotype/subtypes obtained from the NCBI database. A ‘Maximum-Likelihood’ tree was built using the Jukes-Cantor model and 1000 bootstrap events in MEGA6 software. The final tree was edited in Figtree software (Figure 16). This showed that the majority of the individuals had Genotype E (>82%, fluorescence blue) and the rest had Genotype A (15.4%, fluorescence green), except for 4, which may have a recombinant of Genotype E and A (red). Further full length cloning based sequencing studies are needed to confirm (Figure 17). We have been unable to amplify the full length from the 4 samples, due to a limited number of biological specimens and a very low viral load.
Figure 21: Phylogenetic tree reconstruction of full length surface (PreS1/PreS2/S) sequences.
Figure 22: Probable site of recombination in the 2 strains were shown in the picture bellow using web-based NCBI genotyping tools of the full surface gene.
Role of HBV PreS deletion in the development of HCC in SSA:
We have observed a high incidence of PreS deletion (>34%) among the sequence generated in the PROLIFICA samples. To further explore its role in HCC in SSA, we have analysed 217 full surface sequences from the HC4 and WATCH cohorts with genotype E or A in relation to their clinical phenotype. We excluded 4 individuals who may carry recombinant strains, due to its very low prevalence. We divided the 217 individuals into 3 groups: - Chronic HBV (No LC/HCC), Liver Cirrhosis (LC) and Hepatocellular Carcinoma (HCC). We have 42 LC and 36 HCC. Comparing the groups, we have observed PreS2 deletion, Genotype A individuals are associated with the development of HCC in SSA. It is also possible that the lower incidence of HBsAg among the HCC group, compared to the No LC/HCC group (Table 10), may be related to the PreS2 deletion. We are currently exploring these findings.
Table 12: Demographic, Biochemical and HBV profiles
Type of PreS deletion in PROLIFICA samples:
We have mainly identified 3 types of PreS deletion strains among the strains isolated from different individuals (Figure 18). PreS region of HBV is frequently deleted by immune selection pressure, as it contains several T and B cell epitopes and is often found to be associated with liver cirrhosis or HCC (Pollicino et al., 2014). In our study, the most prevalent was the PreS2 deletion (>32%), followed by truncated S (9%) and PreS1 deletion (3%). These deletions were present in individuals with both the genotype (E and A). The most important observation of this sequencing study is the high incidence of PreS2 deletion in a specific region (aa5-20 of preS2) and its strong association with the HCC (>55%). This type of mutant was previously described by Pollicino et al (2014) and Su et al (2008 and 2014) and implicated in ground glass hepatocyte (GGH) formation and liver pathogenesis by oxidative stress.
Figure 23: Representative amino acid sequence alignment to show the different deletion strains
Presence of wild type and mutant PreS2 deletion strains in the same individuals:
To explore the quasispecies structure (since we have had difficulties in direct PCR product sequencing) within the same individuals, we sequenced the PCR products after cloning. The cloning and sequencing revealed that almost all the individuals carrying the PreS2 mutant have wildtype strains. Also, the in frame deletions are of different lengths within the same individual (Figure 19). To further confirm and explore the complex wildtype and mutant ratio, we are currently exploring selected samples for ultra-deep pyrosequencing (UDPS).
Figure 24: Wild type and mutant PreS2 quasispecies within the same individual
Immunohistochemistry for anti-HBs on liver biopsy:
It has previously been described (Su et al., 2008 and 2014) that the PreS1 and PreS2 deletion mutant forms GGH type 1 and type 2 are an important indication of liver pathogenesis. To explore this phenomenon, we also did immunohistochemistry (IHC) on the liver biopsy samples from PROLIFICA study patients and observed the same staining profile (Figure 20). Currently, we are investigating a large series of biopsy samples matched with wild type and mutant strains, to further explore/establish this hypothesis.
Figure 25: Ground Glass hepatocytes and PreS deletion
In-vitro studies on PreS2 deletion mutant:
To further explore the impact of PreS2 deletion mutant on the liver pathology, we designed some in-vitro studies (Figure 21). We cloned both full length Wild type and Mutant strains and amplified them with a primer containing SapI RE site. After SapI digestion and the purification of PCR products, we transfected into the Huh7 cells. Flowing transfection, we collected the supernatant and checked for HBsAg secretion by quantitative ELISA. We found that the mutants have less secretion of HBsAgcompared to the wild type. However, when we co-transfect the wild type and mutant, the secretion of HBsAg was restored to near normal. We also performed RNA blots to confirm normal replication and transcription capacity of both wild type and mutant strains within the cells.
Figure 26: In-vitro characterization of PreS2 deletion mutant
Immunoflourescence studies on in-vitro transfected cells:
In order to explore the intracellular HBsAg present in the wild type and mutant containing cells, we undertook immunofluorescence on the transfected cells after fixation. The confocal visualization of the transfected cells confirmed that PreS2 mutants store a high amount of HBsAg within the cells (Figure 22) and a high probability of having oxidative stress, DNA damage and ER stress. We have unsuccessfully tried to explore this hypothesis by western blot and qPCR (mRNA expression) using several ER stress markers. It is possible that due to very low transfection efficiency (5-15%), we are not able to see the difference on the whole cell extract. However, we are currently performing co-localization of HBV infected cells (with or without mutant) and different oxidative stress and DNA damage markers via IF, to further explore the possibility.
Figure 27: Immuno-fluorescence of transfected Huh7 cell lines
3.6.3 Occult HBV:
We have found 22 patients (22/64, 34%) in the HC4 cohort who are negative for HBsAg and positive for the HBV DNA test. We have also checked for the quantitative HBsAg in these patient’s samples and found all were below the detection limit (>0.025 IU/ml) for confirmation of the sero-negative status. We are currently exploring the HBeAg, HBcAg and Anti-HBs among these samples. All 22 samples were sequenced for the full surface gene and we are currently looking for any genetic variability of the strains..
3.6.4 Current and future virology work:
• We are currently analyzing all the sequences for point mutation.
• Undertaking UDPS to establish the wild type mutant quasispecies ratio within an individual.
• Using IHC on a series of biopsy samples to further explore the role of GGH1 and GGH2.
• Investigating the co-localization of oxidative stress markers on the transfected cells containing wildtype or mutant HBs
• Submitting all the generated sequence data into the INSERM database,
• Exploration of occult HBV and its molecular basis (if any).
• Submission of 2 manuscripts on HBV viral load, HBV virology and PreS deletion from the PROLIFICA study by November 2015.
INSERM has successfully completed their objectives for this period: the implementation of standardized protocols and the technology transfer of HBV qPCR to PROLIFICA partners. They are well placed to complete their project objectives.
3.7 Training & capacity building (WP7)
Objective: to build capacity by creating reference centres
All the partners (except Xeptagen) have been involved in capacity building
3.7.1 Clinical training and capacity building
In The Gambia five clinicians have been trained in hepatology: Dr Gibril Ndow, Dr Louise Saar, Dr Abu Kambi, Dr Momodou-Lamin Tekanyi, and Dr Sheikh-Omar Bittaye and one nurse (Saydiba Tamba).
The training consisted in lectures, supervision of clinical work in the liver clinic. All have been trained on the use of Fibroscan and Liver ultrasound and liver biopsy. They have also been trained in the liver diagnosis, management of chronic hepatitis B and end-stage liver disease.
The nurse has been train to manage HBV patients under treatment.
Two medical officers employed at the EFSTH (previously named RVTH) were recruited to join PROLIFICA as trainee hepatologists.
The trainees dedicated two days of the week entirely for this training. They worked closely with and under supervision of PI and consultant hepatologist Dr Ramou Njie, as well as Dr Maud Lemoine. During the past year they have seen and managed a number of patients with chronic liver disease such as hepatocellular carcinoma, cirrhosis and chronic hepatitis. Consequently, most of the patients seen at the EFSTH (previously named RVTH) with chronic liver disease seen on Wednesday afternoons are referred to the liver clinic at MRC Fajara for specialist review. Their skills in how to properly assess a patient with liver disease, carry out a liver biopsy, Fibroscan and ultrasonography, have been greatly enhanced.
In Senegal, 4 hepatologists (Prof. Ka. Dr Toure, Dr Madoky and Dr Da Veiga)and one nurse have been trained to use the Fibroscan.
Two Fibroscan devices and two Ulrasoound machines have been bought and given to the local partner (MRC in Gambia and University de Thies in Senegal).
As part as the project we also built a liver Unit at the MRC unit in Gambia and in the main teaching hospital in Banjul (Gambia)
Within the PROLIFICA programme, we have been able to settle an endoscopy clinic in The Gambia (this was already present in Senegal) with the support of the World Gastroenterology Society (Dr Des Leddin). Dr Ramou Njie and Maud Lemoine participated to the teaching and training in January 2014 with the support of the MRC clinical services (Dr Suzanne Anderson and Prof. Umberto D’Alessandro).
3.7.2 Lab Training
Local lab technicians and senior virologists have been trained to run qPCR and set up a HBV in-house PCR under the supervision of INSERM, Lyon France (Prof. Zoulim’s team) and Dr Maimuna Mendy (IARC, France). Training has been done locally but also at INSERM (Lyon, France) to train in particular Dr Amina Sow from Senegal and Dr Harr Njai from Gambia. Penda Suso junior Lab tec has been trained in Lyon from Feb to march 2016 to run and interpret serologies and qPCR.
In addition, Imperial has facilitated the setting up of a Mass Spec Laboratory at MRC Gambia in conjunction with Waters Inc and also provided training to MRC staff in the use of metabonomics to identify liver cancer biomarkers. This involved comprehensive training by specialised researchers at Imperial, followed by training on site, once the mass spectroscopy laboratory had been established at the MRC. Dr Njie has been involved with the on-site establishment of the laboratory and MRC staff training.
There is one dedicated trained scientist, Dr Haddy Fye, who managed the facility and was responsible for all local metabonomic research. She will eventually assist in training local staff in the technology. The Imperial metabonomic staff has supervised Dr Haddy Fye’s work and help with any issues, while also providing updates to her training. Imperial is currently investigating funding possibilities to support the future cost of running the facility.
Histopathology: liver biopsy samples went through a typical histopathology laboratory protocol which Penda Suso has documented as an SOP. Histopathological analysis has been done by Prof. Khalil. Initially, only one block was being cut but this was later revised to two so that one was used locally for diagnosis and the other sent to IARC (for archiving). Locally, eight slides were sectioned (2 for diagnosis; 3 sent to UK and 3 were archived).Penda Suso also successfully attended a one month’s training (20th August – 20th September 2013) on histological techniques at Imperial College London, St. Mary’s Hospital, London.
3.7.3 Academic training
The PROLIFICA has been able to train 16 students and junior doctors (please see below) in clinical and basic science research.
Degree Year Name
Bsc in Medicine (n=6) 2013
2014 Yuki Takao
Lauren Berg
Harry Posner
William Stranger
Edmund Donnelly
Mario Lapore
Master in public health (n=1) 2013 Liliane Mpanzani
PhD (n=6)
Epidemiology 2011-2014 Yusuke Shimakawa
Public health 2013-2016 Shevanthi Nayagam
Immunology 2013-2014 Nick
Medicine 2010-2013 Nimzing Ladep
Medicine 2011-2015 Torben Kimhofer
Medicine 2015-2018 Gibril Ndow
Science 2014-2017 Nick Easom
Post Doc (n=3)
2011-2013 Harr Njai
2013- Haddy Fye
2014-2015 Jess Howell
Fellowship (n=1) 2011-2013 Maud Lemoine
Potential Impact:
PROLIFICA has demonstrated that screening and treatment for hepatitis B in sub-Saharan Africa is feasible, cost-effective and necessary in order to control morbidity and mortality related to this infection. Furthermore the work demonstrates that screen and treat will be an essential component of the public health interventions required to eliminate this infection.
4.1 Local impact in Gambia and Senegal
On a number of occasions during the PROLIFICA programme the investigators have held meetings with members of the health ministries in Gambia and Senegal. These meetings have helped to raise the levels of awareness about HBV infection and its role in the development of cirrhosis, liver failure and liver cancer. In Gambia the Minister of Health attended the WHO World Hepatitis Day conference in Glasgow in September 2015 and was a signatory to the Glasgow Declaration on Hepatitis (http://www.who.int/hepatitis/glasgow-declaration-on-viral-hepatitis/en/). He is committed to increasing the public health response to viral hepatitis in Gambia and is currently developing a national plan to counter HBV in the country. It is expected that this plan will encompass the screen and treat strategy that was pioneered in the PROLIFICA project.
The HBV viral load assay developed for use by PROLIFICA is now being used effectively in Gambia to monitor patients on treatment. The skills and laboratory facilities generated by the project will now be available to support the more widespread application of screening, assessment and treatment monitoring anticipated in the Ministry of Health’s plan. During the PROLIFICA project four locally trained doctors and about 20 fieldworkers were trained in the skills required for screening and assessment of patients with HBV infection. These medical staff will prove invaluable once the Government backed plan is implemented.
The Senegalese ministry of health are also now developing a national plan for viral hepatitis. This includes a commitment to purchase Tenofovir for use in HBV mono-infected patients. The clinicians and laboratories participating in the PROLIFICA project are advising the Ministry about strategy for a national programme as well as providing the necessary molecular virology support.
4.2 Natural history of HBV
Using the clinical and laboratory facilities established as part of the PROLIFICA project in Gambia we were able to define the natural history of chronic HBV infection in West Africa. Cohorts of patients recruited over 25 years ago in the villages of Keneba and Manduar had been followed at 4-5 year intervals using liver biochemistry and viral serology. Full clinical assessments were performed on 301 patients. The following key findings were reported:
• Annually, 7.4% (95% CI 6.3% to 8.8%) cleared HBeAg and 1.0% (0.8% to 1.2%) cleared HBsAg.
• The incidence of HCC was 55.5/100 000 carrier-years (95% CI 24.9 to 123.5).
• 5.5% (95% CI 3.4% to 9.0%) had significant liver fibrosis.
• HBV genotype A (versus E), chronic aflatoxin B1 exposure and an HBsAg-positive mother, a proxy for mother-to-infant transmission, were risk factors for liver fibrosis.
• A small proportion (16.0%) of chronic carriers were infected via mother-to-infant transmission; however, this population represented a large proportion (63.0%) of the cases requiring antiviral therapy.
Most importantly this study highlighted the importance of mother to child transmission in sub-Saharan Africa which had previously been ignored due to the low proportion of transmission via this route. Future public health interventions will therefore need to target this mode of infection.
4.3 Epidemiological Modelling
As part of the PROLIFICA project we developed a simulation model of the global HBV epidemic, incorporating data on the natural history of HBV, prevalence, mortality, vaccine coverage, treatment dynamics and demography. We used the model to estimate the impact of current interventions and the impact of scaling up of existing interventions for prevention of infection and introducing population screening and treatment as a novel intervention. The model was populated with data taken directly from the PROLFICA WATCH study.
The model demonstrates the enormous impact that infant and neonatal vaccination programmes have already made in some global regions. Specifically it is estimated that 210 million new chronic HBV infections were averted between 1990 and 2015. The impact of vaccination on HBV related deaths is more modest with approximately 1.3 million deaths averted by 2030. However the model shows clearly that there is a need to scale up existing interventions as well as to introduce novel interventions in order to control mortality. For example if infant vaccination was scaled up from the current average of 70% to a target of 90%, and if birth dose vaccination was scaled up to 80% of children born to mothers with high level viraemia and if population wide testing and treatment was able to access 80% of the eligible population then there would be a global reduction of 90% of new chronic HBV infections and 62% reduction in mortality. Furthermore, this scale up of public health interventions would lead to achievement of a realistic HBV elimination threshold by 2090 which could not otherwise be reached.
The Global Epidemiological model has been shared with the WHO’s Global Hepatitis Programme in order to respond to the World Health Assembly’s request to evaluate what is required and whether it is feasible to eliminate HBV (http://www.who.int/mediacentre/news/releases/2014/WHA-20140522/en/) .
4.4 Return on investment
PROLIFICA data and the Global Epidemiological Model were used to estimate the costs and the return on investment for a comprehensive package of possible public health interventions for Senegal. The package of interventions include infant vaccination, measures to prevent mother-to-child-transmission and screening and treatment for HBV. We assumed an overall societal perspective by considering both direct and indirect costs associated with HBV interventions. The former were mainly the costs associated with seeking and obtaining prevention or treatment interventions (medical costs) while the latter were the costs to society due to the premature death of people infected with HBV. Direct costs were represented by the medical fees charged to patients at either public or private health facilities. In reality, fees are just a representation of the real costs as they depend on several factors including whether the patient is covered by health insurance, whether the patient seeks treatment at a private or a public health facility and on the baseline availability of certain diagnostic tests or treatments
Between 2015 and 2030, investing 265 billion CFA (€400 million) into this comprehensive package of interventions will give an economic return of 631 billion CFA (€963 million). This would amount to a return on investment of 2.38 CFA per CFA invested & could avert 21 000 deaths by 2030.
4.5 Awareness of limitations of vaccination only strategy
Until the World Health Assembly resolution on viral hepatitis in 2010 the WHO had only one strategy to tackle viral hepatitis which was the global implementation of the infant vaccination integrated with the expanded programme of immunisation in all member states. The WHO had also recommended birth dose vaccination for children born to mothers with high levels of viraemia but this policy has been poorly observed. The data emerging from PROLFICA have highlighted the need for additional public health interventions to control new chronic infections and to reduce the mortality associated with HBV
4.6 WHO guidelines
Prof. Mark Thursz and Dr Maud Lemoine are members of the HBV & HCV Guidelines Development groups for the World Health Organisation Global Hepatitis Programme. Dr Nayagam is a member of the WHO guideline development group for hepatitis screening.
4.6 Laboratory enhancement in Gambia
PROLIFICA’s Metabonomics workpackage results have stimulated further research on biomarker development for the early diagnosis of HCC. The Waters Corporation kindly donated a QToF mass spectrometry system for use at the MRC Laboratories facility in Gambia. This allowed Imperial researchers (Prof Taylor-Robinson and Prof Elaine Holmes) to set up a metabonomics laboratory run by Dr Haddy Fye, a post doctoral scientist trained at Imperial. The lab is now analysing urinary metabolic profiles in suspected HCC patients.
4.8 Cohorts of HBV infected patients to study long term outcomes
The WATCH platform generated two cohorts of patients that will be used to answer key questions about the management of HBV in the future. 1. The treated patient cohort will be used to monitor levels of adherence and the ability of treatment to prevent advanced liver disease over a prolonged duration of treatment. 2. The observational cohort of patients who did not meet treatment eligibility criteria will be used to determine the rate of spontaneous viral clearance, the rate of transfer from ineligible to eligible for treatment and the liver related outcomes in this group. This data will inform policy on the need for a frequency of follow-up in this patient group.
List of Websites:
www.prolifica.eu