1 INTRODUCTION

Over 10% of people living with HIV (PLWH) in West Africa also have chronic hepatitis B virus (HBV) infection, the leading cause of liver cirrhosis and cancer in the region.1 HIV infection has a profound impact on the natural history of HBV, including the acceleration of the progression to end-stage liver disease and the increase in liver-related mortality.2, 3 Tenofovir-containing antiretroviral therapy (ART) leads to HBV viral suppression in more than 80% of HIV/HBV-co-infected individuals and reduces the progression of liver fibrosis and the risk of hepatocellular carcinoma (HCC).4-6 Considering its clinical benefits and high genetic barrier to resistance, tenofovir, either as tenofovir disoproxil fumarate (TDF) or tenofovir alafenamide (TAF), should always be included in the ART regimen of HIV/HBV-co-infected individuals, provided there is no contraindication.

Despite international recommendations, the uptake of systematic HBV testing among PLWH has been generally poor in sub-Saharan Africa.7 This is problematic for several reasons: As tenofovir should be included in the ART regimen in the presence of HBV infection, knowledge of the individual hepatitis B surface antigen (HBsAg) status is key in informing the prescription of optimal ART for each patient, including second-line and third-line treatment regimens.8 HBV testing among PLWH is also required to identify those who will need a comprehensive liver disease assessment, including the measurement of liver fibrosis and HCC surveillance. In addition, the systematic testing for HBsAg allows health care providers to offer proper counselling on measures to reduce HBV transmission and screening of household members.9

We aimed to investigate predictors of HBV testing among PLWH in care at a referral HIV outpatient clinic in Senegal and to evaluate their HBV infection status by implementing a simple and systematic HBV testing intervention. We hypothesized that detecting and characterizing HBV infection would allow the improvement of HBV management in this patient population.

2 MATERIALS AND METHODS 2.1 Study setting

We conducted a cross sectional study at the « Service des Maladies Infectieuses et Tropicales/Centre Régional de Recherche et de Formation à la Prise en Charge clinique de Fann» (SMIT/CRCF), an urban adult HIV treatment centre in Dakar, Sénégal. This clinic has been one of the several referral centres for HIV management in Senegal since 1998. Since the publication of the first ART recommendations in 2002, Senegal has been adapting its recommendations in accordance with WHO protocols. Until 2013, recommended initial ART combined two nucleos(t)ide reverse transcriptase inhibitors (NRTI), generally lamivudine (LAM) and zidovudine (AZT), and one non-nucleoside reverse transcriptase inhibitor (NNRTI). In 2013 the Senegalese ART guidelines included the use of TDF alongside LAM or emtricitabine (FTC) as part of the preferred first-line backbones, in line with WHO recommendations.8, 10 However, switching patients with virological suppression to a TDF-based regimen was not a priority. As part of routine clinical care, all PLWH initiating ART undergo clinical and laboratory visits at day 1, 7, 14, 30, followed by quarterly visits. CD4 cell counts are measured every 6 months and HIV-1 viral load (VL) annually. This study was part of a larger research programme, SEN-B, which aims at evaluating the determinants of HBV ‘functional cure’ and liver-related outcomes among HIV/HBV-co-infected and HBV mono-infected individuals in Senegal.

2.2 Study participants and data

All PLWH in routine clinical care at the SMIT/CRCF as of January 2019 were considered for participation in the study. We used the centralized database from the Programme National de la Lutte contre le SIDA and medical records to identify the past availability of HBsAg test results. Individuals without a recorded HBsAg test result were identified and offered voluntary testing for HBV at the next scheduled follow-up visit during the study period (March-July 2019). The following data were retrieved from the clinical chart of all individuals: (a) demographic and clinical characteristics: age, sex, marital status, employment category, date of first positive HIV test, date of ART initiation and date and result of HBV, (b) stage of HIV disease: WHO stage, CD4+ cell counts and HIV-1 VL, and prior AIDS defining illnesses; (c) treatment history: prior and current ART and reason for treatment changes (eg treatment failure, adverse events).

2.3 Study procedures and laboratory measurements

HBsAg testing was performed using a one-step lateral flow assay rapid test (NOVATest®). NOVATest® (Atlas Link Biotech–ISO 13485) One-Step Hepatitis B Virus Surface Antibody (HBsAg) Test Kit is a CE-marked (CE 2265), rapid immunochromatographic in vitro assay for the detection of HBsAg in human serum samples. In a validation study in Senegal using the Elecsys® HBsAg II test (Cobas e-411, Roche Diagnostics) as gold standard, the sensitivity of NOVATest® was 89.5% and its specificity 100%.11 All individuals with a positive HBsAg test result between March and July 2019 were offered additional laboratory and transient elastography measurements, and samples were stored for viral sequencing. We performed HIV-1 RNA (lower limit of detection 20 cp/ml) and HBV DNA (lower limit of detection 20 IU/ml) viral load quantification using a commercial assay (COBAS®Ampliprep Taqman 96, v2.0, Roche Diagnostics GmbH). HBV sequencing was performed in samples of patients with HBV DNA >20 IU/ml using previously published methodologies.12 In brief, polymerase was amplified using specific primer sets and PCR products were purified with the QIAquick PCR purification kit (Qiagen). The Sanger-sequencing reaction was outsourced to a commercial service provider (Microsynth AG). Additional laboratory measurements included aspartate aminotransferase (AST), alanine aminotransferase (ALT), creatinine and CD4+ cell count. We defined ALT as being elevated if ALT >35 IU for men and >25 IU for women, according to AASLD recommendations.13 If the HBsAg test was positive, the ART regimen was reviewed and replaced by a TDF-containing combination if not already prescribed.

We measured liver fibrosis in all HIV/HBV-co-infected individuals using transient elastography (Fibroscan®, Echosens). Liver stiffness measurements (LSM) were performed by a single investigator who received manufacturer-recommended training. Results were expressed as the median (kPa) and the interquartile range (IQR) of all valid measurements. LSM was considered reliable when it included >10 valid measurements with a success rate >60% and IQR/M ≤ 0.30.14 Significant liver fibrosis (Metavir stages F2–F4) was defined as LSM > 7.1 kPa and cirrhosis (Metavir stage F4) as LSM > 11 kPa, as per WHO thresholds.15

2.4 Data analysis

Categorical variables were expressed as percentages with 95% confidence intervals (CI), and continuous variables as median values with IQR. Demographic and clinical characteristics were compared between individuals who had been previously tested and others using the Mann-Whitney test or Kruskal–Wallis test for continuous variables and Pearson chi–square (χ2) test or Fisher exact test for categorical variables. Logistic regression was used to evaluate potential associations between patient characteristics and (a) HBV testing uptake among all PLWH and (b) HBV infection status among individual who underwent HBsAg testing. Multivariable models were adjusted for age, sex and any other variable with a p-value < 0.05 in univariable analyses.

2.5 Ethics

The study was approved by the Senegalese National Health Research Ethics Committee (CNERS) at the Health and Social Action Ministry of Senegal (0061/MSAS/DPRS/CNERS, 576/MSAS/DPRS/CNERS and PSS/BTMB/09611). All SEN-B participants signed an informed consent.

3 RESULTS 3.1 Chronic hepatitis B virus testing during routine care

Of 1076 PLWH in care for a median of 5.6 years (IQR 2.3–10.6) at the time of the HBV testing intervention, 689 (64.0%) did not have a documented HBsAg test in the past (Figure 1). Among 387 individuals previously tested, 49 (12.4%) were co-infected with HBV. Demographic and clinical characteristics of PLWH by HBsAg testing status are shown in Table 1 and Table 2. Overall, 158/387 (41.0%) men and 229/689 (33.2%) women had been previously tested. In multivariable analysis, female patients (adjusted odds ratio (aOR: 0.70, 95% CI 0.50–0.98) and those aged >40 years (aOR: 0.55, 95% CI 0.39–0.78) were less likely to have been previously tested for HBV infection, whereas those who initiated ART after 2014 had increased odds of having a known HBV status (aOR: 2.96, 95% CI 2.12–4.15; Table 2).

Study flow diagram. Abbreviations: HBsAg, hepatitis B virus antigen; HBV, hepatitis B virus; DNA, deoxyribonucleic acid; HIV, human immunodeficiency virus; RNA, ribonucleic acid

TABLE 1. Demographic and clinical characteristics of PLWH, by HBsAg testing status Characteristics HBsAg screening status P

Never tested

n = 689

Previously tested

n = 387

Sex (n[%]) Male 227 [33.0] 158 [40.8] 0.02 Female 460 [68.0] 229 [59.2] Unknown 2 [0.3] 0 [0] Age, years Median (IQR) 47 [39–55] 45 [34–53] <0.001 Employment status (n[%]) Employed 468 [84.8] 299 [77.3] <0.001 Student 33 [6.0] 46 [11.9] Unemployed 50 [7.3] 15 [3.9] Unknown 138 [20.0] 27 [7.0] WHO stage at presentation (n[%]) I–II 228 [33.1] 152 [39.3] <0.001 III–IV 345 [50.1] 215 [55.6] Unknown 116 [16.8] 20 [5.1] Period of ART initiation, years (n[%]) 1998–2013 411 [59.7] 132 [34.1] <0.001 2014–2019 229 [33.2] 254 [65.6] Unknown 49 [7.1] 1 [0.3] CD4 at ART initiation, cel/ul (n[%]) <200 287 [41.7] 187 [48.3] <0.001 200+ 165 [24.0] 129 [33.3] Unknown 237 [34.4] 71 [18.5] Time on ART, years Median (IQR) 9,3 [4.8–12.8] 4,8 [2.5–8.5] <0.001 ART line regimen (n[%]) 1st line 507 [73.6] 334 [86.3] <0.001 2nd/3rd line 117 [17.0] 47 [12.1] Unknown 65 [9.4] 6 [1.6] Abbreviations: ART, Antiretroviral treatment; HBsAg, Hepatitis B virus surface antigen; IQR, interquartile range; WHO, world health organization. TABLE 2. Factors associated with HBV testing during routine HIV care n/N Univariable analysis Multivariable analysis OR (95% CI) p OR (95% CI) p Sex Male 158/385 1.00 0.01 1.00 0.02 Female 229/689 0.72 (0.55–0.93) 0.71 (0.54–0.94) Age group, years 18–40 149/339 1.00 <0.001 1.00 0.01 >40 237/730 0.61 (0.47–0.80) 0.69 (0.51–0.92) WHO stage at presentation (n[%]) I–II 152/380 1.00 0.62 III–IV 215/560 0.79 (0.57–1.09) CD4 at ART initiation, cells/mm3 (n[%]) <200 187/474 1.00 0.23 200+ 129/294 1.19 (0.89–1.61) Period of ART initiation, years (n[%]) 1998–2013 170/632 1.00 <0.001 1.00 <0.001 2014–2019 213/387 3.33 (2.55–4.34) 3.02 (2.27–4.00) ART line regimen 1st line 334/841 1.00 0.01 1.00 0.26 2nd/3rd line 47/164 0.61 (0.42–0.88) 0.80 (0.54–1.18) Abbreviations: ART, antiretroviral treatment; CI, confidence interval; OR, odds ratio; WHO, world health organization. 3.2 Prevalence and risk factors of chronic hepatitis B virus infection

Of 689 previously untested patients, 498 (72.3%) underwent an HBsAg test during the HBV testing intervention. Reasons for not testing were the lack of availability of testing kits or reagents in the laboratory at time of the visit, unwillingness to participate in the study or not attending the follow-up visit, among others. The main demographic and clinical characteristics of individuals who were not tested during our intervention were similar to those of patients who were tested: they were predominantly women (68% vs. 66% among those tested, p = 0.63), aged above 40 years old (69% vs. 72%, p = 0.67) and presenting with less than 200 CD4 cells/µl (45% vs. 40%; p = 0.01). During the HBV testing intervention, 58/498 (11.7%) PLWH had a positive HBsAg test. Thus, the overall prevalence of HBV infection among PLWH in care at our clinic was 12.1% (107 of 884, 95% CI 9.9–14.3). The prevalence of HIV/HBV-co-infection was marginally higher among male participants compared to woman (13.6% vs. 11.2%, p = 0.02), whereas estimates were similar across age categories (Table 3). In multivariable analyses no significant associations were found between potential explanatory variables and HBV infection (Table 4).

TABLE 3. Proportion of people living with HIV co–infected with HBV across sub-groups Characteristics HIV/HBV Prevalence n/N (%; 95% CI) All 107/884 Sex Male 44/323 Female 63/559 Age group, years 18–40 33/281 >40 74/598 WHO stage at presentation I–II 43/304 III–IV 49/463 CD4 at ART initiation, cel/ul <200 48/386 200+ 30/238 Period of ART initiation, years 1998–2013 53/481 2014–2019 45/347 ART line regimen 1st line 76/688 2nd/3rd line 18/130 Abbreviations: ART, antiretroviral treatment; HBV, hepatitis B virus; HIV, human immunodeficiency virus; WHO, World Health Organisation. TABLE 4. Correlates for HIV/HBV co-infection n/N Univariable analysis Multivariable analysis OR (95% CI) p OR (95% CI) p Sex Male 44/323 1.00 0.3 _ Female 63/559 0.81 (0.53–1.22) _ Age group, years 18–40 33/281 1.00 0.79 _ >40 74/598 1.06 (0.61–1.81) _ WHO stage at presentation I–II 43/304 1.00 0.14 _ III–IV 49/463 0.60 (0.36–1.01) _ CD4 at ART initiation, cel/ul <200 48/386 1.00 0.95 _ 200+ 18/153 0.98 (0.52–1.83) _ Period of ART initiation, years 1998–2013 53/481 1.00 0.39 _ 2014–2019 45/347 1.20 (0.79–1.84) _ ART line regimen initiation 1st line 76/688 1.00 0.36 _ 2nd/3rd line 18/130 1.29 (0.74–2.24) _ Abbreviations: ART, antiretroviral treatment; CI, confidence interval; OR, odds ratio; WHO, World Health Organisation. 3.3 Biological and virological characterization of HIV/chronic hepatitis B virus -co-infected individuals

Among 58 individuals who had a newly positive HBsAg test during the HBV testing intervention, 55.2% were women and the median age was 46 years (IQR 39–54). These participants were on ART for a median cumulative time of 7.7 years (IQR 3.0–11.0) and 30% presented to HIV care after 2014, when TDF became available in Senegal. At the time of HBV testing intervention, 54 (93.1%) patients were on a TDF-containing regimen. Overall, 28/55 (51%) co-infected individuals (3 individuals had missing creatinine values) had an eGFR (CKD-EPI formula) below 90 ml/min/1.73m2, including three with values below 60 ml/min. One individual had significant liver fibrosis and one liver cirrhosis, whereas only one participant had an elevated ALT value (Figure 2). Among the 58 HIV/HBV-co-infected individuals, 7 (12.1%) had a detectable HBV VL, of whom 5 were HIV-suppressed (Figure 1, Figure 2). HBV viral suppression was achieved in 40/47 (85.1%) individuals on TDF-containing ART and in 1/4 (25.0%) without TDF. HBV reverse trancriptase sequencing was successful in 4 of 5 HIV/HBV-co-infected individuals with HBV DNA >20 IU/ml, and showed the presence of the triple resistance mutation 180 M/204I/80V in two patients who were treated with lamivudine as the only HBV-active drug (Table 5). The patients without resistance mutations were both on a TDF-containing regimen and had never been exposed to lamivudine monotherapy for HBV infection.

Liver, renal and virological assessment of HIV/HBV-co-infected individuals screened during the study. Abbreviations: ALT, alanine aminotransferase; eGFR, estimated glomerular filtrate rate; CKD-EPI, chronic kidney disease epidemiology collaboration equation; HBV, hepatitis B virus; DNA, deoxyribonucleic acid; HIV, human immunodeficiency virus; RNA, ribonucleic acid; IU, international units

TABLE 5. HBV genotype and drug-resistance patterns among HIV/HBV-co-infected individuals with HBV DNA viral load > 20 IU/ml during ART Patient(ID) Sex Age (years) CD4 counts (cells/mm3) Current ART Time on ART (years) Transient elastography (kPa) ALT (IU) HBV viral load (IU/ml) HIV viral load (cop/ml) HBV genotype Resistance mutations ID 1 M 42 _ TDF 3TC EFV 5.8 4.6 15 33 < 20 a a ID 2 M 52 213 TDF 3TC EFV 3.0 4.8 14 46 < 20 a a ID 3 M 54 399 AZT 3TC LPV/r 18.5 5.5 10 97 554 A1 L180 M–M204I–L80V ID 4 F 36 336 TDF 3TC NVP 7.3 5.3 9 369 33.800 A1 none ID 5 F 45 785 AZT 3TC NVP 12.1 3.1 16 386 < 20 E L180 M–M204I–L80V ID 6 M 54 444 TDF 3TC EFV 10.3 3.1 12 63.600 < 20 a a ID 7 M 60 321 TDF 3TC LPV/r 2.8 5.8 8 > 10.000.000 < 20 E none