Understanding hepatitis B, hepatitis C and HIV among people who inject drugs in South Africa: findings from a three-city cross-sectional survey

Globally, viral hepatitis is responsible for the deaths of approximately 1.34 million people every year, similar to the annual number of deaths from HIV/AIDS (1.3 million), malaria (0.9 million) and tuberculosis (1.3 million) [1–3]. In sub-Saharan Africa, approximately 60 million people are chronically infected with the hepatitis B virus (HBV) and 10.2 million people are chronically infected with hepatitis C virus (HCV) [1]. The region remains an epicentre for HIV with 25.7 million people living with HIV [2]. Most HBV infections occur in childhood through horizontal transmission, while most HIV infections occur during adolescence and adulthood through sexual contact [4–6]. Most HCV infections occur parenterally, through exposure to blood or unsafe medical practice [1, 5, 6]. People who inject drugs (PWID) who cannot access harm reduction services, specifically needle and syringe programmes and opioid substitution therapy (OST), are at particularly high risk of HCV, HIV and if non-immune, for HBV infection [7, 8]. In 1995, universal HBV vaccination was introduced in South Africa’s Expanded Programme of Immunisation [9]. Many current PWID in South Africa were born before this vaccine was introduced and hence a pool of chronic HBV remains, with HCV and HIV providing the bulk of new blood borne infections among PWID [9].

Screening for these infections includes the use of antibody and/or antigen tests, either point-of-care or laboratory-based. HIV confirmatory diagnosis in the sub-Saharan African context (a generalised HIV epidemic) usually involves a second, highly specific, rapid test from a different manufacturer [10]. A positive HBV surface antigen (HBsAg) is the marker for current infection, with further HBV serological assessment used to guide treatment. For HCV, confirmation of active viraemia is usually based on demonstrating HCV ribonucleic acid (RNA) on polymerase chain reaction (PCR) testing. Furthermore, HCV genotyping has until recently been performed to assess which of the six most frequent HCV genotypes is present to inform treatment choice. However, genotyping is not necessarily routinely required if pan-genotypic direct acting antivirals (DAAs) are available [8, 11].

Chronic HCV infection is associated with significant morbidity and mortality due to associated risks for cirrhosis and hepatocellular carcinoma [1]. Hepatitis C is readily treatable, and in most cases curable, with DAAs but access to therapy in South Africa is currently limited as these medications have yet to be registered, and a national viral hepatitis programme has yet to be fully implemented.

We planned to recruit 960 PWID (320 in Cape Town, Durban and Pretoria, respectively). This sample size was based on available resources. People who inject drugs were defined as people having injected a substance for non-therapeutic purposes, irrespective of the type of drug injected or the mode of injection, in the previous 12 months. All participants had to be 18 years or older. Data collection took place between August 2016 and October 2017.

Study activities were integrated into existing HIV prevention and sexual health services provided by non-profit organisations targeting PWID in these cities. MSM were recruited in a targeted sub-study at different sites looking into HBV, HCV and HIV prevalence and risks in that population. These results are not included here.

An overview of study procedures is provided in Fig. 1. Participants who provided written informed consent were administered a confidential health screening questionnaire that was completed in English and clarified in the appropriate language as needed by the peers/nurses administering it. The questionnaire solicited demographic information (age, sex, race and housing), substance use (substances used in previous month and injecting in last month, including injecting frequency, needle and syringe reuse and sharing), use of OST and sexual risk behaviour.

Consenting participants had 20 ml whole blood drawn. HIV, HBV and HCV point-of-care (POC) testing was performed and samples from all HCV POC reactive individuals were sent to the National Institute for Communicable Diseases (NICD) in Johannesburg for confirmatory HCV enzyme-linked immunosorbent assay (ELISA), HCV PCR confirmatory viral load testing and genotyping. POC HIV testing was done in line with national protocols using a serial testing strategy [10]. POC HBsAg and HCV testing was performed using the Determine™ HBsAg (Alere Inc., MA, USA) and OraQuick® HCV rapid antibody test (Orasure Technologies Inc., PA, USA), respectively. All samples that were sent to the laboratory were subjected to anti-HCV testing (ARCHITECT Anti-HCV assay) on the ARCHITECT i1000SR system following the manufacturer’s instructions (Abbott Laboratories, Diagnostics Division, IL, USA). Samples that tested antibody positive were tested quantitatively using the COBAS® AmpliPrep/COBAS® TaqMan® HCV Quantitative Test v2.0 on the COBAS® Ampliprep Taqman® Analyser (Roche Molecular Systems, CA, USA). For HCV genotyping, 10 μl of biotinylated PCR amplicon was applied to nitrocellulose strips using the Versant HCV genotyping 2.0 assay (LiPA, Innogenetics, Ghent, Belgium). Genotyping results were interpreted using the reading card over the strip and the reading chart provided in the assay.

Participants received counselling around viral hepatitis and were provided with condoms, lubricant, information and education material and sterile injecting equipment. Participants diagnosed with HIV were referred to HIV care services. Those with a positive HBsAg screening test were referred to a pre-identified health facility for work-up and assessment for treatment. Participants with a negative HBsAg POC test were offered HBV vaccination. Those with a positive HCV screening test were given a follow-up date to return to obtain the results of additional laboratory testing. At the follow-up visit, the study nurse provided further counselling and linkage to care was arranged for additional work up and management of HCV infection. This was either at a designated gastrointestinal or liver clinic, where these existed, or at a pre-identified district hospital, where specialised clinics did not exist.

Data analysis was performed using STATA 11.2 (Statacorp LLC, Texas, USA). Data was initially analysed using descriptive statistics, stratified by city and biological sex. POC seroprevalence was calculated for HIV, HBV and HCV based on number of reactive POCs divided by the number of PWID participants that were tested as per protocol. HCV viraemic prevalence was calculated by dividing the number of participants with detectable viral load over the number of eligible PWID participants who had a HCV POC test. The HCV viraemic rate was calculated by dividing the number of participants with detectable viral load over the number of HCV POC positives. Bivariate analysis was done between HCV antibody status (reactive/ non-reactive) and selected demographic, drug using and sexual risk factors. Age and frequency of injecting were categorised into binary variables based on their median value. HCV antibody status was the outcome of interest in multivariate analysis. The model was adjusted for biological sex, age, race, city, living on the street, injecting practices (frequency, new needle at last injection, sharing needle at last injection), sexual practices (sexually active in last month, condom use at last penile-vaginal sex) and HIV status.

The study was approved by the Human Research Ethics Committee of the University of Cape Town (ref: 004/2016), the Research Ethics Committee of the University of the Witwatersrand (ref: M160510) and the Western Cape (ref: WC_2016RP19_818) and KwaZulu-Natal (ref: KZN_2016RP59_986) Provincial Department of Health Ethics Committees. Participants were not remunerated for their participation.

This is the largest published quantitative study of HBV, HCV and HIV among PWID in South Africa to date. The study demonstrates a significantly higher prevalence of HCV infection among PWID (43%) than previously known [23]. Notably, HCV prevalence among PWID in Pretoria was substantially higher than in other cities. This study did not explore the reasons for this geographical variation. However, expert consultation in other settings has suggested that injecting practices as a drug using culture have been in existence for longer in Pretoria than other cities, potentially suggesting a reason for this finding [24].

A large proportion of the participants lived on the street, accounted for by the recruitment of participants through existing harm reduction services that target people from lower socio-economic circumstances. People who live on the street often have limited financial resources to purchase injecting equipment, contributing to the increased likelihood of reuse, sharing and use of contaminated injecting equipment [25]. People living on South African streets who use drugs have little or no access to private spaces to do so, and frequently experience human rights violations, including assault, confiscation of possessions (including injecting equipment) and being moved [26–28]. Consequently, conditions for safe injecting are limited. The multivariate analysis found an increased likelihood of HCV infection among people living on the street, reinforcing the importance of addressing social and structural factors to enhance HIV and HCV health outcomes.

The study found that HCV risk is positively associated with increased injection frequency. This is supported by the per injection risk without 100% access to sterile injecting equipment [29]. The high level of reported use of new needles is surprising, and may reflect a social desirability bias, considering that the majority of people inject four or more times a day and access to sterile injecting equipment is limited. Existing harm reduction services in these cities distribute between 10 and 14 sterile needles and syringes at each encounter with a PWID client, with these projects aiming to visit locations where PWID are provided with services once or twice a week.¹ An assessment of needle and syringe service coverage in Pretoria and Cape Town was completed in 2017 and results are forthcoming [24]. The reported levels of needle and syringe sharing highlight the need for increased saturation of needle and syringe services.

Overall, few PWID (5% of the sample) had access to OST in these cities, which is to be expected as OST remains in the pilot stages [16]. It is also unclear if the reported OST access was sustained or over limited periods and if the dosage was aligned to global recommendations [8].

High coverage of OST and needle and syringe services can reduce the risk of HCV infection by 74% [30]. Modelling data from Kenya (PWID population of 50,000 with HCV prevalence of 11%) shows that providing 75% needle and syringe coverage and 50% OST would reduce the risk of HCV incidence among PWID by 69% by 2030, and elimination targets would be reached through the addition of treating chronic HCV infections among 1000 PWID over 5 years [31].

This research confirmed the findings of a previous study that documented elevated HIV prevalence among PWID compared to the general population [21]. In multivariate analysis, HIV infection is significantly associated with HCV serostatus, highlighting the shared transmission routes of these viruses and the need to integrate viral hepatitis and HIV services for PWID. The study also documents HBsAg prevalence similar to the general population [4]. While a relatively small proportion of people were found to be co-infected with HBV/HCV/HIV, the added morbidity and mortality requires that co-infection remains an important area for consideration. The relative exclusive prevalence of HCV genotypes 1a and 3a is similar to that in other countries where HCV infection is predominantly spread through sharing contaminated injecting equipment among PWID [32, 33]. However, of note in South Africa, genotype 5a (a predominant genotype in patients with liver disease [34]) was not found among PWID and suggests that 5a circulates in the general population and not in PWID.

Participants in this study were recruited from among those who currently access or who were reachable to organisations implementing harm reduction services, specifically mobile needle and syringe distribution and collection services in these cities. Unsurprisingly, the participants were overwhelmingly male given that globally, a higher proportion of PWID are male compared to female, with the proportion of female PWID ranging from 3% in South Asia to 33% in Australasia, and 12% in sub-Saharan Africa [7]. Programmatic data from the services that operate in these cities reflects that between 10 and 13% of service users are women [35]. While this is possibly due to higher numbers of men who inject drugs than women in South Africa, it also likely reflects barriers preventing women who inject drugs from accessing harm reduction services—including stigma, discrimination and services that do not address their specific needs [36]. Previous studies among South African and African PWID also included a smaller proportion of women who inject drugs [21, 37]; however, globally, a higher proportion of women who inject drugs are reached [38]. Nevertheless, there is global [39] and sub-Saharan [40] recognition of the need for gender-appropriate and specific HIV and HCV services for women who inject drugs due to their specific needs.

A diverse cross-section of South African racial groups was recruited across sites, indicating that substance use and injecting practices are issues affecting people across race or ethnicity. The relative order of the size of racial groups in the study sites reflected the demographic characteristics of the relevant cities [41]. However, the relative over-representation of white PWID remains (27% of the sample, versus 8% of the general population) [41]. It is not clear from this study if there are relatively more white PWID or if this is due to selection bias. Most of the study sites where centrally located in areas where PWID congregate. Non-white PWID living in poorer, peripheral areas of the city may have not been reached by existing harm reduction services and therefore not recruited in the study. There is no published evidence supporting different levels of stigma around injecting drug use among different racial groups in these cities, but this may also have been a factor. It is also unclear why lower HCV risk was found among black PWID. Data was not collected on length of participant injecting history, so no analysis can be made on associations between injecting history and race groups as a proxy for other factors.

The findings from this study around substance use reflect substance use treatment data that shows the prevalence of poly drug use in South Africa and different usage patterns across the country [35]. Study and treatment data point to heroin being the most commonly injected drug, and methamphetamine/ATS injecting being highest in Cape Town.

Proportionately more women than men reported sexual activity in the previous month and condom use was similarly sub-optimal. More women than men reported transactional sex for drugs, which has previously been identified among women who inject in South Africa [21].

Interestingly, men who reported recent sexual activity showed lower risk drug use practices than their counterparts who did not report being sexually active. This may be because higher frequency of opioid use is associated with reduced libido, and can induce impotence among men [42]. It may therefore be that men engaging in regular sexual activity are those using less opioids, and are therefore able to exercise greater care with each instance of injection. It may also be that sexual activity is indicative of the existence of intimate partnerships that impact on risky injecting behaviours. While intimate couples who inject drugs together often engage in needle sharing and higher risk injecting practices within the couple [43], there is also increasing evidence that drug use tends to lessen when people have close social relationships and that couples set negotiated standards and limits to such use [44]. This can include agreements about where, when, how, and with whom drugs are injected [45]. Furthermore, concerns for placing primary sexual partners at risk of infectious diseases may contribute to safer injecting [13]. These limits on injecting behaviour in intimate relationships may be some of the reasons that recent sexual activity was found to be generally protective against HCV infection.

HCV, and to a lesser extent HIV, are major health threats affecting PWID in South Africa. On-going high risk practices are influenced by limited access to OST and needle and syringe services, as well as the social and structural factors affecting PWID who live on the street in the context of criminalisation of people who use drugs. Furthermore, in the absence of HCV treatment, an increase in HCV infections is inevitable without a rapid, targeted response that is appropriate and acceptable to PWID. Programmatic responses should also focus on prioritising treatment within Pretoria where HCV and HIV prevalence is highest, with intensified prevention messaging in that city. However, hepatitis awareness and prevention activities need to be up-scaled in other cities where HCV prevalence has not yet reached such high proportions.

Intensified efforts are needed to reach and provide services to women who inject drugs and to ensure that these services are inclusive of their sexual and reproductive health needs. Tailored outreach and viral hepatitis testing and treatment services, facilitated by gender appropriate teams should be considered as services scale up.

Broader social issues related to living conditions and social isolation need to be considered in holistic responses to HCV and HIV infection among PWID in these cities. Interventions that address structural issues that may be contributing to the elevated HCV risk among PWID living on the street are needed, which are not limited to exploration of safe injecting facilities, and enhanced support from law enforcement for needle and syringe services.

This study was limited to a description of HCV, HBV and HIV among PWID in these cities that have (potential) access to harm reduction services. Additional research is needed on the epidemiology of these infections among PWID in other South African contexts (urban and rural), as well as within the correctional services system. Greater insights are also needed into factors associated with on-going parenteral risk in the presence of harm reduction services. Models that explore innovative ways to provide integrated HCV-HBV-HIV and other holistic services to PWID in the South African context are needed to inform the granular detail and implementation experience that will be necessary for an effective HCV response. Future research is needed to understand health care providers’ knowledge around viral hepatitis and HIV among PWID.

Eliminating viral hepatitis as a public health threat in South Africa by 2030 will only be an attainable reality once hepatitis prevention, testing and treatment services are provided to all PWID as needed. Services will need to include access to sterile injecting equipment, OST and HIV and HCV testing and treatment. Like many other counties, South Africa has the guidelines and a national action plan to address this issue, but political support, appropriate resource allocation and dedicated and passionate people that include the PWID community will be required to support implementation and reach this ambitious goal.