The impact of direct‐acting antivirals on hepatitis C viraemia among people who inject drugs in England; real‐world data 2011–2018

Abstract Direct‐acting antiviral (DAA) therapy for anybody with viraemic HCV infection has been scaled‐up in England since 2017. To assess early impacts, we investigated trends in, and factors associated with, HCV viraemia among people who inject drugs (PWID). We also examined trends in self‐reported treatment access. Bio‐behavioural data from an annual, national surveillance survey of PWID (2011–2018) estimated trends in viraemic prevalence among HCV antibody‐positive PWID. Multivariable logistic regression identified characteristics independently associated with viraemia. Trends in treatment access were examined for PWID with known infection. Between 2011 and 2016, viraemic prevalence among antibody‐positive PWID remained stable (2011, 57.7%; 2016, 55.8%) but decreased in 2017 (49.4%) and 2018 (50.4%) (both p < 0.001). After adjustment for demographic and behavioural characteristics, there remained significant reduction in viraemia in 2017 (adjusted odds ratio [aOR] 0.79, 95% CI 0.65–0.94) and 2018 (aOR 0.79, 95% CI 0.66–0.93) compared to 2016. Other factors associated with viraemia were male gender (aOR 1.68, 95% CI 1.53–1.86), geographical region, injecting in past year (aOR 1.26, 95% CI 1.13–1.41), imprisonment (aOR 1.14, 95% CI 1.04–1.31) and homelessness (aOR 1.17, 95% CI 1.04–1.31). Among non‐viraemic PWID with known infection, the proportion reporting ever receiving treatment increased in 2017 (28.7%, p < 0.001) and 2018 (38.9%, p < 0.001) compared to 2016 (14.5%). In conclusion, there has been a small reduction in HCV viraemia among antibody‐positive PWID in England since 2016, alongside DAA scale‐up, and some indication that treatment access has improved in the same period. Population‐level monitoring and focus on harm reduction is critical for achieving and evaluating elimination.


| INTRODUC TI ON
The global prevalence of hepatitis C virus (HCV) infection is estimated at 1% 1 and disproportionately affects people who inject drugs (PWID). [2][3][4] In England, injecting drug use is cited as the risk in over 90% of laboratory reports where a risk factor has been disclosed. 5 Severe outcomes such as liver cirrhosis and hepatocellular carcinoma contribute to the high economic and health burden of this disease. 6 The World Health Organisation's (WHO) target to eliminate HCV as a public health threat by 2030 became a possibility in 2011, with the development of highly effective and tolerable direct-acting antiviral (DAA) therapies. [6][7][8][9][10] Mathematical modelling has shown that alongside strengthening of harm reduction services, such as needle and syringe programmes (NSP) and opioid substitution therapy (OST), major scale-up of DAA therapies for PWID is required to meet elimination targets. [11][12][13][14] In 2015, the English National Health Service (NHS) announced their then single largest investment in new treatment, with a budget of £190 million to fund DAA therapies prioritized for patients with severe liver disease. 15 The restriction on disease stage was lifted in 2017 and since then treatment is recommended for all those with viraemic HCV infection. 5,16,17 To drive improvements in access and uptake in local areas, regional operational delivery networks (ODNs) are being utilized to manage HCV treatment decisions and prescribing via a dispersed treatment model. 18,19 ODNs provide a uniform standard of quality treatment, but work in partnership with healthcare providers and local organizations including primary care services, local authorities and services for PWID to meet local priorities and need. [17][18][19] A national treatment database has also been developed by NHS England to track HCV treatment uptake and outcomes. 16,17 Testing for HCV is available in primary care, sexual health clinics and community drug and alcohol services, and more recently has been expanded to community pharmacies. 20 Solidifying its commitment to these above efforts, in 2018, NHS England set out plans to be the first country in the world to eliminate HCV. 16 As countries push towards elimination goals, 'real-world' data on the impact of scaling up HCV DAA therapies among PWID are required to track progress or pitfalls. However, few countries are able to measure population-level changes in HCV viraemia among PWID due to lack of robust data sources. 7,21 England is one of the few countries in a position to do so, having a long-standing national surveillance system utilizing an annual cross-sectional bio-behavioural survey that monitors blood-borne viruses and associated risk behaviours among PWID. 22 We aim to assess the early impact of DAA scale-up in England by examining changes in HCV viraemia among antibody-positive PWID between 2011 and 2018. We also examine factors associated with HCV viraemia and assess self-reported HCV treatment access among those eligible.

| Amplification and detection
The qualitative PCR targets the non-coding region of the HCV genome using the ABI 7500 real-time thermal cycling with Qiagen TaqMan-PCR reagents. The multiplex real-time PCR detects both HCV and MS2 with differently labelled TaqMan probes.
Amplification was performed using 20 μl of extract in a 50 μl vol-  Table S1. The reaction mixture was amplified using the following cycling conditions: 50°C for 30 min for the RT step

| Multiple imputation
Multiple imputation by chained equations (MICE) was performed to assign either an RNA-positive or an RNA-negative status to anti-HCV-positive samples that were insufficient for RNA testing.
Missing data were assumed to be missing at random, such that unbiased imputed values could be obtained conditional on observed covariates. Survey year, age, gender, region, history of homeless and imprisonment, country of birth, hepatitis B (HBV) vaccination history and HBV status were used as predictors in the imputation model, and ten imputed data sets were generated. A sensitivity analysis compared the results from MICE to results from non-imputed data. An additional sensitivity analysis was conducted to include observations for which antibody status was missing (excluded in main analyses) on the basis of the same imputation model. Factors associated with viraemic infection among those anti-HCV positive were explored using multivariable logistic regression.

| Statistical analyses
All variables that had a significant univariable association (Wald p < 0.05) with the outcome were added into the multivariable model. All analyses were conducted on the multiple imputed data sets using Stata 13 (College Station, TX: StataCorp LP). Summary statistics and proportions were calculated by taking the average across imputed data sets. Statistical tests of differences in proportions and trends over time were conducted using logistic regression, with results combined across imputed data sets to appropriately account for within-and between-imputation variability according to Rubin's rules.

| Treatment access
Participants who reported ever testing positive for HCV were asked if they had ever received treatment for their HCV infection; this response was combined with infection status from DBS testing.
A cleared infection result (anti-HCV positive, HCV RNA negative) among people who self-reported ever receiving treatment was assumed to be reflective of successful access to treatment.
Descriptive trends in the proportion of those with non-viraemic (cleared) infection who reported ever receiving treatment were presented for 2011-2018. This was not included in analytical modelling due to treatment being on the causal pathway to infection status.

| Characteristics of the study population
There were a total of 20,637 responses from PWID in England who   Figure 1A). The exclusion of samples not tested for anti-HCV had no impact on these results (Table S2a) or on subsequent multivariable results (Table S2b).

| Factors associated with HCV viraemia
In the adjusted model, HCV viraemia among those anti-HCV positive was associated with survey year (see ' Of the anti-HCV-positive samples available for 2011-2018, n = 1,167 (6.1%) were insufficient for HCV RNA testing and had an RNA result imputed. A sensitivity analysis comparing these findings to a multivariable model without imputation showed no differences in significant associations (Table S3).

Demographic
Year 2011       was also observed in a data linkage study of HCV diagnosed individuals. 31 There has also been a 20% fall in HCV-related deaths in England between 2015 and 2018, exceeding the WHO target 3 years early and twofold. 6,32 However, these are only early signs of improvement, and there is evidence to suggest the number of new infections may have risen in 2018. 5 Our study shows that the number of PWIDs ever exposed to HCV infection continues to rise, indicative that it may be effective treatment, more than successful prevention that is currently controlling HCV viraemia in the PWID population.

| DISCUSS ION
Unless the coverage and intensity of primary prevention interventions increases, there will be re-infections after successful DAA treatment. 33 39,41,42 However, the current indication is that improvements to implementation of these interventions are needed in order for prison and shelter settings in England to be able to adequately support HCV elimination. 39,40,42,43 Changes to funding structures, peer-support systems and policies are required in order to reduce the burden of HCV infection among PWID in these settings. these data are only collected for those who had injected during the month prior to participation, which limits interpretation and testing for differential intervention effects by intensity of injecting risk.
Thirdly, the data on HCV treatment uptake are self-reported and not validated, and represent ever receiving treatment, which should be interpreted with caution, as it reflects perception of care at any point during the individual's HCV infection. We cannot discriminate whether DAA or other therapies were given, although the UAM survey has been modified in 2020 to capture more specific data on HCV treatment. We also cannot comment on reinfection after successful HCV treatment as this information is not currently directly captured through the UAM Survey, although there are plans to collect this in future. Fourthly, the eligibility criteria of the UAM Survey prevent people participating multiple times in 1 year, but due to its anonymous nature those people who participate in more than 1 year cannot be linked over time. Finally, it is possible that antibody and viraemic prevalence in this study is not representative of the general population of PWID due to the sampling method; respondents are only those already in contact with specialist drug or alcohol services. 45 We also excluded a small number of HIV-positive people from analyses, as HIV and HCV-coinfected people may remain HCV-seronegative. 29 Whilst numbers excluded each year were consistently small, it is possible that HCV prevalence and self-reported treatment may be different among this excluded group compared to people who are HIV-negative.
In conclusion, our study adds to the scarce availability of 'realworld' evidence demonstrating that scaling up DAA treatments can lead to reduced HCV viraemic prevalence among PWID. We show treatment uptake has improved in England in recent years but is still suboptimal. There are many challenges that lie ahead for HCV elimination, and whilst revolutionary, DAA treatment is only one element of a comprehensive elimination package. Radical changes to support collaborative work on prevention, testing and linkage to care among PWID is necessary, which will only become more important in light of the impact of the COVID-19 pandemic and its associated restrictions that have negative impacts on service provision and access to harm reduction, testing and treatment. 46,47 Future rounds of the UAM Survey will be critical to be able to evaluate our continued progress towards HCV elimination targets.

CO N FLI C T O F I NTE R E S T
None to declare.

AUTH O R CO NTR I B UTI O N S
The UAM Survey was led by EH, SC, KS and EP and was imple-

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.