1 INTRODUCTION

Global elimination targets set by the World Health Organisation (WHO) aim for the elimination of viral hepatitis as a public health concern by 2030. To achieve elimination, targets call for the treatment of 80% of those eligible and a 90% reduction in incidence of new hepatitis B and C infections by 2030 compared with 2015 levels.¹ Only 12 of 194 countries are reported to be on track to meet the targets,² perhaps even fewer at the time of writing with a significant reduction in hepatitis C virus (HCV) testing and treatment rates during the COVID-19 pandemic.³ Worldwide, HCV was responsible for over a quarter of the 1.1 million deaths caused by viral hepatitis in 2019, due largely to chronic liver disease and liver cancer.¹ Incidence of HCV infection is highest among key populations, with 39% of the 1-year global population attributable fraction of HCV transmission in 2018–19 associated with intravenous drug use.⁴ High incidence of HCV has also been observed in studies of men who have sex with men (MSM) living with HIV and those using HIV pre-exposure prophylaxis (PrEP).⁵ Microelimination programmes targeting key populations suggest that it may be possible to reduce HCV incidence by improving linkage post-diagnosis to care and treatment, to reduce the risk of reinfection or diagnose and treat it at the earliest possible time.^6-11

Treatment-as-prevention (TasP), where risk of onward HCV transmission is lowered through high treatment coverage and reduced prevalence, is a key pillar of global elimination efforts and associated reductions in chronic liver disease-related morbidity and mortality.¹² Modelling has suggested that to achieve the WHO HCV incidence reduction targets, more frequent testing is needed in high-prevalence settings.¹³ While the advent of highly effective and tolerable treatments, known as direct-acting antiviral (DAA) medication, has led to approximately 9.4 million people with HCV infection being treated between 2015 and 2019 worldwide,¹ individuals may remain at risk of reinfection following cure. Reinfection is defined as recurrent viraemia after its clearance either spontaneously or as a result of treatment.¹⁴ People who inject drugs (PWID), MSM and people in custodial settings are among those at highest risk of recurrent viremia.^15-17 Guidelines recommend ‘focused’ testing in these populations, along with ongoing linkage to prevention and care services, and suggest targeted testing among these populations is likely to be cost-effective.¹⁵ Following SVR, the European Association for the Study of the Liver (EASL) recommend at least annual, preferably biannual monitoring for HCV reinfection among PWID and MSM¹⁸ and the American Association for the Study of Liver Diseases (AASLD) recommend annual RNA testing among patients with ongoing risk including intravenous drug use or MSM engaging in unprotected sex.¹⁷ This systematic review was commissioned by the WHO to inform their ‘Consolidated guidelines on HIV, viral hepatitis and STI prevention, diagnosis, treatment and care for key populations’, where key populations included MSM, PWID, people in prisons and closed settings, sex workers and trans and gender diverse people. We aimed to provide specific additional evidence on the association between different reinfection testing intervals and the detection of HCV in the post-SVR follow-up period among highest risk populations.

2 MATERIALS AND METHODS

This systematic review was commissioned and guided by the WHO Global Hepatitis Programme. The review protocol was prospectively registered with PROSPERO (CRD42021249863).

3 RESULTS

3.1 Search results

A total of 14 408 citations were identified from the search strategy, of which 8140 were duplicates. Of the 6268 unique citations screened for eligibility, 220 were eligible for full-text review (Figure 1). A further 11 studies were identified for full-text review by searching conference abstracts and reference lists of included studies. Of the 231 full texts screened, 190 were excluded (study exclusion reasons outlined in Figure 1). The most common reasons for exclusion related to study design, including studies which did not test individuals at discrete testing intervals (n = 91) and studies that sampled populations other than the populations in our inclusion criteria (n = 14).

3.3 Pooled incidence estimates of HCV reinfection

The 38 studies included in the pooled HCV reinfection incidence estimate comprised 8931 participants at risk of reinfection (Table S1). Berenguer et al,¹⁹ and Chen et al²⁰ included two study population arms (PWID and MSM) and HCV reinfection incidence rates were extracted for each arm and assigned to each key population meta-analysis separately. The pooled incidence estimate from all included studies was 4.13 per 100 py (95% confidence interval [CI]: 3.45–4.81) with HCV reinfection incidence ranging from 0.00 per 100 py to 31.00 per 100 py across studies. Heterogeneity was high among all included studies (I² = 93.6%) (Figure 2).

3.3.1 By key population groups

Among 27 studies comprising 4899 participants, where the primary study population were PWID, the pooled HCV reinfection incidence estimate was 2.84 per 100 py (95% CI: 2.19–3.50). Among the nine studies comprising 3269 participants whose primary population were MSM, the pooled incidence estimate was 7.37 per 100 py (95% CI: 5.09–9.65). Among the two studies comprising 763 participants who examined people in custodial settings, the pooled HCV reinfection incidence estimate was 7.23 per 100 py (95% CI: −2.13-16.59). Heterogeneity remained high across each subgroup estimate (Figure 3 and Table S2).

3.3.2 By testing interval

Twenty-three studies comprising 5058 participants were categorised as having testing intervals ≤6 months, with a pooled estimate of 4.26 per 100 py [95% CI: 2.86–5.65). Fifteen studies comprising 3873 participants were categorised as having testing intervals >6 months, with a pooled estimate of 5.19 per 100 py [95% CI: 3.92–6.46). High heterogeneity was observed across both testing interval groups (Figure 4 and Table S2).

3.3.3 By key population groups and testing interval

Among PWID, 17 studies comprising 3520 participants reported testing intervals of ≤6 months with HCV reinfection incidence rates ranging from 0.00 per 100 py to 21.50 per 100 py. Eleven studies comprising 1663 participants reported testing intervals >6 months, with incidence rates ranging from 0.00 per 100 py to 31.00 per 100 py. The pooled HCV reinfection incidence rate estimate was lower among PWID populations with testing intervals ≤6 months (2.97 per 100 py [95% CI: 1.55–4.39]) compared with those with testing intervals >6 months (3.96 per 100 py [95% CI: 2.64–5.29), though noting the presence of overlapping confidence intervals (Figure 5).

Among MSM, seven studies comprising 1945 participants reported a testing interval of ≤6 months with HCV reinfection rates ranging from 5.93 per 100 py to 27.80 per 100 py. Three studies comprising 1608 participants reported testing intervals >6 months, with HCV reinfection rates ranging from 3.46 per 100 py to 17.00 per 100 py. The pooled HCV reinfection incidence rate estimate among MSM populations was higher among studies with testing intervals ≤6 months (7.94 per 100 py [95% CI: 3.30–12.57]) compared with those with testing intervals >6 months (6.86 per 100 py [95% CI: 4.66–9.05]), though noting the presence of overlapping confidence intervals. Moderate-to-high heterogeneity was observed across all PWID and MSM groups (Figure 6). Among people in custodial settings, low study numbers limited the ability to compare testing frequencies.

4 DISCUSSION

In this systematic review of HCV reinfection intervals, we observed higher HCV reinfection incidence rates among studies including MSM compared with PWID. We detected no difference in HCV reinfection incidence based on retesting intervals. There has been interest globally in WHO guidelines development²⁴ to determine optimal testing intervals for people at risk of HCV reinfection. A greater detection of HCV reinfection in studies with shorter testing intervals has been noted and explained previously, in part, due to less time for infections to spontaneously clear.²⁵ Conversely, partial protective immunity from the primary infection may lead to a more rapid resolution of reinfection.²⁶ It is also likely that higher incidence rates in studies which tested more frequently is biased by higher frequency testing protocols in studies of cohorts at greater risk of reinfection. This large meta-analysis has been able to explore both hypotheses and demonstrate no clear difference in very frequent (less than 6 months) or less frequent (more than 6 months) retesting.

Previous systematic reviews have estimated comparative pooled incidence rates of HCV reinfection among key populations and different HCV testing intervals. An earlier review found that rates of HCV reinfection were lower in studies of HCV mono-infected PWID, MSM and prisoners (1.91/100py) compared with HIV/HCV co-infected individuals (3.20/100py).²⁷ Multiple reviews have reported a lower HCV reinfection incidence among individuals receiving opioid agonist therapy (OAT) (ranging from 0.55 to 1.4/100py) compared with those not receiving OAT.^{28, 29} Another meta-analysis among people living with HIV found that HCV reinfection was higher among MSM (5.89/100py) compared with people with recent injection drug use (5.49/100py).³⁰ Another meta-analysis among HIV-infected MSM estimated an overall HCV reinfection rate of 5.27/100 py and found a higher reinfection rate among people with HCV test intervals of less than 6 months (7.59/100py) compared with those tested at greater than 6 month intervals (2.88/100 py).³¹ Among PWID populations, our findings are consistent with previously reported estimates in this group and are higher compared with studies examining HCV reinfection among MSM. A strength of this review is that all individuals at risk were compared using the same methodical approach allowing for better understanding of relative risks in these key populations.

Frequent testing is likely to increase case finding and be beneficial in reducing HCV disease burden for both the individual and community through early detection, treatment and cure as part of a TasP approach, and can be combined with other testing strategies, such as HIV and STI testing. This could potentially improve linkage to care and harm reduction support through increased engagement with healthcare services.^{15, 32} Although a person's risk of reinfection may decline over time, it is difficult to determine when in the post-SVR period any reinfection occurs, and thus if different testing intervals should be offered at different years. Routine HCV testing among people actively using drugs has been shown to be cost-effective in multiple settings, even when repeat testing leads to the need for repeat treatment, due to both the low cost of HCV testing and effective treatment.³³ However, implementing testing for reinfection at regular intervals for all PWID may not be feasible in many settings. As testing for HCV reinfection relies on RNA or antigen testing, limited availability of these testing strategies may pose an additional challenge in certain settings. Further, in resource-constrained settings without universal healthcare or where systems that can deliver effective care post-diagnosis are not available, increased screening may not be cost-effective. Consideration should be given to higher short-term costs of tests, outpatient visits including healthcare personnel, the opportunity cost for time and resources diverted by healthcare staff, and if more re-infections are identified, the short-term costs of increased treatment. Furthermore, while frequent testing may diagnose individuals during the acute HCV phase, many DAAs are only approved for confirmed chronic HCV infection. Additionally, testing must be voluntary and offered alongside counselling to minimise the risk of stigma, discrimination and adverse psychological impacts.^{32, 34} While our analyses show no difference in the reinfection incidence rate based on retesting interval, recommendations for screening frequency should consider country-level or setting-specific contexts, including the availability of appropriate linkage to treatment following diagnosis.

There are important limitations to note when interpreting our findings. First, the purpose of our systematic review was to find RCTs and other comparative studies to determine the association between HCV testing frequency and HCV reinfection among key populations in the post-SVR period; however, only single-arm observational studies were identified. We have thus explored HCV reinfection between pooled estimates of observational studies, limiting our ability to determine the effectiveness of different testing regimes on HCV detection. Second, only studies that tested at discrete time intervals were included, and as such results were more representative of studies from high-income countries. Studies where clinicians self-selected when to offer testing were not included to reduce the influence of higher risk individuals being tested more often. Third, only nine studies were included among MSM and two among people in custodial settings, limiting our ability to discern meaningful differences between testing frequencies for these key populations. Fourth, our analysis could not account for confounding factors, such as age, gender and socioeconomic status since it was inconsistently reported within studies. Fifth, the studies included in this review contain significant heterogeneity among population characteristics, where some studies included participants with life-time drug use, those with only recent drug use or injecting drug use or those on opioid substitution therapy or other risk reduction measures. Due to this heterogeneity, along with incomplete reporting by study authors, we were unable to disaggregate the PWID sub-population by recent and ever injecting drug use, limiting the ability to confidently detect differences between these groups. This heterogeneity has also been observed in previous systematic reviews including the ones described above. Additionally, the inclusion of both recent and past injecting exposure within the PWID cohorts, and the fact that testing frequency within study protocols may have been influenced by the risk profile of included participants (i.e. less frequent testing for those without recent injecting exposure) is likely to be a significant factor altering estimates of incidence of reinfection among this key population. Sixth, our review did not find any studies examining HCV reinfection among transgender people. While it is possible that transgender people were included in our studies, often as part of MSM cohorts, this finding is most likely reflective of deficiencies in the collection and reporting of gender identity in health records and research. The adoption of a gender lens to HCV care and research to explore the impacts of gender disparities on HCV elimination is required.³⁵ Finally, as nearly all included studies were from high-income country settings, any consideration of HCV testing frequency recommendations should be applied to high-income settings only due to differences in HCV risk context including local drivers of HCV transmission and availability of resources. This highlights the crucial need for more context-specific HCV elimination research from low- and middle-income countries.

This systematic review advances our understanding of how different testing intervals influence HCV detection among PWID, MSM, and people in custodial settings. Furthermore, this review updates the HCV reinfection incidence estimates among these key populations. Our findings have highlighted the absence of high-quality trial or cohort data to make direct comparisons on testing frequency, and suggest that future longitudinal studies comparing annual testing with more frequent testing (i.e. 3–6 monthly) among key populations are needed. Our findings have direct implications for clinical practice and have contributed to WHO global testing recommendations for key populations, where people at ongoing risk and a history of previous HCV infection may be offered 3–6 monthly HCV testing where appropriate and available.²⁴ Increasing voluntary testing frequency coupled with offers of HCV treatment among people at ongoing risk could have significant individual and population level benefits, enabling further progress towards global HCV elimination.

AUTHOR CONTRIBUTIONS

Joseph S. Doyle is the guarantor of this article. Authors Joseph S. Doyle, Margaret E. Hellard, Mark A. Stoové, Ned H. Latham, Rachel Baggaley, Virginia MacDonald, Annette Verster, Nandi Siegfried, Stephanie C. Munari, Michael W. Traeger, Virginia MacDonald, Ned H. Latham and Lakshmi Manoharan contributed to the conceptualisation and design of the research study. Nandi Siegfried provided methodological expertise. Brian Conway, Marina Klein and Julie Bruneau contributed to the acquisition and interpretation of data. Stephanie C. Munari, Michael W. Traeger and Vinay Menon performed the search, data collection and analysis. Stephanie C. Munari, Michael W. Traeger and Joseph S. Doyle prepared the original draft manuscript. All authors were involved in reviewing and revising the manuscript and have approved the final version for publication.

FUNDING INFORMATION

This work was supported by the World Health Organisation's Global HIV, Hepatitis and Sexually Transmitted Infections Programmes who commissioned and funded the project. Copyright in the original work on which this Article is based belongs to WHO. The authors alone are responsible for the views expressed in this Article and they do not necessarily represent the views, decisions, or policies of the institutions with which they are affiliated. Burnet Institute acknowledges support from the Victorian Government Operational and Infrastructure Fund. MWT, JSD, MEH and MAS acknowledge fellowship support from the National Health and Medical Research Council.

CONFLICT OF INTEREST STATEMENT

MWT has received speaker's fees and investigator-initiated research funding from Gilead Sciences. NS provides consultancy to act as a methodologist for the WHO guidelines related to this manuscript including contracts and consulting fees. JSD, MEH and MAS's institution has received funding for research and speaking from Gilead Sciences and AbbVie. MAS reports unrelated consultancy fees and advisory board participation for Gilead Sciences and AbbVie. JB receives grants or contracts from Canadian Institutes on Health Research, Gilead Sciences, National Institute on Drug Abuse (USA) and AbbVie and reports advisory board participation for Gilead Sciences and AbbVie. All other authors declare no potential conflicts of interest.