Acute hepatitis C infection among adults with HIV in the Netherlands: a capture-recapture analysis

Background Reliable surveillance systems are essential to assess the national response to eliminating hepatitis C virus (HCV), in the context of the global strategy towards eliminating viral hepatitis. Aim We aimed to assess the completeness of the two national registries of acute HCV infection in people with HIV, and estimated the number of acute HCV infections among adults with HIV in the Netherlands. Methods For 2003-2016, cases of HCV infection and reinfection among adults with a positive or unknown HIV-serostatus were identified in two national registries: the ATHENA cohort, and the National Registry for Notifiable Diseases. For 2013-2016, cases were linked, and two-way capture-recapture analysis was carried out. Results During 2013-2016, there were an estimated 282 (95%CI: 264-301) acute HCV infections among adults with HIV. The addition of cases with an unknown HIV-serostatus increased the matches (from N=104 to N=129), and a subsequently increased the estimated total: 330 (95%CI: 309-351). Underreporting was estimated at 14-20%. Conclusion In 2013-2016, up to 330 cases of acute HCV infection were estimated to have occurred among adults with HIV. National surveillance of acute HCV can be improved by increased notification of infections. Surveillance data should ideally include both acute and chronic HCV infections, and be able to distinguish between acute and chronic infections, and initial and reinfections. Data were obtained from two national registries: the ATHENA national observational HIV cohort at SHM and the National Registry for Notifiable Diseases at the RIVM. We identified cases of acute HCV infection among adults (aged ≥18 years at HCV diagnosis) in the period


Introduction
Hepatitis C virus (HCV) infections are generally uncommon in the Netherlands, with a chronic HCV prevalence of <0.2% [1]. However, an increase in the number of acute HCV infections and reinfections among men who have sex with men (MSM) who are HIV-positive has been reported since early 2000 [2][3][4]. Detection, diagnosis and registration of acute HCV infections are crucial to measure trends in the epidemic, and plan appropriate public health and clinical interventions, such as prevention programs for those at risk, targeted testing, increasing treatment uptake, and contact tracing to reduce subsequent transmission. In addition, treating chronic HCV infections with direct acting antivirals (DAAs) is expected to strongly reduce, but not eliminate, the HCV epidemic among HIV-positive MSM [5,6]. Despite high DAA uptake in HIV/HCV co-infected people in the Netherlands [7], ongoing HCV transmission and reinfection continue to occur.
In the Netherlands, the Dutch HIV Monitoring Foundation (Stichting HIV Monitoring, SHM) [8] registers acute HCV infection among people with HIV since 2000, and the National Registry for Notifiable Diseases at the National Institute for Public Health and the Environment (Rijksinstituut voor Volksgezondheid en Milieu, RIVM) [4] registers acute HCV infections among people irrespective of HIV-serostatus since 2003. While the RIVM and SHM have reported similar trends of acute HCV infection over the years, consistently more cases were registered among HIV-positive MSM by the SHM, compared to the RIVM [4,9].
Reliable surveillance systems, monitoring and evaluation are essential to assess the national HCV response, in the context of the global strategy towards eliminating viral hepatitis [10,11].
The question arises whether SHM and the RIVM register the same cases of acute HCV infection among people with HIV in the Netherlands, and whether cases are missed by one registry or another. The aim of this study is to assess the completeness of these two national registries. In addition, we estimate the number of acute HCV infections among the HIVpositive population in the Netherlands, by means of capture-recapture analysis.

Data sources
Data were obtained from two national registries: the ATHENA national observational HIV cohort at SHM and the National Registry for Notifiable Diseases at the RIVM. We identified cases of acute HCV infection among adults (aged ≥18 years at HCV diagnosis) in the period All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint of 2003-2016 in both registries. Both cases of a primary episode of acute HCV infection and reinfection were included; reinfections were treated as independent cases.

ATHENA cohort at SHM
Since 2000, SHM has managed the ATHENA cohort and is responsible for registering all HIV-positive people in care at the 26 HIV-treatment centres in the Netherlands [8]. At entry in HIV care, people are informed of the cohort and the purpose of data collection by their treating physician; participant consent follows an opt-out procedure (2% opt-out) [8]. After linkage to HIV care and registration at SHM, people are enrolled in the ATHENA cohort, which systematically collects demographic and clinical data from medical records, including information on HCV co-infection. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Additionally, we assessed the influence of the distinction between acute and chronic HCV diagnosis on the registration in both databases. We linked RIVM cases of acute HCV infection among HIV-positive adults to all SHM cases (i.e., acute, chronic, and other).

Case-linkage and statistical analysis
Subsequently, we repeated this analysis containing all RIVM cases of acute HCV infection, including cases with an unknown HIV serostatus. All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint The capture-recapture method The capture-recapture method is derived from ecology and applied to epidemiological studies [17]. To estimate the number of cases in a population, cases are captured in one data source and then independently recaptured in a second data source. The capture-recapture method analyses the degree of overlap between the two data sources (Figure 1). The capture-recapture estimate is adequate when the following assumptions hold [18,19]: i) the study applies to a closed population, i.e., the definition or interpretation of the 'target population' is explicit; ii) linkage between the cases is possible and reliable; iii) the databases are functionally independent and do not rely on each other; vi) every case has the same probability to be included in each database.
For this study, we performed a two-way capture-recapture analysis to estimate the total number of cases of acute HCV among HIV-positive people based on the two databases for the period of 2013 to 2016. We calculated the adjusted Petersen-Lincoln estimate [18,19]  Adjusted Petersen-Lincoln estimated number of cases: With the associated variance of the estimate of the total population: And the 95% confidence interval of the estimate of the total population: Where:

Ethics
At initiation, the ATHENA cohort [8] at SHM was approved by the institutional review board of all participating centres and patient consent is received by opting out. Data are All rights reserved. No reuse allowed without permission.
author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint pseudonymised to investigators and may be used for scientific purposes. The National Registry for Notifiable Diseases at RIVM anonymously registers mandatory communicable disease notifications [20], as a legal obligation enforced by the Dutch public health act [15] and International Health Regulations by the World Health Organisation [21].
For this study, we received approval from the steering committees of both institutions (i.e., RIVM and SHM). To secure non-tractability, personal identifiers were removed from the datasets before case-matching, using surrogate identifiers. A designated quality management coordinator safeguarded privacy protection and compliance with the European General Data Protection Regulation [22]. This work complies with the principles laid down in the Declaration of Helsinki [23].

Acute HCV infections among adults in 2003-2016.
Cases of HCV infection were extracted from both registries as described in Figure 2, and summarized in Figure 3 and Appendix table 1.  Table 1).
We restricted our capture-recapture analysis to the 213 cases of acute HCV infection among  author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint   Table 1 author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint acquired through other sexual contact (N=2, heterosexual and unknown), injecting drug use (N=2), a needle stick or bite incident (N=1), or another or unknown route (N=7).

Chronic HCV infections and other HCV cases
Additionally, we linked all 632 cases of HCV infection among HIV positive adults registered by SHM (i.e., acute, chronic, and other), with the 142 cases of acute HCV infection adults who were HIV positive, registered by the RIVM (Appendix Table 2). All but two RIVM-cases could be linked to an acute, chronic, or other HCV case registered by SHM (98.6%; 140 out of 142). Furthermore, when we added the 58 RIVM-cases of acute HCV infection among people with an unknown HIV serostatus --i.e. when we linked the 632 SHM cases with the All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the The results should be interpreted while taking into account the study's limitations. Notably, capture-recapture analysis is based on several assumptions, which influence the accuracy and reliability of the presented estimates [18,19]. Our study fulfils all four assumptions.
However, incorrect case linkage and non-linkage cannot be ruled out because linkage was based on personal characteristics, and not on personal identifiers.
Our annual estimates of the number of acute HCV infections among people with HIV are in line with HCV incidence studies performed among HIV-positive MSM both in Amsterdam [24] and across Europe [25]. The estimated number of acute HCV infections was lower in 2016 than in to 2013, 2014 and 2015, which could be explained by a delay in registration, a change in the frequency of HCV testing[26], or an actual decrease in HCV incidence among HIV-positive people in the Netherlands, most likely due to increased DAA uptake. The latter is in line with recently published findings covering 17 out of 26 Dutch HIV-treatment centres, relating the high DAA uptake to a lower HCV incidence among HIV-positive MSM in 2016, than in to 2014 [5]. Of note, annual numbers are known to fluctuate over time, also before 2013, and recent numbers reported by the RIVM showed a subsequent increase in the total number of HCV infections for 2017 compared to 2016 [4].
Underreporting at both registries could be explained by several factors. At SHM, underreporting of acute HCV infections could be due to missed or incomplete screening for HCV by the HIV treating physician. Virtually all RIVM cases of acute HCV infection could be identified at SHM, although some RIVM cases of acute HCV were registered as chronic HCV at SHM. Previous analyses of the SHM have data shown substantial variation in routine HCV screening for all HIV-positive people at entry into HIV care, as well as in annual HCV screening among risk groups (i.e., MSM) [2]. This may, to some extent, be explained by All rights reserved. No reuse allowed without permission. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the To control and eventually eliminate HCV from the HIV-positive population in the Netherlands and beyond, all HIV/HCV co-infected people should be diagnosed, linked to, and retained in care [10,11]. Registration is essential to monitor trends in the epidemic, and to assess the All rights reserved. No reuse allowed without permission.
author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint impact of treatment and other interventions on HCV incidence. The high rate of HCV reinfections among HIV-positive MSM [2,31] warrants the need for increased awareness and prevention measures for high risk MSM, as well as structural repeat screening of HCV infection [32,33]. We call for increased attention for both documentation of HCV cases in medical records and disease notification at RIVM, by all health professionals. Surveillance data should ideally include both acute and chronic HCV infections, and be able to distinguish between acute and chronic infections, and initial and reinfections. Robust HCV surveillance is essential to monitor the impact of prevention programs and DAA treatment uptakeon both the individual and public health level. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint author/funder, who has granted medRxiv a license to display the preprint in perpetuity. author/funder, who has granted medRxiv a license to display the preprint in perpetuity. author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/19002097 doi: medRxiv preprint