Hepatitis C virus testing in a clinical HIV cohort in Ontario, Canada, 2000 to 2015

Abstract Background HIV‐positive individuals may acquire HCV via injection drug use (IDU) and condomless anal sex. HIV care provides opportunities for HCV testing and cure with direct‐acting antiviral agents (DAAs). Methods We analyzed data from the Ontario HIV Treatment Network Cohort Study. Among those not HCV‐positive or diagnosed previously (n = 4586), we used Cox regression to test the rates of ever HCV testing (serological or RNA) in HIV care by DAA era (pre‐DAA: 2000‐2010; after DAA: 2011‐2015) and compared the proportion diagnosed with HCV. We identified correlates of annual proportions of serological testing using Poisson generalized estimating equations. Results After DAA vs pre‐DAA, the hazard rate ratio (95% CI) of ever HCV testing was 1.70 (1.59, 1.81). The proportion (95% CI) tested annually increased from 9.2% (8.0%, 10.7%) in 2000 to 39.1% (37.1%, 41.1%) in 2015 (P < 0.0001). The proportion diagnosed with HCV declined by 74% pre‐DAA to 11% after DAAs. Annual testing increased per calendar year (16% steeper slope after DAA vs pre‐DAA) and was more common among men who have sex with men; those more educated (post‐secondary vs ≤ high school); and those positive for syphilis or reporting any IDU. Annual testing decreased per decade of age and time since HIV diagnosis. Discussion Annual HCV testing increased over time with higher testing among those reporting sexual or IDU risk factors, but fell short of clinical guidelines. Targeted interventions to boost testing may be needed to close these gaps and reach WHO 2030 HCV elimination targets.

Discussion: Annual HCV testing increased over time with higher testing among those reporting sexual or IDU risk factors, but fell short of clinical guidelines. Targeted interventions to boost testing may be needed to close these gaps and reach WHO 2030 HCV elimination targets.

K E Y W O R D S
Coinfection/epidemiology, hepatitis C/epidemiology*, hepatitis C virus testing, HIV infections/ epidemiology*, HIV-HCV co-infection

| INTRODUCTION
People living with HIV are vulnerable to the acquisition and consequences of Hepatitis C virus (HCV) co-infection due to biological and social factors. In Canada, 18% to 20% of those living with HIV are co-infected with HCV compared to <1% in the general population. 1,2 Direct-acting antiviral (DAA) drugs, approved by Health Canada in 2011, are highly efficacious and curative (>95%) even in co-infected individuals, though uptake was initially low. Timely HCV testing and treatment with DAA drugs can ameliorate clinical complications and interrupt ongoing HCV transmission, thus making it possible to reach elimination goals set by the World Health Organization (WHO) by 2030 (90% diagnosis, 80% reduction in HCV incidence, and 65% reduction in HCV-related mortality). 3 In 1999, U.S. guidelines first recommended testing all HIVpositive individuals for HCV 4 though HCV testing in Canada remained risk or symptom based even after approval of DAAs in 2011 and interferon-free, all-oral regiments in 2014. 5 It was not until 2016 that new Canadian guidelines recommended one-time HCV testing when first evaluated for HIV, followed by annual retesting for "high risk" individuals such as people actively injecting drugs and sexually active HIV-positive men who have sex with men (MSM) engaging in "high risk" behaviors. 6 In 2011, 44% of those chronically infected with HCV were estimated to be undiagnosed in Canada. 7 Diagnosis rates were thought to be higher for people living with HIV, as attending HIV care provides opportunities for HCV screening and treatment. To our knowledge, there are no published reports of HCV testing in this population in Canada, which could provide historical estimates for inputs of mathematical models of HCV transmission and identify gaps in HCV care cascades and barriers to reaching WHO elimination targets. Therefore, we sought to characterize temporal patterns of HCV testing between 2000 and 2015 in a clinical cohort of HIV patients in Ontario, the province that comprised 38.7% of all reported HCV cases in Canada in 2009 and has the largest population of people living with HIV. 2 We aimed to estimate the annual proportion that had tested for HCV, the frequency of annual serological HCV tests, and the proportion diagnosed with HCV. We hypothesized that HCV testing trends would reflect the testing guidelines and treatment options over time, with higher testing after DAA approval (2011 onwards) and in groups perceived to be at higher risk for HCV acquisition.

| METHODS
We used data from the open, prospective Ontario HIV Treatment Network Cohort Study (OCS), which has been described previously. 8 Briefly, the OCS represents almost 25% of the HIV patients under care in Ontario, consisting of participants aged 16 or older who volunteered to be a part of the study and accessed HIV care at any of nine participating clinics. From 1995 to 2007, participants selfcompleted a questionnaire at enrollment; since 2008, they were interviewed annually. 8 Clinical data was abstracted from medical charts.
The study protocol, research instruments and forms received ethical approval from the University of Toronto Human Subjects Review Committee and from the individual study sites.

| Testing and laboratory data
We obtained testing data for HIV viral load, HCV and syphilis through linkage with OCS clinical records and the provincial Public Health Ontario Laboratories (PHOL), the sole provider of HIV viral load and syphilis serological tests in Ontario. In Ontario, HCV serological testing can be performed by private laboratories, hospital laboratories, or at the PHOL. However, almost all confirmatory HCV serological tests and HCV-RNA detection and quantification are conducted by the PHOL. Testing for HCV antibodies (anti-HCV) is often the first recommended step in the testing algorithm, and if positive, is followed by HCV RNA detection, measurement and genotyping. We defined "ever testers" as those with at least one HCV test (serological or RNA), and "annual testers" as those with at least 1 HCV serological test in a calendar year that they were under observation (Details in Table S1).
HCV diagnosis was classified on the basis of laboratory test results (confirmed antibody test or positive RNA or genotype test) or notation of an HCV diagnosis in a participant's medical record.

| Inclusion criteria for analysis
As of December 2015, a total of 6891 participants had enrolled in the OCS. We restricted the analyses to participants who had at least linked 1 HIV viral load test and were enrolled in the OCS between 2000 and 2015 (1731 removed) so that any HCV testing with the PHOL would be captured. Accumulation of study time began at baseline, which we defined as the later of January 1, 2000, the date of first HIV viral load test or the date of the first OCS visit. Because we were interested in HCV testing patterns among people in HIV care who were not yet known to have HCV, we excluded those who had been diagnosed or tested positive for HCV prior to baseline (574 removed) based on either HCV diagnosis dates on medical records or positive HCV antibody, RNA, or genotype tests (remaining analytic sample, n = 4586). Additional restrictions for specific analyses are described below and summarized in Table S1. All statistical analyses were conducted with Stata v13 (College Station, TX) 9 and we used a complete case analysis strategy to handle missing data. Self-reported binary variables were used for sex and MSM. Ethnicity was categorized as white, black, Aboriginal/indigenous, other, or unknown. Based on previous analyses of HCV seroconversion among MSM in the OCS, 10 we considered having ever had syphilis as a proxy measure of high-risk sexual behavior for HCV acquisition; this was defined as a dichotomous record of any reactive syphilis test result. We analyzed age and HIV duration by decade, where the latter was based on an estimated date of HIV diagnosis. To account for any sociodemographic differences, we used dichotomous variables to classify education into "any postsecondary" vs "high school or less" based on the last reported education level and whether urban-dwelling or not (rural, out-of-province, or unknown) based on residential postal codes.

| Ever testing
Descriptive analyses: We defined "ever testers" as those with at least one HCV test (serological or RNA) from all dates available and used descriptive statistics to characterize participants overall in the analytic sample (n = 4586 individuals) or by specific HCV exposure groups, defined by history of IDU and possible sexual transmission.
Testing rate in HIV care by DAA era: We calculated the annualized rate of having ever had an HCV test (serological or RNA), that is, the cumulative incidence of having ever tested for HCV in each year. To do so, we restricted analytic time to years under OCS follow-up. We excluded participants who had an HCV test prior to baseline (1563 removed) or who had missing HCV test dates (six removed) for an analytic sample of 3017 individuals. Follow-up ended at the date of the first HCV test or was censored at the last viral load test, last OCS visit, last date of OCS site data collection, or December 31, 2015, whichever was earliest. To test the effect of DAA approval on time to first HCV test under HIV care, we used Cox regression with robust standard errors, where DAA era was included as a time-varying covariate. Using the Stata command -estat phtest-, we tested the proportional hazards assumption based on Schoenfeld residuals. To address left truncation, we conducted a sensitivity analysis incorporating delayed entry (year of first HIV viral load test) and year of HIV diagnosis as the origin in a subsample including only those who tested for HCV prior to baseline and without restricting to OCS follow-up (n = 5568).

| Annual serological testing for HCV
Among the analytic sample of 4586 individuals (39 337 person-years), we calculated annual proportions of serological testing as the number of people with at least 1 HCV serological test in the calendar year that they also had a viral load test ("annual testers") in the overall population as well as by HCV exposure groups, defined by history of any IDU. Follow-up ended at the earlier date of December 31, 2015, last viral load test, last OCS visit, last date of OCS site data collection or, for those who eventually tested positive for HCV, the date of their first HCV-positive test (serologic, RNA, or genotyping test). We identified correlates of annual testing using modified Poisson regression in a generalized estimating equations 10 framework with robust standard errors to account for repeated observations per person. Covariates were selected a priori and included in the final model if statistically significant (P < 0.05). Recent IDU, which was available only for a subset of participants who completed annual questionnaires after 2008, was used in a subanalysis.

| Number of serological tests per year
We used descriptive statistics and the nonparametric Wilcoxon signed rank tests for paired data to quantify and test for differences in number of HCV antibody tests per year by DAA era and transmission risk due to IDU or sexual contact. Follow-up time was defined as above for annual serological testing.

| HCV diagnoses
Diagnosis was based on either laboratory tests (confirmed antibody test or positive RNA or genotype test) or medical records. Among participants whose HCV status was unknown or HCV-negative at baseline, we calculated (1) the cumulative incidence of an HCV diagnosis and (2)

| RESULTS
A total of 4586 participants were followed for a median of 9 years (interquartile range 4-12 years; total 39 337 person-years). The majority were male (84%), white (63%), living in urban settings (88%), and classified as MSM (64%) ( Table 1). At baseline, 7.8% had a history of IDU; at follow-up, 2.8% reported ongoing IDU, such that 11% had ever injected drugs by the end of follow-up. At baseline, the median CD4 count was >350 cells and 72% were taking antiretroviral therapy (ART). By the end of follow-up, 95% had initiated ART. reported any IDU and almost a quarter (22%) remained under followup. On average at baseline, they were older than ever testers (43 vs 40) and had been living with HIV longer (7.5 vs 5.3 years).

| HCV testing while in HIV care
Among the 2593 participants whose first HCV test occurred while under OCS follow-up, the annualized rate of having ever had an HCV test was 70% higher after DAA approval (2011)(2012)(2013)(2014)(2015)

| Annual HCV serological testing
The annual proportion tested rose from 9.2% (95% CI 8.    Figure S1). In those with any IDU, the diagnosis rate dropped steeply and then more slowly; in the other subgroups, the HCV diagnosis rate stabilized after DAAs were approved in 2011 ( Figure 2). Dr. Ann Burchell had full access to all of the data in this study and takes complete responsibility for the integrity of the data and the accuracy of the data analysis.

TRANSPARENCY STATEMENT
Dr. Nasheed Moqueet and Ann Burchell (lead and senior author, respectively) affirm that this manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are not publicly available to protect the privacy of the participants but are available from the OHTN Cohort Study upon reasonable request and approval of the Governance Committee.