A population-based screening for hepatitis C antibodies and active infection using a point-of-care test in a low prevalence area

Background Data on the true prevalence of hepatitis C virus (HCV) infection in the general population is essential. We evaluated a program implementing free universal HCV screening using a non-invasive point-of-care test (POCT) (OraQuick-HCV rapid test) in oral fluid in an urban area in Valencia, South-Eastern Spain. Methods A cross-sectional study was performed during 2015–2017. Free HCV screening was offered by regular mail to 11,500 individuals aged 18 and over, randomly selected from all census residents in the Health Department. All responding participants filled in a questionnaire about HCV infection risk factors and were tested in their tertiary Hospital. In those with a positive POCT, results were confirmed by enzyme-immunoassay and HCV-RNA. Results 1,206 persons agreed to participate (response rate: 11.16%). HCV antibodies were detected in 19 (1.60%) cases (age-sex standardized rate: 1.31%; 95%CI: 0.82–2.07), but only 8 showed positive HCV-RNA (age-sex standardized rate: 0.56%; 95%CI: 0.28–1.14). The majority (89%) of the cases were born before 1965 and 74% had at least one known risk factor for HCV infection. All anti-HCV positive individuals were already aware of their infection, and no undiagnosed cases were detected. The performance of the POCT was excellent for detecting active infection. Conclusions These preliminary data suggest that HCV population screening with a POCT is feasible but, in our setting, mailing recruiting is not effective (11% response rate). The low prevalence of HCV antibodies and active infection in the participant population (with no new diagnoses made) suggests that, in our setting, underdiagnosis may be uncommon.


Results
1,206 persons agreed to participate (response rate: 11.16%). HCV antibodies were detected in 19 (1.60%) cases (age-sex standardized rate: 1.31%; 95%CI: 0.82-2.07), but only 8 showed positive HCV-RNA (age-sex standardized rate: 0.56%; 95%CI: 0.28-1.14). The majority (89%) of the cases were born before 1965 and 74% had at least one known risk factor for HCV infection. All anti-HCV positive individuals were already aware of their PLOS  Introduction Chronic Hepatitis C Virus (HCV) infection represents a considerable economic burden on health-care systems. [1][2][3] A reliable estimation of the actual prevalence of HCV infection is needed to meet the World Health Organization first action plan.
[4] In Spain, neither the current prevalence of HCV nor the proportion of undiagnosed cases are well-known. [5,6,7,8] Most studies on HCV prevalence in Spain have been conducted using the classical gold standard immunoassays (EIA/CIA) on serum or plasma samples and have focused on small portions of the population, often those considered to be at high risk [7,8]. The applicability of standard EIA/CIA assays for indiscriminate population screening is hampered by the requirement for venipuncture and specialist laboratory intervention. Screening with point-of-care tests (POCT) guarantees a rapid delivery of information and allows clinical decisions to be taken in a timely manner. One of them (OraQuick HCV) has been approved by the FDA and the EU for the diagnosis of HCV antibodies in venous blood, fingerstick, or oral mucosal transudate (OMT) [9,10]. This test is minimally (fingerstick) or non-invasive (OMT), simple to perform and interpret, and has shown a sensitivity and specificity equal or even superior to the gold standard (EIA/CIA) [11,13], making HCV screening in non-hospital settings feasible [8,[10][11][12][13][14].
The OraQuick HCV rapid test has been used in several countries in small pilot studies to evaluate screening methods for the general population and/or specific risk groups [14][15][16]. We recently reported the performance of the test as a diagnostic technique for a rapid detection of anti-HCV antibodies in OMT and fingerstick blood samples. In our hands, the clinical sensitivity and specificity of the OraQuick HCV rapid test in OMT was up to 99.1% and 100%, respectively [13]. Thus, screening with point-of-care tests (POCT) represents an alternative to venous blood for population screening of active HCV infection in order to increase the number of tested subjects, and to include difficult-to-reach populations. [12][13][14]17] We report a mailing recruiting intervention for universal HCV screening with a POCT, targeting the general population of a health department in the city of Valencia, south-eastern Spain, to estimate the prevalence of anti-HCV, active HCV infection, and undiagnosed cases at the population level.

Study design
We conducted a cross-sectional universal screening study in the general adult population of a Healthcare Department in Valencia (Spain), from September 2015 to October 2017 using the OraQuick anti-HCV rapid test followed, if positive, by gold-standard immunoassay (CIA) and characterized HCV positive cases including patient and virus characteristics.

Setting
The Valencia La Fe Health Department is part of the Valencia Public Healthcare System, a healthcare network territorially organized providing universal healthcare coverage, free of charge except for a co-payment for drugs dispensed outside hospital, for about 95% of the residents in the Valencia Community. The Valencia Healthcare System has an information system with a unique identifier of all people covered (SIP, by its Spanish acronym), linking various administrative data including age, sex, address and the assigned hospital district. From this system, successive random samples of people assigned to our Hospital District were extracted by the informatics office of our Health department. In 2015, the districts assigned to "La Fe Department" included 153,318 inhabitants aged 18 and over, served by one public university hospital and several primary healthcare centres providing inpatient and outpatient care to all residents in its demarcation.

Population and sample
The target population was composed of all individuals aged 18 and over covered by the Valencia Public Healthcare System in the Healthcare Department. Assuming an HCV prevalence of about 2%, we estimated a sample size for HCV screening of about 6,000 individuals (n = 5,934 with a 95% confidence interval and 0.5% precision), expecting a 50% response rate based in our experience with other surveys. Due to low recruitment rates, we later increased the number of individuals we attempted to contact in a second sample population of 30,000 target adults randomly selected by age and sex strata as the previous sample (see Results). From July 2015 to October 2017 a proposal for free screening was offered to 11,500 individuals in successive waves of ordinary mail. Target individuals were adults over 18 years of age, randomly selected by age and sex strata in order to obtain a participant sample close to the overall district population. An informative leaflet of the HCV screening procedure was included in the mailing.

Inclusion and exclusion criteria
Inclusion criteria were persons of 18 years and over who had received a proposal by random population sampling and agreed to participate in the study by signing the informed consent form. Exclusion criteria were those who did not meet the inclusion criteria, and/or refused to participate.

Risk factors questionnaire/demographic characteristics
Before performing the OraQuick HCV rapid test, all participants were invited to sign the informed consent form and fill out a questionnaire regarding demographic characteristics and risk factors for HCV infection, based on the World Health Organization Global Hepatitis Report [4] ( Table 1). Information regarding viral hepatitis, HCV infection, risk factors associated with HCV transmission, and potential consequences of HCV infection were provided to all participants.

OraQuick HCV rapid antibody test
The OraQuick HCV rapid test is a single-use lateral flow indirect immunoassay method approved by the FDA and CE-marked for the detection of anti-HCV antibodies in oral mucosal transudate (OMT), blood, serum or plasma, with a sensitivity and specificity equal or even superior to gold standard EIA/CIA assays [9][10][11][12]. In our hands, the clinical sensitivity and of the OraQuick HCV rapid antibody test in OMT was previously established up to 90,1% and 99.1% for past (non-viremic) or current (viremic) HCV infection, respectively; with a clinical specificity of 100% in both cases [13]. The test was performed in OMT at the Health Department's reference Hospital according to the manufacturer's instructions.

Follow-up of anti-HCV positive patients
Patients who tested anti-HCV positive with the OraQuick HCV test underwent a serological confirmation test for HCV antibodies in venous blood by a gold-standard immunoassay (COBAS Elecsys1 Anti-HCV-II electrochemiluminescence immunoassay Roche diagnostics, Mannheim, Germany). Those with a confirmed positive test were offered appropriate care that included determination of viral genotype (Versant1 HCV Genotype LiPA 2.0, Siemens Healthcare Diagnostics, Berkeley, USA), HCV-RNA levels (real-time PCR-based assay Cobas1 TaqMan HCV v2.0, Roche Diagnostics, Sant Cugat, Spain) and liver fibrosis stage (elastography and/or liver biopsy), as well as indication for antiviral therapy.

Statistical analyses
First, the characteristics of the participants were described using means or proportions according to the type of variable, and including their 95% confidence intervals (CI) when appropriate. Then the rates of individuals with anti-HCV antibodies and with active infection, standardized by age for men and women, were estimated separately, along with the rates standardized by age and sex for the whole population (allowing an adjustment in HCV prevalence in line with the different age-sex distribution between the sample and the reference population), and then according to specific characteristics and risk factors of the population with their corresponding CI. Finally, the characteristics of the cases with active infection were compared with those of non-viremic using Fisher's exact test. All analyses were performed using STATA1 software version 13 (StataCorp, College Station, Texas). A minimal anonymized data set is available in the Dryad Digital Repository <https://datadryad.org/> with the doi:10. 5061/dryad.9ghx3ffdd.

Ethics
This study was conducted in accordance with the Declaration of Helsinki and the study protocol was approved by the Research Ethics Committee of the Hospital Universitario y Politécnico La Fe (2015, April 7; reference number: 2014/0430). Written informed consent was obtained from all participants prior to their enrolment in the study.

Study population
First, we prepared a population sample of 6,000 target adults over 18 years of age, randomly selected by age and sex strata, with the intention of sending them successive waves of invitation letters according to the workloads assumable by the research team. During the first mailshot wave in September, 2015, 300 letters were sent, but only 18 (6%) of the individuals mailed responded agreeing to participate in the study. We speculated that the low response could be due to the recent summer holyday season, but following rates were also was below 10%. Therefore, in the successive mail waves other strategies were tried to improve the response rate, including the incorporation of a pre-determined appointment and direct telephone contact by a doctor.
The incorporation of a pre-determined appointment in the invitation letter (110 letters were sent with pre-determined appointment) procured a response rate of 12.7%, with no substantial increase over the previous strategy and a greater complexity in the management of appointments.
Using a direct telephone contact approach (we attempted to contact 467 people after sending them the corresponding invitation letter), we get to contact to and interview with 178 persons (38.1% of telephoned individuals, details in S1 Table). Only 100 of the interviewed (56.2%) confirmed to have previously received the letter, 53 (29.8%) claimed not to have received the letter, and 25 (14.0%) were not sure. From the telephone interviews, 68 (38.2%) agreed to participate in the study while 110 (61.8%) declined. The most common reasons for declining participation were change of address, schedule incompatibility, elderly person/disability, and refusing to participate in research. Nevertheless, only 53 (11,3% of the 467 telephone attempts) finally approached the Hospital and could be included in the study (23-30.26%-male and 30-29.41%-female). Due to the discreet success of the recruitment protocol modifications, and because of the increase in the time burden dedicated to phone contacts, we decided to continue with the initial mailing format, but increasing the number of individuals we attempted to contact. and prepared second sample population of 30,000 target adults randomly selected by age and sex strata as the previous sample.
In all, from September 2015 to November 2017 a total of 11,500 invitation letters were sent to the target population. A total of 689 (5.99%) letters were returned to sender (change of address, incomplete, erroneous address or death). From the remaining 10,811 invitations, 1,206 individuals agreed to participate in the study with a final response rate of 11.16% (Fig 1).
The main characteristics of the study participants are summarized in were born in Spain, and 24.42% had at least one risk factor for HCV infection. The age-sex structure of the participants was quite different from the age-sex structure of the hospital district, with the youngest and oldest groups being underrepresented, while the intermediate age groups were overrepresented, especially those between 55 and 75 years of age (S2 Table). Women were also slightly overrepresented with respect to men (S2 Table).

Prevalence of HCV antibodies as determined with the OraQuick HCV test
Nineteen out of 1,206 sampled individuals (1.58%) had HCV antibodies according to the Ora-Quick HCV test in OMT, which translates into an age-sex standardized prevalence rate of 1.31 (95%CI: 0.82; 2.07) positive cases by 100 district inhabitants (Table 3) All anti-HCV positive individuals were already aware of their infection, but three of them had previously abandoned medical (hepatology) monitoring (after screening they were reincorporated into the appropriate care schemes). The OraQuick HCV rapid test failed to detect two patients with a past positive result for anti-HCV antibodies. These two patients had been infected in the past (>10 years before). According chart evaluation one had cleared the virus spontaneously while the other had cleared the virus after successful antiviral treatment with interferon plus ribavirin.  Individuals with a positive OraQuick HCV test in OMT were prospectively assessed by a hepatologist. Chart review together with new HCV-RNA testing (limit of detection 15 IU/mL) indicated that, of the 19 OraQuick HCV positive cases, 8 (42.1%) were patients with active

Characteristics of individuals with a positive OraQuick HCV test result
The HCV genotype distribution together with stage of fibrosis and baseline features of Ora-Quick positive patients (both HCV-RNA positive and negative) are shown in Table 4. No substantial differences were found regarding viral genotype distribution, baseline features or stage of fibrosis.
Of the non-viremic individuals, one had spontaneously resolved the infection 15 years before OraQuick HCV testing, one had cleared the virus after interferon treatment 23 years before OraQuick HCV testing, two had cleared the virus following interferon and ribavirin treatment 6 and 16 years before OraQuick HCV testing, while 7 had achieved a sustained virological response (SVR) after receiving a course of direct-acting antiviral (DAA) therapy at a median of 23 months before OraQuick testing. All the 8 viremic patients were also aware of their infection; five were actively being followed up by a specialist and have been now treated with DAA agents. Regarding the other three patients previously lost for follow up, two have started therapy with DAA agents while the last one is awaiting new evaluation of liver fibrosis.

Discussion
To our knowledge, this is the first population-based universal screening study in Spain randomly sampling the total population in the catchment area. The most relevant findings of our study show: 1) a low effectiveness of recruitment based on ordinary mail with a participation  Our recruiting strategy by regular mail showed to be poorly effective, despite trying several changes in the recruitment process. Using a direct telephone contact approach gave us the opportunity to ascertain if the letters reached the target individuals, and to explore the potential reasons of declining participation. Only 56.2% of the interviewed confirmed to have received the letter, suggesting an actual loss of 43.8% of the target. From the 38% who agreed to participate in the study by the phone, 30% finally approached the Hospital and could be included in the study. Some of the justifications given for declining to cooperate by those interviewed were change of address, schedule incompatibility, elderly person/disability, advanced age and/or disability and refusing to participate in research mainly due to a healthy status with lack of interest in the screening proposal. Despite the increased response rates, the time burden imposed by telephone interviewing and citation no-shows precluded extension of this approach to all potential participants. In line with our results, another study in Barcelona, Spain, based on invitation for a blood HCV test by ordinary mail, showed an even lower response rate (4.11%: 238 participants out of 5,793 letters sent) [18]. Successful recruitment in this type of population studies probably requires different approaches and better communication strategies, including the use of the media and social networks, especially in younger populations and in urban areas [19].
The performance of the OraQuick HCV rapid test was excellent for detecting active HCV infection. Because the aim of screening is to detect new HCV infected individuals with active viral replication, testing OMT with the OraQuick HCV rapid test seems sufficiently sensitive. There were 21 individuals with a previous anti-HCV-positive laboratory record (9.5%), but the OraQuick HCV rapid test failed to detect two of them. These two "false-negatives" in fact corresponded to spontaneous or treatment-induced HCV clearance, achieved more than 24 months before OraQuick testing. However, it has been reported that the test sometimes fail in OMT samples from individuals with an altered immune response or under immunosuppression. [13,[20][21][22] In these reports, the test sensitivity raised to 94.7% in OMT when reading was taken after 40 minutes, and up to 99.4% when using fingerstick blood. [13,20,21] Several studies, mainly from the USA, suggest that chronic HCV hepatitis is an under-diagnosed disease and that the proportion of undiagnosed individuals is higher than that of those diagnosed with HCV infection. [23] In contrast, in our study only 19 of the 1,206 individuals tested showed a positive OraQuick HCV test, and all were aware of their infection. The population prevalence of anti-HCV antibodies as detected with the OraQuick HCV test was 1.31%, and the prevalence of active HCV-RNA positive infection was only 0.56%, both figures lower than expected according to the Spanish HCV Strategic Plan [estimated figures of 1.7% (CI: 0.4; 2.6) for anti-HCV prevalence and 1.2% (CI: 0.3; 1.8) for active infection prevalence], [6] but in line with recent studies from other regions in Spain on selected populations reporting low anti-HCV estimates in patients scheduled for elective surgery, [24] colorectal cancer screening, [25] or recruited for a seroepidemiological, [26] or hepatitis surveys. [27] Recent studies from other Western European countries also show prevalence rates below 1%. [28,29] The absence of undiagnosed cases in our study (i.e. individuals unaware of their HCV infection) is at odds with studies from the USA, [23] but consistent with some studies in Spain. [24] Although it is possible that individuals with known risk factors for HCV were less likely to participate, our results are probably related to the free, universal, public healthcare Spanish system and the National Hepatitis C Plan implemented in 2011, where free access to new DAA was offered, first to patients with advanced fibrosis and more recently to all HCV infected individuals.
Several limitations of our study may hamper the generalization of our HCV prevalence rates beyond our region. First, our power calculations were based in a 2% HCV prevalence and 50% response rate, but we finally obtained a response rate of only 11%. Although the estimated anti-HCV prevalence was lower than expected (1.31%), our study is potentially underpowered. In addition, we cannot exclude the existence of a population bias due to the dominance of a "relatively low risk" population on the census from our Health Care Department, with no prisons or districts with drug use problems belonging to the assigned healthcare area. Thus, difficult to reach populations such as homeless and PWID are probably underrepresented in our study (i.e. there were only 0.5% of PWID in the overall screened population). Finally, almost ninety percent of the individuals contacted did not respond to the survey and we have no information regarding their baseline characteristics.
Despite these limitations, our study suggests a low prevalence of HCV in the Department District and the relatively advanced age of the anti-HCV positive population, in line with previous studies in Spain. [1,24,30] The age-cohort with most OraQuick HCV positive cases was that of 60-69 years old (42.1%); and 73.7% of the cases occurred in individuals born between 1945 and 1965, with 89.5% of anti-HCV positive patients being born before 1965. In addition, although anti-HCV positive individuals had a higher prevalence of risk factors compared to those who tested negative, no new diagnoses were made through population screening. This finding is relevant because the value of the screening programs lies in the detection of previously unknown active cases, and our study indicates a very low performance of population screening for this purpose. Collectively, our results-and other data from Spain and other European countries-suggests a different epidemiological situation than in other countries that recommend to screen all individuals born between 1945 and 1965 and/or some groups of those considered to be at high risk of HCV infection. [24,[30][31][32][33][34][35][36][37] Perhaps, and in light of the few previously undiagnosed cases detected by this type of screening, the performance of other strategies should be re-evaluated, including "opportunistic screening" or "case finding" schemes, whereby patients from vulnerable populations and patients attending care who have risk factors are actively tested, retrieval policies to re-evaluate VHC patients lost to follow-up and targeting specific subgroups (i.e. methadone clinic attendees).
Supporting information S1