Elizabeth Clark

Screening for Hepatitis C Virus Infection: A Review of the Evidence for the U.S. Preventive Services Task Force

Hepatitis C virus (HCV), the most common chronic bloodborne pathogen in the United States, is acquired primarily by large or repeated percutaneous exposures to blood (1). In the United States, approximately 2.3% of adults 20 years of age or older are positive for anti-HCV antibody. Between 55% and 84% of these have chronic infection (1-6), but only 5% to 50% of infected adults are thought to know their status (7-9). In the United States, HCV is associated with approximately 40% of cases of chronic liver disease and 8000 to 10000 deaths each year (1). Chronic HCV infection can also cause fatigue and decreased quality of life in the absence of cirrhosis or other complications (5, 10, 11). The natural course of chronic HCV infection varies. Some patients never develop histologic evidence of liver disease even after decades of infection (12, 13). In a meta-analysis of community-based cohort studies, 7% of patients with chronic HCV infection developed cirrhosis after 20 years (14). Factors that may be associated with a more progressive course include older age at acquisition (14, 15); comorbid medical conditions, such as heavy alcohol use (14, 16-21), HIV infection (22-24), and other chronic liver disease (25-27); male gender (14); and longer duration of infection. Mode of acquisition, viral load, aminotransferase levels, and viral genotype have not been consistently established as predictors of disease progression (28-31). The effects of ethnicity on the course of HCV infection have not been well studied in the United States (32). In this systematic review, commissioned by the U.S. Preventive Services Task Force (USPSTF), we focus on whether it is useful to test for anti-HCV antibodies in asymptomatic adults who have no history of liver disease. Methods The analytic framework, definitions used, key questions, literature search, and data extraction methods are described in detail in the Appendix. Briefly, relevant studies were identified from searches of MEDLINE (1989 through February 2003) and the Cochrane Clinical Trials Registry (2002, Issue 2) and from the reference list of a recent evidence report commissioned by the National Institutes of Health (33). Reference lists of retrieved articles, periodic hand searches of relevant journals, and suggestions from experts supplemented the electronic searches. Two readers reviewed all English-language abstracts. We selected studies that provided direct evidence on the benefits of screening and studies on risk factors for HCV infection and the performance of third-generation HCV enzyme-linked immunoassay (ELISA) alone or followed by confirmatory recombinant immunoblot assay (RIBA). We focused on third-generation ELISAs because they are thought to be slightly more sensitive than second-generation tests but included data on second-generation ELISAs from large, good-quality observational studies (34). We also selected studies evaluating noninvasive methods to evaluate active HCV infection and the harms associated with biopsy. For treatment, we focused on trials of pegylated interferon with ribavirin but included studies that examined the effect of other interferon-based treatment regimens on long-term clinical outcomes. We also reviewed studies evaluating effects of counseling on high-risk behaviors and benefits of immunizations. Good-quality meta-analyses were reviewed when available. We excluded studies of pregnant patients; children; and patients with occupational exposures, end-stage renal disease, or HIV infection, as well as studies focusing on patients who had already developed complications of chronic HCV infection. We used predefined criteria developed by the USPSTF, described in detail elsewhere (35), to assess the internal validity of included studies, which we rated as good, fair, or poor. We also rated the applicability of each study to the population likely to be identified by screening. We rated the overall body of evidence for each key question using the system developed by the USPSTF (35). This research was funded by the Agency for Healthcare Research and Quality under a contract to support the work of the U.S. Preventive Services Task Force. Agency staff and Task Force members participated in the initial design of the study and reviewed interim analyses and the final manuscript. Additional reports were distributed for review to content experts and revised accordingly (36). Agency approval was required before this manuscript could be submitted for publication, but the authors are solely responsible for the content and the decision to submit it. Data Synthesis Studies of Screening We identified no randomized trials or longitudinal cohort studies comparing outcomes between patients in the general adult population who were screened and not screened for HCV infection. Risk Factor Assessment The identification of risk factors for the presence of HCV infection could aid in the development of selective screening strategies (8). We identified 4 large population-based studies from the United States and Europe that evaluated rates of HCV infection and risk factors associated with HCV status (2, 3, 37, 38). Among these, the National Health and Nutrition Examination Survey III (NHANES III), a good-quality nationwide sample of U.S. households that was conducted from 1988 to 1994 and had 21241 participants, found that the prevalence of positivity for anti-HCV antibodies was 1.8% overall and 2.3% in adults older than 20 years of age (3). Although NHANES III provides the most reliable estimate of prevalence of HCV infection in U.S. households, it probably underestimates the overall prevalence of the disease because it excluded persons without addresses, institutionalized persons, and those in military service. Independent risk factors for HCV infection found in the 4 large population-based studies are shown in Table 1 (2, 3, 37, 38). Intravenous drug use was the strongest independent risk factor in 3 of these studies (adjusted odds ratios, 18.4 to 29.2). The fourth study, NHANES III, did not assess intravenous drug use. However, it found that cocaine and marijuana use were associated with HCV infection, perhaps because they are surrogate markers for intravenous drug use (Table 1) (3). Many other smaller, cross-sectional studies in a variety of specific populations support the strong association between HCV infection and intravenous drug use (39-51). Cross-sectional studies in intravenous drug users have reported prevalence rates ranging from 50% to more than 90% (52-56). Table 1. Data from Large Observational Studies on Independent Risk Factors for Positive Status on Tests for AntiHepatitis C Virus Antibody All 4 large population-based studies also found an independent association between HCV infection and high-risk sexual behaviors (variably defined, but usually considered sex with multiple partners or sex with an HCV-infected person). In most settings with a low prevalence of intravenous drug use, high-risk sexual behaviors are the strongest risk factor for HCV infection (56-60). It is not clear whether this association is due to a high rate of sexual transmission in specific situations (61-66) or because high-risk sexual behaviors are a marker for unacknowledged drug use. Since 1992, transfusions have not been an important mode of HCV transmission (56, 67, 68). There is insufficient evidence to determine the importance of tattoos, body piercings, shared razors, and acupuncture as risk factors (2, 67, 69-75). Nonpercutaneous risk factors such as gender, ethnicity, and socioeconomic status have inconsistent or weak associations with the prevalence of HCV infection (3, 38, 76). In large U.S. cross-sectional studies, between 33% and 81% of patients with HCV infection reported intravenous drug use (38-40, 77). Other retrospective studies have found that 53% to 88% of infected patients had identifiable risk factors (78, 79). Sample differences, varying stringency of risk factor ascertainment, or variation in the risk factors examined could explain some of the discrepancies between studies (79). No study has prospectively applied a selective screening strategy and determined how many patients were correctly identified by it. Accuracy of HCV Antibody Testing The terminology and interpretation of tests used to diagnose HCV infection are reviewed in the Appendix. A recent fair-quality systematic review of third-generation ELISA (7 studies) and RIBA (3 studies) found that only 10 of 150 studies used appropriate methods for evaluating a diagnostic test (80). We applied the USPSTF quality criteria to 9 of these 10 studies and found that all 9 had at least one important flaw: inclusion of a narrow patient spectrum, failure to perform a reference standard test in all samples, or lack of clarity about whether the reference standard test was interpreted independently of the screening test (81-86). The tenth study, a study of RIBA in 51 patients receiving hemodialysis, was not referenced in the systematic review and we could not find it. Of 7 studies that evaluated the sensitivity of third-generation ELISA and involved 4674 patients, sensitivity ranged from 97.2% to 100% compared with the results of polymerase chain reaction (PCR) (a reference standard for active infection) or RIBA (a reference standard for exposure). We identified 3 additional studies of the sensitivity of third-generation ELISA using PCR as the reference standard (Table 2) (87-89). One of these was a good-quality study that found a sensitivity of 94% (107 of 114) (87). The specificity was 97% (946 of 976), and the positive predictive value (prevalence, 10%) was 78% (107 of 137). Second-generation ELISAs are thought to be slightly less sensitive than third-generation tests but may be more specific (34). In data collected by the Centers for Disease Control and Prevention in 24012 lower-prevalence (2%) patients, the positive predictive value of current second- and third-generation ELISAs without confirmatory RIBA was 42%