NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.
Chou R, Clark E, Helfand M. Screening for Hepatitis C Virus Infection [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2004 Mar. (Systematic Evidence Reviews, No. 24.)
This publication is provided for historical reference only and the information may be out of date.
Arrow 1: Does screening for hepatitis C reduce the risk or rate of harm and premature death and disability?
We identified no randomized trials or observational studies comparing outcomes between patients in the general adult population or high-risk subpopulations screened and not screened for hepatitis C. Although blood donors are routinely screened for HCV infection in order to insure the safety of transfused blood products, we also searched for studies demonstrating improvement in clinical outcomes in volunteer blood donors screened versus those not screened, and found none.
Arrow 2: Can clinical or demographic characteristics identify a subgroup of asymptomatic patients at higher risk for HCV infection?
The identification of risk factors for HCV infection could aid in the development of selective screening strategies. This approach might save resources compared to universal screening because efforts are focused on individuals who would be most likely to benefit from testing and treatment.16 The most frequently cited risk factors for HCV infection are intravenous drug use and high-risk sexual behaviors (variably defined, but usually considered sex with multiple partners or sex with someone with HCV).8, 10, 69, 109, 110
Data regarding risk factors for HCV infection in the United States are available from two large cross-sectional studies that performed multivariate analyses to assess associations between various risk factors and anti-HCV-Ab status. One study, NHANES III (n=21,241), was a nationwide household-based sample conducted from 1988 to 1994 in which the prevalence of anti-HCV-Ab positivity was 1.8% overall, and 2.3% in adults >20 years old.9 Among adults 20 to 49 years old, the prevalence in this study was 4.1% for males and 1.6% for females. The National Hepatitis Screening Survey (NHSS) (n=13,997) was a nationwide screening program in 40 mostly urban centers in September 1992 that found a higher prevalence of 7.0%, but was targeted at persons who were at risk for HCV.111 Data are also available from two large European population-based studies, where the epidemiology of HCV may differ slightly compared to the US. The Dionysos (Italy) study (n=6,917) found rates of 3.2% for anti-HCV antibody and 2.3% for viremia,112 and a French study (n=6,288) found rates of 1.2% and 0.9%. 18 All studies were performed prior to 1994, when many patients with transfusion-associated HCV infection were being identified.
Independent risk factors for HCV infection found in large population-based studies are shown in Table 4. Studies varied in the risk factors assessed. In all studies that assessed intravenous drug use, this was the strongest independent risk factor (adjusted odds ratios [OR], 18.4–29.2). Most studies found an independent association between some high-risk sexual behaviors and HCV infection. Other factors (gender, race, socioeconomic status) showed less consistent or weaker associations. The European studies found similar independent risk factors as the American studies.18, 112
In NHANES III, the strongest independent predictors of HCV positivity were illegal drug use (cocaine and marijuana use) and high-risk sexual behaviors in the absence of illegal drug use. Weaker independent predictors of HCV positivity were marital status, income (above or below the poverty level), and education. Race, ethnic group, and gender did not have significant independent associations with infection. Because NHANES III did not ask specific questions about intravenous drug use, it was postulated that cocaine and marijuana use were surrogate markers for this behavior.
In the National Hepatitis Screening Survey, the strongest independent risk factor for anti-HCV-Ab positivity was intravenous drug use (adjusted OR, 23.34), followed by hemodialysis, sex with an intravenous drug user, a history of blood transfusion, and male gender (Table 4). White or Hispanic race was associated with a decreased risk (adjusted OR, 0.57) for HCV infection. History of sex with multiple partners and age were not associated with HCV infection.
The strong association between HCV infection and intravenous drug use is supported by findings from numerous smaller cross-sectional studies. Intravenous drug use was the strongest risk factor for HCV infection in a variety of settings including US veterans,113, 114 incarcerated persons,115, 116 street youths,117 sexual partners of patients with HCV,118, 119 pregnant women,120 blood donors,121, 122 and patients attending sexually transmitted diseases clinics.123–125 Cross-sectional studies in intravenous drug users have reported prevalence rates ranging from 50 to over 90%.10–13, 126 In addition to being highly prevalent, HCV infection may be associated with rapid spread in intravenous drug users, though estimates of incidence have been higher in cross-sectional than in longitudinal studies. In the US, a cross-sectional study of drug users who reported injecting for 1 year or less found a prevalence of anti-HCV antibodies of 65%,127 but a longitudinal study12 found a 30% rate of seroconversion after a median of 6.5 years. Differences in drug use behaviors could account for some of the variation in incidence and prevalence; one study found that the rate of anti-HCV positivity was higher among intravenous drug users who shared needles (53%) than in all intravenous drug users (38%).123
Sexual contact with infected patients appears to be a less efficient mode of transmitting HCV than intravenous drug use, and the relative importance of sexual transmission remains controversial.109 Despite numerous conflicting studies118, 119, 128–142 about the precise role of sexual transmission, however, it is clear that after intravenous drug use, high-risk sexual behaviors are the most important risk marker for HCV infection. This is probably related to either a high rate of sexual transmission in specific situations (high viral load, specific genotype, presence of other sexually transmitted diseases, or other local factors)128–132, 143 or because high-risk sexual behaviors are a marker for unacknowledged drug use or other risk factors. The rate of unacknowledged drug use is probably higher if risk factor ascertainment is less stringent or if there is a strong stigma against this behavior in the population studied.109
In addition to large population-based studies9, 111, 112 that found independent associations between HCV infection and some high-risk sexual behaviors (Table 4), a widely quoted CDC case-control study in patients who had not had blood transfusion or admitted intravenous drug use within the preceding 6 months found that multiple sexual partners and history of hepatitis in household or sexual contacts were independent risk factors for acute NANB hepatitis.144 In a follow-up CDC study of populations from the same sentinel county surveillance project, 15% of acute HCV cases between 1991 and 1995 reported high-risk sexual practices or exposures (sex with patient with HCV infection or sex with multiple partners) during the 6 months prior to illness in the absence of intravenous drug use; of these 2/3 had a partner positive for HCV infection.10 In most settings with a low prevalence of acknowledged intravenous drug use, high-risk sexual behaviors are the strongest risk factor for HCV infection.145–147 The risk of sexual transmission of HCV from an infected individual in the US has been estimated at 0% to 0.6% per year for those in monogamous relationships, and 1% per year for those with multiple sexual partners.143
Other risk factors for HCV infection can be divided into percutaneous and non-percutaneous risk factors. Non-percutaneous risk factors are not associated with a plausible mechanism of transmission and are probably markers for unsolicited high-risk behaviors or high background prevalence of disease in the specified setting. Studies have not consistently shown independent associations between these risk factors and HCV infection. These risk factors include low socioeconomic status, non-injection drug use, alcohol use, male gender, age 40–59, and non-white race.9, 111, 148 The higher prevalence of HCV infection in incarcerated, emergency room, urban, and veteran populations also appears related to the presence of other risk factors, particularly injection drug use.113, 116, 149
Percutaneous risk factors other than illegal drug use and high-risk sexual behaviors that are plausible modes of HCV transmission in the general population include transfusions, tattoos, body piercing, and sharing razors.150 Recently, the relative importance of transfusions has declined, and since 1990, transfusions have not been an important mode of HCV transmission.10, 151 There is insufficient evidence to determine whether tattoos, body piercing, or sharing razors are risk factors for HCV infection.152 Some observational studies153, 154 in selected populations in the US have found that tattoos are an independent risk factor for HCV infection, while others136, 155, 156 have not. CDC data indicate that less than 1% of patients with newly acquired HCV report recent tattoos; a large CDC study is currently underway to further investigate the association.157 Data on other percutaneous exposures in the US are scant. A study in volunteer blood donors found an association between ear piercing among men (but not women), but this has not been confirmed in other studies.136 A large French population-based study showed no association between acupuncture and HCV infection.18
Other populations with a higher prevalence of HCV infection in which percutaneous exposures may be an important mode of transmission include hemodialysis patients, children of mothers with HCV infection, patients with percutaneous occupational exposures, post-transplantation patients, and HIV-positive patients.8, 158–163 These populations were outside the scope of this review.
We identified no study that applied a selective screening strategy in the general population and measured what proportion of infected patients was identified correctly. Studies that reported the proportion of HCV patients with identified risk factors have not prospectively validated their findings.120, 164, 165 Cross-sectional studies from several settings, including CDC sentinel county surveillance data on acute hepatitis, the National Hepatitis Screening Survey, a population-based study in Olmstead County, and US Veterans Affairs screening programs, have found that between 33 and 81% of patients with HCV reported intravenous drug use.9, 111, 113, 114, 166 CDC data indicate that in over 50% of cases, acute HCV infection was associated with injection drug use <6 months (43%) or >6 months (16%) prior to onset, an additional 5% were associated with snorting, and another 15–25% reported high-risk sexual behaviors in the absence of injection drug use.10 The epidemiology of acute HCV infection, however, may differ from the epidemiology of chronic HCV. In a post-hoc analysis of data from the National Hepatitis Screening Survey, which identified patients with chronic HCV, screening patients using one of three different risk factor models would have identified between 53%–69% of patients with HCV infection.165 Another study found that only 12% of HCV-positive individuals had no admitted risk factor.164 Factors that may explain some of the discrepancies between studies regarding rates of patients with no identified risk factors include population differences, varying stringency of risk factor ascertainment, or broadening what is included as an HCV risk factor (e.g., tattoos, broader sexual practices, or acupuncture).164
Arrow 3: What are the test characteristics of HCV antibody testing?
A recent fair-quality systematic review of third-generation ELISA and RIBA found that only 10 of 150 studies used appropriate methods for evaluating a diagnostic test167 We applied the USPSTF quality criteria to nine of these 10 studies; all nine had at least one of the following flaws: narrow patient spectrum, reference standard not performed in all samples, or unclear if reference standard interpreted independently of screening test.161, 168–172 The tenth, a study of RIBA in 51 patients on hemodialysis, was not referenced in the systematic review and we were unable to find it.
Of these 10 studies, seven evaluated sensitivity of third-generation ELISA (n=4,674) and 3 evaluated third-generation RIBA (n=359). The seven studies on ELISA were performed in blood donors, patients with chronic liver disease, hemodialyzed patients, or from panels of sera. When compared with the results of PCR (four studies) or RIBA (three studies), the sensitivity of third-generation ELISA ranged from 97.2% to 100%. PCR is a reference standard for chronic infection and RIBA is a reference standard for exposure. When compared with the results of PCR (two studies), the sensitivity of third-generation RIBA was 80% and 100%, and when compared with the results of ELISA (one study), the sensitivity was 100%.
We identified three studies (one good-quality) not included in the systematic review that also evaluated the sensitivity of third-generation ELISA using PCR as the reference standard.173–175 All of these studies reported similar findings as the systematic review, with sensitivities of 94–100% for third-generation ELISA compared to PCR. The good-quality study performed third-generation ELISA and PCR tests in 1,090 patients admitted for acute liver disease or suspected chronic hepatitis, and reported the lowest sensitivity at 94% (107/114).173 One of the fair-quality studies evaluated the false-negative rate of third-generation ELISA in blood donors with elevated transaminases (thought to be higher risk for HCV infection) and found that no patients (0/247) had hepatitis C viremia on subsequent testing.175
We found extremely limited data regarding specificity of third-generation ELISA. In the systematic review cited above, specificity of third-generation ELISA was evaluated in three studies using PCR as the reference standard and one study using a combination of second-generation ELISA and RIBA as the reference standard. Because approximately 15–45% of patients who are ELISA- and RIBA-positive do not have evidence of viremia, we expected to find lower specificity in studies using PCR as the reference standard, unless the prevalence of HCV infection was very low in the population tested. In the systematic review, however, the specificity was 100% for both types of studies (84/84 using PCR and 2,099/2,099 using second-generation ELISA/RIBA). In the good-quality study that was not included in the systematic review, specificity of ELISA was 97% (946/976) using PCR as the reference standard.173
Arrow 4: What is the false-positive rate and what are the harms associated with screening for hepatitis C virus?
In ione good-quality study of 1,090 patients admitted to the hospital for acute liver disease or suspected chronic hepatitis (prevalence of hepatitis C viremia 10%), the false positive rate of an ELISA test (without confirmatory RIBA) was 3% when PCR was used as the reference standard. The positive predictive value of third-generation ELISA (without confirmatory RIBA) was 78% (107/137).173 Similar data from good-quality studies are not available for lower-prevalence populations, though for second-generation ELISA positive predictive values of 50–61% have been reported.60
Data on positive predictive values for viremia of a positive confirmatory third-generation RIBA following a positive ELISA test are available from two large population-based cross-sectional studies from France18 and Italy, which may include a more applicable spectrum of patients and prevalence (1.2–3.2%) than the study cited above.112 In these studies a positive confirmatory RIBA test was associated with a likelihood of viremia of 84–86%.
Some patients who have discrepant ELISA and PCR results might be “biological false positives” who have cleared their infection, while others are “analytic” false positives that, in the absence of past infection, have a positive ELISA due to cross-reactivity or another factor that reduces the specificity of the assay. Although all patients receive the “gold standard” test for active HCV infection before they are subjected to invasive tests (i.e., biopsy) or potentially harmful antiviral treatments, false-positive results on third-generation ELISA can result in other harms (labeling, anxiety, effects on close relationships) that are difficult to measure. There are few data regarding potential harms in patients who have false-positive tests or in patients with true-positive tests that are ineligible for or decline treatment. In one small fair-quality cross-sectional report (n=34) of intravenous drug users with chronic HCV infection, those who were aware of their HCV status reported worse quality of life compared to those who were not aware of their status.37 Another small (n=34) controlled trial, published only in abstract form, found that a brief counseling program helped improve sense of well-being in women diagnosed with HCV.176 We found no other studies investigating whether harms associated with HCV screening could be reduced by effective patient education and counseling.69, 177
Arrow 5a: What are the test characteristics of the work-up for treatable disease?
A general work up for treatable chronic HCV disease includes PCR testing for viremia, transaminase levels, and liver biopsy. PCR testing is currently the only readily available method for determining active HCV infection in clinical practice. Pretreatment liver biopsy is currently recommended by the National Institutes of Health.178–180 Several other blood tests, however, have been proposed as less invasive methods of predicting biopsy findings.
A good-quality systematic review recently evaluated studies of blood tests used to predict biopsy results.181 None of these blood tests appeared to predict liver biopsy findings accurately. Transaminase levels are already routinely ordered in patients with HCV infection and 15 studies evaluated their ability to predict biopsy findings. Elevated serum alanine aminotransferase (ALT) was associated with fibrosis in 11/15 studies, and had a sensitivity ranging from 61 to 76% and specificity from 44 to 66%. The AST/aspartate aminotransferate (ALT) ratio, a combination of two transaminase tests, had a sensitivity ranging from 31 to 56% and specificity of 90 to 100% for predicting cirrhosis. The systematic review found little information on the role of the AST/ALT ratio in predicting noncirrhotic stages of fibrosis. The systematic review also evaluated the ability of extracellular matrix tests, cytokines, and panels of five or more blood tests to predict the results of liver biopsy. These tests are not currently part of a standard work-up for treatable chronic HCV infection. The review found that extracellular matrix tests (hyaluronic acid and laminin) may have value in predicting fibrosis on liver biopsy, that cytokines have less value than the extracellular matrix tests, and that panels of tests may have the greatest value in predicting cirrhosis and the absence of more than minimal fibrosis. None of the tests were able to consistently classify intermediate stages of fibrosis.
We identified no study evaluating outcomes in patients receiving antiviral therapy who had pretreatment biopsy compared to those who did not undergo biopsy.
Arrow 5b: In patients found to be positive for hepatitis C antibody, what proportion of patients would qualify for antiviral treatment?
Antiviral therapy is recommended for patients with chronic HCV who are at the greatest risk for progression to cirrhosis. These persons have detectable HCV RNA, often have persistently elevated transaminase levels, and have liver biopsy findings showing at least portal or bridging fibrosis or at least moderate degrees of inflammation and necrosis.8, 68, 182–185 The proportion of patients identified by a screening program who would have clear indications for treatment based on biopsy findings is unclear but is probably lower than would be seen in patients sent to a referral center for treatment. Large-population-based studies have not systematically biopsied patients with chronic HCV infection, but in two cohort studies that included patients with asymptomatic disease, the number of patients with more than minimal inflammation on biopsy (HAI score >4) was 39–56%.32, 33
Treatment is not appropriate for all patients with chronic HCV infection and evidence of more than minimal liver damage. Patients with decompensated cirrhosis are not treated with interferon-based therapies due to the risk of disease exacerbation. These patients, however, are likely to be identified without screening. Although patients with compensated cirrhosis can be considered for treatment, response rates might be lower than in patients without cirrhosis.8, 69 Patients with active alcohol abuse or ongoing illicit drug use are usually not treated until these behaviors have been discontinued for 6 or more months, although new recommendations suggest treatment may be offered in conjunction with substance abuse therapy.8, 182 In addition, patients with serious medical or psychological comorbidities may not be appropriate for treatment due to the serious side effects of current therapies.
In clinical practice, the number of patients who are referred for evaluation and receive treatment will depend on the degree of liver damage, the presence of serious co-morbid conditions or contraindications to treatment, and patient preferences regarding treatment. The clinical trials that we reviewed provided insufficient information about the numbers of persons with anti-HCV antibodies screened, eligible, and enrolled, and the reasons persons were ineligible for treatment. We identified four observational studies that evaluated the number of patients referred for HCV infection who were eligible for and received treatment.186–189 One study was excluded from further review because it did not define reasons for ineligibility or report numbers of patients ineligible for different reasons.189 This study reported high rates of loss to follow-up (39%) after initial evaluation and in those patients who did follow up, a high proportion of patients declined treatment (42%), though reasons were not specified.189 Of the 3 included studies, 2 (n=557 and 100) were performed in Veterans Affairs centers186, 187 and one in a general liver teaching clinic (n=327).188 Because all studies were in patients referred for treatment, these studies may overestimate the proportion of treatment-eligible patients who would be identified by screening. Two of the studies included a small proportion (10%) of patients without viremia.187, 188 Patients were ineligible for treatment if they had psychiatric disease, active alcohol or drug abuse, or significant medical comorbidities. Other reasons for ineligibility varied between studies but included the presence of normal transaminases or nonadherence to the evaluation protocol. These studies generally considered patients with milder forms of liver disease ineligible for treatment.
In all three included observational studies, approximately 30–40% of evaluated patients were eligible for therapy and received treatment. Demographic characteristics did not predict ineligibility in one study.187 Common reasons for ineligibility were ongoing substance abuse (13–44%) and serious co-morbid medical or psychiatric conditions (12–34%). Non-compliance with the evaluation protocol was the most important reason for ineligibility in one study (37%).188 Decompensated cirrhosis was an uncommon (1%) reason for ineligibility. About 10% of patients declined therapy in one study.188 Findings appeared similar in the VA and non-VA studies.
In patients with chronic HCV infection who have minimal or no biopsy abnormalities, indications for treatment are not clear, and the decision for therapy is individualized.68, 69, 150, 190 Many patients identified by screening are likely to be in this category. Information on the proportion of HCV patients with milder liver disease that might be identified by screening is available from community-based cohort studies and population-based cross-sectional studies. In three community-based cohort studies, the rate of chronic hepatitis of minimal grade or no inflammation (HAI score 0–3 or 0–4) was 43–61%.32, 33, 66 In population-based studies, between one-third to one-half of patients with hepatitis C viremia had normal transaminases in population-based studies.18, 111, 112 Biopsies were not performed systematically in these studies, but a review of 15 smaller cross-sectional studies of liver biopsies in patients with normal transaminases found 20% had a normal liver, 4% minimal chronic hepatitis, 52% mild chronic hepatitis, 23% moderate-severe chronic hepatitis, and 1.3% cirrhosis.67
Arrow 6: What are the harms associated with the work-up for active HCV disease?
In the work-up of patients with chronic HCV infection, percutaneous liver biopsy is the most invasive diagnostic procedure and associated with the highest risk of complications. The remainder of the work-up is non-invasive and does not appear associated with major complications, although other difficult-to-measure harms such as anxiety probably occur. Percutaneous liver biopsy can be performed “blind” or using ultrasound guidance, and is generally an outpatient procedure with post-biopsy observation for 4–6 hours.191 Patients who are at high risk for bleeding are generally excluded from the procedure.192 The most common complication of percutaneous liver biopsy is pain, with around 30% of patients requiring strong analgesic medications following the procedure.192 More serious but less common complications include bleeding, the most frequent major complication, and biliary rupture, intestinal perforation, vasovagal hypotension, or infection. There are no standardized definitions for major complications from percutaneous liver biopsy.
Most data on risks of percutaneous liver biopsy come from large series of patients undergoing liver biopsy for a variety of reasons.191, 193–197 Recent smaller studies have evaluated the risks of percutaneous liver biopsy specifically in patients with HCV infection, including suspected cirrhosis.198, 199 No study on percutaneous liver biopsies specifically examines asymptomatic patients with chronic HCV, who may be at lower risk for complications. Clinical trials on treatment of HCV have typically randomized patients after percutaneous biopsy and do not report biopsy complication rates.
Of the studies examining large series of patients biopsied for a variety of indications, the study which appears to be of highest quality evaluated consecutive percutaneous liver biopsies in a nationwide sample of the United Kingdom using a standard form to assess for major complications.195 This was the only study that used independent assessment to ascertain complications. A bleeding rate of 26/1,500 (1.7%) was found, with 11/1,500 ( 0.7%) requiring transfusion. Death was definitely associated with liver biopsy in 2/1,500 patients and possibly in another three, giving a mortality rate of 0.13–0.33%. This report may overestimate the risk of percutaneous biopsy in asymptomatic patients with chronic HCV, because the indication in over one third of the patients was primary or secondary malignancy, conditions that are associated with substantially higher complication rates.197 The rates of major complications reported in this study are comparable to other large series, which reported rates of 0–3.7%.191, 194, 200 Mortality rates in other large series have typically been <0.1%.
Two small studies (n=126 and 166) have reported complication rates from percutaneous biopsy in patients specifically with HCV infection.198, 199 In both studies, no episodes of bleeding, perforation, or death were reported. Both studies included patients with known or suspected cirrhosis.
Small studies and trials suggest that ultrasound-guided biopsies may be associated with fewer complications than blind biopsies, including bleeding and pain.199–203 Several studies have found that increased experience of the person performing the liver biopsy was associated with less risk of complications.
Arrow 7a: How well does antiviral treatment reduce the rate of viremia, improve transaminase levels, and improve histology?
Important clinical outcomes of treatment for chronic HCV infection include reduction in mortality, cirrhosis, and hepatocellular carcinoma, and improvement in quality of life. Because of the large numbers of patients and long duration required to demonstrate improvements in most of these outcomes, intermediate outcomes have been the most common measure of treatment benefit. Measured intermediate outcomes include normalization of liver enzymes, loss of detectable viremia, and improvement in liver biopsy findings. Sustained virologic response rates are currently considered the best indication of successful treatment.
Antiviral treatment for non-A, non-B hepatitis began in 1986 with the use of interferon-alfa.204 A meta-analysis of trials of interferon monotherapy versus placebo found a summary sustained biochemical response rate of 21% (95% CI, 13–28%) with 12-month courses of treatment versus 2% in controls.205 Other meta-analyses have reported sustained virological response rates of 6–16% for interferon monotherapy.206–208 This treatment was considered the standard of care for chronic HCV infection until 1998, when combination interferon plus ribavirin was approved as first line treatment.8, 209 Treatment with pegylated interferon, alone or in combination with ribavirin has been used for only a few years.182
Three recent good-quality systematic reviews evaluated the effect of newer treatment regimens on intermediate outcomes.206–208 One207 evaluated the data for pegylated interferon alone or with ribavirin; all 3 evaluated data for non-pegylated interferon plus ribavirin.
The systematic review of the pegylated interferon literature identified three randomized controlled trials (RCTs) that compared pegylated interferon plus ribavirin to pegylated interferon alone. The review concluded there was grade A evidence that pegylated interferon plus ribavirin was superior to pegylated interferon alone for intermediate outcomes.207 The review also concluded that there was grade A evidence from four RCTs that pegylated interferon alone was superior to regular interferon alone for intermediate outcomes.207
We independently reviewed trials of pegylated interferon plus ribavirin and for pegylated interferon monotherapy. One RCT only available as an abstract when the systematic review was written has since been published with full data.210 The 3 RCTs are summarized in Evidence Table 1a and 1b. Two210, 211 were large (n=1121 and 1530), multi-center, good quality RCTs and the other was a small (n=72), fair quality, dose-finding study.212 All three studies compared pegylated interferon plus ribavirin to pegylated interferon alone. One RCT also compared pegylated interferon plus ribavirin to non-pegylated interferon plus ribavirin.210 Treatment was for 24 to 48 weeks with a 24-week follow-up period. Patients with significant other medical or psychiatric problems, other sources of liver disease, HIV, end-stage liver disease, or active substance abuse were excluded.
The two good-quality trials found that 54% to 56% of all patients achieved a SVR with pegylated interferon plus ribavirin versus 44% to 47% with pegylated interferon alone (p<=0.01) using comparable dosage regimens.210, 211 Patients with HCV genotype 1 responded more poorly (42% to 46% SVR) compared to genotypes 2 and 3 (76% to 82% SVR). In the fair-quality study, combination pegylated interferon plus ribavirin achieved a 60% SVR versus 42% with pegylated interferon alone (significance not reported), in the highest dose treatment groups.212 Histologic outcomes were not evaluated in one trial;210 in the other RCTs, all treatment groups demonstrated improvement in inflammation and fibrosis, but there was no significant difference between treatment arms in histologic outcomes.211, 212
One RCT also compared pegylated interferon plus ribavirin to non-pegylated interferon plus ribavirin.210 In this trial, 56% of patients achieved a SVR with pegylated interferon plus ribavirin versus 44% of patients with non-pegylated interferon plus ribavirin (p<0.001).
We identified the same 4 RCTs evaluating the efficacy of pegylated interferon versus non-pegylated interferon monotherapy in treatment naïve patients as the systematic review.213–216 Data from these studies are summarized in Evidence Tables 1a and 1b. Treatment was for 48 weeks with a 24-week follow-up period in all of these studies. Inclusion and exclusion criteria were similar to the combination pegylated interferon and ribavirin studies. Each of these studies found that a greater percentage of patients achieved a SVR when treated with pegylated interferon monotherapy. The percentage of patients responding ranged from 25%214to 39%.216 All of the studies also evaluated histologic responses to treatment. Between 44% and 70% of patients had histologic improvement with pegylated interferon monotherapy, but the difference between treatments was significant in only one of the studies.213
The three systematic reviews that evaluated interferon plus ribavirin in treatment-naïve chronic HCV patients are summarized in Evidence Table 2.206–208 The most recent systematic review contained trials published through March 2002.207 Each systematic review concluded that there was good evidence that there is an increased response rate for interferon-alfa plus ribavirin versus interferon alone. In one review, the relative risk of SVR was 0.75 (95% CI, 0.65–0.79) with interferon monotherapy compared to interferon plus ribavirin.206 Another systematic review found a SVR of 33% for patients treated for 24 weeks and 41% for 48 weeks with combination therapy compared to 6% and 16% for patients with monotherapy (p<0.05).208 The summary relative risk of achieving a SVR was calculated to be 4.90 (95% CI, 2.63–9.13).208
Table 5 summarizes the relative effects of each of the reviewed treatments. Newer combination treatments with higher-dose pegylated interferon plus ribavirin appeared superior to older combination treatments in two good-quality randomized trials, with a sustained response rate of 54% to 56%, compared to a historical sustained response rate of 2% in untreated patients in placebo-controlled trials. The number needed to treat with newer combination treatments to achieve one sustained response appears to be a little under 2 patients.
Treatment studies may not be directly applicable to the population that would be identified by screening because they evaluated patients referred for treatment who probably had more serious disease. In studies evaluating pegylated interferon, for example, 6210, 211, 213–216 out of 7212 used elevated transaminases as an inclusion criterion. It is difficult to further assess the applicability of treatment studies to the screening population because biopsy findings and numbers screened were not consistently reported. In trials that reported these data, mean HAI and fibrosis scores ranged from 1.5–13.4 and 1.3–5.0, and were generally higher than biopsy scores reported in cohorts that included asymptomatic patients.107 Even if treatment is equally effective for virologic endpoints in patients identified by screening and those studied in clinical trials, the overall clinical benefit would be expected to be smaller since the underlying progression rate is lower.
Arrow 7b: How well does antiviral treatment improve health outcomes in asymptomatic patients with hepatitis C?
Several factors complicate our ability to assess the long-term benefits of treatment of HCV, including the long duration for important complications to develop and the relatively short time period that treatments have been available. There is no data showing long-term benefits after treatment with pegylated interferon, alone or in combination with ribavirin, due to the relatively recent introduction of pegylated interferon.
One recent good quality systematic review evaluated the long-term effects of non-pegylated interferon alone on chronic HCV. This review identified three RCTs and 14 cohort studies.207 An additional 23 studies were also reviewed that assessed the natural history of HCV infection in persons without treatment. The studies reviewed were heterogeneous in design and had some methodological limitations (selection bias, variable lengths of follow-up, variable baseline rates of cirrhosis, little description of the untreated population, and variable alcohol consumption). Nonetheless, included studies were consistent in suggesting that treatment with standard interferon-based therapy produced a moderate decrease in the risk of hepatocellular cancer and cirrhosis in complete responders (persons with both a SBR and SVR). The review also concluded that there was some evidence that persons with a SBR also had a decreased risk of HCC and progression of liver disease compared to untreated patients. The data were inconsistent regarding the long-term impact of treatment on non-responders and relapsers: one long-term randomized controlled trial suggested that all patients treated with standard interferon, regardless of response, derived long-term benefit; other studies suggested that relapsers but not nonresponders may derive some long-term benefit from standard interferon therapy.
We independently reviewed the three RCTs that provided long-term outcomes data after treatment with interferon-alfa monotherapy. These studies are summarized in Evidence Tables 3a and 3b. In an unblinded, good-quality Japanese trial of 90 patients randomized to interferon-alfa for 24 weeks or to symptomatic treatment, after 8.7 years there was a significant long-term reduction in the rate of developing HCC (27% vs. 73%, p<0.0001) and in mortality (11% vs. 58%, p<0.001).217 Cirrhosis was also decreased in the treatment group. Rates of cirrhosis, HCC, and mortality were much higher in this study than in studies in US and European populations. The relative risk of progressing to Child B cirrhosis was 0.230 (0.120–0.505) in the treatment versus the control groups.217
In the two other RCTs, both from Italy, 61 and 149 patients were followed for up to 5 to 6 years after randomization to treatment with either interferon-alfa or interferon-beta.218, 219 No significant differences in long-term outcomes were seen, but one study evaluated interferon-beta219 which has since been found to be less effective than interferon-alfa,220, 221 and the other compared two doses of interferon-alfa to each other rather than to placebo.218
Long-term outcomes data for combination interferon plus ribavirin versus interferon monotherapy are not yet available. Available trials are characterized by a short duration of follow-up in a relatively young study population without cirrhosis or alcohol use. As a result, very few serious morbidities or deaths have occurred. The best data at this time come from a good quality systematic review which found that out of 6,585 patients in 48 trials, only patients patients on combination therapy and 12 on interferon monotherapy had developed histological cirrhosis (none developed clinical cirrhosis); in addition only one case of hepatocellular cancer had occurred, three deaths (one suicide and two accidental), and no transplants.206 There was no significant difference in liver related morbidity plus all-cause mortality after treatment with combination therapy as compared to interferon (Peto odds ratio 0.45; 95% CI, 0.19–1.06, Peto odds ratio 0.29; 95% CI, 0.04–2.10, respectively).
Quality of life outcomes have been evaluated in one randomized controlled trial of treatment-naïve patients that analyzed results according to treatment received.222 Other clinical trials evaluating quality of life outcomes stratified results according to whether sustained virologic response occurred and are reviewed later (Arrow 9). In the trial that evaluated quality of life outcomes according to randomization group, 160 patients randomized to interferon monotherapy (n=106) or placebo (n=54) for 24 weeks were evaluated with the Sickness Impact Profile (SIP) at baseline, at the end of treatment, and 24 weeks post-treatment.222 This study was rated poor-quality because SIP results were only available at the end of treatment for 53/106 patients randomized to interferon, baseline SIP scores appeared significantly different between treated and untreated patients, and it was unclear whether patients were blinded to viral or biochemical markers of response to treatment. Patients randomized to interferon had no significant change in total SIP score compared to baseline but did have significant improvements in 3 out of 7 domains.
The single RCT showing significantly improved long-term outcomes after interferon monotherapy was conducted in Japan.217 Many of the cohort studies evaluating long-term outcomes have also been performed in Japan.107 There is evidence that severe complications from HCV infection including cirrhosis and hepatocellular cancer are substantially more common in Japan.223 Though the reason for this is unclear, it may be related to differences in genotype, longer duration that HCV infection has been present in the country, or co-existing liver disease, including chronic hepatitis B infection. The results of these studies may not be applicable to populations screened in the United States and Europe. Cohort studies that include patients with asymptomatic infection in these settings have shown lower rates of HCC and mortality. In one study in the UK, the rate of HCC was 0.1% and the mortality rate 14% after 10 years, compared to 73% and 58% after 8.7 years in the Japanese RCT reporting improved long-term outcomes.28, 217 In a US study the mortality rate was 41% after 45 years.29
Arrow 7c: How well do counseling and immunizations in asymptomatic patients with hepatitis C improve clinical outcomes or prevent spread of disease?
Asymptomatic patients with HCV infection identified through a screening program could benefit from interventions other than antiviral treatment, including counseling to avoid excess alcohol and to obtain immunizations for hepatitis A (HAV) or hepatitis B (HBV).177 In addition, counseling patients on avoiding high-risk behaviors might help prevent spread of HCV and result in significant public health benefits.152 All patients with viremia could benefit from these interventions.
We identified no studies evaluating the effect of counseling regarding HAV vaccinations after diagnosis of HCV on subsequent clinical outcomes. In a US study, hepatitis A vaccination resulted in a satisfactory immune response after a 2-dose regimen in >90% of patients with HCV.224 Reactions to hepatitis A vaccinations appeared mild and self-limited, and no serious adverse events were reported. The incidence of fulminant co-infection with HAV and HCV in the US is difficult to estimate. In a widely-publicized Italian study, 17/432 patients with HCV acquired HAV over a 7-year period, with 7/17 (43%) having a fulminant course and 6/17 (33%) dying.225 Other studies, including other Italian studies from the same time period, have reported much lower rates of fulminant hepatitis and death from HAV coinfection, and the discrepancy remains unexplained.226, 227 In the US, the CDC reported that between 1983 and 1988, 107 fatalities from HAV occurred in 2,311 patients (4.6%) with underlying chronic liver disease from different conditions, compared to 247/113,009 (0.2%) without chronic liver disease.224 It is not clear how many of the deaths were associated with HCV infection, nor how many patients with HCV infection developed fulminant HAV coinfection while unaware of their status. We did not review several cost-effective analyses of HAV vaccination in patients with HCV.228–230
We identified no studies evaluating the effect of counseling regarding HBV vaccination after diagnosis of HCV on subsequent clinical outcomes. HBV and HCV coinfection is associated with more progressive disease and worse outcomes than HCV infection alone. In a widely publicized US study, HBV vaccination resulted in high rates of protective seroconversion (up to 100%) in patients with HCV infection, and appeared well-tolerated.224 Because HBV and HCV are transmitted by similar routes, some patients diagnosed with HCV will have already been exposed to HBV. We did not identify studies estimating how many patients with HCV acquired HBV after becoming aware of their status.
We identified no studies evaluating the effect of counseling regarding alcohol consumption after diagnosis of HCV on subsequent clinical outcomes. We identified one French observational study that retrospectively asked patients to compare their use of alcohol prior to diagnosis with HCV to current use.231 This study found that out of 25 patients who reported ‘excessive’ alcohol consumption prior to HCV diagnosis, 9/25 had become completely abstinent and 14/25 had cut back to ‘moderate’ intake. The results of this study may have been affected by recall bias or unwillingness to admit to current heavy alcohol use.
We identified one US study evaluating changes in behavior after diagnosis of HCV in young injection drug users.232 In this study, there were no significant differences between patients aware of their HCV status and intravenous drug users without HCV or unaware of their positive status with regard to high-risk behaviors including sharing needles and syringes. We identified no other studies evaluating changes in high-risk behaviors following diagnosis of HCV infection, nor any studies that estimated the risk of spreading HCV in patients aware of their status compared to patients unaware of their status.
Although simply being diagnosed with HCV could alter behaviors without any specific intervention, we identified no studies evaluating differences in rates of serious hepatitis A or B coinfection, complications from excess alcohol use, or decreased spread of infection between patients with HCV aware of their status compared to those not aware.
Arrow 8: What are the harms (including intolerance to treatment) associated with antiviral treatment?
Interferon monotherapy is associated with significant adverse effects. In a good-quality systematic review of 21 studies, interferon monotherapy was associated with significant adverse effects (many dose-dependent) including flulike syndrome (41–76%), alopecia (16–19%), thyroid disease (2%), depression (7–10%), thrombocytopenia (4–8%), leukopenia (9–13%), stopping treatment (4–5%), and dose decreases (9–22%).205 Serious or life-threatening side effects occurred in 1–2% of patients. Some trials have reported side effects that result in discontinuation of treatment in up to 15% of patients.68 Because of the long duration of most interferon treatment regimens (6 months) these adverse effects can have significant (though usually self-limited) impacts on quality of life.
Two systematic reviews evaluated the adverse events with non-pegylated interferon monotherapy versus non-pegylated monotherapy plus ribavirin combination therapy.206, 208 These results are presented in Evidence Table 2. Both reviews found the risk of treatment discontinuation and dose reduction were greater with combination therapy than with interferon alone. Approximately 21% of persons receiving interferon plus ribavirin discontinued treatment compared to 8% of those on interferon monotherapy.208 Approximately 26% of persons receiving combination therapy reduced their treatment dose as compared to 9% on monotherapy.208 Other adverse events were also more common, including anemia, cough, dyspepsia, dyspnea, leukopenia, pharyngitis, and pruritus, with as many as 67% of persons on combination therapy experiencing at least one adverse event.208
We identified four RCTs that compared adverse events in pegylated versus non-pegylated interferon monotherapy. These studies are summarized in table 1a and 1b. In all 4 studies, adverse events occurred at similar rates in the two treatment groups.213–216 The most commonly reported events were fatigue, headache, myalgias, rigors, and fever. The rates of injection site reactions and depression were higher in the pegylated interferon groups in 3 of the 4 studies. No deaths occurred as a result of therapy.
We identified three RCTs that provided data about adverse effects associated with pegylated interferon plus ribavirin versus pegylated interferon monotherapy.210–212 These results are summarized in Evidence Table 1a. In all of the studies, adverse events were reported at similar rates in both groups. Between 50% and 60% of study subjects experienced some adverse events, but these were usually mild and self-limited. In the two large, good-quality trials, the most frequent adverse events were myalgias (42–56%), fatigue (54–64%), headache (47–62%), and fever (43–46%).210, 211 Psychiatric (depression 22–31%), gastrointestinal and dermatological problems were also common. In studies that evaluated different doses of pegylated interferon and ribavirin, flu-like symptoms, headache, and asthenia increased with increasing medication doses in all treatment groups. None of the studies noted any new or unusual adverse events, nor did any serious complications or deaths occur as a result of treatment. Mild, self-limited hematologic abnormalities (decrease in hemoglobin) were noted more frequently in the combination therapy group in one study.212 Fever, nausea, injection site reactions, and dose reductions were more common with combination therapy than with pegylated interferon monotherapy in one trial, but dose discontinuation rates were not different between groups in either study.211 In the large, good-quality studies, the withdrawal rates on pegylated interferon plus ribavirin therapy were 14% and 22%.
In summary, the majority of patients taking interferon-based therapies alone or in combination with ribavirin experienced adverse effects. Withdrawal rates on interferon monotherapy averaged 5%, and were higher at 10–20% on combination therapy. The most common adverse effect was a flu-like syndrome. Serious adverse effects included depression that rarely led to suicide. The addition of ribavirin for combination therapy may increase the rate of mild hematologic side effects. The addition of pegylated interferon does not appear to significantly increase the risk of adverse events, though injection site reactions may be more frequent. Few major adverse events or deaths occurred with any of the treatments.
Arrow 9: Have improvements in intermediate outcomes (liver function tests, viral remission, histologic changes) been shown to reduce the risk or rate of harm from hepatitis C?
A recent good-quality, systematic review identified four cohort studies that examined the relationship between intermediate response to interferon treatment and long-term health outcomes.207 These studies were heterogeneous in design and had methodologic limitations. Specifically, cohort studies are not ideal for addressing this question due to potential bias from underlying population differences. For example, patients who respond to treatment may have less serious disease than those who do not. This might be evident in pre-treatment ALT levels, HCV RNA levels and in histological differences. The systematic review did not provide information about clinical, demographic, or other differences in each treatment response group. The review concluded that the studies were somewhat consistent in suggesting a moderately decreased risk of HCC and cirrhosis in complete responders to treatment, and that there might be a decreased risk of HCC and progression of liver disease associated with a biochemical response to treatment.
We identified one additional study in which long-term outcomes were stratified by the serologic response to interferon treatment. This was a large (1,643) retrospective cohort study from Japan of patients diagnosed with chronic HCV, 72.5% (1,191) of whom received interferon therapy.233 Of those who received treatment, 38.7% (461) were complete responders (CR), 12.2% (145) were biochemical responders (BR) only, and 49.1% (585) were non-responders. Some demographic differences were observed between these groups: patients receiving interferon treatment were younger, less likely to drink alcohol, had higher ALT levels, and were less likely to have genotype 1. The main long-term outcomes were crude and adjusted hepatocellular cancer rates and 10-year survival. Patients were observed for a total of 22.8 years. At the end of 10 years, the crude hepatocellular rates were 1.5% (CR or BR), 14.9% (NR), and 12.4% (NT) (p=0.0011). Ten-year survival was 98.8% (treated) and 95.6% (no treatment) (p<0.001). Multivariate analysis of the hepatocellular cancer rate found a hazard ratio of 0.32 (0.13–0.78), (p=0.012) for CR plus IR versus no treatment after adjustment for other factors.
We identified four clinical trials of treatment-naïve patients with HCV infection that analyzed quality of life outcomes according to response to treatment.222, 234–236 Of these, no study was rated good-quality; three were rated fair-quality (Evidence Table 4)234–236 and one222 poor-quality (see Arrow 7). In all studies, patients could have been aware of results of biochemical or virologic testing prior to repeating quality of life testing. In one fair-quality study, patients were only notified of viral load test results at the end of treatment after completing the quality of life questionnaire, though they may have known the results of earlier tests.234 In another fair-quality study, patients were blinded to viral load testing but could be aware of transaminase test results.235
All three fair-quality studies found improvements in health-related quality of life measures in sustained responders compared to non-responders (Table 6). In two studies, a significant difference between baseline and post-treatment SF-36 scores was found in 5 of 8235 and 8 of 8 domains.234 In the other study, an improvement of at least 2 standard deviations was seen in 7 of 8 SF-36 domains in responders compared to non-responders.236 The range of significant differences between responders and non-responders in SF-36 domain scores was 7 to 22 in one study235 and 3 to 10 in another234 Some experts have suggested that a difference in scores of 3–5 points represents a minimally clinically important change and 10 points a moderately important change.237, 238 In all three studies, the greatest difference in pre- and post-treatment scores was seen in the domain evaluating ability to perform physical roles. In the study in which patients were blinded to viral load results but not transaminase levels, subgroup analysis of patients with normalized transaminases found that sustained virologic responders reported improved quality of life scores compared to non-responders.235 These trials had relatively short duration of follow-up (24-weeks post-treatment); we did not identify studies evaluating whether quality of life improvements in responders are sustained.
- Arrow 1: Does screening for hepatitis C reduce the risk or rate of harm and premature death and disability?
- Arrow 2: Can clinical or demographic characteristics identify a subgroup of asymptomatic patients at higher risk for HCV infection?
- Arrow 3: What are the test characteristics of HCV antibody testing?
- Arrow 4: What is the false-positive rate and what are the harms associated with screening for hepatitis C virus?
- Arrow 5a: What are the test characteristics of the work-up for treatable disease?
- Arrow 5b: In patients found to be positive for hepatitis C antibody, what proportion of patients would qualify for antiviral treatment?
- Arrow 6: What are the harms associated with the work-up for active HCV disease?
- Arrow 7a: How well does antiviral treatment reduce the rate of viremia, improve transaminase levels, and improve histology?
- Arrow 7b: How well does antiviral treatment improve health outcomes in asymptomatic patients with hepatitis C?
- Arrow 7c: How well do counseling and immunizations in asymptomatic patients with hepatitis C improve clinical outcomes or prevent spread of disease?
- Arrow 8: What are the harms (including intolerance to treatment) associated with antiviral treatment?
- Arrow 9: Have improvements in intermediate outcomes (liver function tests, viral remission, histologic changes) been shown to reduce the risk or rate of harm from hepatitis C?
- Results - Screening for Hepatitis C Virus InfectionResults - Screening for Hepatitis C Virus Infection
Your browsing activity is empty.
Activity recording is turned off.
See more...