Guideline 1: Detection and evaluation of HCV in CKD

INTRODUCTION
The prevalence of HCV infection is higher in most subgroups of CKD patients than in the general population.36,57,58 The reasons for testing CKD patients for HCV include diagnostic evaluation of the cause of CKD (specifically, HCV-associated GN), infection control in hemodialysis units, and optimal care before and after kidney transplantation. Treating HCV infection as early as possible is another major reason for testing all CKD patients who may benefit from antiviral treatment. The specifics of diagnostic testing for HCV in various CKD populations are discussed below, taking into account the presumed prevalence of HCV infection in each population and its characteristics (especially immune deficiency).

1.1 Determining which CKD patients should be tested for HCV

1.1.1 It is suggested that CKD patients be tested for HCV. (Weak)
1.1.2 Testing for HCV should be performed in patients on maintenance hemodialysis (CKD Stage 5D) and kidney transplant candidates. (Strong)

1.2 HCV testing for patients on maintenance hemodialysis:

1.2.1 Patients on hemodialysis should be tested when they first start hemodialysis or when they transfer from another hemodialysis facility. (Strong)

• In hemodialysis units with a low prevalence of HCV, initial testing with EIA (if positive, followed by NAT) should be considered (see Algorithm 1). (Moderate)
• In hemodialysis units with a high prevalence of HCV, initial testing with NAT should be considered (see Algorithm 1). (Moderate)

1.2.2 For patients on hemodialysis who test negative for HCV, retesting every 6–12 months with EIA should be considered. (Moderate)
1.2.3 Testing for HCV with NAT should be performed for hemodialysis patients with unexplained abnormal aminotransferase(s) levels. (Strong)
1.2.4 If a new HCV infection in a hemodialysis unit is suspected to be nosocomial, testing with NAT should be performed in all patients who may have been exposed. (Strong)

• Repeat testing with NAT is suggested within 2–12 weeks in initially NAT-negative patients (Weak).

BACKGROUND
Worldwide, HCV is the major etiologic agent of chronic hepatitis that can lead to the development of liver cirrhosis and hepatocarcinoma. According to the American Association for the Study of Liver Disease (AASLD),⁵⁹ individuals who should be tested for HCV infection include those

• who have injected illicit drugs in the recent or remote past;
• with conditions associated with a high prevalence of HCV infection, including
• HIV infection,
• hemophilia treated by clotting factor concentrates before the availability of heat-treated concentrates,
• CKD Stage 5 ever treated by hemodialysis,
• unexplained abnormal aminotransferase levels,
• recipients of a transfusion or organ transplant, specifically those
• who have been notified that they had received blood from a donor who later tested positive for HCV infection,
• who have received a transfusion of blood or blood products before the systematic testing of blood donors with second-generation EIA or more recent tests,
• who have received an organ transplant before the systematic testing of organ donors with secondgeneration EIA or more recent tests;
• who are suspected of having chronic HCV infection;
• children born to HCV-infected mothers;
• health-care, emergency-medical, and public safety workers after a needlestick injury or mucosal exposure to HCV-positive blood;
• sexual partners of HCV-infected persons.

RATIONALE
1.1 Determining which CKD patients should be tested for HCV:

1.1.1 It is suggested that CKD patients be tested for HCV. (Weak)
1.1.2 Testing for HCV should be performed in patients on maintenance dialysis (CKD Stage 5D) and kidney transplant candidates. (Strong).

HCV infection has been associated with various types of GN, such as cryoglobulinemic GN,⁶⁰ MPGN, focal and segmental glomerulosclerosis, and membranous GN.31,⁶¹ Thus, testing for HCV appears logical in CKD patients with either hematuria or proteinuria indicative of the possibility of GN. In addition, in diabetic patients with CKD, HCV infection is an independent predictor of a more rapid decrease of GFR.^62,63 Finally, the prevalence of HCV infection tends to be higher in patients with CKD not yet on dialysis than in the general population.^64,65 Thus, testing such patients with CKD is associated with the additional advantage of potentially offering antiviral treatment to all those who are able to benefit from it.

The prevalence of HCV infection is much higher in hemodialysis patients (CKD Stage 5D) than in the general population⁵⁸ and is associated with an increased mortality rate.^66,67 Although HCV infection results in an increase in ALT, levels are generally lower in hemodialysis patients^66–68 and kidney transplant patients⁶⁹ than in the general population, and the reasons for this are unknown. Therefore, a single measurement of ALT level is not a good tool to detect or rule out either acute or chronic HCV infection in these patients. Treating HCV infection after kidney transplantation is associated with an increased risk of rejection. This, taken together with the need for optimal selection of kidney transplant recipients, makes testing for HCV mandatory as part of the pretransplant evaluation. Moreover, patients in all stages of CKD frequently have abnormal liver enzyme levels and also an increased prevalence of HCV infection. Thus, testing for HCV is also mandatory in the post-transplant period.

1.2 HCV testing for patients on maintenance hemodialysis:

1.2.1 Patients on hemodialysis should be tested when they first start hemodialysis or when they transfer from another hemodialysis facility. (Strong)

• In hemodialysis units with a low prevalence of HCV, initial testing with EIA (and, if positive, followed by NAT) should be considered (see Algorithm 1). (Moderate)
• In hemodialysis units with a high prevalence of HCV, initial testing with NAT should be considered (see Algorithm 1). (Moderate).

The prevalence of HCV infection in patients on hemodialysis is highly variable but clearly much higher than in the general population of the respective countries. In phase one of the Dialysis Outcomes and Practice Patterns Study (DOPPS)- a prospective, observational study of adult hemodialysis patients randomly selected from 308 representative dialysis facilities in France, Germany, Italy, Japan, Spain, the United Kingdom, and the United States-an overall HCV prevalence of 13% was found in 8615 patients.⁴⁰ The HCV prevalence varied from 3% (the United Kingdom, Germany) to 23% (Italy, Spain). In some developing countries, the prevalence of HCV in dialysis patients is even higher: Brazil 24–47%,⁷⁰ India 12–45%,⁴² Jordan 35%,⁷¹ Saudi Arabia 43%.⁷²

The prevalence of HCV infection is influenced, among other factors, by

• dialysis modalities, that is, hemodialysis (center4home hemodialysis)4peritoneal dialysis;
• hemodialysis vintage;
• history of blood transfusions or organ transplantation before effective screening of donors with secondgeneration EIA or more recent tests;
• prevalence of HCV infection in the dialysis unit.

Testing for HCV in all patients initiating hemodialysis or transferring to a new facility should be part of the infection control strategy of each hemodialysis facility (see Guideline 3).

The detection of anti-HCV antibodies is based on the use of third-generation EIA that detects antibodies directed against various HCV epitopes. EIA tests are reproducible, inexpensive, and suitable for use in the diagnosis of HCV infection. Given the good performance of third-generation EIA tests, immunoblot tests have become obsolete in clinical practice.³ The increased sensitivity of the last generation of HCV assays has dramatically reduced the risk of HCV transmission by blood components and reduced the time between acquisition of infection and detection of anti-HCV antibodies (the ‘serologic window') from 82 to 66 days.^4,5 In the near future, fourth-generation tests will be available, allowing the simultaneous detection of HCV antibodies and HCV core protein. These tests should further reduce the serologic window. In some populations with frequent polyclonal hypergammaglobulinemia, there may be discrepancies among third-generation EIA tests. Thus, in pregnant women in Cameroon, HCV-positive determination using only one third-generation EIA test was 4.9%, but it decreased to 1.9% when requiring positive results of two third generation EIA tests; HCV RNA was present in 75% of women having concomitantly two positive EIA tests and 0% in those having only one positive EIA test.⁶

NAT is based either on qualitative HCV RNA detection or on HCV RNA quantitation. Qualitative detection assays are based on the principle of target amplification using conventional PCR, real-time PCR, or TMA. All commercially available assays can detect 50 IUml^-1 or less of HCV RNA, and have equal sensitivity for the detection of all HCV genotypes.⁷ The lower limit of detection of the qualitative conventional PCRbased assays or their semiautomated version is 50 IUml^-1; that of real-time PCR assays, which are able at the same time to qualify and quantify HCV RNA, is 10–30 IUml^-1; and that of TMA-based assay is 10 IUml^-1. Quantitative assays are based either on target amplification techniques (conventional PCR or real-time PCR) or on signal amplification techniques (branched DNA). Branched DNA and most quantitative conventional PCR-based assays have detection limits higher than those of qualitative detection assays.

NAT should be performed in laboratories that have facilities specifically designed for that purpose. Serum or plasma samples must be collected, processed, and stored in a manner suitable for minimizing false-negative results obtained from NAT. Serum or ethylenediaminetetraacetic acid (EDTA) plasma must be separated from cellular components within 2–6 h after collection. Storage of serum or EDTA plasma at 2–5 °C should be limited to 72 h; for longer storage, freezing at -20 or -70 °C is recommended. Samples collected for serologic testing can be used only if these conditions are met.⁷³ Since heparin is an inhibitor of PCR,^74,75 samples from hemodialysis patients should be obtained before the dialysis session and from a peripheral vein in patients with a central catheter locked with heparin.

Tests other than classical EIA or NAT may become clinically available in the relatively near future.⁷⁶ Among the potential test candidates is one for the core protein, which is a structural HCV protein whose sequence is highly conserved across HCV genotypes.⁷⁷ The HCV core antigen test, in hemodialysis patients, has a sensitivity and specificity of 84 and 89%, respectively.⁷⁸ There is now an HCV test that combines the simultaneous detection of HCV core antigen and anti-HCV antibodies, and also enables an early detection of HCV infection during the so-called ‘window period' compared to anti-HCV assays. This test could be a useful alternative to HCV RNA detection or HCV core antigen assays for diagnosis or blood screening when NAT or HCV core antigen detection is not implemented.⁵ However, these tests are not yet routinely available.

How should hemodialysis patients be tested?
The consequences of variable diagnostic accuracy are not purely academic. Patients with positive diagnostic tests will be further assessed, including, if indicated, a liver biopsy, and eventually will receive appropriate treatment, with its own associated benefits and harm. Patients with false-positive results might be subjected to further inappropriate testing- in particular, liver biopsies-and eventually unnecessary treatment. Patients with false-negative results might lose an opportunity for intervention (with possible increased morbidity and mortality of undiagnosed HCV infection) while on dialysis and after kidney transplant. The sensitivity and specificity of EIA as compared with the reference standard of NAT have been examined in a meta-analysis of relevant published papers (Tables 2–4). This analysis made several assumptions. While acknowledging that EIA and NAT measure different conditions (antibody response to present or past infection and viremia, respectively), in practice these two tests are both used primarily as markers of HCV infection. This analysis assumes that EIA is used as a cheaper, more readily accessible alternative to the more definitive NAT. Thus, sensitivity and specificity of EIA are based on NAT as the reference standard. It is important to realize that there are a number of conditions where EIA and NAT accurately disagree (such as continued antibody response after a cleared infection or immunodeficiency with active viremia). However, for the purpose of this analysis, it is assumed that both tests are being considered as alternatives for making new diagnoses of HCV infection. Patients for whom EIA is considered an inaccurate test for active infection, due to severe immunodeficiency, should be tested with NAT alone. In the relevant published studies, the sensitivity of EIA varied from 53 to 100% and the specificity from 85 to 100%, with pooled sensitivity and specificity of 75 and 95%, respectively (Figure 1). Across studies, there was no association between HCV prevalence and reported sensitivity and specificity.

Implications with consideration of unit prevalence
Figure 2 depicts how the estimates of HCV prevalence can change after an EIA test and vary depending on the actual prevalence of HCV and the sensitivity and specificity of EIA. In each graph, the upper curve indicates the prevalence of HCV among patients with a positive EIA over the full range of true HCV prevalence (that is, the positive predictive value of EIA). The lower curves indicate the prevalence of HCV among patients with a negative EIA (or 1-negative predictive value). Figure 2b uses a theoretical midpoint in both sensitivity and specificity from the summary receiver operating characteristics curve, based on the summary estimates among the 12 studies (sensitivity 75%, specificity 95%), whereas approximate extreme values are used in Figure 2a and c.

The graphs also highlight several theoretical pretest prevalence values. The pretest values refer to the best guess estimate of the likelihood of HCV infection before the test is performed. A starting point for this estimate can be the recent prevalence of disease in a dialysis unit or in a region. The pretest prevalence values graphed include the mean (13%) and extreme values of 3% and 23% in DOPPS40 as well as a 40% prevalence taken to represent higher prevalence settings. Using Figure 2b, at low pretest prevalence values (3%, 10%) when the EIA is negative, the post-test prevalence (the likelihood of infection after the test) remains less than 5%. However, as pretest prevalence increases to 23% or 40%, the post-test prevalence among those with a negative EIA increases to 15% (in Figure 2b). Thus, with higher baseline prevalence rates of disease, there is an increasing proportion of false-negative EIA results.

Furthermore, in the low pretest prevalence setting, a positive EIA results in only a 30–65% post-test prevalence of HCV. As pretest prevalence increases, a positive EIA results in increasing post-test prevalence of HCV. For example, at a pretest prevalence of 40%, a positive EIA results in about a 90% post-test prevalence of HCV infection. Thus, in low-prevalence settings, there is a relatively high risk of false-positive EIA results.

However, uncertainty exists about the true sensitivity and specificity of EIA testing of hemodialysis patients. Figure 2a demonstrates the post-test prevalence estimates of HCV as a function of pretest prevalence when EIA has a low sensitivity (53%) and high specificity (99%). Using these performance characteristics, the curves are shifted to the upper left. With increasing pretest prevalence of HCV, the post-test prevalence of HCV when EIA is negative increases to as high as 24% in the highest prevalence setting. As a result, there is a marked increase in false-negative EIA test results as pretest prevalence rises. Conversely, there is a low rate of false-positive testing (that is, nearly all positive EIA results represent true HCV infection).

Figure 2c depicts the scenario where sensitivity is high and specificity relatively low (sensitivity 99%, specificity 87%), where the curves are shifted to the lower right. Here, a negative EIA result likely represents a true negative in all but the highest (480%) prevalence settings. However, a positive EIA result has a high likelihood of being a false positive in the range of prevalence common in dialysis units. For instance, in the 3–13% range, a positive EIA results in only a 15–50% prevalence of disease.

In summary, in low-prevalence settings, EIA is adequate to rule out HCV infection when the test is negative, but a positive EIA would need confirmation with NAT. In this setting, only relatively few patients will require NAT testing, as most patients without HCV will test negative on EIA. In higher prevalence settings, a negative EIA becomes increasingly unreliable to rule out HCV infection; thus, initial testing with NAT becomes appropriate to avoid missing HCV infections. However, given the uncertainties regarding the sensitivity of EIA to predict NAT-positive patients and given the differing preferences regarding the acceptable risks of missing HCV infections in hemodialysis patients, no single threshold can be used to distinguish between high- and low-prevalence settings.

In addition, it is important to note that this discussion refers only to patients who have not been recently tested for HCV. Patients who were previously tested negative for HCV (by EIA or NAT) and who have not had an intervening event placing them at increased risk can be considered to be at low risk for HCV and should be tested with EIA when necessary (see rationale for Guideline 1.2.2). Conversely, patients with recently elevated ALT/AST levels or with another event that has placed them at high risk for HCV infection (where their likelihood of infection is greater than the threshold used for high prevalence) should be tested with NAT.

Hemodialysis patients are tested for HCV to identify infected patients, who may be treatment candidates, and to identify newly infected cases for the purposes of infection control. The consideration of which test (EIA or NAT) to use should depend on the prevalence of HCV in the dialysis unit to minimize false-positive and false-negative results. The implications of a false-negative EIA test (whether due to test error or to immune dysfunction in the setting of viremia) include a delay or failure of diagnosis of HCV infection. Not identifying HCV-infected hemodialysis patients has important implications, as HCV infection is associated with increased mortality (relative risk (RR)=1.57)⁷⁹ on hemodialysis and after kidney transplantation (RR=1.79).⁸⁰ Moreover, non-recognition of HCV-infected hemodialysis patients can increase the risk of transmission to other dialysis patients (particularly if the infection was due to, in part, poor infection control in the dialysis unit). On the other hand, a false-positive EIA test (based on test error as opposed to a previous history of HCV infection) will lead to unnecessary additional testing (NAT).

In high-prevalence settings, high false-negative rates call into question the value of EIA as a screening test for HCV in a dialysis unit. In these settings, a high percentage of patients with a negative EIA are, in fact, HCV RNA-positive. For example, in a setting with a pretest prevalence of HCVof 40%, of those who test negative by EIA, 15% will be HCV RNA positive when a NAT is performed. Furthermore, in high-prevalence settings, a large percentage of patients will require additional testing with NAT due to positive EIA tests (both true positives and false positives), mitigating the cost savings of starting with EIA.

In low-prevalence settings, there is an increased likelihood of false-positive EIA testing (whether due to test error or to previously cleared HCV infection). However, owing to the low prevalence of disease, the actual number of false-positive results (as well as the number of true positives) will be low; thus, there will be only a small additional expense of further testing. For example, in a setting with 5% prevalence, if specificity is 95%, less than 5% of patients will require confirmation with NAT.

1.2.2 For patients on hemodialysis who test negative for HCV, retesting every 6–12 months with EIA should be considered. (Moderate).

How should regular testing be performed?
Patients who are EIA-negative and NAT-negative should be considered in a low-risk (that is, low prevalence) group unless a change in clinical status that increases the likelihood of acute HCV infection occurs, for example, elevated ALT/ AST levels or other traditional risk factors for acquiring HCV such as intravenous drug user (IVDU), and so on. In most countries, the incidence rate of new HCV infection varies from 0 to 3.6% in most units (DOPPS 1 (1996–2001): Italy, 3.6%; the United States, 3.1%; the United Kingdom, 1.1%; Japan and Spain, 3%; France, 1.9%; Germany, 1.7%). Other, more recent studies indicate incidence rates of Italy, 2%; Japan, 0.3%;⁸¹ Tunisia, 0.5%.⁵⁰ However, in some units prevalence is as high as 75%, suggesting a persistently high incidence (for example, Casablanca, Morocco).⁸² Thus, in most countries and units, there is a very low likelihood of acquiring new HCV infection in a 6-month to 1-year period even in dialysis units with the highest prevalence. Consequently, repeat NAT testing on an annual or biannual basis will likely detect very few cases and will not be cost-effective. The proposed interval of 6–12 months between tests for HCV may probably be extended in the few patients treated exclusively by home hemodialysis. Monthly ALT testing to identify those patients with increased likelihood of acute HCV infection should be adequate when combined with biannual EIA testing. These patients are in a low-prevalence state if they have had a onetime negative NAT. A patient whose ALT/AST levels increase acutely would be managed as having a higher likelihood of HCV infection in the period in which the ALT/AST levels increase. Also, travel to regions where the prevalence of HCV is high in dialysis centers is unlikely to be associated, in a brief period of time on hemodialysis, with such an increase of the risk that the high-prevalence threshold is reached. Thus, NAT testing on return to the primary unit is not necessary (in the absence of elevated aminotransferase levels). This approach may be adapted if the incidence/prevalence in the hemodialysis unit where the patient is visiting is extremely high and/or the stay prolonged (several months).

This usually raises the issue of previous HCV testing with EIA. If one-time NAT testing of the entire cohort of EIAnegative hemodialysis units is adopted, does a history of previous EIA negativity decrease the likelihood of a positive NAT? In theory, repeated EIA-negative results might be hypothesized to lower the likelihood of a positive NAT, but this would require that the performance characteristics of this test on sequential samples are independent. However, it is likely that this is not the case. False-negative EIA results are relatively rare in immunocompetent individuals; thus, the false-negative results in hemodialysis patients likely represent an impaired immune response to HCV infection. Consequently, an EIA-negative patient who is actually NAT-positive would be more likely to have repeated false-negative results during ongoing testing. Thus, NAT testing of EIA-negative patients from high-prevalence settings would be anticipated to detect additional HCV RNA patients.

1.2.3 Testing for HCV with NAT should be performed for hemodialysis patients with unexplained abnormal aminotransferase(s) levels. (Strong).

In suspected acute HCV infection, a negative anti-HCV test does not exclude HCV infection. After an exposure to HCV, HCV RNA can be detected within 1–2 weeks, whereas antibodies to HCV are detectable only, on average, 8 weeks later in immunocompetent subjects.⁵⁹

ALT levels increase in acute HCV infection. Although a single measurement is not useful as a screening method for HCV, ALT levels are regularly measured in CKD Stage 5D patients, and an elevation of ALT levels compared to baseline values may suggest a recent infection.^83–85 Even though in hemodialysis patients the ALT levels are lower than those in patients without kidney disease,⁵⁹ an unexplained elevation of aminotransferase levels from baseline should prompt testing by NAT.

In CKD Stage 5D patients, the serologic window, that is, the time lag between acute HCV infection and seroconversion, may have a duration of up to several months.^83,84 In a recent study in CKD Stage 5D patients, the median interval between NAT positivity and EIA positivity was 246 and 154 days for second- and third-generation EIA, respectively.⁸⁶

1.2.4 If a new HCV infection in a hemodialysis unit is suspected to be nosocomial, testing with NAT should be performed in all patients who may have been exposed. (Strong)

• Repeat testing with NAT is suggested within 2–12 weeks in initially NAT-negative patients (Weak).

If the nosocomial transmission of HCV to a patient on hemodialysis is suspected, the early recognition of other cases of acute HCV infection within the facility would provide an additional clue to ongoing HCV transmission; hence, there is urgency to audit and reinforce the basic hygienic precautions to prevent HCV transmission. In addition, the early diagnosis of acute HCV infection should prompt early treatment in suitable candidates, with a much better chance of therapeutic success (see Guideline 2).

A negative sensitive NAT test in a person with a positive EIA most likely indicates that the HCV infection has resolved.⁵⁹ Other interpretations are that the anti-HCV immunoassay is falsely positive, the NAT test is falsely negative, or rarely, that a person has intermittent or low-level viremia. The latter situation is most unlikely when using TMA. As the implications of missing actual HCV viremia may be substantial, a repeat testing of EIA-positive NATnegative patients is recommended.

Summary of recommendations

• Dialysis units with a known high prevalence of HCV should ensure that all patients have been tested once with NAT (as it is likely that some EIA-negative patients are actually HCV RNA-positive).
• Incident dialysis patients with a high likelihood of being HCV infected and without a documented NAT should have NAT testing performed on admission to the unit.
• EIA-negative patients who are believed to be at high risk (using the same threshold as for high prevalence) of HCV infection due to changes in risk factors or exposures should be tested with NAT.
• Patients in low-prevalence units, from low-prevalence regions or countries, and those who remain at low risk of infection (below the threshold for high prevalence) should be tested with EIA.

RESEARCH RECOMMENDATIONS

• Additional studies are required to better delineate the actual prevalence of HCV infection and distribution of HCV genotypes at various stages of CKD.
• The sensitivity of EIA testing for the detection of HCV infection along the various (especially early) stages of CKD should be investigated.
• The impact of CKD stage on the natural history of HCV viremia should be studied further.