Community viral load and hepatitis C virus infection: Community viral load measures to aid public health treatment efforts and program evaluation

Abstract

Background

Hepatitis C virus (HCV) infection is the most prevalent blood-borne infection and causes more deaths than any other infectious disease in the US. Incident HCV infection in the US increased nearly 300% between 2010 and 2015, Community viral load (CVL) measures have been developed for HIV to measure both transmission risk and treatment engagement in programs or areas.

Objective

This paper presents a systematic review exploring the published literature on CVL constructs applied to HCV epidemiology and proposes novel CVL measures for HCV.

Study Design and Setting

A systematic review was conducted of electronic databases; the search sought to identify published literature on HCV which discussed or applied CVL measures to HCV epidemiology. Novel CVL measures were constructed to apply to HCV.

Results

No reports examining quantitative measures of HCV CVL were identified. Using the HIV CVL literature and the specific characteristics of HCV epidemiology, five HCV CVL measures are proposed. Narrower measures focusing on those engaged-in-care may be useful for program evaluation and broader measures including undiagnosed people may be useful for surveillance of HCV transmission potential.

Conclusion

Despite their potential value, CVL constructs have not yet formally been developed and applied to HCV epidemiology. The CVL measures proposed here could serve as valuable HCV program and surveillance measures. There is a need for informative surveillance measures to enhance policy and public health responses to achieve HCV control. Further study of these proposed HCV CVL measures to HCV epidemiology is warranted.

1.0. Background

There is a public health crisis in the United States (US) of opioid-related morbidity and mortality that can be best understood as a syndemic of opioid misuse, overdose, HIV, hepatitis C virus (HCV).¹ HCV is the most prevalent blood-borne infection and causes more deaths than any other infectious disease in the US.^2,3 HCV is a preventable infection most efficiently spread via parenteral exposure through non-sterile injection practices.⁴ Incident HCV infection in the US increased nearly 300% between 2010 and 2015,⁵ an increase which has been found to be temporally linked to the opioid epidemic in longitudinal cohort studies.^6–8 These findings highlight the need for informative surveillance measures to enhance policy and public health responses to achieve HCV control.

Community viral load (CVL)—a group-, program-, or area-level aggregate measure of individual quantitative (e.g., mean or total virion copies/ milliliter) viral loads (VLs)—is a construct that has been applied to HIV to monitor progress towards prevention and antiretroviral therapy (ART) goals; HIV CVL is central the concept of ‘treatment as prevention’ (TasP).⁹ At the individual-level, VL magnitude is a potent independent predictor of HIV disease progression;¹⁰ it is also directly associated with the probability of vertical,¹¹ sexual,¹² and needlestick transmission through any given exposure event.^13–14 This, along with evidence that HIV VL suppression reduces individual-level transmission risk, supports the relevance of the magnitude of HIV VL at the area- or group-level (i.e., CVL) on HIV incidence.^15–17

Data indicate strong population-level associations between increased ART coverage and subsequent decreases in HIV CVL and decreases in HIV incidence.^18–21 CVL constructs have been used as predictors of HIV incidence,^20,22,23 to examine the population-level impact of HIV prevention and treatment,^15,18,19,24 and as measures of area-level disparities in which higher area-level CVLs are associated with higher rates of area-level poverty.^15,23,25 HIV CVL measures have been shown to be associated with decreases in HIV incidence.²⁶

Data suggest associations between the magnitude of individual HCV VLs and infectivity.^27–29 The risk of HCV transmission through nosocomial needlesticks is associated with higher source patient HCV VL;¹³ these data and modeling studies led the Society for Healthcare Epidemiology of America to advise that HCV infected healthcare workers with HCV VLs ≥10,000 IU (international units)/mL be restricted from procedures that might risk exposure.³⁰ Further, data have shown higher volumes of blood present in used syringes are directly associated with higher risk of HCV transmission.²⁷ Since VL is the concentration of viral particles per unit volume, the combination of a low volume blood exposure and a low concentration of virus in that blood (i.e., a low VL) translates to a low likelihood of exposure to any viral particles.²⁷

Combined, these data suggest that both individual HCV VLs and aggregate HCV VLs (i.e., HCV CVLs) may impact transmission risk. This further suggests that measures of HCV CVL may be useful to HCV control efforts.^27,30–32 In the absence of effective area-level or group-level treatment, a high prevalence of active HCV infection yields a high CVL which then serves as a large HCV reservoir and increases the probability that a non-sterile injection episode will contribute to onward HCV transmission.^33–35 However, HCV treatment results in cure; since cured individuals can no longer transmit infection, treatment can be viewed as cure as prevention (CasP).

1.1. Objective

Despite similarities in modes of transmission for HIV and HCV, the CVL construct has not been applied to HCV. We conducted a systematic review to formally search the published literature for applications of CVL constructs to HCV epidemiology. Based on the results from the review and specific considerations of HCV epidemiology, we propose five novel CVL measures specifically constructed for HCV that could be used in program and intervention evaluation, public health surveillance, and studies of HCV incidence.

2.0. Methods and Study Design

The objective of the systematic review was to characterize the published data on applications of CVL constructs to HCV epidemiology.^36,37 This systematic review was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).³⁸

2.1. Systematic review methods

Electronic searches were conducted to identify potentially eligible reports published between 01/01/1991-12/31/2017. MEDLINE, EMBASE, the Cochrane Database of Systematic Reviews and Register of Controlled Trials, and PubMed were searched. The search strings included, as key terms: “HCV”, “RNA”, and “community viral load”. Primary data reports of quantitative HCV VL data aggregated at the group-, area-, or program-level and used to examine HCV transmission, incidence, or program evaluation were eligible for inclusion in the systematic review.

2.2. Bridging the HIV literature to construct HCV CVL measures

There is a robust and growing literature of HIV CVL measures.^{9,10,15,16,20,23,39–41} We used the HIV literature to inform the development of CVL measures applied to HCV. The Centers for Disease Control and Prevention (CDC) HIV CVL measures were used as a starting point to reconceptualize CVL to fit HCV epidemiology.⁹ Each measure reflects considerations at different levels of analysis including when all (or most) CVL data would be expected to be available, as well as estimating the reservoir of virus in larger networked populations or areas, where directly measured quantitative VL data are likely to be incomplete.^{10,15–19,21–24,40,42}

3.0. Results

3.1. Systematic review results

Searches retrieved a total of 727 potentially eligible reports (Figure 1). After removing duplicate records, 514 report abstracts were assessed for eligibility; 11 reports were ascertained as requiring full-text assessment to determine final eligibility. After full-text review, no report was ultimately eligible as none presented HCV CVL as an area-level measure and none expressed group mean or median VLs studied in relation to incidence or transmission risk.

Figure 1.

PRISMA flow diagram of systematic review of hepatitis C virus community viral load

The reports requiring full-text review (n=11) used mean or median HCV VLs for purposes including analyses of progression of hepatic injury, of differences in clinical presentation by specific characteristics, of VL testing methods, of relationships with other biomarkers and of outcomes of clinical trials of investigational anti-viral agents.^{43,44,44–51} Some reports examined the impact of individual HCV VL magnitude on disease progression, some demonstrating a direct relationship between individual HCV VL and various HCV-induced biologic outcomes.^{52,53,29,54–56} In the previous era of HCV treatment using pegylated interferon,⁵⁷ the magnitude of the pre-treatment individual HCV VL was a consistent predictor of treatment response, with higher VLs predicting lower rates of sustained virologic response (SVR), and longer times to viral clearance.^58–60

Only one report specifically mentioned HCV CVL in the context of enumerating barriers to HCV elimination; it noted that “combined, these barriers contribute to the persistence of a high community HCV viral load fueling ongoing transmission” but it did not provide a specific HCV CVL definition or present any quantitative data.⁶¹

One domain of exception were reports examining vertical HCV transmission of HCV which compared mean or median HCV VL among groups of HCV-infected pregnant women who did and did not vertically transmit HCV.^62–64 A critical distinction between injection-mediated and vertical transmission is that while any uninfected people who inject drugs (PWID) could engage in risk behavior with any PWID with active HCV infection, giving meaning to aggregates of group VLs for that individual’s risk, the risk of vertical transmission to any given fetus is not impacted by aggregate VLs of a group of pregnant women but only by that of the fetus’ mother.

3.2. Conceptualizing HCV CVL constructs

The systematic review revealed the absence of quantitative CVL constructs applied to HCV epidemiology as a gap in the published literature; no applications of HCV CVL constructs were published in the study period and no report directly developed HCV CVL constructs even from a conceptual standpoint. Therefore, published work on HIV CVL measures were used to bridge the gap in the HCV CVL literature by blending methods used for HIV CVL measures with the distinct aspects of HCV epidemiology, informing the proposed five HCV CVL.

One critical distinction between HIV and HCV infection is that unlike for HIV, for HCV, both spontaneous clearance and treatment can result in cure, and therefore some individuals remain HCV antibody positive but VL negative and therefore, do not contribute to transmission risk.^65–68 (Table 1) These individuals thus contribute a zero HCV VL in the proposed HCV CVL measures.

Table 1.

Differences between HIV and Hepatitis C virus (HCV) infection that are important in conceptualizing CVL constructs for HCV

Testing and Diagnosis
HIV	HIV antibody testing is sufficient to make a diagnosis of active infection.
HCV	HCV testing requires at least a two-step process to determine active infection (i.e., a positive anti-HCV and a detectable HCV VL). Note that not all quantitative VL tests are approved by the Food and Drug Administration for diagnostic testing. Qualitative HCV VL RNA tests (a dichotomous measure of the presence or absence of detectable viremia) provide an inexpensive and reliable measure of active HCV infection and are approved by the FDA for diagnostic testing. If qualitative HCV VL RNA tests are used when HCV antibody tests are positive then three testing steps are needed to determine quantitative HCV VL. There is significant heterogeneity among programs as to whether a positive HCV antibody test prompts referral elsewhere for VL testing or whether antibody testing prompts on-site VL testing with subsequent referral only of those with detectable VLs. These two- and three-step testing processes are often associated with losses at each step and create important categories of those incompletely diagnosed.
Natural History and Viremia
HIV	While acute HIV infection is associated generally with an initial very high HIV VL followed by a generally lower HIV VL, it is virtually never followed by spontaneous viral clearance and all those HIV infected and untreated remain viremic forever.
HCV	Approximately 25% of those who are infected with HCV spontaneously clear the virus (without intervention). Those who have spontaneously cleared HCV infection, remain antibody positive, do not contribute to transmission risk but do remain susceptible to re-infection with HCV.
Anti-Viral Treatment
HIV	HIV treatment does not result in cure. Consequently, any HIV infected person requires lifelong care, having distinct measures for those engaged-in-care reflects an important population because of the ongoing need for treatment and viral suppression.
HCV	HCV treatment is a transient state (typically 8-12 weeks) after which most people are done with treatment and are cured, retaining a positive anti-HCV but no detectable VL. Given that treatment durations are finite, measures reflecting the brief period of treatment may not be particularly relevant metrics for HCV. Those treated and cured do not contribute to CVL measures but do remain susceptible similar to those who spontaneously clear HCV infection. As HCV treatment is highly efficacious and treatment durations are short, group HCV CVL measures will consist of VL data contributed by those not treated (whether in or out of care).

Another important consideration in the conceptualization HCV CVL measures is the process of diagnosing infection. For HCV infection, different settings employ diverse strategies for identifying active infection among those HCV antibody positive,^69–72 either directly performing a quantitative VL test⁷³ or first performing a qualitative RNA test,⁷⁴ followed by a quantitative VL test for those found to have active infection.⁷⁵ Several factors may inform whether an HCV testing setting performs two- or three-step HCV testing. Quantitative VL tests are not all currently approved by the Food and Drug Administration (FDA) for diagnostic testing (as opposed to use for management of patients undergoing treatment) and are more expensive than qualitative RNA tests. While less expensive, qualitative RNA tests prompt the need for a quantitative VL test an additional step in those found to have active viremia. The proposed HCV CVL constructs are designed to be potentially applicable to all jurisdictions regardless of whether a two- or three-step HCV testing process is utilized.

Figure 2 visually depicts the types of HCV testing data to be used to construct the HCV CVL measures, including when data need to be estimated. Table 2 further delineates the five proposed HCV CVL. The five measures reflect increasingly inclusive denominators where more inclusive denominators include undiagnosed and unidentified individuals who are transiently infected and spontaneously clear HCV or who achieved SVR after treatment.

Figure 2.

The proposed five hepatitis C virus (HCV) community viral load (CVL) measures

Dotted outlines indicate that the enclosed component of the measure requires imputation for data which are unavailable.

* No current lab results, including viral load.

** The right-hand panel depicts two triangular figures, with the triangle on the left reflecting transmission potential and the triangle on the right reflecting the degree of engagement in care. The width of the triangular figure depicts the transmission potential which increases as the measures become more broadly defined, including more people in groups that are either diagnosed but untreated or undiagnosed, reflecting increasing engagement in and quality of care. For the right-hand triangle, width of the triangular figure depicts the degree of engagement in care which decreases as the measures become more broadly defined, including more people in groups that are either diagnosed but untreated or undiagnosed, reflecting increasing engagement in and quality of care.

*** It is important to note that if the two-step testing process is utilized, there is no need for box B as depicted in Figure 3, and the proposed diagnosis-based CVL and active CVL reduce to be the same.

Table 2.

Five proposed hepatitis C virus (HCV) community viral load (CVL) constructs

Enaaaed in care CVL Measure 1	Includes those with documented Quantitative HCV VL test results. This is a directly measured construct which is constructed by aggregating quantitative VL tests. No imputation or estimation of data are required for this measure. The measure provides programs with a minimum estimate of the number with active HCV infection and the magnitude of CVL; it characterizes the effectiveness of the program’s prevention and treatment efforts and identifies the extent of the need for treatment.
Hypothetical scenario	In a population of 10,000 in a drug treatment program where 60% have active HCV and 75% had quantitative HCV VLs available, the total number of individuals contributing to the HCV CVL would be 4,500 people
Active infection CVL Measure 2	Includes those with active HCV infection bv documentation of Qualitative VL tests, with or without quantitative VL tests. Imputation of VLs is required for those without quantitative VL tests. This measure provides programs with a fuller estimate of those who have active infection and who need fuller evaluation and treatment.
Hypothetical scenario	In a population of 10,000 in a drug treatment program where 60% have active HCV and 80% had qualitative HCV VL testing, but may or may not have had quantitative VL testing, the total number of individuals contributing to the HCV CVL would be 4,800 people
Diaanosis-based CVL Measure 3	Includes those with documented exposure to HCV bv virtue of havina been found to have a positive anti-HCV test; these people may or may not have had VL testing (qualitative or quantitative, or both). Imputation will be needed to estimate the proportion of those with positive anti-HCV tests who in fact have active infection. For those who did not have any VL tests, imputation will be needed to estimate the proportion who are viremic; this proportion can be reliably imputed informed by well-established data that 25% of those exposed will have spontaneously cleared the virus and thus will not have active infection and will not contribute to CVL. Further, quantitative VLs will need to be imputed for those who only had qualitative VL tests as well as for the proportion of those who did not have any VL tests but were anti-HCV positive and are projected to have active infection (i.e, to not have spontaneously cleared).
Hypothetical scenario	In a population of 10,000 in a drug treatment program where 80% have been exposed to HCV and 95% had antibody HCV testing, but may or may not have had quantitative VL testing, the total number of individuals contributing to the HCV CVL would be 7,600 people
Prevalence-based CVL Measure 4	Includes those with or without quantitative VLs, those with qualitative positive VLs but not quantitative test results, those with anti-HCV positive tests but without any VL tests; and for those without any documented anti-HCV testing. Imputation will need to be conducted, as described above for the diagnosis-based CVL, with the addition of imputation for the proportion who were not anti-HCV tested (i.e., undiagnosed cases) but are truly HCV infected. In settings where a large proportion have received antibody testing, there should not be a substantial difference between the prevalence- and diagnosis-based HCV CVL measures. This measure is analogous to the “population HIV VL” measure proposed in the CDC HIV CVL family of measures.5 Like the population VL, this measure could be valuable to public health authorities to monitor the population risk of transmission and the state of treatment engagement in this area. Further, it could be useful to public health authorities to construct CVLs for different areas and to identify disparities in CVL that would benefit from public health intervention.
Hypothetical scenario	In a population of 10,000 in a drug treatment program where 80% have been exposed to HCV, the total number of individuals contributing to the HCV CVL would be 8,000 people
Pooulation CVL Measure 5	Includes those with or without quantitative VLs, those with qualitative positive VLs but not quantitative test results, those with anti-HCV positive tests but without any VL tests; and for those without any documented anti-HCV testing as well as those who are HCV negative in the population. This uninfected group consisting of the following known and measurable groups: 1a) those testing anti-HCV negative, 1b) those testing anti-HCV positive and VL negative; and two unmeasured groups which require imputation: 2a) those anti-HCV positive and not VL tested who are truly VL negative, and 2b) those not tested for antibody but truly anti-HCV negative. Imputation will be needed as described above for the prevalence-based CVL; in addition, imputation will be need for groups 2a and 2b as above. Where data can be reliably estimated or measured directly, the population VL may be the best measure for assessing transmission and could be used by health departments, policy makers, or programs to plan of the allocation of resources to address and prevent infections and to predict trends in incidence at the population-level.
Hypothetical scenario	The population of 400,000 in an area where 10% have active HCV, the total number of individuals contributing to the HCV CVL would be 400,000 people

Engaged-in care HCV CVL measure:

The first and narrowest HCV CVL measure consists of people with active HCV infection who are engaged-in-care with directly measured HCV CVL data. This measure also would include those successfully treated who achieved SVRs and hence have undetectable VLs. Engaged-in-care VL estimates can be directly useful to programs by providing information on the degree to which the population in-care is virally suppressed. For HCV, where treatment durations are finite and result in CasP, CVL will generally consist of VL data contributed by those untreated.

Active infection HCV CVL measure:

The second HCV CVL measure includes those with a quantitative HCV VL result and people without a quantitative VL result but known to be viremic from a positive qualitative RNA test. This measure therefore may be directly useful to programs as a performance measure and to public health authorities in identifying groups engaged-in-care who remain capable of transmitting the virus and who need treatment.

Diagnosis-based HCV CVL measure:

Individuals who have documented anti-HCV positive tests with or without documented HCV VL tests comprise the third proposed HCV CVL measure. For this measure, in addition to using directly observed data from those who had quantitative VL tests, values need to be imputed based on assumptions for both those who had qualitative RNA but not quantitative VL tests and for those without any documented VL tests.

Prevalence-based HCV CVL measure:

The fourth proposed measure includes a broader population in that it also includes individuals with documented HCV antibody positive tests but without VL tests as well as those who remain undiagnosed. This measure could be used to monitor the potential for HCV transmission. If a two-step testing process is utilized, the diagnosis-based HCV CVL and the prevalence-based HCV CVL measures become identical.

Population HCV CVL measure:

This measure is comprised of the prevalence based CVL however it includes those who are HCV negative; this measure resembles the prevalence CVL employed for HIV by Solomon, et al.⁴¹ Recognizing that there are both infected and uninfected people in any setting or area, and that the proportion infected significantly affects transmission dynamics, this measure could be concretely useful to programs and policymakers as it has the potential to most fully reflect HCV transmission risk in an area or program.

4.0. Discussion

As an index to monitor the potential for viral transmission, reflecting the reservoir of infection capable of being transmitted within any area or group, CVL incorporates both directly measured and estimated (e.g., for those undiagnosed or those not in-care) data for the number of infected or viremic people and the magnitude of VL for each person.⁹ CVL can be used for area-level surveillance among populations linked by transmission-relevant factors (e.g., within PWID networks) or as a measure of program-level effectiveness. It is important that CVLs are measured in and applied to clearly defined populations, be able to reflect variations in CVL over time, and reflect transmission risks and the degrees of treatment engagement.⁹

An important consideration when constructing CVL measures for epidemiology and program evaluation is the need to consider whether the data are serial cross-sectional or longitudinal. Published studies on HIV CVL have varied in design including both single cross-sectional assessments,¹⁰ serial cross-sectional designs,²⁰ and longitudinal designs.⁷⁶ While having the limitation of not being able to account for individual-level variations in risk, such serial cross-sectional designs may have the advantage of being better able to detect changes in area-level factors over time (e.g., access to sterile syringes or HCV treatment).⁷⁷

There are some practical limitations to the use of CVL measures relating to issues of data availability, and the need for specific methods to handle missing and duplicate data, respectively, and to account for those who may have moved or died. Such methods are particularly important when combining data across programs and jurisdictions. Relevant approaches to doing so for HIV are discussed in the CDC document and in subsequent work.^10,41,42,78

There are the related issues of who or what constitutes a population and of what constitutes an appropriately defined area to which the CVL construct is applied, and whether the people therein are truly interconnected in ways that create the potential for transmission and the issue of the distribution of risk behaviors in any population so defined.^26,79,80 A related potential limitation of place-based inferences is that people spend different proportions of time in different places doing different things.⁸¹ An individual’s behavior may be influenced by, or in fact, constrained to different degrees by, the presence or absence of risk, prevention, or HCV services in their area of residence.

Further, if those with active infection are not engaging in transmission risk behaviors, then the magnitude of the VL in that individual does not truly contribute to transmission potential so that there is the potential introduction of bias in the exposure measurement. This consideration however also applies to analyses of transmission based on the prevalence of viremic persons where those who are viremic (regardless of the magnitude of viremia) only truly contribute to transmission risk if they engage in risk behaviors. Since risk behaviors may change over time, the population of all infected persons constitutes an important pool from which outbreaks can develop.^82–84 Studies have found that the prevalence of viremic HIV infected persons and the magnitude of HIV CVL are nonetheless associated with incident HIV infections.^15,18,20,39

HCV CVL measures may be useful to program and policy decision makers to evaluate the impact of and need for HCV prevention and treatment services. They may be useful to monitor area-level trends in HCV epidemiology including degrees of treatment implementation and engagement and a measure that reflects HCV transmission potential among susceptible populations. These measures could also be used in quantitative evaluations of any area- or program-level phenomena including the impact of the implementation of specific interventions (e.g., impact of HCV prevention and testing and CasP) on HCV incidence and prevalence. These measures can be used to identify area-level disparities in HCV CVL and to examine associations between HCV CVL and incidence. Further study of these proposed HCV CVL measures to HCV epidemiology is warranted.

Highlights.

• Novel metrics are needed to control the ongoing HCV epidemic.

• Community viral load measures have proven useful to HIV public health control efforts.

• Community viral load measures have not previously been applied to HCV epidemiology.

• 5 HCV community viral load measures of varying inclusiveness are proposed.

• Further study of these proposed HCV community viral load measures is warranted.