An estimated 3.2 million people in the US have chronic hepatitis C (HCV) infection, making the pathogen the most common chronic blood-borne infection nationwide [Armstrong et al., 2006]. Chronic HCV infection is a strong risk factor for the development of hepatocellular carcinoma (HCC) [Colombo, 1999; Di Bisceglie, 1997; El-Serag and Mason, 2000; Thomas and Seeff, 2005; Tsai et al., 1994] and cirrhosis of the liver [Thomas and Seeff, 2005]. The progression of chronic HCV infection to serious liver disease is associated with age at infection, male gender, immunosuppression, coinfection with HIV and hepatitis B virus (HBV), and heavy alcohol intake [Thomas and Seeff, 2005].

There has been considerable debate regarding differences in pathogenicity and hepatocarcinogenic potential of the various genotypes of HCV [e.g., Brechot, 1997; Bruno et al., 1997; Bruno et al., 2007; Fattovich et al., 2004; Romeo et al., 1996; Sangiovanni & Iavarone, 2008]. A recent meta-analysis of 21 age-corrected studies by Raimondi and colleagues (2009) found increased risk of HCC in patients infected with genotype 1b, with or without the presence of cirrhosis. Unfortunately, many of the studies included are from geographical regions where genotype 1b predominates [Zein, 2000], and there are few studies from HCV genotype 1a-predominant areas for comparison. Other studies have found similar associations with HCC risk for both 1a and 1b subtypes within genotype 1 [Lee et al., 2010; Tanaka & Hirohata, 1998; Zein et al., 1996; Zekri et al., 2000;], and even within genotypes 2a [Han et al., 1997], 3a [Khan et al., 2009], and 4, with no difference between subtypes [Ryu et al., 2009]. However, differences in relative risk reported in some studies may be due to regional differences in genotype prevalence [Haydon et al., 1997; Mangia, 1997; Reid, 1999; Widell, 2000] and/or a cohort effect reflecting confounding age-related and disease-progression variables [Di Tomasso et al., 2003; Lopez-Labrador et al., 1997; Sartori et al., 2006]. Thus while the association between HCV genotype and the severity of hepatic disease remains controversial, a majority of studies have found both subtypes within genotype 1 to confer heightened risk of histopathological damage, concomitant progression to HCC, and to a lesser extent, persistence of infection [Lehmann et al., 2004; Mehta et al., 2002; Strasfeld et al., 2003a; Thomas et al., 2000; Villano et al., 1999].

In addition, those infected with HCV genotype 1 have also shown a diminished sustained virologic response (SVR) to treatment with pegylated interferon-alpha (IFN-α) and ribavirin, and these patients typically require a longer duration and a greater dosage of combination therapy than do genotypes 2 and 3 [Manns et al., 2001; National Digestive Diseases Information Clearinghouse (NDDIC), 2006; Poynard et al., 1998; Scott and Perry, 2002; Shepherd et al., 2004; Zein et al., 1996]. However, the new class of direct-acting antivirals (DAAs) such as telaprevir and boceprevir have enabled a new “triple-therapy” standard-of-care combining the new agents with standard IFN-α and ribavirin therapy[Pawlotsky, 2011; Nelson, 2011]. Several studies have demonstrated improved outcomes and shortened treatment regimes amongst individuals infected with genotype 1 in both previously treated [Bacon et al., 2001; McHutchison, 2010] and untreated [Poordad et al., 2011] cases, with an increase in SVR from 50% up to 70% [Asselah et al., 2010] and even raising the possibility of interferon-free treatment options in the future [Dore et al., 2011]. However, as with ribavirin and IFN-α, DAA therapy is not without its own threat of adverse effects and resistance mutations [Fonseca-Coronado et al., 2011; Fried, 2011; Khalid & Bacon, 2011], and the high cost could present a barrier to uninsured and underinsured patients such as injection drug users (IDUs), who already tend to have significant barriers to HCV treatment [Grebely et al., 2008; Swan et al., 2010].

Despite earlier recommendations to avoid treating only former and recovering IDUs infected with HCV [National Institutes of Health (NIH), 1997], recent evidence suggests that standard HCV treatment can be successful among active IDUs [e.g. Dore et al., 2010; Hellard et al., 2009; Manolakopoulos et al., 2010; Papadopoulos et al., 2010; Schulte et al., 2010;]. However, in addition to HCV genotype, HCV viral concentration can also impact treatment success. For example, previous evidence has suggested that patients with high HCV RNA concentrations are less responsive to pegylated interferon-alpha (IFN-α) and ribavirin combination therapy [Lam, 1999] and to IFN-α and ribavirin combination therapy [Munir et al., 2010]. Other studies have indicated that patients with high RNA levels, particularly those who are non-responders to other treatment regimens, may be more responsive to consensus interferon [Alberti, 1999; Witthöft, 2008; Witthöft et al., 2007], though there does appear to be some variability in cost effectiveness depending on the intensity of treatment [Sjogren et al., 2007]. The NDDIC suggests that a HCV viral concentration of less than 800,000 IU/mL can determine a better response to IFN-α and ribavirin treatments [NDDIC, 2006].

In addition to the clinical and epidemiological implications, HCV is associated with great economic burden in terms of increased absence from job-related activities and productivity, as well as increased healthcare service utilization and healthcare costs [Davis et al., 2011; DiBonaventura et al., 2010; Su et al., 2010]. In fact, Su and colleagues (2010) found that healthcare-related costs for HCV-infected individuals can be over $8000 per year more than that for non-infected individuals. A study by Grant and colleagues (2005) examining trends in healthcare service utilization for HCV infection in the US found that costs increase considerably with the aging of the HCV-infected population and with HCV disease progression. At this juncture, the authors note that, “as patients continue to age and disease burden progresses, suboptimal decisions regarding HCV treatments will bring increasing opportunity costs for the health care system and society” [Grant et al., 2005, p. 1406]. Moreover, a growing body of evidence suggests that decision-making regarding HCV treatment and its cost-effectiveness will be influenced by HCV RNA concentration and genotype. Higher HCV viral load and infection with genotype 1 have negative impacts on the cost-effectiveness of various treatments regimens [Hartwell et al., 2011; Salomon et al., 2003; Shepherd et al., 2004; Sullivan et al., 2004; Wong et al., 1998; Yeh et al., 2007] and of early intervention strategies [Shepherd et al., 2007]. The issue is complicated further in the active IDU population, in which reinfection following treatment is a serious concern, although the frequency of reinfection in IDUs following SVR to drug therapy is low [Grebely et al., 2010]. Furthermore, using mathematical modeling, Martin and colleagues (2011) have shown that treating active IDUs is more cost-effective than leaving them untreated, provided the prevalence of chronic HCV infection is under 60%.

Substantial geographic and demographic variation can exist in the distribution of HCV genotypes [Blatt et al., 2000; Zein and Rakela, 1996]. HCV genotypes 1a/1b, and to a lesser degree 2b and 3a, are predominant in the US population [Nainan et al., 2006; Zein, 2000; Zein and Rakela, 1996], and HCV genotypes and subtypes within the country vary according to ethnicity and age. In a nationally representative sample, increased age and non-Hispanic black race/ethnicity were associated with an increased likelihood of infection with HCV genotype 1b [Nainan et al., 2006]. Research examining HCV genotype distribution has been used for various purposes, including the identification of populations’ origin of HCV infection [Panduro et al., 2010; Spada et al., 2008], examination of historical trends in characteristics of HCV infection [Magiorkinis et al., 2009], and making inferences regarding routes of HCV transmission [van den Berg et al., 2009; Zein, 2000; Zhou et al., 2009]. In a review of the HCV literature on drug users published between 1989 and 2006, Stern et al. (2008) identified 123 studies which reported genotype distribution [Stern et al., 2008], and in the past two to three years, there has been growing recognition of the value of descriptive studies examining HCV genotype distribution in drug using populations [e.g. Lee et al.; Liao et al., 2011; Micalessi et al., 2008; Sereno et al., 2009; Viazov et al., 2010; Zhou et al., 2009]. However, nearly all of the existing studies to date have been conducted either outside of the U.S. or in urban populations.

In light of the epidemiologic, clinical, and economic implications of HCV genotypes and their uneven dispersion within certain populations, more research is needed to elucidate the genotype distribution and prevalence of HCV viremia in high-risk, HCV-exposed groups from under-resourced settings in the US. As of 2007, Kentucky had the 7^th highest HCV incidence rate in the nation [Daniels et al., 2009], yet the HCV burden has been largely uncharacterized. Given the implications of HCV viremia and genotype on treatment cost-effectiveness, this gap in knowledge can present a major barrier to the planning of HCV intervention strategies in the state’s economically distressed region of Appalachia [Appalachian Regional Commission (ARC), 2011], where cost-effectiveness is a paramount consideration in healthcare system decision-making. Efforts to understand the HCV burden in Appalachia are made more relevant by the region’s distinguishing burden of drug use. Unlike other populations that have been the focus of other HCV-related studies on drug users, Appalachia drug users are predominantly abusing prescription drugs [Drug Enforcement Agency, 2002; Havens et al., 2007; Young et al., in press], and have distinctive profiles of HCV-related drug risk behavior [Havens et al., 2007; Young and Havens, 2011; Young et al., 2010].

There is emerging evidence that demonstrates that IDUs in Appalachia are indeed much different than IDUs in other populations in terms of their risk behaviors and drug use characteristics. In fact, studies have demonstrated that drug users in Appalachia have substantially different trajectories of drug use than residents of urban areas in Kentucky, with significantly earlier ages of onset for use of oxycodone, hydrocodone, benzodiazepines, cocaine, and crack. Adjusting for age, gender, and race, the rural drug users had higher odds of lifetime and recent nonmedical prescription opioid use (specifically methadone, OxyContin^®, and oxycodone) and had higher odds of lifetime cocaine and crack use [Young et al., in press]; the latter of which have been identified to be associated with HCV RNA concentration [Operskalski et al., 2008]. Evidence also demonstrates that compared an urban sample of drug users in the state, rural Appalachian drug users are more likely to use alternative routes of administration (e.g. snorting and injecting) for many drugs. In fact, among the Appalachian participants in one study, snorting was the most frequent route of administration for hydrocodone, methadone, OxyContin^®, and oxycodone, and injection was most common for hydromorphone and morphine [Young et al., 2010]. These findings were corroborated by another study which demonstrated that prescription opiate misuse and injection drug use was higher among rural Appalachian probationers compared to their urban counterparts [Havens et al., 2007]. The prevalence of use of alternative routes of administration by drug users in the Appalachian region increases their risk for acquiring HCV and underscores the importance of examining HCV in the population.

A study based on recent retrospective survival analyses of drug users in Appalachia found that this population differs drastically from those involved in other studies of IDUs, and that nonmedical prescription opioid use may be the differentiating factor. OxyContin^® was involved in nearly as many initiations to injection (48%) as were stimulants, other prescription opioids, and heroin combined. The rate from which users transitioned from first OxyContin^® use to initiation of injection with OxyContin^® was extremely fast, just 3 years on average. The sample’s nonmedical prescription drug use played a significant role in their hazard for initiating injection, with five prescription drugs (benzodiazepines, illicit methadone, oxycodone, OxyContin^® and other opiates) associated with an increased hazard for transitioning, adjusting for demographics characteristics [Young and Havens, in press].

Historically and culturally, there are also reasons to believe that the epidemiological characteristics of HCV in Appalachia could differ from that of other regions in the US. For example, molecular evidence has demonstrated that HCV type 1 was likely first introduced into the US in approximately 1909 and more widely disseminated throughout the US population in the 1960’s [Mizokami et al., 2006; Tanaka et al., 2002]. During the diffusion of HCV in the 1960’s, the Appalachian region was becoming increasingly recognized as being heavily isolated and distinct from other populations in the US, though efforts were underway to “integrate” and “modernize” its residents [Photiadis and Schwarzweller, 1970; Whisnant, 1980]. The degree to which Appalachian residents were isolated from larger society has been debated [Lewis and Billings, 1997], and the extent to which the isolation continues is not well-established and likely highly variable across communities. Nevertheless, there continues to be sociological and demographic evidence that suggests the region experiences extensive out-migration, but little in-migration of new residents [ARC, 2011; Bell, 2009]. The role that Appalachia’s regional isolation and its unique patterns of drug use play in influencing the profile of HCV are speculative, but certainly cause for further investigation. Thus, the purpose of this exploratory study was to (1) assess the prevalence of HCV viremia (i.e. HCV RNA positivity), (2) examine quantitatively the HCV RNA viral concentration, and (3) determine the HCV genotype distribution among a random sample of anti-HCV positive drug users in a rural Appalachian region of Kentucky. Behavioral and demographic correlates to HCV RNA status and genotype were also assessed.

Among a random sample of anti-HCV positive prescription drug users drawn from a longitudinal cohort study in Appalachia, the majority (69%) had active HCV infection. HCV RNA positive participants were more likely to have a history of blood transfusion and more likely to have injected drugs in the past 6 months and past 30 days. Significantly more participants who were HCV RNA positive were white and uninsured than those who were HCV RNA negative. HCV genotype 1a was detected in 66% of the HCV RNA positive sample, while types 2b (16%) and 3a (13%) were less common.

The sample’s genotype distribution and RNA status have clinical implications. The most prevalent HCV genotype in the sample, type 1, is less responsive to treatment, requires a longer duration and higher dosage of combination therapy [Manns et al., 2001; NDDIC, 2006; Poynard et al., 1998; Scott and Perry, 2002; Zein and Rakela, 1996], and has potential association with higher RNA levels and decreased likelihood of spontaneous clearance of the virus [Berger et al., 1996; Kobayashi et al., 1996]. Also, in the majority of studies HCV genotypes 1a and 1b seem to be more strongly associated with HCC than genotypes 2 and 3 [Raimondi et al., 2009; Tanaka & Hirohata, 1998; Zein et al., 1996;]. The proportion of HCV-positive individuals who would require a long duration of treatment could present an especially difficult situation in rural Appalachia, given marked health and economic disparities and barriers in access to care. The sample’s high RNA concentration may also have implications for treatment [NDDIC, 2006] and prognosis [Hisada et al., 2005]. Ability to extrapolate from the RNA concentration data is limited by the study’s cross-sectional design, but it warrants mention that the RNA concentration for 59% of the sample exceeded the range which usually determines a better response to peginterferon and ribavirin treatments (< 800,000 IU/mL) [NDDIC, 2006]. However, new direct acting antivirals boceprevir and telaprevir may mitigate some of the efficacy lost due to heightened viral load in genotype 1-infected individuals [Asselah, 2010; Dore et al., 2011; Nelson, 2011].

The median viral concentration in this sample is nearly double that found other recent studies of injection and non-IDUs [Schulte et al., 2010; Strasfeld et al., 2003a], similar to that found in others [Boodram et al., 2011; Hisada et al., 2005], and lower than that found in HIV co-infected cohorts [Obermeier et al., 2011]. Given the prevalence of high risk behavior in this sample, the findings regarding elevated RNA concentration are not unexpected; previous research has found that HCV viral concentration is positively associated with risky drug use behaviors, including lifetime cocaine, marijuana, amphetamine, or heroin use [Operskalski et al., 2008] and frequency of sharing unclean syringes [Boodram et al., 2011]. The high viral concentration is concerning not only given participants’ continued syringe sharing, but also given the prevalence of straw sharing (96%). There is limited evidence to suggest that drug users with high HCV viral loads can have detectable levels of HCV RNA in nasal secretions, which could have consequences for HCV transmissibility through the sharing of straws [McMahon et al., 2004] and other drug preparation equipment [Hagan et al., 2001].

The serological profile of this sample of Appalachian drug users, in some respects, mirrors that of the US general population, and in other ways, is unique. The proportion of anti-HCV positive individuals who tested positive for HCV RNA in our study (69%) was less than that found in a nationally-representative sample (78%) [Nainan et al., 2006]. However, the samples were similar in terms of HCV genotype. Among white, non-Hispanic individuals in the national study, the weighted percentage of HCV RNA positive individuals infected with genotype 1 was 70%, comparable to the prevalence of 66% reported in this study [Nainan et al., 2006]. Similarly, the prevalence of genotypes 1, 2, and 3 in this study were comparable to that found in samples from a large cohort of chronic HCV patients [Blatt et al., 2000]. In terms of subtype, the present study found a higher percentage of individuals infected with type 1a than did the national study or large clinic-based study (66% versus 52% and 39%, respectively) and a lower percentage infected with type 1b (0% versus 27% and 30%, respectively) [Blatt et al., 2000; Nainan et al., 2006]. In the present study, participants with HCV genotype 1 were more likely to have graduated from high school than participants with genotypes 2 and 3, though the difference was marginally significant (p=.04) and the confidence interval was wide. This association between education and genotype remained statistically significant even after controlling for age. The study’s finding that HCV genotype 1 was associated with higher educational status is curious. To explore explanations for the association of education and genotype, demographic and behavioral correlates to education were examined. Educational status was positively associated with age (p=.018), but not associated with any other demographic or behavioral variables (data not shown). When analysis of the association between education and genotype was adjusted for age, the association between education and genotype remained statistically significant (OR: 3.87, 95% CI: 1.03 – 14.53). To our knowledge, no other study to date has reported an association between education and genotype, thus further research is necessary to verify and explore this potential association.

The disparity in the subtype findings may be attributed to differences in racial composition of the samples, as type 1b is more common among non-Hispanic blacks. The differences may also be attributable to the likely infection route for the source populations. Genotypes 1a and 3a are more common among those with a history of injection drug use [Serra et al., 2003; Zein, 2000; Zeuzem et al., 1996], a behavior which characterizes 95% of our sample. Some research has suggested that genotype may be correlated with year of infection [Chlabicz et al., 2008; Pybus et al., 2001; Zeuzem et al., 1996; Zhou et al., 2009], or correspondingly, with the year of the genotype’s introduction into the population [Gérard et al., 2005; Katsoulidou et al., 2006]. The vast majority of participants began injecting drugs within the last 15 years, over half between 1995 and 2005, and over a quarter since 2005 (data not shown). The exploratory nature of this study precludes our ability to make inference regarding the mechanisms that produced the identified genotype distribution. However, further research could explore whether the relatively recent initiation of drug-related risk behavior within the sample corresponds with the national, temporal shift in prevalence from genotype 1b to 1a [Nainan et al., 2006].

This study not only contributes to epidemiologic knowledge about HCV in rural Appalachia, but also offers insight for intervention. An alarming proportion of individuals with active HCV infection were engaging in risk behaviors which could promulgate the infection through the study population. Among those with active infection, one in seven had engaged in recent distributive needle sharing (i.e. had sold, returned, or given away a needle after using it). On the flipside, among individuals without active infection (i.e. HCV RNA negative), nearly one in eight had engaged in recent receptive needle sharing. This population’s involvement in risk behavior is likely not attributable to inadequate knowledge about HCV transmission. In fact, 100% of the sample was aware that HCV could be transmitted through sharing needles and syringes and 99% was aware HCV could be transmitted through sharing other injection equipment.

These findings demonstrate a need for behavioral, clinical, and structural interventions to prevent secondary transmission and increase access to treatment. Intervention may be particularly difficult in rural Appalachia due to a lack of resources. Previous research has shown that syringe exchange programs (SEPs) can substantially decrease risk behavior [Des Jarlais et al., 2005; NIH, 1997, 2002], HCV prevalence [van den Berg et al., 2007], and HCV incidence when coupled with drug treatment [Des Jarlais et al., 2009]. SEPs also provide a strategy for providing drug users with condoms, referrals to substance abuse treatment, and HCV testing and counseling [NIH, 1997, 2002]. Nationally, the implementation of SEPs is difficult due to legal and regulatory restrictions; Kentucky is no exception. While Kentucky does not directly prohibit the establishment of SEPs, statutes regulate the sale and disposal of hypodermic syringes and needles and criminalize the possession of drug paraphernalia [Kentucky Legislative Research Commission (KLRC), 1992; KLRC, 2005]. In addition to regulatory barriers, economic circumstances may impede the implementation of a SEP in the region. According to the Centers for Disease Control and Prevention (CDC), 28% of existing SEPs operate on a budget of less than $25,000, 37% operate on a budget between $25,000 and $100,000, and 35% on budgets greater than $100,000 [Centers for Disease Control and Prevention (CDC), 2005]. Given the impact of chronic HCV infection on quality of life [Bernstein et al., 2002; Foster et al., 1998] and healthcare costs [Davis et al., 2011; DiBonaventura et al., 2010; Su et al., 2010], the cost-effectiveness of a SEP in the region is worth consideration. In lieu of SEPs, behavioral and clinical interventions may be applicable. Peer-driven and other behavioral interventions could also be considered as a way to reduce HCV risk behavior in this population [Garfein et al., 2007; Latka et al., 2008; Sacks-Davis et al., 2011]. Programs and funding to improve residents’ access to health care and to drug treatment also could help to prevent and/or mitigate the long-term consequences of HCV infection. Clinical interventions, however, must be coupled with strategies which address structural barriers, such as lack of transportation.

While the study provides foresight into possibilities for intervention in this population, it is not without limitations. The study’s reliance on self-report subjects the data to potential information bias, which if present, would likely bias the results toward the null hypothesis and would make the reported findings more conservative than if no bias existed. Like other recent descriptive studies regarding HCV genotype distribution [e.g. Lee et al., 2010; Liao et al., 2011; Micalessi et al., 2008; Sereno et al., 2009; Viazov et al., 2010; Zhou et al., 2009], the study’s sample size and cross-sectional design present limitations. Though the sample was drawn randomly from and was representative of a larger pool of HCV antibody positive drug users, the sample’s size and homogeny in terms of behavioral and demographic characteristics limits the power to distinguish correlates to HCV RNA positivity and genotype. The cross-sectional design also precludes the ability to make causal inferences and to draw conclusions about dynamics of fluctuation in viral concentration, as well as reinfection, coinfection, and superinfection. The inability of the study to provide more detailed phylogenetic analysis also limits inference about the dynamics of transmission.

Despite limitations, this study provides important insights into a high-risk, under-studied, rural population. The prevalence of active HCV infection and distribution of certain genotypes in this sample of HCV-exposed drug users highlight a need for primary and secondary prevention strategies as well as interventions aimed at increasing access to appropriate treatment. Though more research is needed to determine the underpinning factors associated with HCV infection and subtype distribution within this population, it is clear that a combination of structural, behavioral, and clinical intervention is necessary to mitigate the consequences of HCV in this region with marked health disparities.