Primary Care-Based Hepatitis C Treatment Outcomes With First-Generation Direct-Acting Agents

Abstract

Objectives

Vulnerable, urban populations with a history of substance use disorders have a high prevalence of hepatitis C virus (HCV). Primary care-based treatment has been proposed to improve access to care. In this study, we present outcomes from our urban, primary care-based HCV treatment program in patients treated with telaprevir or boceprevir in combination with pegylated-interferon and ribavirin (‘‘triple therapy’’).

Methods

We collected data from 126 consecutive patients with genotype 1 HCV monoinfection seen in our treatment program (2011–2013). Among the 40 who initiated treatment, we analyzed factors associated with achieving a sustained viral response (SVR).

Results

During the study period, 40 patients initiated triple therapy (32%), 80% with recent or past substance use disorders. Patients initiating treatment were younger than untreated patients (P = 0.002), but otherwise did not differ demographically, or in the severity of their liver fibrosis (P >0.05). An SVR was achieved in 18 patients (45%) and was less likely in patients with recent or past substance use disorders or psychiatric illness (both P <0.01).

Conclusions

Nearly one third of patients initiated triple therapy with SVR rates comparable to other HCV treatment settings, despite a significant burden of mental illness and substance dependence. Our experience demonstrates that a primary care-based practice can successfully deliver HCV care to a vulnerable population. Additional interventions may be needed to improve outcomes in patients with recent or past substance use disorders or psychiatric illness.

Between 3 and 4 million people are estimated to have chronic hepatitis C virus (HCV) infection in the United States (Holmberg et al., 2013b), where HCV is now the leading indication for orthotopic liver transplantation (Kim et al., 2014). Chronic HCV infection increases all-cause mortality (Moorman et al., 2013), and treatment has been shown to be cost-effective (Liu et al., 2012; Blazquez-Perez et al., 2013). Despite this, the small supply of clinicians providing HCV care limits access to evaluation and treatment (Morrill et al., 2005; McGowan et al., 2013).

The majority of chronic HCV infections occur in under-served, urban populations in the United States–populations who also experience many barriers to medical care and to the initiation and completion of antiviral treatment for HCV (Rousseau et al., 2008). Previous studies have demonstrated marked treatment disparities among these members of the HCV-infected population, with minority patients experiencing less frequent diagnostic assessment and treatment (Rousseau et al., 2008). Successful treatment completion has been challenging in all patient groups, as it requires consistent adherence to a complicated regimen for a lengthy period.

The standard-of-care for treatment of chronic genotype 1 (GT1) HCV, the most prevalent in the United States, changed in 2011 with the introduction of the protease inhibitors telaprevir and boceprevir. In clinical trials, these direct-acting agents (DAAs) demonstrated higher rates of sustained viral response (SVR) compared with the previous standard-of-care, pegylated interferon and ribavirin, but also contributed additional toxicities to the burdensome side-effect profile of pegylated interferon and ribavirin, with increased risk of hematologic abnormalities and rash (Chou et al., 2013). Addition of the new agents also increased the dosing frequency and pill burden of the regimen.

Numerous barriers prevent patients with chronically monoinfected HCV from initiating and completing antiviral therapy. Fear of treatment-related toxicities (Evon et al., 2010), high rates of medical and psychiatric comorbidities (Evon et al., 2007), current and past substance use disorders (Grebely et al., 2009), and the limited availability of HCV specialists (Rousseau et al., 2008)— all contribute to low rates of HCV treatment initiation, and serve as barriers to the successful completion of treatment. Several innovative approaches have been proposed to mitigate these barriers, including HCV-specific case management, telemedicine, and group treatment (Bruggmann, 2012).

The limited supply of HCV specialists, as well as the complex comorbid disease profile of patients in need of HCV care, necessitates new care delivery approaches. Expanding primary care-based HCV treatment has been proposed as one method of increasing the HCV clinician pool, which may provide additional benefits for some HCV-infected patients who require clinicians experienced in the management of comorbid chronic conditions (Arora et al., 2011). To be effective, primary care-based treatment programs must demonstrate noninferiority, providing comparable treatment initiation and success rates to specialty-based care. Therefore, in this study, we evaluate the rates of treatment initiation and treatment outcomes associated with the first generation of HCV triple therapy in our novel, urban, primary care-based HCV treatment program.

METHODS

Program Description

The Primary Care Liver Clinic (PCLC) at the Icahn School of Medicine at Mount Sinai in New York City was established in 2002 to improve access to HCV monoinfection treatment for the urban, underserved patient population of our internal medicine clinic (www.mountsinai.org/PrimaryCar-eHCV). The PCLC is housed within this primary care practice, which serves more than 12,000 patients who are mostly residents of East and Central Harlem and the South Bronx. Most referrals to the PCLC are made by primary care practitioners in the internal medicine practice, although some patients are referred by outside clinicians and community outreach programs. Patients with evidence of advanced cirrhosis or liver disease of non-HCV (including hepatitis B coinfection) etiologies are referred to a hepatology clinic, and patients with HIV/HCV coinfection are referred to an infectious disease specialty clinic.

The PCLC uses a multidisciplinary approach, as this has shown benefit in other programs for treating patients with HCV who might otherwise be considered ‘‘poor treatment candidates’’ (Ho et al., 2013). The foundation of our evaluation and treatment model is aggressive case management by HCV-specific patient navigators. These patient navigators support all facets of the evaluation and treatment process, including patient visit scheduling and reminder calls, facilitation of tests and consultations, navigation during the clinic session, telephone support, adherence monitoring, and prior authorization. The model includes fully integrated mental health and social work services. Patient navigators also help coordinate care efforts of other members of the treatment team, which includes primary care physicians and a nurse practitioner, who work in a colocated fashion with a clinical psychologist, a nutritionist, and a social worker. The treatment program includes a protocol-based pretreatment evaluation of the patients’ health status, including noninvasive assessment of the degree of liver fibrosis (protocol available by request from the corresponding author), as well as a web-based structured assessment of psychosocial readiness for HCV treatment (Psychosocial Readiness Evaluation and Preparation for hepatitis C treatment; www.prepc.org). A unified care plan is developed for each patient at case conferences with input from all multidisciplinary team members. Patients initiated on antiviral therapy were treated according to a standard treatment protocol developed by our program, on the basis of response-guided therapy as described in the DAA drug labeling. A component of the treatment protocol included frequent assessments for drug toxicity and medication adherence. Our clinic also served as a site for the Hepatitis C Assistance Program (HepCAP), a New York state-funded program providing financial support for HCVevaluation and treatment for uninsured patients.

Study Cohort

We used an electronic treatment-tracking database to identify all patients with HCV GT1 who were viremic and were evaluated for treatment in our clinic, starting from the date of our first use of triple therapy, September 20, 2011, through April 1, 2013. Patients were considered medically ineligible for HCV treatment in the PCLC if they had any of the following: advanced cirrhosis at high risk of decompensation (Child-Pugh class B or C), more than 69 years of age, severe cardiac disease, significant chronic renal insufficiency, or any end-stage or terminal illness. These patients were referred to a hepatology specialty clinic for further evaluation. If patients had modifiable medical issues that prevented initiation of HCV treatment, such as poorly controlled diabetes, we intervened and continued to follow up those patients and considered them eligible for treatment. From our treatment-tracking database, we identified 126 viremic GT1 patients who were medically eligible for treatment during the study period (Fig. 1).

FIGURE 1.

Flowchart for study cohort.

Data Collection

We extracted data from electronic medical records on all 126 patients including demographics, primary language, insurance type, HCV treatment history, laboratory values, presence of diabetes, a history of clinical depression, anxiety disorder, psychotic disorders, recent and past (>6 months prior) substance use disorders, and HCV evaluation outcome. All patients evaluated in the PCLC underwent a comprehensive initial visit with a corresponding electronic medical record template, prompting the collection of all of these data points. To ascertain recent or past psychiatric illnesses from the medical record, we identified past use of psychiatric medications or relevant diagnostic codes to establish diagnoses. These diagnoses were then confirmed by a clinical psychologist. Using laboratory values for aspartate amino-transferase, alanine aminotransferase, and platelet count at the time of initial evaluation in our clinic, we calculated a fibrosis index based on the 4 factors (FIB-4) score for each patient to noninvasively assess the degree of liver fibrosis (Holmberg et al., 2013a). For patients who initiated HCV treatment, we also collected data on previous HCV treatment use as well as the specific DAA used in the treatment course. We also collected data on patients’ treatment course and outcome. Our primary treatment outcome, SVR, was assessed in all patients and was defined as an undetectable HCV viral load 6 months or more after completing antiviral therapy. Additional non-SVR treatment outcomes were categorized as viral breakthrough or nonresponse (as defined as failure to achieve viral undetectability or return of viremia after undetectability while on treatment), viral relapse (after treatment completion), failure because of poor adherence to antiviral medications or loss to follow-up, or discontinuation because of drug toxicity.

Data Analysis

We compared baseline characteristics of the cohort by treatment initiation status using the t test for normally distributed continuous variables and the χ² test for categorical variables, as appropriate. We then assessed predictors of SVR in our primary care treatment cohort using a multivariable logistic regression model that included demographics (age, sex, and race/ethnicity), FIB-4 score, and the presence of psychiatric illness. All analyses were performed with STATA 10 (Stata Corporation, College Station, TX). This study was approved by the Institutional Review Board of the Icahn School of Medicine at Mount Sinai (GCO# 10-0032).

RESULTS

During the study period, 126 patients with GT1 were evaluated and considered potentially eligible for HCV treatment. Our cohort was largely African American and Hispanic (>84%) and had a high prevalence of previous substance use disorders (78%; Table 1) and significant psychiatric illness (46%). During the study period, 40 patients (32%) in the PCLC treatment program initiated triple therapy (Table 1). Patients who initiated treatment were younger than those who were not started on treatment (P = 0.002) and were more likely to be uninsured (P = 0.01), but otherwise did not differ demographically (all other comparisons P >0.05). There was also no difference in the prevalence of psychiatric illness (depression, anxiety, and/or psychotic disorders) or substance use disorders when comparing patients who were initiated on treatment to those who were not (all P values >0.1). We estimated the stage of liver fibrosis for patients in the cohort, using the FIB-4 index, and 17% had scores suggestive of advanced fibrosis; however, the distribution of FIB-4 scores did not differ between those who initiated treatment and those who did not.

TABLE 1.

Characteristics of All Genotype 1 Chronic Hepatitis C Patients Evaluated During Triple Therapy Era Eligible for Treatment

Characteristic	Eligible for HCV Treatment n = 126	Not Initiated on HCV Treatment n = 86	Initiated on HCV Treatment n = 40	P
Age, y, mean (SD)	52 (11)	54 (10)	48 (12)	0.002
Female, n (%)	55 (44)	36 (42)	19 (49)	0.3
Race/ethnicity, n (%)
White	11 (9)	7 (8)	5 (13)	0.9
African American	56 (45)	38 (44)	18 (45)
Hispanic	49 (39)	35 (41)	14 (35)
Other	9 (7)	6 (7)	3 (8)
Primary language, n (%)
English	110 (89)	76 (89)	35 (88)	0.7
Spanish	10 (8)	7 (8)	3 (7)
Other	4 (3)	2 (2)	2 (5)
Insurance type, n (%)
Medicaid	87 (70)	59 (69)	28 (72)	0.01
Medicare	15 (12)	15 (17)	0 (0)
Private insurance	12 (10)	7 (8)	6 (15)
Uninsured	11 (8)	5 (6)	6 (15)
FIB-4 score, n (%)
0–1.25	38 (30)	26 (30)	13 (33)	0.9
1.25–3.25	66 (53)	46 (54)	20 (51)
>3.25	21 (17)	14 (16)	7 (18)
Diabetes, n (%)	27 (12)	22 (26)	5 (13)	0.1
Any psychiatric illness, n (%)	57 (46)	38 (44)	19 (49)	0.6
Depression	48 (39)	33 (38)	15 (38)	0.9
Anxiety	10 (8)	8 (9)	2 (5)	0.4
Psychotic disorder	17 (14)	9 (11)	8 (20)	0.2
Substance use disorder ever, n (%)	97 (78)	65 (76)	32 (80)	0.6
Substance use in last 6 mos, n (%)	30 (24)	22 (26)	8 (20)	0.5
Reason treatment not initiated, n (%)
Still being evaluated for treatment		34 (40)
Patient unwilling to risk treatment toxicity		5 (6)
Patient with personal or work commitments		3 (4)
Patient wants to wait for next-generation therapy		9 (11)
Harms exceed benefits with current generation of treatments; waiting for next-generation drugs		35 (41)

FIB-4, fibrosis index based on the 4 factors; HCV, hepatitis C virus; SD, standard deviation.

At the end of follow-up, all 40 patients had either completed treatment and reached an SVR timepoint or discontinued treatment. For the patients who initiated treatment, we assigned mutually exclusive treatment outcomes. Of these 40 patients, 18 (45%) achieved SVR at 24 weeks after treatment termination and 2 (5%) patients had virologic relapse (Table 2). Significant treatment toxicity led 10 (25%) patients to discontinue therapy early. Viral breakthrough or nonresponse led 6 (16%) patients to discontinue therapy early. Four patients (10%) were lost to follow-up or nonadherent to treatment.

TABLE 2.

Treatment Outcomes for Patients Initiated on Triple HCV Therapy

Treatment Outcome	Primary Care-Based Practice n = 40
Sustained viral response, n (%)	18 (45)
Viral relapse after treatment completion, n (%)	2 (5)
Viral breakthrough or nonresponse, n (%)	6 (16)
Treatment stopped for toxicity, n (%)	10 (25)
Lost to follow-up or nonadherent, n (%)	4 (10)

We evaluated the characteristics of patients who achieved SVR in our PCLC treatment program to identify predictors of treatment success (Table 3). The only difference in demographics when comparing patients with and without SVR was that all 5 patients in our program who had a primary language other than English achieved SVR. No prior substance use disorder and a lack of a history of psychiatric illness were both significantly associated with SVR (both P <0.01). However, none of these predictors maintained significance after adjusting for demographic covariates.

TABLE 3.

Characteristics of Patients With All Genotype 1 Chronic Hepatitis C Initiated on Treatment by Presence of Posttreatment Viral Suppression

Characteristic	No-SVR n = 22	SVR n = 18	P
Age, y, mean, (SD)	51 (11)	45 (13)	0.1
Female, n (%)	10 (46)	9 (53)	0.6
Race/ethnicity, n (%)
White	3 (14)	2 (11)	0.5
African American	12 (55)	6 (33)
Hispanic	6 (27)	8 (45)
Other	1 (5)	2 (11)
Health coverage, n (%)
Medicaid	18 (82)	10 (56)	0.1
Private insurance	3 (14)	3 (17)
No insurance	1 (5)	5 (28)
Primary language, n (%)
English	22 (100)	13 (72)	0.04
Spanish	0 (0)	3 (17)
Other	0 (0)	2 (11)
FIB-4 score, n (%)
0–1.25	5 (23)	8 (44)	0.3
1.25–3.25	12 (55)	8 (44)
>3.25	5 (23)	2 (11)
Diabetes, n (%)	4 (18)	1 (6)	0.2
Previously treated, n (%)	2 (9)	2 (11)	0.9
Any psychiatric illness, n (%)	15 (68)	4 (22)	0.004
Depression	12 (55)	3 (17)	0.01
Anxiety	2 (9)	0 (0)	0.2
Psychotic disorder	6 (29)	2 (11)	0.2
Substance use disorder ever, n (%)	21 (95)	11 (61)	0.007
Substance use in prior 6 mos, n (%)	5 (23)	3 (17)	0.6
Protease inhibitor used, n (%)
Telaprevir	16 (73)	14 (78)	0.7
Boceprevir	6 (27)	4 (22)

FIB-4, fibrosis index based on the 4 factors; SD, standard deviation; SVR, sustained viral response.

DISCUSSION

Expansion of primary care-based HCV programs has been proposed as a strategy for improving access to HCV treatment. Studies from the dual-therapy era have shown that primary care-based HCV treatment can be an effective means of safely delivering treatment (Hill et al., 2008; John-Baptiste et al., 2009). This study provides the first data regarding treatment initiation and outcomes for first-generation direct-acting HCV treatments in a primary care setting. We found that a multidisciplinary HCV treatment program staffed by primary care practitioners had high treatment initiation frequency (compared to historical values) and treatment outcomes similar to those published from other ‘‘real-world’’ settings, including community-based specialty practices, Veterans Administration hepatology clinics, and academic specialty practices (Bichoupan et al., 2014; Ioannou et al., 2014; Sterling et al., 2015).

Estimates of HCV treatment initiation proportions in monoinfected patients vary widely, and are affected by the characteristics of the clinical populations studied. Low treatment efficacy and poor tolerability contributed to infrequent treatment initiation in the dual-therapy era. Community-based data from Canada demonstrated that less than 1.1% of potentially treatment-eligible HCV patients initiated treatment in the early 2000 s. Studies of patients engaged in HCV care have demonstrated a higher frequency of treatment initiation, with dual-therapy era treatment initiation among ‘‘eligible’’ patients ranging from 19% (Charlebois et al., 2012; Chen et al., 2013) to 37% (Clark et al., 2012). Comorbid mental illnesses seem to influence HCV treatment initiation, as patient cohorts with high prevalence of mental illness and substance use disorders have reported lower HCV treatment initiation (Sockalingam et al., 2013). We observed more frequent treatment initiation among eligible HCV patients in our primary care-based treatment program than in previous studies of similar populations. It is possible that this was influenced by the multidisciplinary care offered in the primary care setting. One study evaluating the integration of HCV treatment into primary care in HIV coinfected patients found that treatment initiation increased from 16% to 24% after implementing a primary care-based treatment model (Cachay et al., 2013). It is also possible that the increased frequency of treatment initiation, we observed in our clinic population, was related to the advent of triple therapy. Higher SVR incidence reported in the registration trials of the first-generation DAAs may have served as an additional incentive for patients to start treatment. Last, we observed higher treatment initiation rates among uninsured patients in our study cohort. We hypothesize that this finding was due to our inclusion of patients from the HepCAP, who were highly motivated to initiate HCV treatment.

The SVR frequency in our study (45%) of patients with GT1 HCV treated with triple therapy was much lower than frequencies observed in registration trials (Chou et al., 2013). There were marked differences in the patient demographics of our patient population and those enrolled in the phase III trials of boceprevir and telaprevir (Jacobson et al., 2011; Poordad et al., 2011). The presence of previous psychiatric illness and a history of substance use disorders were not reported for participants in these trials, but were likely much less prevalent than in the patients served by our program. Diagnosed psychiatric illness and previous substance use disorders were associated with less frequent SVR in our cohort, whereas substance use in the 6 months before treatment initiation was not. These data add to the wealth of data that argues against requiring patients to be abstinent from illicit substances in order to be treated for hepatitis C infection (Schulte et al., 2010; Sasadeusz et al., 2011; North et al., 2012). We were limited in our ability to assess the association of psychiatric and substance use on treatment toxicity. As such, we noted high frequencies of adherence among the patients who were treated in our program, but significant treatment discontinuation from drug-related toxicity. Treatment discontinuation because of drug toxicity led to triple-therapy SVR frequencies in our study that were lower than several dual-therapy SVR frequencies reported in similar settings and patient populations (Hill et al., 2008; Brew et al., 2013; Seidenberg et al., 2013).

HCV treatment outcomes in underserved, primary care settings have been described in several previous studies. Arora et al. described treatment by primary care clinicians in rural settings with remote access to specialist assistance, with proportion of SVR similar to contemporaneous studies (Arora et al., 2011). Ho et al. demonstrated efficacy of an intensive, multidisciplinary approach in a group setting, also based in a primary care clinic, in a population with a high prevalence of substance abuse, and psychiatric illness (Ho et al., 2013). Opioid substitution programs have also been used as a vehicle for HCV treatment, yielding reasonable frequencies of treatment success (Stein et al., 2012; Alavi et al., 2013).

The primary care setting provides many advantages for conducting hepatitis C treatment, including patient comfort and familiarity, and easier integration into patients’ existing treatment plans. The complexity of pretreatment evaluation and treatment, however, is challenging in this setting, and is one reason the treatment has traditionally been administered in specialized settings. To mitigate these barriers, our clinical model relies heavily on the use of hepatitis C-specific patient navigation and integrated mental health and social work services. The use of patient navigation has a strong basis in improving outcomes in a number of primary care-based practices, including HIV (Bradford et al., 2007) and cancer screening (Freeman et al., 1995). Integrated mental health has been shown to improve outcomes for patients in numerous primary care settings (Huffman et al., 2014).

Our study has several strengths and limitations. Our sample reflected clinical outcomes in an underserved, understudied demographic. We had high rates of complete follow-up for treated patients. The study was observational in nature, however, and therefore all clinical decisions, such as the initiation of treatment, and the decision to terminate treatment were based on the judgment of the administering clinician, instead of a preexisting study protocol. The purpose of this study was to illustrate real-world practice and outcomes, and therefore this approach was appropriate for our research question.

In our clinical cohort of patients with GT1 HCV from an underserved, largely racial and ethnic minority population, with a significant burden of substance dependence and mental illness, we observed a moderate frequency of HCV treatment initiation within our primary care-based treatment program during the era of first-generation DAAs. Treatment success (SVR) was significantly affected by pharmaceutical toxicity in our cohort. The newest-generation HCV DAAs, now in widespread use, do not require the use of interferon or ribavirin for the majority of patients, and therefore have dramatically reduced drug toxicity. Further research on optimizing the candidacy of patients with substance use and psychiatric disorders should be undertaken. Nonetheless, as HCV treatment becomes more tolerable and convenient, primary care-based HCV treatment may be an increasingly effective means of delivering HCV care to marginalized clinical populations that are disproportionately affected by this infection.