FormalPara Key Summary Points

Why carry out this study?

Clinical trials have shown that an 8-week regimen of glecaprevir/pibrentasvir (GLE/PIB) is safe and effective in patients with chronic hepatitis C (HCV) and cirrhosis, but there are few studies of real-world US patients.

We compared the rate of sustained virological response (SVR) and adverse events among patients with cirrhosis from the Chronic Hepatitis Cohort Study (CHeCS), a multisite study of patients with viral hepatitis from four US health systems.

What was learned from the study?

There were no significant differences in the rates of SVR nor adverse outcomes between patients receiving 8 or 12 weeks of GLE/PIB treatment.

An 8-week regimen of GLE/PIB demonstrated a similar rate of SVR and adverse outcomes compared to the originally approved 12-week regimen, which may reduce related costs and patient burden.

Introduction

Glecaprevir/pibrentasvir (GLE/PIB, trade name Mavyret™) is a fixed-dose combined regimen of a new-generation NS3/4A inhibitor and an NS5A inhibitor [1]. An 8-week regimen of GLE/PIB was approved by the US Food and Drug Administration (FDA) in August 2017 for treatment of chronic hepatitis C (HCV) genotypes (GT) 1–6 in treatment-naïve patients without cirrhosis [2]. In September 2019, the FDA granted approval for its use in patients with cirrhosis [3], given positive results of the EXPEDITION-8 study [4]—which demonstrated a 99.7% rate of sustained virological response (SVR) among patients randomized to the shorter regimen—but issued a Drug Safety Communication warning regarding rare occurrences of serious liver injury related to use of GLE/PIB, elbasvir-grazoprevir, or sofosbuvir-velpatasvir-voxilaprevir in some patients with advanced liver disease [5]. Although there are reports from outside the USA regarding the effectiveness of the 8-week GLE/PIB regimen in a clinical setting, they may not be generalizable to a multiracial US population [6]. Likewise, there is little data regarding the safety and effectiveness of an 8-week GLE/PIB regimen among US patients with cirrhosis. One USA-based study of an 8-week regimen in treatment-naïve patients with cirrhosis demonstrated a 96% SVR rate [7], but no safety data were collected and there was no comparison group. Likewise, a recent study using a large administrative database found that an 8-week regimen was effective in treatment-naïve patients with cirrhosis but did not include data regarding adverse events [8]. Another study in a Veterans Administration population [9] was overwhelmingly (97%) male and included a large proportion of patients without cirrhosis.

Our goal was to address this knowledge gap by investigating the safety and effectiveness an 8-week regimen of GLE/PIB in USA-based patients with HCV receiving routine clinical care. Using data from the Chronic Hepatitis Cohort Study (CHeCS), an observational study of over 16,000 patients with HCV receiving routine clinical care from one of four large US health systems, we evaluated (a) treatment response, and (b) on-treatment adverse events and laboratory abnormalities in patients with HCV and compensated cirrhosis receiving an 8-week regimen of GLE/PIB. Our endpoints were (1) rates of SVR at 12 weeks (SVR12) after end of treatment (EOT); and (2) adverse events, based on an intention-to-treat (ITT)/modified ITT approach.

Methods

CHeCS is a dynamic longitudinal study of patients with HCV from four large US health systems. The study follows all Declaration of Helsinki and US Department of Health and Human Services guidelines regarding the protection of human subjects. The CHeCS protocol was approved and is renewed annually by the institutional review boards at each of the four sites (Henry Ford Health [HFH], Detroit, MI; Geisinger Health System [GHS], Danville, PA; Kaiser Permanente Northwest, Portland, OR; and Kaiser Permanente Hawaii, Honolulu, HI). The overarching CHeCS study protocol was approved by the HFH IRB (IORG0000148). Due to the de-identified nature of this observational study, requirements for written informed consent were waived. CHeCS methods have been described elsewhere [10, 11]. Patients’ HCV and treatment status were confirmed using medical chart abstraction.

For this sub-study, patients from two CHeCS sites (HFH and GHS) were included if they had cirrhosis and were prescribed GLE/PIB treatment from August 2017 through June 2020. Given that there is variation in the amount and type of data collected in “real-world” routine-care settings, we have implemented a hierarchical classification algorithm to identify patients with cirrhosis; sources include liver biopsy, vibration-controlled transient elastography (VCTE), Fibrosis-4 (FIB4), and diagnosis codes from the electronic health record (EHR) [12]. To characterize history of decompensated cirrhosis prior to GLE/PIB initiation, we used a diagnosis- and procedure-code-based, automated classification model that was previously derived from CHeCS data, using a Classification Regression Tree (CART) method; predictive accuracy (area under the receiver operating characteristic [AUROC] curve) of this method is > 0.85 for decompensated cirrhosis [12]. Patients were excluded from this analysis if they were co-infected with hepatitis B virus or had a history of decompensated cirrhosis.

Data Collection

Index date (baseline) was defined as the date of first GLE/PIB treatment initiation. Data collected at/before the index date included sex; ethnicity (Hispanic, non-Hispanic); race (White; Black/African American; Asian American/ Pacific Islander [AAPI]); age category (≤ 50; 51–60; 61–70; > 70); previous history of HCV treatment with interferon-based (IFN) drugs or direct acting antivirals (DAAs); HCV GT; and Charlson-Deyo comorbidity index [13], excluding liver-related comorbidities.

Detailed treatment course data were collected, including: planned treatment duration; actual treatment start and end dates; treatment interruptions; and reasons for interruptions or early treatment termination (including adverse events). SVR12 was defined as undetectable HCV RNA results at least 12 weeks after the EOT; patients with and without RNA testing data were included. Safety data were collected by medical chart abstraction for the on-treatment period; mild/moderate adverse events (AEs) included: anemia; fatigue; nausea/diarrhea; and skin rash. Serious adverse events (SAEs) included any that resulted in hospitalization, emergency department visits, or death. The incidence of on-treatment hepatic decompensation (AE or SAE) was also captured. Indications of decompensation were defined as any of the following while on GLE/PIB treatment: (1) Any diagnosis or procedure codes associated with decompensated cirrhosis: hepatic encephalopathy; ascites/paracentesis procedures; gastrointestinal hemorrhage; esophageal varices complications (bleeding) and related procedures portal hypertension and portal decompression procedures; hepatorenal syndrome; liver cancer; liver transplant; or other sequelae of chronic liver disease (codes 572.8, K72.10, K72.90); (2) Hyperbilirubinemia (total bilirubin > 3.0 mg/dL) or diagnosis codes for jaundice (782.4, R17); (3) A greater than 25% decline in albumin levels from baseline; (4) A tenfold increase over baseline for alanine aminotransferase (ALT) or aspartate aminotransferase (AST); (5) Any prescription for lactulose or rifaximin; or (6) Documentation on-treatment hepatic decompensation during treatment (based on treating physician’s notes). All possible indications of decompensation were reviewed by an experienced hepatologist for confirmation of the diagnosis and onset during treatment.

Study Design and Statistical Analysis

To optimize the rigor of our analysis, the study protocol was developed using pragmatic clinical trial approaches. These included pre-specified inclusion/exclusion criteria; primary and secondary hypotheses regarding SVR among patients planned to receive an 8-week regimen; frequency of AEs among patients planned to receive an 8-week regimen; and an exploratory hypothesis regarding differences in SVR rates between patients planned to receive 8-week, 12-week, and other treatment duration regimens.

The primary endpoint was SVR12, defined as undetectable HCV RNA ≥ 12 weeks after the EOT, with a one-sided 95% confidence limit (95%CL). We used an intention-to-treat (ITT) approach, which means that patients were included if it was planned that they would receive a specific course of GLE/PIB. Lack of data regarding post-treatment HCV RNA levels was classified as non-SVR. A modified ITT (mITT) analysis assessed SVR12 results only among patients with available post-treatment HCV RNA data. Patient baseline characteristics were described with mean and standard deviation (std) for continuous variables, and proportions for categorical variables. Actual treatment duration was reported.

We also explored the impact of planned treatment duration as well as other risk factors (GT, type 2 diabetes [T2D], prior treatment experience, and comorbidities) on rates of SVR using Fisher’s exact test for categorical variables and t test for continuous variables.

Safety and toxicity during treatment was the secondary endpoint. Data collected included: adverse events (AEs, defined as health conditions/events that resulted in an emergency room visit, hospitalization, a new prescription for treatment of a new onset or exacerbated health condition, or alteration in the course of antiviral therapy); serious adverse events (SAEs, the subset of AEs that resulted in an emergency room visit, hospitalization, or death during the treatment period); decompensation events during treatment; actual treatment duration; and early treatment terminations or interruptions. We used an ITT approach with predefined criteria; patients were stratified on the basis of the planned treatment duration (8 weeks; 12 weeks; > 12 weeks). We also collected data regarding treatment courses that were abbreviated or interrupted. Safety analyses were further stratified by use/non-use of ribavirin, given the high rate of side effects [14].

Results

We identified a total of 342 patients with compensated cirrhosis who were prescribed GLE/PIB. After review for case confirmation, 176 (51.5%) were excluded; reasons included: (1) failure to initiate prescribed treatment (n = 33); (2) treatment received at an outside facility (n = 23); (3) unconfirmed cirrhosis status/non-cirrhotic (n = 117); and (4) evidence of decompensation prior to treatment initiation (n = 3).

The remaining sample of 166 patients was included in the ITT analysis (Table 1) with 43, 116, and 7 patients in the 8-week, 12-week, and > 12-week planned treatment groups, respectively. For the mITT analysis, six patients with unknown HCV RNA (two in the planned 8-week and four in the planned 12-week groups) were excluded, for a sample of 160 patients with known HCV RNA status at 12 weeks post-treatment (Table 1; with 42, 112, and 6 patients in the 8-week, 12-week, and > 12-week planned treatment groups). Again, roughly one-quarter (26.3%) were included in the planned 8-week treatment group. Only two patients received adjuvant ribavirin—one in the 8-week group and the other in the 12-week group.

Table 1 Patients characteristics by planned treatment duration (8 weeks, 12 weeks, other) for all intention-to-treat patients (ITT, N = 166) and modified intention-to-treat (mITT, n = 160)
Full size table

The univariate analysis demonstrated some significant differences in patient characteristics by planned treatment duration. Patients in the planned 8-week treatment group were younger and had lower rates of private health insurance; FIB4 scores were also lower in the 8-week group.

Table 2 displays the results of the primary efficacy analysis. In the ITT sample (including patients with unknown HCV RNA status at 12 weeks post-end of treatment), 159 of 166 patients had confirmed SVR (95.8%, 95% lower confidence limit [LCL] 93.2%). In the mITT group (excluding patients with unknown HCV RNA), 159 of 160 patients achieved SVR (99.4%, LCL 98.3%). Among patients in the 12-week treatment group, rates of SVR were 95.7% (LCL 92.6%) and 97.6% (LCL 97.6%) for the ITT and mITT samples, respectively, while SVR rates in the 8-week group were 97.7 (LCL 93.8%) and 100.00% (LCL 100%). Confidence limits overlapped for all groups, indicating that there were no significant differences in rates of SVR between the 8- and 12-week regimens. The descriptive analysis showed no association between patient characteristics and SVR in the ITT or mITT samples.

Table 2 Unadjusted rates of confirmed SVR12 for patients with planned 8- and 12-week treatment regimens
Full size table

As shown in Table 3, the unadjusted analysis showed similar rates of non-SVR rates in the 8- and 12-week groups in the ITT sample (p = 0.17). Likewise, there remained no significant differences in SVR rates between the 8-week and 12-week groups after propensity score adjustment (p = 0.08, odds ratio [OR] 0.27, 95% confidence interval [CI] 0.06–1.14). No other significant associations were observed between patient characteristics and SVR. Similar results were observed in the mITT sample; there were no non-SVR patients in the 8-week group (0.0%) and one in the 12-week group (0.9%; p = 0.86 [unadjusted] and 0.93 [propensity score weighted]). Again, there were no other significant associations between patient characteristics and SVR in the mITT sample.

Table 3 Achievement of confirmed SVR12 and association with patient characteristics for all patients (ITT sample) and sample with unknown HCV RNA excluded (mITT)
Full size table

Adverse Events

As shown in Table 4, 24 of 166 patients in the ITT sample (14%) experienced at least one adverse event (AE)—3 (7%) in the 8-week group, 19 (16%) in the 12-week group, and 2 (29%) in the “other” group. All three patients in the 8-week group experienced only a single AE, one of which was serious (SAE); 18 patients (16%) in the 12-week group experienced a single AE, six of which were SAEs. One patient in the 12-week group experienced two AEs, neither serious. Two patients in the “other” group each experienced two AEs, both of which were SAEs.

Table 4 Summary of toxicity and protocol deviation data by planned treatment duration
Full size table

SAEs were experienced by only one patient (2%) in the 8-week group, seven (5%) in the 12-week group (including one death), and two (29%) in the > 12-week group. Only four patients experienced SAEs or hepatic events that were “likely attributable” to GLE/PIB treatment; all four of these patients were from the planned 12-week group. One of the four developed hepatic encephalopathy approximately 30 days after commencing therapy, but completed therapy and achieved SVR12. The other three patients developed hyperbilirubinemia that began after initiation of therapy and ended after therapy was completed (although one of these patients was suspected of relapsing alcohol misuse while on therapy). Two of these three patients discontinued treatment early, while the third completed the full planned course; all three achieved SVR12. None of the four patients had an associated hospitalization or emergency department visit.

A fifth patient (from the 12-week group) experienced hyperbilirubinemia while on GLE/PIB treatment. Bilirubin was normal during the first 30 days of treatment, but the patient was diagnosed with hepatocellular carcinoma shortly after initiating therapy; they subsequently underwent ablation therapy. After the ablation procedure, bilirubin increased to 5.6 mg/dL then gradually declined to 1.7 mg/dL throughout the course of treatment, fully normalizing after treatment ended. As a result of the timing of the hyperbilirubinemia relative to the ablation procedure, we suspect that the ablation procedure, rather than the GLE/PIB treatment, was the primary driver of the development of hyperbilirubinemia.

One patient in the planned 12-week treatment group died while on therapy, roughly 9.5 weeks after starting treatment. The patient, who had recurrent GT1a HCV after a 1999 liver transplant, had a number of comorbidities, including chronic kidney disease. They were treatment naïve and had a liver stiffness of 21.6 kPa before starting treatment, without prior evidence of decompensation. The patient died at home alone of unexplained causes. The treating hepatologist noted that he spoke with the patient 3 days before their death and the patient reported no new complaints, and that recent labs indicated the patient was doing well.

Two patients were prescribed adjuvant ribavirin: one in the planned 8-week treatment group, and one in the 12-week group. There were no AEs/SAEs observed in either of these patients. Both completed their treatment and achieved SVR. A summary and detailed clinical review of AEs are included in Appendix 1.

Discussion

This study was a real-world evaluation of an 8-week regimen of GLE/PIB in a USA-based routine clinical care setting among patients with HCV and cirrhosis. High rates of SVR were observed across HCV genotypes—and these cure rates did not depend on any patient characteristics present before treatment. The effectiveness and abbreviated treatment course may simplify care and allow non-specialist healthcare professionals to treat these patients, contributing to global efforts to eliminate HCV.

Among patients with known HCV RNA results at 12 weeks post-end of therapy (the mITT sample), 100% of patients achieved SVR; inclusion of patients with unknown response (ITT sample) still resulted in a 97.7% rate of SVR. These results are consistent with those of the EXPEDITION-8 clinical trial of an 8-week GLE/PIB regimen in treatment-naïve patients without cirrhosis (99.7% in the per-protocol sample and 97.7% in the ITT sample) [4]. They are likewise consistent with a number of “real-world” studies of patients in Taiwan [6] and the US Veterans’ Administration [7, 9], as well as a recent study drawn from a large US administrative database [8].

Our study population in the planned 8-week treatment group was roughly representative of US routine care patients with HCV and compensated cirrhosis. Overall, there were few significant differences in patient characteristics between the 8- and 12-week groups; the 8-week patients tended to be younger, less likely to have private insurance, and had lower FIB4 scores (Table 1). We used propensity score adjustment to address any potential treatment selection bias. Likewise, given that there is variation in the amount and type of data collected in “real-world” routine-care settings, we have implemented a hierarchical classification algorithm to identify patients with viral hepatitis-related cirrhosis with data sourced from liver biopsy, VCTE, FIB4 [2], and diagnosis codes from the EHR.

The 8-week course of GLE/PIB was well tolerated in our patient population. Only 7% and 2% of patients in the 8-week group experienced an AE or SAE, respectively, during the treatment period (see Supplemental Table 1). Although these rates are considerably lower than the rates (AE 46%, SAE 2%) observed in the EXPEDITION-8 trial, safety data for that trial were also collected for 30 days after the end of treatment [4].

Overall, no new safety signals were observed during this real-word study, and the safety profile was consistent with the approved label [5]. Only one SAE was observed among patients receiving the 8-week regimen (a patient experienced suicide ideation/attempt and was hospitalized) but no serious liver injury related to use of GLE/PIB was observed.

We did observe some significant differences between patients who were offered the 8-week versus 12-week regimens. Patients in the 8-week group were younger and had lower FIB4 scores. Notably, we also observed that patients in the 8-week group were less likely to have private insurance than those in the 12-week group (14% versus 32%, respectively; p = 0.02). However, these differences were not associated with differences in rates of SVR12.

There are some limitations of this study. As a result of the strict inclusion criteria, only 166 (48%) of the 342 patients prescribed GLE/PIB were included in the study sample. This may have limited our ability to detect some significant associations between patients’ demographic and clinical characteristics and likelihood of SVR. Likewise, 96% of our sample was composed of patients with HCV GT 1, 2, and 3; as a result, we are unable to draw any conclusions about efficacy in patients with other or unknown GTs. Subtyping data was unavailable for a large portion of the sample, which made it impossible to compare outcomes by subtype; however, there were no indications of differences in SVR by subtype among those patients with available data (Supplemental Table S5).

Strengths of this study include the geographically and racially diverse sample, as well as detailed demographic and clinical data, which allowed us to use propensity scores to adjust for potential treatment selection bias, as well as to evaluate whether these variables were associated with likelihood of achieving SVR. Although this was a retrospective study, our protocol was designed to be rigorous, using principles of pragmatic clinical trials adapted for “read-world” cohorts, including pre-defined primary and secondary objectives with consideration of treatment safety before data collection commenced, as well as use of ITT and mITT cohorts, and assessment of side effects and AEs for all treated patients.

Conclusion

Our study demonstrated that an 8-week regimen of GLE/PIB is well tolerated and highly effective among US patients with HCV and cirrhosis receiving routine clinical care (SVR 97.7% [ITT] and 100% [mITT]). This shorter treatment duration can reduce the complexity and expense of the HCV care pathway, furthering progress towards HCV elimination.