Sofosbuvir was the first direct-acting antiviral drug to be approved by the US Food and Drug Administration (FDA) for the treatment of hepatitis C virus (HCV) infection.1 The Centers for Disease Control and Prevention estimates that between 2.7 million and 3.9 million people in the United States have chronic hepatitis C.2 Approximately 60% to 70% of patients with HCV show signs and symptoms of liver disease. Patients infected with HCV may have chronic liver disease, cirrhosis, or hepatocellular carcinoma, which are notable reasons for a liver transplant, other serious morbidities, and death.3,4
This article provides an early evalution of the real-world effectiveness of 3 sofosbuvir-based treatment regimens; sofosbuvir is a direct-acting agent that was approved by the FDA in December 2013.1 Sofosbuvir-based regimens have shown high sustained virologic response (SVR) rates in clinical trials,5,6 in addition to an SVR rate of 90% in a phase 3 single-group study (95% confidence interval [CI], 87%-93%).7
The effectiveness of sofosbuvir-based treatments,7,8 as well as their high cost,9,10 have attracted considerable stakeholder attention. Patients, providers, and payers are keen to learn how these costly treatment regimens perform outside of a randomized controlled trial.
Information on treatment outcomes with sofosbuvir combinations has been limited to the results of tightly controlled clinical trials with small patient samples.5-7 Although one larger study has provided a glimpse into the postapproval observational results of sofosbuvir treatment among military veterans,11 there are almost no data on the use of this agent in commercially insured populations.
Fully cognizant of the downsides associated with the descriptive approach based on claims data (as discussed later: the intrinsic limitations of claims data, the high likelihood of selection bias, and the inability to provide definitive answers because of missing and overall limited availability of clinical data on the study patients), we have avoided comparisons between treatment regimens and have only cited cursory parallels with the results of previous clinical trials. In this descriptive analysis, we sought to provide an evaluation of the real-world effectiveness of sofosbuvir-based regimens in clinical practice within a commercially insured population. The primary objective was to evaluate patients’ response to a sofosbuvir-based regimen using HCV RNA viral load results after the completion of treatment using real-world data from integrated pharmacy and medical claims, as well as clinical data (ie, laboratory results) from patients’ medical records and network laboratory providers.
Study Design and Data Sources
This is an observational cohort study from a large national administrative claims database, the HealthCore Integrated Research Database, which is a repository of medical and pharmacy claims and laboratory results from 14 US-based commercial health plans, with approximately 13 million researchable members during the study period. Patients with chronic HCV who started sofosbuvir therapy (general product identifier code starting with 12353080) between the intake period of December 1, 2013, and April 30, 2014, were identified. Patients aged ≥18 years who had ≥1 RNA viral load tests after commencing treatment were included in the population available for medical record abstraction.
To enhance the study data with clinical information that is not typically captured in administrative claims (such as HCV genotype, treatment regimen, planned treatment duration, and HCV RNA viral load test results), the medical records of the identified patients, along with laboratory results from several network laboratory providers, were abstracted from the start of treatment (between December 1, 2013, and April 30, 2014) to the end of the study (December 31, 2014). Figure 1 lists the variables and data sources.
From among the 629 patients who were identified as having received sofosbuvir therapy during the intake period and were selected for medical records review, 278 patients’ records were abstracted. The reasons for patient’s noninclusion in the study included insubstantial contact information, provider refusal or unresponsiveness, or no documentation of sofosbuvir in the medical records. A total of 249 of the 278 indexed patients with available medical records had ≥1 viral load test results after initiating treatment, and among them only 200 (80%) records had viral load results available after the end of treatment for analysis of posttreatment response and relapse rates. This represents an effort to comply with the American Association for the Study of Liver Diseases and Infectious Diseases Society of America (AASLD/IDSA) joint recommendation for the testing, management, and treatment of HCV regarding HCV viral load quantitative measurements after 4 weeks and again at 12 weeks after the completion of treatment.12
The posttreatment response rates—the primary outcome measure—were calculated using the latest HCV RNA viral load test results from laboratory results and medical records after treatment completion. HCV viral load response was defined as an HCV RNA viral load of <25 IU/mL, which is consistent with the rates used in pivotal randomized control trials, using the most recent result after the end of treatment.7,13 In addition, various laboratories may have different thresholds for undetectable viral loads, so this value takes into account the disparities between these laboratories. The response rate in this study provides insights to the patients’ responses after the end of treatment, but it is not sufficient to derive results on the sustained viral load response (typically laboratory testing is done 12 weeks after the completion of treatment), because not all patients at the time of the medical record review had viral load testing occurring at least 12 weeks after the completion of treatment.
The posttreatment response rates were evaluated for the 3 sofosbuvir-based regimens—sofosbuvir plus peginterferon alfa and ribavirin, sofosbuvir plus ribavirin, and sofosbuvir plus simeprevir with or without ribavirin; this was further broken down by HCV genotypes 1 through 3 wherever the sample size was sufficient for analysis.
The disease relapse rates were measured to determine the proportion of patients who had a response (ie, viral load of <25 IU/mL) at any point during treatment, but who then had detectable viral loads (ie, viral load of ≥25 IU/mL) when measured after the end of treatment.7,14 Treatment discontinuation was defined as the abandonment of treatment for any reason before the end of the intended regimen, and was based on documentation in the patients’ medical records.
The secondary outcome was the number of patients who had a response to treatment within the first 4, 6, and 8 weeks of therapy to determine whether a lack of early response during treatment is suggestive of a lack of response after the end of treatment. To do this, we examined the viral load results among patients who had at least 1 viral load test within the first 4, 6, and 8 weeks of treatment to determine the proportion of patients who had viral loads of <25 IU/mL as an indicator of early response. For patients who did not have a response within these periods, we evaluated their posttreatment response.
Descriptive statistics were calculated to report the means for continuous variables and the frequencies and percentages for categorical variables. To evaluate the treatment response and disease relapse rates, 95% CIs using binomial proportion were used. Alpha was set at 0.05 for a 2-tailed significance test. The data analyses were conducted with SAS version 9.2 (SAS Inc; Cary, NC) and Stata version 12 (Stata Corp LP; College Station, TX).
More patients received sofosbuvir plus ribavirin (N = 116) than sofosbuvir plus peginterferon alfa and ribavirin (N = 93) or sofosbuvir plus simeprevir with or without ribavirin (N = 40; 3 received ribavirin, and 37 did not). The average patient age was 56.7 years; patients receiving sofosbuvir plus simeprevir with or without ribavirin were the oldest (ie, 61.6 years), and those receiving sofosbuvir plus peginterferon alfa and ribavirin were the youngest (ie, 55 years). The mean body mass index (BMI) ranged between 27 kg/m2 and 29 kg/m2 across the treatment groups.
Overall, 43% of all patients had documented status of alcohol abuse, and 32% had documented status of illicit intravenous drug use in their medical records. Table 1 lists clinical and demographic characteristics. The documented rates of liver fibrosis or cirrhosis, liver cancer, and liver transplant history in the patients’ medical records were 29%, 3%, and 9%, respectively.
Overall, 63% of the patients had genotype 1 HCV. The proportion of patients with genotype 1 HCV was higher in some treatment groups, including sofosbuvir plus simeprevir with or without ribavirin cohort (96%) and the sofosbuvir plus peginterferon alfa and ribavirin cohort (92%). The remaining sample was comprised of patients with genotype 2 (21%), genotype 3 (12%), and genotype 4 (1%) HCV (Table 1).
Treatment Duration and Discontinuation
Overall, more patients received 12 weeks of treatment (78%) than 24 weeks (22%; Table 1), which was based on the recommended lengths of therapy resulting from multiple patient factors, including genotype, the type of previous treatment received, and the extent of liver disease. The proportions of patients receiving a 12-week regimen were 96%, 57%, and 100% across the cohorts receiving sofosbuvir plus peginterferon alfa and ribavirin, sofosbuvir plus ribavirin, and sofosbuvir plus simeprevir with or without ribavirin, respectively. Overall, 10% of patients in our study discontinued treatment, which was higher than the 1% to 2% discontinuation rate reported in clinical trials5-7; the reason for discontinuation was unknown in 8% of patients. Less than 10% of the patients responding while receiving treatment had disease relapse after the completion of treatment, and thus would not have had an SVR.
Response rates were reported for all patients after the end of the planned treatment duration, and were stratified by regimen and by genotype, as is shown in Table 2.
The overall response rate was 88% (95% CI, 84%-93%). The response rates were 93% (95% CI, 88%-98%) in patients receiving sofosbuvir plus ribavirin, 86% (95% CI, 78%-94%) in patients receiving sofosbuvir plus peginterferon alfa and ribavirin, and 81% (95% CI, 68%-94%) in patients receiving sofosbuvir plus simeprevir with or without ribavirin.
Among patients with genotype 1 HCV, the response rates were 92% (95% CI, 81%-100%), 89% (95% CI, 81%-96%), and 83% (95% CI, 70%-95%) in patients receiving sofosbuvir plus ribavirin, sofosbuvir plus peginterferon alfa and ribavirin, or sofosbuvir plus simeprevir with or without ribavirin, respectively, but the differences were not significant (P = .54). The response rate among patients with genotype 1 HCV who received sofosbuvir plus peginterferon alfa and ribavirin in a previous single-group, open-label study (89%)7 were the same as our study (89%).
Table 2 shows a response rate of 89% (95% CI, 83%-95%) in patients with genotype 1a HCV and 83% (95% CI, 70%-97%) in patients with genotype 1b HCV. The response rates in patients with genotypes 2 and 3 HCV were 98% (95% CI, 93%-100%) and 88% (95% CI, 71%-100%), respectively.
Among patients receiving sofosbuvir plus ribavirin, the response rates were 92% (95% CI, 81%-100%), 98% (95% CI, 93%-100%), and 86% (95% CI, 67%-100%) for patients with genotypes 1, 2, and 3 HCV, respectively (Table 2), but there was no significant difference between the groups (P = .24). The response rates among patients with genotype 2 HCV receiving sofosbuvir plus ribavirin in previous clinical studies (95%) were similar to our findings (98%).5,7,14 The response rate among the patients with genotype 3 HCV receiving sofosbuvir plus ribavirin in our study matched the results from an open-label, randomized controlled trial (86% vs 84%, respectively) that had the same treatment duration.14
Two-thirds of patients with any viral load test result by week 4 across all regimens responded to sofosbuvir treatment within the first 4 weeks (Figure 2). The proportions of patients responding within 6 weeks and 8 weeks were 82% and 89%, respectively. Of the patients without an early response (within 4 weeks) to sofosbuvir treatment who had any viral load test result after week 12, 81% responded after week 12. Late response to sofosbuvir treatment after week 12 was 68% within 6 weeks and 53% within 8 weeks.
Disease Relapse Rates
Among patients who completed their treatment, 11 (8%) of those who responded while receiving treatment had relapsed disease after the treatment ended, including 11% of patients receiving sofosbuvir plus peginterferon alfa and ribavirin, 5% receiving sofosbuvir plus ribavirin, and 8% receiving sofosbuvir plus simeprevir with or without ribavirin.
A majority (88%) of patients in this study sample responded to the 3 HCV treatment regimens, as determined by the viral RNA load results after the end of treatment. The response rates after the end of treatment were similar to those from clinical studies with sofosbuvir.5,7,14 This is a cursory observation; we did not conduct a statistical comparison of our descriptive results with those from applicable clinical trials, because this was not the focus of the study. This is also true for the results of a recent observational study by Backus and colleagues using the Veterans Affairs (VA) Clinical Case Registry for HCV.11 Although the findings of the study by Backus and colleagues are instructive and represent the first large real-world study in patients receiving sofosbuvir-based therapy,11 no rational basis existed for a direct comparison with our results.
Backus and colleagues evaluated a patient population of veterans, and used a different definition of response and a different patient population.11 In our study, we defined response as an HCV viral RNA load of <25 IU/mL to allow for different thresholds used to determine undetectable viral loads, which was consistent with values used in clinical trials. Because our study showed slightly higher response rates than those in the VA study, it is possible that various factors could have contributed to these differences, including the thresholds used for undetectable viral loads. The VA study population included more male patients, older patients, and patients with a higher body mass index than the patients in our study,11 which may suggest additional reasons for the different response rates. In addition, we could not compare the extent of cirrhosis between these 2 populations.
Notably, 34% of patients who had test results during the first 4 weeks after initiating treatment did not respond, raising questions about the effectiveness of continued treatment (Figure 2). Among those patients, however, more than 80% responded after week 12, alleviating concerns about effectiveness and supporting the importance of treatment persistence for the intended duration of treatment.
Less than 10% of the patients responding while receiving treatment had a relapse after treatment completion. To our knowledge, this could be the first time that relapse rates were reported in relation to usual care. Future studies are needed to identify patients who are at risk for relapse and to determine the optimal treatment for successful outcomes.
Approximately 20% of patients who received treatment did not have viral testing after treatment to evaluate their response. This reiterates the existing challenge in the care management of tracking patients for outcomes. As such, it would be advisable for clinicians to monitor viral loads during and after active treatment and to identify patients who may need further management.13
The results of our study provide some key insights about an important advancement in the treatment of HCV, which has an immense impact on society. Interest in direct-acting treatments has intensified since the completion of this study in 2014, which concentrated on sofosbuvir-based direct-acting HCV agents available before 2014.
After the completion of our study, new sofosbuvir-based therapies have been approved by the FDA for the management of patients with HCV, including the combination of elbasvir and grazoprevir for the treatment of genotype 1 or 4 HCV in adults15,16; daclatasvir for the treatment of genotype 1 or 3 HCV17,18; the combination of ombitasvir, paritaprevir, and ritonavir for the treatment of genotype 4 HCV19,20; the combination of ombitasvir, paritaprevir, and ritonavir for the treatment of genotype 1 HCV21,22; and the combination of ledipasvir and sofosbuvir for the treatment of genotype 1, 4, 5, or 6 HCV.23,24
These results must be considered with caution, because in addition to the small sample size, the volume of missing medical records data was substantial, and there are inherent limitations associated with claims data.
The study is also limited by a common challenge associated with population-based studies, with 20% of the patients not having test results after their treatment ended, rendering their response status unknown. This is well below the required levels in the AASLD/IDSA joint recommendations for viral load tests at 4 weeks and again at 12 weeks after treatment completion.12 This presents a clear opportunity for providers to improve the rates of testing and documentation during the course of managing patients with HCV who receive these agents.
This observational study equated prescription fill with the start of treatment, although this may not necessarily be true; however, the seriousness of this disease and the associated high costs of treatment could have limited medication discontinuation in this population.
Finally, all data were drawn from a commercially insured population, so the results may not be generalizable to patients with HCV who have another type of insurance (ie, Medicaid or Medicare), or who are uninsured.
The high response rates among patients in this study were comparable with the rates seen in clinical trials. Although a sizable population of patients in our study still had detectable viral loads within the first 4 to 6 weeks of treatment, the vast majority responded by the end of treatment. With the treatment of HCV evolving quickly with increasing medications coming to market, payers and providers will be looking for future studies, with more completely documented patient information profiles, to generate data-driven evidence regarding which regimens are the most effective in specific populations with HCV.
The rapid emergence of new treatments in this field presents exciting opportunities for additional research, and holds important clinical and economic implications for patients and their families, healthcare providers, and critically, for payers, who have to accommodate the new pricing models associated with these treatments.
The authors thank Jinan Liu, formerly of HealthCore, who was part of the team that initiated, designed, and wrote the protocol for this study.
This study was funded by Anthem, a health insurance company. HealthCore is an outcomes research subsidiary of Anthem.
Author Disclosure Statement
Dr Barron, Dr White, Dr Singer, and Dr Rosenberg hold stocks of Anthem. Dr Xie, Dr Wu, and Mr Tulsi reported no conflicts of interest.
1. US Food and Drug Administration. Approval of Sovaldi (sofosbuvir) tablets for the treatment of chronic hepatitis C. Updated September 25, 2014. www.fda.gov/ForPatients/Illness/HepatitisBC/ucm377920.htm. Accessed August 26, 2016.
2. Centers for Disease Control and Prevention. Hepatitis C FAQs for health professionals. Updated July 21, 2016. www.cdc.gov/hepatitis/hcv/hcvfaq.htm. Accessed August 26, 2016.
3. Kowdley KV, Gordon SC, Reddy KR, et al; for the ION-3 Investigators. Ledipasvir and sofosbuvir for 8 or 12 weeks for chronic HCV without cirrhosis. N Engl J Med. 2014;370:1879-1888.
4. Westbrook RH, Dusheiko G. Natural history of hepatitis C. J Hepatol. 2014;61(1 suppl):S58-S68.
5. Jacobson IM, Gordon SC, Kowdley KV, et al. Sofosbuvir for hepatitis C genotype 2 or 3 in patients without treatment options. N Engl J Med. 2013;368:1867-1877.
6. Lawitz E, Lalezari JP, Hassanein T, et al. Sofosbuvir in combination with peginterferon alfa-2a and ribavirin for non-cirrhotic, treatment-naive patients with genotypes 1, 2, and 3 hepatitis C infection: a randomised, double-blind, phase 2 trial. Lancet Infect Dis. 2013;13:401-408.
7. Lawitz E, Mangia A, Wyles D, et al. Sofosbuvir for previously untreated chronic hepatitis C infection. N Engl J Med. 2013;368:1878-1887.
8. Stedman C. Sofosbuvir, a NS5B polymerase inhibitor in the treatment of hepatitis C: a review of its clinical potential. Therap Adv Gastroenterol. 2014;7:131-140.
9. Patel KR. Sovaldi ushers in wave of costly hepatitis C drugs. Manag Care. 2014;23:51.
10. The price of good health. Nat Med. 2014;20:319.
11. Backus LI, Belperio PS, Shahoumian TA, et al. Effectiveness of sofosbuvir-based regimens in genotype 1 and 2 hepatitis C virus infection in 4026 U.S. Veterans. Aliment Pharmacol Ther. 2015;42:559-573.
12. American Association for the Study of Liver Diseases; Infectious Diseases Society of America. Recommendations for testing, managing, and treating hepatitis C. Updated July 6, 2016. http://hcvguidelines.org/sites/default/files/HCV-Guidance_July_2016_b.pdf. Accessed July 16, 2016.
13. Hsu CS. Sofosbuvir for previously untreated chronic hepatitis C infection. Letter to the editor. N Engl J Med. 2013;369:678-679.
14. Foster GR, Pianko S, Brown A, et al; for the BOSON Study Group. Efficacy of sofosbuvir plus ribavirin with or without peginterferon-alfa in patients with hepatitis C virus genotype 3 infection and treatment-experienced patients with cirrhosis and hepatitis C virus genotype 2 infection. Gastroenterology. 2015;149:1462-1470.
15. Zepatier (elbasvir and grazoprevir) tablets [prescribing information]. Whitehouse Station, NJ: Merck & Co; January 2016.
16. US Food and Drug Administration. FDA approves Zepatier for treatment of chronic hepatitis C genotypes 1 and 4. Press release; January 28, 2016. www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm483828.htm.
17. Daklinza (daclatasvir) tablets [prescribing information]. Princeton, NJ: Bristol-Myers Squibb Company; April 2016.
18. US Food and Drug Administration. FDA approves new treatment for chronic hepatitis C genotype 3 infections. Press release; July 24, 2015. www.fda.gov/newsevents/newsroom/pressannouncements/ucm455888.htm. Accessed August 31, 2016.
19. US Food and Drug Administration. FDA approves Technivie for treatment of chronic hepatitis C genotype 4. www.fda.gov/newsevents/newsroom/pressannouncements/ucm455857.htm. Press release; July 24, 2015. Accessed August 31, 2016.
20. Technivie (ombitasvir, paritaprevir, and ritonavir) tablets [prescribing information]. North Chicago, IL: AbbVie; June 2016.
21. US Food and Drug Administration. FDA approves Viekira Pak to treat hepatitis C. Press release; December 19, 2014. www.fda.gov/newsevents/newsroom/pressannouncements/ucm427530.htm. Accessed August 31, 2016.
22. Viekira Pak (ombitasvir, paritaprevir, and ritonavir tablets; dasabuvir tablets) [prescribing information]. North Chicago, IL: AbbVie; June 2016.
23. US Food and Drug Administration. FDA approves first combination pill to treat hepatitis C. Press release; October 10, 2015. www.fda.gov/newsevents/newsroom/pressannouncements/ucm418365.htm. Accessed August 31, 2016.
24. Harvoni (ledipasvar and sofosbuvir) tablets [prescribing information]. Foster City. CA: Gilead Sciences; June 2016.