About AVT
Browse Articles
Authors
Customer Services

Original article

Comparative effectiveness of 8- and 12-week ledipasvir/sofosbuvir regimens for HCV infection

Rohit P Ojha1,2,*, Brooke R MacDonald1, Tzu-Chun Chu1, Esther O Fasanmi3, Jonathan D Moore2, Rachel A Stewart4

1Center for Outcomes Research, JPS Health Network, Fort Worth, TX, USA
2Department of Biostatistics and Epidemiology, UNT Health Science Center School of Public Health, Fort Worth, TX, USA
3Pharmacy Services Administration, JPS Health Network, Fort Worth, TX, USA
4Acclaim Gastroenterology, JPS Health Network, Fort Worth, TX, USA

*Corresponding author e-mail: rojha@jpshealth.org

Citation: Antiviral Therapy 2018; 23:585-592
doi: 10.3851/IMP3249

Date accepted: 18 June 2018
Date published online: 03 July 2018

Copyright (c) 2018 International Medical Press, all rights reserved.

Abstract

Background: Real-world studies have aimed to compare the effects of 8- and 12-week ledipasvir/sofosbuvir regimens on sustained virological response (SVR) among HCV infection genotype-1 (HCV-1) treatment-naive patients. Nevertheless, real-world comparative effectiveness studies pose unique challenges, such as confounding by indication, that were not adequately addressed in prior studies. We thus aimed to address limitations in prior studies and compare overall- and subgroup-specific effectiveness of 8- and 12-week ledipasvir/sofosbuvir regimens among HCV-1 treatment-naive patients.

Methods: Patients eligible for our study were aged ≥18 years and initiated 8- or 12-week ledipasvir/sofosbuvir regimens for treatment-naive HCV-1 at an urban public hospital network. We excluded patients with HIV or cirrhosis. We used marginal structural models to estimate overall and subgroup-specific risk ratios (RRs) and 95% confidence limits (CL) comparing the effect of 8- and 12-week ledipasvir/sofosbuvir regimens on 12-week SVR.

Results: Our study population comprised 191 patients. Among both regimens, the majority were aged >50 years, non-Hispanic White and uninsured. The overall risk of SVR was comparable between the 8- and 12-week regimens (RR=1.01, 95% CL: 0.92, 1.11). The risk of SVR did not vary by race/ethnicity (non-Hispanic Black: RR=1.01, 95% CL: 0.84, 1.21; non-Hispanic White: RR=1.01, 95% CL: 0.89, 1.04).

Conclusions: Our real-world results suggest that 8- and 12-week ledipasvir/sofosbuvir have comparable effects on SVR among HCV-1 patients without cirrhosis or HIV. In addition, the comparable effectiveness of 8- and 12-week regimens among non-Hispanic Black individuals adds to the growing body of evidence that supports the removal of race-based treatment guidelines.

Introduction

The United States Food and Drug Administration (US FDA) approved a fixed-dose ledipasvir/sofosbuvir combination for treating HCV genotype 1 (HCV-1) infections in October 2014 [1]. This approval was based on results of the ION-1 [2] and ION-3 [3] randomized controlled trials (RCTs). In particular, the ION-3 trial restricted eligibility to treatment-naive HCV-1 patients without cirrhosis and compared 8- and 12-week ledipasvir/sofosbuvir regimens. Sustained virological response (SVR) frequencies were similar between the regimens [3]. Nevertheless, results from RCTs do not always generalize to real-world (that is, observational) settings, where practical considerations override ideal conditions. In addition, guidelines from the American Association for the Study of Liver Diseases (AASLD) and Infectious Diseases Society of America (IDSA) currently recommend 8-week ledipasvir/sofosbuvir for patients who are non-Black, HIV-negative and have HCV RNA levels <6 million IU/ml [4], but these guidelines are sometimes modified in practice, which raises further questions about real-world effectiveness.

Several real-world studies have assessed the effect of 8- and 12-week ledipasvir/sofosbuvir regimens on SVR [513], but real-world comparative effectiveness studies pose unique challenges that were not adequately addressed in prior studies [14]. In particular, confounding by indication is a bias specific to observational studies of treatment effects and occurs because patients may have favourable or unfavourable characteristics for which certain treatments are indicated. The results of real-world studies can be misleading because of confounding by indication and guidelines consistently emphasize addressing this bias in comparative effectiveness studies [15]. Most prior studies [5,712] presented results of unadjusted comparisons of the 8- and 12-week regimens, which could overestimate the effectiveness of the 8-week regimen because of more favourable characteristics in this group, but the magnitude of potential overestimation is unclear. Attempts to reduce biases and improve interpretation are particularly important given the implications if 8- and 12-week ledipasvir/sofosbuvir regimens have comparable effectiveness in real-world settings. For example, a shorter duration treatment could reduce cost and improve adherence [16,17]. Therefore, we aimed to compare overall and subgroup-specific effectiveness of 8- and 12-week ledipasvir/sofosbuvir regimens among treatment-naive HCV-1 patients.

Methods

Study population

We queried electronic health records to develop a clinical cohort with prospectively documented treatment (at baseline) and outcome (at follow-up) [18,19]. Electronic health record data were validated and supplemented by manual record review. The cohort was derived from patients treated for HCV infection at the JPS Hepatology Clinic, which is part of the JPS Health Network, an urban public hospital network with a well-defined catchment area that covers Tarrant County, Texas (15th largest county in the United States with approximately 2 million residents). The 573-bed main hospital in Fort Worth, TX, and 40 satellite clinics comprise the health-care safety-net and serve as the primary source of care for over 261,000 uninsured individuals aged 18 to 65 in the county. Patients in the network are provided health care regardless of ability to pay. Patients eligible for our study were aged ≥18 years and initiated 8- or 12-week ledipasvir/sofosbuvir regimens for treatment-naive HCV-1 between February 2014 and June 2016. We excluded patients with HIV or cirrhosis (assessed by biopsy, ultrasound or computed tomography) for compatibility with current guidelines and to reduce the potential for deterministic non-positivity, which is a barrier to causal inference from observational data. Non-positivity occurs when certain characteristics preclude eligibility for a comparison treatment (that is, in this case ineligibility for the 8-week regimen) [20]. We did not restrict eligibility based on race/ethnicity for this analysis because this factor was not used for treatment decisions in our setting. In addition, we did not restrict eligibility based on viral load (<6 million IU/ml) because this criterion was not universally applied for selecting the 8-week regimen during the study period. Given payer restrictions, drug testing was required for the majority of patients to rule out active substance use. This study was approved by the North Texas Regional Institutional Review Board.

Variables

Our outcome of interest was SVR, which was defined as no detectable HCV RNA, measured by standard assays at least 12 weeks post-treatment completion [21]. The duration of ledipasvir/sofosbuvir was based on the final treatment duration. We also ascertained age, sex, self-reported race/ethnicity (non-Hispanic White, non-Hispanic Black or Hispanic) and self-reported history of illicit drug use (yes/no). We ascertained insurance type (private, government-based [Medicare or Medicaid] or uninsured) based on documented payer. Lastly, we ascertained body mass index (BMI) based on provider-measured height and weight, and diabetes status based on ICD-10 codes.

Data analysis

We used a counterfactual framework and marginal structural models [22,23] to compare effectiveness between 8- and 12-week regimens of ledipasvir/sofosbuvir for achieving SVR. In contrast to conventional methods that estimate conditional measures of association (that is, comparing patients who received 8-week ledipasvir/sofosbuvir with patients who received 12-week ledipasvir/sofosbuvir), a counterfactual approach compares SVR frequencies if everyone in the study population were exposed to 8-week ledipasvir/sofosbuvir and if everyone in the study population were exposed to 12-week ledipasvir/sofosbuvir (that is, average causal effects). We fit a logistic regression model to compute stabilized inverse probability of treatment weights [24], where covariates to reduce confounding bias were selected by applying the back-door criterion in a directed acyclic graph [25,26] of dependencies between duration of ledipasvir/sofosbuvir and SVR relevant to our setting (Directed acyclic graph illustrating dependency assumptions in Additional file 1). Given that eligibility of either regimen is based on similar criteria, confounding by HCV genotype, cirrhosis status and HIV status were effectively addressed by the study design (that is, restriction). The remaining covariates, age, race/ethnicity and insurance status were included in the treatment model to compute stabilized weights (mean =1.0, range =0.40–1.9). We fit a separate logistic regression model to compute stabilized inverse probability of censoring weights [24] to reduce potential bias from loss to follow-up and corresponding missing SVR status. Age, race/ethnicity, insurance status, current alcohol use, history of illicit drug use and ledipasvir/sofosbuvir treatment regimen were included in the censoring model to compute stabilized weights (mean =1.0, range =0.35–5.7). The product of stabilized weights from the treatment and censoring models (mean =1.0, range =0.32–8.3) were initially used in generalized linear models with a log-binomial distribution and log link to estimate risk ratios (RRs) and 95% confidence limits (CL), but some models did not consistently converge. Therefore, we used a Poisson distribution, log link and robust variance, which is a standard alternative for the log-binomial model when estimating RRs [27].

Effect heterogeneity

We assessed potential effect heterogeneity in the effect of 8- and 12-week ledipasvir/sofosbuvir on SVR using subgroup analyses for a priori selected characteristics: age category (<50 years, ≥50 years), sex, race/ethnicity (non-Hispanic Black, non-Hispanic White [insufficient sample size for Hispanic]) and diabetes (yes/no). To facilitate interpretation, we estimated the ratio of risk ratios (RRR) and corresponding 95% CL, where departures from RRR=1.0 may indicate effect heterogeneity between strata [28].

Sensitivity analysis

Confounding by indication (that is, characteristics that indicate or contraindicate a particular treatment regimen that also affect the outcome) is a concern in observational studies of treatment effect and no method entirely removes this bias [29]. Consequently, quantitative bias analysis related to unmeasured confounding is a useful approach for assessing the durability of the observed estimate [30]. We assessed the sensitivity of our estimates to unmeasured confounding using formulae by Ding and Vanderweele [31,32]. We varied the strength of effects of the unmeasured confounder on treatment and SVR to estimate the RR after adjustment for unmeasured confounding conditional on adjustment for covariates already included in the original estimation of the RR, which compared the effect of 8- and 12-week ledipasvir/sofosbuvir regimens on SVR.

Results

We identified 198 patients in the cohort who initiated 8- or 12-week ledipasvir/sofosbuvir regimens, of whom 7 were Asian or mixed ethnicity and excluded because of insufficient sample sizes for comparison (Figure 1). Table 1 summarizes the characteristics of our study population. The majority of the 12-week group were male (65%), whereas gender was more evenly distributed among the 8-week group. Among both treatment groups, the majority were aged ≥50 years (8-week: 71%; 12-week: 77%), non-Hispanic White (8-week: 57%; 12-week: 52%), obese (8-week: 42%; 12-week: 42%), uninsured (8-week: 84%; 12-week: 73%), and did not have pre-existing diabetes (8-week: 76%; 12-week: 72%) or a history of illicit drug use (8-week: 86%; 12-week: 78%). HCV viral load at 12 weeks post-treatment completion was not available for 23% of patients, but all of these patients completed their respective courses of treatment.

-
Figure 1.
Figure 1. Selection of HCV-infected patients treated with 8- and 12-week ledipasvir/sofosbuvir regimens
-
Table 1.  Characteristics of patients who initiated 8- and 12-week ledipasvir/sofosbuvir regimens for HCV infection
Table 1. Characteristics of patients who initiated 8- and 12-week ledipasvir/sofosbuvir regimens for HCV infection

a Missing value for some patients.

Table 2 summarizes the overall and subgroup-specific estimates comparing the effect of 8- and 12-week ledipasvir/sofosbuvir regimens on SVR. Overall, the risk of SVR was comparable between the 8- and 12-week regimens (RR=1.01, 95% CL: 0.92, 1.11). The risk of SVR modestly varied by subgroups of age (RRR=0.88, 95% CL: 0.74, 1.04) and diabetes (RRR=0.87; 95% CL: 0.72, 1.04) but not gender (RRR=1.01, 95% CL: N/A) or race/ethnicity (RRR=1.00, 95% CL: 0.82, 1.22).

-
Table 2.  Overall and subgroup-specific RRs and RDs comparing sustained virological response between 8- and 12-week ledipasvir/sofosbuvir regimensa
Table 2. Overall and subgroup-specific RRs and RDs comparing sustained virological response between 8- and 12-week ledipasvir/sofosbuvir regimensa

a 12-week ledipasvir/sofosbuvir is the reference category for the comparisons. b Ratio of risk ratios (RRR) comparing subgroup estimates for assessment of effect heterogeneity. c Unable to estimate confidence limits (CL). RD, risk difference; RR, risk ratio.

Figure 2 illustrates the sensitivity of our effect estimates to unmeasured confounding. The observed estimate assumes no unmeasured confounding (that is, RR=1.01). As the effect of the unmeasured confounder on treatment and SVR increases, the comparative effectiveness of the 8-week regimen decreases relative to the 12-week regimen. For example, an unmeasured confounder with an RR of 1.5 on treatment regimen and SVR, if adjusted, would result in an RR of 0.90 for 8-week ledipasvir/sofosbuvir therapy relative to 12-week therapy. An unmeasured confounder with an RR of 3.5, if adjusted, would result in an RR of 0.49 for 8-week ledipasvir/sofosbuvir relative to 12-weeks.

-
Figure 2.
Figure 2. Sensitivity of the estimated risk ratio to unmeasured confounding for varying strengths of effect of the unmeasured confounder on treatment and SVR

SVR, sustained virological response.

Discussion

We used a counterfactual framework to compare the effectiveness of 8- and 12-week ledipasvir/sofosbuvir regimens on SVR, where the eligibility criteria were based on current guidelines for 8-week ledipasvir/sofosbuvir except without restrictions by race/ethnicity or universal application of the HCV viral load criterion during the study period. This framework allows the following question to be asked: ‘What would be the average causal effect on SVR if everyone in the study population initiated 8-week ledipasvir/sofosbuvir compared with if everyone in the study population initiated 12-week ledipasvir/sofosbuvir?’ Our results suggest that the effect of the 8-week regimen on SVR is comparable to the 12-week regimen among treatment-naive HCV-1 patients who are non-cirrhotic and HIV-negative. We observed modest effect heterogeneity by age group and diabetes status, but the frequency of SVR exceeded 90% in each subgroup and small samples within subgroups preclude definitive interpretation of effect heterogeneity. Most importantly, 8- and 12-week regimens had comparable effectiveness among non-Hispanic Black individuals.

Several prior studies assessed the effectiveness of 8- and 12-week ledipasvir/sofosbuvir regimens in real-world settings (Summary of studies pertaining to the effectiveness of 8- and 12-week ledipasvir/sofosbuvir in Additional file 1). These individual studies and a meta-analysis [33] reported comparable effectiveness between regimens. Nevertheless, the evidence from prior studies is obscured by systematic errors that were common across studies and would only be exacerbated by summarizing estimates in a meta-analysis [14,34,35]. In particular, confounding by indication is an unavoidable bias in real-world (that is, observational) studies of treatment effects but was not addressed in prior studies. This bias is not a concern in randomized controlled trials because treatment allocation is a random process. Confounding by indication occurs when pre-treatment characteristics that indicate or contraindicate a particular regimen also affect the outcome of interest (that is, treatment allocation is non-random). For example, age, race/ethnicity and insurance status can influence treatment decisions and SVR, but these covariates were not adjusted in prior studies, where applicable. Several guidelines emphasize adjustment for pre-treatment characteristics that influence treatment decisions and the outcome to reduce bias in comparative effectiveness studies [15]. The unadjusted comparisons in prior studies may thus overestimate the effectiveness of 8-week ledipasvir/sofosbuvir on SVR because of better prognostic characteristics in this group. Only Marcus et al. [13] addressed confounding by indication and our estimates are consistent with estimates reported in their study.

We recently illustrated that prior studies may also be sensitive to selection bias from inappropriate exclusion of patients who had missing SVR status at 12 weeks post-treatment because of non-adherence to treatment or loss to follow-up [14]. The consequence of such exclusions is overestimated frequency of SVR. This overestimation may be particularly severe for studies where loss to follow-up is high [36]. More importantly, the estimates comparing these regimens would be biased if the magnitude of SVR overestimation is different between 8- and 12-week regimens. We used inverse probability of censoring (that is, loss to follow-up) weights to address this issue in the current study. Lastly, if treatment decisions for patients in prior studies were based on published guidelines, certain patients would be ineligible for the 8-week regimen because of host or viral characteristics and would have received the 12-week regimen. The inclusion of patients who were ineligible for 8-week ledipasvir/sofosbuvir when comparing with the 12-week regimen would create strata of the population without patients to compare, which is a violation of the positivity assumption [20]. Positivity is implicit in the design of randomized controlled trials because each individual has an equal probability of being allocated to the treatments of interest. A comparison may not have a meaningful interpretation when the positivity assumption is violated [20]. Consequently, our eligibility criteria were based on reducing potential non-positivity.

Our finding of comparable effectiveness between 8- and 12-week regimens for non-Hispanic Black individuals opposes most prior real-world studies, but comparisons with prior studies are tenuous given the aforementioned limitations. In addition, prior studies compared SVR between Black individuals and non-Black individuals across or within treatment regimens [6,3739], which does not address the question of whether 8- and 12-week ledipasvir/sofosbuvir regimens are comparable within racial/ethnic subgroups. Marcus et al. [13] directly addressed comparative effectiveness between 8- and 12-week regimens among Black patients and reported an RR consistent with our estimate (Black: RR=1.00, 95% CL: 0.95, 1.04). In addition, Wilder et al. [40] aimed to combine data from the ION trials to address potential differences within racial/ethnic groups, but the simple pooling method used in their analysis is well-known to generate spurious and even counterintuitive results [4143]. Consequently, we re-analysed data from the ION-3 trial [3] to provide further insight because this trial was specifically designed to assess non-inferiority of 8- and 12-week ledipasvir/sofosbuvir and the randomized design reduces concerns about confounding bias. The RR comparing 8- and 12-week ledipasvir/sofosbuvir without ribavirin in the ION-3 trial was 0.96 (95% CL: 0.85, 1.07) among Black individuals and 0.99 (95% CL: 0.95, 1.04) among non-Black individuals (Re-analysis of comparative effectiveness of 8- and 12-week ledipasvir/sofosbuvir in Additional file 1).

For greater precision in estimates comparing 8- and 12-week ledipasvir/sofosbuvir regimens by race/ethnicity, we used fixed effect meta-analysis to summarize estimates from our study, the Marcus et al. [13] study, and ION-3 trial [3], which were the few that addressed confounding by indication as part of the design or analysis. The summary RRs (Fixed effect meta-analysis of results in Additional file 1) were similar for Black individuals (RR=1.00, 95% CL: 0.96, 1.03) and non-Black individuals (RR=1.00, 95% CL: 0.99, 1.02). Heterogeneity in study designs, patient populations and categories for racial/ethnic groups notwithstanding, the summary estimate does not provide evidence of lower SVR among Black individuals. More importantly, unlike interferon-based treatments that were sensitive to IL28B genotype, biological differences between racial/ethnic groups in response to ledipasvir/sofosbuvir would not be anticipated because of the mechanism of action of direct-acting antivirals [44].

Despite addressing some limitations of prior studies, limitations of our analysis should also be considered. We initially deliberated designing the study to address whether 8-week ledipasvir/sofosbuvir was non-inferior to the 12-week regimen but anticipated insufficient power to use a non-inferiority approach. Larger sample sizes would allow more flexible designs. We addressed confounding by indication through inverse probability weighting of relevant pre-treatment characteristics, but no current approach entirely removes this bias from real-world studies of treatment effects [29]. HCV viral load was not universally applied for selecting the 8-week regimen during the study period in our setting. If HCV viral load was systematically different between treatment regimens and affects SVR, then our estimates could be sensitive to unmeasured confounding, but the influence of HCV viral load on SVR is debated [45]. Nevertheless, other provider perceptions and preferences could influence treatment selection, which could manifest as confounding by indication [46]. For example, concerns about regimen adherence because of factors such as mental health or transportation availability for the patient could influence treatment selection [47]. We explored the sensitivity of our estimates to unmeasured confounding and observed that even modest unmeasured confounding (Figure 2) could be an alternate explanation for the apparent comparability of 8- and 12-week regimens not only in our study but also prior studies. For example, an unmeasured confounder with an RR of 1.5 on treatment regimen and SVR, if measured and adjusted, would result in a 10% lower SVR (that is, RR=0.90) for 8-week ledipasvir/sofosbuvir relative to 12-weeks. Future studies may need to consider explicitly measuring and adjusting for provider preferences to empirically assess the magnitude of bias.

Similar to prior studies, our study may be sensitive to misclassification of treatment regimen. For example, if patients were initially scheduled for 12-week ledipasvir/sofosbuvir, but early virological response was observed, then providers could have truncated the treatment duration. Similarly, if patients were initially scheduled for the 8-week regimen, but virological response was not observed, then providers could have extended the treatment duration to 12 weeks. Such misclassification could exacerbate confounding by indication and overestimate the effect of 8-week ledipasvir/sofosbuvir on SVR. Nevertheless, we anticipate negligible impact of treatment misclassification because only one patient was extended from 8-weeks to 12-weeks in our study. Future studies should consider an initial documentation of treatment regimen, which would allow for assessing the magnitude of misclassification at the end of study.

In summary, assuming no substantial impact of biases on our study, our findings could have several implications. Our results suggest that 8- and 12-week ledipasvir/sofosbuvir have comparable effects on SVR in a real-world setting among HCV-1 patients without cirrhosis or HIV. The comparable effectiveness of 8- and 12-week regimens among non-Hispanic Black individuals based on evidence from our study, Marcus et al. [13] and simulations that explored cost-effectiveness of these approaches without the need for risk-stratification [16,17], adds to the growing body of evidence that supports the removal of race-based treatment guidelines. Alternatively, the 8-week regimen of glecaprevir/pibrentasvir, which does not have race-based restriction in the treatment guidelines [4], may be a more compelling option than ledipasvir/sofosbuvir because of high effectiveness and lower cost. Nevertheless, several payers established agreements for discounted ledipasvir/sofosbuvir and maintain it as a preferred treatment. The emergence of generic ledipasvir/sofosbuvir also dramatically reduces cost to $600 for an 8-week regimen and emerging reports suggest high frequency of SVR [48]. In addition, glecaprevir/pibrentasvir increases pill burden, which may be a barrier to adherence, particularly for HCV populations with a high prevalence of comorbidities. Therefore, ledipasvir/sofosbuvir may remain relevant in the evolving landscape of HCV treatment, but the apparent advantages of a short-course regimen may need to be tempered by the inevitable challenges of interpreting real-world studies and balancing trade-offs.

Acknowledgements

The authors are grateful to Sajid Shaikh (Information Technology, JPS Health Network, Fort Worth, TX, USA) for assistance with data collection and to the anonymous reviewers for helpful suggestions.

Disclosure statement

This study did not receive funding support from any source. The authors declare no financial or non-financial conflicts of interest.

Additional file

Additional file 1: Additional materials can be found at https://www.intmedpress.com/uploads/documents/4340_Ojha_Addfile1.pdf

References

-
1. Qian XJ, Zhu YZ, Zhao P, Qi ZT. Entry inhibitors: new advances in HCV treatment. Emerg Microbes Infect 2016; 5: e3 Medline doi:10.1038/emi.2016.3
-
2. Afdhal N, Zeuzem S, Kwo P, et al. Ledipasvir and sofosbuvir for untreated HCV genotype 1 infection. N Engl J Med 2014; 370: 1889-1898. Medline doi:10.1056/NEJMoa1402454
-
3. Kowdley KV, Gordon SC, Reddy KR, et al. Ledipasvir and sofosbuvir for 8 or 12 weeks for chronic HCV without cirrhosis. N Engl J Med 2014; 370: 1879-1888. Medline doi:10.1056/NEJMoa1402355
-
4. AASLD-IDSA. HCV guidance: recommendations for testing, managing, and treating hepatitis C: AASLD-IDSA; 2017. (Accessed 16 February 2018.) Available from http://www.hcvguidelines.org/full-report/hcv-testing-and-linkage-care
-
5. Backus LI, Belperio PS, Shahoumian TA, Loomis TP, Mole LA. Real-world effectiveness of ledipasvir/sofosbuvir in 4,365 treatment-naive, genotype 1 hepatitis C-infected patients. Hepatology 2016; 64: 405-414. Medline doi:10.1002/hep.28625
-
6. Backus LI, Belperio PS, Shahoumian TA, Loomis TP, Mole LA. Real-world effectiveness and predictors of sustained virological response with all-oral therapy in 21,242 hepatitis C genotype-1 patients. Antivir Ther 2017; 22: 481-493. Medline doi:10.3851/IMP3117
-
7. Terrault NA, Zeuzem S, Di Bisceglie AM, et al. Effectiveness of ledipasvir-sofosbuvir combination in patients with hepatitis C virus infection and factors associated with sustained virologic response. Gastroenterology 2016; 151: 1131-1140.e5. Medline doi:10.1053/j.gastro.2016.08.004
-
8. Buggisch P, Vermehren J, Mauss S, et al. Real-world effectiveness of 8-week treatment with ledipasvir/sofosbuvir in chronic hepatitis C. J Hepatol 2017; 68: 663-671. Medline doi:10.1016/j.jhep.2017.11.009
-
9. Johnson SW, Ammirati SR, Hartis CE, et al. Effectiveness of ledipasvir/sofosbuvir in real-world patients with chronic hepatitis C: a collaborative treatment approach. Int J Antimicrob Agents 2017; 49: 778-781. Medline doi:10.1016/j.ijantimicag.2017.01.016
-
10. Lai JB, Witt MA, Pauly MP, et al. Eight- or 12-week treatment of hepatitis C with ledipasvir/sofosbuvir: real-world experience in a large integrated health system. Drugs 2017; 77: 313-318. Medline doi:10.1007/s40265-016-0684-y
-
11. Nguyen MH, Trinh H, Do S, Nguyen T, Nguyen P, Henry L. Open label study of 8 vs. 12 weeks of ledipasvir/sofosbuvir in genotype 6 treatment naive or experienced patients. Am J Gastroenterol 2017; 112: 1824-1831. Medline doi:10.1038/ajg.2017.399
-
12. Curry MP, Tapper EB, Bacon B, et al. Effectiveness of 8- or 12-weeks of ledipasvir and sofosbuvir in real-world treatment-naive, genotype 1 hepatitis C infected patients. Aliment Pharmacol Ther 2017; 46: 540-548. Medline doi:10.1111/apt.14204
-
13. Marcus JL, Hurley LB, Chamberland S, et al. No difference in effectiveness of 8 vs 12 weeks of ledipasvir and sofosbuvir for treatment of hepatitis C in Black patients. Clin Gastroenterol Hepatol 2018; 16: 927-935. Medline doi:10.1016/j.cgh.2018.03.003
-
14. Ojha RP, Steyerberg EW. Real-world data on antiviral treatments for hepatitis C virus infections: can we define intention to treat or per protocol analyses? J Hepatol 2018; 69: 551-553. Medline doi:10.1016/j.jhep.2018.02.037
-
15. Morton JB, McConeghy R, Heinrich K, Gatto NM, Caffrey AR. Consensus of recommendations guiding comparative effectiveness research methods. Pharmacoepidemiol Drug Saf 2016; 25: 1354-1360. Medline doi:10.1002/pds.4051
-
16. Chahal HS, Marseille EA, Tice JA, et al. Cost-effectiveness of early treatment of hepatitis C virus genotype 1 by stage of liver fibrosis in a US treatment-naive population. JAMA Intern Med 2016; 176: 65-73. Medline doi:10.1001/jamainternmed.2015.6011
-
17. Morgan JR, Kim AY, Naggie S, Linas BP. The effect of shorter treatment regimens for hepatitis C on population health and under fixed budgets. Open Forum Infect Dis 2017; 5: ofx267 Medline
-
18. Mathes T, Pieper D. Study design classification of registry-based studies in systematic reviews. J Clin Epidemiol 2018; 93: 84-87. Medline doi:10.1016/j.jclinepi.2017.09.016
-
19. Casey JA, Schwartz BS, Stewart WF, Adler NE. Using electronic health records for population health research: a review of methods and applications. Annu Rev Public Health 2016; 37: 61-81. Medline doi:10.1146/annurev-publhealth-032315-021353
-
20. Westreich D, Cole SR. Invited commentary: positivity in practice. Am J Epidemiol 2010; 171: 674-677. Medline doi:10.1093/aje/kwp436
-
21. Yoshida EM, Sulkowski MS, Gane EJ, et al. Concordance of sustained virological response 4, 12, and 24 weeks post-treatment with sofosbuvir-containing regimens for hepatitis C virus. Hepatology 2015; 61: 41-45. Medline doi:10.1002/hep.27366
-
22. Danaei G, Garcia Rodriguez LA, Cantero OF, Logan RW, Hernan MA. Electronic medical records can be used to emulate target trials of sustained treatment strategies. J Clin Epidemiol 2018; 96: 12-22. Medline doi:10.1016/j.jclinepi.2017.11.021
-
23. Hernán MA, Robins JM. Estimating causal effects from epidemiological data. J Epidemiol Community Health 2006; 60: 578-586. Medline doi:10.1136/jech.2004.029496
-
24. Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol 2008; 168: 656-664. Medline doi:10.1093/aje/kwn164
-
25. Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology 1999; 10: 37-48. Medline doi:10.1097/00001648-199901000-00008
-
26. Pearl J. Causality: models, reasoning, and inference. 2 ed. Cambridge, England: Cambridge University Press 2009.
-
27. Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159: 702-706. Medline doi:10.1093/aje/kwh090
-
28. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ 2003; 326: 219 Medline doi:10.1136/bmj.326.7382.219
-
29. Bosco JL, Silliman RA, Thwin SS, et al. A most stubborn bias: no adjustment method fully resolves confounding by indication in observational studies. J Clin Epidemiol 2010; 63: 64-74. Medline doi:10.1016/j.jclinepi.2009.03.001
-
30. Lash TL, Fox MP, MacLehose RF, Maldonado G, McCandless LC, Greenland S. Good practices for quantitative bias analysis. Int J Epidemiol 2014; 43: 1969-1985. Medline doi:10.1093/ije/dyu149
-
31. VanderWeele TJ, Ding P. Sensitivity analysis in observational research: introducing the E-value. Ann Intern Med 2017; 167: 268-274. Medline doi:10.7326/M16-2607
-
32. Ding P, VanderWeele TJ. Sensitivity analysis without assumptions. Epidemiology 2016; 27: 368-377. Medline doi:10.1097/EDE.0000000000000457
-
33. Kowdley KV, Sundaram V, Jeon CY, et al. Eight weeks of ledipasvir/sofosbuvir is effective for selected patients with genotype 1 hepatitis C virus infection. Hepatology 2017; 65: 1094-1103. Medline doi:10.1002/hep.29005
-
34. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ 1994; 309: 1351-1355. Medline doi:10.1136/bmj.309.6965.1351
-
35. Lau J, Ioannidis JP, Schmid CH. Summing up evidence: one answer is not always enough. Lancet 1998; 351: 123-127. Medline doi:10.1016/S0140-6736(97)08468-7
-
36. Hernán MA, Hernandez-Diaz S, Robins JM. A structural approach to selection bias. Epidemiology 2004; 15: 615-625. Medline doi:10.1097/01.ede.0000135174.63482.43
-
37. Su F, Green PK, Berry K, Ioannou GN. The association between race/ethnicity and the effectiveness of direct antiviral agents for hepatitis C virus infection. Hepatology 2017; 65: 426-438. Medline doi:10.1002/hep.28901
-
38. O’Brien TR, Lang Kuhs KA, Pfeiffer RM. Subgroup differences in response to 8 weeks of ledipasvir/sofosbuvir for chronic hepatitis C. Open Forum Infect Dis 2014; 1: ofu110 Medline doi:10.1093/ofid/ofu110
-
39. Ioannou GN, Beste LA, Chang MF, et al. Effectiveness of sofosbuvir, ledipasvir/sofosbuvir, or paritaprevir/ritonavir/ombitasvir and dasabuvir regimens for treatment of patients with hepatitis C in the Veterans Affairs National Health Care System. Gastroenterology 2016; 151: 457-471.e5. Medline doi:10.1053/j.gastro.2016.05.049
-
40. Wilder JM, Jeffers LJ, Ravendhran N, et al. Safety and efficacy of ledipasvir-sofosbuvir in black patients with hepatitis C virus infection: a retrospective analysis of Phase 3 data. Hepatology 2016; 63: 437-444. Medline doi:10.1002/hep.28334
-
41. Hernán MA, Clayton D, Keiding N. The Simpson’s paradox unraveled. Int J Epidemiol 2011; 40: 780-785. Medline doi:10.1093/ije/dyr041
-
42. Altman DG, Deeks JJ. Meta-analysis, Simpson’s paradox, and the number needed to treat. BMC Med Res Methodol 2002; 2: 3 Medline doi:10.1186/1471-2288-2-3
-
43. Bravata DM, Olkin I. Simple pooling versus combining in meta-analysis. Eval Health Prof 2001; 24: 218-230. Medline doi:10.1177/01632780122034885
-
44. Jensen CM, Holle LM. Ledipasvir-sofosbuvir: a once-daily oral treatment option for chronic hepatitis C virus genotype 1 infection. Pharmacotherapy 2016; 36: 562-574. Medline doi:10.1002/phar.1748
-
45. O’Brien TR, Feld JJ, Kottilil S, Pfeiffer RM. No scientific basis to restrict 8 weeks of treatment with ledipasvir/sofosbuvir to patients with hepatitis C virus RNA <6,000,000 IU/ml. Hepatology 2016; 63: 28-30. Medline doi:10.1002/hep.28292
-
46. Brookhart MA, Sturmer T, Glynn RJ, Rassen J, Schneeweiss S. Confounding control in healthcare database research: challenges and potential approaches. Med Care 2010; 48: S114-S120. Medline doi:10.1097/MLR.0b013e3181dbebe3
-
47. Rogal SS, McCarthy R, Reid A, et al. Primary care and hepatology provider-perceived barriers to and facilitators of hepatitis C treatment candidacy and adherence. Dig Dis Sci 2017; 62: 1933-1943. Medline doi:10.1007/s10620-017-4608-9
-
48. Zeng QL, Xu GH, Zhang JY, et al. Generic ledipasvir-sofosbuvir for patients with chronic hepatitis C: a real-life observational study. J Hepatol 2017; 66: 1123-1129. Medline doi:10.1016/j.jhep.2017.01.025

Copyright © 2019 Nucleus Holdings Ltd. Part of Nucleus Global.
Design and Technology by Nucleus Global
Company registration No. 321 0712 (England & Wales). Registered Office address: Admiral House 76-78 Old Street, London EC1V 9AZ.