Obeticholic Acid for the Treatment of Primary Biliary Cholangitis: Comparative Clinical Effectiveness, Value, and Value-Based Price Benchmarks Evidence Report July 26, 2016 Institute for Clinical and Economic Review ©Institute for Clinical and Economic Review, 2016 AUTHORS Massachusetts General Hospital, Institute for ICER Staff Technology Assessment Modeling Group Steven D. Pearson, MD, MSc Jagpreet Chhatwal, PhD President, Institute for Clinical and Economic Review Assistant Professor, Harvard Medical School Senior Scientist, MGH Institute for Technology Reiner Banken, MD, MSc Assessment Research Fellow, Institute for Clinical and Economic Review Chin Hur, MD, MPH Associate Professor, Harvard Medical School Rick Chapman, PhD, MS Director, GI Health Outcomes Research, Massachusetts Director of Health Economics, Institute for Clinical and General Hospital Economic Review Associate Director, MGH Institute for Technology Assessment Sarah K. Emond, MPP Chief Operating Officer, Institute for Clinical and Matthew Klebanoff, BA Economic Review Research Analyst, MGH Institute for Technology Assessment Anne M. Loos, MA Research Associate, Institute for Clinical and Economic Sumeyye Samur, PhD Review Post-doctoral Fellow, Harvard Medical School Research Fellow, MGH Institute for Technology Elizabeth T. Russo, MD Assessment Research Scientist, Institute for Clinical and Economic Review The role of the Massachusetts General Hospital, Institute for Technology Assessment is limited to the development of the cost- effectiveness model, and the resulting ICER reports do not Celia S. Segel, MPP necessarily represent the views of the MGH. Program Manager, New England CEPAC, Institute for Clinical and Economic Review Daniel A. Ollendorf, PhD Chief Scientific Officer, Institute for Clinical and Economic Review DATE OF PUBLICATION: July 26, 2016 We would also like to thank Erin Lawler, MA and Shanshan Liu, MS, MPH for their contributions to this report. ©Institute for Clinical and Economic Review, 2016 Page i Draft Evidence Report – OCA for the Treatment of PBC About ICER The Institute for Clinical and Economic Review (ICER) is an independent non-profit research organization that evaluates medical evidence and convenes public deliberative bodies to help stakeholders interpret and apply evidence to improve patient outcomes and control costs. ICER receives funding from government grants, non-profit foundations, health plans, provider groups, and health industry manufacturers. For a complete list of funders, visit http://www.icer- review.org/about/support/. Through all its work, ICER seeks to help create a future in which collaborative efforts to move evidence into action provide the foundation for a more effective, efficient, and just health care system. More information about ICER is available at http://www.icer- review.org About New England CEPAC The New England Comparative Effectiveness Public Advisory Council (New England CEPAC) – a core program of ICER – provides a public venue in which the evidence on the effectiveness and value of health care services can be discussed with the input of all stakeholders. New England CEPAC seeks to help patients, clinicians, insurers, and policymakers interpret and use evidence to improve the quality and value of health care. The New England CEPAC is an independent committee of medical evidence experts from across New England, with a mix of practicing clinicians, methodologists, and leaders in patient engagement and advocacy. All Council members meet strict conflict of interest guidelines and are convened to discuss the evidence summarized in ICER reports and vote on the comparative clinical effectiveness and value of medical interventions. More information about New England CEPAC is available at http://icer-review.org/programs/new-england-cepac/. ©Institute for Clinical and Economic Review, 2016 Page ii Draft Evidence Report – OCA for the Treatment of PBC External Input The following individuals and organizations provided input and feedback that helped guide the ICER team as we shaped our scope and report. None of these individuals or organizations is responsible for the final contents of this report or should be assumed to support any part of this report, which is solely the work of the ICER team and its affiliated researchers. External Input Received from: Clinical Experts: Keith Lindor, MD, Mayo Clinic Daniel Pratt, MD, Massachusetts General Hospital* Harvard Pilgrim HealthCare Intercept Pharmaceuticals * Also provided input into the development of the economic model ©Institute for Clinical and Economic Review, 2016 Page iii Draft Evidence Report – OCA for the Treatment of PBC Table of Contents Executive Summary ................................................................................................................................ 1 1. Background ........................................................................................................................................ 1 1.1 Introduction ................................................................................................................................. 1 2. The Topic in Context .......................................................................................................................... 5 3. Summary of Coverage Policies ........................................................................................................... 9 4. Comparative Clinical Effectiveness .................................................................................................. 10 4.1 Overview .................................................................................................................................... 10 4.2 Methods ..................................................................................................................................... 10 4.3 Results ........................................................................................................................................ 13 5. Other Benefits or Disadvantages ..................................................................................................... 26 6. Comparative Value ........................................................................................................................... 27 6.1 Overview .................................................................................................................................... 27 6.2 Prior Published Evidence on Costs and Cost-Effectiveness of Obeticholic Acid ........................ 27 6.3 Incremental Costs per Outcome Achieved ................................................................................ 27 6.4 Potential Budget Impact ............................................................................................................ 43 6.5 Value-based Benchmark Prices .................................................................................................. 46 6.6 Summary and Comment ............................................................................................................ 47 7. Voting Results .................................................................................................................................. 49 7.1 About the New England CEPAC Process .................................................................................... 49 7.2 Clinical Effectiveness Voting Results .......................................................................................... 51 7.3 Care Value Voting Results .......................................................................................................... 52 8. Roundtable Discussions and Key Policy Implications....................................................................... 53 References ........................................................................................................................................... 58 Appendix A. Evidence Review Methods and PRISMA .......................................................................... 63 Appendix B. Summary Evidence Table................................................................................................. 68 Appendix C. Comparative Value Supplemental Information ............................................................... 78 Appendix D. Clinical Guidelines ........................................................................................................... 79 Appendix E. Previous Systematic Reviews and Technology Assessments ........................................... 80 ©Institute for Clinical and Economic Review, 2016 Page iv Draft Evidence Report – OCA for the Treatment of PBC Appendix F. Ongoing Studies ............................................................................................................... 81 Appendix G. Policy Roundtable Conflict of Interest Disclosures ......................................................... 83 Appendix H. Public Comments ............................................................................................................. 85 Appendix I. Coverage Policy at Harvard Pilgrim Health Care ............................................................... 88 ©Institute for Clinical and Economic Review, 2016 Page v Draft Evidence Report – OCA for the Treatment of PBC List of Acronyms Used in this Report AE Adverse event ALP Alkaline phosphatase ALT Alanine aminotransferase AST Aspartate aminotransferase AMA Antimitochondrial antibodies DB Double-blind FDA Food and Drug Administration GGT Gamma-glutamyl transferase HCC Hepatocellular carcinoma HDL High-density lipoprotein ICER Increased cost effectiveness ratio ITT Intent-to-treat LTSE Long-term safety extension LDL Low-density lipoprotein mITT Modified intent-to-treat OCA Obeticholic acid PBC Primary biliary cholangitis, also known as primary biliary cirrhosis QALY Quality-adjusted life year QoL Quality-of-life RCT Randomized controlled trial SAE Serious adverse event TB Total bilirubin UDCA Ursodeoxycholic acid ULN Upper limit of normal US United States USPSTF United States Preventive Services Task Force ©Institute for Clinical and Economic Review, 2016 Page vi Draft Evidence Report – OCA for the Treatment of PBC Executive Summary Background Primary biliary cholangitis (PBC), which has until recently been referred to as primary biliary cirrhosis,1 is a rare, chronic, progressive autoimmune liver disease that affects around 130,000 individuals in the US,2 mainly middle-aged women.3 Diagnosis is increasingly occurring among asymptomatic patients triggered through an investigation of increased levels of alkaline phosphatase (ALP) on routine blood tests.4 The disease process progresses over many years from autoimmune damage to small intrahepatic bile ducts with chronic cholestasis and portal inflammation, to fibrosis that can lead to cirrhosis and, ultimately, liver failure.5 Fatigue and pruritus are the most common symptoms of PBC, and both can be debilitating in some patients, especially fatigue.6 The disease progression can range from no development of cirrhosis over nearly 20 years of median follow-up in patients who were asymptomatic at diagnosis,7 to a median survival of 6 to 10 years without liver transplantation among patients who are diagnosed once symptomatic.6 Hepatocellular carcinoma (HCC) is an infrequent yet critical consequence of PBC.8 Topic in Context Until recently, UDCA was the only FDA approved treatment for PBC. At a daily dose of 13–15 mg/kg, treatment with UDCA improves hepatic biochemistry, slows histological progression to fibrosis, delays and reverses portal hypertension (which develops prior to cirrhosis in PBC), and delays the requirement for liver transplantation.6,9 However, treatment with UDCA does not improve fatigue and pruritus.6 Lifetime treatment is recommended for all patients with PBC. Patients with early stage disease treated with UDCA have an overall survival similar to the general population. For patients with moderate to severe disease, UDCA treatment significantly improves average time to requirement for liver transplantation.10 UDCA has minimal side effects.4 The most frequently reported adverse events (AEs) of the drug include loose stool (2–9%), headache, and mild weight gain, but these rarely lead to discontinuation.6 Levels of intolerance to UDCA are not well-documented in the scientific literature. According to Intercept market research, around 3% of patients are intolerant to UDCA.2 However, the clinical practice guideline of American Association for the Study of Liver Diseases does not mention intolerance to UDCA.4 Improvement in ALP levels usually starts within a few weeks on UDCA, and 90% of improvement occurs within six to nine months. For some patients, the treatment effect does not appear until up ©Institute for Clinical and Economic Review, 2016 ES 1 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents to five years on UDCA.4 Most of the criteria for defining a positive biochemical response to UDCA use a cut-off point of 12 months to determine whether a response has been achieved.6 Patients with an inadequate response to UDCA are at a higher risk of disease progression compared to those who show improvement.6 Reported nonresponse rates vary between 24% and 79% depending on the criteria used,11 but a rate of approximately 40% of patients with an inadequate response to treatment with UDCA is generally considered an appropriate estimate.6 Obeticholic Acid Obeticholic acid (OCA) (Ocaliva™, Intercept Pharmaceuticals, Inc.) is a novel oral bile acid analogue that has shown positive effects on biochemical markers of liver function in patients with PBC. Following an accelerated approval pathway, on May 27, 2016 the FDA approved the use of OCA for the treatment of PBC in combination with ursodeoxycholic acid (UDCA) in adults with an inadequate response to UDCA, or as monotherapy in adults unable to tolerate UDCA.12,13 OCA is taken once daily in doses of 5-10 mg. The purpose of this review was to assess the comparative clinical effectiveness and comparative value of adding OCA to UDCA in patients with an inadequate response to UDCA, or as monotherapy in patients intolerant to UDCA, relative to standard care for PBC. Comparative Clinical Effectiveness The timeframe for our search for clinical evidence spanned the period from January 1996 to June 20, 2016. The evidence base comes from three industry-sponsored clinical trials, which are presented in further detail in Table ES1. Enrolled patients were had a mean age of 55 years, 92% were female, and 95% had normal bilirubin at baseline which is characteristic of early stage disease. All RCTs included patients with elevated ALP (≥1.5xULN), and excluded patients with a history or presence of hepatic decompensation and bilirubin greater than twice the ULN. Two trials focused on combination therapy with OCA and UDCA in patients with inadequate response to UDCA, while one small trial that has not yet been published examined OCA monotherapy in patients who had discontinued UDCA a minimum of three months prior to enrollment. All trials were designed primarily to measure changes in ALP, including percentage or absolute mean reductions in ALP from baseline, as well as ALP normalization. This means that they do not measure clinical outcomes directly, but rather a change in ALP as a surrogate measure of clinical benefit. Other liver biomarkers, including, aspartate aminotransferase (AST), alanine aminotransferase (ALT), gamma-glutamyl transferase (GGT) and bilirubin, were also measured in all studies; these tests indicate liver function and are used by clinicians to follow the disease process. Treatment response was compared to different published risk models (e.g., ALP <1.67 x the upper limit of normal [ULN] and total bilirubin [TB] <ULN) that may reasonably predict important clinical ©Institute for Clinical and Economic Review, 2016 ES 2 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents outcomes for PBC patients (e.g., transplant-free survival and mortality). We also described the most frequently-reported harms as reported in the trials, with a particular focus on drug-related adverse events (e.g., pruritus). Importantly, the outcomes reported from one of these studies (Study 202) were primarily based on the published study,14 while results from the other two key trials of OCA for PBC (Studies 201 and 301) were only available in the grey literature. Table ES1. Key Trials Study 202 Study 301 RCT Study 201 Hirschfield et al. (POISE) N=59 (ITT) N=161 (ITT, N=165) N=216 (ITT, N=217) Mean age: 55 years Mean age: 55 years Mean age: 56 years Female: 85% Female: 95% Female: 91% Patient Characteristics Mean ALP: 432.6 U/L Mean ALP: 286.9 U/L Mean ALP: 323.2 U/L Mean UDCA dose: 15.9 Mean UDCA dose: 15.4 mg/day mg/day OCA 10 mg OCA 10 mg OCA 5-10 mg OCA 50 mg OCA 25 mg OCA 10 mg Interventions/Comparators Placebo OCA 50 mg Placebo Placebo DB: 3 months DB: 3 months DB: 1 year Duration LTSE: Up to 4.5 years LTSE: 1 year LTSE: 5 years (ongoing) OCA as monotherapy OCA+UDCA for patients OCA±UDCA for patients with for patients not on with inadequate inadequate response to UDCA UDCA for ≥3 months response to UDCA (on UDCA for ≥12 months and (stable dose ≥6 stable dose ≥3 months), or Inclusion criteria ALP 1.5-10xULN months) intolerant to UDCA (7%) ALP 1.5-10xULN ALP ≥1.67xULN or bilirubin >1xULN but <2xULN OCA 10 mg: -44.5 OCA 10 mg: -23.7 OCA 5-10 mg: -33.0 OCA 50 mg: -37.6 OCA 25 mg: -24.7 OCA 10 mg: -39.1 Placebo: +11.7 OCA 50 mg: -21.0 Placebo: -4.8 % Change in ALP Placebo: -3.1 All OCA groups from All OCA groups vs. placebo, baseline, p<0.0001 All OCA groups from p<0.0001 baseline, p<0.0001 ITT = intent-to-treat; DB = double-blind; LTSE = long-term safety extension; ULN = upper limit of normal; UDCA = ursodeoxycholic acid; OCA = obeticholic acid; ALP = alkaline phosphatase ©Institute for Clinical and Economic Review, 2016 ES 3 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Results Clinical Benefits In all of these randomized controlled trials there was a statistically-significant reduction in ALP from baseline for the OCA-treated groups compared to controls (i.e., placebo plus UDCA or placebo alone). After the randomized controlled phase of 12 months in Study 301 (POISE), this improvement was sustained for another nine months of the open-label long-term safety extension (LTSE) period. In the open-label LTSE of Study 201 of OCA monotherapy, a statistically significant mean absolute reduction in ALP from baseline was sustained for the 11 patients who completed the 4.5 years of follow-up.15 Treatment response can be measured by different published risk models that may reasonably predict important clinical outcomes for PBC, which are also used to define inadequate response to UDCA.6 Table ES2 shows the proportion of patients achieving the primary endpoint in POISE, which was defined as mean ALP <1.67xULN with ≥15% reduction, and normal bilirubin. Table ES2. Proportion of Trial Patients Achieving the POISE Primary Endpoint16 % of Patients Study 201, n=43 Study 202, n=76 Study 301, n=146 Pooled, n=306 Placebo 4 8 10 8 OCA 10 mg 40 42 47 45 (all OCA groups) p-value p=0.0026 p=0.0002 p<0.0001 p<0.0001 Overall, other measures of liver function, such as ALT, AST and GGT, were favorably influenced by treatment with OCA. Results from POISE demonstrated statistically significant improvements with OCA compared to controls for ALT (-35.5% and -41.7% vs. -4.7% for placebo, respectively) and GGT (-50.3% and -63.7% vs. 0.8% for placebo, respectively) (all outcomes, p<0.0001).17 Harms Treatment-related adverse effects were generally mild to moderate in nature, and primarily related to pruritus. The main reason for discontinuation in the trials was severe pruritus, with increased frequency in patients receiving higher doses. Although pruritus is a symptom of PBC, the available evidence suggests there is increased frequency and severity of pruritus due to OCA treatment. Among the different dosages used in the trials, the incidence of pruritus was lowest in the titration dosage from 5 mg to 10 mg now recommended in the FDA label.13 Changes in lipid levels, mostly due to reductions in high-density lipoprotein (HDL), were also common in all OCA-treated patients. Although treatment-related reductions are concerning given the protective nature of HDL against cardiovascular morbidity, it is unclear if these changes were clinically significant. Other serious adverse events (SAEs) occurred with much less frequency and ©Institute for Clinical and Economic Review, 2016 ES 4 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents included hepatic decompensation, gastrointestinal disorders, and hyperbilirubinemia; questions remain as to whether these events were correlated with OCA or manifested independently as a result of progressing disease. With the exception of some pruritus outcomes in Study 202, 14 these events were not evaluated statistically, likely due to the overall limited occurrence. Controversies and Uncertainties Several limitations in the body of evidence reduce our ability to make judgments regarding the comparative net health benefits of OCA. Data on clinically-relevant outcomes, including transplant- free survival and mortality, are not yet available. Surrogate endpoints, such as reduction in ALP, have not been previously adopted by the FDA as acceptable criteria for regulatory approval of new PBC regimens. As discussed at the FDA Advisory Committee meeting, there remains some uncertainty surrounding the appropriate definition of treatment response (and specifically the cut- off points used in POISE) for PBC patients, particularly for those patients with elevated ALP and normal bilirubin, as well as the clinical significance of treatment-related reductions in HDL. Concerning the use of OCA as monotherapy, data on the use of OCA without concomitant use of UDCA is primarily limited to results from two trials in conference abstracts and regulatory documents. Additionally, baseline values of ALP were higher for the Phase II monotherapy cohort, and only 16 patients from the POISE trial were considered intolerant to UDCA and taking OCA alone. Some experts suggest that patients taking UDCA with inadequate response may see additional benefits of the drug for up to five years of ongoing treatment.4 However, because there were no additional safety concerns in these patients even after 4.5 years of follow-up, it is unlikely that OCA taken as monotherapy would represent a unique issue for patients intolerant to UDCA, particularly since there is no other available treatment option for these patients in the US.15 Perhaps the greatest amount of uncertainty in comparative net health benefit lies in the lack of data available for patients in later stages of their disease. Across the three trials of PBC, only 46 patients (11%) had abnormal bilirubin at baseline. Therefore, further study should elucidate OCA’s performance at different points during the disease course. Comparative Clinical Effectiveness: Summary and Comment Considering that no evidence is available on clinically-meaningful outcomes for treatment with OCA, we assigned ICER evidence ratings based on the reporting of improvements on surrogate endpoints (e.g., ALP reductions) balanced with treatment-related incidence and severity of AEs. Given the potential variabilities in outcomes for different patient populations (i.e., early vs. moderate- advanced disease) and regimens (i.e., monotherapy vs. combination therapy) being studied in the clinical trials of OCA for PBC, our ratings are based on the evidence for these distinguishing factors. ©Institute for Clinical and Economic Review, 2016 ES 5 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Because only a small minority (11%) of patients included in clinical trials of OCA for PBC had moderate disease based on abnormal bilirubin, and no patients had advanced disease, we judge the evidence to be “insufficient” (“I”) for both patient populations. For OCA taken as monotherapy, we judge the evidence to be “promising but inconclusive.” Across two clinical trials, only 75 patients (17%) received OCA as monotherapy, and outcomes for these patients are only available in the grey literature. For OCA used as combination therapy with UDCA, we have moderate certainty of an incremental or better net health benefit. The one published RCT evaluating patients on a stable dose of UDCA plus OCA demonstrated statistically significant improvements in ALP levels on all liver biomarkers compared to UDCA alone, with the exception of bilirubin. Taking into consideration an increased incidence of treatment-related pruritus, we therefore assign the evidence a “B+” rating. Although not yet demonstrated in clinical trials, OCA has the potential to improve clinically-relevant outcomes, particularly for patients with no other treatment option (i.e., patients with inadequate response or intolerant to UDCA) and constitute an important advancement for treating this rare disease. Other Benefits or Disadvantages As previously discussed, prior to OCA approval, UDCA was currently the only FDA-approved treatment for PBC, on which an estimated 40% of patients have inadequate response to therapy; these patients are at an increased risk of liver transplant and death.6 Although not yet demonstrated in clinical trials, OCA has the potential to improve clinically-relevant outcomes, particularly for patients with no other treatment option (i.e., patients with inadequate response or intolerant to UDCA). Comparative Value A cost-effectiveness analysis was conducted by developing a microsimulation model that simulated the long-term outcomes of patients receiving OCA in addition to UDCA as observed in the Phase III POISE study and as a comparator a placebo (±UDCA) arm of the trial. Model parameters were estimated from published studies and calibrated when assumptions were required. The outcomes of the model included total costs, quality-adjusted life years (QALY), incremental cost-effectiveness ratios, transplant-free survival, and cumulative incidence of advanced disease stages. Outputs from this model were also used to inform a population-based analysis of the one- and five- year potential budgetary impact of OCA at a national level. Potential budgetary impact included estimates of costs saved from averted liver-related events (e.g., transplant, HCC) and was calculated assuming an uptake pattern for OCA if covered for the FDA-labeled indication without payer or provider efforts to restrain utilization. Based on long-term incremental cost-effectiveness ratios and a threshold for potential budget impact related to net health care cost growth at the national level, ©Institute for Clinical and Economic Review, 2016 ES 6 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents we also define a “value-based price benchmark” for OCA. Details on methods and inputs for all analyses can be found in the full report and appendices. Incremental Costs per Outcomes Achieved: Results In patients who had an inadequate response to UDCA, treatment with OCA would decrease the 15- year cumulative incidence of decompensated cirrhosis from 12.2% to 4.5%, hepatocellular carcinoma from 9.1% to 4.0%, liver transplant from 4.5% to 1.2%, and liver-related deaths from 16.2% to 5.7%, respectively. In addition, treatment with OCA increased 15-year transplant-free survival from 61% to 73%. The transplant-free survival in PBC patients as predicted by our model was similar to that predicted by the PBC Global study algorithm. Compared with the UDCA strategy, treating 10,000 patients using OCA plus UDCA could prevent 770 cases of decompensated cirrhosis, 510 cases of hepatocellular carcinoma, 330 liver transplants and 1,050 liver-related deaths. The average life years per patient treated with UDCA versus OCA plus UDCA were 19.97 and 22.23, respectively (increment = 2.26 years). The corresponding average discounted QALYs gained were 10.74 and 11.78, respectively (increment = 1.04 years). The average lifetime discounted cost per patient treated with UDCA was $142,300. Assuming that the price of OCA is $69,350/year, the average lifetime cost of a patient treated with OCA plus UDCA was $633,900 (an increment of $491,400). The incremental cost-effectiveness of OCA plus UDCA was approximately $473,400 per QALY gained (Table ES4). Table ES4. Cost-effectiveness of OCA when the Annual Cost of OCA is $69,350 per Year UDCA* OCA + UDCA Undiscounted Life Years 19.97 22.23 Discounted QALYs 10.74 11.78 Discounted Total Cost ($) 142,300 633,900 ICER ($/QALY) 473,400 *Results correspond to inadequate response to UDCA, as observed in POISE study Next we conducted a price threshold analysis to determine the price of OCA that would meet commonly cited thresholds for cost-effectiveness (Table ES5). We found that OCA would meet thresholds of $50,000, $100,000, and $150,000 per QALY gained if priced below $11,629, $18,445, and $25,261 per year, respectively. Table ES5. OCA Price Threshold Analysis Willingness to Pay ($/QALY) Annual Price of OCA $50,000 $11,629 $100,000 $18,445 $150,000 $25,261 ©Institute for Clinical and Economic Review, 2016 ES 7 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Sensitivity Analyses We conducted one-way sensitivity analyses to identify the parameters to which the model was most sensitive. We have plotted a tornado diagram showing the 15 most sensitive parameters (Figure ES1). We found that the incremental cost-effectiveness ratios were most sensitive to the cost of OCA. For all other parameters, the cost-effectiveness thresholds remained above $300,000/QALY when varied across a plausible range. Figure ES1. Tornado Diagram Showing 15 Most Sensitive Model Parameters Potential Budget Impact: Results We used the cost-effectiveness model to estimate the potential total budgetary impact of OCA for PBC patients, based on assumed patterns of uptake of OCA in clinical practice. This analysis included the entire candidate population for treatment, which was considered to be adults with PBC who have either inadequate response to UDCA or are unable to tolerate UDCA. Applying the PBC prevalence of 40.2 patients per 100,000 of population from the Rochester Epidemiology Project 18 to the projected 2016 US population provides an estimation of approximately 130,000 individuals with PBC in the US. The Kim at al. study also reported that 43.5% of PBC patients had received UDCA treatment.18 We assumed that 40% of the treated population with PBC would have inadequate response to UDCA therapy,19 and that another 3% would be unable to tolerate UDCA.2 Applying ©Institute for Clinical and Economic Review, 2016 ES 8 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents these percentages resulted in a candidate population size of approximately 24,350 individuals in the US. Using ICER’s methods for estimating budget impact, we assume a high uptake scenario in which 50% of the candidate population of 24,350 individuals with PBC would be treated within a five-year period. The estimated total potential budget impact is shown in Table ES6. Table ES6. Estimated Total Potential Budget Impact (BI) of OCA Analytic Horizon = 1 Year Analytic Horizon = 5 Years Eligible Number Annual BI Total BI Number Weighted BI Average BI Populati Treated per (millions) Treated per Patient* per year on Patient* (millions) OCA 24,350 2,440 $69,350 $169.2 12,200 $128,500 $312.9 *Weighted budget impact calculated by subtracting cost offsets from drug costs for one-year horizon. For five- year horizon, drug costs and cost offsets apportioned assuming 20% of patients in uptake target initiate therapy each year. Those initiating in year 1 receive full drug costs and cost offsets, those initiating in year 2 receive 80% of drug costs and cost offsets, etc. Total potential budgetary impact over five years is thus approximately $1.6 billion, with an average budget impact per year of approximately $312.9 million. This annualized potential budget impact is 35% of the budget impact threshold of $904 million for a new drug that should trigger policy action to manage affordability. Figure ES2 illustrates the budget impact scenarios with different pricing and patient uptake scenarios over a five-year period. The vertical axis shows the annualized budget impact, and the horizontal axis represents the percentage of eligible patients treated over a five-year period. The colored lines demonstrate how quickly the annual budget impact increases with increasing percentages of patients treated at four different prices: those at which the cost/QALY = $50,000, $100,000, and $150,000; and the list price used in this analysis (i.e., $69,350 annually for OCA). ©Institute for Clinical and Economic Review, 2016 ES 9 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure ES2. Combined Cost-effectiveness and Potential Budget Impact Graph for OCA in PBC Patients Value-based Price Benchmarks Considering that the potential budget impact of OCA according to an intermediate scenario of uptake is below the threshold of $904 million for a new drug that should trigger policy action to manage affordability, the draft value-based benchmark prices for OCA reflects only the price range that would achieve cost-effectiveness ratios between $100,000 and $150,000 per QALY gained. Our draft value-based benchmark prices for OCA are provided in Table ES7. Table ES7. Value-based price benchmarks for OCA in PBC patients WAC Price per Cost to Achieve Cost to Achieve Draft Value-Based Year $100K/QALY $150K/QALY Price Benchmark OCA $69,350 $18,445 $25,261 $18,445 to $25,261 Therefore, the draft ICER value-based price benchmark for OCA, with all the assumptions mentioned previously regarding five-year uptake patterns and net costs, is approximately $18,400 to $25,300 per year. This price represents a 64%-73% discount from the annual cost of OCA at list price ($69,350). ©Institute for Clinical and Economic Review, 2016 ES 10 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Comparative Value: Summary and Comment We estimated that, in patients who had an inadequate response to UDCA, treatment with OCA would decrease the 15-year cumulative incidence of decompensated cirrhosis from 12.2% to 4.5%, hepatocellular carcinoma from 9.1% to 4.0%, liver transplant from 4.5% to 1.2%, and liver-related deaths from 16.2% to 5.7%. Using the price of OCA as $69,350/year, the incremental cost- effectiveness of OCA plus UDCA was estimated to be approximately $473,400 per QALY, which exceeds commonly-cited thresholds of $100,000 to $150,000 per QALY. While the potential budget impact of OCA based on assumed candidate population size and uptake pattern does not exceed ICER’s $904 million annual threshold, a discount of 64%-73% from the annual cost of OCA at list price ($69,350) would nevertheless be required to achieve cost-effectiveness thresholds of $100,000-$150,000 per QALY. As with any model, ours has some limitations. A key limitation of this model is the lack of data to inform the natural history of PBC. There were limited data on the transition between advanced disease health states, such as compensated cirrhosis, decompensated cirrhosis, and hepatocellular carcinoma. For this reason, we calibrated the model to meet hard clinical endpoints, including the 15-year transplant-free survival for different ALP and bilirubin levels, and transplant-free survival for UDCA treatment. Another limitation is that our model draws on hepatitis C data to inform cost and quality-of-life parameters. Given the lack of economic and quality of life data for PBC health states, our model assumed that PBC-associated cirrhosis and hepatitis-associated cirrhosis involve similar treatment costs and PBC patients experience similar decrements in quality of life as in hepatitis C patients. Sensitivity analysis on these parameters showed that model outcomes were robust to uncertainty surrounding these cost and quality-of-life inputs. ©Institute for Clinical and Economic Review, 2016 ES 11 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 1. Background 1.1 Introduction Background Primary biliary cholangitis (PBC), which has until recently been referred to as primary biliary cirrhosis,1 is a rare, chronic, progressive autoimmune liver disease that mainly affects middle-aged women.3 The prevalence varies between different countries and regions;20 in the US, up to 130,000 individuals are estimated to have PBC.2 Diagnosis is increasingly occurring in up to 60% of asymptomatic patients,4 triggered through an investigation of increased levels of alkaline phosphatase (ALP) on routine blood tests. The presence of antimitochondrial antibodies (AMAs) confirms the diagnosis. AMAs are highly sensitive and specific, being present in 95% of patients and with specificity close to 100% when tested with recombinant antigens.5 A liver biopsy can be used to further substantiate the diagnosis if needed, but is typically not required. The disease process starts with autoimmune damage to small intrahepatic bile ducts, resulting in chronic cholestasis, portal inflammation, and fibrosis that can lead to cirrhosis and, ultimately, liver failure.5 On a biochemical level, the disease can be divided into early stage disease with an elevated alkaline phosphatase (ALP) and normal total bilirubin (TB), moderately advanced disease with either low albumin or high TB, and advanced disease with both low albumin and high TB.21 Figure 1 represents the natural history of PBC over time. ©Institute for Clinical and Economic Review, 2016 Page 1 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 1. Natural History of PBC5 Fatigue and pruritus are the most common symptoms of PBC, and both can be debilitating in some patients, especially fatigue.6 Among patients who are diagnosed once symptomatic, median survival has been estimated to range from 6-10 years without liver transplant.6 In contrast, in a cohort of 29 patients who were asymptomatic at diagnosis, none developed cirrhosis during a median follow-up of 17.8 years.7 An elevation in the levels of ALP and TB are predictive of average time to liver transplantation.22 Hepatocellular carcinoma (HCC) is an infrequent yet critical consequence of PBC, especially in male patients and those who do not respond adequately to standard treatment.8 Osteoporosis occurs in up to one-third of patients with PBC, but is usually asymptomatic and does not present an increased risk for fractures.4,23 Obeticholic acid (OCA) (Ocaliva™, Intercept Pharmaceuticals, Inc.) is a novel bile acid analogue that has shown positive effects on biochemical markers of liver function in two Phase II trials (NCT00550862 and NCT00570765) and in the one-year double-blind (DB) POISE Phase III trial that measured the impact of OCA on levels of ALP and TB (NCT02308111). OCA has received orphan drug designation in both the United States (US) and Europe for the treatment of PBC, and a priority review has been granted by the Food and Drug Administration (FDA). At a meeting of the FDA Gastrointestinal Drugs Advisory Committee Meeting on April, 7th, 2016, the committee unanimously agreed (17 to 0) that based on its effect on ALP, there is substantial evidence to support accelerated approval of OCA for the treatment of PBC in combination with ursodeoxycholic acid (UDCA) in adults with an inadequate response to UDCA, or as monotherapy in adults unable to tolerate UDCA,13 and was approved under the Prescription Drug User Fee Act (PDUFA) on May 27, ©Institute for Clinical and Economic Review, 2016 Page 2 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 2016 for the recommended indication. Three additional post marketing randomized, placebo- controlled clinical trials are required to confirm outcomes for patients who are cirrhotic or taking OCA as monotherapy, and to confirm that OCA-induced improvements in ALP and TB are associated with clinical outcomes such as liver transplantation and death.24 Scope of the Assessment This assessment evaluates the health and economic outcomes of OCA as a second-line treatment of PBC in combination with UDCA in adults with an inadequate response to UDCA or as monotherapy in adults unable to tolerate UDCA. The scope is described using the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, and Settings) framework.25 Analytic Framework The analytic framework for this assessment is depicted in Figure 2. Figure 2. Analytic Framework Treatment with Adverse Events Obeticholic • Pruritus Acid • Dyslipidemia Surrogate Outcomes Clinical and Patient Relevant PBC patients with inadequate response Biochemical Response: Outcomes Alkaline Phosphatase • Disease Progression (bleeding, to UDCA or Bilirubin cirrhosis, liver transplantation) intolerant to UDCA • Pruritus • Fatigue Populations The population of focus for the review included adults with PBC ages 18 years and older who have had an inadequate response to UDCA or who are unable to tolerate UDCA. Interventions The intervention of interest was OCA added to UDCA for patients with inadequate response to UDCA, or as a monotherapy for patients unable to tolerate UDCA. ©Institute for Clinical and Economic Review, 2016 Page 3 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Comparators Comparators included continued use of UDCA in patients able to tolerate such therapy and usual care for patients intolerant to UDCA. Outcomes This review examined key clinical outcomes related to PBC and its treatment, including surrogate outcomes in available clinical trials. Outcomes of interest included:  Biochemical response (e.g., ALP, bilirubin)  Other markers of liver function (e.g., alanine aminotransferase [ALT], aspartate aminotransferase [AST], gamma-glutamyl transpeptidase [GGT])  Measures of liver fibrosis  Bleeding from portal hypertension  Cirrhosis  Liver transplantation  Survival  Health-related quality of life  Adverse events (e.g., pruritus, fatigue, effects on cholesterol) Timing Evidence on intervention effectiveness and harms was derived from studies of any duration. Settings All relevant settings were considered, including inpatient, clinic, and office settings. ©Institute for Clinical and Economic Review, 2016 Page 4 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 2. The Topic in Context A relatively small number of patients have PBC, and it can take more than a decade for this chronic liver disease to progress from the early, typically asymptomatic stage to liver failure and need for liver transplantation. Both of these factors are major obstacles for the development of new therapeutic approaches. Regulatory authorities have developed programs to advance the evaluation and development of therapeutic products for diagnosis and treatment of rare diseases or conditions. According to the US Orphan Drug Act, a rare disease is defined as any disease which affects less than 200,000 persons in the US.26 With an estimate of 130,000 patients in the US, PBC is considered a rare disease and the manufacturer has received an orphan drug designation for OCA.27 The different disease stages from autoimmune damage to cholestasis, portal inflammation, and fibrosis could constitute targets for specific therapeutic interventions. For the early autoimmune stage, trials with most immunosuppressive agents have been inconclusive; the use of budesonide is currently being investigated in a large randomized controlled trial (RCT) (NCT00746486). UDCA mainly acts by decreasing cholestasis and protecting hepatocytes from the toxic effects of the bile acids. Specific antifibrotic agents could be of interest, but the cholestatic conditions seem to limit their potential.14 UDCA is a bile acid that is present in human bile at a concentration of approximately 3%. In patients with PBC, a daily dose of 13–15 mg/kg increases the percentage of UDCA in bile to about 40-50%.28 At this dose it has been shown to improve hepatic biochemistry, including ALP and TB, slow histological progression to fibrosis, delay and reverse portal hypertension (which develops prior to cirrhosis in PBC), and delay the requirement for liver transplantation.6,9 Lifetime treatment is recommended for all patients with PBC. Patients with early stage disease treated with UDCA have an overall survival similar to the general population. For patients with moderate to severe disease, UDCA treatment significantly improves average time to requirement for liver transplantation.10 However, treatment with UDCA does not improve fatigue and pruritus, and around 40% of patients with PBC do not achieve adequate improvement in biochemical measures of liver function.6 The introduction of UDCA has had other positive effects. PBC was the leading indication for liver transplantation in the US in the mid-1980s, but due to treatment with UDCA the number of patients with PBC requiring transplant has declined by 20% and it now ranks sixth.4 UDCA has minimal side effects.4 The most frequently reported adverse events (AEs) of the drug include loose stool (2–9%), headache, and mild weight gain, but these rarely lead to discontinuation.6 Patients who experience diarrhea should try to take the total daily dose in more frequent, smaller doses.29 Levels of intolerance to UDCA are not well-documented in the scientific literature. According to Intercept market research, around 3% of patients are intolerant to UDCA.2 ©Institute for Clinical and Economic Review, 2016 Page 5 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents However, the clinical practice guideline of American Association for the Study of Liver Diseases does not mention intolerance to UDCA.4 Improvement in ALP levels usually starts within a few weeks on UDCA and 90% of improvement occurs within six to nine months. For some patients, the treatment effect does not appear until up to five years on UDCA.4 Different criteria have been proposed to define biochemical response to UDCA, with six criteria being referred to most often. The cut-off points in time vary between six months for the Mayo Clinic criteria and two years for the Toronto criteria. The other four biochemical response criteria (Barcelona, Paris I and II, Rotterdam) use a cut-off point of 12 months to define inadequate response to UDCA.6 Overall, patients with an inadequate response to UDCA are at a higher risk of disease progression compared to those who show improvement.6 In one study of 192 patients treated with UDCA over 1.5 to 14 years, applying the Barcelona criteria of response as normalization or 40% decrease in ALP levels after one year, patients without biochemical response had a relative risk of death or liver transplantation of 5.51 (95% CI, 1.70 to 15.99) compared to those with a response.30 According to the different biochemical response criteria in UDCA-treated PBC patients, the reported nonresponse rates vary between 24% and 79%,11 but a rate of approximately 40% of patients with an inadequate response to treatment with UDCA seems to be considered an appropriate estimation.6 At a meeting of the FDA Gastrointestinal Drugs Advisory Committee Meeting on April 7, 2016, the committee unanimously agreed there is substantial evidence to support the effect of OCA on ALP as a surrogate outcome for the treatment of PBC in combination with UDCA in adults with an inadequate response to UDCA or as monotherapy in adults unable to tolerate UDCA. However, “the majority of the committee agreed that the data are limited on the use of OCA in moderately advanced stage PBC patients, and absent in advanced stage PBC patients, to support the use of OCA in moderately advanced and advanced stages of PBC.”13 For treatment of advanced stage patients, the FDA suggests a reduction to the dosing regimen proposed by the manufacturer.13 For the Phase III trial of OCA, also known as POISE, response was defined as ALP less than 1.67 times the upper limit of normal (ULN), at least a 15% decrease in ALP, and normalization of TB after 12 months of treatment with OCA in combination with UDCA (NCT02308111). The Global PBC Study Group is an international non-profit collaboration between medical centers that was established to search for reliable surrogate endpoints for PBC.15 Using a statistical analysis of this database, the FDA recommends the following definition of response after 12 months of treatment: ©Institute for Clinical and Economic Review, 2016 Page 6 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 3. FDA Proposed Stratified Cutoff for PBC Patients Indicated for OCA13 The FDA analysts estimate that the above cutoff criteria for defining treatment response improve the validity of a treatment effect on ALP as a surrogate outcome for clinical benefit in PBC. The validity of using ALP as a surrogate outcome for clinical benefit can be different across disease stages. As discussed in Section 4, 92% of the patient population in the Phase III trial for OCA were in the early stage of their disease. Patients with an inadequate response should receive a second-line treatment after monotherapy with UDCA, and there are several agents besides OCA that are currently being investigated for this purpose. Budesonide, fibrates, and OCA are currently considered as the most promising second-line treatments to be used in addition to UDCA.14 Both budesonide (NCT00746486) and bezafibrate (NCT01654731) are currently being studied in Phase III trials in Europe. Budesonide is recommended by European experts as a valid add-on to UDCA for patients with early stage disease and inadequate response to UDCA.14 According to clinical experts consulted, budesonide is rarely used as a second line treatment in the US. Bezafibrate, the fibrate most studied as a second-line treatment in PBC, is not licensed for sale in the US. In addition, on March 31, 2016 Genfit ©Institute for Clinical and Economic Review, 2016 Page 7 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents announced plans for a Phase II trial evaluating a new agent, elafibranor, as a second-line treatment for PBC patients with an inadequate response to UDCA, and is set to start before the end of 2016.31 ©Institute for Clinical and Economic Review, 2016 Page 8 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 3. Summary of Coverage Policies OCA Given the very recent FDA approval of obeticholic acid, very few coverage policies were available at the time of this report. Harvard Pilgrim Health Care offers a coverage policy which can be found in Appendix I. This section will be updated as coverage policies become available. UDCA UDCA is widely covered by most public and private payers in the New England region and nationally. Many insurers cover both generic and brand name formulations, with brand names falling into higher tiers and sometimes requiring prior authorization. ©Institute for Clinical and Economic Review, 2016 Page 9 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 4. Comparative Clinical Effectiveness 4.1 Overview To inform our analysis of the comparative clinical effectiveness of OCA as a second-line treatment for PBC, we abstracted evidence from available clinical studies, whether in published, unpublished, or abstract form. Regimens of interest included:  OCA taken once daily in combination with UDCA for patients with an inadequate response to UDCA; and  OCA taken once daily as monotherapy in patients unable to tolerate UDCA As described previously in the Background section, the comparators of interest included continued use of UDCA in patients able to tolerate such therapy and usual supportive care for patients intolerant to UDCA. Our review focused on surrogate markers of clinical benefit (i.e., biochemical response, other markers of liver function) as well as potential harms (drug-related AEs). The outcomes we addressed in detail are as follows:  Clinical Benefits o Changes in ALP o Treatment response (based on composite endpoints of ALP and other liver biomarkers) o Other measures of liver function (GGT, AST, ALT, and bilirubin)  Harms o Pruritus o Dyslipidemia o Other adverse events (e.g., hepatic decompensation) 4.2 Methods We included evidence from Phase II and III RCTs and supplemented our review of published studies with data from conference proceedings, regulatory documents, information submitted by the manufacturer, and other grey literature that met ICER standards for review (for more information, see http://icer-review.org/methodology/icers-methods/icer-value-assessment-framework/grey- literature-policy/). ©Institute for Clinical and Economic Review, 2016 Page 10 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Data Sources and Searches Procedures for the systematic literature review assessing the evidence on OCA with or without UDCA followed established best methods used in systematic review research.32 We conducted the review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.26 The PRISMA guidelines include a checklist of 27 items, further detail of which is available in Appendix Table A1. The timeframe for our search spanned the period from January 1996 to June 20, 2016 and focused on MEDLINE, EMBASE, and Cochrane-indexed articles. We limited each search to studies of human subjects and excluded articles indexed as guidelines, letters, editorials, narrative reviews, case reports, or news items. To supplement the above searches and ensure optimal and complete literature retrieval, we performed a manual check of the references of recent relevant reviews and meta-analyses. Other grey literature sources included data submissions from the manufacturer of OCA that were not otherwise publically available. Further details on the search algorithms, methods for study selection, data extraction, quality assessment, assessment for publication bias, and our approach to meta-analyses of the data are available in Appendix A. Assessment of Level of Certainty in Evidence We used the ICER Evidence Rating Matrix (see Figure 2) to evaluate the evidence for a variety of outcomes. The evidence rating reflects a joint judgment of two critical components: The magnitude of the difference between a therapeutic agent and its comparator in “net health benefit” – the balance between clinical benefits and risks and/or adverse effects AND The level of certainty in the best point estimate of net health benefit.33 ©Institute for Clinical and Economic Review, 2016 Page 11 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 4. ICER Evidence Rating Matrix ©Institute for Clinical and Economic Review, 2016 Page 12 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 4.3 Results Study Selection Our literature search identified 110 potentially relevant references, of which 16 met our inclusion criteria. Only one published study14 was available from our survey of the literature, with 15 additional conference abstracts and poster presentations; these references were related to three individual DB RCTs. Primary reasons for exclusion included: not a population of interest (i.e., not PBC patients), pharmacokinetic studies of OCA, abstracts/posters which contained duplicative data, or inappropriate study type (e.g., comment/opinion or review). We prioritized the reporting of information we identified as part of our literature search, and supplemented our presentation of the evidence with data available from regulatory documents. Details of the included references are described in Appendix B, and the three key trials are summarized in Table 1. Key Studies We identified only one published study reporting on outcomes from the Phase II RCT (Study 202) of OCA in combination with UDCA for patients with inadequate response to UDCA.14 Given the limited information available in the peer-reviewed literature, we included summaries of the other key RCTs (Study 202 and 301) below. All trials included patients with elevated ALP (≥1.5xULN) and excluded patients with a history or presence of hepatic decompensation and bilirubin greater than twice the ULN. Study 301 (POISE)34,35 The only Phase III trial we identified included 216 patients (mean age 56, 91% female, mean ALP 323.2 U/L, mean UDCA dose 15.4 mg/day) with ALP ≥1.67xULN or bilirubin <2xULN who were randomized to placebo, OCA 10 mg, or an OCA titration group who received 5 mg at the beginning of treatment and were titrated based on response after six months to a maximum dose of 10 mg. Most of the participants had been on UDCA for at least 12 months and were on a stable dose for at least three months prior to enrollment; 7% were UDCA intolerant and received OCA as monotherapy, and 92% had normal bilirubin at baseline. The primary endpoint was treatment response based on the proportion of patients achieving the Global PBC Study Group’s definition of response that has been shown to be correlated with clinical benefit (ALP <1.67xULN with a ≥15% ALP reduction and normal bilirubin). Safety measures were also assessed, and a long-term safety extension (LTSE) phase is ongoing for patients who completed the one-year DB phase (n=193). Study 20214 A Phase II trial evaluated three dosing regimens of OCA (10 mg, 25 mg, 50 mg) compared to placebo in 161 adults (mean age 55, 95% female, mean ALP 286.8 U/L, mean UDCA dose 15.9 mg/day) on a ©Institute for Clinical and Economic Review, 2016 Page 13 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents stable dose of UDCA (≥6 months) with ALP 1.5-10xULN over 85 days. Most of the participants (96%) who enrolled had mean bilirubin levels in the normal range across all treatment groups. The primary endpoint was reduction in ALP in the modified intent-to-treat (mITT) population (n=161) who received at least one OCA dose; secondary endpoints were based on changes in liver chemistry in the intent-to-treat (ITT) population (n=165), including AST, ALT, GGT, conjugated bilirubin, and blood serum levels. The authors also evaluated these effects for the proportion of patients meeting five published definitions of response criteria based on surrogate outcomes (e.g., ALP and TB levels above ULN), which have been shown to predict the risk of progression. An open-label safety extension trial was also conducted up to a year following enrollment, with 78 patients from the DB trial restarting treatment at a maximum 10 mg dose of OCA and titrating based on response. Study 20115,36 Another Phase II trial was conducted in 59 adult patients who received OCA as monotherapy in 10 mg and 50 mg doses compared to placebo over 12 weeks, followed by an off-treatment phase (day 86-99). This study was conducted subsequent to another Phase II trial (Study 202) evaluating the use of OCA taken concomitantly with UDCA to determine if the treatment effect of OCA was independent of UDCA. Included participants (mean age 55, 85% female, mean ALP 432.6 U/L) had ALP levels between 1.5-10xULN and had not been taking UDCA for at least three months prior to enrollment. Outcome measures included changes in biochemical response from baseline, mean reduction in ALP above the ULN, and harms. At the end of the DB phase of the trial, 28 patients (mean age 60, 84% female) enrolled in the LTSE and were followed up to 4.5 years; 43% of patients were taking a maximum 10 mg dose of OCA and eight eventually added UDCA. ©Institute for Clinical and Economic Review, 2016 Page 14 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Table 1. Key Trials Study 202 Study 301 RCT Study 201 Hirschfield et al. (POISE) N=59 (ITT) N=161 (ITT, N=165) N=216 (ITT, N=217) Mean age: 55 years Mean age: 55 years Mean age: 56 years Female: 85% Female: 95% Female: 91% Patient Characteristics Mean ALP: 432.6 U/L Mean ALP: 286.9 U/L Mean ALP: 323.2 U/L Mean UDCA dose: 15.9 Mean UDCA dose: 15.4 mg/day mg/day OCA 10 mg OCA 10 mg OCA 5-10 mg OCA 50 mg OCA 25 mg OCA 10 mg Interventions/Comparators Placebo OCA 50 mg Placebo Placebo DB: 3 months DB: 3 months DB: 1 year Duration LTSE: Up to 4.5 years LTSE: 1 year LTSE: 5 years (ongoing) OCA as monotherapy OCA+UDCA for patients OCA±UDCA for patients with for patients not on with inadequate inadequate response to UDCA UDCA for ≥3 months response to UDCA (on UDCA for ≥12 months and (stable dose ≥6 stable dose ≥3 months), or Inclusion criteria ALP 1.5-10xULN months) intolerant to UDCA (7%) ALP 1.5-10xULN ALP ≥1.67xULN or bilirubin >1xULN but <2xULN OCA 10 mg: -44.5 OCA 10 mg: -23.7 OCA 5-10 mg: -33.0 OCA 50 mg: -37.6 OCA 25 mg: -24.7 OCA 10 mg: -39.1 Placebo: +11.7 OCA 50 mg: -21.0 Placebo: -4.8 % Change in ALP Placebo: -3.1 All OCA groups from All OCA groups vs. placebo, baseline, p<0.0001 All OCA groups from p<0.0001 baseline, p<0.0001 ITT = intent-to-treat; DB = double-blind; LTSE = long-term safety extension; ULN = upper limit of normal; UDCA = ursodeoxycholic acid; OCA = obeticholic acid; ALP = alkaline phosphatase Quality of Individual Studies Using criteria from U.S. Preventive Services Task Force (USPSTF),37 we rated the one published RCT identified for this review to be of good quality.14 We judged this report to be of good quality because study arms were comparable at baseline, the authors used valid instruments to evaluate outcomes, and no differential attrition was observed. We did not assign a quality rating to the remaining 13 documents, which included results from the Phase II trial (Study 201) evaluating the use of OCA as monotherapy, and the Phase III trial (Study 301) evaluating the use of titrating the OCA dose (5-10 mg) in combination with UDCA, as well as pooled analyses from all three trials, ©Institute for Clinical and Economic Review, 2016 Page 15 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents because they were obtained from conference proceedings and regulatory packages rather than peer-reviewed publications. Clinical Benefits A detailed review of each outcome of interest is presented in the sections that follow. All trials were designed primarily to measure changes in ALP, including percentage or absolute mean reductions in ALP from baseline, as well as ALP normalization. Other liver biomarkers, including, AST, ALT, GGT and bilirubin, were also measured in all studies; these tests indicate liver function and are used by clinicians to follow the disease process. Treatment response was compared to different published risk models (e.g., ALP <1.67xULN and TB <ULN) that may reasonably predict important clinical outcomes for PBC patients (e.g., transplant-free survival and mortality). We also described the most frequently-reported harms as reported in the trials, with a particular focus on drug-related AEs (e.g., pruritus). Importantly, the outcomes reported from Study 202 were primarily based on the Hirschfield publication, while results from the other two key trials of OCA for PBC (Study 201 and 301) were only available in the grey literature. Changes in ALP The RCT publication for the Phase II trial evaluating OCA taken concomitantly with UDCA evaluated mean reduction in ALP levels from baseline to day 85 in the mITT group as the primary endpoint.14 Among those receiving OCA, there were ALP reductions of 24%, 25%, and 21% in the 10 mg, 25 mg, and 50 mg groups, respectively, while the placebo group saw a 3% reduction; these results were also statistically significant for the ITT and completer populations over three months, and these improvements were sustained during the nine month open-label extension period (OCA groups vs. baseline, p<0.0001). Overall, more OCA-treated patients saw at least a 10% reduction (87% vs. 14% for placebo) or at least a 20% reduction (69% vs. 8% for placebo) compared to placebo (both outcomes, p<0.0001). While only 7% of patients in the OCA-treated groups achieved ALP normalization, no patients in the placebo group did. A greater number of patients in the 25 mg group achieved ALP normalization (12%) than either the 50 mg (11%) or 10 mg (3%) groups, but these results were not evaluated statistically.14 Findings from Grey Literature Similarly, at the end of the one-year DB phase of the POISE study, there was a statistically significant mean percent reduction in ALP from baseline for the OCA-treated groups (-33.0% and -39.1% vs. - 4.8% for 5-10 mg, 10 mg, and placebo, respectively), coinciding with a statistically significant mean absolute ALP reduction (-106 U/L and -127 U/L vs. -6 U/L for 5-10 mg, 10 mg, and placebo, respectively) compared to placebo (both outcomes, p<0.0001).17 Notably, the OCA 10 mg group had a greater effect in the Phase II trial when taken as monotherapy (-44.5%) than for the Phase III trial evaluating UDCA combination therapy (-39.1%), which may in part be due to higher baseline ALP in ©Institute for Clinical and Economic Review, 2016 Page 16 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents monotherapy patients. After an additional six months of follow-up, ALP continued to be statistically significant from baseline in both the titration and 10 mg OCA groups, but placebo (i.e., UDCA alone) also demonstrated a statistically significant reduction (-111.3 U/L, -106.8 U/L, and -97.8, respectively; p<0.001 from baseline); OCA groups were not compared statistically to placebo, however.38 At the end of Study 201, patients in the OCA monotherapy groups experienced a statistically significant mean absolute reduction (10 mg, -233.5 U/L; 50 mg, -161.3 U/L) and mean percent reduction (10 mg, -44.5%; 50 mg, -37.6%) in ALP from baseline (all outcomes, p<0.0001), while ALP increased for placebo (+11.7 U/L, +0.4%). ALP also decreased from a mean 3.9xULN to 1.9xULN in the 10 mg group. Differences between groups were not evaluated statistically.36 In the open-label LTSE trial, OCA monotherapy patients (43% on ≤10 mg OCA) completing 4.5 years of follow-up (n=11) sustained a statistically significant mean absolute reduction in ALP from baseline (-182 U/L, p=0.0105). When including those patients who added UDCA (n=19), the mean absolute reduction in ALP was larger (-244 U/L, p=0.0034).15 Treatment Response In Study 202, treatment response was evaluated against several published criteria, which differed according to specific thresholds of biochemical response.14 Although the 25 mg OCA group saw statistically significant response rates across all algorithms (all p<0.05), the 10 mg and 50 mg groups did not demonstrate a statistically significant difference against placebo on two of the six criteria being evaluated; these included Paris I (ALP ≤3xULN, AST ≤2xULN, and TB ≤1 mg/dL), and a modified version of the Toronto I criteria that required normalization of bilirubin as well as ALP levels ≤1.67xULN.14 Findings from Grey Literature The Phase III POISE trial evaluated the proportion of patients achieving the primary endpoint (i.e., a mean ALP <1.67xULN with ≥15% reduction, and normal bilirubin), as well as the proportion of patients meeting varied definitions of biochemical response according to published predictive risk models.34 Statistically significantly more patients in the OCA-treated groups achieved the primary endpoint (46% and 47% vs. 10% in the 5-10 mg, 10 mg, and placebo groups, respectively; p<0.0001). Although Study 201 did not evaluate outcomes against a prescriptive definition of treatment response, a secondary analysis evaluated the proportion of patients achieving the POISE primary endpoint at 12 weeks across all three available trials.16 Table 2 shows that for the OCA 10 mg group in each trial, as well as the pooled results for all OCA groups, the proportion of patients in the OCA- treated cohorts achieving the POISE definition of treatment response was statistically significantly greater compared to placebo. ©Institute for Clinical and Economic Review, 2016 Page 17 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Table 2. Proportion of Trial Patients Achieving the POISE Primary Endpoint16 % of Patients Study 201, n=43 Study 202, n=76 Study 301, n=146 Pooled, n=306 Placebo 4 8 10 8 OCA 10 mg 40 42 47 45 (all OCA groups) p-value p=0.0026 p=0.0002 p<0.0001 p<0.0001 Other Measures of Liver Function In Study 202, improvements in other measures of liver function were statistically significant, including GGT (range: -48% to -63%) and ALT (-9% to -17%), across all OCA groups compared to the placebo group which experienced a negligible change at day 85 (all outcomes, p<0.05).14 However, the OCA 50 mg group did not produce a statistically significant reduction in AST, nor did the OCA 10 mg group produce a statistically significant reduction in conjugated bilirubin. The authors reported that three patients discontinued the DB trial per the mandated protocol as a result of elevated conjugated bilirubin (>2x average predose value),14 one in the 10 mg and two in the 50 mg groups.39 Findings from Grey Literature Results from the Phase III POISE trial demonstrated statistically significant improvements for the 5- 10 mg and 10 mg groups on ALT (-35.5% and -41.7% vs. -4.7% for placebo, respectively) and GGT (- 50.3% and -63.7% vs. 0.8% for placebo, respectively) compared to placebo, with greater reductions in the 10 mg group (all outcomes, p<0.0001).17 TB was also statistically significantly better than placebo for the OCA groups (1.2% and -0.2% vs. 19.5% for the 5-10 mg, 10 mg, and placebo groups, respectively; p<0.005).17 Finally, absolute changes in AST were statistically significantly improved from baseline for both OCA groups, but were not compared statistically to placebo (-13 U/L and -15 U/L vs. 1 U/L for 5-10 mg, 10 mg, and placebo, respectively, p<0.0001).39 During the LTSE, all groups, including placebo, demonstrated statistically significant reductions on GGT, ALT, and AST, though these results were numerically higher in the OCA-treated cohorts (p<0.05). While bilirubin increased in the placebo groups over 18 months of follow-up, both OCA groups demonstrated a reduction; differences from baseline were not statistically significant, however.38 For Study 201, other measures of liver function in the 10 mg OCA and 50 mg OCA groups were statistically significantly improved from baseline at the end of the DB phase, including ALT (-38% and -35%, respectively) and GGT (-73% and -65%, respectively), with overall greater reductions for the 10 mg group (all outcomes, p<0.0001).36 Changes from baseline continued to be statistically significantly improved in the LTSE phase (mean 4.5 years of follow-up) for the combined OCA- treated groups (ALT: -54%, GGT: -70%, AST: -29%, p<0.01). However, mean change in TB was not statistically significant at 4.5 years, and differences between groups were not compared statistically at any point.15 ©Institute for Clinical and Economic Review, 2016 Page 18 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Harms Treatment-related adverse effects were generally mild to moderate in nature, and primarily related to pruritus. The main reason for discontinuation in the trials was severe pruritus, with increased frequency in patients receiving higher doses. Changes in lipid levels, mostly due to reductions in high-density lipoprotein (HDL), were also common in all OCA-treated patients. Although treatment- related reductions are concerning given the protective nature of HDL against cardiovascular morbidity, it is unclear if these changes represent clinical significance. Other serious adverse events (SAEs) occurred with much less frequency and included hepatic decompensation, gastrointestinal disorders, and hyperbilirubinemia; questions remain as to whether these events were correlated with OCA or manifested independently as a result of progressing disease. With the exception of some pruritus outcomes in the Hirschfield trial, these events were not evaluated statistically, likely due to the overall limited occurrence. Pruritus Although pruritus is a symptom of PBC, the available evidence suggests there is increased frequency and severity of pruritus due to OCA treatment, particularly at higher doses. In the Hirschfield RCT, the incidence of pruritus was 85% (p<0.0003) and 80% (p<0.006) in the 25 mg and 50 mg OCA groups, respectively, compared to 50% in the placebo group. In the OCA 10 mg group, the incidence of pruritus was numerically lower than in the placebo group (47%), but the difference was not statistically significant.14 Severe pruritus, which was the primary reason for trial discontinuation, occurred in 16%, 24%, and 37% of the 10 mg, 25 mg, and 50 mg cohorts, respectively, compared to no patients in the placebo group. Overall, 10% of the OCA-treated population discontinued the DB phase due to pruritus. Although the authors reported that incidence of severe pruritus in the open-label LTSE phase was lower, 10 patients (13%) discontinued the trial due to severe pruritus.14 No additional details on the distribution of pruritus severity or trial discontinuation across treatment groups from the open- label LTSE phase were available, however. Findings from Grey Literature The POISE trial demonstrated similar dosing effects of OCA related to pruritus. In the DB phase, pruritus occurred more frequently in the OCA-treated groups relative to placebo, with 38%, 56%, and 68% in the placebo, 5-10 mg, and 10 mg groups, respectively.34 However, pruritus was generally well-managed using bile acid sequestrants and/or antihistamines, with 13% in the titration group and 20% in the 10 mg group requiring a change in their dosing schedule.40 Pruritus was less severe in the titration group, with only 1% (1 patient) discontinuing due to pruritus compared to 10% (7 patients) in the 10 mg group.34,35 Overall, less than six percent of patients discontinued due to pruritus.34 With up to two years of follow-up in the LTSE phase, new incidence of pruritus was lower ©Institute for Clinical and Economic Review, 2016 Page 19 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents than in the DB trial, occurring at the rate of 15% in the 5-10 mg group and 21% in the 10 mg group, with a discontinuation rate of less than 1%.34,35 In Study 201, incidence of pruritus occurred at a rate of 30%, 70%, and 94% in the placebo, 10 mg, and 50 mg groups, respectively.36 These cases were generally mild to moderate; three patients (15%) in the 10 mg arm and six patients (38%) in the 50 mg arm discontinued due to pruritus in the DB trial.41 After 4.5 years of follow-up, three (10%) of the 28 remaining patients in the trial discontinued due to pruritus.15 Dyslipidemia Hirschfield et al. reported that changes in lipid levels were dose-related and mainly due to a reduction in HDL levels. While total cholesterol decreased 3-13% across all OCA-treated groups, with incrementally greater reductions correlated with higher doses, all OCA groups experienced a drop in HDL while the placebo group remained relatively stable over time.14 Mean reductions in HDL levels ranged from a 0.47 mmol/L in the 50 mg group, to 0.25 mmol/L in the 10 mg group. 41 During the off-treatment phase of the DB study (day 86-99), HDL levels increased for all OCA- treated groups, suggesting an OCA-induced effect. With the exception of the 50 mg treatment arm, mean low-density lipoprotein (LDL) levels were similar for all groups at the end of the study,14 and one patient in the 10 mg group had a clinically significant increase in LDL.41 Findings from Grey Literature HDL effects were also observed in the POISE trial, with reductions of 16% in the 5-10 mg group, 26% in the 10 mg group, and 3% in the placebo group.17,41 The mean percent increases in LDL were 4%, 1%, and 2% in the 5-10 mg, 10 mg, and placebo arms, respectively.41 Additionally, of those patients that had normal HDL at baseline, seven in the 10 mg group compared to four in the placebo group had below normal HDL by the end of the study.41 After two years of treatment, the HDL decrease that occurred during the DB trial was unchanged, while LDL returned to baseline levels in all groups.35 Although all groups in Study 201 had modest changes in serum lipids, HDL levels decreased by 0.16 mmol/L, 0.09 mmol/L, and 0.14 mmol/L in the 10 mg, 50 mg and placebo groups, respectively. The reduction was noticeably less in the 50 mg group than in the placebo groups, but the investigators noted that this was likely due to the high dropout rate due to pruritus. There were also corresponding increases in LDL levels in the 10 mg (0.10 mmol/L) and 50 mg (0.23 mmol/L) arms, while there was a decrease in the placebo arm (0.08 mmol/L).41 The principal investigators did not consider these changes to be clinically meaningful.15 ©Institute for Clinical and Economic Review, 2016 Page 20 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Other Adverse Events Other commonly-reported adverse effects of treatment include skin and subcutaneous tissue disorders (e.g., rash and gastrointestinal issues), infections (e.g., headache), and respiratory disorders. Table 3 lists the three most frequently reported AEs (other than pruritus) across the treatment groups in the DB study.14 Table 3. Top 3 Most Frequently Reported AEs in Hirschfield et al.14 Placebo OCA, 10 mg OCA, 25 mg OCA, 50 mg Incidence of AEs n, % n, % n, % n, % Fatigue 5, 13 7, 18 3, 6 5, 12 Headache 4, 11 3, 8 5, 10 11, 27 Gastrointestinal disorders 10, 26 17, 45 17, 35 17, 41 Overall, 23 patients (14%) discontinued the trial after experiencing an AE of which more than half were in the OCA 50 mg arm.14 Serious AEs occurred in seven patients (4%) (of which five were in the OCA 50 mg cohort), including dyspnea, salivary gland neoplasm, angina pectoris and angioedema. Two hepatic events occurred in two subjects, requiring discontinuation: one had a GI hemorrhage (gastrointestinal bleed related to esophageal varices that were present prior to treatment), and the other on the 50 mg dose of OCA had advanced PBC with cirrhosis and developed jaundice.14,41 In the open-label extension trial, fatigue, insomnia, and upper respiratory infection were most common and reported with the same frequency across the OCA-treated groups (13%), but only 3 patients (4%) discontinued the trial due to an AE.14 Findings from Grey Literature The Phase III trial of OCA (POISE) reported a higher proportion of patients experiencing treatment- related AEs across OCA-treated groups during the first 12 months (39%, 51%, and 28% for 5-10 mg, 10 mg, and placebo, respectively).40 Overall, 22 patients (10%) experienced a serious AE;34 15 events occurred in 11 patients in the titration group, two of which were hepatic-related; five events occurred in eight patients in the 10 mg group, one of which was hepatic-related.41 Eight patients experienced cardiovascular-related AEs, most of which were palpitations and occurred with more frequency in the OCA-treated groups; one patient died of cardiac failure but this was likely unrelated to treatment.41 Bone fractures were also more common in the OCA-treated groups (3 vs. 1 in the placebo group); a subgroup analysis was performed to determine whether this was related to treatment, the details of which are presented in the section below.41 In the DB phase of the trial for OCA monotherapy patients (Study 201), placebo patients experienced the same rate of adverse events as the OCA-treated groups; only one patient in the placebo group experienced a SAE, and no deaths occurred.36,41 A total of eight patients experienced ©Institute for Clinical and Economic Review, 2016 Page 21 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 20 events in the long-term extension phase,15 of which four OCA-treated subjects that had serious hepatic-related AEs, including jaundice, liver decompensation, esophageal variceal hemorrhage, and hyperbilirubinemia.41 Subgroup Analyses We identified six conference abstracts and poster presentations evaluating subgroup analyses for the available trials, three of which pooled data from all three RCTs. Two of three pooled analyses stratified patients according to disease severity with one evaluating those patients within ALP ranges above the ULN, and another for those patients with abnormal bilirubin levels;42,43 a third analysis evaluated subpopulations based on age at diagnosis of PBC, age, and gender.44 Two additional analyses evaluated subgroups of patients in the POISE trial based on those remaining on 5 mg versus those who titrated to 10 mg in the titration treatment arm, and the proportion of patients who had dual-emission x-ray absorptiometry (DEXA) scans to determine the effect of OCA on bone mineral density (BMD).45,46 The final subgroup analysis used the data from POISE to determine if pre-treatment characteristics were associated with pruritus severity.47 In a pooled analysis of the three available RCTs, those patients with abnormal bilirubin at baseline (n=46, mean age 53, 83% female) achieved statistically significant reductions in ALP for the OCA 10 mg and all OCA-treated groups relative to placebo at three months of follow-up (-212 U/L and -189 U/L vs. -31 U/L, in 10 mg, all-OCA, and placebo cohorts, respectively; p<0.005); these changes were also statistically significant at month 12 (-110 U/L and -120 U/L vs. -20 U/L, in 10 mg, all-OCA, and placebo cohorts, respectively; p<0.05).42 Although mean reduction in bilirubin was not statistically significant at three months, for patients that had follow-up at month 12 (n=21) there was a statistically significantly greater effect of OCA compared to placebo on TB levels (-8.9 µmol/L vs. - 0.7 µmol/L, p<0.05). As with patients with normal bilirubin, there was a dose-related increase for the incidence of pruritus (86% and 73% vs. 30% in 10 mg, all OCA, and placebo cohorts, respectively). Hepatic events were lower in the 10 mg group (7%) compared to the all-OCA (14%) and placebo (10%) cohorts, but the number of patients in these groups were too small to determine any association with treatment.42 In a similar subgroup analysis that patients stratified by ranges of ALP above the ULN, a linear regression model adjusting for UDCA dose at baseline indicated that patients receiving a 10 mg dose across the three trials (n=114, mean age 56, 88% female) who were within the ALP ranges of 1-2xULN, >2-3xULN, >3-4x ULN, and >4xULN had a statistically significant (p<0.0001) ALP reduction of 23%, 29%, 44%, and 45%, respectively, from baseline which corresponded to a statistically significant decrease of 3.5% for every incremental level above ULN (p=0.0046).43 Finally, there was a statistically significant incremental effect on TB levels, with a greater reduction for higher ALP ranges (-4%, -5%, -6%, and -17%; p<0.05). Harms associated with these ALP ranges were not reported.43 ©Institute for Clinical and Economic Review, 2016 Page 22 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents The third subgroup analysis evaluating the potential differential effect of OCA according to age at PBC diagnosis (<50 or ≥50 years), age (<65 or ≥65 years), and gender (male or female) on outcomes found that all groups experienced similar reductions in ALP and a comparable proportion of patients meeting the POISE primary endpoint based on three-month results pooled across the three trials for patients taking a maximum 10 mg dose of OCA; incidence of pruritus was also similar.44 For those patients in the titration arm of the POISE trial, a fourth subgroup analysis evaluated whether efficacy and tolerability were comparable for those who remained at 5 mg (n=33) relative to those who titrated to 10 mg (n=36) after six months if the composite endpoint was not met.34 Patients in both groups had similar response rates (with an additional 13 responders who titrated to 10 mg), reductions in ALP, and incidence of pruritus; however, differences between the two subgroups were not evaluated statistically.34 The fifth analysis assessed those patients in the POISE trial (n=122, 85% female, 52% postmenopausal) who had DEXA scans at baseline and study end to determine the effect of OCA on BMD.46 While the 10 mg and placebo groups had statistically significant decreases in femoral T- scores (p=0.01 and p=0.03, respectively), OCA-treated participants had statistically significantly smaller decreases compared to placebo (-0.06 g/cm² and -0.07 g/cm² vs. -0.33 g/cm² for 5-10 mg, 10 mg, and placebo groups, respectively, p<0.05) suggesting a possible beneficial effect of OCA on BMD.46 The final subgroup analysis evaluated if any baseline characteristics were associated with a higher probability of treatment-induced pruritus.47 Based on the 5D Itch Score and VAS scales, higher pretreatment ALP, GGT, and bilirubin were statistically significantly associated with pruritus severity for OCA groups relative to placebo using an ANCOVA model (all outcomes, p<0.03).47 Controversies and Uncertainties Several limitations in the body of evidence reduce our ability to make judgments regarding the comparative net health benefits of OCA. First, data on clinically-relevant outcomes, including transplant-free survival and mortality, are not yet available. Surrogate endpoints, such as reduction in ALP, have not been previously adopted by the FDA as acceptable criteria for regulatory approval of new PBC regimens. As discussed at the FDA Advisory Committee meeting, there remains some uncertainty surrounding the appropriate definition of treatment response (and specifically the cut- off points used in POISE) for PBC patients, particularly for those patients with elevated ALP and normal bilirubin, as well as the clinical significance of treatment-related reductions in HDL. Although the committee voted unanimously to recommend approval of the drug, FDA authorization sets a precedent for the use of ALP reduction alone as a measure of improvement in PBC patients before a correlation between changes in liver biomarkers and clinically-relevant outcomes across the disease spectrum has been corroborated; only long-term studies will validate this association. ©Institute for Clinical and Economic Review, 2016 Page 23 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Our certainty in the efficacy of OCA is also hampered by the lack of peer-reviewed data of the dose regimens selected for marketing approval. For example, only one Phase III trial (POISE) evaluated a starting dose of 5 mg titrating to a maximum of 10 mg based on response, which is the regimen suggested by the manufacturer to reduce discontinuation due to pruritus. Although interim data of this trial is available in the grey literature, such information not yet been subject to the adjudication process employed for journal publications. Yet another area of uncertainty is the use of OCA as monotherapy. Although pooled data demonstrated similar efficacy to patients on combination therapy with regards to reduction in ALP (38% vs. 41% for OCA+UDCA),41 data on the use of OCA without concomitant use of UDCA is primarily limited to results from two trials in conference abstracts and regulatory documents. Additionally, baseline values of ALP were higher for the Phase II monotherapy cohort, and only 16 patients from the POISE trial were considered intolerant to UDCA and taking OCA alone. Some experts suggest that patients taking UDCA with inadequate response may see additional benefits of the drug for up to five years of ongoing treatment,4 which may explain the continued improvement of ALP levels on UDCA alone in the POISE study after 18 months; crossover to OCA may also have contributed to the increase in the placebo cohort.38 Given that no head-to-head trial has been conducted, the true effect of OCA relative to UDCA is uncertain. However, because there were no additional safety concerns in these patients even after 4.5 years of follow-up, it is unlikely that OCA taken as monotherapy would represent a unique issue for patients intolerant to UDCA, particularly since there is no other available treatment option for these patients in the US.15 Perhaps the greatest amount of uncertainty in comparative net health benefit lies in the lack of data available for patients in later stages of their disease. Across the three trials of PBC, only 46 patients (11%) had abnormal bilirubin at baseline. Although a subgroup analysis of these patients demonstrated a statistically significant reduction in ALP and other liver biomarkers for OCA-treated patients compared to placebo at one year, less than half of these patients completed the studies (n=21).42 In addition, because those patients with higher bilirubin at baseline tended to experience more SAEs, there may be a safety concern for moderately advanced patients that would warrant a reduced dosing schedule.41 Therefore, further study should elucidate OCA’s performance at different points during the disease course. Summary As previously mentioned, no trial has yet assessed clinically-meaningful outcomes associated with OCA as a second-line treatment. Therefore, we assigned ICER evidence ratings based on the reporting of improvements on surrogate endpoints (e.g., ALP reductions) balanced with treatment- related incidence and severity of AEs. Given the potential variabilities in outcomes for different patient populations (i.e., early vs. moderate-advanced disease) and regimens (i.e., monotherapy vs. combination therapy) being studied in the clinical trials of OCA for PBC, our ratings are based on the ©Institute for Clinical and Economic Review, 2016 Page 24 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents evidence for these distinguishing factors. Because only a small minority (11%) of patients included in clinical trials of OCA for PBC had moderate disease based on abnormal bilirubin, and no patients had advanced disease, we judge the evidence to be “insufficient” (“I”) for both patient populations. Although a pooled subgroup analysis demonstrated statistically significant improvements in biochemical response for those patients with abnormal bilirubin at baseline, questions remain regarding the safety profile of OCA in patients who are in the later stages of their disease course. For patients with early disease, we have moderate certainty that OCA with UDCA provides an incremental or better net health benefit (“B+”). All three RCTs of OCA for PBC, of which a large majority were early stage PBC patients, demonstrated statistically significant ALP reductions from baseline and compared to placebo for all doses under investigation. Although incidence and severity of pruritus increased while taking OCA, these outcomes appeared to be mostly associated with higher doses, and were reduced in patients on a titration regimen (5-10 mg). Table 4. ICER Evidence Ratings by Regimen Regimen Comparators Evidence Rating OCA as monotherapy Placebo P/I OCA plus UDCA Placebo plus UDCA B+ For OCA taken as monotherapy, we judge the evidence to be “promising but inconclusive.” Across two clinical trials, only 75 patients (17%) received OCA as monotherapy, and outcomes for these patients are only available in the grey literature. Nevertheless, results from conference abstracts and regulatory documents suggest a statistically significant improvement from baseline and compared to placebo on most liver biomarkers. In addition, in a pooled analysis of OCA 10 mg monotherapy patients from the two RCTs, reductions in ALP were similar to those patients who were also taking UDCA, and no additional safety concerns arose over 4.5 years of follow-up. Finally, we have moderate certainty of an incremental or better net health benefit for OCA used as combination therapy with UDCA. The one published RCT evaluating patients on a stable dose of UDCA plus OCA demonstrated statistically significant improvements on all liver biomarkers compared to UDCA alone, with the exception of bilirubin. Taking into consideration an increased incidence of treatment-related pruritus, we assign the evidence a “B+” rating. ©Institute for Clinical and Economic Review, 2016 Page 25 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 5. Other Benefits or Disadvantages Our reviews seek to provide information on other benefits or disadvantages offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Examples include but are not limited to: 1. Methods of administration that improve or diminish patient acceptability and adherence 2. A public health benefit, e.g., reducing new infections 3. Treatment outcomes that reduce disparities across various patient groups 4. More rapid return to work or other positive effects on productivity (if not considered a benefit as part of comparative clinical effectiveness) 5. New mechanisms of action for treatments of clinical conditions for which the response to currently available treatments vary significantly among patients for unknown reasons (substantial heterogeneity of treatment effect) As previously discussed, prior to OCA approval, UDCA was currently the only FDA-approved treatment for PBC, on which an estimated 40% of patients have inadequate response to therapy; these patients are at an increased risk of liver transplant and death.6 Although not yet demonstrated in clinical trials, OCA has the potential to improve clinically-relevant outcomes, particularly for patients with no other treatment option (i.e., patients with inadequate response or intolerant to UDCA). ©Institute for Clinical and Economic Review, 2016 Page 26 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 6. Comparative Value 6.1 Overview The primary aim of this analysis was to estimate the cost-effectiveness of OCA treatment for patients with PBC who have an inadequate response to conventional (i.e., UDCA) treatment. We conducted a cost-effectiveness analysis by developing a microsimulation model that simulated the long-term outcomes of patients receiving OCA in addition to UDCA as observed in the Phase III POISE study; as a comparator, we also simulated the placebo (±UDCA) arm of the trial. Model parameters were estimated from published studies and calibrated when assumptions were required. The outcomes of the model included total costs, quality-adjusted life years (QALY), incremental cost-effectiveness ratios, transplant-free survival, and cumulative incidence of advanced disease stages. 6.2 Prior Published Evidence on Costs and Cost-Effectiveness of Obeticholic Acid We did not identify any published articles or public presentations pertaining to the cost and/or cost- effectiveness of OCA. To the best of our knowledge, this report is the first publicly available analysis that estimates the cost-effectiveness and long-term impact of OCA for the treatment of patients with PBC. 6.3 Incremental Costs per Outcome Achieved Cost-Effectiveness Model: Methods Model Structure We developed a microsimulation model (individual-level state transition model) using Java, a general purpose programming language. Figure 5 describes the model structure. For each treatment regimen, a hypothetical patient population begins the model in the PBC health state (defined as PBC stage 1–3), where they remain until they experience either: (a) disease progression or (b) death from all-cause or liver-related mortality. Patients who transition from the PBC stage 1– 3 to compensated cirrhosis state remain there until they transition to another advanced liver disease health state or die from liver-related mortality or from other causes. Patient survival, quality-adjusted survival, and health care costs were estimated for each model cycle and then summarized over the entire time horizon for each treatment option. We used an annual cycle length in the model. ©Institute for Clinical and Economic Review, 2016 Page 27 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 5. Model Structure for the Natural History of PBC Target Population The population of focus for the review was adults with PBC whose disease has not adequately responded to UDCA treatment (i.e., is refractory). Demographic characteristics for the modeled cohort were selected to match those of patients in the POISE study. The mean age of patients was 55.8, and 91% were female. At the beginning of the model simulation, 90% of the cohort started in stages 1–3 and 10% started with compensated cirrhosis (Table 5). ©Institute for Clinical and Economic Review, 2016 Page 28 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Table 5. Model Cohort Characteristics Value Primary Source Mean age 55.8 POISE study PBC stage distribution POISE study Stage 1-3 90% Stage 4 (cirrhosis) 10% Sex: Female / male 91% / 9% POISE study Treatment Strategies The interventions of interest were OCA plus UDCA compared to UDCA monotherapy. The 12-month Phase III trial of OCA included a 10 mg arm and a titration arm; for patients in the titration arm, the dose was initiated at 5 mg once daily and increased to 10 mg once daily if they had not yet achieved the primary composite endpoint and were tolerating treatment. As the titration arm included fewer discontinuations due to adverse events, we chose to use the titration arm to model the “OCA plus UDCA” intervention in our analysis, which is consistent with our conversations with PBC experts. Key Model Choices and Assumptions  Health states in the model include PBC, compensated cirrhosis (CC), decompensated cirrhosis (DC), hepatocellular carcinoma (HCC), liver transplant (LT), liver-related death, and death from other causes.  The model primarily utilized data from one Phase III trial of OCA in patients with PBC. We derived estimates of the proportion of patients in each of the following categories at baseline and at 12 months: normal bilirubin, ALP ≤ 1.67 x ULN; normal bilirubin, ALP > 1.67xULN; abnormal bilirubin, ALP ≤ 1.67xULN; and abnormal bilirubin, ALP > 1.67xULN. These criteria have been clinically validated as predictors of disease progress.  We also derived estimates of the proportion of patients experiencing pruritus.  Survival was weighted by health state utilities to estimate QALYs  The model also included a disutility for pruritus; however, we assumed pruritus would resolve in all patients after the first year of treatment.  We assumed that treatment effects persisted after 12 months, i.e., if a patient moves from the “normal bilirubin, ALP > 1.67xULN” group to the “normal bilirubin, ALP ≤ 1.67xULN” group, then the patient will continue to follow the disease progress for an individual in the “normal bilirubin, ALP ≤ 1.67xULN” group for the rest of his/her life.  Patients who discontinued OCA, either because of pruritus or lack of adherence, entered the comparator arm of the model (i.e., the UDCA arm).  All future quality-adjusted life years and health care costs were discounted at 3% per year. ©Institute for Clinical and Economic Review, 2016 Page 29 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Clinical Inputs Levels of ALP and bilirubin will be used as reported in the clinical trial. Table 6 below summarizes the proportion of patients in different categories after 12 months of treatment. Levels of ALP and bilirubin are correlated with disease course in patients with PBC. Table 6. Efficacy of OCA versus placebo (UDCA) after 12 months Primary Endpoint OCA + UDCA UDCA Range Normal Bilirubin & ALP ≤ 1.67xULN 45.7% 9.6% ±25% Normal Bilirubin & ALP > 1.67xULN 50.0% 69.9% ±25% Abnormal Bilirubin & ALP ≤ 1.67xULN 2.1% 2.5% ±25% Abnormal Bilirubin & ALP > 1.67xULN 2.2% 18.0% ±25% Source: POISE study (Intercept) Adverse Events The model includes pruritus of varying severity, classified as mild, moderate, or severe. Table 7 provides the percentage of patients who experienced mild, moderate and severe pruritus in the OCA and placebo (±UDCA) arms of the trial. The overall probability of pruritus was derived from the POISE trial, to which we applied the distribution of pruritus severity (i.e., whether pruritus was mild, moderate, or severe) for doses of 10 mg or less using data from the Phase II trial by Hirschfield et al. Table 7. Adverse Event Inputs Value (%) Reference Range Pruritus with UDCA40 Mild 22% Table 1 (placebo) ±25% Moderate 10% Table 1 (placebo) ±25% Severe 7% Table 1 (placebo) Pruritus with OCA plus UDCA40 Mild 16% Table 1 (≤10mg dose) ±25% Moderate 21% Table 1 (≤10mg dose) ±25% Severe 19% Table 1 (≤10mg dose) ±25% Costs Health state costs associated with advanced stages of disease were based on reported costs in hepatitis C patients.48 The cost of care for early stage PBC was assumed to be similar to that of hepatitis C patients with moderate fibrosis. Table 8 summarizes the costs associated with each health state. All costs were converted to a 2015 baseline using the medical care component of the Consumer Price Index. ©Institute for Clinical and Economic Review, 2016 Page 30 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents For patients who have pruritus, we assumed there will be additional costs associated with two primary care visits; these costs are based on the fees associated with HCPCS code 99213 (equivalent to a 15-minute established patient office visit) in the physician fee schedule.49 We also applied the cost of one year of hydroxyzine treatment, based on the Red Book value for a 25 mg dose three times per day for one year.50 The annual cost of UDCA was estimated using an average UDCA dosage of 16 mg/kg/day, as well as an average BMI of 26 kg/m2 and height of 164.3 cm, as reported in the Phase II data currently available to us.14 Accordingly, the average daily dose for patients in the model was 1,123 mg. We assumed that patients in the trial were taking doses of either 1,000 mg (two 500 mg capsules) or 1,250 mg (two 500 mg capsules and one 250 mg capsule). Through this line of reasoning, we determined that the cost of UDCA treatment was $8.66/day or $3,163/year (assuming 365.25 days/year), using Red Book wholesale acquisition costs. Table 8. Cost Inputs Associated with Health States and Management of Adverse Events Range Parameter Values Early stage PBC $737 ±25% Compensated cirrhosis $5,752 ±25% Decompensated cirrhosis $40,141 ±25% Hepatocellular carcinoma $88,383 ±25% Liver transplant-1st year $179,080 $134,310-$739,100 Liver transplant-subsequent year $44,074 ±25% Cost of OCA $69350 $15,000-$100,00 Cost of UDCA $3,163 $2,372-$20,000 Cost of Pruritus (doctor’s office visit) $103 ±25% Cost of Pruritus (ongoing hydroxyzine treatment) $712 ±25% ©Institute for Clinical and Economic Review, 2016 Page 31 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Utilities We assigned health-related quality-of-life (QoL) utilities to each individual in our model, with 0 denoting death and 1 denoting perfect health. Health utilities were adjusted by age and sex to accurately reflect comorbidities occurring with aging as well as differences by sex. Health state utilities from publicly available literature (Table 9), with consistent health state utility values across treatments, were used in the model. Because PBC-specific utilities by different stages of disease were not available, we used utilities of health states for patients with hepatitis C virus infection. Specifically, we used health-state specific utility weights from a previously published study using the EuroQol-5D instrument,51,52 and adjusted these weights to the U.S. population norm (Table 10).53 We also applied a disutility for patients who experience pruritus; to determine the overall utility for a patient with pruritus, we took the product of each health state utility and the pruritus utility. Table 9. Utilities for Health States and Adverse Events Health State Base Case Range Health States51 PBC (UDCA+OCA) 0.93 0.84-0.99 PBC (UDCA) 0.93 0.84-0.93 Compensated cirrhosis 0.90 0.81-0.99 Decompensated cirrhosis 0.80 0.57-0.99 Hepatocellular carcinoma 0.79 0.54-0.99 Transplant-first year 0.84 0.77-0.93 Transplant-subsequent year 0.93 0.84-0.99 Adverse events (multiplicative factor)54 Pruritus – mild 0.93 ±25% Pruritus – moderate 0.87 ±25% Pruritus – severe 0.79 ±25% Table 10. Health-Related Quality of Life Utilities for the US Population Age Group Male Female 20–29 0.928 0.913 30–39 0.918 0.893 40–49 0.887 0.863 50–59 0.861 0.837 60–69 0.84 0.811 70–79 0.802 0.771 80–89 0.782 0.724 53 Source: Hammer et al. ©Institute for Clinical and Economic Review, 2016 Page 32 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Transition Probabilities The transition probabilities between advanced disease states (i.e., compensated cirrhosis, decompensated cirrhosis, hepatocellular carcinoma, and liver transplant) are derived from the natural history of patients with hepatitis C virus infection (Table 8). We took this approach because there are limited data on the natural history of PBC in advanced stages of the disease. In addition, clinical evidence indicates that the risk of hepatocellular carcinoma in PBC patients with cirrhosis is similar to that in patients with hepatitis C, which was confirmed with communications with clinical experts. We also believe that this approach is biologically appropriate.55,56 Therefore, we used transition probabilities from a published cost-effectiveness model of hepatitis C.57 Transition probabilities from early stages of PBC (stages 1–3) to compensated cirrhosis, compensated cirrhosis to decompensated cirrhosis, compensated cirrhosis to hepatocellular carcinoma, and liver-related death in PBC stages 1–3 and compensated cirrhosis were calibrated such that the predicted 15-year transplant-free survival from the model matches the reported values based on patients’ levels of ALP and bilirubin at 12 months, as reported by the Global PBC study.22 We also incorporated age- and sex-specific background mortality using U.S. life tables.58 Disease progression from PBC stages 1–3 to compensated cirrhosis and liver-related death, and from compensated cirrhosis to decompensated cirrhosis and liver-related death, was dependent on the response to OCA (i.e., bilirubin and ALP levels after 12 months of treatment). Specifically, we converted the bilirubin and ALP levels into meaningful clinical endpoints using a meta-analysis that links liver function tests to the likelihood of liver transplantation or death.22 This meta-analysis describes the 15-year transplant-free survival of patients who have not responded to UDCA treatment compared to those who achieved improvements in ALP or bilirubin. For each response category, we estimated the corresponding hazard ratio that would result in a 15-year transplant-free survival as reported by the Global PBC study and adjusted transition probabilities based for each response category (Table 11). The hazard ratio was used to adjust transition probabilities to account for slower disease progression because of ALP reduction or an improvement in bilirubin levels. Figures 6–9 show the 15-year transplant free survival from our model and Global PBC study. ©Institute for Clinical and Economic Review, 2016 Page 33 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Table 11. Estimates of Transition Probabilities Annual Transition Range Parameters Probability Disease progression with Normal Bilirubin & ALP ≤ 1.67xULN (calibrated) PBC stages 1–3 to Compensated Cirrhosis 0.0207 ±25% PBC stages 1–3 to liver transplant 0 - PBC stages 1–3 to liver-related death 0 - Compensated cirrhosis to decompensated cirrhosis 0.0059 ±25% Compensated cirrhosis to HCC 0.0021 ±25% Compensated cirrhosis to liver transplant 0 - Compensated cirrhosis to liver-related death 0 - Disease progression with Normal Bilirubin & ALP > 1.67xULN (calibrated) PBC stages 1–3 to Compensated Cirrhosis 0.0929 ±25% PBC stages 1–3 to liver transplant 0 - PBC stages 1–3 to liver-related death 0 - Compensated cirrhosis to decompensated cirrhosis 0.0275 ±25% Compensated cirrhosis to HCC 0.0098 ±25% Compensated cirrhosis to liver transplant 0 - Compensated cirrhosis to liver-related death 0 - Disease progression with Abnormal Bilirubin & ALP ≤ 1.67xULN (calibrated) PBC stages 1–3 to Compensated Cirrhosis 0.2431 ±25% PBC stages 1–3 to liver transplant 0.023 ±25% PBC stages 1–3 to liver-related death 0.0070 ±25% Compensated cirrhosis to decompensated cirrhosis 0.0765 ±25% Compensated cirrhosis to HCC 0.0278 ±25% Compensated cirrhosis to liver transplant 0.023 ±25% Compensated cirrhosis to liver-related death 0.0278 ±25% Disease progression with Abnormal Bilirubin & ALP > 1.67xULN (calibrated) PBC stages 1–3 to Compensated Cirrhosis 0.3415 ±25% ©Institute for Clinical and Economic Review, 2016 Page 34 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents PBC stages 1–3 to liver transplant 0.023 ±25% PBC stages 1–3 to liver-related death 0.0119 ±25% Compensated cirrhosis to decompensated cirrhosis 0.1125 ±25% Compensated cirrhosis to HCC 0.0414 ±25% Compensated cirrhosis to liver transplant 0.023 ±25% Compensated cirrhosis to liver-related death 0.0237 ±25% 59 Decompensated cirrhosis to HCC (Planas et al.) 0.068 ±25% 60,61 Decompensated cirrhosis to transplantation (Davis et al., Thuluvath et al.) 0.023 ±25% Decompensated cirrhosis (first year) to liver-related death (Planas et al.)59 0.182 ±25% Decompensated cirrhosis (subsequent year) to liver-related death (Planas et al.)59 0.112 ±25% 48,62 HCC to liver transplant (Lang et al., Saab et al.) 0.040 ±25% 63 HCC to liver-related death (Fattovich et al.) 0.427 ±25% 64 Liver transplant (first year) to liver-related death (Wolfe et al.) 0.116 ±25% 64 Post-liver transplant to liver-related death (Wolfe et al.) 0.044 ±25% ©Institute for Clinical and Economic Review, 2016 Page 35 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 6. Transplant-free Survival in Patients with Normal Bilirubin and ALP ≤ 1.67xULN Normal Bilirubin & ALP <=1.67xULN 100% Transplant Free Survival 80% 60% 40% Lammers 20% PBC Model 0% 0 5 10 15 Year Figure 7. Transplant-free Survival in Patients with Normal Bilirubin and ALP > 1.67xULN Normal Bilirubin & ALP >1.67xULN 100% 80% Transplant Free Survival 60% 40% Lammers 20% PBC Model 0% 0 5 10 15 Year ©Institute for Clinical and Economic Review, 2016 Page 36 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 8. Transplant-free Survival in Patients with Abnormal Bilirubin and ALP ≤ 1.67xULN Abnormal Bilirubin & ALP <=1.67xULN 100% Transplant Free Survival 80% 60% 40% Lammers 20% PBC Model 0% 0 5 10 15 Year Figure 9. Transplant-free Survival in Patients with Abnormal Bilirubin and ALP > 1.67xULN Abnormal Bilirubin & ALP >1.67xULN 100% Transplant Free Survival 80% 60% 40% Lammers 20% PBC Model 0% 0 5 10 15 Year ©Institute for Clinical and Economic Review, 2016 Page 37 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Sensitivity Analyses We performed one-way sensitivity analyses to identify the key drivers of model outcomes. In addition, probabilistic sensitivity analysis was also performed by simultaneously varying all model parameters over 100,000 simulations, then calculating 95% credible range estimates for each model outcome based on the results. We also performed threshold analyses to identify drug prices necessary to achieve a range of incremental cost-effectiveness ratios, from $50,000 to $150,000 per QALY. Incremental Costs per Outcomes Achieved: Results Figures 10-13 show 15-year cumulative incidence of decompensated cirrhosis, hepatocellular carcinoma, liver transplants, and liver-related deaths in the simulated cohort of patients treated with OCA plus UDCA versus UDCA only. In patients who had an inadequate response to UDCA, treatment with OCA would decrease the 15-year cumulative incidence of decompensated cirrhosis from 12.2% to 4.5%, hepatocellular carcinoma from 9.1% to 4.0%, liver transplant from 4.5% to 1.2%, and liver-related deaths from 16.2% to 5.7%, respectively. In addition, treatment with OCA increased 15-year transplant-free survival from 61% to 73% (Figure 10). The transplant-free survival in PBC patients as predicted by our model was similar to that predicted by the PBC Global study algorithm.65 Compared with the UDCA strategy, treating 10,000 patients using OCA plus UDCA could prevent 770 cases of decompensated cirrhosis, 510 cases of hepatocellular carcinoma, 330 liver transplants and 1,050 liver-related deaths. ©Institute for Clinical and Economic Review, 2016 Page 38 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 10. Cumulative Incidence of Decompensated Cirrhosis Decompensated Cirrhosis 15.0% 12.2% 12.0% UDCA Cumulative Incidences OCA+UDCA 9.0% 6.0% 4.5% 3.0% 0.0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Year Figure 11. Cumulative Incidence of Hepatocellular Carcinoma Hepatocellular Carcinoma 15.0% UDCA 12.0% Cumulative Incidences OCA+UDCA 9.1% 9.0% 6.0% 4.0% 3.0% 0.0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Year ©Institute for Clinical and Economic Review, 2016 Page 39 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 12. Cumulative Incidence of Liver Transplants in PBC Patients Liver Transplant 15.0% UDCA 12.0% Cumulative Incidences OCA+UDCA 9.0% 6.0% 4.5% 3.0% 1.2% 0.0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Year Figure 13. Cumulative Incidence of Liver-related Deaths in PBC Patients Liver-Related Deaths 18.0% 16.2% 15.0% Cumulative Incidences UDCA 12.0% OCA+UDCA 9.0% 5.7% 6.0% 3.0% 0.0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Year ©Institute for Clinical and Economic Review, 2016 Page 40 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 14. Transplant-free Survival in PBC Patients who had Inadequate Response to UDCA 100% 73% 80% Transplant Free Survival UDCA 60% OCA+UDCA 61% 40% 20% 0% 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Year The average life years per patient treated with UDCA versus OCA plus UDCA were 19.97 and 22.23, respectively (increment = 2.26 years). The corresponding average discounted QALYs gained were 10.74 and 11.78, respectively (increment = 1.04 years). The average lifetime discounted cost per patient treated with UDCA was $142,300. Assuming that the price of OCA is $69,350/year, the average lifetime cost of a patient treated with OCA plus UDCA was $633,900 (an increment of $491,400). The incremental cost-effectiveness of OCA plus UDCA was approximately $473,400 per QALY (Table 12). Using the willingness to pay threshold of $100,000 per QALY, OCA at a $69,350/year price is not a cost-effective option in PBC patients who have inadequate response to UDCA. Table 12. Cost-effectiveness of OCA when the Annual Cost of OCA is $69,350 per Year UDCA* OCA + UDCA Undiscounted Life Years 19.97 22.23 Discounted QALYs 10.74 11.78 Discounted Total Cost ($) 142,300 633,900 ICER ($/QALY) 473,400 *Results correspond to inadequate response to UDCA, as observed in POISE study Next we conducted a price threshold analysis to determine the price of OCA at which it becomes cost-effective, using three different willingness-to-pay thresholds (Table 13). We found that OCA would meet thresholds of $50,000, $100,000, and $150,000 per QALY gained if priced below $11,629, $18,445, and $25,261 per year, respectively. ©Institute for Clinical and Economic Review, 2016 Page 41 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Table 13. OCA Price Threshold Analysis for PBC Patients Willingness to Pay ($/QALY) Annual Price of OCA $50,000 $11,629 $100,000 $18,445 $150,000 $25,261 Sensitivity Analyses We conducted one-way sensitivity analyses to identify the parameters to which the model was most sensitive. We have plotted a tornado diagram showing the 15 most sensitive parameters (Figure ES7). We found that the incremental cost-effectiveness ratios were most sensitive to the cost of OCA. For all other parameters, the cost-effectiveness thresholds remained above $300,000/QALY when varied across a plausible range. Figure 15. Tornado Diagram Showing 15 Most Sensitive Model Parameters ©Institute for Clinical and Economic Review, 2016 Page 42 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 6.4 Potential Budget Impact We also used the cost-effectiveness model to estimate the potential total budgetary impact of OCA for PBC patients, based on assumed patterns of product uptake. Potential Budget Impact Model: Methods We used results from the same model employed for the cost-effectiveness analyses to estimate total potential budget impact. Potential budget impact was defined as the total incremental cost of the OCA therapy for the treated population, calculated as incremental health care costs (including drug costs) minus any offsets in these costs from averted disease progression. All costs were undiscounted and estimated over one- and five-year time horizons. The five-year timeframe was of primary interest, given the potential for cost offsets to accrue over time. The potential budget impact analysis included the entire candidate population for treatment, which was considered to be adults with PBC who have either inadequate response to UDCA or are unable to tolerate UDCA. To estimate the size of the potential candidate population for OCA, we first applied the estimated prevalence of PBC in the US. We located one estimate of PBC prevalence in the US. Kim et al.18 used the Rochester Epidemiology Project diagnostic indexes to estimate an age- and sex-adjusted prevalence in Olmsted County, Minnesota, of 40.2 per 100,000. Applying this prevalence to the projected 2016 US population would imply approximately 130,000 individuals with PBC. The Kim at al. study also reported that 43.5% of PBC patients had received UDCA treatment. We assumed that 40% of the treated population with PBC would have inadequate response to UDCA therapy,19 and that another 3% would be unable to tolerate UDCA.2 Applying these percentages resulted in a candidate population size of approximately 24,350 individuals in the US. ICER’s methods for estimating potential budget impact and calculating value-based benchmark prices are described in detail elsewhere. Briefly, our calculations assume that the utilization of new drugs occurs without any payer, provider group, or pharmacy benefit management controls in place, to provide an estimate of “unmanaged” drug uptake by five years after launch. In general, we examine six characteristics of the drug or device and the marketplace to estimate “unmanaged” uptake. These characteristics are listed below:  Magnitude of improvement in clinical safety and/or effectiveness  Patient-level burden of illness  Patient preference (ease of administration)  Proportion of eligible patients currently being treated  Primary care versus specialty clinician prescribing/use  Presence or emergence of competing treatments of equal or superior effectiveness ©Institute for Clinical and Economic Review, 2016 Page 43 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Based on our assessment of these criteria, we assign a new drug or device to one of four categories of unmanaged drug uptake patterns: 1) very high (75% uptake by year 5); 2) high (50% uptake by year 5); 3) intermediate (25% uptake by year 5); and 4) low (10% uptake by year 5). In this analysis, we assumed a 50% uptake pattern for OCA in PBC patients. We assumed that uptake would be high for this drug because of the lack of therapeutic alternatives for PBC patients with inadequate response or intolerance to UDCA. Using this approach to estimate potential budget impact, we then compared our estimates to a budget impact threshold that represents a potential trigger for policy mechanisms to improve affordability, such as changes to pricing, payment, or patient eligibility. As described in ICER’s methods presentation, this threshold is based on an underlying assumption that health care costs should not grow much faster than growth in the overall national economy. From this foundational assumption, our potential budget impact threshold is derived using an estimate of growth in US gross domestic product (GDP) +1%, the average number of new drug approvals by the FDA each year, and the contribution of spending on retail and facility-based drugs to total health care spending. Calculations are performed as shown in Table 14. For 2015-16, therefore, the five-year annualized potential budget impact threshold that should trigger policy action to manage affordability is calculated to total approximately $904 million per year for new drugs. Table 14. Calculation of Potential Budget Impact Threshold Item Parameter Estimate Source 1 Growth in US GDP, 2015-2016 (est.) +1% 3.75% World Bank, 2015 2 Total health care spending ($) $3.08 trillion CMS NHE, 2014 3 Contribution of drug spending to total health 13.3% CMS National Health care spending (%) Expenditures (NHE), Altarum Institute, 2014 4 Contribution of drug spending to total health $410 billion Calculation care spending ($) (Row 2 x Row 3) 5 Annual threshold for net health care cost $15.4 billion Calculation growth for ALL new drugs (Row 1 x Row 4) 6 Average annual number of new molecular 34 FDA, 2014 entity approvals, 2013-2014 7 Annual threshold for average cost growth per $452 million Calculation individual new molecular entity (Row 5 ÷ Row 6) 8 Annual threshold for estimated potential $904 million Calculation budget impact for each individual new molecular entity (doubling of Row 7) ©Institute for Clinical and Economic Review, 2016 Page 44 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Potential Budget Impact Model: Results Table 15 below presents the potential budget impact of one year and five years of OCA in the candidate population, assuming the uptake patterns previously described. (Undiscounted costs per patient for years 1 through 5 are provided in Appendix Table G1.) Results are presented for both one-year and five-year time horizons. Results from the potential budget impact model showed that, with the uptake pattern assumptions mentioned above, an estimated 2,440 individuals would receive OCA in the first year. After one year of treatment, with net annual costs of approximately $69,350 per patient, one-year budget impact is estimated to be $169.2 million. Over the entire five-year time horizon, we estimate that “unmanaged” uptake would lead to approximately 12,200 persons taking OCA. Across the full five-year time horizon, the weighted potential budgetary impact (i.e., adjusted for differing periods of drug utilization and associated cost-offsets) is approximately $128,500 per patient. Total potential budgetary impact over five years is approximately $1.6 billion, with an average budget impact per year of approximately $312.9 million. This annualized potential budget impact is 35% of the budget impact threshold of $904 million for a new drug. Table 15. Estimated Total Potential Budget Impact (BI) of OCA Analytic Horizon = 1 Year Analytic Horizon = 5 Years Eligible Number Annual BI per Total BI Number Weighted BI Average BI Population Treated Patient* (millions) Treated per Patient* per year (millions) OCA 24,350 2,440 $69,350 $169.2 12,200 $128,500 $312.9 *Weighted budget impact calculated by subtracting cost offsets from drug costs for one-year horizon. For five- year horizon, drug costs and cost offsets apportioned assuming 20% of patients in uptake target initiate therapy each year. Those initiating in year 1 receive full drug costs and cost offsets, those initiating in year 2 receive 80% of drug costs and cost offsets, etc. Figure 16 provides findings of multiple analyses that give perspective on the relationship between varying possible drug prices, cost-effectiveness ratios, uptake patterns, and potential budget impact. The vertical axis shows the annualized budget impact, and the horizontal axis represents the percentage of eligible patients treated over a five-year period. The colored lines demonstrate how quickly the annual budget impact increases with increasing percentages of patients treated at four different prices: those at which the cost/QALY = $50,000, $100,000, and $150,000; and the list price used in this analysis (i.e., $69,350 annually for OCA). ©Institute for Clinical and Economic Review, 2016 Page 45 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure 16. Combined Cost-effectiveness and Potential Budget Impact Graph for OCA in PBC Patients 6.5 Value-based Benchmark Prices Our draft value-based benchmark prices for OCA are provided in Table 16. As noted in the ICER methods document, the draft value-based benchmark price for a drug is defined as the price range that would achieve cost-effectiveness ratios between $100,000 and $150,000 per QALY gained, without exceeding the $904 million budget impact threshold for new drugs. As noted previously, the potential budget impact of OCA does not exceed our stated threshold when annualized over a five-year time horizon. That is, the price for OCA that could be charged and not exceed the $904 million annual benchmark is higher than the price range that would achieve $100,000 to $150,000 per QALY gained. Therefore, the draft ICER value-based price benchmark for OCA, with all the assumptions mentioned previously regarding five-year uptake patterns and net costs, is approximately $18,400 to $25,300 per year. This price represents a 64%-73% discount from the annual cost of OCA at list price ($69,350). ©Institute for Clinical and Economic Review, 2016 Page 46 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Table 22. Draft value-based price benchmarks for OCA in PBC patients WAC Price per Cost to Achieve Cost to Achieve Draft Value-Based Year $100K/QALY $150K/QALY Price Benchmark OCA $69,350 $18,445 $25,261 $18,445 to $25,261 6.6 Summary and Comment We conducted a cost-effectiveness analysis by developing a microsimulation model that simulated the long-term outcomes of PBC patients with an inadequate response to UDCA receiving OCA (in addition to UDCA) compared to UDCA alone. We estimated that, in patients who had an inadequate response to UDCA, treatment with OCA would decrease the 15-year cumulative incidence of decompensated cirrhosis from 12.2% to 4.5%, hepatocellular carcinoma from 9.1% to 4.0%, liver transplant from 4.5% to 1.2%, and liver-related deaths from 16.2% to 5.7%. Using the price of OCA as $69,350/year, the incremental cost-effectiveness of OCA plus UDCA was estimated to be approximately $473,400 per QALY, which does not meet the commonly used willingness-to-pay thresholds of $100,000 to $150,000 per QALY. We found that the cost-effectiveness results were most sensitive to the cost of OCA and age of PBC patients. All other parameters did not have a substantial impact on the incremental cost- effectiveness ratios. As with any model, ours has some limitations. A key limitation of this model is the lack of data to inform the natural history of PBC. There were limited data on the transition between advanced disease health states, such as compensated cirrhosis, decompensated cirrhosis, and hepatocellular carcinoma. For this reason, we calibrated the model to meet hard clinical endpoints, including the 15-year transplant-free survival for different ALP and bilirubin levels, and transplant-free survival for UDCA treatment. Another limitation is that our model draws on hepatitis C data to inform cost and quality-of-life parameters. Given the lack of economic and quality of life data for PBC health states, our model assumed that PBC-associated cirrhosis and hepatitis-associated cirrhosis involve similar treatment costs and PBC patients experience similar decrements in quality of life as in hepatitis C patients. Sensitivity analysis on these parameters showed that model outcomes were robust to uncertainty surrounding these cost and quality-of-life inputs. We also used the cost-effectiveness model to estimate the potential total budgetary impact of OCA for PBC patients over five years. Assuming that “unmanaged” uptake would lead to 50% of eligible patients (or approximately 12,200 persons) taking OCA, total potential budgetary impact over five years is approximately $1.6 billion, with an average budget impact per year of approximately $312.9 ©Institute for Clinical and Economic Review, 2016 Page 47 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents million. This annualized potential budget impact is 35% of the budget impact threshold of $904 million for a new drug. ©Institute for Clinical and Economic Review, 2016 Page 48 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 7. Voting Results 7.1 About the New England CEPAC Process During public meetings of the New England CEPAC, the Council deliberates and votes on key questions related to the systematic review of the clinical evidence, an economic analysis of the applications of the medical technologies or treatments under examination, and the supplementary information presented. Council members serve for three year terms and are intentionally selected to represent a range of expertise and diverse perspectives. To maintain the objectivity of New England CEPAC and ground the conversation in the interpretation of the published evidence, members are not pre-selected based on the topic being addressed. Acknowledging that any judgment of evidence is strengthened by real-life clinical and patient perspectives, clinical representatives with expertise in the subject matter are recruited for each meeting topic and provide input to Council members before the meeting to help clarify their understanding of the interventions being analyzed in the evidence review. The same clinical experts serve as a resource to the Council during their deliberation, and they help form recommendations with the Council on ways the evidence can be applied to policy and practice. At each meeting, after the Council votes, a Policy Roundtable discussion is held with the Council, clinical experts, and representatives from relevant manufacturers, provider groups, payers, and patient groups. This is intended to bring stakeholders into the discussion on how best to apply the evidence to guide patient education, clinical practice, and coverage and public policies. Participants on Policy Roundtables are selected for their expertise on the specific meeting topic, are different for each meeting, and do not vote on any questions. At the July 15, 2016 meeting, the Council discussed issues regarding the application of the available evidence to help patients, providers, and payers address the important questions related to the treatment of primary biliary cholangitis and non-alcoholic steatohepatitis with obeticholic acid. Following an evidence presentation and public comments, the Council voted on key questions concerning the clinical effectiveness and value of obeticholic acid. These questions are developed by the ICER research team for each assessment, with input from the New England CEPAC Advisory Board to ensure that the questions are framed to address the issues that are most important in applying the evidence to support clinical practice and medical policy decisions. The voting results are presented in the section below, along with comments reflecting considerations mentioned by the Council members during the voting process. In its deliberations and voting related to value, the Council made use of a value assessment framework with four different components of care value, a concept which represents the long-term perspective, at the individual patient level, on patient benefits and the incremental costs to achieve ©Institute for Clinical and Economic Review, 2016 Page 49 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents those benefits. The four components of care value are comparative clinical effectiveness, incremental cost per outcomes achieved, additional benefits or disadvantages, and contextual considerations regarding the illness or therapy. Figure 17. Care Value Framework There are four elements to consider when deliberating on care value: 1. Comparative clinical effectiveness is a judgment of the overall difference in clinical outcomes between two interventions (or between an intervention and placebo), tempered by the level of certainty possible given the strengths and weaknesses of the body of evidence. New England CEPAC uses the ICER Evidence Rating Matrix as its conceptual framework for considering comparative clinical effectiveness. 2. Incremental cost per outcomes achieved is the average per-patient incremental cost of one intervention compared to another to achieve a desired “health gain,” such as an additional stroke prevented, case of cancer diagnosed, or gain of a year of life. Alternative interventions are compared in terms of cost per unit of effectiveness, and the resulting comparison is presented as a ratio: a “cost per outcome achieved.” Relative certainty in the cost and outcome estimates continues to be a consideration. As a measure of incremental costs per outcomes achieved, ICER follows common academic and World Health Organization (WHO) standards by using cost per quality-adjusted life years (QALYs) and adopting thresholds at $100,000 per QALY and $150,000 per QALY as guides to reasonable ratios of incremental costs per outcomes achieved. 3. Other benefits or disadvantages refers to any significant benefits or disadvantages offered by the intervention to the individual patient, caregivers, the delivery system, other patients, or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Examples of other benefits include mechanisms of treatment delivery that require many fewer visits to the clinician’s office, treatments that reduce disparities ©Institute for Clinical and Economic Review, 2016 Page 50 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents across various patient groups, and new potential mechanisms of action for treating clinical conditions that have demonstrated low rates of response to currently available therapies. Other disadvantages could include increased burden of treatment on patients or their caregivers. For each intervention evaluated, it will be open to discussion whether other benefits or disadvantages such as these are important enough to factor into the overall judgment of care value. There is no quantitative measure for other benefits or disadvantages. 4. Contextual considerations include ethical, legal, or other issues (but not cost) that influence the relative priority of illnesses and interventions. Examples of contextual considerations include whether there are currently any existing treatments for the condition, whether the condition severely affects quality of life or not, and whether the condition affects priority populations. There is no quantitative measure for the role of contextual considerations in an overall judgment of care value. 7.2 Clinical Effectiveness Voting Results 1. For patients with PBC (primary biliary cholangitis or primary biliary cirrhosis), who fail to achieve an adequate reduction in alkaline phosphatase on ursodeoxycholic acid (UDCA) monotherapy, is the evidence adequate to demonstrate a net health benefit with the addition of obeticholic acid to continuing therapy with UDCA? Yes: 10 votes No: 4 votes Comments: Ten members of the New England CEPAC voted that the evidence was adequate to demonstrate a net health benefit for the combination of obeticholic acid and UDCA versus UDCA alone. Members voting “yes” felt that the surrogate endpoint (reduction in ALP levels) could be reasonably expected to correlate with positive clinical outcomes (e.g., delayed progression of the disease or reduced need for liver transplantation). Of the four members who voted that there was not sufficient evidence to demonstrate a net health benefit, several expressed concern about the use of surrogate endpoints over long-term clinical endpoints, and did not feel that there was sufficient evidence to demonstrate that reduction in ALP improves long-term clinical outcomes. ©Institute for Clinical and Economic Review, 2016 Page 51 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 7.3 Care Value Voting Results 2. Given the available evidence for patients with PBC, what is the care value of adding obeticholic acid to UDCA alone? Low: 8 votes Intermediate: 6 votes High: 0 votes Comments: The majority of the discussion around care value centered on the high incremental cost-effectiveness ratio for OCA in PBC, as well as contextual considerations. Those voting for “intermediate” value highlighted the potential for productivity gains from effective PBC treatment since many patients are working-age women. They also highlighted the relatively high percentage of patients who do not respond to usual care with UDCA alone. The primary rationale for those voting “low” value was the high incremental cost-effectiveness ratio for OCA, driven by the high list price of OCA. Many council members felt that if the price was lower, they would have voted for higher care value. ©Institute for Clinical and Economic Review, 2016 Page 52 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 8. Roundtable Discussions and Key Policy Implications Following its deliberation on the evidence, the New England CEPAC engaged in a moderated discussion about the use of OCA for the treatment of PBC with a Policy Roundtable that included two clinical experts, one patient, a public payer representative, a private payer representative, and a manufacturer representative. This discussion reflected multiple perspectives and opinions, and therefore, none of the recommendations below should be taken as a consensus view held by all participants. The names of the Policy Roundtable participants are shown below. All participants completed potential conflict of interest forms in advance of the meeting; their written statements can be found in Appendix G. Roundtable Participant Association Kathleen Corey, MD, MPH Director, MGH Fatty Liver Clinic, Massachusetts General Hospital Gastrointestinal Unit Judith Donovan Patient with PBC Barbara Henry, RPh Senior Clinical Pharmacy Coordinator, Harvard Pilgrim Health Care Juan Carlos Lopez- Senior Vice President, Global Medical Affairs, Intercept Pharmaceuticals Talavera, MD, PhD Daniel S. Pratt, MD Clinical Director, Liver Transplantation; Director, Autoimmune & Cholestatic Liver Center, Massachusetts General Hospital Tom Simpatico, MD Chief Medical Officer, Vermont Department of Health Access The roundtable discussion was facilitated by Dr. Steven Pearson, MD, MSc, President of ICER. The main themes and recommendations from the discussion are organized by audience and summarized below. Patients Take a leadership role in advocating for the inclusion of patient-relevant outcomes in clinical studies. Measures available in clinical trials for PBC are not necessarily defined in a patient-centric manner. For example, fatigue is a well-known consequence of PBC and its presence is documented in trials. However, patients are most interested in the impact of fatigue on their activities of daily living and overall quality of life as well as whether use of a new treatment improves or worsens the situation. Patient groups can work with speciality societies, manufacturers, and researchers to identify functional, quality of life, and other measures that more accurately document their experience with PBC as well as the impact of treatment on that experience. ©Institute for Clinical and Economic Review, 2016 Page 53 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Clinicians Propose standardized criteria for defining inadequate response to treatment with UDCA and OCA. Clinicians agree that UDCA should be used for at least 12 months before prescribing OCA for patients with inadequate biochemical response to UDCA, as reflected in the FDA label.13 Unfortunately, different thresholds for reduction in ALP levels are used in the literature, with widely differing resultant rates of response, so there is a need for the clinical community to develop standardized criteria for biochemical response that can be used to guide treatment with UDCA and OCA. There is an opportunity for organizations such as the American Association for the Study of Liver Diseases (AASLD) to lead consensus development efforts around this goal. In a related vein, there is not enough current research or clinical experience to define rules for stopping treatment after inadequate response to OCA. A period of 12 months of OCA use has been discussed, but no biochemical criteria have yet been proposed. The AASLD and other clinical organizations can work with the manufacturer to develop and update guidelines for treatment cessation as appropriate data are acquired. Employ strategies to improve medication adherence to UDCA and diminish side effects. The clinical experience presented at the policy roundtable suggests that rates of real intolerance to UDCA may be substantially lower than the estimated 3%. In addition, as with OCA, patients who do not achieve full biochemical response still may see some clinical benefit from continued UDCA treatment. Strategies should be developed based on expert opinion to maintain adherence to UDCA and isolate true cases of intolerance. These strategies could include methods for managing common and bothersome side effects, which are mainly gastrointestinal. Consider use of histological data from liver biopsies for important baseline and prognostic information. Clinical experts on the roundtable agreed that liver biopsy is not necessary for diagnosis in all cases and may represent a burden to patients. However, biopsies may provide important prognostic information and establish a baseline so that future histology studies could help determine the level of benefit provided by therapy. ©Institute for Clinical and Economic Review, 2016 Page 54 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Insurers When developing coverage criteria for OCA: 1. Determine the diagnosis of primary biliary cholangitis in accordance with current clinical guidelines. According to guidelines from the American Association for the Study of Liver Diseases (AASLD),4 a patient must meet two of three criteria in order to be diagnosed with PBC:  Biochemical evidence of cholestasis based on alkaline phosphatase elevation  Presence of antimitochondrial antibodies (AMA)  Histologic evidence of nonsuppurative destructive cholangitis and destruction of interlobular bile ducts Insurers should therefore not mandate the use of liver biopsy for diagnosis. While clinical experts mentioned that histological data may provide useful prognostic information, liver biopsy is not necessary for diagnosis and some patients may be reluctant to undergo the procedure. 2. Use clinical experts’ opinions and consensus to define criteria for starting and for stopping OCA treatment, as well as determining intolerance to UDCA. As noted above, there is general consensus that the adequacy of biochemical response to UDCA be assessed after 12 months of therapy. There is currently no agreement, however, on what constitutes an inadequate biochemical response. Nevertheless, insurers should document specific criteria for determining inadequate response to UDCA (e.g., a 12-month trial plus inadequate response using “Paris” criteria) as an appropriate rule for initiating OCA treatment. Once more empiric data are available on a suitable duration for a trial of therapy on OCA, a similarly-defined “stopping rule” for OCA can be defined. In addition, given that true intolerance to UDCA is likely to be quite low, criteria for management of side effects and other concerns should be developed to isolate true cases of intolerance to UDCA. 3. Do not exclude patients with moderate-severe PBC from coverage for OCA. While clinical experts acknowledged that there is currently very limited evidence on the effects of OCA in patients with moderate-severe PBC, there is no clear reason to assume that these patients would not also see benefit from the drug, and therefore no reason to exclude them from coverage for OCA. ©Institute for Clinical and Economic Review, 2016 Page 55 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Avoid outcomes-based contracting for OCA in the absence of clearly defined clinical endpoints. Because not all patients in the OCA trials achieved the response endpoint (pooled rate: 45%), 16 outcomes-based contracting is a potential consideration for insurers to manage the drug’s cost. However, a very large proportion of patients experienced some level of benefit from treatment in clinical trials, so the creation of a single threshold for response is challenging. In addition, there are no clear data linking different levels of ALP response to important and measurable clinical outcomes such as liver transplantation, making consideration of outcomes-based contracting problematic. Manufacturers Include a broad spectrum of PBC patients in future studies of OCA. As noted in our report and at the FDA Advisory Committee discussions, patients in the POISE trial were predominantly in earlier stages of PBC. For example, only 11% of patients had abnormal bilirubin levels at baseline. There is an ongoing clinical trial of OCA in patients with more severe disease that should address this concern, but additional studies should seek to enroll patients who are at varying levels of disease progression. Assess the use of non-invasive tests for obtaining information on disease status. There is broad clinical interest in understanding the role of non-invasive tests like transient elastography in diagnosis, prognosis, and ongoing monitoring disease. Many patients are reluctant to undergo initial and/or repeat liver biopsy because of the inconvenience, discomfort, and potential risks of the procedure. The clinical development program for OCA represents an opportunity to assess the accuracy and clinical utility of these non-invasive tests for PBC. Seek solutions to the barriers that prevent manufacturers from becoming more transparent about the basis for their pricing decisions. Much attention is being paid to a need for increased transparency around drug pricing in state legislatures around the country and in the media. Manufacturers need to overcome institutional barriers that prevent concrete demonstration of the rationale for their pricing decisions. For a drug like OCA, the only medication available to treat patients with a serious condition who have not responded to alternative treatment, there is no market competition. Justification of the pricing strategy in this type of situation is therefore even more important for patients, payers, and policymakers to understand. ©Institute for Clinical and Economic Review, 2016 Page 56 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Clinical Research Community Increase efforts to include patient-relevant outcomes in clinical studies. Treatment with OCA will be lifelong for a majority of patients and pruritus is known to be not only a side effect of the medication but also a major determinant of quality of life for this population. It is not sufficient to document patient-relevant outcomes as discrete adverse events. Every effort must be made to include in all trials the impact of PBC and its treatment on activities of daily living, overall quality of life and the overall patient experience. Develop long-term clinical studies to enable greater understanding of the relationship between intermediate endpoints and important clinical outcomes. Data are currently limited on the relationship between surrogate markers such as ALP and longer- term outcomes in PBC, including the development of cirrhosis and requirements for liver transplant. Long-term studies seeking to clarify the natural history of PBC should be made a clear research priority. Design and conduct studies to ascertain the validity of biomarkers and other factors in identifying which patients are likely to undergo rapid disease progression. There are currently very few predictors of the progression of PBC, particularly when the disease is diagnosed in its earliest, asymptomatic stages. The research community should prioritize study designs that can measure the predictive power of multiple clinical factors over relatively short periods of follow-up (e.g., nested case-control designs) and/or take advantage of existing registries for this purpose. **** This is the first ICER review of obeticholic acid for PBC. ©Institute for Clinical and Economic Review, 2016 Page 57 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents References 1. Beuers U, Gershwin ME, Gish RG, et al. Changing Nomenclature for PBC: From 'Cirrhosis' to 'Cholangitis'. The American journal of gastroenterology. 2015;110(11):1536-1538. 2. Intercept. Corporate Presentation. 2016. 3. Boonstra K, Beuers U, Ponsioen CY. Epidemiology of primary sclerosing cholangitis and primary biliary cirrhosis: a systematic review. Journal of hepatology. 2012;56(5):1181-1188. 4. Lindor KD, Gershwin ME, Poupon R, Kaplan M, Bergasa NV, Heathcote EJ. Primary biliary cirrhosis. Hepatology (Baltimore, Md.). 2009;50(1):291-308. 5. Selmi C, Bowlus CL, Gershwin ME, Coppel RL. Primary biliary cirrhosis. Lancet (London, England). 2011;377(9777):1600-1609. 6. Carey EJ, Ali AH, Lindor KD. Primary biliary cirrhosis. Lancet (London, England). 2015;386(10003):1565-1575. 7. Metcalf JV, Mitchison HC, Palmer JM, Jones DE, Bassendine MF, James OF. Natural history of early primary biliary cirrhosis. Lancet (London, England). 1996;348(9039):1399-1402. 8. Trivedi PJ, Lammers WJ, van Buuren HR, et al. Stratification of hepatocellular carcinoma risk in primary biliary cirrhosis: a multicentre international study. Gut. 2016;65(2):321-329. 9. Huet PM, Vincent C, Deslauriers J, et al. Portal hypertension in primary biliary cirrhosis (PBC): A reversible condition? Yes, but not in all UDCA treated patients. Hepatol Res. 2009;39(10):1032-1038. 10. Poupon RE, Lindor KD, Cauch-Dudek K, Dickson ER, Poupon R, Heathcote EJ. Combined analysis of randomized controlled trials of ursodeoxycholic acid in primary biliary cirrhosis. Gastroenterology. 1997;113(3):884-890. 11. van Buuren HR, Lammers WJ, Harms MH, Hansen BE. Surrogate Endpoints for Optimal Therapeutic Response to UDCA in Primary Biliary Cholangitis. Digestive diseases (Basel, Switzerland). 2015;33 Suppl 2:118-124. 12. FDA. FDA approves Ocaliva for rare, chronic liver disease. FDA News Release 2016; http://www.fda.gov/newsevents/newsroom/pressannouncements/ucm503964.htm. Accessed 2016-06-20. 13. FDA. 2016 Meeting Materials, Gastrointestinal Drugs Advisory Committee. 2016; http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/Gastrointes tinalDrugsAdvisoryCommittee/ucm486125.htm. Accessed 2016-05-05. 14. Hirschfield GM, Mason A, Luketic V, et al. Efficacy of obeticholic acid in patients with primary biliary cirrhosis and inadequate response to ursodeoxycholic acid. Gastroenterology. 2015;148(4):751-761.e758. 15. Kowdley KV, Shah H, Mason A, et al. Long-term safety and efficacy of obeticholic acid treatment in primary biliary cirrhosis after more than 4 years of treatment. Hepatology (Baltimore, Md.). 2015;62:521A-522A. 16. Kowdley K, Nevens F, Andreone P, et al. The first new trials for primary biliary cirrhosis in a decade: An international program evaluating the farnesoid X receptor agonist obeticholic acid. Gastroenterology. 2015;148(4):S991-S992. ©Institute for Clinical and Economic Review, 2016 Page 58 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 17. Nevens F, Andreone P, Mazzella G, et al. An international phase 3 study of the FXR agonist obeticholic acid in PBC patients: Effects on markers of cholestasis associated with clinical outcomes and hepatocellular damage. Hepatology (Baltimore, Md.). 2014;60(8). 18. Kim WR, Lindor KD, Locke GR, 3rd, et al. Epidemiology and natural history of primary biliary cirrhosis in a US community. Gastroenterology. 2000;119(6):1631-1636. 19. Poupon R. Primary biliary cirrhosis: a 2010 update. Journal of hepatology. 2010;52(5):745- 758. 20. Eaton JE, Lindor KD. Primary Biliary Cirrhosis. In: Feldman M, Friedman LS, Brandt LJ, eds. Sleisenger and Fordtran's gastrointestinal and liver disease: pathophysiology/diagnosis/management. Tenth edition. ed. Philadelphia, PA: Saunders/Elsevier; 2016:1512-1523. 21. Kuiper EM, Hansen BE, de Vries RA, et al. Improved prognosis of patients with primary biliary cirrhosis that have a biochemical response to ursodeoxycholic acid. Gastroenterology. 2009;136(4):1281-1287. 22. Lammers WJ, van Buuren HR, Hirschfield GM, et al. Levels of alkaline phosphatase and bilirubin are surrogate end points of outcomes of patients with primary biliary cirrhosis: an international follow-up study. Gastroenterology. 2014;147(6):1338-1349.e1335; quiz e1315. 23. Boulton-Jones JR, Fenn RM, West J, Logan RF, Ryder SD. Fracture risk of women with primary biliary cirrhosis: no increase compared with general population controls. Aliment Pharmacol Ther. 2004;20(5):551-557. 24. FDA. Letter of approval for OCALIVA (obeticholic acid). 2016; http://www.accessdata.fda.gov/drugsatfda_docs/appletter/2016/207999Orig1s000ltr.pdf. Accessed 2016-06-20. 25. Whitlock EP, Lopez SA, Chang S, Helfand M, Eder M, Floyd N. AHRQ series paper 3: identifying, selecting, and refining topics for comparative effectiveness systematic reviews: AHRQ and the effective health-care program. Journal of clinical epidemiology. 2010;63(5):491-501. 26. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Int J Surg. 2010;8(5):336-341. 27. FDA. Developing Products for Rare Diseases & Conditions. http://www.fda.gov/ForIndustry/DevelopingProductsforRareDiseasesConditions/ucm20055 25.htm. Accessed 2016-05-07. 28. Paumgartner G, Beuers U. Mechanisms of action and therapeutic efficacy of ursodeoxycholic acid in cholestatic liver disease. Clin Liver Dis. 2004;8(1):67-81, vi. 29. MedicineNet. Primary Biliary Cirrhosis Treatment (PBC). http://www.medicinenet.com/primary_biliary_cirrhosis_treatment_pbc/page2.htm#ursode oxycholic_acid_udca. Accessed 2016-05-07. 30. Pares A, Caballeria L, Rodes J. Excellent long-term survival in patients with primary biliary cirrhosis and biochemical response to ursodeoxycholic Acid. Gastroenterology. 2006;130(3):715-720. 31. GENFIT. GENFIT to begin elafibranor clinical program in PBC. http://www.genfit.com/wp- content/uploads/2016/03/2016.03.31-PR-GENFIT-RD-News.pdf. Accessed 2016-05-07. 32. Cook DJ, Mulrow CD, Haynes RB. Systematic reviews: synthesis of best evidence for clinical decisions. Annals of internal medicine. 1997;126(5):376-380. ©Institute for Clinical and Economic Review, 2016 Page 59 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 33. Ollendorf DA, Pearson SD. An integrated evidence rating to frame comparative effectiveness assessments for decision makers. Med Care. 2010;48(6 Suppl):S145-152. 34. Beuers U, Jones DE, Mayo MJ, et al. Fxr agonist obeticholic acid: Pruritus, a common side effect ameliorated by dose titration. Hepatology (Baltimore, Md.). 2014;60(8). 35. Peters Y, Hooshmand-Rad R, Pencek R, et al. Long-term safety of obeticholic acid in patients with primary biliary cirrhosis. Digestive and Liver Disease. 2016;48:e117-e118. 36. Kowdley KV, Luketic VA, Jones DE, et al. The first new monotherapy therapeutic PBC study in a decade? An international study evaluating the farnesoid X receptor agonist obeticholic acid in PBC. Hepatology (Baltimore, Md.). 2011;54:416A-417A. 37. U.S. Preventive Services Task Force. Procedure Manual. Agency for Healthcare Research and Quality;2008. 38. Trauner MH, Nevens F, Andreone P, et al. Durable response in the markers of cholestasis through 18 months of open-label long term safety extension study of obeticholic acid in primary biliary cirrhosis. Gastroenterology. 2016;150(4):S1073-S1074. 39. Intercept Pharmaceuticals Inc. New Drug Application for Obeticholic Acid (OCA) for the Treatment of Primary Biliary Cirrhosis (PBC) (New Drug Application [NDA] 207999). In: Committee GDA, ed: U.S. Food and Drug Administration; March 14, 2016. 40. Mayo M, Kremer AE, Beuers U, et al. Mitigation of pruritus during obeticholic acid treatment in patients with primary biliary cirrhosis: Strategies and successes. Gastroenterology. 2016;150(4):S1072. 41. U.S. Food and Drug Administration. FDA Briefing Package. In: Committee GDA, ed: Office of Drug Evaluation III; Division of Gastroenterology and Inborn Errors of Metabolism (DGIEP); and Office of Translational Sciences, Office of Biostatistics, Division of Biometrics 3; April 7, 2016. 42. Jones D, Pencek R, Hooshmand-Rad R, Shapiro D. Efficacy and safety of obeticholic acid (OCA) in PBC patients with advanced disease as evidenced by abnormal bilirubin: An integrated analysis. Hepatology (Baltimore, Md.). 2015;62:519A-520A. 43. Mayne TJ, Pencek R, Marmon T, Shapiro D. Obeticholic acid (OCA) has incremental efficacy in patients with higher baseline alkaline phosphatase (ALP) levels. Hepatology (Baltimore, Md.). 2015;62:921A-922A. 44. Pencek R, Lutz K, Marmon T, MacConell L, Shapiro D. Integrated analysis of efficacy of obeticholic acid evaluating subgroups of patients less likely to respond to UDCA (age, age at time of PBC diagnosis, and sex). Hepatology (Baltimore, Md.). 2015;62:530A. 45. Bowlus CL, Pockros PJ, Drenth J, et al. Obeticholic acid in PBC patients: The utility of titration based on therapeutic response and tolerability. Hepatology (Baltimore, Md.). 2014;60(8). 46. Pares A, Pencek R, Peters Y, et al. Fxr agonism with obeticholic acid may attenuate bone mineral density decrease in subjects with primary biliary cirrhosis. Journal of hepatology. 2015;62(22). 47. Malecha ES, Beuers U, Liberman A, Hooshmand-Rad R, Peters Y, Pencek R. Baseline factors predicting obeticholic acid induced pruritus in patients with PBC. Gastroenterology. 2016;150(4):S1058. 48. Saab S, Hunt DR, Stone MA, McClune A, Tong MJ. Timing of hepatitis C antiviral therapy in patients with advanced liver disease: A decision analysis model. Liver Transpl. 2010;16(6):748-759. ©Institute for Clinical and Economic Review, 2016 Page 60 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 49. Centers for Medicare & Medicaid Services. Physician Fee Schedule. Retrieved from: https://www.cms.gov/medicare/medicare-fee-for-service-payment/physicianfeesched/ (last accessed: May 1, 2016). 2016. 50. Truven Health Analytics. RedBook Online. 2016; http://www.redbook.com/redbook/online/. 51. Chong CAKY, Gulamhussein A, Heathcote EJ, et al. Health-state utilities and quality of life in hepatitis C patients. The American journal of gastroenterology. 2003;98(3):630-638. 52. Siebert U, Sroczynski G, Rossol S, et al. Cost effectiveness of peginterferon-2b plus ribavirin versus interferon -2b plus ribavirin for initial treatment of chronic hepatitis C. Gut. 2003;52(3):425. 53. Hanmer J, Lawrence WF, Anderson JP, Kaplan RM, Fryback DG. Report of nationally representative values for the noninstitutionalized US adult population for 7 health-related quality-of-life scores. Medical Decision Making. 2006;26(4):391-400. 54. Hawe E, McBride D, Balp MM, Tian H, Halliday A, Stull DE. EQ-5D Utilities in Chronic Spontaneous/Idiopathic Urticaria. PharmacoEconomics. 2016;34(5):521-527. 55. Bruix J, Sherman M. Management of hepatocellular carcinoma: An update. Hepatology (Baltimore, Md.). 2011;53(3):1020-1022. 56. Caballeria L, Pares A, Castells A, Gines A, Bru C, Rodes J. Hepatocellular carcinoma in primary biliary cirrhosis: similar incidence to that in hepatitis C virus-related cirrhosis. The American journal of gastroenterology. 2001;96(4):1160-1163. 57. Chhatwal J, Kanwal F, Roberts MS, Dunn MA. Cost-effectiveness and budget impact of hepatitis C virus treatment with sofosbuvir and ledipasvir in the United States. Annals of internal medicine. 2015;162(6):397-406. 58. Arias E. United States Life Tables, 2011. Natl Vital Stat Rep 2015;64:1-63. 59. Planas R, Ballesté B, Antonio Álvarez M, et al. Natural history of decompensated hepatitis C virus-related cirrhosis. A study of 200 patients. Journal of hepatology. 2004;40(5):823-830. 60. Davis G, Alter M, El-Serag H, Poynard T, Jennings L. Aging of hepatitis C virus (HCV)-infected persons in the United States: a multiple cohort model of HCV prevalence and disease progression. Gastroenterology. 2010;138(2):513-521. 61. Thuluvath P, Guidinger M, Fung J, Johnson L, Rayhill S, Pelletier S. Liver transplantation in the United States, 1999–2008. American Journal of Transplantation. 2010;10(4p2):1003- 1019. 62. Lang K, Danchenko N, Gondek K, Shah S, Thompson D. The burden of illness associated with hepatocellular carcinoma in the United States. Journal of hepatology. 2009;50(1):89-99. 63. Fattovich G, Giustina G, Degos F, et al. Morbidity and mortality in compensated cirrhosis type C: a retrospective follow-up study of 384 patients. Gastroenterology. 1997;112(2):463- 472. 64. Wolfe R, Roys E, Merion R. Trends in Organ Donation and Transplantation in the United States, 1999–2008. American Journal of Transplantation. 2010;10(4p2):961-972. 65. Harms MH, Lammers WJ, Marmon R, et al. Improvement in Estimated Liver Transplantation Free Survival After 1 Year of Obeticholic Acid Treatment. Journal of hepatology. 2016;64:S631-S832. ©Institute for Clinical and Economic Review, 2016 Page 61 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents APPENDICES ©Institute for Clinical and Economic Review, 2016 Page 62 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix A. Evidence Review Methods and PRISMA Table A1. PRISMA 2009 Checklist # Checklist item TITLE Title 1 Identify the report as a systematic review, meta-analysis, or both. ABSTRACT Structured summary 2 Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number. INTRODUCTION Rationale 3 Describe the rationale for the review in the context of what is already known. Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS). METHODS Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number. Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale. Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched. Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated. Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis). Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators. Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made. ©Institute for Clinical and Economic Review, 2016 Page 63 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Risk of bias in individual 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done studies at the study or outcome level), and how this information is to be used in any data synthesis. Summary measures 13 State the principal summary measures (e.g., risk ratio, difference in means). Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I2) for each meta-analysis. Risk of bias across studies 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies). Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified. RESULTS Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram. Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot. Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. Risk of bias across studies 22 Present results of any assessment of risk of bias across studies (see Item 15). Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]). DISCUSSION Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers). Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias). Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. From: Moher D, Liberati A, Tetzlaff J, Altman DG. The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097 ©Institute for Clinical and Economic Review, 2016 Page 64 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Search Strategies Table A2. Search Strategy of Medline 1996 to Present with Daily Update, EBM Reviews - Cochrane Database of Systematic Reviews, EBM Reviews - Cochrane Central Register of Controlled 1 exp Liver Cirrhosis, Biliary/ 3,905 2 (primary biliary cirrhosis or primary biliary cholangitis or PBC).mp. 5,371 3 1 or 2 6,095 4 (obeticholic acid or OCA or INT-747).mp. 665 5 3 and 4 34 Date of Search: April 12, 2016 Table A3: Search Strategy of Embase on April 12, 2016 #4 #3 AND [humans]/lim AND [english]/lim NOT [medline]/lim 69 #3 #1 AND #2 128 #2 'obeticholic acid' OR oca OR 'int 747' 2,774 #1 'primary biliary cirrhosis'/exp OR 'primary biliary cirrhosis' OR 'primary biliary 26,047 cholangitis' OR pbc Study Selection We performed screening at both the abstract and full-text level. Two investigators screened abstracts identified through electronic searches according to the inclusion and exclusion criteria described earlier. All exclusions were validated by a third reviewer. We did not exclude any study at abstract-level screening due to insufficient information. For example, an abstract that did not report an outcome of interest would be accepted for further review in full text. We retrieved the citations that were accepted during abstract-level screening for full text appraisal. Two investigators reviewed full papers and provided justification for exclusion of each excluded study; a third investigator resolved any discrepancies in selection as necessary. Data Extraction and Quality Assessment Summary tables of extracted data are available in Appendix E. We abstracted data from conference abstracts and posters affiliated with the clinical trials included in the evidence review. We used criteria published by the U.S. Preventive Services Task Force (USPSTF) to assess the quality of RCTs using the categories “good,” “fair,” or “poor.”37 Guidance for quality ratings using these criteria is presented below. Good: Meets all criteria: Comparable groups are assembled initially and maintained throughout the study; reliable and valid measurement instruments are used and applied equally to the groups; ©Institute for Clinical and Economic Review, 2016 Page 65 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents interventions are spelled out clearly; all important outcomes are considered; and appropriate attention is paid to confounders in analysis. In addition, intention to treat analysis is used for RCTs. Fair: Studies were graded "fair" if any or all of the following problems occur, without the fatal flaws noted in the "poor" category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred with follow-up; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are addressed. Intention to treat analysis is done for RCTs. Poor: Studies were graded "poor" if any of the following fatal flaws exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied equally among groups (including not masking outcome assessment); and key confounders are given little or no attention. For RCTs, intention to treat analysis is lacking. Assessment of Bias As part of our quality assessment, we evaluated the evidence base for the presence of potential publication bias. Given the emerging nature of the evidence for newer treatments, we performed an assessment of publication bias using the clinicaltrials.gov database of trials. We scanned the site to identify studies completed more than two years ago that would have met our inclusion criteria and for which no findings have been published. No studies were identified using this criterion. Data Synthesis and Statistical Analyses Given the small numbers of relevant studies for OCA in PBC, we judged there to be no role for formal meta-analysis to generate pooled estimates of treatment effect. ©Institute for Clinical and Economic Review, 2016 Page 66 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Figure A1. PRISMA Flow Chart Showing Results of Literature Search for OCA in PBC and NASH 234 potentially relevant references screened 176 citations excluded 18 Population 16 Intervention/Comparator 77 Study Type 65 Duplicates 58 references for full text review 36 citations excluded (not a population of interest, unrelated to OCA, abstracts/posters with duplicated data, inappropriate 22 TOTAL study type) 3 RCTs (1 PBC, 2 NASH) 19 conference abstracts/posters (15 PBC, 4 NASH) ©Institute for Clinical and Economic Review, 2016 Page 67 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix B. Summary Evidence Table Author & Interventions Inclusion and Year of Study Design (n) & Dosing Exclusion Patient Characteristics Outcomes Harms Publication Schedule Criteria (Trial) Beuers, 2014 See Luketic See Luketic See Luketic See Luketic See Luketic Pruritus 1) 2) 3) Outcomes n, n, n, Abstract in Patients in % % % Hepatology the OCA, 5- Baseline 47, 37, 44, Pruritus 64 53 60 10mg were TE 28, 39, 50, (POISE) titrated Pruritus 38 56 68 based on 0-6mo 11, 22, 38, biochemical 15 31 52 response and >6-12mo 8, 14, 9, tolerability 11 20 14 Pruritus 0, 1, 7, discon’d 0 1 10 Mean 25, 21, 20, Baseline 28 26 25 VAS, SD 0-6mo 21, 23, 26, 23 27 29* >6-12mo 25, 24, 23, 25 27 30 *p=0.0314; OCA 10mg vs. placebo (did not persist at 1 year) Bowlus, See Luketic See Luketic See Luketic See Luketic Outcomes % Change in *p<0 Outcomes @ 12 Pruritus (%) 2014 @ 12 mo responders ALP (U/L) .000 mo Subgroup 7% UDCA intolerant Remained 53* -80 (12)** 1 for Remained @ 58 Abstract in analysis for @ 5mg OCA 5mg (n=36) Hepatology the titration (n=36) vs. Titrated to 10mg 55 arm (n=69) Titrated to 39* -126 (17)** place (n=33) (POISE) 10mg (n=33) bo p=NR ©Institute for Clinical and Economic Review, 2016 Page 68 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents from CMH test ** p<0.0001 for change from baseline from paired 1 subject discontinued due to t-test pruritus in the 10mg arm between 0-6 months Results reported at 6 months were also SS 4 subjects who were able to titrate did not due to pruritus Hirschfield, Phase II 1) Placebo Inclusion: Sex (%): Mean ALP change: Overall AEs: 2015 RCT (38) -Adults aged Male- 5 mITT- Placebo- 84% 18-75 with Female: 95 1) -3% (95% CI -7%, +2%) OCA- 96% (747-202) 41 North 2) OCA, 10mg PBC 2) -24% (95% CI -30%, -18%) American and (38) -Stable does of Age (years): 3) -25% (95% CI -30%, -20%) Overall SAEs due to pruritus (n, %): Publication European UDCA for 6 55 4) -21% (95% CI -30%, -12%) 30/37, 81 in Gastro- Centers 3) OCA, 25mg months before p<0.001 for all OCA groups vs. placebo, and for ITT enterology (48) screening Total UDCA dose (mg/kg): and completer populations Incidence of pruritus (%): -Increased APL 15.9 1) 50 4) OCA, 50mg levels for 6 % achieving 10% ALP reduction: 2) 47 (41) months Weight (kg): OCA- 87% 3) 85 -Mean ALP 72.8 Placebo- 14% 4) 80 All doses baseline value % achieving 20% ALP reduction: once daily for between 1.5- BMI, laboratory markers, and OCA- 69% OCA 10mg vs. placebo- 3 months (in 10x ULN PBC inclusion criteria also Placebo-8% p=NS combination reported Both outcomes: p<0.001 OCA 25mg vs. placebo- p<0.0003 with UDCA) Exclusion: OCA 50mg and 50mg- p<0.006 -Elevated AST No statistically significant ALP normalization: or ALT >5xULN differences between groups OCA- 7% Incidence of severe pruritus (%): -TB >2x Placebo- 0% 1) NR -Serum p=NR 2) 16 creatinine 3) 24 >1.5mg/dl For all algorithms, OCA had a higher response rate 4) 37 -Use of p=NR colchine, OCA results maintained in long-term extension: methotrexate, 210 U/L after 3 months Cholesterol reduction (%)*: azathioprine, 202 U/L after 12 months 1) NR or systematic Daily mean dose of ≤10mg restarted (range 3- 2) -3% corticosteroids 60mg) 3) -5% during 3 4) -13% ©Institute for Clinical and Economic Review, 2016 Page 69 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents months before Reductions in liver biochemistry: *Due to HDL lowering screening GGT- 48-63% -Patients with ALT- 21-35% concomitant AST- 9-17% Other AEs: liver diseases All SS vs. placebo 7 patients, 4% Bilirubin also significantly reduced in 25mg and Discontinuation: 50mg groups (not in 10mg group) 27 patients, 23 due to AEs 3 had elevated bilirubin 1 had elevated ALT/AST levels 15/27 (56% in the OCA 50mg group) Long-term extension: Pruritus- 87% (less severe) Discontinuation: 19 patients, 24% Due to pruritus: 10 patients, 13% Jones, 2014 Pooled 1) Placebo + Patients with Age Female White ∆ALP ∆Bilirubin ∆AST ∆ALT ∆GGT All TEAEs (n, % [events]): analysis of all UDCA (10) abnormal (years) (%) (%) 1) 3 mths -31 1.3 -1 -4 -90 1) 9, 90 1) 12 mths -20 -0.1 10 -5 -110 Abstract in OCA trials (2 2) 10mg OCA bilirubin 1) 50.2 100 90 2) 3 mths -212** -3.4 -25 -38 -425** 2) 13, 93 Hepatology Phase II trials, ± UDCA (14) pooled across 2) 54.8 79 93 2) 12 mths -140* -13** -33* -4 -375* 3) 21, 95 3) 3 mths -189** -3.3 -10 -32 -400** and POISE) 3) Total OCA 3 trials at 12 3) 53.6 77 95 4) 12 mths -100* -9** -15 -30 -300* Abstract and ± UDCA (22) weeks Weighted Means All absolute values Pruritus (n, % [events]): poster Age (years): 52.8 *p<0.05, **p<0.01 for comparing active treatment to 1) 3, 30 Female (%): 86.9 placebo using ANCOVA model with baseline value as 2) 12, 86 White (%): 92.2 covariate and fixed effects for treatment 3) 16, 73 Baseline ALP (U/L): 421.3 Patients with decreased bilirubin @ 12 months Fatigue (n, % [events]): Baseline GGT (U/L): 551.0 Placebo 23/70 1) 2, 20 Baseline ALT (U/L): 94.9 Titration OCA 34/64 2) 1, 7 Baseline AST (U/L): 83.9 10mg OCA 37/62 Baseline bilirubin (µmol/L): 3) 3, 14 28.8 1) 29.0 All serious TEAEs (n, % [events]): 2) 28.9 1) 1, 10 ©Institute for Clinical and Economic Review, 2016 Page 70 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents 3) 27.9 2) 1, 7 3) 3, 14 p=NR Other AEs also reported, including more granular data for TEAEs Kowdley Phase II DB DB DB DB DB DB: 2011 RCT 1) Placebo -Patients who Age (years): 55 Endpoint 1) 2) 3) Pruritus (%): LTSE: 2015 Multicenter (23) had not been Female: 84% ∆% ALP +0.4 -44.5* -37.6* 1) 30 Double- taking UDCA Caucasian: 95% ∆ U/L ALP +11.7 -233.5* -161.3* 2) 70 Abstracts in blinded (DB) 2) OCA, 10mg for at least 6 AP: 433 U/L ∆% GGT -3 -73* -65* 3) 94 Hepatology (20) months GGT: 527 U/L ∆% ALT -4 -37** -35** DB: 12 weeks -Mean ALP ALT: 81 U/L Compared to baseline: Discontinued due to pruritus (%): (747-201) f/u 3) OCA, 50mg baseline value AST: 68 U/L *p<0.0001; **p<0.01 1) 0 (16) between 1.5- 2) 15 LTSE: 4.5 10x ULN LTSE LTSE (all OCA patients); n=19 3) 38 years f/u LTSE: 28 Age (years): 60 Endpoint Mean change P value enrolled and LTSE Female: 85% ∆ U/L ALP -244 0.0034 LTSE 19 43% taking All patients ∆ U/L GGT -317 0.0019 AEs: 8 patients (3 pruritus-related); completed ≤10mg OCA ALP: 435 U/L ∆ U/L ALT -37 0.0010 20 events f/u (8 added GGT: 455 U/L ∆ U/L AST -22 0.0043 UDCA) ALT: 68.21U/L 7 patients discontinued ∆ µmol/L -2 0.1314 AST: 60 U/L Bilirubin Bilirubin: 13 µmol/L LTSE (OCA monotherapy); n=11 Monotherapy: Endpoint Mean change P value ALP: 435 U/L GGT: 455 U/L ∆ U/L ALP -182 0.0105 ALT: 68.21U/L ∆ U/L GGT -235 0.0188 AST: 60 U/L ∆ U/L ALT -32 0.0049 Bilirubin: 13 µmol/L ∆ U/L AST -17 0.0205 ∆ µmol/L -1 0.7990 Bilirubin Kowdley, Pooled Reported Reported Reported elsewhere ALP Change from BL (end of DB study): Reported elsewhere 2015 analysis of all elsewhere elsewhere 201 Study- OCA trials (2 Placebo (n=23): -19 Abstract in Phase II trials, OCA, 10mg (n=20): -251 ©Institute for Clinical and Economic Review, 2016 Page 71 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Gastro- and POISE) 202 Study- enterology Placebo (n=38): 12 OCA, 10mg (n=38): -58 POISE (301) Study- Placebo (n=73): -0.06 OCA, 5-10mg (n=70): -98 OCA, 10mg (n=23): -110 Pooled (all 3 trials): Placebo (n=134): -26 OCA (all groups, n=306): -128 p<0.001 for OCA vs. placebo % of patients achieving the POISE Composite Endpoint: normal bilirubin and ALP reduction <1.67 ULN and ≥15% reduction 201 Study- Placebo (n=23): 4 OCA, 10mg (n=20): 40 p=0.0026 for OCA vs. placebo 202 Study- Placebo (n=38): 8 OCA, 10mg (n=38): 42 p=0.0002 for OCA vs. placebo Pooled (all 3 trials, including POISE)- Placebo (n=134): 8 OCA (all groups, n=306): 45 p<0.0001 for OCA vs. placebo Luketic, Phase III 1) Placebo Patients on a Age (years): 55.8 POISE Paris I Paris II AE incidence by group (%): 2014 RCT (73) stable dose of Female (%): 91 Tx 1 yr Baseline 1 yr Baseline 1 yr 1) 90 Multicenter UDCA with Caucasian (%): 94 1) 10 53 48 0 4 2) 89 Abstract in Double- 2) OCA, 5- ALP Median UDCA dose: 15.4mg 2) 46* 49 79*** 0 27** 3) 86 Hepatology blinded (DB) 10mg (70) ≥1.67xULN or 3) 47* 52 70*** 0 26** p=NR ©Institute for Clinical and Economic Review, 2016 Page 72 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents bilirubin *p<0.0001, **p<0.0002, ***p<0.02 (POISE) 3) OCA, 10mg <2xULN POISE Endpoint = normal bilirubin and ALP Although most AEs were related to (73) reduction <1.67 ULN and ≥15% reduction pruritus, <6% discontinued due to Paris I Endpoint = ALP ≤3x ULN, AST≤3x ULN, and pruritus 1 year f/u normal bilirubin Paris II Endpoint = ALP ≤1.5x ULN, AST≤3xULN Overall SAEs in 22 (10%) of patients 91% completed No SS differences for Rotterdam criteria across all See Beuers for more detailed info severity of illness (likely due to high % of patients on harms associated with pruritus with normal bilirubin) Malecha, Secondary Reported Reported Reported elsewhere Univariate results: Reported elsewhere 2014 analysis to elsewhere elsewhere Parameter Measure Change P-value identify pre- GGT VAS 0.039 0.0008 Abstract in treatment 5D Itch 0.008 0.0005 Gastro- characteristics PBC40 Itch 0.0003 0.0136 Direct bilirubin VAS 1.309 0.0278 enterology which may be 5D Itch 0.316 0.0051 associated PBC40 Itch 0.192 0.0072 (POISE) with ALP VAS 0.087 0.0022 treatment- 5D Itch 0.014 0.0164 induced PBC40 Itch 0.007 0.0576* pruritus *Not statistically signifncant Data for multivariate results not reported Mayne, 2014 Pooled 1) 1-2x ULN Patients Age Female White ∆ALP ∆Bilirubin ∆AST ∆ALT ∆GGT Reported elsewhere analysis of all (39) receiving the (years) (%) (%) (%) (%): (%): (%): (%): 1) 57.5 77 95 1) -23* -4 -18* -30* -60* Abstract in OCA trials (2 2) >2-3x ULN 10mg dose 2) 55.0 95 95 2) -29* -5 -19* -32* -56* Hepatology Phase II trials, (44) across 3 trials 3) 54.1 82 94 and POISE) 3) >3-4x ULN at 12 weeks 3) -44* -6 -25* -46* -66* 4) 53.1 100 100 4) -45* -17** -32* -46* -62* Abstract and (17) stratified by Weighted Means poster 4) >4x ULN ALP *p<0.0001 from baseline; **p<0.05 from baseline Age (years): 55.5 (14) Female (%): 87.5 For each incremental increase 1x ULN increase in White (%): 95.5 baselines ALP, there was an incremental 3.5% Baseline ALP (U/L): 314.7 reduction in ALP in patients treated with 10mg (p=0.0046). Mayo, 2016 See Luketic See Luketic See Luketic See Luketic Reported elsewhere Patients reporting at least 1 TEAE (n, %): Abstract in Analysis to 1) 28, 38 ©Institute for Clinical and Economic Review, 2016 Page 73 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Gastro- study the 2) 39, 56 enterology effect of an 3) 51, 70 OCA titration Abstract and group on Median time to first onset of poster pruritus pruritus (days): 1) 50.5 (POISE) 2) 24.0 3) 9.0 Results also reported for those experiencing severe pruritus Discontinuations due to pruritus: 1) 0 2) 1, 1 3) 7, 10 Pruritus events by maximum severity (n, %): Mild 1) 16, 22 2) 11, 16 3) 15, 21 Moderate 1) 7, 10 2) 15, 21 3) 19, 26 Severe 1) 5, 7 2) 13, 19 3) 17, 23 Also reports % of patients receiving concomitant medication Nevens, See Luketic See Luketic See Luketic See Luketic %∆ ALP Bili GGT ALT HDL mean reduction (%): 2014 from U/L µmoL U/L U/L 1) NR baseline 2) 16 ©Institute for Clinical and Economic Review, 2016 Page 74 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Abstract in 1) -4.8 19.5 0.8 -4.7 3) 26 Hepatology 2) -33.0** 1.2* -50.3** -35.5** p=NR 3) -39.1** -0.2* -63.7** -41.7** (POISE) OCA vs. placebo: *p<0.005, **p<0.0001 Also see Luketic and Beuers for additional harms outcomes Absolute values at baseline and 12 mo also reported Pares, 2015 See Luketic See Luketic See Luketic See Luketic BMD Placebo OCA, 5- OCA, Reported elsewhere ∆ (SD) 10mg 10mg Abstract in J Analysis to 122 had ∆ (SD) ∆ (SD) Hepatology study the DEXA scans Lumbar -0.01 -0.01 -0.01 effect of OCA at baseline L2-L4 (0.01) (0.01) (0.01) (POISE) vs. placebo on and at 12 mo (g/cm²) bone density Lumbar T- -0.26 -0.01 -0.09 score (0.14) (0.14) (0.14) (g/cm²) Femoral 0.80 -0.01 -0.04 Neck (0.12) (0.01) (0.01) Femoral -0.33* -0.06ⱡ -0.07ⱡ* T-Score (0.11) (0.11) *(0.11) *p<0.05, OCA vs. placebo ⱡp<0.03 end of DB vs. DL Baseline and study end scores also reported Pencek, Pooled 1) Placebo Reported Reported elsewhere Age <65 ≥65 <50 ≥50 Incidence of pruritus (%): 2014 analysis of all (134) elsewhere BL ALP (U/L) 334 315 338 318 Age OCA trials (2 2) OCA End of DB -133* -109* -120* -120* <65: 63 Abstract in Phase II trials, ≤10mg (201) ∆ALP (U/L) ≥65: 51 Hepatology and POISE) Responders 75, 17, 50, 42, n, % 47* 41** 42* 51* Age at diagnosis Outcomes for age (<65|≥65) and age at diagnosis <50: 62 (<50|≥50) ≥50: 57 *p<0.0001, **p=0.0004 for OCA vs. placebo using ANCOVA model Sex Male: 57 Sex Male Female Female: 61 BL ALP 294 334 End of DB ∆ALP -111** -136* ©Institute for Clinical and Economic Review, 2016 Page 75 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Responders 9, 43ⱡ 83, 46* n, % Outcomes for sex *p<0.0001, **p=0.0009; ⱡp= 0.0131 for OCA vs. placebo using ANCOVA model Peters, 2014 See Luketic See Luketic See Luketic See Luketic Reported elsewhere Overall AE incidence rate (DB, %): 2) 56 Abstract in Long-term 3) 68 Hepatology safety extension Overall AE incidence rate (LTSE, %): (POISE) trial 2) 15 outcomes up 3) 21 to 24 months (n=193) Discontinuation due to pruritus (DB, %): 1) 0 2) 1 3) 10 Discontinuation due to pruritus (LTSE, %): 3% overall LDL remained comparable to baseline; HDL lowering remained stable and unchanged Hepatic disorders (DB trial, n): 0 Hepatic disorders (LTSE trial, n): 2 Overall SAEs incidence (%): 1) NR 2) 7 3) 3 1 death occurred due to cardiac failure ©Institute for Clinical and Economic Review, 2016 Page 76 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Trauner, See Luketic See Luketic See Luketic See Luketic Treatment Placebo OCA OCA Pruritus incidence decreased in the 2015 group (UDCA 5-10 mg 10 mg 6 months following the DB phase 18-month along) from 56-68% to 19-36% Abstract in follow up of ALP U/L 310.3 314.7 307.7 Gastro- POISE ∆ 12 mths -12.1 -106.1ⱡ -122.0ⱡ 1 patient died from sepsis enterology efficacy and ∆ 18 mths -97.8ⱡ -111.3ⱡ -106.8ⱡ secondary to endocarditis after a safety Bilirubin 11.0 10.6 10.9 prosthetic aprtic valve replacement (POISE) µmol/L ∆ 12 mths 1.5* -0.6 -1.2 ∆ 18 mths 1.9 -0.3 -1.3 GGT U/L 329.4 267.0 277.0 ∆ 12 mths -16.0 -149.4ⱡ -188.3ⱡ ∆ 18 mths -176.7◊ -161.7ⱡ -143.3ⱡ ALT U/L 55.7 65.4 57.0 ∆ 12 mths -5.3 23.8ⱡ -25.0ⱡ ∆ 18 mths -22.1ⱡ 25.8ⱡ -17.5* AST U/L 47.3 54.1 51.1 ∆ 12 mths 1.9 -12.9ⱡ -14.3 ∆ 18 mths -10.4ⱡ -13.1ⱡⱡ -12.7◊ *p<0.05; ◊p<0.01; ⱡp<0.0001 ©Institute for Clinical and Economic Review, 2016 Page 77 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix C. Comparative Value Supplemental Information Table C1. Undiscounted Budget Impact Cost per Patient from 1 to 5 Years: Payer Perspective OCA + UDCA UDCA Treatment Costs Non-Treatment Costs Treatment Costs Non-Treatment Costs 1 year $68,574 $1,241 $3,441 $1,240 2 years $101,064 $2,719 $6,525 $2,934 3 years $133,204 $4,431 $9,518 $5,100 4 years $164,971 $6,364 $12,415 $7,704 5 years $196,361 $8,488 $15,214 $10,672 ©Institute for Clinical and Economic Review, 2016 Page 78 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix D. Clinical Guidelines The American Association for the Study of Liver Diseases (2009) http://www.aasld.org/sites/default/files/guideline_documents/PrimaryBillaryCirrhosis2009.pdf The AASLD recommends a daily 13-15mg/kg dose of UDCA for treatment of PBC. UDCA treatment should be continued indefinitely, with liver tests every three to six months. In addition, the AASLD notes that additional medications may be needed to aid in management of symptoms of PBC, including use of bile acid sequestrants for treatment of pruritus. For pruritis that does not respond to treatment with bile acid sequestrants, patients can try rifampicin, oral opiate antagonists, or Sertaline. European Association for the Study of the Liver (2009) http://www.easl.eu/medias/cpg/issue2/Report.pdf Patients with PBC should be treated on a long-term basis with UDCA dosed at 13-15mg/kg per day. Biochemical response should be assessed after one year. If biochemical response is not optimal, there is no consensus on the next treatment options. One option may be adding budesonide to UDCA in patients without cirrhosis. Liver transplant should be considered for patients with advanced disease indicated by bilirubin over 6mg/dL or decompensated cirrhosis with an unacceptable quality of life or prognosis of less than one year of life without transplant. ©Institute for Clinical and Economic Review, 2016 Page 79 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix E. Previous Systematic Reviews and Technology Assessments Because obeticholic acid was only very recently approved, we were not able to identify any previous HTAs or systematic reviews on OCA for PBC. However, NICE has set forth plans of an appraisal on OCA for PBC, with expected publication in February 2017; more details on this report can be found here. ©Institute for Clinical and Economic Review, 2016 Page 80 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix F. Ongoing Studies Estimated Title/ Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Date A Phase 3 Double-Blind, Phase 3 double-  Obeticholic Acid N=217 Primary June 2018 Placebo Controlled Trial blind RCT (5-10 mg and Inclusion criteria  Composite and Long-term Safety 10mg doses)  Definite or probably PBC diagnosis endpoint of ALP Extension of Obeticholic After completion  Placebo  At least one of the following: ALP ≥1.67xULN, TB <1.67ULN, ≥15% Acid in Patients with >ULN but <2xULN decrease in ALP, of the 1-year Primary Biliary Cirrhosis  Age ≥18 years and normal TB double-blind  Taking UDCA ≥12 months (stable dose ≥3 months), or (POISE) study, subjects unable to tolerate UDCA prior to Day 0 Secondary will be offered the  Females must be postmenopausal, surgically sterile,  ALP response Sponsor: Intercept opportunity to or if premenopausal, be prepared to take ≥1 rates of 10, 20 Pharmaceuticals enter an open- contraceptive and 40% change label LTSE trial  ALP ≤3xULN and Exclusion criteria ≤2xULN  History of presence of other concomitant liver  ALP ≤1.5xULN diseases and aspartate  Presence of complications of PBC or clinically aminotranserase significant hepatic decompensation ≤1.5xULN and  Patients with a history of severe pruritus or prior normal TB systematic treatment for pruritus  Bilirubin and albumin ©Institute for Clinical and Economic Review, 2016 Page 81 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Estimated Title/ Trial Sponsor Study Design Comparators Patient Population Primary Outcomes Completion Date A Phase 3b Double-Blind, Phase 3b double-  Obeticholic Acid N=350 Primary April 2023 Randomized, Placebo- blind RCT (5-10 mg dose) Inclusion criteria  Composite Controlled Multicenter  Placebo  Definite or probably PBC diagnosis endpoint of any Study Evaluating the  A mean TB of >ULN and ≤3xULN of: death, liver Effect of Obeticholic Acid  Age ≥18 years transplant, MELD on Clinical Outcomes in  Taking UDCA ≥12 months (stable dose ≥3 months), or score ≥15, Patients with Primary unable to tolerate UDCA prior to Day 0 uncontrolled Biliary Cirrhosis  Females must be postmenopausal, surgically sterile, ascites, HCC, or if premenopausal, be prepared to take ≥1 hospitalization (COBALT) contraceptive for new onset or reoccurrence of Sponsor: Intercept Exclusion criteria variceal bleed, Pharmaceuticals  History of presence of other concomitant liver encephalopathy, diseases spontaneous  Presence of complications of PBC or clinically bacterial significant hepatic decompensation problems  Mean TB <3xULN Secondary  First occurrence of any one of the above- mentioned for the primary endpoint  Changes from baseline on liver biomarkers Source: www.ClinicalTrials.gov (NOTE: studies listed on site include both clinical trials and observational studies) ©Institute for Clinical and Economic Review, 2016 Page 82 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix G. Policy Roundtable Conflict of Interest Disclosures Conflict of Interest Disclosure Language: Do you have a relationship with a health care company (life sciences manufacturer, health plan, health plan association, life sciences manufacturer association, etc.) that falls under any of the following categories? A relationship extends to immediate family members and/or any entity in which you have an interest. Please mark all that apply within the last 12 months. Multiple selections are permitted. Kathleen Corey, MD, MPH, MMSc √ Manufacturer support of research in the clinical Director, MGH Fatty Liver Clinic area of this meeting in which you are participating Massachusetts General Hospital Gastrointestinal Unit Please describe: I am a principle investigator for trials including those from Novartis, Intercept, Galectin, Conatus and Tobira. Judith Donovan I have no conflicts of interest as defined in the above Patient with PBC categories Barbara Henry, RPh √ Receipt or potential receipt of anything of monetary Lead Clinical Pharmacy Specialist value, including but not limited to, salary or other Harvard Pilgrim Health Care payments for services such as consulting fees or honoraria in excess of $5,000 Please describe: Salary from a Health Plan; Employee of Health Plan Juan Carlos Lopez-Talavera, MD, √ Receipt or potential receipt of anything of monetary PhD value, including but not limited to, salary or other Senior Vice President, Global payments for services such as consulting fees or Medical Affairs honoraria in excess of $5,000 Intercept Pharmaceuticals √ Equity interests such as individual stocks, stock options or other ownership interests in excess of $10,000. Ownership of stock in a mutual fund over which an individual has no trading control does not count toward this item. √ Manufacturer support of research in the clinical ©Institute for Clinical and Economic Review, 2016 Page 83 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents area of this meeting in which you are participating √ Any other relationship that could reasonably be considered a financial conflict of interest Please describe: I am actively employed by Intercept Pharmaceuticals. Daniel Pratt, MD I have no conflicts of interest as defined in the above Clinical Director, Liver categories Transplantation; Director of the Autoimmune Cholestatic Liver Center Massachusetts General Hospital Thomas Simpatico, MD √ Receipt or potential receipt of anything of monetary Chief Medical Officer value, including but not limited to, salary or other Vermont Department of Health payments for services such as consulting fees or Access honoraria in excess of $5,000 Please describe: Salary from a Public Health Plan; Employee of Public Health Plan ©Institute for Clinical and Economic Review, 2016 Page 84 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix H. Public Comments This section includes summaries of the public comments prepared for the New England CEPAC Public Meeting on July 15, 2016 in Portland, ME. These summaries were prepared by those who delivered the public comments at the meeting. Conflict of interest disclosures are included at the bottom of each statement. (1) DONNA R CRYER, JD, PRESIDENT & CEO, GLOBAL LIVER INSTITUTE SUMMARY REMARKS FOR REPORT I am Donna Cryer, CEO of the Global Liver Institute, a patient who has lived with biliary disease and liver transplant for over 20 years. I serve on the ABIM GI Board, Executive Committee of the Liver Forum, WHO/PAHO Technical Expert Panel, and FDA patient representative, in fact serving on the OCA advisory committee. I am not hard to find. Value of ICER: Illegitimate You were not mandated by Congress, clinicians don’t look to you instead of their medical societies, patients haven’t delegated our voice or confidence to determine affordability. Due to tortured terminology, muddled methodology, and lack of transparency, even members of your advisory board have criticized you. Purpose of ICER: Payer Voice You’ve admitted that timing is to give payers leverage to negotiate with manufacturers. Be honest about your purpose/effect– justify access restrictions, be self-appointed “pricing police” and provide a chilling effect to innovation which we certainly cannot afford in liver disease. Stakeholder Engagement: Substantially None There are no patients on the voting panel. Scheduling the policy roundtable after the vote is also anti-informed engagement. Examples of patient engagement -- UNOS, FDA, PCORI, National Health Council and Faster Cures. PBC Comments OCA for PBC falls below your budget impact threshold. Question 1 invalid unless panel purports to have superior clinical expertise than the FDA. Question 2 usurps the personalized judgement and shared decisionmaking of physicians and patients. NASH Comments There are no FDA-approved therapies for NASH. Your questions 3 and 4 are premature and should be removed. Conflict of Interest Disclosure. My husband and I serve on several boards of nonprofit organizations that receive funding from healthcare companies as defined above. Neither of us receive any funds directly or indirectly for work related to PBC or NASH. The Global Liver Institute has received funding in the preceding 12 months from Abbvie, Amgen, Genentech, and Gilead. CryerHealth has received money from Amgen and Esperion. (2) TERRY WILCOX, EXECUTIVE DIRECTOR, PATIENTS RISING NEW ENGLAND CEPAC PUBLIC COMMENT SUMMARY ©Institute for Clinical and Economic Review, 2016 Page 85 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Patients Rising believes value frameworks like the one ICER is trying to develop are by any objective measure meant to control access first. Not advance treatment first. We gathered in Portland, Maine at the New England CEPAC to discuss two liver dieases PBC and NASH. One disease is being treated effectively by current therapy for many patients. The other has no effective treatment for the most vulnerable cases. NASH is set to pass Hepatitis C as the leading cause of liver transplants by 2020. By 2025 more than 25 million Americans will be living with the disease, 10% of those being children; mostly Hispanic. Tina Dooley is a PBC patient from Houston. She meets the criteria for OCA as she is only a partial responder to the standard of care. Patients like Tina deserve fair access to the best of medical innovation that is right for them. If we as a society focus on the precision capabilities of medical innovation, figure out how to monitor the efficacy of these treatments in real world settings and ultimately create value structures that look at costs in totality across all sectors of healthcare – now that is a care model Patients Rising can stand with. We can do better than the value structure being used in a country with the poorest overall cancer survival rates in Western Europe. We can win the economic battle without sacrificing the access to new therapies patients in this country need. Conflict of Interest Disclosure. Patients Rising has received funding from Amgen, Celgene and PhRMA for specific projects: Voices of Value, Right Patient, Right Treatment, Right Now and Check Yourself to Protect Yourself screenings. We have no conflict of interest relationship with Intercept Pharmaceuticals. (3) JUAN CARLOS LOPEZ-TALAVERA, SVP, HEAD GLOBAL MEDICAL AFFAIRS ICER COMMENTS SUMMARY Intercept is dedicated to developing innovative treatments for progressive, non-viral liver diseases with high unmet need, and we are committed to pricing our medicines based on the value they provide to patients and the healthcare system. Since ICER announced its intended review of Ocaliva (obeticholic acid) in PBC and NASH, we have had a constructive dialogue with the organization regarding their proposed economic models, and provided our feedback on its draft report. PBC is a rare, autoimmune liver disease that progresses to severe life-threatening complications such as liver cirrhosis, liver failure, and ultimately death from liver failure or hepatocellular carcinoma in the absence of liver transplantation. Despite revisions in the final draft report, ICER’s PBC model still does not accurately reflect the risks associated with disease progression, costs of managing complications, or the impact of PBC on patient quality of life. Thus, ICER’s resulting analysis of Ocaliva significantly undervalues the benefit it provides to patients. While Intercept has completed Phase 3 development of Ocaliva in PBC, it is still being evaluated for NASH in the Phase 3 REGENERATE trial. It is inappropriate to conduct a value analysis of a drug candidate prior to having data ©Institute for Clinical and Economic Review, 2016 Page 86 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents from a Phase 3 program. We were pleased to see that ICER’s final draft report acknowledged that the introduction of a value-based price benchmark for OCA in NASH is premature at this time. Finally, while we appreciate ICER’s effort, the evaluation of Ocaliva was compromised by a lack of patient perspective, as reflected in the critical public comments from patient advocacy groups. Conflict of Interest Disclosure. I am actively employed by Intercept Pharmaceuticals. ©Institute for Clinical and Economic Review, 2016 Page 87 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents Appendix I. Coverage Policy at Harvard Pilgrim Health Care HARVARD PILGRIM HEALTH CARE RECOMMENDED MEDICATION REQUEST GUIDELINES Generic Brand HICL GCN Exception/Other OBETICHOLIC ACID OCALIVA 43438 GUIDELINES FOR USE 1. Is the request for a member with the diagnosis of primary biliary cholangitis (PBC)? If yes, continue to #2. If no, do not approve. DENIAL TEXT: Per your health plan's Ocaliva guideline, this medication is only covered for the diagnosis of primary biliary cholangitis. Your physician did not indicate that you have this condition, and therefore your request was not approved. 2. Has the patient tried therapy with ursodiol for a duration of at least one year, or does the patient have a contraindication to ursodiol (ursodeoxycholic acid; UDCA; generic Actigall)? If yes, continue to #3. If no, do not approve. Please use status code #238 and the provided denial text. DENIAL TEXT: Per your health plan's Ocaliva guideline, a trial with ursodiol for at least one year is required prior to approving coverage for Ocalia. Your physician did not indicate that you have been treated with ursodiol for at least one year, and therefore your request was not approved. 3. Please approve for 12 months by HICL as follows. Please use status code #050. For a member on an Open Formulary, approve for 12 months by HICL. (A quantity of one tablet per day is hard-coded) APPROVAL TEXT: Your request for Ocaliva has been approved for one tablet per day for 12 months. For a member on a Closed Formulary (determined by the word 'closed' in the benefit description), approve for 12 months by HICL at the appropriate tier listed in the table below. Please enter a quantity of one tablet per day. Benefit Drug Description Exception Tier 3 Tier All non-formulary 3 4 Tier All non-formulary 4 5 Tier All non-formulary 5 APPROVAL TEXT: Your request for Ocaliva has been approved at your highest cost- share tier for one tablet ©Institute for Clinical and Economic Review, 2016 Page 88 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents per day for 12 months. Refer to your Harvard Pilgrim ID card for the amount you pay for drugs on that tier. RATIONALE: To ensure the appropriate and cost-effective use of Ocaliva. FDA APPROVED INDICATIONS Ocaliva is indicated for the treatment of primary biliary cholangitis (PBC) in combination with ursodeoxycholic acid (UDCA) in adults with an inadequate response to UDCA, or as monotherapy in adults unable to tolerate UDCA. REFERENCES HARVARD PILGRIM HEALTH CARE RECOMMENDED MEDICATION REQUEST GUIDELINES Intercept Pharmaceuticals, Inc. Ocaliva package insert. New York, NY. May 2016. Created:Effective: 6/13/16 Client Approval: P&T Approval: ©Institute for Clinical and Economic Review, 2016 Page 89 Draft Evidence Report – OCA for the Treatment of PBC Return to Table of Contents