Cost-Effectiveness Analysis of Screening for Hepatitis B Virus Infection in Patients With Solid Tumors Before Initiating Chemotherapy

Journal Pre-proof Cost-Effectiveness Analysis of Screening for Hepatitis B Virus Infection in Patients with Solid Tumors before Initiating Chemotherapy Gauree G. Konijeti, Sirisha Grandhe, Monica Konerman, Jill Lane, Mark Shrime, Siddharth Singh, Rohit Loomba

PII: DOI: Reference:

S1542-3565(19)31241-8 https://doi.org/10.1016/j.cgh.2019.10.039 YJCGH 56826

To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 25 October 2019 Please cite this article as: Konijeti GG, Grandhe S, Konerman M, Lane J, Shrime M, Singh S, Loomba R, Cost-Effectiveness Analysis of Screening for Hepatitis B Virus Infection in Patients with Solid Tumors before Initiating Chemotherapy, Clinical Gastroenterology and Hepatology (2019), doi: https:// doi.org/10.1016/j.cgh.2019.10.039. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 by the AGA Institute

Konijeti et al. Cost-Effectiveness Analysis of Screening for Hepatitis B Virus Infection in Patients with Solid Tumors before Initiating Chemotherapy Gauree G. Konijeti1, Sirisha Grandhe2, Monica Konerman3, Jill Lane1, Mark Shrime4, Siddharth Singh5, Rohit Loomba5 1

Division of Gastroenterology, Scripps Clinic, La Jolla, CA, United States. Division of Gastroenterology, University of California, Davis, Sacramento, CA, United States. 3 Division of Gastroenterology, University of Michigan, Ann Arbor, MI, United States. 4 Program in Global Surgery and Social Change, Harvard Medical School, Boston, MA, United States. 5 Division of Gastroenterology, University of California, San Diego, La Jolla, CA, United States. 2

Manuscript Number: CGH 18-01861 Word Count: 2936 Financial Support: None Potential Conflicts of Interest: Dr. Loomba has received research funding and served on the advisory board for Gilead and Bristol-Meyers Squibb. Otherwise the other authors report no potential conflicts of interest. Konijeti: study concept and design, analysis and interpretation of data, drafting of the manuscript, study supervision Grandhe: study concept and design, drafting of the manuscript Konerman: critical revision of the manuscript for important intellectual content Lane: critical revision of the manuscript for important intellectual content Shrime: study concept and design, statistical analysis Singh: study concept and design, critical revision of the manuscript for important intellectual content Loomba: study concept and design, critical revision of the manuscript for important intellectual content, study supervision Corresponding Author: Rohit Loomba, MD, MHSc Professor of Medicine, Director of Hepatology Vice Chief, Division of Gastroenterology Director, NAFLD Research Center Adjunct Professor, Division of Epidemiology University of California at San Diego ACTRI 1W202, 9500 Gilman Drive La Jolla, CA, 92037-0887 Ph: 858-246-2201 Email: [email protected]

1

Konijeti et al. ABSTRACT Background & Aims: Patients with solid tumors who undergo chemotherapy have an increased risk of hepatitis B virus (HBV) reactivation, but a low proportion of these patients are screened for HBV infection and guidelines make conflicting recommendations. Further, the cost-effectiveness of newer treatments for HBV prophylaxis has not been examined for this population. We aimed to analyze the costeffectiveness of HBV screening before chemotherapy for patients with solid tumors.

Methods: We compared 3 HBV screening strategies (screen all, screen only high-risk patients, or screen none) using a Markov model of a population of adults in the United States who initiated chemotherapy for a solid tumor. We modeled use of entecavir prophylaxis for HB surface antigen (HBsAg)-positive patients and surveillance for HBsAg-negative patients who are positive for HBV core antibody. The Markov cycle length was 1 year, with model simulation for up to 5 years.

Results: The screen all strategy was the most cost effective, with an incremental cost-effectiveness ratio of $42,761 compared to screening only high-risk patients. The screen none strategy was less effective and less costly than screening all patients or only high-risk patients. The screen-all strategy was the most cost effective for all estimates of prevalence of HBsAg-positive patients and estimates of HBV reactivation in HBsAg-positive patients. Screening only high-risk patients was the most cost-effective strategy when more than 25% of high-risk patients were screened for HBV infection.

Conclusions: In a Markov model analysis, we found screening all patients with solid tumors for HBV infection before chemotherapy to be the most cost-effective strategy. Guidelines should consider recommending HBV tests for patients initiating chemotherapy.

KEY WORDS: ICER; cancer treatment; detection; viral infection; immune suppression

2

Konijeti et al. INTRODUCTION

According to the Centers for Disease Control and Prevention, over 7 million people in the United States were diagnosed with a solid tumor malignancy (excluding hematopoietic malignancy) between 201120151. Among patients with newly diagnosed cancer, rates of prior hepatitis B virus (HBV) infection range from 5.6% to 7.4%, and chronic HBV infection from 0.4% to 1.0%2. Many of these infected individuals are unaware that they have chronic hepatitis B and are therefore at high risk of HBV reactivation when they start immunosuppressive therapy or cancer chemotherapy. HBV reactivation occurs in multiple stages, including an initial rise in HBV viremia followed by an acute rise in serum transaminases leading to acute hepatitis3. However, in a subset of cases, reactivation of HBV can be extremely severe and can potentially lead to persistent liver injury with fulminant hepatic failure and even death3.

The 2008 U.S. Centers for Disease Control (CDC) recommendations, as well as 2009 American Association for the Study of Liver Diseases (AASLD) guidelines, include routine screening for HBV by testing for HBV surface antigen (HBsAg) and HBV core antibody (HBcAb) for all patients at risk for HBV reactivation, including those undergoing chemotherapy4,5. This recommendation is also included in the 2017 American College of Physicians (ACP) Clinical Guideline6. However, the most recent 2015 American Society of Clinical Oncology only recommends HBV screening prior to highly immunosuppressive therapy (anti-CD20 chemotherapy and hematopoietic stem cell transplantation)7, despite an increased risk of HBV reactivation among most solid tumor chemotherapies8. As a result, HBV screening rates prior to solid tumor chemotherapy are much lower (8-10%) than for hematologic malignancies (37-56%)9,10. Additionally, prophylactic therapy is strongly recommended for patients at high or medium risk of HBV reactivation, whereas those with a low risk of reactivation have the option to be closely monitored or receive HBV prophylaxis3. Given the variation in guidelines and

3

Konijeti et al. recommendations, only 13 to 19% of oncologists in surveys from the United States and Australia have adopted universal HBV screening prior to chemotherapy11-13.

Guidelines for HBV screening in all patients prior to receiving solid tumor chemotherapy are conflicting given insufficient evidence14,15. The cost-effectiveness of a screening strategy depends both on costs associated with screening as well as health and cost benefits through preventing complications via appropriate screening and treatment. Results from prior cost-effectiveness analyses of HBV screening prior to chemotherapy in patients with solid tumors have been inconsistent. One analysis16 in breast cancer supported universal HBV screening, compared to screening high-risk or no screening. In contrast, one analysis14 in breast cancer and non-small-cell lung cancer, and another analysis17 in sarcoma and gastrointestinal stromal tumors, both favored no screening compared to universal screening. These analyses vary in model design and thresholds, population, and incorporate use of lamivudine for prophylaxis. The cost-effectiveness of newer treatments first-line for HBV prophylaxis in the setting of immunosuppression has not been examined. The aim of this study was to analyze the cost-effectiveness of HBV screening prior to solid tumor chemotherapy to inform clinical practice.

METHODS Model Design Case scenario We created a decision-analytic state-transition (Markov) model for a hypothetical population of adults (age range, 20-80 years, median 50 years) in the United States newly initiating chemotherapy for solid tumor, comparing 3 screening strategies for HBV infection: screen all, high-risk only, or none (Supplementary Figure 1).

Screening tests and primary analysis

4

Konijeti et al. Screening tests for patients with solid tumor included measurements of HBV surface antigen (HBsAg), HBV surface antibody (HBsAb), and HBV core antibody (HBcAb)6. Health states in the model for the patient with solid tumor included chronic, prior, or no HBV infection (Figure 1). Those with chronic or prior HBV infection were at risk for HBV reactivation with non-severe hepatitis, which could progress to severe hepatitis and acute liver failure. Possible causes of death included acute liver failure or non-HBV related death in the context of solid tumor. The simulation followed patients for up to 5 years, with a Markov cycle length of 1 year. Table 1 lists transition annual probabilities, costs, and utilities. Our model was created in conformity with the CHEERS checklist17. For our primary analysis, we modeled use of entecavir for prophylaxis in HBsAg-positive patients, with use of tenofovir disoproxil fumarate (TDF) either for those with HBV reactivation despite entecavir prophylaxis or those without prophylaxis who progressed to severe hepatitis despite use of entecavir for non-severe hepatitis. We modeled surveillance for HBsAg-negative/HBcAb-positive patients. All analyses were performed using TreeAge Pro 2019 (Williamstown, MA).

HBV prevalence and chemotherapy specific modelling data Patients considered to be high-risk for HBV infection were those with a history of injection drug use or those residing in the United States but born in endemic countries with HBsAg-positive prevalence of 8% or greater, whereas intermediate risk was considered to be 2-7%4,10,18-21. Based on data from populationbased cohorts in the United States, for our base-case scenario we modeled a 10% rate of HBV screening in high-risk patients9,10, and performed sensitivity analyses varying this rate. Estimated rates of HBV reactivation for solid tumor chemotherapy according to HBV infection status and antiviral prophylaxis are listed in Table 13,15,22,23. Most estimates for risks of HBV reactivation without prophylaxis during solid tumor chemotherapy involve cancer of the breast (0.05-0.68)22-33, gastrointestinal tract (0.06-0.50)22,28-29,3236

, lung (0.14-0.37)22,28-29,32-33,37, head and neck (0.17-0.29)29,33,38, or other (0.06-0.50)22,28-29,32-33 sites15.

Across these tumor subtypes, risk of HBV reactivation with antiviral prophylaxis is approximately 3.5%

5

Konijeti et al. (0.023-0.05)15. Median time to HBV reactivation with and without antiviral prophylaxis was estimated at 9 and 3 months, respectively (Supplementary Methods).

Though it would be ideal to include risks and costs related to reduced chemotherapy in the setting of HBV reactivation, we could not build this into the model due to limited estimates regarding survival following HBV reactivation according to tumor subtype in the setting of standard vs. reduced dose chemotherapy. Annual death rate due to non-HBV related causes was calculated using 5-year solid tumor cancer survival rates from the Surveillance, Epidemiology, and End Results (SEER) Program24.

Health Utility States and Costs Based on available data, utility estimates for patients with solid tumors during or after chemotherapy are presented in Table 116,25-30. Additional model assumptions, including average durations in given health states, are provided in the Supplementary Methods.

All costs were adjusted to 2018 US dollars using the consumer price index31. Costs for laboratory testing were taken from the Clinical Diagnostic Laboratory Fee Schedule form the Centers for Medicare and Medicaid Services (Table 1). Costs for HBV medications were taken from the lowest cost of average wholesale pricing of generic medications, with sensitivity analyses of drug pricing varied to include higher generic and/or brand costs. Initial and continuing costs of cancer care were averaged for the above solid tumor types from the National Cancer Institute Cancer Prevalence and Cost of Care Projections32. All costs and utilities were discounted at a rate of 3% per year.

Outcomes and Data Analysis Primary outcome The primary outcome of the Markov model using Monte Carlo simulation of 10,000 independent trials and probabilistic sensitivity analysis (100-fold distribution sampling of all variables) was the incremental 6

Konijeti et al. cost-effectiveness ratio (ICER), calculated as incremental cost divided by incremental effectiveness. Effectiveness was measured as quality adjusted life-years (QALY), calculated as the product of time and utility for a given health state. Willingness-to-pay (WTP) threshold was set at $50,000 per QALY, though results are also presented across a range of thresholds31.

Sensitivity analyses Probabilistic sensitivity analysis was performed in addition to Monte Carlo simulation to account for heterogeneity in the model assumptions. To compare differences between expected incremental effectiveness and costs between two screening strategies, we examined incremental cost-effectiveness scatter plots resulting from probabilistic sensitivity analyses. Uncertainty around model estimates is characterized by the 95% confidence ellipse, with the dashed line on the scatter plot indicating the WTP threshold of $50,000 per QALY. One-way sensitivity analyses using microsimulation of all variables were conducted to evaluate for individual factors affecting model outcomes (Supplementary Figure 2). Where ranges were not available from published literature, the range for model input varied between 25% below and above average values for probabilities, costs, and utilities32-35. Analysis of the model was done from the healthcare payer perspective.

RESULTS Model Analysis In our primary model comparing screen all, screen high-risk only, or screen none as HBV screening strategies prior to solid tumor chemotherapy, the screen all strategy was the most cost-effective at a WTP of $50,000 per QALY, with an ICER of $41,078 compared to the screen high-risk only strategy (Table 2). The screen none strategy was both less effective and less costly than screening high-risk only patients and everyone.

7

Konijeti et al. Probabilistic Sensitivity Analysis The acceptability curve (Figure 2), examining screening strategies across a range of WTP thresholds up to $50,000/QALY, indicates that the screen all strategy is more likely to be cost-effective than the screen high-risk only strategy at a WTP threshold greater than $40,275. At WTP thresholds between $16,203 and $40,275, the screen high-risk only strategy was more likely to be cost-effective, whereas at a WTP threshold less than $16,203, the screen none strategy was more likely to be cost-effective.

When comparing the incremental cost-effectiveness of two competing screening strategies, at a WTP threshold of $50,000 per QALY, the screen all approach was preferred in 54% of trials compared with the screen high-risk only approach, and in 63% of trials compared with the screen none approach (Supplementary Figure 3).

Sensitivity Analyses The screen all strategy remained the preferred strategy across all estimates of prevalence of HBsAg positivity among all or high-risk patients, as well as all estimates of prevalence of HBsAgnegative/HBcAb-positive status in high-risk patients. Among all patients, the screen all strategy was costeffective at HBsAg-negative/HBcAb-positive prevalence estimates exceeding 4% (Table 3), and below this threshold a screen high-risk only strategy was preferred. Among HBsAg-positive patients, the screen all strategy was preferred across all probabilities of HBV reactivation without prophylaxis. For the primary model analysis, we assumed 100% probability of patients receiving prophylaxis if they were identified as HBsAg-positive. All other factors unchanged, the screen all strategy was preferred when at least 13% of HBsAg-positive patients received antiviral prophylaxis; below this, the screen none strategy was preferred.

The model assumed that patients considered high risk were screened for HBV at a rate of 10%, consistent with current population estimates9,10. Our model suggests that, at this current estimate, a screen all 8

Konijeti et al. strategy is preferred. However, screening high-risk only patients became more cost-effective than screening everyone when rates of screening high-risk patients was greater than 25% (Figure 3, Table 3). The model also assumed that patients identified as HBsAg-negative/HBcAb-positive would not receive prophylaxis. Sensitivity analyses demonstrated that a screen all strategy remained robust across all ranges of prophylaxis for patients identified as HBsAg-negative/HBcAb-positive. A screen all strategy was also preferred across all ranges for utilities and costs.

DISCUSSION The risk of HBV reactivation in patients with chronic HBV infection receiving solid tumor chemotherapy has recently been shown to be similar to risks associated with other types of immunosuppressive therapy for which HBV screening and prophylaxis are currently recommended. Whether to screen all or only high-risk patients in countries with low HBV prevalence (<2%) remains controversial. Guidelines conflict with respect to conducting universal or only high-risk screening for patients receiving solid tumor chemotherapy3-5,7,11. Therefore, determination of a cost-effective strategy for screening in this population is an important research need and clinical priority. This study fills that gap in knowledge. By utilizing a Markov model, incorporating recently published estimates and modeling use of newer therapies for HBV prophylaxis, we demonstrate that an approach of screening all patients for HBV infection prior to solid tumor chemotherapy is the most cost-effective strategy for the management of this patient population.

Two prior studies examined the cost-effectiveness of HBV screening prior to solid tumor chemotherapy at the population level, incorporating published estimates. Day et al14 found that universal HBV screening, compared with no screening, was not cost-effective for a cohort of Australian patients beginning adjuvant chemotherapy for early breast cancer (ICER $88224/LY) or palliative chemotherapy for advanced nonsmall cell lung cancer (ICER $1,344,251/LY). This study modeled prophylaxis with entecavir if patients were HBsAg-positive, HBV DNA+ (31% cohort), and lamivudine for HBsAg-positive only (69%). In 2015, Wong et al14 performed a cost-effectiveness analysis of HBV screening strategies in patients 9

Konijeti et al. receiving chemotherapy for breast cancer, with the use of either lamivudine/tenofovir or entecavir as prophylaxis to prevent HBV reactivation. They concluded that the screen all strategy was the most costeffective strategy (ICER $47,808-$76,527/LY). Tan et al17 published a cost-effectiveness analysis of HBV screening for patients in Singapore starting chemotherapy for sarcomas or gastrointestinal stromal tumors, but utilized estimates from 485 patients rather than population estimates. In this study, they found that universal HBV screening was not cost-effective at a WTP of $100,000 Singapore dollars per QALY. Upon comparing our results with these prior decision analyses, differences in outcomes may be attributable to the use of more narrowly defined treatment population, incorporation of specific chemotherapy regimens as well as cessation, and variations in model design. Importantly, Day et al14 found that universal screening using HBsAg-positive only was cost-effective (ICER $30,126/LY) for the adjuvant cohort, near cost-effective for the pooled cohort ($51,201/LY) at a WTP threshold of $50,000/LY.

Our model analyses and sensitivity analyses suggest that our findings remain robust to a wide range of prevalence estimates for HBsAg-positive or HBsAg-/cAb+ in high-risk patients and HBsAg-positive in all patients. Only where the prevalence of HBsAg-negative/HBcAb-positive varied among all patients did the model identify differences in preferred screening strategy. However, this one-way sensitivity analysis assumes all other factors (e.g., probabilities, costs, and utilities) are the same, which is less likely, and must be taken into consideration when interpreting results. Additional assumptions of our model were that laboratory testing was 100% accurate in detecting HBV exposure, patients identified as HBsAg-positive were treated with and took antiviral prophylaxis with 100% compliance, and patients identified as HBsAg-/cAb+ were not given HBV prophylaxis. Our analyses remained robust to these assumptions.

Importantly, we modeled a 10% rate of HBV screening in high-risk patients, consistent with population based estimates9,10. Even at rates of HBV screening more than twice this value, a screen all strategy was preferred. However, if HBV screening rates in this population increased to 25% or greater, screening 10

Konijeti et al. high-risk only patients became the most cost-effective approach. In patients with hematologic malignancies and those undergoing bone marrow transplantation (BMT), the risk and severity of HBV reactivation is substantial. It is estimated that the risk of HBV reactivation is extremely high in this clinical scenario. Therefore, there is greater acceptance of routine screening prior to hematologic malignancies and BMT. However, the current guidelines in patients with solid-organ malignancies are not consistent, increasing the challenge for providers to identify patients at high risk for HBV. Recent data among patients with solid tumors suggests that early identification of HBV status before initiation of cancer chemotherapy, compared with after starting chemotherapy, is associated with significantly lower risks of hepatitis flare, liver failure, and death36. Our data would suggest that we should adopt a prevention-based strategy of screening all patients prior to solid tumor chemotherapy, particularly given dismal rates of HBV screening and knowledge gap that is resulting in high morbidity and mortality in this patient population.

In our model utilizing 2018 cost estimates, we modeled the use of tenofovir disoproxil fumarate (TDF) rather than tenofovir alafenamide (TAF), which was approved by the U.S. Food and Drug Administration in November 2016. Based on our model and sensitivity analyses showing use of TDF to be cost-effective at a broad range of drug costs, we would not expect that the use of TAF instead of TDF would alter the primary results. Further, we would not expect that use of TDF instead of entecavir for HBV prophylaxis or reactivation would alter the primary results, particularly given lower monthly cost of TDF (all other factors being equal).

Strengths of our study include a well-developed decision-analytic Markov model evaluating competing strategies based upon the baseline prevalence of HBV infection, risk of reactivation in a clinical scenario that mimics clinical practice and the linked outcomes related to HBV reactivation in patients with solid tumor receiving chemotherapy. Probabilistic sensitivity analyses and individual variable sensitivity analyses, utilizing microsimulation, further add to the robustness of this model. 11

Konijeti et al.

One limitation of our study includes lack of evaluation of the model by tumor subtype. We did account for recent estimates by pooling the results from individual studies. A recent systematic review and metaanalysis15 found a variable risk of HBV reactivation without prophylaxis ranging from 5% to 68%, which we included in our estimate. Though we did average initial cancer costs for the first year across various tumor subtypes, sensitivity analysis suggests that the screen all strategy remained preferred across a range of first year cancer costs. Based on ranges for probability estimates, costs and utilities, we feel that our model can be applied to a variety of solid tumors, particularly those from which the estimates were derived, including breast, gastrointestinal tract, lung, and head and neck. Compared to published studies, the outcome with our model assessing the risk of HBV reactivation across all solid tumors would likely not vary based on tumor subtype. Finally, we did not examine cost-effectiveness of HBV screening in settings where HBV burden is higher, particularly in resource-limited regions. Future studies should also be performed examining cost-effectiveness of HBV screening in these settings.

In conclusion, our cost-effectiveness analysis examining HBV screening strategies prior to solid tumor chemotherapy, where entecavir (or TDF) is used first-line for HBV prophylaxis, indicates that the screen all strategy is the most cost-effective approach. Based on our results, guidelines should be reexamined and consider recommending HBV testing prior to all solid tumor chemotherapy.

12

Konijeti et al. REFERENCES 1. U.S. Cancer Statistics Working Group. U.S. Cancer Statistics Data Visualizations Tool, based on November 2017 submission data (1999-2015): U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute; www.cdc.gov/cancer/dataviz, June 2018. 2. Ramsey SD, Unger JM, Baker LH, et al. Prevalence of Hepatitis B Virus, Hepatitis C Virus, and HIV Infection Among Patients With Newly Diagnosed Cancer From Academic and Community Oncology Practices. JAMA Oncol 2019. 3. Loomba R, Liang TJ. Hepatitis B Reactivation Associated With Immune Suppressive and Biological Modifier Therapies: Current Concepts, Management Strategies, and Future Directions. Gastroenterology 2017;152:1297-309. 4. Lok AS, McMahon BJ. Chronic hepatitis B: update 2009. Hepatology 2009;50:661-2. 5. Weinbaum CM, Williams I, Mast EE, et al. Recommendations for identification and public health management of persons with chronic hepatitis B virus infection. MMWR Recomm Rep 2008;57:1-20. 6. Abara WE, Qaseem A, Schillie S, et al. Hepatitis B Vaccination, Screening, and Linkage to Care: Best Practice Advice From the American College of Physicians and the Centers for Disease Control and Prevention. Ann Intern Med 2017;167:794-804. 7. Hwang JP, Somerfield MR, Alston-Johnson DE, et al. Hepatitis B Virus Screening for Patients With Cancer Before Therapy: American Society of Clinical Oncology Provisional Clinical Opinion Update. J Clin Oncol 2015;33:2212-20. 8. Voican CS, Mir O, Loulergue P, et al. Hepatitis B virus reactivation in patients with solid tumors receiving systemic anticancer treatment. Ann Oncol 2016;27:2172-84. 9. Wi CI, Loo NM, Larson JJ, et al. Low level of hepatitis B virus screening among patients receiving chemotherapy. Clin Gastroenterol Hepatol 2015;13:970-5; quiz e51. 10. Kwak YE, Stein SM, Lim JK. Practice Patterns in Hepatitis B Virus Screening Before Cancer Chemotherapy in a Major US Hospital Network. Dig Dis Sci 2018;63:61-71. 11. Day FL, Link E, Thursky K, Rischin D. Current hepatitis B screening practices and clinical experience of reactivation in patients undergoing chemotherapy for solid tumors: a nationwide survey of medical oncologists. J Oncol Pract 2011;7:141-7. 12. Khokhar OS, Farhadi A, McGrail L, Lewis JH. Oncologists and hepatitis B: a survey to determine current level of awareness and practice of antiviral prophylaxis to prevent reactivation. Chemotherapy 2009;55:69-75. 13. Tran TT, Rakoski MO, Martin P, Poordad F. Screening for hepatitis B in chemotherapy patients: survey of current oncology practices. Aliment Pharmacol Ther 2010;31:240-6. 14. Day FL, Karnon J, Rischin D. Cost-effectiveness of universal hepatitis B virus screening in patients beginning chemotherapy for solid tumors. J Clin Oncol 2011;29:3270-7. 15. Paul S, Saxena A, Terrin N, Viveiros K, Balk EM, Wong JB. Hepatitis B Virus Reactivation and Prophylaxis During Solid Tumor Chemotherapy: A Systematic Review and Meta-analysis. Ann Intern Med 2016;164:30-40. 16. Wong WW, Hicks LK, Tu HA, et al. Hepatitis B virus screening before adjuvant chemotherapy in patients with early-stage breast cancer: a cost-effectiveness analysis. Breast Cancer Res Treat 2015;151:639-52. 17. Tan G, Zhou K, Tan CH, et al. Cost Effectiveness of Universal Hepatitis B Virus Screening in Patients Beginning Chemotherapy for Sarcomas or GI Stromal Tumors. J Glob Oncol 2016;2:186-99. 18. Roberts H, Kruszon-Moran D, Ly KN, et al. Prevalence of chronic hepatitis B virus (HBV) infection in U.S. households: National Health and Nutrition Examination Survey (NHANES), 1988-2012. Hepatology 2016;63:388-97. 13

Konijeti et al. 19. Schweitzer A, Horn J, Mikolajczyk RT, Krause G, Ott JJ. Estimations of worldwide prevalence of chronic hepatitis B virus infection: a systematic review of data published between 1965 and 2013. Lancet 2015;386:1546-55. 20. Kowdley KV, Wang CC, Welch S, Roberts H, Brosgart CL. Prevalence of chronic hepatitis B among foreign-born persons living in the United States by country of origin. Hepatology 2012;56:422-33. 21. Shing JZ, Ly KN, Xing J, Teshale EH, Jiles RB. Prevalence of Hepatitis B Virus Infection among US Adults Aged 20-59 Years with a History of Injection Drug Use: National Health and Nutrition Examination Survey, 2001-2016. Clin Infect Dis 2019. 22. Yun J, Kim KH, Kang ES, et al. Prophylactic use of lamivudine for hepatitis B exacerbation in postoperative breast cancer patients receiving anthracycline-based adjuvant chemotherapy. Br J Cancer 2011;104:559-63. 23. Chen WC, Cheng JS, Chiang PH, et al. A Comparison of Entecavir and Lamivudine for the Prophylaxis of Hepatitis B Virus Reactivation in Solid Tumor Patients Undergoing Systemic Cytotoxic Chemotherapy. PLoS One 2015;10:e0131545. 24. Noone AM, Howlader N, Krapcho M, et al. SEER Cancer Statistics Review, 1975-2015, National Cancer Institute. Bethesda, MD, https://seer.cancer.gov/csr/1975_2015/, based on November 2017 SEER data submission, posted to the SEER web site, April 2018. 25. Woo G, Tomlinson G, Yim C, et al. Health state utilities and quality of life in patients with hepatitis B. Can J Gastroenterol 2012;26:445-51. 26. Kanwal F, Gralnek IM, Martin P, Dulai GS, Farid M, Spiegel BM. Treatment alternatives for chronic hepatitis B virus infection: a cost-effectiveness analysis. Ann Intern Med 2005;142:821-31. 27. Shiroiwa T, Fukuda T, Shimozuma K. Cost-effectiveness analysis of trastuzumab to treat HER2positive advanced gastric cancer based on the randomised ToGA trial. Br J Cancer 2011;105:1273-8. 28. Poole CD, Connolly MP, Chang J, Currie CJ. Health utility of patients with advanced gastrointestinal stromal tumors (GIST) after failure of imatinib and sunitinib: findings from GRID, a randomized, double-blind, placebo-controlled phase III study of regorafenib versus placebo. Gastric Cancer 2015;18:627-34. 29. May AM, Bosch MJ, Velthuis MJ, et al. Cost-effectiveness analysis of an 18-week exercise programme for patients with breast and colon cancer undergoing adjuvant chemotherapy: the randomised PACT study. BMJ Open 2017;7:e012187. 30. Georgieva M, da Silveira Nogueira Lima JP, Aguiar P, Jr., de Lima Lopes G, Jr., Haaland B. Costeffectiveness of pembrolizumab as first-line therapy for advanced non-small cell lung cancer. Lung Cancer 2018;124:248-54. 31. Sanders GD, Neumann PJ, Basu A, et al. Recommendations for Conduct, Methodological Practices, and Reporting of Cost-effectiveness Analyses: Second Panel on Cost-Effectiveness in Health and Medicine. JAMA 2016;316:1093-103. 32. Mariotto AB, Yabroff KR, Shao Y, Feuer EJ, Brown ML. Projections of the cost of cancer care in the United States: 2010-2020. J Natl Cancer Inst 2011;103:117-28. 33. Zurawska U, Hicks LK, Woo G, et al. Hepatitis B virus screening before chemotherapy for lymphoma: a cost-effectiveness analysis. J Clin Oncol 2012;30:3167-73. 34. Cholankeril G, Perumpail RB, Hu M, Skowron G, Younossi ZM, Ahmed A. Chronic Hepatitis B Is Associated with Higher Inpatient Resource Utilization and Mortality Versus Chronic Hepatitis C. Dig Dis Sci 2016;61:2505-15. 35. Allen AM, Kim WR, Moriarty JP, Shah ND, Larson JJ, Kamath PS. Time trends in the health care burden and mortality of acute on chronic liver failure in the United States. Hepatology 2016;64:2165-72. 36. Hwang JP, Suarez-Almazor ME, Cantor SB, et al. Impact of the timing of hepatitis B virus identification and anti-hepatitis B virus therapy initiation on the risk of adverse liver outcomes for patients receiving cancer therapy. Cancer 2017;123:3367-76. 14

Table 1. Transition Probabilities, Costs, and Utilities Variable

Base Case

ANNUAL PROBABILITIES HBV Prevalence in United States Chronic HBV 0.003 HBsAg-positive in high-risk patients 0.0646 HBsAg-negative/HBcAb-positive in high0.1 risk patients HBsAg-positive in all patients 0.004 HBsAg-negative/HBcAb-positive in all 0.05 patients Prevalence of high-risk patients 0.112 HBV-related hepatitis in HBsAg-positive patients HBV prophylaxis if identified as HBsAg1 positive HBV prophylaxis if HBcAb-positive only 0 Reactivation without prophylaxis resulting 0.25 in nonsevere hepatitisa Reactivation with HBV prophylaxis (ETC, 0.035 TEN) resulting in nonsevere hepatitisa

Range

Distribution

References

0.002-0.004 0.041-0.2 0.061-0.12

Beta Beta Beta

18,19 10, 20, 21 10, 20, 21

0.002-0.01 0.028-0.077

Beta Beta

2, 10, 18, 19, 21 10, 18, 21

Beta

20, 21

0-1

Beta

0-1 0.04-0.68

Beta Beta

15

0.023-0.05

Beta

15

0.001-0.006

Beta

Expert opinion

Severe hepatitis if develop nonsevere 0.9 0.690-0.950 hepatitis, no prophylaxis; initiate treatment when nonsevere hepatitis recognizeda HBV-related hepatitis in HBsAg-negative/HBcAb-positive patients Reactivation without prophylaxis, resulting 0.03 0.003-0.09 in nonsevere hepatitisa Severe hepatitis if develop nonsevere 0.9 0.690-0.950 hepatitis, no prophylaxis; initiate treatment when nonsevere hepatitis recognizeda Mortality Non-HBV related death in patients 0.1 0.02-0.29 undergoing solid tumor chemotherapy ALF/death due to HBV reactivationa 0.02 0.01-0.20 Reactivation without prophylaxis by tumor subtype Gastrointestinal tract 0.16 0.06-0.50 Breast 0.21 0.05-0.68 Lung 0.20 0.14-0.37 Head and Neck 0.28 0.17-0.29 Other 0.13 0.06-0.50

Beta

Expert opinion

Beta

15

Severe hepatitis if develop nonsevere hepatitis despite prophylaxis for HBVa

0.005

UTILITIES 1

Expert opinion

Beta

24

Beta

15

Beta Beta Beta Beta Beta

15 15 15 15 15

Patient with solid tumor before chemotherapy Patient with solid tumor during chemotherapy Chronic HBV infection HBV reactivation - nonsevere hepatitis (assumed to be similar to compensated cirrhosis) HBV reactivation - severe hepatitis (assumed to be same as decompensated cirrhosis) Acute liver failure Patient with solid tumor in HBV recovery

0.95

0.85-0.98

Triangular

29

0.85

0.70-0.90

Triangular

16, 28-30

0.87 0.81

0.83-0.91 0.75-0.86

Triangular Triangular

25 25, 26

0.55

0.22-0.75

Triangular

25, 26

0.2 0.7

0.10-0.30 0.60-0.80

Triangular Triangular

Expert opinion 16, 27

15.07 367.27 14.74 15.04 16.8 11.09 5.61 81.29 60.5

Gamma Gamma Gamma Gamma Gamma Gamma Gamma Gamma Gamma

CMS CMS CMS CMS CMS CMS CMS GoodRxb GoodRxb

82.87

Gamma

Physician office visit - follow-up

54

Gamma

Physician office visit - acute hepatitis

117.37

Gamma

HCPCS 99214, CMS HCPCS 99213, CMS HCPCS 99215, CMS 34 32

COSTS, 2018 US$ Complete metabolic panel (CPT 80053) HBV DNA PCR (CPT 87912) Hepatitis B sAg (CPT 87340) Hepatitis B sAb (CPT 86706) Hepatitis B cAb total (CPT 86704) Complete blood count (CPT 85025) PT/INR (CPT 85610) Entecavir 0.5mg/d x 30 days Tenofovir disoproxil fumarate 300mg/d x 30 days Physician office visit - initial

Hospitalization for severe HBV hepatitis 12405 Gamma Initial costs cancer care (first year of 50843 26677-70176 Gamma diagnosis) Continuing costs of cancer care (after first 5306 2551-9087 Gamma 32 year) Terminal costs (Other cause) 13647 865-25376 Gamma 32 End of life organ failure (acute liver 54921 Gamma 33, 35 failure) Abbreviations: HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; HBcAb, hepatitis B core antibody; ALF, acute liver failure; CMS, Centers for Medicare and Medicaid Services; CPT, Current Procedural Terminology; HCPCS, Healthcare Common Procedural Coding System a See Supplementary Methods b

Drug pricing obtained from GoodRx.com, accessed December 2, 2018

2

Table 2. Primary Analysis of HBV Screening Strategies in Patients prior to Solid Tumor Chemotherapy Scenario Screen None Screen High-Risk Screen All

Cost (95% CI) $55,931 ($34,479 - $83,968) $55,976 ($34,486 - $84,125) $56,205 ($34,571 - $84,229)

Incremental Cost

$44.36 $229.03

1

QALY (95% CI) 1.831 (1.714 - 1.952) 1.834 (1.722 - 1.952) 1.839 (1.723 - 1.960)

Incremental QALY

ICER

0.003

$14,152

0.005

$42,761

Table 3. One-Way Sensitivity Analyses (Willingness to Pay $50,000) Probabilities

HBV Prevalence in United States sAg+ in high-risk patients sAg-/cAb+ in high-risk patients sAg+ in all patients sAg-/cAb+ in all patients

Base Case Value

Sensitivity Analysis Range

Threshold Preferred Strategy Below Threshold

Preferred Strategy Above Threshold

0.0646 0.10 0.004 0.05

0.041-0.2 0.061-0.12 0.002-0.01 0.028-0.077

n/a n/a n/a 0.04

Screen All Screen All Screen All Screen High Risk

Screen All Screen All Screen All Screen All

0.001-1.0

0.13

Screen None

Screen All

0.04-0.68

n/a

Screen All

Screen All

0.001-1.0

n/a

Screen All

Screen All

0.030

0.003-0.09

0.04

Screen All

Screen None

0.10

0.02-0.29

0.03

Screen None

Screen All

0.1

0.001-1.0

0.25

Screen All

Screen High Risk

$81.29 $60.50

$50-$200 $50-$200

n/a n/a

Screen All Screen All

Screen All Screen All

0.95

0.85-0.98

n/a

Screen All

Screen All

0.85

0.70-0.90

n/a

Screen All

Screen All

0.87 0.81

0.83-0.91 0.75-0.86

n/a n/a

Screen All Screen All

Screen All Screen All

0.55

0.22-0.75

n/a

Screen All

Screen All

0.2 0.7

0.10-0.30 0.60-0.80

n/a n/a

Screen All Screen All

Screen All Screen All

HBV related hepatitis in sAg+ patients HBV prophylaxis if identified as 1.00 HBsAg+ Reactivation without 0.25 prophylaxis, resulting in nonsevere hepatitis HBV related hepatitis in sAg- cAb+ patients HBV prophylaxis if identified as 0 HBsAg-/HBcAb+ Reactivation without prophylaxis, resulting in nonsevere hepatitis Mortality Non-HBV related death in patients undergoing solid tumor chemotherapy Probability of screening high risk patients Costs Entecavir 0.5mg/d x 30 days Tenofovir disoproxil fumarate 300mg/d x 30 days Utilities Patient with solid tumor before chemotherapy Patient with solid tumor during chemotherapy Chronic HBV infection HBV reactivation – non-severe hepatitis HBV reactivation - severe hepatitis Acute liver failure Patient with solid tumor in HBV recovery

1

2

What You Need to Know

Background: We aimed to analyze the cost-effectiveness of HBV screening before chemotherapy for patients with solid tumors. Findings: In a Markov model analysis, we found screening all patients with solid tumors for HBV infection before chemotherapy to be the most cost-effective strategy, compared to screening high-risk patients or no screening. Implications for Patient Care: Guidelines should recommend HBV tests for patients initiating chemotherapy for solid tumors.

Supplementary Methods 1. One month duration between diagnostic evaluation of solid tumor and initiation of chemotherapy a. For those undergoing HBV screening, assume testing for HBV sAg, sAb, and core IgG Ab 2. Probability of a physician identifying a patient at high risk for HBV: 0.1 3. Probability of receiving HBV prophylaxis if identified as HBsAg+/HBcAb±: 1.00 4. Probability of receiving HBV prophylaxis if identified as HBsAg-/HBcAb+: 0 5. For patients who are HBsAg+/HBcAb±: a. If HBsAg positivity is identified on screening: i. Assume initial office visit with Hepatologist for evaluation prior to chemotherapy initiation ii. Assume all patients who are HBsAg+ are given HBV prophylaxis with entecavir iii. Base case HBV reactivation rate for entecavir 3.5% (Table 1). If HBV reactivation occurs on entecavir, assume patient is switched to tenofovir disoproxil fumarate. iv. Assume median time to reactivation (presenting as non-severe hepatitis) with HBV prophylaxis is 9 months v. For sensitivity analysis, where patients may not be given HBV prophylaxis despite detection of HBsAg+ on screen, assume median time to reactivation (presenting as non-severe hepatitis) without prophylaxis is 3 months vi. Assume following median times to progression: non-severe hepatitis to severe hepatitis (2 months), severe hepatitis to acute liver failure or recovery (2 weeks) b. If HBsAg positivity is identified in the setting of HBV reactivation (i.e., patients were not screened prior to starting chemotherapy): i. Assume median time to reactivation (presenting as non-severe hepatitis) without prophylaxis is 3 months

ii. Assume entecavir given for HBV reactivation when identified iii. Assume following median times to progression: non-severe hepatitis to severe hepatitis (1 month), severe hepatitis to acute liver failure (2 weeks) 6. For patients who are HbsAg-/cAb+: a. If HBsAg-/cAb+ identified on screening: i. Assume HBsAg-/cAb+ patients are not started on HBV prophylaxis ii. Assume surveillance labs for HBV reactivation every 3 months iii. Assume median time to reactivation (presenting as non-severe hepatitis) without prophylaxis is 7 months iv. Assume entecavir given for HBV reactivation when identified v. Assume following median times to progression: non-severe hepatitis to severe hepatitis (2 months), severe hepatitis to acute liver failure (2 weeks) b. If HBsAg-/cAb identified in setting of HBV reactivation (i.e., patients were not screened prior to starting chemotherapy): i. Assume median time to reactivation (presenting as non-severe hepatitis) without prophylaxis is 4 months ii. Assume entecavir given for HBV reactivation when identified iii. Assume following median times to progression: non-severe hepatitis to severe hepatitis (1 month), severe hepatitis to acute liver failure (2 weeks) 7. Considering patients who are either HBsAg+/HBcAb± or HBsAg-/HBcAb+ and do not undergo screening prior to chemotherapy are recognized to have HBV reactivation more commonly in states of severe hepatitis or acute liver failure, we assume they carry a 2-fold relative risk above base case estimates for HBV reactivation. 8. All patients on HBV prophylaxis undergo follow-up office visits with Hepatology, and monitoring of surveillance labs (HBV DNA, CBC, CMP, and PT/INR) every 3 months

9. Those initiated on HBV prophylaxis or treatment at any point in the Markov model remain on HBV therapy throughout the 5 year time horizon 10. If patients develop HBV reactivation (regardless if they receive prophylaxis), they follow a course of non-severe hepatitis, potentially followed by severe hepatitis, potentially followed by acute liver failure resulting in death (Table 1, Figure 1). a. Assume chemotherapy interruption (range, 1-4 months) during management of HBV reactivation 11. Patients with HBV reactivation with non-severe hepatitis: Managed as outpatients, with office visits and surveillance laboratories done monthly. a. If patients recover from non-severe hepatitis, assume surveillance labs every month for 3 months, then every 2 months for 4 months, then every 3 months thereafter. 12. Patients developing severe hepatitis due to HBV reactivation are hospitalized for an average of 2 weeks. a. If patients recover from severe hepatitis, assume surveillance labs done every month for 3 months, then every 2 months for 4 months, then every 3 months. During HBV recovery, same antiviral treatment as given as was received during initial identification of HBV reactivation 13. Patients progressing to acute liver failure following severe hepatitis are hospitalized for an average of 2 weeks, with ongoing antiviral treatment. Assume 100% risk of mortality with acute liver failure. 14. All patients either started on HBV prophylaxis or requiring HBV treatment due to reactivation at any point during the 5-year time horizon are modeled to continue HBV treatment until model completion.

Supplementary Figure 2. Tornado Diagram (Microsimulation)

Variable Name

Variable Definition

Variable Name

Variable Definition

cCANinit

Cost - Initial costs of cancer care

pREACnoppxSAG

uChemo

Utility - patient with solid tumor during chemotherapy

cCANother

Probability - HBV reactivation without prophylaxis if sAg+, not screened Cost - terminal costs (other costs)

cCANcont

Cost - Continuing costs of cancer care

pREACsevnoppx

uSolTum

Utility - patient with solid tumor before chemotherapy

pREACppxSAG

pCABall

Probability - sAg-/cAb+ in all patients

cHOSalf

pnoPPXcAB

Probability of no prophylaxis if cAb+ only

uHBVreacnonsev

Utility - HBV reactivation non-severe hepatitis

uCHB

Utility - Chronic HBV infection

uHBVreacsev

Utility - HBV reactivation severe hepatitis

Probability - severe hepatitis if develop non-severe hepatitis, no prophylaxis Probability - HBV reactivation with prophylaxis if sAg+ Cost - End of life organ failure (acute liver failure)

pREACnoppxCAB

Probability - HBV reactivation without prophylaxis if cAb+ only

cTen

Cost - tenofovir (x 30 days)

pScreen

Probability of screening high-risk patients

uHBValf

Utility - Acute liver failure

pSAGall

Probability - sAg+ in all patients

pCHB

Probability - chronic HBV

pScrREACnoppxC AB

Probability - HBV reactivation without prophylaxis if cAb+ only, identified on screening Probability of non-HBV related death in patients undergoing solid tumor chemotherapy Utility - patient with solid tumor in HBV recovery

pSAGhr

Probability - sAg+ in highrisk patients

pCABhr

Probability - sAg-/cAb+ in high-risk patients

pPPXcAB

Probability of prophylaxis if cAb+ only

pPPXsAG

Probability of prophylaxis if sAg+

pScrREACnoppxSAG

pALFreac

Probability of ALF/death due to HBV reactivation

pNoScrALFreac

cEnt

Cost - entecavir (x 30 days)

pREACsev

Probability - HBV reactivation without prophylaxis if sAg+ identified on screening Probability of ALF/death due to HBV reactivation in patients not screened Probability - severe hepatitis if develop non-severe hepatitis, with prophylaxis

cHOSshep

Cost - hospitalization for severe hepatitis

pDEATHnonhbv

uPostChemo