The NEW ENGLAND JOURNAL of MEDICINE ORIGINAL ARTICLE Safety and Efficacy of Vadadustat for Anemia in Patients Undergoing Dialysis K.-U. Eckardt, R. Agarwal, A. Aswad, A. Awad, G.A. Block, M.R. Bacci, Y.M.K. Farag, S. Fishbane, H. Hubert, A. Jardine, Z. Khawaja, M.J. Koury, B.J. Maroni, K. Matsushita, P.A. McCullough, E.F. Lewis, W. Luo, P.S. Parfrey, P. Pergola, M.J. Sarnak, B. Spinowitz, J. Tumlin, D.L. Vargo, K.A. Walters, W.C. Winkelmayer, J. Wittes, R. Zwiech, and G.M. Chertow ABSTRACT BACKGROUND Vadadustat is an oral hypoxia-inducible factor prolyl hydroxylase inhibitor, a class The authors' full names, academic of compounds that stimulate endogenous erythropoietin production. METHODS We conducted two randomized, open-label, noninferiority phase 3 trials to evalu- ate the safety and efficacy of vadadustat, as compared with darbepoetin alfa, in patients with anemia and incident or prevalent dialysis-dependent chronic kidney disease (DD-CKD). The primary safety end point, assessed in a time-to-event analysis, was the first occurrence of a major adverse cardiovascular event (MACE, a composite of death from any cause, a nonfatal myocardial infarction, or a non- fatal stroke), pooled across the trials (noninferiority margin, 1.25). A key second- ary safety end point was the first occurrence of a MACE plus hospitalization for either heart failure or a thromboembolic event. The primary and key secondary efficacy end points were the mean change in hemoglobin from baseline to weeks 24 to 36 and from baseline to weeks 40 to 52, respectively, in each trial (non- inferiority margin, -0.75 g per deciliter). RESULTS A total of 3923 patients were randomly assigned in a 1:1 ratio to receive vadadus- tat or darbepoetin alfa: 369 in the incident DD-CKD trial and 3554 in the prevalent DD-CKD trial. In the pooled analysis, a first MACE occurred in 355 patients (18.2%) in the vadadustat group and in 377 patients (19.3%) in the darbepoetin alfa group (hazard ratio, 0.96; 95% confidence interval [CI], 0.83 to 1.11). The mean differences between the groups in the change in hemoglobin concentration were -0.31 g per deciliter (95% CI, −0.53 to −0.10) at weeks 24 to 36 and -0.07 g per deciliter (95% CI, −0.34 to 0.19) at weeks 40 to 52 in the incident DD-CKD trial and -0.17 g per deciliter (95% CI, −0.23 to −0.10) and −0.18 g per deciliter (95% CI, −0.25 to -0.12), respectively, in the prevalent DD-CKD trial. The inci- dence of serious adverse events in the vadadustat group was 49.7% in the incident DD-CKD trial and 55.0% in the prevalent DD-CKD trial, and the incidences in the darbepoetin alfa group were 56.5% and 58.3%, respectively. CONCLUSIONS Among patients with anemia and CKD who were undergoing dialysis, vadadustat was noninferior to darbepoetin alfa with respect to cardiovascular safety and cor- rection and maintenance of hemoglobin concentrations. (Funded by Akebia Therapeutics and Otsuka Pharmaceutical; INNOVATE ClinicalTrials.gov numbers, NCT02865850 and NCT02892149.) degrees, and affiliations are listed in the Appendix. Address reprint requests to Dr. Eckardt at the Department of Medical Intensive Nephrology and Care, Charité-Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany, or at nephro-intensiv@charite.de. A list of investigators is provided in the Supplementary Appendix, available at NEJM.org. N Engl J Med 2021;384:1601-12. DOI: 10.1056/NEJMoa2025956 Copyright 2021 Massachusetts Medical Society. N ENGL J MED 384;17 NEJM.ORG APRIL 29, 2021 The New England Journal of Medicine Downloaded from nejm.org on May 7, 2024. For personal use only. No other uses without permission. Copyright © 2021 Massachusetts Medical Society. All rights reserved. 1601 1602 The NEW ENGLAND JOURNAL of MEDICINE A NEMIA, A COMMON COMPLICATION OF chronic kidney disease (CKD), is associ- of life, an increase in red-cell transfusions, and a heightened risk of cardiovascular events. 1-3 Erythropoiesis-stimulating agents (ESAs) (i.e., recombinant human erythropoietin and its de- rivatives) are standard care for the management of anemia in patients with CKD. The use of ESAs to target hemoglobin concentrations in the nor- mal or near-normal range in patients with CKD has been shown to increase the risks of stroke, vascular access thrombosis, and death 4-6 ; such findings have resulted in recommendations for caution in the use of ESAs and for only partial correction of anemia. Hypoxia-inducible factor (HIF), a transcription factor that regulates physiological responses to hypoxia, stimulates erythropoietin production by the liver and kidneys. 8-10 HIF is regulated by oxy- gen-dependent proteasomal degradation through a family of prolyl hydroxylases that serve as oxy- gen sensors. 9,¹¹ HIF prolyl hydroxylase inhibitors comprise a recently developed class of com- pounds that stabilize HIF, in turn stimulating endogenous erythropoietin production and ulti- mately erythropoiesis. 10,12-15 Vadadustat, an oral HIF prolyl hydroxylase in- hibitor, is an investigational drug that is in de- velopment for the treatment of anemia associ- ated with CKD. In phase 2 trials, vadadustat was reported to increase and maintain hemoglobin concentrations in patients with CKD and anemia both those who were undergoing dialysis and those who were not undergoing dialysis." 16-19 The vadadustat phase 3 program includes four phase 3 international, randomized, controlled trials: two trials involving patients with non-dialysis-depen- dent CKD (the PROTECT trials) and two trials involving patients with dialysis-dependent CKD (DD-CKD) (the INNOVATE trials). Here, we re- port the results of the two INNOVATE trials, which evaluated the cardiovascular safety and hematologic efficacy of vadadustat as compared with the ESA darbepoetin alfa. METHODS TRIAL DESIGN AND OVERSIGHT Details regarding the rationale, design, and methods of the two INNOVATE trials were de- scribed previously, 20 and the protocols for the trials are available with the full text of this article at NEJM.org. Both trials were randomized, open- were designed to evaluate the cardiovascular safety and hematologic efficacy of vadadustat, as compared with darbepoetin alfa, for the treat- ment of anemia in patients undergoing hemo- dialysis or peritoneal dialysis. In both trials, personnel at Akebia Therapeutics and Otsuka Pharmaceutical were unaware of the treatment as- signments. In the incident DD-CKD trial (correc- tion-conversion trial; ClinicalTrials.gov number, NCT02865850), which involved the correction and maintenance of hemoglobin concentrations, patients were to have initiated dialysis within 16 weeks before screening and were to have had limited exposure to ESAs. In the prevalent DD-CKD trial (conversion trial; NCT02892149), which involved the maintenance of hemoglobin concentrations, patients were to have been under- going maintenance dialysis for at least 12 weeks before screening and were to have been receiving treatment with an ESA.20 Akebia Therapeutics designed the trials and protocols with input from an executive steering committee, and Otsuka Pharmaceutical partici- pated in the executive steering committee meet- ings and provided input on trial protocol amend- ments. An independent ethics committee approved the informed consent forms. The trial investiga- tors conducted the trials in collaboration with Akebia Therapeutics. The executive steering com- mittee supervised the conduct and progress of the trials. The trials were monitored by an inde- pendent data and safety monitoring committee (lists of the members of these committees are provided in the Supplementary Appendix, avail- able at NEJM.org). The first and last authors wrote the first draft of the manuscript and made final decisions regarding the content of the sub- mitted manuscript. All the authors had access to the trial data, critically reviewed earlier drafts of the manuscript, and approved the manuscript for submission. The investigators vouch for the ac- curacy and completeness of the data; the authors and Akebia Therapeutics vouch for the fidelity of the trials to the respective protocols and for the analysis of the data. PARTICIPANTS In both trials, eligible patients were at least 18 years of age, had CKD and were undergoing di- N ENGL J MED 384;17 NEJM.ORG APRIL 29, 2021 The New England Journal of Medicine Downloaded from nejm.org on May 7, 2024. For personal use only. No other uses without permission. Copyright © 2021 Massachusetts Medical Society. All rights reserved. VADADUSTAT FOR ANEMIA IN PATIENTS UNDERGOING DIALYSIS alysis, had a serum ferritin concentration of at least 100 ng per milliliter and a transferrin satu- ration of at least 20%, and had not received a red-cell transfusion within the previous 8 weeks. In addition, patients had a hemoglobin concen- tration between 8 and 11 g per deciliter (incident DD-CKD trial) or a hemoglobin concentration between 8 and 11 g per deciliter (patients in the United States) or between 9 and 12 g per deci- liter (patients in other countries) (prevalent DD-CKD trial). Patients were excluded if they had anemia that was considered to be due to causes other than CKD or if they had uncon- trolled hypertension or had had a recent cardio- vascular event.20 All the patients provided written informed consent. RANDOMIZATION AND TRIAL PERIODS Eligible patients were randomly assigned, in a 1:1 ratio, to receive vadadustat or darbepoetin alfa. Randomization was stratified according to geographic region (United States vs. Europe vs. other regions), New York Heart Association (NYHA) heart failure classification (class 0 or I vs. class II or III), and hemoglobin concentration at trial entry (incident DD-CKD trial, <9.5 vs. ≥9.5 g per deciliter; prevalent DD-CKD trial, <10 vs. ≥10 g per deciliter). Both trials had four defined trial periods: a correction or conversion period (weeks 0 to 23); a maintenance period (weeks 24 to 52), comprising the primary (weeks 24 to 36) and secondary (weeks 40 to 52) efficacy evaluation periods; a long-term treatment period (weeks 53 to the end of treatment); and a 4-week safety follow-up period.20 END POINTS The primary safety end point, assessed in a time-to-event analysis, was the first occurrence of an adjudicated major adverse cardiovascular event (MACE) ― a composite end point of death from any cause, a nonfatal myocardial infarc- tion, or a nonfatal stroke - pooled across the two trials. Key secondary safety end points, which were also pooled across the trials, were the first occurrence of "expanded MACE" (a MACE plus hospitalization for either heart fail- ure or a thromboembolic event, excluding vas- cular access failure); death from cardiovascular causes, a nonfatal stroke, or a nonfatal myocar- dial infarction combined; death from cardio- vascular causes; and death from any cause. Adjudication of MACE was performed by an independent clinical end-point committee whose members were unaware of the treatment as- signments (a list of the committee members is provided in the Supplementary Appendix). The primary efficacy end point was the mean change in the hemoglobin concentration from baseline to the average concentration during the primary evaluation period, and the key sec- ondary efficacy end point was the mean change in the hemoglobin concentration from baseline to the average concentration during the sec- ondary evaluation period.20 INTERVENTION The starting dose of vadadustat was 300 mg orally once daily, with doses of 150, 450, and 600 mg available for adjustment of the dose to a maximum of 600 mg daily. Darbepoetin alfa was administered subcutaneously or intrave- nously; the initial dose was based on the previ- ous dose or, in the case of patients who had not received darbepoetin alfa before random- ization, on information in the product label. Doses of the trial medications were adjusted ac- cording to protocol-specified dose-adjustment algorithms to achieve target hemoglobin con- centrations (in the United States, 10 to 11 g per deciliter; in other countries, 10 to 12 g per deciliter). 20 We encouraged the use of iron supplementation (intravenous, oral, or intradi- alytic administration) to maintain a serum fer- ritin concentration of at least 100 ng per mil- liliter or a transferrin saturation of at least 20%. The provision of red-cell transfusions did not necessitate discontinuation of the trial med- ication. Starting at week 6, patients in both treatment groups could receive ESAs as rescue therapy if they had worsening symptoms of anemia and a hemoglobin concentration of less than 9.5 g per deciliter. In addition, in the dar- bepoetin alfa group, an ESA was defined post hoc as a rescue medication if the dose was at least double that of the previous dose of darbe- poetin alfa. Patients temporarily discontinued the trial medication while receiving ESAs as rescue therapy. STATISTICAL ANALYSIS The prespecified noninferiority margins, which were selected in consultation with regulatory agencies, were an upper bound of the 95% con- N ENGL J MED 384;17 NEJM.ORG APRIL 29, 2021 The New England Journal of Medicine Downloaded from nejm.org on May 7, 2024. For personal use only. No other uses without permission. Copyright © 2021 Massachusetts Medical Society. All rights reserved. 1603 1604 The NEW ENGLAND JOURNAL of MEDICINE fidence interval of 1.25 for the primary safety end point and a lower bound of -0.75 g per deciliter for the primary efficacy end point. 20 To evaluate the primary and key secondary efficacy end points, an analysis of covariance with mul- tiple imputation for missing data was used, with baseline hemoglobin concentration, geographic region, and NYHA class as covariates. The analy- sis of the time to a first occurrence of a MACE was performed with the use of a Cox regression model stratified according to trial. The model included the following covariates: baseline hemo- globin concentration, geographic region (United States vs. Europe vs. other regions), NYHA class (0 or I vs. II or III), sex, age (≤65 years vs. >65 years), race (White vs. non-White), cardiovascu- lar disease (yes vs. no), and diabetes mellitus (yes vs. no). Subgroup analyses were performed with the use of the same Cox model as that used in the primary analysis. PATIENTS RESULTS Across the two trials, 3923 patients underwent randomization: 369 in the incident DD-CKD trial and 3554 in the prevalent DD-CKD trial (Fig. S1 in the Supplementary Appendix). The median duration of follow-up was 1.2 years (in- terquartile range [25th to 75th percentile], 0.8 to 1.7) in the incident DD-CKD trial and 1.7 years (interquartile range, 1.2 to 2.2) in the prevalent DD-CKD trial. patients (18.2%) in the vadadustat group and in 377 of the 1955 patients (19.3%) in the darbepo- etin alfa group (hazard ratio, 0.96; 95% confi- dence interval [CI], 0.83 to 1.11). The numbers and percentages of patients in whom the first MACE was death from any cause, a nonfatal myocardial infarction, or a nonfatal stroke were 253 (13.0%), 76 (3.9%), and 26 (1.3%), respec- tively, in the vadadustat group and 253 (12.9%), 87 (4.5%), and 37 (1.9%), respectively, in the darbepoetin alfa group (Table S2). The results within each trial were qualitatively consistent with the pooled analysis, but the confidence in- terval in the incident DD-CKD trial, which was a much smaller trial than the prevalent DD-CKD trial, was wide (incident DD-CKD trial: hazard ratio of a first MACE, 0.97; 95% CI, 0.54 to 1.76; prevalent DD-CKD trial: hazard ratio, 0.96; 95% CI, 0.83 to 1.12) (Table S3). Figure 1B shows the cumulative probability of a first expanded MACE, which occurred in 420 of the 1947 patients (21.6%) in the vadadustat group and in 449 of the 1955 patients (23.0%) in the darbepoetin alfa group (hazard ratio, 0.96; 95% CI, 0.84 to 1.10). Figures 1C and 1D show the pooled cumulative probability of death from cardiovascular causes (hazard ratio, 0.96; 95% CI, 0.77 to 1.20) and the cumulative probability of death from any cause (hazard ratio, 0.95; 95% CI, 0.81 to 1.12), respectively. Figure S2 shows the cumulative probability of a composite of death from cardiovascular causes, a nonfatal myocar- dial infarction, or a nonfatal stroke (hazard ra- tio, 0.95; 95% CI, 0.80 to 1.14). Table S2 shows similar incidences of MACE, expanded MACE, and their components in the two treatment groups across both trials. The results of pre- specified subgroup analyses of a first occurrence of a MACE and a first occurrence of an expand- ed MACE were consistent across subgroups, as shown in Figure 2 and Figure S3, respectively. The baseline characteristics at randomization have been described previously.20 The demo- graphic, clinical, and laboratory characteristics of the two treatment groups were generally well balanced in both trials (Table 1 and Table S1), except that in the incident DD-CKD trial, the percentage of patients with diabetes mellitus was higher among the patients randomly as signed to receive vadadustat than among those randomly assigned to receive darbepoetin alfa PRIMARY, KEY SECONDARY, AND OTHER EFFICACY (58.0% vs. 51.1%). PRIMARY AND KEY SECONDARY SAFETY END POINTS Analyses of the safety end points were based on the pooled safety population, which included 1947 patients in the vadadustat group and 1955 patients in the darbepoetin alfa group. Figure 1A shows the cumulative probability of a first MACE. A first MACE occurred in 355 of the 1947 END POINTS Incident DD-CKD Trial The median doses of vadadustat and darbepoetin alfa over the course of the trial are shown in Figure S5. Figure 3A shows the mean changes in hemoglobin concentrations in the two treat- ment groups over time. The least-squares mean (±SE) change in hemoglobin concentration from baseline to the average of the values during N ENGL J MED 384;17 NEJM.ORG APRIL 29, 2021 The New England Journal of Medicine Downloaded from nejm.org on May 7, 2024. For personal use only. No other uses without permission. Copyright © 2021 Massachusetts Medical Society. All rights reserved. VADADUSTAT FOR ANEMIA IN PATIENTS UNDERGOING DIALYSIS Table 1. Selected Demographic, Clinical, and Laboratory Characteristics of the Patients at Baseline in the Randomized Populations.* Incident DD-CKD Trial Prevalent DD-CKD Trial Characteristic Vadadustat (N=181) Darbepoetin Alfa (N=188) Vadadustat (N=1777) 56.5±14.8 55.6±14.6 57.9±13.9 Darbepoetin Alfa (N=1777) 58.4±13.8 Age - yr Male sex no. (%) 107 (59.1) 113 (60.1) 990 (55.7) 1004 (56.5) Racial or ethnic group ― no. (%)† 129 (71.3) 143 (76.1) 1135 (63.9) 1096 (61.7) White 38 (21.0) 35 (18.6) 432 (24.3) 444 (25.0) Black 12 (6.6) 8 (4.3) 76 (4.3) 99 (5.6) Asian American Indian 1 (0.6) 0 19 (1.1) 30 (1.7) 0 0 13 (0.7) 6 (0.3) Native Hawaiian 0 1 (0.5) 42 (2.4) 45 (2.5) Other group Multiple groups 1 (0.6) 0 8 (0.5) 5 (0.3) 0 1 (0.5) 52 (2.9) 52 (2.9) Not reported Hispanic ethnic group — no. (%)† Hispanic Non-Hispanic 71 (39.2) 66 (35.1) 682 (38.4) 674 (37.9) 104 (57.5) 118 (62.8) 1043 (58.7) 1040 (58.5) Not reported Unknown Time since dialysis initiated Type of dialysis — no. (%) - 5 (2.8) 3 (1.6) 36 (2.0) 47 (2.6) 1 (0.6) 1 (0.5) 16 (0.9) 0.14±0.09 0.15±0.28 4.00±4.02 16 (0.9) 3.94±4.01 - yr In-center hemodialysis 158 (87.3) 169 (90.9) 1652 (93.0) 1633 (91.9) Peritoneal dialysis 22 (12.2) 16 (8.6) 137 (7.7) 143 (8.0) Unknown or combination 3 (1.7) 1 (0.5) 17 (1.0) 18 (1.0) Disease history- — no. (%) Diabetes mellitus 105 (58.0) 96 (51.1) 971 (54.6) 998 (56.2) Cardiovascular disease 69 (38.1) 73 (38.8) 868 (48.8) 932 (52.4) Hemoglobin concentration - g/dl 9.4±1.1 9.2±1.1 10.6±0.9 10.2±0.8 * Plus-minus values are means ±SD. Percentages may not total 100 because of rounding. DD-CKD denotes dialysis- dependent chronic kidney disease. * Racial and ethnic groups were reported by the patient. *This analysis was performed in the safety population in both trials. Patients could have been included in more than one category. weeks 24 to 36 was 1.26±0.11 g per deciliter in the vadadustat group and 1.58±0.11 g per deci- liter in the darbepoetin alfa group. The corre- sponding changes from baseline to the average of the values during weeks 40 to 52 were 1.42±0.13 g per deciliter and 1.50±0.14 g per deciliter, respectively. The mean (±SE) differ- ences between the groups in the change in hemo- globin concentration were −0.31±0.11 g per deci- liter (95% CI, −0.53 to -0.10) at weeks 24 to 36 and −0.07±0.13 g per deciliter (95% CI, -0.34 to 0.19) at weeks 40 to 52. The percentages of patients who had an aver- age hemoglobin concentration within their country-specific target range during weeks 24 to 36 were 43.6% in the vadadustat group and 56.9% in the darbepoetin alfa group. During weeks 40 to 52, the percentages were 39.8% and 41.0%, respectively (Table S5). The percentages of patients who received red- cell transfusions during weeks 24 to 36 were 1.3% in the vadadustat group and 1.8% in the darbepoetin alfa group. During weeks 40 to 52, the percentages were 2.4% and 0.7%, respectively. N ENGL J MED 384;17 NEJM.ORG APRIL 29, 2021 The New England Journal of Medicine Downloaded from nejm.org on May 7, 2024. For personal use only. No other uses without permission. Copyright © 2021 Massachusetts Medical Society. All rights reserved. 1605