Roxadustat Versus Epoetin Alfa for Treating Anemia in Patients with Chronic Kidney Disease on Dialysis: Results from the Randomized Phase 3 ROCKIES Study

Abstract

Visual Abstract

Background 

Concerns regarding cardiovascular safety with current treatments for anemia in patients with dialysis-dependent (DD)-CKD have encouraged the development of alternatives. Roxadustat, an oral hypoxia-inducible factor prolyl hydroxylase inhibitor, stimulates erythropoiesis by increasing endogenous erythropoietin and iron availability.

Methods 

In this open-label phase 3 study, patients with DD-CKD and anemia were randomized 1:1 to oral roxadustat three times weekly or parenteral epoetin alfa per local clinic practice. Initial roxadustat dose depended on erythropoiesis-stimulating agent dose at screening for patients already on them and was weight-based for those not on them. The primary efficacy end point was mean hemoglobin change from baseline averaged over weeks 28‒52 for roxadustat versus epoetin alfa, regardless of rescue therapy use, tested for noninferiority (margin, −0.75 g/dl). Adverse events (AEs) were assessed.

Results 

Among 2133 patients randomized (n=1068 roxadustat, n=1065 epoetin alfa), mean age was 54.0 years, and 89.1% and 10.8% were on hemodialysis and peritoneal dialysis, respectively. Mean (95% confidence interval) hemoglobin change from baseline was 0.77 (0.69 to 0.85) g/dl with roxadustat and 0.68 (0.60 to 0.76) g/dl with epoetin alfa, demonstrating noninferiority (least squares mean difference [95% CI], 0.09 [0.01 to 0.18]; P<0.001). The proportion of patients experiencing ≥1 AE and ≥1 serious AE was 85.0% and 57.6% with roxadustat and 84.5% and 57.5% with epoetin alfa, respectively.

Conclusions 

Roxadustat effectively increased hemoglobin in patients with DD-CKD, with an AE profile comparable to epoetin alfa.

Clinical Trial registry name and registration number: 

Safety and Efficacy Study of Roxadustat to Treat Anemia in Patients With Chronic Kidney Disease, on Dialysis. ClinicalTrials.gov Identifier: NCT02174731.

Patients with anemia of CKD have inadequate erythropoietin synthesis in response to low oxygen and anemia.^1,2 These patients may also experience functional and/or absolute iron deficiency.³ Anemia is associated with reduced quality of life and increased transfusion, hospitalization, and mortality rates.^4,5

Iron supplementation and erythropoiesis-stimulating agents (ESAs) are currently recommended for treating anemia of CKD in patients on dialysis.⁶ Safety concerns noted with ESAs and the potential convenience of oral administration have stimulated the development of alternative treatments for anemia.^78–9

Elucidation of cellular pathways responsible for protection from hypoxia led to development of hypoxia-inducible factor prolyl hydroxylase (HIF-PH) inhibitors. Roxadustat, the first-in-class HIF-PH inhibitor, stabilizes HIF-α and promotes an erythropoietic response, including increased synthesis of endogenous erythropoietin, and may promote iron absorption, transport, and mobilization.¹⁰ HIF-PH inhibitors may have advantages over ESAs, including oral administration, anemia correction with smaller increases in serum erythropoietin, improved iron metabolism, and efficacy regardless of inflammation.^11,12 Roxadustat is approved to treat anemia in patients with CKD on or not on dialysis in China, Japan, and Europe.^13,14

Here, we report the results of ROCKIES, an international, phase 3, randomized, open-label trial to assess the efficacy of roxadustat compared with epoetin alfa in 2133 patients with anemia of CKD on dialysis.

Methods

Trial Design and Oversight

This phase 3, international, randomized, open-label, active-comparator study assessed roxadustat versus epoetin alfa for treating anemia in patients with CKD on dialysis (NCT02174731). The study comprised a 6-week screening period, followed by a treatment period of up to 4 years (Figure 1) and was conducted between July 1, 2014, and September 26, 2018. Treatment end date was based on accruing a predefined number of patients with adjudicated cardiovascular events for pooled analysis of three phase 3 roxadustat studies in patients on dialysis (NCT02174731, NCT02273726, and NCT02052310).

Figure 1.: Study design. ^aEOT is defined as the patient’s next scheduled visit after discontinuing study treatment. ^bEOS is defined as the last visit of the last patient undergoing the study. CV, cardiovascular; EOS, end of study; EOT, end of treatment; Hb, hemoglobin; R, randomization; TIW, three times a week.

The study was performed in accordance with the Declaration of Helsinki, the International Council for Harmonisation Good Clinical Practice guidelines, applicable regulatory requirements, and AstraZeneca’s policy on Bioethics and Human Biologic Samples. All patients provided written consent before screening. The final study protocol and informed consent form were approved by the applicable Independent Ethics Committee or Institutional Review Board for each site.

Patients

Eligible patients were aged ≥18 years, on hemodialysis or peritoneal dialysis for ≥30 days before randomization (modified to >2 weeks to ≤4 months after a protocol amendment to facilitate the recruitment of incident dialysis patients), with documented hemoglobin (Hb) <10 g/dl if ESA-untreated or Hb <12 g/dl if ESA-treated at two assessments obtained at least 7 days apart, ferritin levels ≥100 ng/ml, and transferrin saturation (TSAT) ≥20%. Patients were excluded if they had the following: New York Heart Association Class III or IV congestive heart failure at enrollment; myocardial infarction, acute coronary syndrome, stroke, seizure, or a thrombotic/thromboembolic event within 12 weeks before randomization; or uncontrolled hypertension at the time of randomization. Full inclusion/exclusion criteria are provided in the Supplemental Appendix.

Treatment

Eligible patients were randomized 1:1 to either oral roxadustat three times weekly (TIW) or parenteral epoetin alfa per local clinic standard of care. Randomization was stratified by country and codes were computer-generated by AstraZeneca R&D using GRand and loaded into the Interactive Web Response System database, with a block size of 4.

For patients randomized to roxadustat who were receiving an ESA at entry, starting doses were 70–200 mg TIW, based on prior ESA dose (Supplemental Table 1). ESA-naive patients at entry started roxadustat on either 70 mg (patients with dry weights 45–70 kg) or 100 mg (dry weights >70–160 kg) TIW. A roxadustat-specific dosing algorithm directed dose adjustment to maintain Hb at 11±1 g/dl (Supplemental Table 2). For patients receiving hemodialysis, it was recommended that roxadustat was taken after completion of the hemodialysis session. Patients randomized to epoetin alfa were dosed based on ESA dose before study entry, or with 50 IU/kg TIW if ESA-naive (Supplemental Appendix), and dose was adjusted according to local prescribing information.

Oral iron supplementation was permitted in both groups without restriction. In the roxadustat group, intravenous (IV) iron was permitted if a patient’s Hb did not increase sufficiently after ≥2 dose increases, and their ferritin or TSAT values were <100 ng/ml or <20%, respectively, and according to standard of care in the epoetin alfa group. Red blood cell (RBC) transfusion was permitted in either group if rapid anemia correction was required to stabilize the patient’s condition or if the investigator considered transfusion to be medically necessary. Rescue ESA therapy was temporarily allowed for roxadustat-treated patients if the following criteria were met: Hb was <8.5 g/dl on two consecutive measurements taken at least 5 days apart despite ≥2 roxadustat dose increases or reaching the maximum roxadustat dose; reducing the risk of alloimmunization in transplant-eligible patients and/or reduction of other RBC transfusion-related risks was a goal; and no indication of iron deficiency or bleeding as the cause of lack of response. Rescue ESA therapy was to be stopped when Hb increased to >9 g/dl or after one 4-week course.

Patients receiving IV iron supplementation and/or RBC transfusion rescue therapy could continue roxadustat. Patients receiving temporary rescue ESA therapy were to hold roxadustat until rescue ESA was stopped. For permitted and prohibited concomitant medications, see the Supplemental Appendix.

Reasons for permanent study drug discontinuation included patient/investigator decision, an adverse event (AE), which the investigator thought put the patient at undue risk, severe noncompliance with study protocol as determined by the investigator that could affect the validity of the data, pregnancy, and if a patient received an organ transplant during the study. Patients who received a course of ESA rescue therapy, and who met criteria for a second course, were to be permanently discontinued from study drug. Patients who discontinued study drug and remained in the study were to continue to undergo prespecified in-person study visits or modified follow-up, such as in-person study visits occurring at different frequencies and/or with fewer study procedures (e.g., without study laboratory assessments), or periodic telephone contact. ESA therapy was not prohibited after permanent study drug discontinuation.

End Points

The primary efficacy end point was the mean change from baseline in Hb averaged over weeks 28–52, regardless of rescue therapy use, in the intent-to-treat (ITT) analysis set, which was assessed for noninferiority (NI). Prespecified subgroup analyses (Supplemental Appendix) were also performed.

Secondary efficacy end points were assessed in a hierarchical, fixed sequence approach, examined in the prespecified analysis sets (ITT, on-treatment [OT] + 3 days, or the per-protocol set) and assessed for NI and/or superiority (Supplemental Table 3). The hierarchical order for the secondary end points was as follows: mean change in Hb from baseline averaged over weeks 28–36 without need for rescue therapy (censoring all Hb values within 6 weeks after rescue therapy; NI); mean change in LDL cholesterol (LDL-C) from baseline to week 24 (superiority); mean change in Hb from baseline averaged over weeks 28–52 in patients with baseline high-sensitivity C-reactive protein (hsCRP) greater than the upper limit of normal (ULN; i.e., 5 mg/l; superiority); proportion of total time with Hb ≥10 g/dl and 10–12 g/dl over weeks 28–52 regardless of rescue therapy use (NI); mean monthly IV iron use from week 36 until end of study (EOS; superiority), and need for first RBC transfusion (NI). If criteria for an end point were not met, formal statistical testing was stopped and nominal P values were reported for subsequent end points.

Exploratory end points included change in serum iron, ferritin, TSAT, total iron binding capacity (TIBC), total cholesterol, HDL cholesterol (HDL-C), and triglycerides from baseline averaged over week 24 until end of treatment (EOT), and change in hepcidin from baseline until week 24 (Supplemental Table 3). Change in LDL-C/HDL-C ratio from baseline until week 24 was determined post hoc. Total cholesterol and HDL-C were captured as laboratory safety variables. Hepcidin and hsCRP values were quantified from stored biomarker specimens. Treatment compliance calculations are described in the Supplemental Appendix.

The primary safety objective was to contribute adjudicated cardiovascular safety data for a separate pooled analysis across the phase 3 roxadustat program for patients receiving dialysis. AEs were assessed in all patients who received ≥1 dose of treatment and censored 28 days after treatment discontinuation (OT + 28 days), and reported according to the AE terms associated with investigator reports. Treatment exposure by dose was also examined (OT + 7 days). Treatment-emergent AEs were spontaneously reported and coded using the Medical Dictionary for Regulatory Activities (MedDRA) version 20.0. AE rates by AE term (per 100 patient-years) were calculated to account for exposure to study treatment as follows: $$\begin{array}{l}[\mathrm{\text{Number of patients with AE}}\\ /(\mathrm{\text{total number of days at risk for that AE}}\\ \mathrm{\text{across all patients in group}}÷ 365.25\left)\right]\times 100\end{array}$$

Patients with more than one event for the same AE category/term were counted once in that category/term. For patients who withdrew consent, public record searches were used by site staff and/or search agency to confirm vital status (alive or dead) at EOS, as appropriate in accordance with local regulations.

Statistical Analysis

A sample size of ≥600 patients was determined sufficient to provide ≥99% power to demonstrate NI of roxadustat versus epoetin alfa for the primary efficacy end point, assuming a difference of −0.30 g/dl with a NI margin of −0.75 g/dl and an SD of 1.25 g/dl. To ensure sufficient adjudicated major adverse cardiac events (defined as a composite of death from any cause, nonfatal myocardial infarction, and nonfatal stroke; target: 615) for the preplanned pooled cardiovascular safety analysis, a population of approximately 2000 patients was prespecified for this study.

Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). For the primary efficacy end point, multiple imputation analysis of covariance (ANCOVA) was used with missing data imputed through a missing at random imputation (Supplemental Appendix). Baseline Hb, cardiovascular/cerebrovascular/thromboembolic history (“yes versus no”), geographical region (US versus ex-US), and incident versus prevalent dialysis (dialysis duration ≤4 months versus >4 months from randomization) were used as covariates in the efficacy analyses. Observed or imputed values up to the EOT visit if treatment was completed, the EOS visit if the patient discontinued treatment, or date of withdrawal of consent, last contact, or death if the patient withdrew consent, was lost to follow-up or died, and imputed values up to death of the patient were used to derive the mean during weeks 28–52. Sensitivity analyses were also performed on the primary efficacy end point. NI was declared if the lower bound of the two-sided 95% confidence interval (CI) of the difference between roxadustat and epoetin alfa exceeded −0.75 g/dl. The null hypothesis for NI in the primary efficacy end point analysis was that the mean change in Hb from baseline averaged over weeks 28–52 in the roxadustat group would be at least 0.75 g/dl less than the change in the epoetin alfa group. A NI margin of −0.75 g/dl was chosen for comparison of roxadustat and epoetin alfa; this is more rigorous than the −1.0 g/dl margin used historically for the peginesatide development program.¹⁵ The choice of other NI margins is outlined in the Supplemental Appendix. For subgroup analyses, the mixed model of repeated measures was used, with baseline Hb, treatment arm, visit, and treatment by visit interaction as factors in the model. NI margins for secondary end points are specified in Supplemental Table 3. ANCOVA was used to assess change from baseline in lipid and iron parameters. For time-to-event analyses, treatments were compared using a Cox proportional hazards model, with baseline Hb as a continuous variable and treatment group and other stratification factors as fixed effects. Analyses of differences in proportions were performed using the approach by Miettinen and Nurminen,¹⁶ with a two-sided 95% CI adjusted for stratification factors. The Wilcoxon rank sum test was used to compare mean monthly IV iron usage.

Results

Patients

Overall, 2941 patients were screened; 2133 patients were randomized to either roxadustat (n=1068) or epoetin alfa (n=1065; Figure 2). The ITT set comprised 1051 patients treated with roxadustat and 1055 with epoetin alfa, with 99.8% receiving either treatment (n=1048 and n=1053, respectively). Between randomization and the EOS, 352 of 1048 (33.5%) patients on roxadustat and 257 of 1053 (24.4%) patients on epoetin alfa prematurely discontinued treatment (hazard ratio on post hoc analysis, 1.54 [95% CI, 1.31 to 1.81]; P<0.001; Figure 2; Supplemental Figure 1). In both groups, patient decision was the most common reason for treatment discontinuation (roxadustat: n=135 [12.8%]; epoetin alfa: n=88 [8.3%]), followed by kidney transplantation (roxadustat: n=63 [6.0%]; epoetin alfa: n=80 [7.6%]). During the study, including the treatment and follow-up periods, 232 (22.1%) and 212 (20.1%) patients died in the roxadustat and epoetin alfa groups, respectively. When including patients found to have died, using a search of public records, there were 247 (23.6%) deaths among roxadustat-treated patients and 231 (21.9%) deaths among epoetin alfa–treated patients. Vital status at EOS was known in 2098 of 2101 (99.9%) patients who received treatment.

Figure 2.: Patient disposition. ^aExcluded from the study due to major Good Clinical Practice violations (i.e., significant deviations from best practice in obtaining or recording the data that might affect the validity of the data), or being phantom patients (not physically existing) due to system technical issues. ^bNeed for more than one cycle of ESA in patients treated with roxadustat. ^cKidney transplant (n=63), investigator decision (n=22), patient moved/relocated (n=14), death (n=6), miscellaneous (n=20). ^dKidney transplant (n=80), investigator decision (n=8), patient moved/relocated (n=18), death (n=14), miscellaneous (n=22). The analysis sets used in this study are detailed in Supplemental Table 3.

At baseline, 89.1% and 10.8% of patients were on hemodialysis and peritoneal dialysis, respectively (Table 1). Cardiovascular history, presence of diabetes mellitus, and dialysis duration before randomization were similar between groups (Table 1). Diabetic kidney disease was the most common cause of CKD in both treatment groups (Table 1).

Table 1. - Baseline characteristics

Baseline Characteristic	Roxadustat (n=1051)	Epoetin Alfa (n=1055)
Age, yr
Mean (SD)	53.5 (15.3)	54.5 (15.0)
Sex, n (%)
Female	426 (40.5)	429 (40.7)
Male	625 (59.5)	626 (59.3)
Race, n (%) a
White	597 (56.8)	598 (56.7)
Asian	208 (19.8)	198 (18.8)
Black or African American	148 (14.1)	158 (15.0)
Other	98 (9.3)	101 (9.6)
Ethnicity, n (%) a
Hispanic or Latino	268 (25.5)	271 (25.7)
Not Hispanic or Latino	783 (74.5)	784 (74.3)
Geographical region, n (%)
United States	385 (36.6)	391 (37.1)
Europe	345 (32.8)	343 (32.5)
Asia and Australia	202 (19.2)	200 (19.0)
Latin America	92 (8.8)	93 (8.8)
Weight, kg
Mean (SD)	75.1 (21.2)	75.1 (19.7)
BMI, kg/m2
Mean (SD)	27.0 (6.8)	26.9 (6.4)
Total time on dialysis, mo, n (%) b
≤4	198 (18.9)	213 (20.2)
>4	852 (81.1)	841 (79.8)
Type of dialysis, n (%)
Peritoneal dialysis	111 (10.6)	117 (11.1)
Hemodialysis	938 (89.4)	938 (88.9)
Duration of dialysis at the time of randomization, current, mo, median (IQR)
Peritoneal dialysis	10.2 (3.9–30.2)	13.5 (4.5–34.5)
Hemodialysis	23.2 (5.8–53.8)	22.0 (5.8–54.2)
ESA use before randomization, n (%)	908 (86.4)	915 (86.7)
Dose of ESA before randomization	n=908	n=915
Overall, as epoetin alfa equivalent, c IU/wk, median, (Q1–Q3)	4500 (2000–8000)	4500 (2000–8000)
Comorbidities, n (%)
Type 2 diabetes	421 (40.1)	423 (40.1)
Coronary artery disease	220 (20.9)	231 (21.9)
Congestive heart failure	185 (17.6)	189 (17.9)
Cerebrovascular d events	89 (8.5)	80 (7.6)
Hypertension	1018 (96.9)	1009 (95.6)
Hb value, g/dl
Median (IQR)	10.2 (9.3–10.9)	10.3 (9.2–11.0)
LDL-C, mg/dl
Median (IQR)	82.0 (59.0–108.0)	82.0 (59.2–109.0)
hsCRP, mg/dl e
Median (IQR)	0.4 (0.2–1.1)	0.4 (0.2–1.0)
hsCRP >0.5 mg/dl (ULN), n (%)	306 (29.1)	319 (30.2)
SBP, mmHg
Mean (SD)	140.8 (16.9)	140.6 (17.2)
DBP, mmHg
Mean (SD)	78.2 (10.1)	77.9 (10.3)
Most likely etiology of CKD, n (%)	n=1046	n=1045
Diabetic nephropathy	342 (32.7)	315 (30.1)
Chronic glomerulonephritis	189 (18.1)	179 (17.1)
Other primary or secondary glomerulonephritis	103 (9.8)	106 (10.1)
Focal segmental glomerulosclerosis	32 (3.1)	21 (2.0)
IgA nephropathy	12 (1.1)	15 (1.4)
Lupus nephritis	8 (0.8)	1 (<0.1)
Membranous nephropathy	6 (0.6)	14 (1.3)
Minimal change	3 (0.3)	4 (0.4)
Ischemic/hypertensive nephropathy	179 (17.1)	204 (19.5)
Cystic kidney disease	51 (4.9)	60 (5.7)
Chronic pyelonephritis (infectious)	45 (4.3)	47 (4.5)
Chronic interstitial nephritis	15 (1.4)	19 (1.8)
Obstructive nephropathy	13 (1.2)	15 (1.4)
Renal artery stenosis	2 (0.2)	0
Unknown	137 (13.1)	121 (11.6)
Other	73 (7.0)	85 (8.1)
Missing	5 (0.5)	10 (1.0)
Not specified	25 (2.4)	18 (1.7)

ITT analysis set, unless specified. BMI, body mass index; MedDRA, Medical Dictionary for Regulatory Activities; Q, quartile; SBP, systolic blood pressure; DBP, diastolic blood pressure.

^aRace/ethnicity were determined by the investigator according to fixed categories in the case report form.

^bTime on dialysis at randomization. Patients on dialysis ≤4 mo were considered incident dialysis patients, and on dialysis >4 mo were considered prevalent dialysis patients.

^cEpoetin alfa equivalents were calculated based on the following equivalence: 60 μg/wk darbepoetin alfa or 180 μg/mo Mircera were equivalent to 12,500 IU/wk.

^dMedical history events, defined by MedDRA version 20.0: ischemic stroke, hemorrhagic stroke, cerebrovascular accident, transient ischemic attack.

^eOT + 28 days analysis set.

Baseline median (interquartile range [IQR]) Hb was 10.2 (9.3–10.9) g/dl in the roxadustat group and 10.3 (9.2–11.0) g/dl in the epoetin alfa group. Among patients with prior ESA use (n=1823 [86.6%]), median (IQR) epoetin alfa equivalent dose in both groups was 4500 (IQR: 2000–8000) IU/wk.

In the roxadustat group, 214 patients were documented to have received ESA after the last dose of the study treatment.

Dosing

Mean (SD) duration of exposure was 20.6 (13.4) months with roxadustat and 23.2 (12.7) months with epoetin alfa. Mean roxadustat dose decreased over the first 52 weeks of treatment and mean epoetin alfa dose was stable (Supplemental Figure 2). The most common roxadustat doses were in the range of 50–150 mg TIW (Supplemental Table 4). Mean (SD) weekly doses of roxadustat and epoetin alfa were 280.60 (49.49) mg and 8656.26 (6491.65) IU, respectively. Overall, 972 (92.7%) patients on roxadustat and 957 (90.9%) patients on epoetin alfa had at least one dose modification. In the roxadustat group, 916 (87.4%) patients had at least one dose reduction and 895 (85.4%) patients had at least one dose increase; the corresponding number of patients in the epoetin alfa group was 841 (79.9%) and 901 (85.6%) patients, respectively. Compliance data were obtained for roxadustat but data for epoetin alfa were not obtained for sites where epoetin alfa was locally supplied (Supplemental Table 5). Significant noncompliance (<50%) was observed in only nine (0.9%) patients receiving roxadustat.

Primary Efficacy End Point: Mean Hb Change

Roxadustat was noninferior to epoetin alfa (P<0.001; Table 2) and was associated with a numerically higher adjusted least squares mean (LSM) increase in Hb from baseline averaged over weeks 28–52 than epoetin alfa (0.77 g/dl [95% CI, 0.69 to 0.85] versus 0.68 g/dl [95% CI, 0.60 to 0.76], respectively; LSM difference, 0.09 g/dl [95% CI, 0.01 to 0.18], P=0.036 for superiority), at the doses selected for comparison. This was consistent across subgroups, including dialysis modality and duration (Figure 3A). Mean Hb levels for both groups rose within the first 12 weeks of treatment and subsequently remained within 10.5–11.0 g/dl for the remainder of the evaluation period (Figure 3B). Primary efficacy results were consistent in the sensitivity analyses performed, including OT analysis (Supplemental Table 6). The results of superiority testing are reported in Supplemental Table 7.

Table 2. - Prespecified efficacy end points, presented in hierarchical order

End Point (Analysis Set)	Roxadustat (n=1051)				Epoetin Alfa (n=1055)				Difference in LSM Changes (95% CI)/HR (95% CI) c	P Value for Superiority	NI P Value
	n	BL Value	Final Value	LSM Change/Adjusted LSM/Mean Monthly Value a /Event Rate b	n	BL Value	Final Value	LSM Change/Adjusted LSM/Mean Monthly Value a /Event Rate b
Change in Hb from BL to mean during weeks 28–52, g/dl (ITT)	1003	10.01	10.78	LSM change: 0.77	1016	10.04	10.72	LSM change: 0.68	Difference in LSM changes: 0.09 (0.01 to 0.18)	–	<0.001 d
Change in Hb from BL to mean during weeks 28–36, g/dl (PPS)	836	9.98	10.86	LSM change: 0.88	864	10.04	10.78	LSM change: 0.74	Difference in LSM changes: 0.14 (0.03 to 0.25)	–	<0.001 d
Mean change in LDL-C from BL to week 24, mg/dl (ITT)	902	87.70	73.16	LSM change: −14.54	937	87.83	86.07	LSM change: −1.76	Difference in LSM changes: −12.79 (−15.08 to −10.49)	<0.001	–
Change of Hb from BL to mean value during weeks 28–52 in patients with BL hsCRP >ULN, g/dl (ITT)	280	10.05	10.85	LSM change: 0.80	301	9.96	10.55	LSM change: 0.59	Difference in LSM changes: 0.20 (0.04 to 0.36)	0.012	–
Proportion of total time of interpolated Hb ≥10 g/dl from week 28–52 e (ITT)	896	NA	NA	Adjusted LSM: 0.79	941	NA	NA	Adjusted LSM: 0.76	Difference in LSM: 0.03 (0.00 to 0.05)	–	<0.001 f
Proportion of total time of interpolated Hb values between 10–12 g/dl from week 28–52 e (ITT)	896	NA	NA	Adjusted LSM: 0.65	941	NA	NA	Adjusted LSM: 0.63	Difference in LSM: 0.02 (−0.01 to 0.05)	–	<0.001 f
Mean monthly IV iron use during week 36 to EOS (ITT)	885	NA	NA	Mean monthly value: 58.71	920	NA	NA	Mean monthly value: 91.37	NA	<0.001	–
Event rate for first RBC transfusion (OT+3)	1048	NA	NA	Event rate: 6.0	1053	NA	NA	Event rate: 7.2	HR: 0.83 (0.64 to 1.07)	–	<0.001 g

NI P value is one-sided. BL, baseline; HR, hazard ratio; PPS, per-protocol set; NA, not applicable; OT+3, OT + 3 days.

^aMean monthly IV iron use in milligrams.

^bRate of RBC transfusion per 100 patient-years.

^cHR (95% CI) comparing risk of RBC transfusion for roxadustat versus epoetin alfa.

^dNI margin is −0.75.

^eProportion of total time of interpolated Hb values ≥10 or 10–12 g/dl was calculated as the time the linearly interpolated curve between measurements was ≥10 or 10–12 g/dl, respectively, divided by the time between measurements from weeks 28–52.

^fNI margin is −0.15.

^gNI margin is 1.8.

Figure 3.: Hemoglobin endpoints. (A) Hb change from baseline to the mean Hb in weeks 28–52 by subgroup. P value for superiority is modeled using ANCOVA with baseline Hb as covariate and CV history, geographical region (US versus ex-US), incident versus prevalent dialysis (≤4 versus >4 months), treatment group, subgroup, and treatment by subgroup interaction as fixed effects. ^aCV/cerebrovascular/thromboembolic history defined by MedDRA version 20.0 dictionary-derived terms: cardiac failure congestive, myocardial infarction, percutaneous coronary intervention, coronary artery bypass, ischemic stroke, hemorrhagic stroke, and cerebrovascular accident. (B) Mean Hb over time to week 164. ITT analysis set. Symbol inside box: mean; symbols outside box: outliers. Whiskers extend to the most extreme observation within 1.5 times the IQR from the nearest quartile, so that all outliers >1.5 times the IQR are individually displayed with clear circles. Baseline Hb is defined as the mean of the last three central laboratory Hb values obtained from screening and randomization. ANCOVA, analysis of covariance; BMI, body mass index; PD, peritoneal dialysis; HD, hemodialysis.

Secondary Efficacy End Points

All secondary efficacy outcomes met the criteria for successful statistical testing using the fixed sequence approach. Roxadustat was noninferior to epoetin alfa for the adjusted LSM change from baseline Hb averaged over weeks 28–36 without rescue therapy (0.88 g/dl [95% CI, 0.79 to 0.97] for roxadustat versus 0.74 g/dl [95% CI, 0.65 to 0.82] for epoetin alfa; LSM difference, 0.14 g/dl [95% CI, 0.03 to 0.25], P<0.001 for NI; Table 2).

To understand the treatment effect among patients with possible inflammation, the mean change in Hb from baseline averaged over weeks 28–52 was studied in patients with elevated hsCRP. Roxadustat resulted in statistically significant greater increases in Hb from baseline averaged over weeks 28–52 in patients with hsCRP >ULN compared with epoetin alfa (LSM difference, 0.20 g/dl [95% CI, 0.04 to 0.36]; P=0.012 for superiority; Table 2). This increase in Hb with roxadustat was similar to the increase observed in the primary end point for the overall roxadustat-treated population.

Roxadustat was noninferior to epoetin alfa for the proportion of total time with Hb ≥10 g/dl and 10–12 g/dl over weeks 28–52 (P<0.001 for NI for both, Table 2). The proportion of time patients treated with roxadustat had an Hb ≥10 g/dl from weeks 28–52 was numerically greater compared with those treated with epoetin alfa (0.79 [95% CI, 0.76 to 0.81] versus 0.76 [95% CI, 0.73 to 0.78], respectively; LSM difference, 0.03 [95% CI, 0.00 to 0.05]; Table 2).

Low-Density Lipoprotein Cholesterol

Patients receiving roxadustat had an LSM reduction in LDL-C from baseline to week 24 of 14.54 mg/dl. A statistically significant greater reduction in LDL-C from baseline to week 24 was observed with roxadustat, compared with epoetin alfa (LSM difference, −12.79 mg/dl [95% CI, −15.08 to −10.49]; P<0.001 for superiority; Table 2).

Iron Use and Rescue Therapy

Mean monthly IV iron use from week 36 until EOS was significantly lower with roxadustat versus epoetin alfa (58.71 mg versus 91.37 mg, respectively; P<0.001 for superiority; Table 2). Concomitant oral iron use during the study was similar (roxadustat: 218 of 1051 [20.7%]; epoetin alfa: 190 of 1055 [18.0%]).

Roxadustat was noninferior to epoetin alfa for use of first RBC transfusion (hazard ratio of roxadustat to epoetin alfa, 0.83 [95% CI, 0.64 to 1.07]; P<0.001 for NI; Figure 4, Table 2). RBC transfusion was received by 103 (9.8%) and 139 (13.2%) roxadustat- and epoetin alfa–treated patients, respectively.

Figure 4.: Time-to-first administration of RBC transfusion. RBC transfusion was allowed if rapid correction of anemia was required or if deemed a medical necessity by the investigator. On-treatment + 3 days analysis set. HR, hazard ratio.

Overall, 132 (12.6%) patients on roxadustat and 139 (13.2%) patients on epoetin alfa received ≥1 rescue therapy. In addition to RBC transfusion as rescue therapy reported above, ESA rescue was administered to 39 (3.7%) patients on roxadustat and two (0.2%) patients on epoetin alfa.

Exploratory End Points: Hepcidin and Markers of Iron Metabolism

A reduction in serum hepcidin from baseline to week 24 was observed with roxadustat versus epoetin alfa (LSM difference, −28.21 ng/ml [95% CI, −41.98 to −14.45]; Table 3). A reduction in ferritin was observed in the roxadustat and epoetin alfa groups, with a greater reduction observed with roxadustat (Figure 5A, Supplemental Figure 3A, Table 3). Roxadustat increased serum iron and TIBC from baseline and this change was maintained over 52 weeks (Figure 5, B and C, Supplemental Figure 3, B and C, Table 3). TSAT was unchanged in both treatment groups (Figure 5D, Supplemental Figure 3D, Table 3).

Table 3. - Hepcidin and iron profile exploratory end points

End Point a	Roxadustat (n=1051)			Epoetin Alfa (n=1055)			LSM Treatment Difference (95% CI)	P Value
	n	Mean Baseline	Adjusted LSM Change (95% CI)	n	Mean Baseline	Adjusted LSM Change (95% CI)
Hepcidin, ng/ml	608	275.61	−44.99 (−57.52 to −32.46)	625	269.62	−16.77 (−29.15 to −4.40)	−28.21 (−41.98 to −14.45)	<0.001
Serum iron, µg/dl	877	75.11	6.58 (4.47 to 8.69)	946	73.35	−5.54 (−7.57 to −3.50)	12.12 (9.78 to 14.46)	<0.001
Serum ferritin, µg/l	875	542.96	−104.47 (−126.16 to −82.77)	946	555.78	−41.18 (−62.09 to −20.27)	−63.29 (−87.38 to −39.19)	<0.001
TSAT, %	866	36.00	−1.92 (−2.78 to −1.06)	939	34.95	−2.44 (−3.27 to −1.61)	0.52 (−0.44 to 1.48)	0.287
TIBC, µg/dl	874	208.16	34.98 (31.77 to 38.18)	946	208.64	−2.41 (−5.50 to 0.68)	37.38 (33.82 to 40.95)	<0.001

Hepcidin was quantified from stored biomarker specimens obtained at baseline and week 24. Mean change from baseline was analyzed using an ANCOVA model with baseline value and baseline Hb as covariates and cardiovascular/cerebrovascular/thromboembolic history, geographical region (US versus ex-US), incident versus prevalent dialysis (≤4 versus > 4 mo) and treatment group as fixed effects. TSAT was calculated as: TSAT (%) = (serum iron level × 100)/TIBC. Intent-to-treat analysis set.

^aLSM change from baseline to week 24 for hepcidin; LSM change from baseline to mean during week 24 to EOT for iron, ferritin, TSAT, and TIBC.

Figure 5.: Serum iron parameters by visit. Mean iron levels for (A) ferritin, (B) iron, (C) TIBC, and (D) TSAT. Intent-to-treat analysis set. Error bars are 95% CIs. Baseline is defined as the last measurement before randomization. 95% CI of the mean is based on the normal distribution.

Exploratory End Points: Lipid Parameters

A reduction in total cholesterol, HDL-C, and triglycerides from baseline to the mean over week 24 to EOT was observed in both treatment groups, with a greater reduction observed with roxadustat versus epoetin alfa (Supplemental Table 8). Similarly, the ratio of LDL-C/HDL-C was lowered to a greater extent with roxadustat versus epoetin alfa (Supplemental Table 8).

Exploratory End Points: Hb before Early Treatment Discontinuation

Among patients with early treatment discontinuation, the mean (SD) of the most recent Hb before discontinuation was 10.3 (1.5) g/dl and 10.4 (1.5) g/dl in the roxadustat and epoetin alfa groups, respectively.

Adverse Events

The proportion of patients experiencing ≥1 AE with roxadustat compared with epoetin alfa was 85.0% and 84.5%, respectively; corresponding event rates per 100 patient-years were 168.2 and 132.5, respectively (Table 4). The most commonly reported AEs in both treatment groups were diarrhea, hypertension, and pneumonia (Table 5). No individual malignancies were identified as common AEs. Differences in event rates between groups were <2.0 events per 100 patient-years for each reported AE (Table 5). However, a greater proportion of patients experienced arteriovenous fistula thrombosis (7.4% versus 5.4%) and arteriovenous fistula site complication (5.4% versus 3.5%) with roxadustat compared with epoetin alfa, respectively.

Table 4. - Adverse events summary

AE Category	Roxadustat (n=1048)			Epoetin Alfa (n=1053)
	N Pts w/Event	%	N Per 100 P-Y	N Pts w/Event	%	N Per 100 P-Y
Any AE	891	85.0	168.2	890	84.5	132.5
Any AE leading to discontinuation of drug	57	5.4	3.1	26	2.5	1.2
Any AE leading to interruption of drug	97	9.3	5.5	44	4.2	2.2
Any SAE (including events with an outcome of death)	604	57.6	49.2	606	57.5	43.5
Any AE in a patient with an outcome of death	167	15.9	9.0	187	17.8	8.9
Any AE in the cardiac disorders SOC a	245	23.4	15.0	277	26.3	15.3
Any SAE in the cardiac disorders SOC a	153	14.6	8.7	169	16.0	8.6

On-treatment + 28 days analysis set. Any AE that was not present before the first dose of the study treatment and that occurred during study treatment or within 28 days of the last dose of study drug was deemed treatment emergent. Patients with multiple events in the same category were counted only once in that category. Patients with events in more than one category were counted once in each of those categories. Pts, patients; w/, with; P-Y, patient-years; SOC, system organ class.

^aAnalyses performed post hoc.

Table 5. - AEs ≥5% incidence in either treatment group by AE term

AE Category	Roxadustat (n=1048)			Epoetin Alfa (n=1053)
	N Pts w/Event	%	N Per 100 P-Y	N Pts w/Event	%	N Per 100 P-Y
Diarrhea	117	11.2	6.8	107	10.2	5.5
Hypertension	92	8.8	5.3	94	8.9	4.8
Pneumonia	91	8.7	5.1	101	9.6	5.0
Headache	82	7.8	4.7	57	5.4	2.8
AV fistula thrombosis	78	7.4	4.4	57	5.4	2.8
Hypotension	75	7.2	4.2	69	6.6	3.4
Upper RTI	72	6.9	4.1	56	5.3	2.8
Vomiting	62	5.9	3.5	55	5.2	2.7
Nausea	61	5.8	3.4	65	6.2	3.2
Cough	61	5.8	3.4	62	5.9	3.1
AV fistula site complication	57	5.4	3.2	37	3.5	1.8
Pyrexia	55	5.2	3.1	48	4.6	2.4
Dyspnea	51	4.9	2.8	59	5.6	2.9
Fluid overload	51	4.9	2.8	53	5.0	2.6
Bronchitis	47	4.5	2.6	66	6.3	3.3
Atrial fibrillation	31	3.0	1.7	59	5.6	2.9

On-treatment + 28 days analysis set. AEs were reported by investigators and AE terms were assigned to these events according to MedDRA version 20.0. AV, arteriovenous; Pts, patients; w/, with; P-Y, patient-years; RTI, respiratory tract infection.

The proportion of patients experiencing ≥1 serious AE (SAE) was 57.6% with roxadustat and 57.5% with epoetin alfa; event rates per 100 patient-years were 49.2 and 43.5 for roxadustat and epoetin alfa, respectively (Table 4). The most commonly reported SAEs in both groups were pneumonia, sepsis, and acute myocardial infarction (Table 6; a full list of SAEs can be found on ClinicalTrials.gov).¹⁷ Deep vein thrombosis was reported for 11 (1.0%) patients receiving roxadustat, compared with 0 (0.0%) patients receiving epoetin alfa. Seizure was reported for 13 (1.2%) patients in the roxadustat group and 7 (0.7%) patients in the epoetin alfa group (Table 6).

Table 6. - SAEs ≥1% in either treatment group by system organ class and SAE term

AE Category	Roxadustat (n=1048)			Epoetin Alfa (n=1053)
	N Pts w/Event	%	N Per 100 P-Y	N Pts w/Event	%	N Per 100 P-Y
Infections and infestations
Pneumonia	70	6.7	3.9	78	7.4	3.8
Sepsis	40	3.8	2.2	40	3.8	1.9
Peritonitis	24	2.3	1.3	24	2.3	1.2
Cellulitis	12	1.1	0.7	16	1.5	0.8
Septic shock	12	1.1	0.6	13	1.2	0.6
Device-related infection	11	1.0	0.6	11	1.0	0.5
Gastroenteritis	11	1.0	0.6	6	0.6	0.3
Gangrene	9	0.9	0.5	11	1.0	0.5
Urinary tract infection	8	0.8	0.4	13	1.2	0.6
Metabolism and nutrition disorders
Fluid overload	32	3.1	1.7	29	2.8	1.4
Hyperkalemia	27	2.6	1.5	21	2.0	1.0
Hypoglycemia	19	1.8	1.0	11	1.0	0.5
Nervous system disorders
Cerebrovascular accident	14	1.3	0.8	12	1.1	0.6
Seizure	13	1.2	0.7	7	0.7	0.3
Ischemic stroke	10	1.0	0.5	11	1.0	0.5
Syncope	5	0.5	0.3	11	1.0	0.5
Cardiac disorders
Acute myocardial infarction	39	3.7	2.1	41	3.9	2.0
Cardiac failure congestive	24	2.3	1.3	29	2.8	1.4
Coronary artery disease	17	1.6	0.9	16	1.5	0.8
Myocardial infarction	14	1.3	0.8	6	0.6	0.3
Cardiac failure	13	1.2	0.7	10	0.9	0.5
Atrial fibrillation	10	1.0	0.5	18	1.7	0.9
Cardiac failure acute	10	1.0	0.5	5	0.5	0.2
Vascular disorders
Hypertensive crisis	24	2.3	1.3	36	3.4	1.8
Hypertensive emergency	24	2.3	1.3	31	2.9	1.5
Hypotension	17	1.6	0.9	18	1.7	0.9
Hypertension	12	1.1	0.7	12	1.1	0.6
Deep vein thrombosis	11	1.0	0.6	0	0.0	0.0
Injury, poisoning and procedural complications
Arteriovenous fistula thrombosis	37	3.5	2.0	31	2.9	1.5
Vascular access site thrombosis	16	1.5	0.9	13	1.2	0.6
Respiratory, thoracic and mediastinal disorders
Pleural effusion	15	1.4	0.8	14	1.3	0.7
Acute respiratory failure	13	1.2	0.7	9	0.9	0.4
Pulmonary edema	10	1.0	0.5	20	1.9	1.0
Gastrointestinal disorders
Gastrointestinal hemorrhage	22	2.1	1.2	21	2.0	1.0
Gastritis	11	1.0	0.6	5	0.5	0.2
Abdominal pain	10	1.0	0.5	10	0.9	0.5
General disorders and administrative site conditions
Death	21	2.0	1.1	23	2.2	1.1
Noncardiac chest pain	12	1.1	0.7	9	0.9	0.4
Renal and urinary disorders
Azotemia	13	1.2	0.7	11	1.0	0.5

On-treatment + 28 days analysis set. SAEs were reported by investigators and SAE terms were assigned to these events according to MedDRA version 20.0. All SAEs are listed in the ClinicalTrials.gov listing for this study (https://clinicaltrials.gov/ct2/show/results/NCT02174731). Pts, patients; w/, with; P-Y, patient-years.

In an overview of cardiac AEs, the proportion of roxadustat- and epoetin alfa–treated patients with AEs was 23.4% versus 26.3%, respectively; corresponding event rates per 100 patient-years were 15.0 and 15.3 for roxadustat and epoetin alfa, respectively (Table 4). The proportion of patients experiencing cardiac SAEs was 14.6% versus 16.0%, respectively, in the cardiac disorders system organ class; event rates per 100 patient-years were 8.7 and 8.6 for roxadustat and epoetin alfa, respectively (Table 4). The complete listing of SAEs in the cardiac disorders system organ class is provided in Supplemental Table 9.

The most common AE leading to discontinuation in the roxadustat group, and the event with the largest imbalance as the cause of study drug discontinuation for roxadustat versus epoetin alfa, was nausea (n=5 for roxadustat and n=0 for epoetin alfa; corresponding event rates per 100 patient-years were 0.27 and 0.00 for roxadustat and epoetin alfa, respectively). One patient in the roxadustat group, and no patients in the epoetin alfa group, discontinued due to vomiting, and one patient in the epoetin alfa and no patients in the roxadustat group discontinued due to decreased appetite. A full list of AEs leading to discontinuation is presented in Supplemental Table 10.

From week 28 until EOT, no clinically significant changes in vital signs were observed (Supplemental Table 11).

During the treatment period, distribution of serum potassium levels was comparable between treatment groups, with 42.6% and 45.8% of roxadustat- and epoetin alfa–treated patients, respectively, having at least one serum potassium value of ≥6.0 mmol/l (Supplemental Table 12). Overall mean change in serum potassium from baseline was 0.07 mmol/l for roxadustat and 0.05 mmol/l for epoetin alfa.

Discussion

The phase 3 clinical development program for roxadustat in the United States for patients with anemia on dialysis consists of three studies (ROCKIES [NCT02174731], SIERRAS [NCT02273726], and HIMALAYAS [NCT02052310]), and includes approximately 3900 patients. ROCKIES is the largest study within this program. In this study, roxadustat resulted in a noninferior increase in Hb compared with epoetin alfa as used in patients with CKD on dialysis. These findings support the efficacy of roxadustat versus epoetin alfa previously observed in phase 3 trials.^12,18

Analyses of efficacy in increasing Hb across various patient subgroups were generally consistent with the primary efficacy end point, although analysis of some subgroups was limited by small sample size. This study showed that roxadustat was efficacious in patients with inflammation. Roxadustat may be a useful tool in the management of patients with inflammation, which is a common cause of ESA hyporesponsiveness.^192021–22

IV iron use was significantly reduced with roxadustat even though similar proportions of patients received oral iron in both groups. Protocolized differences in treatment strategies between the roxadustat and epoetin alfa groups may have influenced IV iron use, and so IV iron use should be interpreted in this context. As treated in this study, oral roxadustat effectively increased and maintained Hb levels despite lower doses of IV iron. Consistent with roxadustat’s mode of action, hepcidin was decreased with roxadustat compared with epoetin alfa. Despite restrictions on the use of IV iron in roxadustat-treated patients, serum iron and TIBC increased among these patients.^18,23 Reduced ferritin was also noted with roxadustat, which could be reflective of iron utilization for erythropoiesis.¹⁰ Higher rates of ESA rescue therapy in the roxadustat group were observed, and this would be expected as hyporesponsiveness to epoetin alfa in the epoetin alfa group may be more likely to be addressed by a dose increase, rather than “rescue” with a different ESA.

The majority of the AEs identified in this study were consistent with the AEs expected in patients with CKD on dialysis, such as pneumonia, hypoglycemia, peritonitis, and cardiovascular events.^2425–26 Although overall event rates of AEs and SAEs were numerically higher with roxadustat versus epoetin alfa, rates of individual and common AEs were generally similar between treatment groups. Serious noncardiovascular AEs that appeared to be more common with roxadustat than epoetin alfa in this study included gastroenteritis, hypoglycemia, seizure, and hyperkalemia. Concerns have been raised previously that hyperkalemia is related to HIF-PH inhibitor use.¹² Although an imbalance in the rate of hyperkalemia reported as an AE by investigators was observed, the incidence of the objective finding of abnormal potassium values was comparable between treatment groups, suggesting that comparison of hyperkalemia risk using investigator-reported event rates may be less reliable.

AEs leading to discontinuation were more common among roxadustat-treated patients, but the events were generally reflective of the disease burden in patients with CKD on dialysis, and no clustering of AEs was identified. The most common AE leading to discontinuation in the roxadustat group was nausea (n=5).

Because of the increased cardiac risk in this patient population, event rates for AEs and SAEs in the cardiac disorders system organ class were assessed. Rates were generally comparable between roxadustat and epoetin alfa, both overall and for individual cardiac SAEs. Seizure was reported in more patients treated with roxadustat compared with those treated with epoetin alfa, and there was an increase in the rate of deep vein thrombosis and vascular access thrombosis with roxadustat compared with epoetin alfa in the ROCKIES study. However, ROCKIES by design was not powered to assess cardiovascular safety, and AE and SAE data are reported using the AE terms associated with investigator reports. The roxadustat program was designed to assess the cardiovascular safety of roxadustat in a pooled analysis of three studies in patients on dialysis.

A proposed advantage of roxadustat is its oral mode of administration; this may be preferred over parenteral medication in a home dialysis setting or for patients not on dialysis. However, real-world adherence to roxadustat has not yet been assessed, and the effect of changing from a parenteral to oral medication on adherence in this patient population requires further investigation.

This study has limitations. Although the roxadustat dosing algorithm was the same for all participating sites, epoetin alfa dosing and Hb targets varied according to the local standard of care and prescribing information, which may have contributed to the greater increases in mean Hb values in the roxadustat group. The route of administration of epoetin alfa has been reported to affect its potency; doses 25% higher are needed when given IV compared with subcutaneous to achieve equivalent Hb response.²⁷ However, data regarding route of ESA administration were not routinely collected as part of this study. Secondly, the open-label trial design comparing a novel, orally administered agent to a parenteral standard of care used by most patients at baseline may have contributed to differences in study drug discontinuation between treatment groups, and potentially led to differences in AE reporting.²⁸ For instance, patients receiving epoetin alfa may have been less likely to discontinue study treatment and revert to receiving nonstudy ESA. Indeed, there were higher rates of study drug discontinuation and AEs leading to discontinuation, and lower exposure in the roxadustat group than the epoetin alfa group. Nonetheless, common AEs for both treatment groups were those common in dialysis patients, roxadustat was noninferior in prespecified Hb end points, and Hb values before discontinuation were similar between groups, suggesting that differences in treatment discontinuation may not be explained by differences in efficacy or AE profiles between roxadustat and epoetin alfa. Finally, the majority of patients were ESA-exposed, baseline ESA doses were generally low, and the proportion of patients with hsCRP >ULN was lower than reported in observational studies of dialysis-dependent populations,²⁹ suggesting the population in ROCKIES may be more likely to have been ESA responsive.

In conclusion, among patients with anemia of CKD on dialysis, treatment with oral roxadustat increased Hb as effectively as epoetin alfa, with reduced IV iron and noninferior reduction in RBC transfusion, according to the present algorithm of administration. The AE profile of roxadustat was comparable with epoetin alfa in this study; however, the roxadustat global phase 3 program was designed to assess cardiovascular safety using a pooled analysis of similarly designed trials, including ROCKIES. Overall, the findings of the ROCKIES study suggest that treatment protocols with roxadustat may be an acceptable option for the treatment of anemia of CKD in patients on dialysis.

Disclosures

M. El-Shahawy received research support and consulting fees from AstraZeneca; reports ownership interest with East LA Dialysis Center and Paramount Hope Dialysis Center; reports research funding from Bayer, Pfizer, Sanofi, and UCB; reports honoraria from AstraZeneca; reports scientific advisor or membership with AstraZeneca and Bayer; and is on the speakers bureau for AstraZeneca. E.T. Escudero received consulting fees from AstraZeneca and FibroGen. S. Fishbane received research support and consulting fees from AstraZeneca; received research support from Akebia, Ardelyx, Cara Therapeutics, Corvidia Therapeutics, Gilead Sciences, FibroGen, and MegaPro Biomedical; reports consultancy agreements with Akebia, Cara Therapeutics, and FibroGen; reports honoraria from Akebia and AstraZeneca; and reports scientific advisor or membership with AstraZeneca and FibroGen. P.E. Pergola received research support and consulting fees from Akebia, Ardelyx, AstraZeneca, Bayer, Corvidia Therapeutics, FibroGen, Gilead Sciences, Otsuka, Reata Pharmaceuticals, Tricida, and Unicycive Therapeutics; reports ownership interest with Unicycive Therapeutics; reports research funding as prinicipal investigator or sub-investigator on multiple clinical trials with his practice (not as an individual); reports scientific advisor or membership with Ardelyx and Unicycive Therapeutics; and is on the speakers bureau for AstraZeneca. C.A. Pollock has received consulting fees from AstraZeneca, Eli Lilly, FibroGen, Johnson & Johnson, Janssen-Cilag, Novartis, and Otsuka Pharmaceutical; reports honoraria with Amgen, AstraZeneca, Boehringer Ingelheim, Eli Lilly, Merck Sharp & Dohme, Novartis, Otsuka, Sanofi, and Vifor; reports copyright as book editor with Elsevier; reports scientific advisor or membership with AstraZeneca, Boehringer Ingelheim, Eli Lilly, Janssen-Cilag, Merck Sharp & Dohme, Novartis, Otsuka Pharmaceutical, Pharmaxis, and Vifor; is on the speakers bureau for AstraZeneca, Janssen-Cilag, Otsuka Pharmaceutical, and Vifor; and has other interests/relationships: chair of Kidney Health Australia, chair of the New South Wales Bureau of Health Information, deputy chair of Australian Organ Tissue and Transplant Authority, chair of NSW Cardiovascular Research Network, and director of Certa Therapeutics. A. Rastogi has received research support and consulting fees from AstraZeneca; reports research support from GlaxoSmithKline; reports consultancy agreements with Akebia, Ardelyx, Aurinia, Chiesi, Fresenius Medical Care-Vifor, GlaxoSmithKline, Sanofi, and Vifor (Relypsa); reports research funding from Alnylam, Bayer, Gilead Sciences, Idorsia, Kadmon, National Institutes of Health, Novo Nordisk, Omeros, Pfizer, Protalix, Reata, Sanofi, and Summit; reports scientific advisor or membership with AstraZeneca, Akebia, Ardelyx, Aurinia, Chiesi, Fresenius Medical Care-Vifor, GlaxoSmithKline, Sanofi, and Vifor (Relypsa); and is on the speakers bureau for Amgen, AstraZeneca, Aurinia, Baxter, Fresenius Medical Care-Vifor, Janssen, Natera, Sanofi, Spire Learning, Tricida, and Vifor (Relypsa). B.P. Van received research support and consulting fees from AstraZeneca; reports honoraria from Astellas, AstraZeneca, Boehringer Ingelheim, DKSH, Kalbe International, Nanogen (Vietnam), Otsuka, Pfizer, Servier, and Tedis; reports scientific advisor or membership with Nguyen Tri Phuong University Hospital (Vietnam); and is on the speakers bureau for Astellas, AstraZeneca, Boehringer Ingelheim, DKSH, Kalbe International, Otsuka, Pfizer, Sanofi, Servier, and Tedis. L. Frison, N. Guzman, D.J. Little, and M. Pola are employees and shareholders of AstraZeneca. M. Houser is a former employee and shareholder of AstraZeneca. D. Little reports being a volunteer nephrologist at Walter Reed National Military Medical Center.

Funding

This study was supported by AstraZeneca (grant number not applicable).