Roxadustat (FG-4592): Correction of Anemia in Incident Dialysis Patients

Abstract

Safety concerns with erythropoietin analogues and intravenous (IV) iron for treatment of anemia in CKD necessitate development of safer therapies. Roxadustat (FG-4592) is an orally bioavailable hypoxia-inducible factor (HIF) prolyl hydroxylase inhibitor that promotes coordinated erythropoiesis through HIF-mediated transcription. We performed an open-label, randomized hemoglobin (Hb) correction study in anemic (Hb≤10.0 g/dl) patients incident to hemodialysis (HD) or peritoneal dialysis (PD). Sixty patients received no iron, oral iron, or IV iron while treated with roxadustat for 12 weeks. Mean±SD baseline Hb was 8.3±1.0 g/dl in enrolled patients. Roxadustat at titrated doses increased mean Hb by ≥2.0 g/dl within 7 weeks regardless of baseline iron repletion status, C-reactive protein level, iron regimen, or dialysis modality. Mean±SEM maximal change in Hb from baseline (ΔHb(max)), the primary endpoint, was 3.1±0.2 g/dl over 12 weeks in efficacy-evaluable patients (n=55). In groups receiving oral or IV iron, ΔHb(max) was similar and larger than in the no-iron group. Hb response (increase in Hb of ≥1.0 g/dl from baseline) was achieved in 96% of efficacy-evaluable patients. Mean serum hepcidin decreased significantly 4 weeks into study: by 80% in HD patients receiving no iron (n=22), 52% in HD and PD patients receiving oral iron (n=21), and 41% in HD patients receiving IV iron (n=9). In summary, roxadustat was well tolerated and corrected anemia in incident HD and PD patients, regardless of baseline iron repletion status or C-reactive protein level and with oral or IV iron supplementation; it also reduced serum hepcidin levels.

Keywords: anemia; chronic kidney disease; dialysis; erythropoietin.

Figure 1.

Patient disposition. Screened refers to the inclusion and exclusion criteria listed under "Selection of Study Population" in the Supplemental Material. FU, follow-up period. *The adverse events were dizziness in the no-iron cohorts and worsening gastritis in the IV iron cohort. ^Recovery of renal function. Most study patients (54 of 60 [90.0%]) were enrolled in Russia. The most common causes of renal failure (all patients combined) were unspecified GN (41.7%), pyelonephritis (21.7%), hypertensive nephropathy (18.3%), and diabetic nephropathy (15.0%).

Figure 2.

Mean hemoglobin levels over time are similar through 7 weeks for all treatment groups and thereafter lower for the no-iron vs oral or IV iron groups. Data are for the EE population using last-observation-carried-forward imputation for missing data and are expressed as the mean±SEM Hb value at each time point. Week 0 (baseline) is the mean of three predosing Hb values. *P<0.05 in comparisons between no-iron cohort to the pooled iron cohorts based on the repeated-measures analysis of covariance model with baseline Hb and iron repletion status as covariates, using all observed data collected during treatment.

Figure 3.

Cumulative Hb responses are similar among the cohorts. Data are for the EE population. Hb response was defined as the first Hb increase from baseline of ≥1.0 g/dl. Median time to first response was 3.0 weeks in all cohorts.

Figure 4.

ΔHb_max during last 10 weeks of treatment (g/dl) is independent of baseline CRP level. Data are for the EE population. Baseline CRP was defined as the last value before the first dose administration. LR, linear regression on log(CRP) with adjustment of treatment cohort, baseline Hb, and baseline iron repletion. ΔHb_max versus log(baseline CRP): regression slope on log(CRP) = −0.31 (P=0.30). All patients with CRP greater than the upper limit of normal (ULN=5 g/ml) responded to treatment.

Figure 5.

Weekly roxadustat dose during the last 2 weeks of the treatment period is independent of baseline CRP level. Data are for the EE population. Baseline CRP was defined as the last value before the first dose administration. LR, linear regression on log(CRP) with adjustment of treatment cohort, baseline Hb, and baseline iron repletion. Maintenance dose was defined as total dose during last 2 weeks of treatment. Maintenance dose versus log (baseline CRP): regression slope on log(baseline CRP) = −0.43 (P=0.38). ULN, upper limit of normal (5 ng/ml).