The Search for the Perfect Agent for Anemia Management in Chronic Kidney Disease

Erythropoiesis-stimulating agents (ESAs) are generally effective and relatively safe in correcting the anemia of CKD. However, maintaining hemoglobin (Hb) levels and providing maximal benefit can be challenging since nearly 5%–10% of the patients are hyporesponsive to treatment. Considering that high ESA doses could increase risk of venous and arterial thrombotic events, mortality, and possibly cancer progression, improving safety and overcoming hyporesponsiveness are the main reasons for developing new drugs to treat anemia associated with CKD. Hypoxia-inducible factor (HIF) prolyl hydroxylase inhibitors (HIF-PHIs) are a category of new drug candidates that mimic the response of the body to moderate hypoxia by increasing the stability of the HIF alfa subunit. It is followed by an increased production of endogenous erythropoietin (EPO) in the kidneys and, to a lesser extent, in the liver. The HIF system regulates several other pathways, including iron absorption and mobilization, angiogenesis, lipid and glucose metabolism, glycolysis, mitochondrial function, inflammation and immunity, cell growth and survival, vasodilation, and cell migration.1 In this issue of JASN, Fishbane et al. report the findings of the Safety and Efficacy Study of Roxadustat to Treat Anemia in Patients With Chronic Kidney Disease, on Dialysis (ROCKIES) study.2 This is an international, open-label, phase 3 trial testing the effects of roxadustat, an oral HIF-PHI, in 2133 dialysis-dependent (DD) patients. As shown in similar trials, roxadustat was found to be noninferior compared with epoetin alfa on the mean Hb change from baseline. Safety was a secondary endpoint, as the primary safety objective was to contribute to adjudicated cardiovascular safety endpoints for separate pooled analyses of three phase 3 trials (including ROCKIES). These data have not been published yet. Even if the analysis of ROCKIES2 was just descriptive, roxadustat failed to show a better safety profile than the comparator ESA.1 Despite a similar percentage of adverse events in the two groups, the patients randomized to roxadustat had an increased rate of deep vein and vascular access thrombosis (Figure 1). This is an unexpected finding considering that HIF-PHIs do not usually expose patients to high EPO peaks. A possible explanation for the higher thrombotic risk associated with roxadustat could be a relatively higher dose selected for the comparison with epoetin alfa. Indeed, Hb levels were higher in the roxadustat group throughout this and other trials, especially during the first weeks of treatment. A meta-analysis by the Food and Drug Administration (FDA) of the roxadustat trials showed a weak association between roxadustat dose and the probability of thrombotic events in DD patients; no clear association was shown for epoetin alfa instead.3 Considering that thrombosis has been associated with hypoxia, a possible pathogenic mechanism linking prolyl hydroxylase inhibition to thrombosis cannot be excluded.4 In ROCKIES,2 the roxadustat- and epoetin alfa-treated patients had similar percentages of cardiac adverse events. A meta-analysis of a mixed population of prevalent and incident DD patients showed that roxadustat was noninferior to ESA for risks for major adverse cardiovascular events (MACE) (Figure 1), MACE plus unstable angina or chronic heart failure requiring hospitalization, and chronic heart failure (MACE+).5 However, the prevalent patients randomized to roxadustat had a slightly elevated hazard ratio for time to MACE and MACE+, reaching statistical significance in the case of all-cause mortality. The FDA's meta-analysis also confirmed noninferiority for MACE risk in the specific case of DD patients when censoring events at the end of treatment plus seven days; however, it showed a trend toward increased risk for MACE in those receiving roxadustat in comparison to epoetin alfa in the on-study sensitivity analysis. The interpretation of these findings is complex. As explained in the FDA's meta-analysis, the choice of the ascertainment window could influence the likelihood of detecting adverse events depending on their latency and possible relationship with true drug exposure. Of note, in the DD population, mean drug exposure was just slightly shorter for roxadustat compared with epoetin alfa.3 It is also possible that particular characteristics of the patients who are incident to dialysis treatment may differentiate the hypothetical risk of MACE or MACE+ when receiving roxadustat or an ESA. Of note, baseline Hb levels, together with the proportion achieving Hb response, were lower in this subgroup.5 The Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT) trial6 and the non-dialysis patients enrolled in the Efficacy and Safety Study to Evaluate Vadadustat for the Correction of Anemia in Subjects With Non-dialysis-dependent Chronic Kidney Disease (NDD-CKD) (PRO2TECT) trial with vadadustat7 showed a different geographical and racial distribution, with different practice patterns, which could have also influenced the findings. The Hb target (10–12 g/dl) was consistent in the case of the roxadustat trials, differing from the PRO2TECT7 and TREAT trials.6 Among ancillary effects, HIF-PHIs also have the theoretical advantage of efficacy in patients with inflammation (Figure 1).8 In ROCKIES,2 among the patients with high-sensitive C-reactive protein above the upper limit of normality, those who were randomized to receive roxadustat had significantly greater Hb increases from baseline in comparison to those receiving epoetin alfa. This result could be clinically relevant considering that the cutoff for defining patients as having inflammation was not very high. However, a relative higher potency of the roxadustat dose in respect to that of epoetin alfa in the comparison could explain the difference. Unfortunately, data on roxadustat and epoetin alfa doses, and iron requirement, in patients with and without inflammation, were not reported.2 Better iron absorption and, above all, mobilization could be an additional beneficial ancillary effect of roxadustat for overcoming hyporesponsiveness. Several observations, including an exploratory analysis of the ROCKIES2 findings, go in this direction (Figure 1). However, iron protocols are not standardized in the trials, precluding definitive conclusions. Another off-target effect of roxadustat is the decrease of low-density lipoprotein3 (and, to a lesser extent, of high-density lipoprotein cholesterol) (Figure 1).8 However, this could translate into a clinical advantage more in the earlier stages of CKD since in dialysis patients, atherosclerotic complications are more due to vascular calcifications than to cholesterol deposition.Figure 1.: Roxadustat use in the dialysis-dependent population. Single boxes represent possible benefits, room for further investigation, and possible safety signals.At present, no definitive information is available on other positive or negative pleiotropic effects, including those on cancer, due to a relatively short follow-up. In addition to safety and efficacy, the potential advantages of oral administration deserve a comment. In hemodialysis patients, ESAs are usually given intravenously, directly in the line of the dialyzer, and an extra pill could impact treatment adherence. Oral administration may then benefit more patients on peritoneal dialysis, home hemodialysis, or with CKD not on dialysis. In ROCKIES,2 33.5% of the patients on roxadustat and 24.4% on epoetin alfa prematurely discontinued the treatment. Even if the patient's decision was the most common reason (roxadustat, 12.8%; epoetin alfa, 8.3%), significant noncompliance was observed in only nine (0.9%) patients receiving roxadustat.3 Nausea, the most common adverse event leading to roxadustat discontinuation, was present in only five patients. The open-label trial design comparing a novel, orally administered agent to the parenteral standard-of-care used by most patients at baseline should be considered for its positive or negative effect on adherence. However, the reasons why patients could favor a given treatment instead of another are often neglected, poorly investigated, and could be surprising. In conclusion, the availability of a new class of drugs with an entirely different mechanism of action compared with present ESAs increases available options for doctors and patients to treat anemia in CKD. However, because of the unique mechanism of action of HIF-PHI, we would have expected more in terms of improved outcomes in comparison to present ESAs. Disclosures F. Locatelli reports consultancy agreements with Amgen and Baxter; honoraria from Amgen, Astellas, Baxter, and Roche Vifor-Pharma as speaker at sponsored meetings; advisory or leadership role with Accelsior, Amgen, Astellas, AstraZeneca, GSK, Norgine, Otsuca Roche, Travere, and Vifor Pharma; speakers bureau for Amgen, Astellas, and Roche; and other interests or relationships with ERA, SIN, DOPPS, and KDIGO. L. Del Vecchio reports honoraria from Astellas for speaker fee; and advisory or leadership role with Astellas. Funding None.