Navigating the Complexity of Clinical Outcomes with HIF-PH Inhibitor for Anemia Management

Anemia is a common complication of CKD, and active management of anemia related to CKD has been an important part of nephrology practice for more than two decades. Treatment for anemia has evolved over this time, with a gradual transition toward lower hemoglobin targets, lower doses of therapies that raise hemoglobin (erythropoiesis-stimulating agents [ESAs]/hypoxia-inducible factor prolyl hydroxylase [HIF-PH] inhibitors), and more liberal use of intravenous iron.¹ A major success in nephrology is that these decisions/changes in treatment paradigm have been a direct result of high-quality randomized controlled trials (positive or negative) that directly affected clinical practice.² Nonetheless, the ideal anemia treatment that improves hemoglobin and quality of life and is safe from a cardiovascular treatment remains elusive.

In the early 2000s, evidence from the large randomized trials of ESAs changed the management of anemia in CKD.^3,4 These trials helped us recognize the cardiovascular risks with ESA therapy and lowered hemoglobin targets in the United States and worldwide. An unmet need for a safe and effective oral agent was discovered, and Nobel prize–winning research from nephrologists and significant investment from the biotech/biopharmaceutical industry led to the creation of a new drug class—oral HIF-PH inhibitors for the management of anemia in CKD.⁵ Given the known cardiovascular risks with ESAs, these drugs were required to show evidence of noninferiority for cardiovascular risks compared with ESAs, in addition to the ability to raise hemoglobin to target and avoid transfusions. Large randomized trials in both the dialysis-dependent (DD) and non–dialysis-dependent (NDD) CKD population were conducted for three of these agents (roxadustat, vadadustat, daprodustat), and none met the criteria for approval in the nondialysis CKD population in the United States.^6–8

It is important to note that roxadustat is authorized and available in Europe and has been used for several years in Japan, China, and other countries. Vadadustat is also available in Europe and Japan, and other agents are being marketed in India and other geographies. These key distinctions in market authorizations and availability highlight the differences in practice in anemia management worldwide and show that regulatory bodies pay attention to regional differences in efficacy and safety that may otherwise appear as subgroups in the reporting of a large clinical trial.

In this issue of JASN, Chertow et al. present findings from a prespecified analysis of US and non-US patient subgroups from the vadadustat clinical program.⁹ The program consisted of four large clinical trials: two trials in patients in DD-CKD (INNO₂VATE) and two trials in NDD-CKD (PRO₂TECT). The dialysis-dependent patient population trials consisted of one focused on an incident dialysis population (n=369) and the other dedicated to a prevalent dialysis population (n=3554). In the non–dialysis-dependent program, the trials were similar in size, with one focused on ESA-naïve patients (n=1751) and the other focused on ESA-treated patients (n=1725). All four trials recruited adults and had achievement of hemoglobin as the primary efficacy end point and noninferiority for major adverse cardiovascular events (MACE) as the primary safety end point. Fifty-five percent of the participants in these trials came from US sites, and 45% came from sites outside the United States, including more than 100 patients each from Ukraine, Brazil, and South Africa.

There were important differences in US and non-US patients at baseline in all four clinical trials. In the INNO₂VATE program, serum ferritin concentrations in US patients were double that of non-US patients, and in the PRO₂TECT program, the number of participants with eGFR <10 ml/min at the time of randomization was substantially higher in the non-US population (15% versus 7%). Safety outcomes also differed by region; in US patients, there was no increased risk for any MACE in DD-CKD (hazard ratio [HR], 1.00; 95% confidence interval [CI], 0.84 to 1.18) or NDD-CKD (HR, 1.06; 95% CI, 0.87 to 1.29). By contrast, the non-US patients had a higher risk of any MACE in NDD-CKD group (HR, 1.29; 95% CI, 1.03 to 1.60), but no effect of MACE risk in DD-CKD population (HR, 0.88; 95% CI, 0.67 to 1.17).

There were no differences in hemoglobin efficacy in the US or non-US populations in the NDD or the DD trials, and the risk of treatment-emergent adverse events and serious adverse events was similar in all four trials across the US and non-US regions. The authors suggest that differences in clinical practice and imbalances in randomization may have contributed. These differences included a disproportionate number of patients with eGFR <10 who ended up with death from kidney failure, as well as higher than expected rate of MACE events in European patients who switched from darbepoeitin to vadadustat. Together, these findings may have led to failure to meet the prespecified noninferiority margin in the NDD-CKD trials.

Large randomized trials in kidney and cardiovascular diseases must recruit outside of US populations. The cost of conducting trials in the United States has continued to rise, and in both kidney and cardiovascular trials, and US patients have the highest discontinuation rates of therapy. In fact, therapy discontinuation rates in the INNO₂VATE and PRO₂TECT programs were 11% in US participants versus 6% in the non-US population. However, US regulatory authorities make approval decisions based on US patient outcomes, and the standard of care for concomitant therapy, which includes BP management, statin, renin-angiotensin-aldosterone system inhibitors, and sodium-glucose cotransporter 2 inhibitors use, and availability of treatment for kidney failure does differ from US to non-US settings. It is also important to note that not all non-US settings are the same—countries such as Canada, United Kingdom, and Germany as an example may have very similar practice patterns to US health care, whereas other European sites may differ.¹⁰ Furthermore, it is possible that differences in genotype, diet, or even the estimation of kidney function (eGFR) is different in other countries (e.g., China, Japan, and India), and these factors can be critical in determining the efficacy of an intervention in kidney disease.¹¹

The findings from Chertow et al. are prespecified and show that these differences do matter and had a conclusive effect on the overall safety estimate from the trial program. However, they do not answer the important question on whether the differences in MACE risk for non-US populations were incidental or related to practice patterns and leave us with a hypothesis-generating finding about noninferiority in the United States, which must be tested in a new clinical trial. Nephrologists want and deserve large, well-run randomized trials for the therapies we prescribe, and these trials will always have a substantial non-US population. Regulators and sponsors should ensure that the non-US sites follow a similar standard-of-care background therapy compared with US sites, and wherever possible, objective outcomes (e.g., eGFR decline >40% or reaching eGFR <10 ml/min) should be favored compared with subjective outcomes (need for KRT). Overinterpretation of subgroup analyses will always remain a bigger risk than missing a true region-by-treatment interaction, but the article from Chertow et al. reminds us that differences in practice patterns across regions do matter and can be the appropriate reasons for drug approvals that are different in different countries.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/F390.

Author Contributions

Conceptualization: Navdeep Tangri.

Writing – original draft: Navdeep Tangri.

Funding

None.