- The Evolution of Evidence Generation in CKD-Anemia: A History of Agents and Trial Designs
-
Wolfgang Winkelmayer
2022 ; 2022(1):
- 논문분류 :
- 춘계학술대회 초록집
Just 35 years ago, severe anemia was pervasive in the hemodialysis population, where blood transfusions were commonly administered to alleviate the most extreme cases of anemia. The postulation and subsequent discovery of erythropoietin and identification of its gene, heralded a new era of treatments for CKD anemia. Recombinant erythropoietin alfa (epoetin alfa) came to the U.S. and other markets based on a pivotal single‐arm trial that did not have a comparison (placebo) group. While it was evident that the rather effective treatment of anemia led to a swift reduction in transfusions relative to the rate observed prior to epoetin alfa initiation, other benefits such as patient reported outcomes could not be demonstrated convincingly.
Further, demonstrating the postulated cardiovascular and mortality benefits required a different trial design. Subsequently, several trials comparing higher to relatively lower hematocrit or hemoglobin targets were conducted: the Normal Hematocrit Trial in maintenance hemodialysis patients, as well as the CREATE and CHOIR trials in patients with moderate to advanced chronic kidney disease and anemia. These experiments were executed using superiority trial designs to test for differences in the rates of all‐cause mortality and non‐fatal cardiovascular events. None of these trials was able to corroborate the alternative hypothesis of improved mortality and cardiovascular risk with more aggressive anemia treatment targets, with the Normal Hematocrit Trial being stopped early for futility and the CHOIR trial for safety.
Finally, the TREAT trial, which used a placebo‐controlled superiority design in patients with diabetes, moderate to advanced CKD, and anemia did not only fail to demonstrate superiority of darbepoetin alfa over placebo with rescue erythropoiesis‐stimulating agent (ESA) regarding the composite of mortality and nonfatal cardiovascular events but found a significant doubling of ischemic stroke in the darbepoetin group. In aggregate, these findings not only led to substantial changes in the FDA‐approved labels of ESAs, but it also appears to have altered the burden of evidence required for approval of new therapies for CKD anemia. All subsequent non‐iron investigational new drugs for the treatment of CKD anemia were subjected to non‐inferiority trials compared with ESA or, in one specific case, against placebo.
While non‐inferiority trials are increasingly common, their interpretation is more challenging. Understanding their limitations and biases is key to drawing appropriate conclusions. This lecture will explain key differences between superiority and non‐inferiority trials and explore how violations of certain key assumptions in design and analysis may induce biases that lead to false positive results, corresponding to overly optimistic adoption of non‐inferiority. Recent trials of novel treatments for CKD‐anemia are suitable vehicles to explore and explain the strengths, limitations, and fallacies of non‐inferiority designs.