Erythropoietic response to erythropoiesis-stimulating agents and outcome: should we give up the haemoglobin target approach?

Abstract

The hypothesis that complete anaemia correction with erythropoiesis-stimulating agents (ESA) would reduce the risk of death and cardiovascular and renal endpoints among patients with chronic kidney disease (CKD) has been largely tested in the last 15 years. Despite expectations, no study has demonstrated any clinical benefit except some improvement in quality of life by targeting at high haemoglobin (Hb) levels in comparison to partial anaemia correction. Some of these studies even found a trend towards increased risk of cardiovascular endpoints (i.e. death, stroke, acute myocardial infarction and hospitalization for congestive heart failure in different combinations) at primary or secondary analyses [1–4]. A lively debate has developed focussing on possible explanations why clinical trials have not confirmed what has been shown by observational studies: the higher the Hb the better the outcome. The Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT) study randomized 4038 patients with Type 2 diabetes and CKD Stages III–IV to darbepoetin alfa (Hb of 13 g/dL) or to placebo (with rescue darbepoetin alfa for Hb <9.0 g/dL) [5]. Results were neutral on primary composite cardiovascular and renal outcomes, but a higher risk of stroke, death for cancer in patients with previous history of malignancies and thromboembolic events were found at secondary analyses in those randomized to complete anaemia correction. This parallelled a reduction in cardiac revascularization procedures, transfusion rates and improvement in quality of life in this group [6]. Possible implications of these findings have already been discussed [7]. In this editorial, we would like to examine carefully and comment on the recent secondary analysis of the TREAT data by Solomon et al. [8], which investigated the relationship between initial response to darbepoetin alfa, patient characteristics and outcomes. For this purpose, patients in the darbepoetin-alfa group having a Hb change <2% in the first month after the initial weight-based two doses of darbepoetin alfa were defined as having a poor response and compared with the remaining ones (upper three quartiles of Hb changes). After adjusting for outcome-related covariates, this subgroup had a higher risk of reaching the cardiovascular composite endpoint or death from any cause than the better Hb responders who had an event rate similar to the placebo group. Conversely, both subgroups had similar stroke rates that were higher than the placebo group. This post hoc analysis suggests that after just two doses of darbepoetin alfa, it is possible to classify the patients in good and bad responders and this may translate into a different outcome risk. This may be an easy method to identify high-risk patients and tailor ESA treatment accordingly, avoiding excessive doses. However, as admitted by the authors, these findings cannot identify the culprit of increased cardiovascular risk following ESA treatment. Is this a consequence of too high ESA doses especially in those who are hyporesponsive to treatment or does this merely reflect patient characteristics?