Introduction
Anemia of chronic kidney disease (CKD) is associated with increased risk of cardiovascular events and mortality, increased healthcare resource utilization (HCRU), and reduced health-related quality of life (HRQoL) [
1‐
6]. While the prevalence of anemia increases with CKD severity [
7‐
9], anemia also imparts a substantial healthcare burden in early-stage CKD [
7].
Traditionally, management of anemia of CKD has relied on the variable use of iron replacement therapy and erythropoiesis-stimulating agents (ESAs), according to geographical region and CKD severity [
10‐
12]. ESAs have demonstrated efficacy in correcting hemoglobin (Hb) levels, improving HRQoL, and reducing the need for red blood cell (RBC) transfusions [
5,
13]. However, clinical trial data have showed inconsistent improvement in mortality, renal, or cardiovascular outcomes with ESA use in patients with CKD not on dialysis [
14‐
16]. Notably, reports from several large randomized trials of an increased risk of mortality, thrombotic events, and cardiovascular outcomes in ESA-treated patients when targeted to normalize Hb [
15‐
17] led the Food and Drug Administration and the Kidney Disease: Improving Global Outcomes (KDIGO) group to advise against using ESAs in patients with Hb levels > 10 g/dL [
12,
18]. Recommendations from the European Renal Best Practice group suggest ESA initiation at Hb levels between 9 and 10 g/dL in high-risk patients (e.g. older patients and those with comorbidities), with the potential to initiate treatment at higher Hb levels in low-risk patients (e.g. younger patients with very few comorbidities) [
19].
The optimal management of anemia of CKD remains controversial [
3]. Marked reduction in ESA use has resulted in a shift towards less intensive therapy and lower Hb treatment targets in patients with anemia of CKD [
8,
20,
21]. Notably, many patients with CKD not on dialysis do not receive any anemia treatment [
7,
22], despite some evidence that untreated versus treated anemia is associated with higher HCRU and decreased HRQoL [
5,
6]. Moreover, treating anemia of CKD can reduce the requirement for RBC transfusions, and thus associated risks from transfusion-related reactions and the development of alloantibodies [
23].
Contemporary, real-world insights into the burden of anemia of CKD, HCRU, and long-term clinical outcomes for patients with CKD not on dialysis are needed to inform clinical practice. In particular, as novel treatment strategies for anemia with unique mechanisms of action and potentially improved cardiovascular safety emerge, information on treatment patterns and outcomes in patients with CKD will be particularly valuable to help physicians assess the benefits and risks of treatment on an individual patient basis. In this study, we used the Henry Ford Health System (HFHS) database to gather descriptive real-world information, including selected clinical outcomes of anemia, on US patients with CKD who were not receiving dialysis.
Discussion
This retrospective cohort study investigated the baseline prevalence, use of anemia treatments, post-baseline incidence of anemia, and adverse clinical outcomes in patients with CKD not on dialysis at baseline, using data originating from primary and secondary care settings within a large US healthcare system. This analysis confirmed that potentially treatable anemia (i.e. Hb < 10 g/dL) was prevalent among patients with CKD but treated only infrequently (31.7% of the population with anemia at baseline were receiving any anemia treatment, predominantly iron replacement therapies). In addition, the cumulative incidence of anemia in patients with CKD without anemia at baseline was 7.4 and 38.8% at 1 year and 5 years, respectively, while the presence of anemia at baseline was independently associated with increased risks of adverse renal and cardiovascular outcomes at 1 and 5 years.
The baseline prevalence of severe anemia was 23.0% and increased with CKD stage, confirming previous reports [
7,
9,
27,
28]. It is, however, difficult to compare with published prevalence estimates, given the differences in study populations, the time period in which prevalence is assessed, and in the thresholds used to define anemia. Among patients without anemia at baseline who developed anemia during follow-up, the decline in Hb levels occurred in parallel to a decline in eGFR over the same period. This finding emphasizes the importance of periodic screening of patients with CKD for anemia early in the course of the disease, as well as monitoring of Hb levels and symptoms of anemia as CKD progression occurs.
Despite the high prevalence of anemia in the study cohort, the frequency of prescriptions for anemia treatments, particularly ESAs, was relatively low. These findings concur with two other studies of patients with CKD stages 3–5 (either not on dialysis [
1] or with dialysis status not specified [
7]) using US datasets between 2007 and 2013. In both studies, < 40% of patients with anemia received treatment for this condition [
1,
7]. Similarly, in a cross-sectional study of patients with CKD not on dialysis in China, approximately one-third of patients with anemia were receiving treatment with erythropoietin and/or iron products [
27]. Other longitudinal studies have highlighted the potential undertreatment of anemia in patients with CKD and anemia. In one European study, there was a notable lack of anemia therapy modification, particularly with regard to iron supplementation, over 6 months in patients with Hb < 11 g/dL, despite unmet treatment goals [
29]. Additionally, in a recent analysis from the prospective, multinational Chronic Kidney Disease Outcomes and Practice Patterns Study, the proportions of treatment-naive patients with stages 3–5 CKD and Hb < 10 g/dL prescribed any anemia therapy and ESAs during 12 months of follow-up were 40 and 28%, respectively [
22].
The low ESA use observed in the present study may be attributable in part to the definition of anemia, which required at least one Hb value < 10 g/dL to be recorded in an outpatient setting, with exclusion of acute and common major causes of anemia, including active cancer and bleeding. Some patients may have subsequently had higher Hb values recorded, which is why they did not have an ESA prescription recorded at baseline. Other potential clinical factors include avoidance of associated cardiovascular events and uncertainty regarding the need for treatment prior to starting dialysis [
1,
30‐
32]; the need for patients to meet certain thresholds in terms of Hb levels or other laboratory parameters before treatment is given; under-recognition of anemia in non-dialysis-dependent CKD as a treatable condition; operational challenges of administering parenteral drugs to patients not on dialysis; and cost. In particular, uncertainty regarding whether or not to treat patients with ESAs before dialysis initiation may be one reason for these findings. Studies have also highlighted substantial differences in anemia treatment practices, both between countries and also at the clinic level within individual countries [
11,
28]. These differences may reflect several factors, including variation in adherence to different CKD management guidelines, selection of Hb target thresholds for ESA and iron prescribing, and use of anemia algorithms to guide clinical practice. Irrespective of the reason, the findings from this study suggest an incongruity between guidelines for the treatment of anemia [
12,
19] and clinical practice, and a need for further exploration of optimal treatment practices for patients with anemia of CKD. This is particularly important in the context of evidence that suggests untreated anemia of CKD is associated with reduced HRQoL and increased HCRU [
6]. However, it is important to acknowledge the potential risks associated with some currently available treatments for anemia, and the need for treating physicians to carefully balance the potential benefits of treatment against the risk of adverse events.
Incidence rates of adverse renal and cardiovascular outcomes in the study cohort were comparable with those seen in previous studies of patients with advanced CKD and at high cardiovascular risk [
33,
34], and likely reflect the multimorbid patient population in this study. The incidence of these outcomes observed in patients with baseline anemia compared with those without anemia is consistent with findings from other studies that showed an increased risk of adverse outcomes, including mortality, renal events, and adverse cardiovascular outcomes in patients with CKD with versus without anemia [
4,
35,
36]. In particular, our findings concur with those from a Danish cohort study of patients with severe CKD (eGFR < 30 mL/min/1.73m
2), where the adjusted hazard ratios for incident dialysis, all-cause mortality, and MACE were markedly increased in non-dialysis-dependent patients with anemia (Hb < 12 [women] or < 13 g/dL [men]) versus without anemia, and increased with anemia severity [
4]. As in the study by Toft et al. [
4], the hazard ratios for renal and cardiovascular outcomes for patients with versus without anemia were higher at 1 year versus 5 years. This apparent attenuation in risk over time may in part reflect changes in both patient populations over time, where more ill patients died, and the survivors were selected for less severe disease. It could also be attributed to a greater likelihood of specialist care and an increased use of targeted renal and cardiovascular interventions in patients with the most severe disease. The risk of renal and cardiovascular outcomes was similar using both a Kaplan–Meier survival approach (to estimate the risk of events) and CIF (to estimate a patient’s risk of events prior to mortality); results using the latter method provide valuable additional insight from a population perspective, when taking into account mortality as a competing risk.
Whether anemia contributes directly or is merely associated with adverse renal and cardiovascular outcomes is unclear. On the one hand, it has been proposed that oxidative stress, inflammation, and diminished biological capacity contribute simultaneously to both anemia and adverse clinical outcomes without a clear linear relationship being present [
37]. In particular, diminution of renal function leading to depressed erythropoietin production is a major cause of anemia [
37]. Conversely, decreased oxygen delivery to tissues has been proposed to drive renal fibrosis through hypoxia-mediated signaling pathways [
38], and chronic oxygen deprivation may also contribute to cardiac dysfunction by driving compensatory adaptations such as increased cardiac output [
39].
Analysis of baseline factors associated with clinical outcomes showed that advancing CKD stage was generally associated with a greater risk of renal and cardiovascular events in all patients, regardless of whether they had anemia at baseline. Hazard ratios for outcomes assessed varied widely, which likely reflects differences in the risk of experiencing each outcome during the CKD disease course. Our findings also demonstrated an association between white ethnicity and an increased risk of renal, cardiovascular, and bleeding outcomes in both the anemia and non-anemia cohorts, although this was not evident at all time points, or for all outcomes. This contrasts with prior reports that suggested an increased rate of CKD progression associated with non-white ethnicity [
40]. Notably, these studies differ in adjustment for baseline factors, including socioeconomic status, in analyses, necessitating caution in their interpretation.
Major strengths of this study include the large and diverse population, the extensive follow-up period, and comprehensive data source, including laboratory test results and prescription medication orders. Limitations include those that are inherent to many retrospective database analyses. Administrative data are not collected for research purposes and are subject to coding errors, and complete datasets were not available for all variables. Medications administered without a prescription (e.g. oral iron) were not captured; some patients were lost to follow-up; and there was no link between EHRs and retrospective claims. Information regarding the etiology of anemia was not available; thus, it was not possible to verify that all patients in the study had anemia attributable to CKD, although major causes of anemia other than CKD (e.g. bleeding and cancer) were excluded. Information on cause of death recorded on death certificates is not always informative, and this information was not readily accessible from other sources, due to Institutional Research Board regulations.
Although we adjusted for key baseline covariates when assessing associations between patient factors and clinical outcomes, there are other patient factors that may have also influenced the results, but were not well captured in the database, such as body mass index. Moreover, the anemia status of some patients (for example the small proportion of patients who were treated with ESAs at baseline) may have changed during the study follow-up period, which could have impacted clinical outcomes assessments.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.