Timing of erythropoiesis-stimulating agent initiation and adverse outcomes in nondialysis CKD: a propensity-matched observational study

Abstract

Background and objectives: The severity of anemia at which to initiate erythropoiesis-stimulating agent (ESA) treatment in nondialysis chronic kidney disease (CKD) patients is unclear. Risk of mortality, hospitalizations, and blood transfusion were compared among nondialysis CKD patients with "early" versus "delayed" ESA initiation.

Design, setting, participants, & measurements: A retrospective cohort study was conducted on CKD (estimated GFR <60 ml/min/1.73m(2)) outpatients in the national Veterans Administration who were initiated on ESAs. Patients with ESRD, gastrointestinal bleeding, chemotherapy, or hematologic malignancy were excluded. Patients were characterized as having early [hemoglobin (Hb) 10.0 to 11.0 g/dl] or delayed (Hb 8.0 to 9.9 g/dl) ESA initiation. A propensity score comprising demographic, clinical, and laboratory variables was used to select a 1:1 matched cohort. Cox survival and negative binomial regression were used to compare the matched groups for all-cause mortality, hospitalizations, and blood transfusions.

Results: Of 1837 patients who met inclusion criteria, 1410 (77%) were successfully matched. The groups did not differ significantly in 31 characteristics reflecting sociodemographics, comorbidity, healthcare utilization, and renal function. There was no significant difference in mortality with early initiation. Those initiated early had a 17% lower risk of initial hospitalization and a 29% lower risk of transfusion compared with delayed initiation patients. Results did not differ between those with and without pre-ESA transfusion or hospitalization.

Conclusions: In nondialysis CKD, ESA initiation at Hb 10.0 to 11.0 g/dl compared with 8.0 to 9.9 g/dl is associated with reduced risk of blood transfusion and initial hospitalization.

Figure 1.

Flow of subjects included in analysis.

Figure 2.

(a) Cumulative survival, (b) proportion without hospitalization, and (c) proportion without RBC transfusion among propensity-matched subgroups with early and delayed ESA initiation.