Objective: To determine whether an independent association exists between anaemia and chronic kidney disease (CKD) outcomes in a quasi-incidence cohort when patients' most recent laboratory values are considered.
Methods: We conducted a dynamic, retrospective cohort study among patients with incident CKD in a large health maintenance organization administrative data set. CKD was defined by two estimated glomerular filtration rates (eGFR). We measured the absolute rates for all-cause mortality, cardiovascular hospitalizations and end-stage renal disease.
Results: Our completed cases Cox regression model followed 5885 patients with both CKD and haemoglobin measures. For patients with the most severe anaemia (haemoglobin <10.5 g/dL), we estimated an increased rate of mortality (hazard ratio (HR)=5.27, CI 4.37-6.35), cardiovascular hospitalizations (HR=2.18, CI 1.76-2.70) and end-stage renal disease (HR=5.46, CI 3.38-8.82) when compared with patients who were not anaemic; the HR reflect time-varying haemoglobins and eGFR.
Conclusion: Anaemia is a predictor of excess mortality, excess cardiovascular hospitalizations and excess end-stage renal disease even when the progression of CKD is considered by controlling for time-varying eGFR values.