Normalization of Hemoglobin in Patients With CKD May Cause Harm: But What Is the Mechanism?

In this issue of the American Journal of Kidney Diseases, Streja et al1 hypothesize that the increased mortality in patients with kidney failure with hemoglobin levels increased into the normal range by using erythropoietic-stimulating agents (ESAs) is caused by reactive thrombocytosis that, in turn, is caused by iron deficiency. Their study was a retrospective look at more than 40,000 patients with kidney failure receiving outpatient hemodialysis. Relative thrombocytosis (defined as platelet count > 300 × 109/L) was seen in 15% of the patient population and was associated with a 30% greater weekly dose of ESA, lower transferrin saturation, and lower serum ferritin concentration. Patients with a high hemoglobin level (>13.0 g/dL) who also had relative thrombocytosis had a relative risk of all-cause mortality of 1.22 compared with similar patients without thrombocytosis. Patients with thrombocytosis had an increased death rate at all hemoglobin levels compared with patients without thrombocytosis.

ESAs have been the treatment of choice for patients with anemia of chronic kidney disease (CKD) since 1986, when Winearls et al2 published a report of 10 patients with end-stage kidney failure treated with recombinant erythropoietin. This pilot study was rapidly followed by several multicenter trials that confirmed the efficacy of ESAs at increasing hemoglobin levels, decreasing transfusion need and improving quality of life.3, 4 Complications of treatment with ESAs, such as hypertension and thrombosis in arteriovenous accesses, were recognized and managed effectively.

There is controversy about the optimum target hemoglobin level. Observational studies of patient with kidney failure, as well as those with earlier stages of CKD, have shown lower mortality with anemia correction to hemoglobin levels of around 12.0 g/dL. Clinical trials of increasing hemoglobin levels into the normal range (>13.0 g/dL) have shown harm.5, 6, 7 A recently published meta-analysis confirmed these findings8 and that the target hemoglobin level probably should be 11.0 to 12.5 g/dL. Interestingly, similar findings have been reported in ESA-treated patients with anemia of cancer or cancer chemotherapy. Correction of hemoglobin levels from less than 10.0 g/dL toward 12.0 g/dL improves quality of life and decreases transfusion need in these patients.9, 10 Recent studies suggested that increasing hemoglobin levels into the normal range might be deleterious, particularly in patients not receiving concomitant chemotherapy.11, 12 This adverse effect has not been seen in patients in whom hemoglobin levels are maintained in the 11.0- to 12.0-g/dL range, and in some studies, there is a suggestion of improved life expectancy compared with more anemic patients.13

What is the scale of the problem and what are the possible mechanisms by which normalization of hemoglobin levels results in excess mortality in patients with CKD? The meta-analysis8 showed that the relative increase in all-cause mortality was approximately 20% in patients with a high hemoglobin level, with an increase in arteriovenous access thrombosis and hypertension of around 30%. Much of the excess mortality can be attributed to an increase in cardiovascular events. Possible explanations previously have included an increase in viscosity through an increase in hematocrit, adverse effects of hypertension, and also direct effects of the ESA.

In the study by Streja et al,1 only patients with thrombocytosis had a greater mortality risk at a hemoglobin concentration greater than 13.0 g/dL, suggesting that the increase in hematocrit per se is not sufficient explanation to account for the increased risk of death.

Patients with CKD who receive an ESA have a lower mortality rate than those not receiving an ESA, but mortality increases incrementally with greater monthly doses.14 There are 2 possible explanations for the increased mortality at greater doses. First, that patients needing greater doses are those with concomitant infectious, inflammatory, or malignant conditions resulting in relative ESA resistance and that might contribute to increased mortality risk. Second, greater ESA doses might be deleterious because of the excess increase in hemoglobin levels or through their extensive nonhemopoietic effects. ESAs have been reported to increase thrombotic risk through increased inflammation and antifibrinolytic activity,15 although how much of an impact this has on hemostasis in patients with kidney failure has been questioned.16 ESAs also have cardioprotective properties after myocardial infarction and may be of clinical value in the management of heart failure.17 In summary, there are no convincing data to suggest that the nonhematopoietic effects of ESAs are to blame for the increased mortality risk.

The suggestion by Streja et al1 is that the increased mortality risk is caused by thrombocytosis as a result of real or functional iron deficiency induced by ESA treatment. Although the increased platelet count probably is secondary to iron deficiency, it is possible that other factors that can cause thrombocytosis (eg, infection, inflammation, or blood loss) are at play here. Whatever the cause(s) of the increased platelet count, can such a modest increase be blamed for a 20% increase in mortality, particularly when platelet function is abnormal in patients with kidney failure? Data from patients with malignancy suggest a correlation between platelet count and mortality,18, 19 but it is likely that patients with the highest platelet counts were those with the most aggressive or advanced malignancy or caused by other comorbidities (such as bleeding), which will have contributed to an increased mortality risk. Nevertheless, it is possible that in a group of patients with kidney failure, with a high prevalence of cardiovascular disease, even a small increase in platelet numbers could trigger an increase in mortality.

Further support for the notion that reactive thrombocytosis might increase the thrombotic risk comes from a trial of anemic chemotherapy-treated patients with cancer receiving ESAs who were randomly assigned to receive intravenous iron, oral iron, or no iron.20 In the intravenous iron group, there were 7 thrombotic events in 6 patients (of 63 patients) compared with 22 events in 13 patients in the oral or no iron control groups (total population, 124 patients). Overall, there was a 4-fold increase in thrombotic events in patients with a platelet count that increased from baseline to greater than 350 × 109/L at any time in the study.

There are also observational data suggesting a link between iron status and survival in patients with CKD.21 At low serum iron levels, mortality increases; there is similarly high mortality in patients with a low iron saturation level (calculated by dividing serum iron by total iron-binding capacity), but also an increase in mortality at high iron saturation levels. A low serum iron level is found in patients with iron deficiency, but also in those with anemia of chronic disease. The association between low iron level and mortality therefore may be a reflection of associated infectious or inflammatory illness in these patients.

In this study by Streja et al,1 approximately 60% of patients had received intravenous iron. Intravenous iron is superior to oral iron in overcoming functional iron deficiency in renal patients treated with ESAs22 (the same findings have now been convincingly shown in patients with cancer-related anemia).23, 24 If the hypothesis of Streja et al1 is correct, more use of intravenous iron to help prevent iron deficiency will reduce the number of patients with reactive thrombocytosis and hence reduce thrombotic complications and mortality. Data from the Dialysis Patients' Response to IV Iron With Elevated Ferritin (DRIVE) Study25 support the value of intravenous iron in improving hemopoietic response, even in patients with a high baseline ferritin level. A decrease in mortality has been suggested in patients receiving any intravenous iron up to a maximum dose of 400 mg/mo.21 Concerns about excess mortality from iron overload in patients with kidney failure have not been substantiated, although in the study by Besarab et al,5 which showed an increase in mortality in patients with high hemoglobin levels, those who received intravenous iron had a relative mortality risk of 2.4 against those not receiving intravenous iron.

ESAs are a good thing for patients with the anemia of kidney failure, but too much of a good thing can be bad. We know for certain that normalization of hemoglobin levels increases mortality in this group of vulnerable patients. The mechanism for this increase is uncertain, but Streja et al1 have suggested a novel and interesting possible explanation.