The clinical diagnosis of erythropoietin deficiency is usually prompted by new onset anemia on a background of elevated serum creatinine, and is then confirmed by measurement of serum erythropoietin. In patients with sickle cell anemia (SCA), who are already anemic, and in whom base-line creatinine and erythropoietin levels tend to run low, it is easy to miss this diagnosis. We demonstrate the clinical significance of erythropoietin deficiency in SCA and suggest alternative criteria for considering this diagnosis. Compensatory reticulocytosis maintains hemoglobin levels compatible with life in the chronic hemolytic anemia of SCA. Relative reticulocytopenia (RR) has been defined by the Multi-Center Study of Hydroxyurea Follow-Up Study (MSH-FU) as reticulocytes <250×109/L despite hemoglobin <9g/dl and is associated with decreased survival. However, to effectively address RR as a possible cause of early death, its underlying cause(s) must be defined. In a population of 433 SCA adult patients followed in our outpatient clinic between 1997 and 2008, 47 patients had consecutive out-patient laboratory values that met the MSH-FU definition of RR. In these patients, the mortality rate for those with RR was 55% compared to 7% in the patients without RR. Median survival from the initial clinic visit was 383 weeks for patients with RR but was not reached for patients without RR. Despite their increased risk for mortality, patients with RR did not have more frequent vaso-occlusive crises or emergency room (ER) visits (ie., ER visit frequency was a poor indicator of this population at risk).

SCA causes multi-organ damage, therefore, RR could result from decreased bone marrow function, from renal damage causing a decrease in erythropoietin levels, or from both processes. In univariate analyses, RR was significantly associated with thrombocytopenia, leucopenia, elevated serum creatinine and proteinuria (p<0.001 Chi-Square). This suggested that both global bone marrow function and renal damage causing erythropoietin deficiency could be contributing to RR. This was confirmed in multi-variate analyses, in which both platelet counts and proteinuria were significantly associated with RR (p<0.0001, logistic regression). Erythropoietin levels, and the erythropoietin/reticulocyte ratio, were especially low in RR patients with hemoglobin levels <8g/dl, suggesting that erythropoietin deficiency was the major cause of RR in patients with the greatest declines in hemoglobin.

Importantly, the erythropoietin/reticulocyte ratios that implicated erythropoietin deficiency in the etiology of RR were mostly noted in patients who did not meet standard clinical definitions of chronic renal failure: < 30% of the patients with RR had serum creatinine levels >1.1 or calculated creatinine clearance <60ml/min. However, 87% of the patients with RR had proteinuria as evidence of renal damage. Therefore, early renal damage, sufficient to cause proteinuria, but insufficient to cause azotemia, was associated with clinically significant erythropoietin deficiency in SCA.

In conclusion, RR in adult SCA is not associated with more frequent vaso-occlusive crises but nonetheless has a major association with mortality. An important and potentially correctable cause of RR is erythropoietin deficiency, which can be diagnosed in patients with early renal damage manifest by proteinuria but without azotemia. This sub-set of SCA patients requires more study, including evaluation for the causes of death, and the possible benefits of erythropoietin replacement therapy, using end-point criteria such as survival or progression of chronic organ damage (for eg., left-ventricular dysfunction) rather than vaso-occlusive crisis frequency.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author