To the Editor:  

In their December 2007 case report, Damon L. Baker, DO, and colleagues¹ concluded that their patient's elevated lactate dehydrogenase level and reticulocyte count indicated hemolytic anemia, and that the results from a peripheral blood smear suggested normocytic anemia with granulocytosis and thrombocytosis. 

The authors reported that laboratory tests showed that the patient's lactate dehydrogenase level was 263 U/L, and her reticulocyte count was 3.08%.¹ However, both of these measurements should be used only as screening tests for hemolytic anemia—not as confirmatory tests—because of their lack of diagnostic specificity. 

More appropriate confirmatory tests for hemolytic anemia would be measurements of serum indirect bilirubin and haptoglobin. In fact, Marchand et al² reported that a haptoglobin level of 25 mg/dL or less was 96% specific and 83% sensitive in a 100-patient retrospective study. Also arguing against a diagnosis of hemolytic anemia in the case report by Baker et al¹ was the lack of spherocytosis, anisocytosis, and polychromatophilia on the patient's peripheral blood smear.³ 

In this patient's case,¹ even if she was experiencing autoimmune hemolysis, the reticulocyte count was inappropriately low to fully explain her degree of anemia (hemoglobin, 9.6 g/dL after transfusion of 4 units of packed red blood cells). These laboratory results indicate a primary bone marrow or erythropoietin problem. The inappropriately low reticulocytosis may better be appreciated by calculating the absolute reticulocyte count or the corrected reticulocyte count.⁴ The absolute reticulocyte count is determined by the product of the red blood cell count and the percent reticulocytosis. The corrected reticulocyte count is equal to the absolute reticulocyte count divided by the reticulocyte maturation time (in days).⁴ The basis of this calculation arises from the fact that reticulocytes may spend as long as 2.5 days in the peripheral blood of very anemic patients, resulting in a reticulocyte count that more than doubles without any increase in red blood cell production. 

The authors also stated that a Coombs test was not administered because the patient had already received corticosteroids.¹ The Coombs test is an essential diagnostic test in the clinical approach to hemolytic anemia because of its substantial positive predictive value and its ability to differentiate warm from cold autoantibodies.³ This differentiation is important as the treatments associated with each condition vary. The administration of corticosteroids does not preclude obtaining a Coombs test in this case¹ for two reasons. First and most importantly, several different mechanisms of corticosteroid action in hemolytic anemia have been described in the literature, but only decreased antibody production directly affects the Coombs test.³ However, several weeks of therapy are usually required before this effect is observed.³ Second, the patient in this case¹ received only 40 mg/d of prednisone—below the current weight-dependent guidelines of 1 mg/kg prednisone (usually administered in dosages of 60-100 mg/d) for patients with hemolytic anemia.³ Most patients with hemolytic anemia will require this dosage for 1 to 3 weeks, while 20% to 30% of patients will be refractory to steroids and will require other treatment modalities, such as rituximab, cyclophosphamide, intravenous immunoglobulin, or a splenectomy.³ 

One final point of concern is that the authors¹ state that the suspected presence of hemolytic anemia also supported a diagnosis of seronegative rheumatoid arthritis. Interestingly, however, both the incidence and prevalence of hemolytic anemia in patients with seronegative rheumatoid arthritis do not exceed that of the general population, and only a few cases have been reported of the coexistence of these two diseases.⁵ 

In conclusion, the anemia of the patient in the case report by Baker et al¹ was most likely not caused by hemolysis. Rather, it was probably a multifactorial normocytic anemia with inappropriate erythropoietin response, which is one mechanism seen in the anemia of chronic disease.⁶ This diagnosis is supported by the patient's elevated erythrocyte sedimentation rate (106 mm/h) in the presence of rheumatoid arthritis.¹ Normocytic anemia with inappropriate erythropoietin response is also seen in cases of reduced erythropoietin levels stemming from chronic kidney disease⁵—a condition that matches the patient's reported creatinine value of 1.8 mg/dL and her 24-hour urine protein loss of more than 2 g.¹ These two types of normocytic anemia can be easily diagnosed by obtaining results of iron studies (eg, total serum iron-binding capacity, serum ferritin levels, percent transferrin saturation), as well as soluble transferrin receptor and serum erythropoietin levels.⁷