Utility of the anemia index in investigating the ‘3S's' of anemia

A case-based report examines the ideal workup.


A 37-year-old woman presented to the hospital with a 2-week history of generalized weakness and lightheadedness. Her medical history was significant for treatment-refractory pemphigus vulgaris on prednisone therapy, steroid-induced hyperglycemia, and chronic anemia on iron supplementation. Her autoimmune disorder, pemphigus vulgaris, had been diagnosed at another hospital 3 years earlier. Despite a variety of treatments, she continued to have diffuse weeping and bleeding skin lesions, often leading to blood transfusions. Home medications included mycophenolate mofetil, prednisone, amitriptyline, lorazepam, metformin, insulin, pantoprazole, hydromorphone, furosemide, folic acid, potassium and iron supplementation, and laxatives.

Physical examination was remarkable for tachycardia, bipedal edema, decreased muscle strength in the lower extremities, and tearful affect. Ulcerated bullae were noted on the face, neck, chest, and back. Her skin was discolored on the lateral legs due to prior excoriations, consistent with hemosiderosis hyperpigmentation. Laboratory results were as follows:

  • hemoglobin, 8.5 g/dL (reference range, 12.0 to 15.5 g/dL);
  • hematocrit, 29.4 g/dL (reference range, 34.9 to 44.5 g/dL);
  • mean corpuscular volume, 86 fL (reference range, 81.6 to 98.3 fL);
  • red cell distribution width, 18% (reference range, 11.9% to 15.5%);
  • reticulocyte count, 3.36% (reference range, 0.77% to 2.36%) or 94.1 × 109 cells/L; (reference range, 38.1 to 112.6 × 109 cells/L);
  • iron level, 24 µg/dL (reference range, 35 to 145 µg/dL);
  • total iron-binding capacity, 282 µg/dL (reference range, 250 to 400 µg/dL);
  • percentage of transferrin saturation, 9% (reference range, 14% to 50%);
  • ferritin level, 28 µg/L (reference range, 11 to 307 µg/L);
  • soluble transferrin receptor level, 5.5 mg/L (reference range, 1.8 to 4.6 mg/L);
  • vitamin B12 level, 260 ng/L (180 to 914 ng/L); and
  • methylmalonic acid level, 0.43 nmol/mL (reference range, 0 to 0.4 nmol/mL).

Peripheral smear showed nonspecific red blood cell changes and polychromasia. Results of intrinsic factor antibody and fecal hemoglobin testing were negative.

The evidence showed that this patient's normocytic anemia had a multifactorial cause. There was a component of iron deficiency anemia (IDA), consistent with the prescribed iron supplementation, possibly due to blood loss from her weeping wounds. Elevated red cell distribution width (and soluble transferrin receptor level) supported this diagnosis. However, other laboratory findings were inconsistent with iron deficiency as the sole cause of the anemia: normal total iron-binding capacity, normal ferritin (although on the low side of normal; ferritin is generally <20 µg/L in uncomplicated iron deficiency anemia), and normal mean corpuscular volume. Elevated percentage reticulocyte count with normal absolute count and polychromasia suggested her bone marrow was producing red blood cells, perhaps as a response to recent blood loss. The autoimmune disorder also raised suspicion for anemia of chronic inflammation, also known as anemia of chronic disease (ACD).

Anemia workup

How do we determine the true cause of her persistent anemia? A systematic anemia workup addresses the disrupted synthesis, decreased survival, or sequestration (3S's) of mature red blood cells.

S1: For synthesis, close attention should be given to reticulocyte count, red blood cell distribution width, and the peripheral blood smear. These give us clues about the ability of the bone marrow to synthesize red blood cells.

S2: For sequestration, an abdominal ultrasound of the left upper quadrant should be considered, particularly if splenomegaly is noted on the abdominal examination.

S3: For survival, lactate dehydrogenase (LDH), haptoglobin, indirect bilirubin, and peripheral blood smear could be assessed to look for hemolysis.

Another method of anemia workup is classification by the size of the red blood cells: the mean corpuscular volume will diagnose macrocytic, microcytic, or normocytic anemia. For a high mean corpuscular volume indicating macrocytosis, we should screen for vitamin B12 deficiency, alcohol use, liver disease, thyroid disease, or medications that interfere with DNA synthesis leading to macrocytosis. In cases of microcytosis, we screen for IDA, thalassemia, and inflammation (ACD). Iron studies include iron, total iron-binding capacity, percentage saturation, and ferritin. The differential for normocytic anemia includes hemolysis, renal insufficiency, ACD, multifactorial causes, and others. If the diagnosis of IDA versus ACD is equivocal, we should then calculate the anemia index.

The anemia index was first described in 1997 (1) and has been used only sparingly since then. This is perhaps because soluble transferrin receptor testing was previously not widely available and still has limited standardization across various laboratories (2). In addition, uncomplicated IDA is often easily diagnosed with iron, total iron-binding capacity, and ferritin levels. In iron deficiency, serum (Fe) and storage (ferritin) iron levels are low. Total iron-binding capacity is high, as less iron is available to bind its receptors and therefore the binding capacity is increased. These generally are sufficient to diagnose IDA; however, ferritin is an acute-phase reactant and can rise during illness or inflammation. Most patients are admitted to the hospital with acute illness and/or inflammation, often leading to unclear results.

The anemia index is described as:

a1

Algorithm for differentiating among iron deficiency anemia IDA anemia of chronic disease ACD or a combination The anemia index which is the ratio of soluble transferrin receptor level to the l
Algorithm for differentiating among iron deficiency anemia (IDA), anemia of chronic disease (ACD), or a combination. The anemia index, which is the ratio of soluble transferrin receptor level to the logarithm to the base 10 of ferritin, indicates ACD if low (<1) or a combination of ACD and IDA if high (>2) and suggests IDA alone if very high (>>2) (1). High cytokine levels would be consistent with ACD or a combination of ACD and IDA; low cytokines would be consistent with IDA. For a very high anemia index (>>2), cytokine levels could be used to distinguish between IDA and a combination of IDA and ACD. From N Engl J Med, Weiss G and Goodnough LT, Anemia of chronic disease, 2005;352:1011-23. Copyright © 2005 Massachusetts Medical Society. Reprinted with permission from Massachusetts Medical Society.

Transferrin receptors on the plasma membrane bind circulating iron-transferrin complexes, leading to internalization and release of free iron in the cytoplasm. Transferrin receptors on the cell surface also undergo proteolytic cleavage, with soluble transferrin receptor (sTfR) released into the serum. In IDA, sTfR levels rise as transferrin receptor expression is upregulated. Our patient's sTfR level was only slightly high. The anemia index is based on the premise that in iron deficiency, sTfR levels should be markedly elevated for cells to import as much serum iron as possible, while ferritin should be low, reflecting decreased iron storage. Taking the log of ferritin refines this ratio. A ratio greater than 2 suggests true iron deficiency alone or coexisting with ACD (Figure) (1, 3). A smaller ratio suggests that ferritin level is elevated or falsely normal, as in infection, inflammation, or malignancy. Thus, a ratio of less than 1 is consistent with ACD without concurrent iron deficiency (1). If a certain diagnosis absolutely must be made, cytokine levels can be checked and are usually increased in ACD (3).

Our patient's normocytic anemia was associated with low iron level and percentage transferrin saturation and normal total iron-binding capacity and ferritin level, consistent with IDA, ACD, or both. Her anemia index was calculated as:

a2

The index of 3.8 is greater than 2, suggesting a diagnosis of combined IDA and ACD, as suspected. Fecal hemoglobin was negative, suggesting that the iron deficiency was likely due to blood loss from skin wounds. In addition, given her long-term use of metformin and pantoprazole, vitamin B12 was also checked and was low-normal with high levels of methylmalonic acid and negative intrinsic factor antibody. Her anemia was truly of multifactorial etiology, including ACD, IDA, and vitamin B12 deficiency.

The primary reason to calculate the anemia index is to help determine whether appropriate management needs to be commenced or adjusted to treat IDA and/or ACD, in accordance with the diagnosis resulting from index calculation. In our patient's case, an anemia index of 3.8 informed us of suboptimal treatment of iron deficiency. Accordingly, the amount and frequency of her iron supplementation were increased. Vitamin B12 supplementation was also initiated, and folate supplementation was continued. Therapy for her autoimmune disease was escalated, with the initiation of biologic therapy and continuation of prednisone. Pneumocystis jiroveci pneumonia prophylaxis was also started. Her wounds were treated topically with a new regimen. With these changes, weeping and bleeding improved, generalized weakness improved with physical therapy, lightheadedness resolved, and anemia remained stable prior to discharge. Two months later, her hemoglobin level was 15.4 g/dL.

The anemia index is useful in investigating the 3S's of anemia, to help distinguish among IDA, ACD, or both. Nevertheless, clinical acumen is required to determine the most likely cause of anemia in such patients and to decide on appropriate therapy.