Anemia refers to a low red blood cell (RBC) count – measured as either hemoglobin (HGB) concentration or hematocrit (HCT). Both of these measures are concentrations, so changes in hydration status or plasma volume will affect them. Pregnancy, for example, can result in an expanded plasma volume with falling HCT despite stable total body red cell mass.
A normal HGB is 12 g/dL for a woman and 14 g/dL for a man. There is also some variation expected with age, but in general, values less than these require attention. Although anemia is more common in the elderly, “anemia of the elderly” is not a diagnosis and an evaluation should be performed. Of elderly patients with anemia, one third is secondary to chronic disease, one third is secondary to a nutritional deficiency, and the remaining third is unexplained.
Not every anemic patient appears to be anemic. The evaluation of an anemia may be needed in a patient whose HCT is “normal” but not normal for the patient, e.g. in adults with congenital heart disease, severe chronic obstructive pulmonary disease, or chronic polycythemias. The cause of anemia is frequently multifactorial. For example, a patient with chronic anemia (e.g. due to cirrhosis) presents for evaluation with acute worsening.
Occasionally, an anemia evaluation is pursued in someone with a stable or normal hematocrit, but an elevated reticulocyte count indicating that red cells are being lost or destroyed at an increased rate.
The most common error in approaching anemias is assuming a single etiology, and trying to fit a single diagnosis to the situation. Forgetting uncommon diagnoses, missing simultaneous multiple etiologies or simply being overwhelmed by the possibilities are all common outcomes as well. Fortunately, the evaluation of anemia lends itself to a systematic approach.
The first step in evaluating an anemia is to consider at the reticulocyte count – either by ordering the test, or doing the thought experiment of thinking through the causes of both low, and high reticulocyte counts. Anemias are either hypoproliferative (low reticulocyte count), or hyperproliferative (high reticulocyte count). There are three possibilities:
Hyperproliferative. A rapidly falling HGB, or high reticulocyte count is seen with shortened red cell survival, or blood loss.
Hypoproliferative. Inappropriately normal, or low reticulocyte count is seen when bone marrow is unable to produce adequate red cells.
Mixed. The reticulocyte count may be elevated, consistent with a shortened red cell survival (e.g. autoimmune hemolytic anemia secondary to lymphoma), but not as elevated as one would expect (e.g. owing to lymphoma infiltrating the bone marrow).
If the reticulocyte count is high or the HGB is rapidly falling (normal RBC survival is 120 days, or roughly 1% per day), then the anemia is likely caused by loss or destruction.
Blood loss is usually obvious if it is rapid enough to cause an acute drop in HGB or rise in reticulocyte count. This is not the occult gastrointestinal (GI) loss that can result in iron deficiency, but a rapid GI bleed or trauma related loss. One must remember that large amounts of blood can be lost without overt signs or symptoms into the quadriceps and gluteal muscles, and the retroperitoneum.
Red cell destruction is due to a problem intrinsic to the red cell or in the red cell environment. Intrinsic problems can occur in one of the three major components of the RBC:
Hemoglobinopathies, which include sickle cell, thalassemias, or other unstable HGBs.
Membranopathies, which include hereditary spherocytosis, paroxysmal nocturnal hemoglobinuria, and some rare fragmentation syndromes.
Enzymopathies, which prevent the red cell from maintaining the membrane in a stable state. Glucose-6-phosphate dehydrogenase (G6PD) deficiency is a frequent cause of episodic hemolysis, but there are many other enzymopathies which present as stable anemia.
Extrinsic etiologies of hemolysis include:
Autoimmune hemolytic anemias
Drug, or toxin induced hemolysis
Sepsis and burns with multifactorial shortening of RBC survival
Cirrhosis, and hypersplenism
Mechanical destruction, or microangiopathic anemia
If the reticulocyte count is inappropriately low then bone marrow production of RBCs is inadequate.
Etiologies of bone marrow failure can be categorized by evaluating the mean cellular volume (MCV) of RBCs. The MCV is tightly regulated in healthy patients, and even modest variation is of clinical significance.
Microcytic anemia (low MCV) is commonly caused by iron deficiency, or a thalassemia; although lead poisoning, congenital sideroblastic anemia, or severe chronic disease anemia can present this way.
Macrocytic anemia (high MCV) is caused by the following:
Interference with normal RBC nuclear maturation, e.g. B12/folate deficiency, chemotherapy-like agents, immunosuppressive drugs (e.g. azathioprine, cytoxan) or myelodysplasia. These causes are often identified by the presence of megaloblasts on peripheral smear.
Specific dyslipidemias, which enlarge the RBC lipid bilayer membrane, e.g. liver disease, ethanol use, severe hypothyroidism.
Artifacts of testing such as red cell clumping caused by cold agglutinins, or the missed diagnosis of severe reticulocytosis (large, immature RBCs in peripheral blood).
Normocytic anemia (normal MCV) is a more complex anemia to sort out, and may be:
Mixed macrocytic and microcytic process (e.g. B12 with atrophic gastritis causing iron deficiency). This can often be identified by a large Red Cell Distribution Width (RDW). Normal RDW indicates RBCs are of uniform size, but high RDW indicates high variability in RBC size.
Chronic disease anemia. (Described in a separate chapter.)
Renal, or liver disease related.
Intrinsic bone marrow disease including: myeloma, aplastic anemias, lymphoma, infections, and metastatic non-hematologic malignancies. Often accompanied by thrombocytopenia, or leukopenia, and diagnosed by a bone marrow biopsy.
The most common error in approaching the anemic patient is to jump to specific tests without understanding the context. A normal G6PD level or mild iron deficiency is not helpful without the reticulocyte count, and may send you down the wrong diagnostic path. Always check the reticulocyte count, unless the cause of the anemia is very obvious (e.g. young woman with microcytic anemia and iron deficiency or obvious blood loss anemia). Always check for normalization of the anemia after therapy for the presumed cause is completed.
One way to approach the anemic patient is to think of a bucket with a hole in the bottom and a faucet running in at the top. The bucket is the patient, and the water level is the patient’s hematocrit. The rate of the faucet is the reticulocyte count, and the hole is the rate of blood loss, or hemolysis. Because RBCs are not immortal, the hole cannot be sealed. If the water level in the bucket is stable, then the rate of loss at the bottom is the same as the rate of water running in from the faucet.
If something happens to slow down the faucet, or increase size of the hole in the bottom of the bucket, then the water level will fall.
It isn’t possible to say you have a hemolytic anemia with a low reticulocyte count, and a stable HGB, just like it isn’t possible to say the faucet is turned off, but the water level is magically staying stable in spite of the hole in the bucket. The faucet and the water level, or the reticulocyte count and the stability of the HGB, are always the key to differentiating marrow failure from blood loss or RBC destruction.
The reticulocyte count can be assessed in two ways. The absolute reticulocyte count is the easiest to interpret, and requires no correction. The normal number is about 60, and reflects the production of about 1% of a patient’s RBC count per day. The reticulocyte index corrects for a patient’s level of anemia to determine if their reticulocyte count is an appropriate response. The index is calculated by multiplying the reticulocyte count by the patient’s HCT and dividing by a normal HCT.
The reticulocyte index allows the reticulocyte count to be interpreted in context. The more anemic the patient, the greater the expected reticulocyte count. A healthy woman with a HGB of 11 may only mount a modest absolute reticulocyte count of 120, and would be reflected by a normal index, whereas the same woman with a HGB of 4 would be expected to have a reticulocyte count of 400, or greater. This is an important distinction because failure to mount the expected reticulocyte count suggests either an intrinsic bone marrow defect, or an acute anemia without time to mobilize new reticulocytes.
In patients with severe anemia, reticulocytes leave the marrow early and last longer in the peripheral blood. A simple way to correct for this is to divide the reticulocyte count in half if the HGB is less than 10 (and the HCT less than 30). A more complex solution is the “reticulocyte proliferation index” which requires a table of reticulocyte maturation times for various hematocrits.
There are many layers to the history in the anemic patient. Anemia is a manifestation of many illnesses, so a thorough history is surprisingly important.
Previous history of anemia – Was it episodic with resolution vs. chronic vs. episodic without confirmation of resolution?
Family history of anemia – Did they present in childhood and did it vary in severity among members, and over time? Is there a family history of hemoglobinopathies such as thalassemia or sickle cell disease?
Chronicity – When did symptoms begin (or lack of symptoms with chronic anemias)?
Medications being taken (prescribed and over the counter), and medications prescribed but not being taken (e.g. iron, B12). This is particularly important if the anemia is hypoproliferative.
Herbal/non-western medications – This is especially important in hemolytic anemias.
Note any neurologic/behavioral changes which would force emergent consideration of thrombotic thrombocytopenic purpura (TTP) and systemic symptoms (e.g. weight loss or edema, B symptoms) suggesting underlying malignancy/renal failure, etc. Often these are only discovered in a thorough Review of Systems.
Anemia is a common manifestation of many illnesses, so a very thorough physical exam is needed. Be especially attentive to the following:
Skin manifestations of disease, vitamin deficiency, or bleeding disorders (cheilosis, tongue changes, nail changes, petechial rashes or bruising).
Abdominal exam including masses, aortic expansion, hepatosplenomegaly, testicular masses.
Splenomegaly may be a key piece of the diagnosis in a difficult case, and ultrasound confirmation may be needed.
Cardiac exam for tachycardia (acute blood loss or sepsis), murmurs (hemolysis due to endocarditis or a mechanical valve), or signs of volume overload.
Neurological exam looking for signs of B12 deficiency, neuropathy (e.g. myeloma) and TTP/atypical hemolytic uremic syndrome (aHUS) related changes.
Your first task is to determine whether you have a hypo- vs. a hyper-proliferative anemia by closely evaluating a complete blood count, and a reticulocyte count as described above. Without initial characterization of the anemia, more specific testing can lead you astray.
The peripheral blood smear is useful, but does not rule out hemolysis. It can only provide clues to the type of hemolysis, or anemia when combined with other clinical data. A smear is particularly critical in the acutely anemic patient with thrombocytopenia where a microangiopathic process (e.g. TTP, is possible. TTP and aHUS are treatable disorders), which can be rapidly fatal if a diagnosis is not made. The identification of schistocytes is critical in that setting. The peripheral smear can also be useful for identifying megaloblasts (large, dysfunctional and immature RBCs), or hypersegmented neutrophils which are both seen in megaloblastic macrocytic anemias.
An iron panel and ferritin level are often ordered to evaluate a patient with microcytic anemia for the presence of iron deficiency. Obtaining a serum ferritin is the most cost effective way to diagnose iron deficiency. A ferritin of less than 15 ng/mL almost always indicates low iron stores and a level greater than 200 ng/Ml generally indicates adequate iron stores regardless of underlying conditions. However, because ferritin is an acute phase reactant (increasing in inflammatory states, liver and renal disease, and malignancy), levels between this range can be difficult to interpret. In these settings, an iron panel may be useful. Iron deficiency is characterized by low to low normal serum iron with high total iron binding capacity (TIBC), and low transferrin saturation and ferritin. Anemia of chronic disease is characterized by low iron, and normal to low TIBC and saturation with normal, or high ferritin. (Note that anemia of chronic disease is described in more detail in a separate chapter.)
Lactate dehydrogenase (LDH) can be very useful, especially in diagnosing or ruling-out TTP, but can be elevated from many other causes and can be normal despite moderate extravascular hemolysis.
Haptoglobin, if low with an elevated LDH, is very suggestive of hemolytic anemia. Be aware, however, that haptoglobin does not assess the rate of hemolysis (as the reticulocyte count does), and can be misleading. Haptoglobin, a protein that binds free, extracellular HGB, is congenitally low in 2% of the population and may be low in the setting of liver disease. It may be appropriately low in any setting where HGB is being released-post operatively, secondary to trauma/hematomas, or with minimal hemolysis.
Specific serologic testing for hemolytic anemia, especially autoimmune hemolytic anemia, should be performed only in the acute setting when there is high suspicion for the diagnosis, or when other causes have been thoughtfully considered, and excluded in a stable patient.
Specific testing can be useful if obtained thoughtfully, but may not be needed in some clinical scenarios. A young woman with microcytic anemia probably has iron deficiency, and either treating, or testing can be appropriate in that situation. If the anemia is severe or there are issues with follow-up, then initial testing is necessary. In any case, rechecking the HGB in 6-8 weeks to confirm normalization is critical to avoid missing other causes of anemia.
Macrocytic anemia secondary to folate deficiency is uncommon in the developed world, and levels should not be checked routinely unless other diagnoses (including medications causing disruptions in folate synthesis such as methotrexate) have been ruled out. B12 deficiency can cause megaloblastic macrocytic anemia, but should not be tested routinely in other forms of anemia. In cases of borderline low B12 levels, serum methylmalonic acid, and homocysteine levels should both be elevated in B12 deficiency.
Bilirubin is an often overused test in the evaluation of anemia. It may be useful to confirm the bilirubin is normal when hemolysis is not suspected, and mildly elevated when it is. It may also be useful in the acute setting when HGB is falling, and the reticulocyte count has not yet risen, or cannot rise owing to marrow failure. Elevated bilirubin in this setting may point to hemolysis rather than bleeding as a cause for acute anemia. Bilirubin should be interpreted with caution, however, as it is very dependent on liver function and the site of hemolysis (intravascular vs. extravascular). Bilirubin may be normal with moderate extravascular hemolysis, or abnormal in a patient with hepatic dysfunction without any hemolysis.
In an acute anemia, stabilizing the patient is the primary concern. In the euvolemic patient, hypotension may be a late, pre-terminal event and tachycardia, anxiety, and tachypnea may be important signs of instability. Maintaining optimal hydration may make the HGB appear to drop, but it improves survival.
Severely anemic patients can develop high output heart failure, so careful monitoring is required. In an actively bleeding or rapidly hemolyzing patient, it may be prudent to trigger the massive transfusion protocol in your institution. This improves availability of products and, in the bleeding setting, may automate the availability of plasma/cryoprecipitate. Massive transfusion of RBCs (greater than 4 units) also requires transfusion of other blood products (e.g., platelets, and clotting factors) to prevent dilution which could exacerbate bleeding.
Transfusion should not be delayed to allow for diagnostic testing in an unstable patient. Only a few pieces of data cannot be reconstructed after a transfusion. When the diagnosis is unclear, order iron studies, a smear, and a direct and indirect coombs test prior to the transfusion, but do not delay needed therapy. Other tests can be obtained, and correctly interpreted later.
The goal of transfusion is to improve tissue oxygen delivery, not solely to increase serum HGB. Routine transfusion in patients with HGB concentration less than 7.0 g/dL is not recommended unless there is concern for ongoing or expected bleeding. There is ongoing debate regarding blood transfusion for patients with acute medical problems caused by poor oxygen delivery (e.g., active cardiac or cerebral infarction). In this population, goal HGB concentration ranges from 8 to 10 g/dL depending on the literature. Routine transfusion to HGB greater than 10 g/dL has been associated with increased mortality.
Blood transfusions are not without potential adverse events. Complications from blood transfusion include volume overload, Transfusion Related Acute Lung Injury, and infection.
A chronic anemia is typically not a medical emergency although a patient with a very stable hemolytic anemia (e.g. hereditary spherocytosis, sickle cell disease) can deteriorate rapidly if they develop a superimposed hypo-proliferative process or blood loss. In patients with known or suspected hemolytic anemias, the reticulocyte count (which should be appropriately elevated at the patient’s usual level if that is known), and close follow-up are critical. In patients with previously autoimmune hemolytic anemia, the cross-match may be difficult and blood that is given may be hemolyzed at an even a greater rate than the native blood. Nonetheless, if the patient is unstable/at risk, they should be transfused.
See above discussions for common pitfalls in work-up and management.
Cappellini, MD, Motta, I.. “Anemia in Clinical Practice- Definition and Classification: Does Hemoglobin Change with Age?”. Seminars in Hematology. vol. 52. October 2015. pp. 261-9.
Garratty, G. “Drug Induced Hemolytic Anemia”. ASH. vol. 1. 2009. pp. 73-79.
Goodnough, LT, Maggio, P.. “Restrictive blood transfusion practices are associated with improved patient outcomes”. Transfusion. vol. 54. 2014. pp. 2753-9.
Green, R, Dwyre, DM.. “Evaluation of Macrocytic Anemias”. Seminars in Hematology. vol. 52. October 2015. pp. 279-86.
Keel, S. “Anemia Chronic Inflammation”. NEJM. vol. 361. 2009. pp. 1904
Lopez, A, Cacoub, P. “Iron deficiency anaemia”. Lancet. vol. 387. Feb 2016. pp. 907-16.
Odom, D.. “Approach to Anemia”. The Washington Manual: General Internal Medicine Consult. May 2016.
Tefferi, A. “Anemia in adults: a contemporary approach to diagnosis”. Mayo Clin Proc. vol. 78. 2003. pp. 1274-1280.
Copyright © 2017, 2013 Decision Support in Medicine, LLC. All rights reserved.
No sponsor or advertiser has participated in, approved or paid for the content provided by Decision Support in Medicine LLC. The Licensed Content is the property of and copyrighted by DSM.