A 77-year-old woman who had hadanorexia and weakness for 3 monthswas seen after a syncopal episode. Sheappeared pale but alert. Heart rate was110 beats per minute; respiration rate,22 breaths per minute; and blood pressure,170/70 mm Hg. Her hematologicindices were: hemoglobin level, 4.3 g/dL;mean corpuscular volume (MCV), 60fL; mean corpuscular hemoglobin concentration(MCHC), 29 g/dL; red bloodcell count, 1.6 million/μL; white bloodcell count, 7500/μL; and platelet count,452,000/μL.
3 CASES OFMICROCYTIC ANEMIA
Case 1:Older Woman WithRecent SyncopeA 77-year-old woman who had hadanorexia and weakness for 3 monthswas seen after a syncopal episode. Sheappeared pale but alert. Heart rate was110 beats per minute; respiration rate,22 breaths per minute; and blood pressure,170/70 mm Hg. Her hematologicindices were: hemoglobin level, 4.3 g/dL;mean corpuscular volume (MCV), 60fL; mean corpuscular hemoglobin concentration(MCHC), 29 g/dL; red bloodcell count, 1.6 million/μL; white bloodcell count, 7500/μL; and platelet count,452,000/μL.
The patient's syncopal episodewas orthostatic; it resulted from her severeanemia and was reflected in hervital signs. The blood smear revealedhypochromic, microcytic red bloodcells, a few target cells, and profoundred blood cell heterogeneity (Figure1). The MCV-60 fL-was well belowthe normal range (82 to 98 fL).
The first step in the evaluationof microcytosis is to measure thelevel of serum ferritin, the iron storageprotein (Algorithm I). Theserum ferritin level provides a reliable measure of storage iron.1 Theresult was 6 μg/L. A level below 20μg/L is diagnostic of iron deficiencyanemia. Levels between 20 and 100μg/L are equivocal, and levels above100 μg/L essentially rule out iron deficiency.2 Since ferritin is an acutephase reactant, levels may be falselyelevated when cytokines are releasedin inflammation, infection, or malignancy.Levels may be falsely low inpregnancy, especially during the secondand third trimester.
If this patient's ferritin levelhad been equivocal, that is, between20 and 100 μg/L, then the next stepwould have been to order a measurement of the serum transferrin receptorlevel. In iron deficiency, that levelis elevated. An elevated serum transferrinreceptor level precedes anychanges in the MCV or red blood cellvolume distribution width (RDW). Itis a valuable noninvasive test for thediagnosis of iron depletion (ie, tissueiron deprivation).3 A normal serumtransferrin receptor level, like an elevatedserum ferritin level, rules outfunctional iron deficiency.
The sensitivity and specificity ofa low serum ferritin level, as a testfor iron deficiency anemia, are 25%and 99%, respectively. The sensitivityand specificity of an elevated serumtransferrin receptor level, as a testfor the same entity, are 71% and 74%.The sensitivity and specificity of thetwo together, calculated using a sequentialalgorithm, are 67% and 93%.4
Because the most commoncause of iron deficiency is chronicblood loss, a careful history, physicalexamination, and appropriate laboratoryand diagnostic tests were indicated.Two potential sources of bleedingwere identified. The patient's stoolswere heme-negative; however, anupper endoscopy revealed a nonbleedinggastric ulcer, and colonoscopyshowed angiodysplastic areas in thececum and splenic flexure.
The anemia resolved with oraliron replacement. Measurementsof the patient's serum ferritin levelmay be used to monitor the adequacyof the patient's iron stores.
Case 2:Heavy Smoker With Symptomsof Lung DiseaseA 58-year-old man who smokedheavily (80 pack-year) complained ofhemoptysis and weakness. Wheezingwas noted over the right thorax. Hematologicindices were: hemoglobinlevel, 10 g/dL; MCV, 80 fL; MCHC,33 g/dL; RDW, 13%; red blood cellcount, 3.3 million/μL; white blood cellcount, 6800/μL; and platelet count,350,000/μL.
Bronchoscopy and biopsy confirmedthe suspected diagnosis of carcinomaof the right stem bronchus.The blood smear revealed microcyticcells with very little red blood cellheterogeneity. The MCV was low.
To narrow the differential diagnosisof this patient's microcytosis,the serum ferritin level was measured;the result was 120 μg/L. Thisvalue was just above the 100 μg/Llevel required to exclude the diagnosisof iron deficiency. Next, a measurementof the serum transferrin receptorlevel was ordered. The resultwas 4 mg/L (normal, 2 to 9 mg/L).This result was diagnostic of anemiaof chronic disease; no further testingwas necessary.5
The transferrin receptor is atransmembrane glycoprotein thatbinds with circulating diferric andmonoferric transferrin. It cycles theiron through the cytoplasm, unloadingthe iron atoms, then returns to the cell surface and disassociates.The serum level of transferrin receptoris measured by immunoassay.6
Progressive decrease in storageiron is reflected by a proportional decreasein the serum ferritin level.Progressive decrease in tissue ironis reflected by a proportional increasein the serum transferrin receptorlevel. In the anemia of chronicdisease, the serum transferrin receptorlevel remains unchanged.Transferrin receptor is not an acutephase reactant and thus remains inthe normal range in infection, inflammation,and malignancy. It also remainsin the normal range throughoutpregnancy but rises when irondeficiency supervenes.7
The serum ferritin level reflectsthe amount of storage iron, whilethe serum transferrin receptor levelreflects the amount of iron functionallyavailable in tissues. Thus, thestatus of iron in a patient's body canbe shown by combining these measurementsin a ratio of receptorlevel to ferritin level.8,9 This ratio isbelow 50 in the anemia of chronicdisease and over 100 in iron deficiencyanemia. If a patient with anemiaof chronic disease has iron deficiency,the ratio rises over 50.10When a defined cut point betweeniron deficiency and anemia of chronicdisease was used, serum transferrinreceptor levels correctly classified83% of patients considered tohave a mixture of both entities, eitheras having anemia of chronic diseaseand iron deficiency or as havinganemia of chronic disease withoutiron deficiency; the receptor-ferritinratio correctly classified 97%.11 Anotherstudy found the serum transferrinreceptor level 94% sensitiveand specific for anemia of chronicdisease coupled with iron deficiency,and the ratio 88% sensitive and 100%specific.12
The average cost of measuringserum ferritin level and serum transferrin receptor level is only a fractionof the cost of a bone marrow aspiratewith iron stain.
Case 3:Woman With Lifelong AnemiaA 60-year-old woman who was anative of India had had mild anemiaall her life and sought an explanationfor her condition. Physical examinationrevealed no splenomegaly. Her hematologicindices were: hemoglobin, 10.5g/dL; MCV, 56 fL; MCHC, 32 g/dL;RDW, 14%; red blood cell count, 8.2million/μL; white blood cell count,8000/μL; platelet count, 240,000/μL.
Although the patient had beenanemic since childhood, she had hadno symptoms of anemia. Numerousattempts had been made to resolvethe anemia through iron supplementation,with no success. The familyhistory of anemia was unknown.
The blood smear revealed microcyticred blood cells, plentiful targets,and little heterogeneity of the redcells (Figure 2). The MCV was verylow. The serum ferritin level was 250μg/L, reflecting ample storage iron.The serum transferrin receptor levelwas elevated (14 mg/L).
In this patient, the elevatedserum transferrin receptor level wasnot the result of iron deficiency, as inCase 1, but of the accelerated erythropoiesisof thalassemia.6 Since redblood cell precursors contain moretransferrin receptors than all othertissues, an increase in the number ofred cell precursors results in an increasedserum transferrin receptorlevel. This is true whether the acceleratederythropoiesis is effective orineffective, as in this patient.
The increased production ofsmaller red blood cells in thalassemiaskews the normal ratio of red cells(in millions/μL) to hemoglobin (ing/dL). In healthy persons the ratio is5:15. In iron deficiency, the ratio typicallyremains normal. In this patientthe ratio was 8.2:10.5. When patientswho are iron deficient have an ethnicbackground that suggests thalassemia(as this woman's Indian origins do) ora family history of thalassemia, thisratio can be an indication that theyhave both disorders. The presence ofboth iron deficiency and another conditionthat incites erythrocytosiswould produce a similar result.
Electrophoresis of this patient'shemoglobin revealed increases in hemoglobinF and hemoglobin A2, resultswhich are characteristic of β-thalassemia.α-Thalassemia with 3 gene deletionsalso has an abnormal hemoglobinelectrophoretic pattern (hemoglobin His present). The patterns are normal inα-thalassemia with 1 or 2 gene deletions;however, Southern blotting of endonuclease-digested genomic DNAmay be useful in the diagnosis of theselatter types of thalassemia.
2 CASES OFMACROCYTIC ANEMIA
Case 4:Man With Weakness andMild JaundiceA 58-year-old man presented withweakness, dyspnea, pallor, and mildjaundice. His liver measured 11 cmand his spleen was felt 2 cm below theleft costal margin. Other than slightjaundice, he showed no symptoms ofliver disease. Hematologic indices were:hemoglobin level, 7 g/dL; MCV, 110 fL;MCHC, 38 g/dL; RDW, 20%; red bloodcell count, 2.6 million/μL; white bloodcell count, 10,000/μL; and plateletcount, 452,000/μL.
The patient had elevated serumbilirubin and lactate dehydrogenaselevels with normal aspartate aminotransferaseand serum albumin levels.The blood smear revealed many blue,larger red blood cells, many deformedred cells, and a few spherocytes (Figure3). The MCV was high.
The first step in the evaluation ofa patient with an elevated MCV is toobtain a reticulocyte count (AlgorithmII). This patient's reticulocyte countwas 8%. The absolute reticulocytecount was 0.08 X 2.6 million/μL, or208,000/μL. In the setting of severeanemia, this level of reticulocytosis reflectsa brisk marrow response to theperipheral destruction of red cells.This patient's blood smear was consistentwith the diagnosis of hemolyticdisease. There was evidence of redcell damage (deformed red cells andspherocytes) and of a brisk bone marrow response (excessive production oflarger, blue-tinged younger red cells).
Once the diagnosis of hemolyticanemia was confirmed, a directCoombs test, or direct antiglobulintest, was needed to determinewhether the hemolysis was caused byautoimmunity.13,14 The test is performedby adding anti-immunoglobulinand anticomplement antibodies(contained in the Coombs reagent) toa patient's washed red blood cells. Ifthese cells have immunoglobulin orcomplement attached to red cell surfaceantigens, then agglutination willoccur. In this patient, the results of adirect Coombs test were positive.
A positive result on a directCoombs test may be further clarifiedby testing for anti-IgG and anti-C3 separately.15 Anti-IgG positivityalone is associated with a reaction topenicillins or methyldopa, and anti-C3 positivity alone with an anti-IgM antibody--as is sometimes producedby ingestion of quinidine or asulfonylurea. In this patient, resultsof both tests were positive, whichis characteristic of autoimmune hemolyticanemia (or of hemolysisassociatedsystemic lupus erythematosusor chronic lymphatic leukemia).Results of an antinuclearantibody test and a white blood cellcount were normal; autoimmune hemolyticanemia was diagnosed. Thepatient responded favorably to steroidtherapy.
Treating patients with autoimmunehemolytic anemia can be extremelychallenging. For example,in spite of strenuous efforts to locatecompatible blood for transfusions,none may be available. These patientsare best referred to a hematologist.
Case 5:Elderly Woman WithEcchymoses on Her LimbsA 75-year-old woman complainedof weakness, easy fatigability,and dyspnea on exertion. She appearedfrail and pale, and ecchymoseswere evident on her arms and legs.Hematologic indices were: hemoglobinlevel, 6 g/dL; MCV, 130 fL; MCHC,33 g/dL; RDW, 22%; red blood cellcount, 2.2 million/μL; white blood cellcount, 2600 /μL; and platelet count,95,000/μL.
This patient had elevated serumbilirubin and lactate dehydrogenaselevels and, in addition to an elevatedMCV, pancytopenia. The peripheralblood smear revealed large, oval redblood cells, a decreased number ofplatelets, and polymorphonuclearwhite blood cells with hypersegmentation(Figure 4). Many 5- and 6-lobed cells were seen. The bloodsmear findings were consistent witha maturation defect.
Because of the macrocytosis, areticulocyte count was ordered; theresult was 3%. The absolute reticulocytecount was 0.03 X 2.2 million/μL,or 66,000/μL. The lack of a reticulocyteresponse to the severe anemiareflected ineffective erythropoiesisand delivery of red cells to the peripheralblood. This finding pointedto a diagnosis of myelodysplastic disease,either primary or secondary.
Measurements of serum vitaminB12 and serum folate levels were orderedto help determine the cause ofthe patient's myelodysplasia. Theserum B12 level was 10 pg/mL (normal160 to 960 pg/mL) and theserum folate level was 10 ng/mL(normal 2 to 25 ng/mL). These resultsindicated a secondary cause ofmyelodysplasia, namely vitamin B12deficiency.16
The serum vitamin B12 test ispreferred for cobalamin deficiencyscreening. However, there are seriousproblems with both the sensitivityand specificity of the test. The resultis normal in 3% to 5% of patientswith established deficiency.17,18 It isnormal in transcobalamin II congenitaldefect, myeloproliferative disorders,and jejunal bacterial overgrowth.A low serum B12 level maynot indicate deficiency in 25% to 40%of patients.19
Metabolite assays providegreater sensitivity and specificity fora deficiency state. In one study, theurinary methylmalonic assay had asensitivity of 100% and a specificityof 99%.20 In another study, the methylmalonicserum assay was elevatedin 98.4% of patients with cobalamindeficiency and the serum homocysteineassay was elevated in 95.5%.21However, the use of metabolic assaysfor primary screening is probablynot indicated at the present timebecause of unresolved issues aboutthe tests and their cost. They areuseful in patients in whom B12 deficiencyis suspected but who have anormal or only mildly abnormalserum B12 level, and in patients whohave a low value but no convincingevidence of disease. Normal resultson the serum B12 test and either ofthe metabolic tests rule out cobalamindeficiency.
If results of both the serum B12and the serum folate tests had beennormal, then primary myelodysplasticdisease would have been the presumptivediagnosis. A bone marrowexamination along with cytogeneticstudies would have been ordered forconfirmation.22
If the patient had had an isolatedmacrocytic anemia, that is, macrocyticanemia with normal white blood cellsand platelets, a measurement of herserum thyroid-stimulating hormonelevel would have been ordered. Themechanism that produces macrocyticanemia in hypothyroidism without B12or folate deficiency is not understood;however, thyroid replacement therapyis an effective treatment.
Consider drug therapy andalcohol ingestion as possible causesof macrocytosis. If liver disease iscomplicated by macrocytosis,thrombocytopenia or pancytopeniamay occur.
Nutritional deficiency is a commoncause of folic acid deficiencybut is rarely a cause of B12 deficiency.Pernicious anemia, a disease ofgastric atrophy with autoantibodiesto intrinsic factor, may be accompaniedby a family history. Gastric atrophywithout autoantibodies to intrinsicfactor may present in elderly patientsas B12 deficiency; the causativedefect is failure to absorb proteinboundB12. In such patients, resultsof a Schilling test are normal andoral therapy is often effective; however,B12 levels must be monitored toensure the ongoing effectiveness oftherapy. Surgery or disease of thestomach or terminal ileum can alsoresult in B12 deficiency. In a similarmanner, diseases that affect thejejunum can result in folic aciddeficiency.
This patient needed no furtherstudies. She was treated with vitamin B12 injections and had a complete remissionof her pancytopenia. She willcontinue to receive replacementtherapy.
2 CASES OFNORMOCYTIC ANEMIA
Case 6:Woman With Breast CancerAfter an abnormality was detectedon a 53-year-old woman's mammogram,aspiration revealed carcinomaand surgery was scheduled. The patient'shematologic indices were: hemoglobinlevel, 10.8 g/dL; MCV, 86 fL;MCHC 33 g/dL; RDW, 12.8%; redblood cell count, 3.7 million/μL; whiteblood cell count, 6800/μL; and plateletcount, 250,000/μL.
The cause of this patient's anemianeeded to be clarified. The clinicalsetting pointed to anemia ofchronic disease caused by the releaseof cytokines associated with a malignancy.The blood smear revealed redblood cells with normal morphologyand little heterogeneity. The MCVwas normal.
First, the serum ferritin levelwas measured (Algorithm III). Theresult was 175 μg/L. Next, theserum transferrin receptor level wasmeasured (3 mg/L). These normalserum ferritin and serum transferrinreceptor levels were consistent withthe diagnosis of anemia of chronicdisease.
Other disorders associated withnormal serum ferritin and serumtransferrin receptor levels includerenal and liver disease, hypoadrenalism,hypothyroidism, hypoplasia, andrecent hemorrhage. Appropriate laboratoryinvestigation usually revealsthe underlying cause.
If the serum transferrin receptorlevel in this normocytic patient hadbeen elevated (but the serum ferritinlevel still normal), the following diseaseswould have been considered inthe differential:
Man With Chronic Knee Pain
A year ago, a 56-year-old manhad presented with a right knee effusion,and analysis revealed the presenceof calcium pyrophosphate crystals.Daily NSAID therapy had been prescribed.A year later, the pain in theknee was improved, but the patientcomplained of lassitude and mild shortnessof breath. Hematologic indiceswere: hemoglobin level, 11 g/dL; MCV,92 fL; MCHC, 30 g/dL; RDW, 17%;red blood cell count, 3.6 million/μL;white blood cell count, 10,400/μL; andplatelet count, 456,000/μL.
These laboratory results wereobtained after the patient had beentreated with NSAIDs for a year. Hishemoglobin level at the beginning oftherapy had been 14 g/dL. A currentblood smear revealed moderate redcell heterogeneity with a few targetcells.
Because this patient's anemiawas normocytic, the first laboratorytest ordered to investigate its cause-as with the patient in Case 6-was ameasurement of the serum ferritinlevel. The result was 80 mg/L, which,because it was between 20 and 100μg/L, was suggestive of early-stageiron deficiency anemia.
Next, the patient's serum transferrinreceptor level was measured.This value-elevated at 11 mg/dL-confirmed the diagnosis of iron deficiency.If the patient's anemia hadbeen anemia of chronic disease, hisserum transferrin receptor levelwould have been normal.
Stool specimens were intermittentlypositive for blood. An upperGI endoscopy revealed a 2.5-cm, nonbleedinggastric ulcer, which waspresumed to be the source of thebleeding. After 3 months of appropriatetherapy, a repeated endoscopyrevealed that the gastric ulcer hadhealed; the patient's hemoglobinlevel had returned to normal.
A similar combination of serumferritin level (between 20 and 100μg/L) and elevated serum transferrinreceptor level is also seen at times inpatients with iron deficiency accompaniedby folate or B12 deficiency.When these deficiencies coexist, theMCV tends to remain in the normalrange, and the RDW is almost alwayselevated. The blood smear can helpto confirm the diagnosis; typically itreveals hypersegmentation of thepolymorphonuclear leukocytes. Specifictesting for serum B12 and serumfolate levels is indicated.
As in Case 6, a serum ferritinlevel between 20 and 100 μg/L togetherwith a normal serum transferrinreceptor level would have pointedto anemia of chronic disease. Whensuch values are seen in a patient, it isuseful to look at the ratio of serumtransferrin receptor level to serumferritin level. In patients with wellestablishedchronic inflammatory disease,this ratio is usually 50 or less.When iron deficiency supervenes, asit did here, the ratio rises above 50;the ratio in this patient was 125.
REFERENCES:1. Cook JD, Lipschitz DA, Miles LE, et al. Serumferritin as a measure of iron status in normal subjects.Am J Clin Nutr. 1974;27:681-687.
2. Cook JD, Skikne BS. Iron deficiency: definitionand diagnosis. J Intern Med. 1989;226:349-355.
3. Cook JD. The measurement of serum transferrinreceptor. Am J Med Sci. 1999;318:269-276.
4. Means RT Jr, Allen J, Sears DA, et al. Serumsoluble transferrin receptor and the prediction ofmarrow aspirate iron results in a heterogeneousgroup of patients. Clin Lab Haemotol. 1999;21:161-167.
5. Ferguson BJ, Skikne BS, Simpson KM. Serumtransferrin receptor distinguishes the anemia ofchronic disease from iron deficiency anemia. J LabClin Med. 1992;19:385-390.
6. Cook JD, Skikne BS, Baynes RD. Serum transferrinreceptor. Annu Rev Med. 1993;44:63-74.
7. Carriaga MT, Skikne BS, Finley B, et al. Serumtransferrin receptor for the detection of iron deficiencyin pregnancy. Am J Clin Nutr. 1991;54:1077-1081.
8. Baynes RD, Cook JD. Current issues in iron deficiency.Curr Opin Hematol. 1996;3:145-149.
9. Baynes RD. Assessment of iron status. Clin Biochem.1996;29:209-215.
10. Pettersson T, Kivivuori SM, Siimes MA. Isserum transferrin receptor useful for detecting irondeficiency in anemic patients with chronic inflammatorydiseases? Br J Rheumatol. 1994;33:740-744.
11. Remacha AF, Sarda MP, Parellada M, et al. Therole of serum transferrin receptor in the diagnosisof iron deficiency. Haematologica. 1998;83:963-966.
12. Punnonen K, Irjala K, Rajamaki A. Serum transferrinreceptor and its ratio to serum ferritin in the diagnosisof iron deficiency. Blood. 1997;89:1052-1057.
13. Engelfriet CP, Overbeeke MAM, von demBorne AEG Kr. Autoimmune hemolytic anemia.Semin Hematol. 1992;29:3-12.
14. Domen RE. An overview of immune hemolyticanemias. Cleve Clin J Med. 1998;65:89-98.
15. Wright MS, Smith LA. Laboratory investigationof autoimmune hemolytic anemias. Clin Lab Sci.1999;12:119-122.
16. Toh BH, van Driel IR, Gleeson PA. Perniciousanemia. N Engl J Med. 1997;337:144-148.
17. Lindenbaum J, Savage DG, Stabler SP, AllenRH. Diagnosis of cobalamin deficiency, II: relativesensitivities of serum cobalamin, methylmalonicacid and total homocysteine concentrations. Am JHematol. 1990;34:99-107.
18. Lindenbaum J, Healton EB, Savage DB, et al.Neuropsychiatric disorders caused by cobalamindeficiency in the absence of anemia or macrocytosis.N Engl J Med. 1988;318:1720-1728.
19. Stabler SP, Allen RH, Savage DG, et al. Clinicalspectrum and diagnosis of cobalamin deficiency.Blood. 1990;76:871-881.
20. Matchar DB, Feussner JR, Millington DS, et al.Isotope-dilution assay for urinary methylmalonicacid in the diagnosis of vitamin B12 deficiency. AnnIntern Med. 1987;106:707-710.
21. Savage DG, Lindenbaum J, Stabler SP, et al. Sensitivityof serum methylmalonic acid and total homocysteinedeterminations for diagnosing cobalaminand folate deficiencies. Am J Med. 1994;96:239-246.
22. Heaney ML, Golde DW. Myelodysplasia. NEngl J Med. 1999;340:1649-1660.