CHAPTER 59 ACUTE BLOOD LOSS ANEMIA
CHAPTER 59 ACUTE BLOOD LOSS ANEMIA
ROBERT S. HILLMAN
Volume Loss and Replacement
Red Cell Loss and Replacement
Homologous Blood Transfusion
Maximizing Red Cell Output
The clinical manifestations of acute blood volume loss reflect adjustments in cardiac output and vascular tone that help prevent circulatory collapse and maintain oxygen supply to vital organs. The first requirement in the management of a patient with acute hemorrhage is to maintain an adequate blood volume and prevent shock. This can be accomplished by intravenous infusion of crystalloid solutions or, when available, whole blood. When blood loss is relatively slow and the total blood volume is maintained by natural or artificial means, anemia becomes a problem. The importance of this problem depends on a number of variables, including the patient’s general condition, the nature of the complicating illness, the ability of the cardiovascular system to compensate, and the flow characteristics of vital vascular pathways. A decision on blood transfusion is not based on any specific hemoglobin level but rather on a thoughtful evaluation of the anemic individual. Pre-existing cardiac or pulmonary disease, advanced age, hypertension, a history of heavy smoking, or the use of beta-adrenergic antagonists may all indicate increased morbidity risk and justify a more liberal approach to blood transfusion. Once hemorrhage has ceased, the recovery of the red cell mass to normal is usually accomplished gradually by increased red cell production.
A hemorrhage of major proportions represents a double threat to the homeostasis of the organism. First, acute severe blood loss can decrease the blood volume to a point of cardiovascular collapse, irreversible shock, and death. In this situation, the loss of circulating red cells is of far less importance than the sudden depletion of the blood volume. Second, when blood loss is more gradual, the circulating red cell mass may be so depleted as to impair oxygen delivery to vital organs. The response to these threats involves a number of physiologic mechanisms, including adjustments in cardiovascular dynamics, blood volume, red cell production, and oxygen transport by erythrocytes.1
VOLUME LOSS AND REPLACEMENT
The clinical manifestations of acute blood volume loss reflect adjustments in cardiac output and vascular tone that help prevent circulatory collapse and maintain oxygen supply to vital organs. As outlined in Table 59-1, a normal person can rapidly lose up to 20 percent of the blood volume without signs or symptoms of anemia or cardiovascular collapse. If the hemorrhage exceeds 20 percent, signs of cardiovascular distress appear. At first, this is limited to tachycardia with exercise and postural hypotension. When the blood loss exceeds 30 to 40 percent of the blood volume, there is a fall in cardiac output and the gradual onset of shock: The patient becomes immobile and exhibits air hunger; a rapid, thready pulse; and cold, clammy skin. Unless further hemorrhage is prevented and effective therapy is begun, organ damage and death ensue. A very rapid blood loss that exceeds 50 percent of the patient’s blood volume carries a high mortality rate unless immediate volume replacement therapy is initiated. With acute hemorrhage, the hemoglobin or hematocrit will not reflect the quantity of blood lost.
TABLE 59-1 REACTION TO ACUTE BLOOD LOSS OF INCREASING SEVERITY
With more gradual blood loss, sufficient restoration of plasma volume can occur to permit losses of even larger volumes of blood without the onset of shock. However, unless the physician intercedes with volume replacement therapy, plasma volume expansion is a relatively slow process. Following a sudden loss of 20 percent of the total volume, it requires 20 to 60 h to restore a normal blood volume by endogenous plasma replacement.2,6,7 In humans, this is accomplished acutely by mobilizing albumin-containing fluid from extracellular sites.7 For this reason, the hematocrit falls gradually over a period of 2 to 3 days after a sudden, single hemorrhage (Fig. 59-1). At the same time, normal individuals can produce enough albumin to tolerate chronic blood losses of 1000 ml or more each week.
FIGURE 59-1 After a sudden loss of whole blood, the fall in hematocrit is a gradual process that depends on the rate of mobilization of albumin from extravascular sites.7 Full expansion of the blood volume and the lowest he-matocrit value may not be appreciated for up to 72 h.
The first requirement in the management of a patient with acute hemorrhage is to maintain an adequate blood volume and prevent shock. This can be accomplished by intravenous infusion of crystalloid (electrolyte) solutions; colloid solutions of plasma protein, albumin, or hydroxyethyl starch; or, when available, whole blood. The choice of solution depends on the clinical setting, including such factors as the severity and rate of hemorrhage, the patient’s age and cardiovascular status, and the duration of hypotension. With hemorrhagic shock of short duration, losses are primarily from the intravascular space, with little change in extracellular and intracellular fluid compartments. In this situation, the infusion of a crystalloid solution can rapidly restore blood volume and circulation. With more prolonged hypotension, extracellular fluid shifts into both the intravascular and the intracellular fluid spaces. The latter reflects a failure of the active ATPase-dependent membrane sodium pump, with a resultant increase in intracellular levels of sodium, chloride, and water, and by an increase in extracellular potassium.8,9 To adequately resuscitate a patient suffering from severe hemorrhagic shock, large volumes of crystalloid and colloid solutions must be given quickly to replete both intravascular and exracellular fluid compartments and restore circulation to the point where cellular membrane transport can recover.
Based on this scenario, a crystalloid solution—isotonic saline or Ringer’s lactate—is the first choice in the emergency treatment of an acutely hemorrhaging patient.10 Since crystalloid solutions are rapidly distributed between the intravascular and extravascular compartments, they need to be infused in a volume of two to four times the estimated blood loss. In patients with relatively normal cardiovascular status, this will quickly return hemodynamic parameters toward normal, including the mean arterial pressure, cardiac output, systemic vascular resistance, and tissue oxygen consumption. When large volumes of crystalloid are given to elderly patients, or patients with heart disease, there is a risk of fluid overload and pulmonary edema. However, it is still debatable whether colloid solutions are any better than crystalloid solutions in supporting the blood volume.11 In pathologic states such as respiratory distress syndrome, capillary membrane integrity is altered, resulting in increased permeability of fluids with leakage of albumin into the pulmonary interstitial space. Consequently, in such patients the administration of colloidal fluids may result in the development of pulmonary edema.12
When the volume of blood lost is very large, it may be necessary to treat with a colloid solution such as 5% albumin or hydroxyethyl starch.13 Both 5% albumin in isotonic saline and a comparable product, “purified protein fraction,” provide volume-for-volume expansion in hypovolemic patients. Neither product has been found to transmit hepatitis B, hepatitis C, or HIV. An infusion of a 6% solution of hydroxyethyl starch produces a volume expansion slightly larger than the volume infused and maintains its effect as long as 24 to 36 h. The starch polymer solution contains a spectrum of molecules with different molecular weights, the smaller of which are rapidly excreted in the urine, while larger molecules require molecular degradation. The half-life of hydroxyethyl starch is 17 days, and traces of the material can be detected in the circulation for many months.14 Hydroxyethyl starch solutions are used frequently in surgery when patients undergo elective cardiac procedures and as a volume replacement fluid in pheresis therapy. Acute reactions to the starch polymer are unusual, and volumes of 2 to 3 liters of 6% hydroxyethyl starch can be administered with only minor impact on platelet function and coagulation. For the emergency situation, it is a reliable, readily available colloid expander and is relatively inexpensive.
Reliance on whole blood or packed red cells plus fresh frozen plasma for the emergency treatment of acute blood loss should be discouraged. Its use would require that large amounts of type O Rh-negative whole blood or type-specific blood be constantly available. If typing and cross-match procedures are required prior to transfusion, an unnecessary and possibly dangerous delay in therapy is introduced. In addition, whole blood cannot always be relied upon to produce adequate volume expansion. A reaction to allergenic substances within the plasma or to the cells in whole blood can interfere with volume expansion and even produce plasma volume contraction.15 Therefore, transfusion of whole blood or red blood cells should be reserved for specific treatment of a low red cell mass where tissue hypoxia is a potential threat.
RED CELL LOSS AND REPLACEMENT
With precipitous hemorrhage the immediate effects of volume depletion are more important than the loss of circulating red blood cells. Only when blood loss is relatively slow and the total blood volume is maintained by natural or artificial means does anemia become a problem. How much of a problem depends on a number of variables, including the patient’s general physical condition, the nature of the complicating illness, the ability of the cardiovascular system to compensate, and the flow characteristics of vital vascular pathways.1
While the change in hematocrit after hemorrhage occurs relatively slowly, there can be a rapid increase in the numbers of circulating leukocytes and platelets during the bleeding episode. The leukocyte count can rise to levels between 10,000 and 30,000/µl (10 and 30 × 109/liter) within a few hours as a result of a shift of marginated leukocytes into the circulation and a release of white cells from the marrow. The platelet count can rise to levels approaching 1,000,000/µl (1000 × 109/liter). In severe hemorrhage accompanied by shock and tissue hypoxia, immature elements—metamyelocytes, myelocytes, and nucleated red blood cells—may enter the circulation.
As there is no ready reserve of mature red cells to replace the lost red cell mass, oxygen supply to tissues is initially maintained by a shift in the hemoglobin oxygen dissociation curve and adjustments in cardiovascular dynamics. With sudden blood volume loss, there is reflex arteriolar constriction in oxygen-insensitive areas such as skin and kidneys and a decrease in vascular resistance in sensitive organs where oxygen delivery is essential. At the tissue level, changes in pH result in a shift of the oxygen dissociation curve to the right, the Bohr effect, and a greater release of oxygen. Over the next several hours and days, red cell levels of 2,3-bisphosphoglycerate (2,3-BPG) increase to sustain the shift in the curve. Although this mechanism may be of importance in chronic anemias,16 its effectiveness as a compensatory mechanism immediately after a hemorrhage remains to be defined. Plasma levels of erythropoietin also increase according to the severity of the anemia; a linear fall in the hemoglobin is accompanied by a logarithmic rise in plasma erythropoietin.17 This hormone is responsible for the subsequent increase in red cell production by the erythroid marrow (see Chap. 29).
Replacement of the red cell mass by increased red cell production is a gradual process. In response to erythropoietin stimulation, marrow progenitor cells must first proliferate and then mature over a period of 2 to 5 days prior to their delivery to the circulation as adult red cells. There is, therefore, a considerable time lag before red cell production can appreciably increase the red cell mass.
Erythropoietin has a specific effect on the progenitor cells, and a rising tide of erythropoietin initiates proliferation and maturation of early erythroblasts. The response of the erythroid marrow may be recognized as early as the second day by examination of a marrow aspirate. A surge in erythropoietin also appears to cause premature delivery of marrow reticulocytes to the circulation.18,19,20 The latter event may be detected within 6 to 12 h of the onset of a hemorrhagic anemia by an increase in reticulocyte counts.20 A full level of marrow production as estimated from the absolute reticulocyte count occurs only after 8 to 10 days, at which time the erythroid hyperplasia of the marrow and the absolute reticulocyte count are increased to the same extent.21
The severity of the anemia is important in determining the degree of marrow response. As long as the marrow structure is intact and iron supply to the red cell precursors is not rate-limiting, the observed increase in red cell production will usually reflect the severity of the anemia. However, damage to the kidneys, inflammation, or a hypometabolic state can markedly interfere with the response.22,23 A normal individual with an intact erythropoietin mechanism will increase marrow production by a factor of two to three times normal when the hematocrit falls below 30 percent. With progressively more severe anemia, plasma erythropoietin levels rise even higher, and marrow production can increase to levels of three to five times normal if iron supply is sufficient.21
In the majority of individuals where marrow structure and erythropoietin response mechanisms are normal, the amount of iron available to the erythroid marrow is the prime determinant of the level of marrow production (Fig. 59-2).21,22,23 With increasing anemia, the level of marrow response directly reflects the number of available iron supply pools and the rate of iron delivery from those pools.21 For example, following a gastrointestinal hemorrhage, a normal individual is able to deliver sufficient iron to support a marrow production level of no greater than three times normal despite increasingly severe anemia. This reflects the maximum rate of mobilization and delivery of storage iron from the monocyte-macrophage system. Furthermore, if these iron stores are exhausted, as is often seen with chronic blood loss, the subject is unable to increase red cell production even to this level, and the proliferative response of the marrow is severely restricted. This effect on marrow production is the earliest sign of absolute iron deficiency. It antedates by weeks or months the typical microcytosis and hypochromia of long-standing iron deficiency. In contrast, when additional iron supply pools are available, as in a subject who bleeds internally and can mobilize iron from the degraded red cells, marrow production may attain levels of four to five times normal. When large numbers of red cells are destroyed in the monocyte-macrophage system, as with a hemolytic anemia, the iron recovered from the degraded hemoglobin is even more rapidly returned to the erythroid marrow so as to permit marrow production levels exceeding five times normal. These characteristics of marrow production must be recognized in order to predict the rate of recovery of the patient’s hematocrit and plan proper therapy.
FIGURE 59-2 The rate of red blood cell production after hemorrhage reflects both the severity of the anemia and the rate of iron delivery from various sources. With red cell mass depletions of 20 percent or less, marrow production will increase to two to three times normal regardless of the source of iron. However, at lower hematocrit levels, production reflects the type of iron supply. A normal individual who must rely on hemosiderin stores in the monocyte-macrophage system is unable to increase production further (solid circles, shaded area). In contrast, in patients with a hemolytic process (circled dots) or with more than one source of iron supply (open circles), production can increase to levels of four to seven times normal when the hematocrit falls to 25 percent. Iron-deficient patients fail to show a marrow production increase at either hematocrit level (triangles).
The primary objective of red blood cell transfusions is the restoration of normal oxygen delivery to tissues.24 However, the hemoglobin level is only one of several variables determining oxygen delivery (Fig. 59-3). The addition of oxygen to inhaled respiratory gases and an increase in cardiac output achieved by optimizing cardiopulmonary hemodynamics with fluid therapy or pharmacologic intervention can compensate for acute blood loss. These interventions and the subsequent evaluation of therapeutic response should precede the decision to transfuse blood.
FIGURE 59-3 Effect of hemoglobin concentration, oxygen saturation, and cardiac output on oxygen delivery.24 The area between the two horizontal lines represents normal oxygen delivery. With increasing severity of anemia (Hb = hemoglobin in g/dl) cardiac output needs to be increased proportionally to maintain normal oxygen delivery. Increasing oxygen saturation (% Sao2) will also offer a limited degree of compensation, but the major compensatory force is cardiac output. Conversely, failure to increase cardiac output in the presence of severe anemia will lead to inadequate oxygen delivery.
HOMOLOGOUS BLOOD TRANSFUSION
The level of hemoglobin at which a blood transfusion is justified is flexible. The traditional practice of transfusing blood preoperatively when hemoglobin concentration is lower than 10 g/dl or hematocrit less than 30 percent can no longer be supported. An expert NIH panel has proposed a “transfusion trigger” of less than 7.0 g/dl with recommendations for more liberal transfusion criteria in patients at increased risk of suffering damage from decreased oxygen-carrying capacity.25 Indeed, in resting healthy subjects, isovolemic reductions of blood hemoglobin concentration to 5.0 g/dl produce no evidence of inadequate oxygen delivery because of effective compensation by a shift in the hemoglobin oxygen dissociation curve, a decrease in systemic vascular resistance, and increases in heart rate and stroke volume.26
However, patients presenting with acute blood loss are not healthy resting subjects. Hence, a decision on blood transfusion cannot be based on any specific hemoglobin level but rather on a thoughtful evaluation of the anemic individual. Preexisting cardiac or pulmonary disease, advanced age, hypertension, a history of heavy smoking, or the use of beta-adrenergic antagonists may all indicate increased morbidity risk and justify a more liberal approach to blood transfusion.27 Similarly, increased temperature, heart rate, sympathetic activity, or metabolic state may alter the balance between oxygen delivery and oxygen consumption, resulting in an increased transfusion requirement.28
Packed red blood cells are the preferred component for restoring oxygen-carrying capacity in patients with a normal coagulation status and a stable blood volume. Each unit contains about 200 ml of red blood cells with a hematocrit of about 70 to 80 percent. The infusion of one unit should raise the average-sized adult’s hematocrit by 2 to 3 percent. Packed red cells do not provide significant amounts of coagulation factors or platelets.
For massive transfusion therapy, whole blood or packed red cells together with fresh frozen plasma and platelets is preferable to packed red blood cells alone. In emergency situations, large volumes of blood may be administered rapidly by using large-bore intravenous catheters, multiple infusion sites, and infusion under pressure. The rate of infusion may be further increased by mixing red blood cells with normal saline. When transfusing large volumes of blood, careful hemodynamic monitoring and frequent hematocrit measurements are mandatory.29 Hypothermia during massive blood transfusion can be prevented by warming the blood to 37°C using high-flow blood-warming devices. Citrate intoxication may occur when massive amounts of blood are given; it can be prevented by the infusion of calcium gluconate (Chap. 140). The impact of homologous blood transfusion on survival in general and in surgical patients in particular has been eloquently demonstrated by a major study of survival in 1958 surgical patients who declined blood transfusion for religious reasons.30 The 30-day postoperative mortality was 1.3 percent in patients with a hemoglobin of 12 g/dl or greater, and 33.3 percent in patients with a hemoglobin of less than 6 g/dl. The adjusted odds ratio for mortality by cardiovascular disease according to preoperative hemoglobin showed only a modest increase in patients without cardiovascular disease with hemoglobin levels decreasing from 12 to 6 g/dl but increasing 16-fold in patients with cardiovascular disease defined by a history of angina, myocardial infarction, congestive heart failure, or peripheral vascular disease. This study illustrates the importance of identifying patients at risk in whom the ability to compensate for anemia by increased cardiac output is limited, and in whom the correction of anemia by blood transfusion may be life-saving.
Refusal of a patient on religious grounds to receive blood transfusion may result in an apparent conflict between the right of a person not to accept a service and the professional values of the physician involved in his or her management. It is useful to remember, however, that if surgical blood loss is limited to less than 500 ml, low hemoglobin levels may be well tolerated.31 Likewise, minimizing perioperative diagnostic phlebotomies, effective use of combined iron and erythropoietin treatment to correct anemia preoperatively, and the use of intraoperative blood salvage methods (see below) may limit significantly the risks of “bloodless surgery.”32 Finally, if a patient insists on avoiding transfusion after being informed of the possible consequences of such refusal, as in the case of severe anemia in a patient with cardiovascular disease, a physician is not obligated to violate his or her own professional and moral values, and arrangements can be made to transfer responsibility to another physician who is more comfortable with the patient’s decision.33,34
MAXIMIZING RED CELL OUTPUT
Every effort should be made to evaluate the adequacy of the patient’s marrow production response and institute appropriate therapy to maximize red cell output. Primarily, this involves an evaluation of iron supply and the use of oral or parenteral iron preparations when indicated. In selected patients, for example, individuals with an impaired erythropoietin response due to renal disease or chronic inflammation (the anemia of chronic disease), treatment with recombinant erythropoietin can speed recovery.35
Studies of the rate of hemoglobin regeneration in iron-deficient patients given either oral or parenteral iron have shown no significant advantage for either form of iron.36,37 Marrow production studies21,38 do show a greater increase in red cell production immediately after intravenous infusions of large amounts of iron dextran than is seen with oral iron. However, this is sustained for only 10 to 14 days; the major portion of the injected iron dextran is made available by the action of macrophages at a rate no greater than the level of iron absorbed from four oral iron tablets containing 60 mg of elemental iron each per day. Therefore, in the final analysis a single source of iron, whether normal macrophage storage iron, oral iron, or parenteral iron injections, will provide approximately the same iron supply, enough for a maximum red cell production level of three times normal. In order to exceed this limit, it is necessary to provide several sources of iron at one time. Thus, a combination of an oral iron supplement and macrophage or parenchymal iron deposits may improve iron delivery and permit marrow production to increase to levels of four to five times normal.
Once hemorrhage has ceased, the recovery of the red cell mass to normal is usually accomplished gradually without inconvenience to the patient. Serious attempts at increasing iron supply by combination therapy should therefore be reserved for those situations where a rapid maximum response is essential, as in preparation of a patient for surgery or in the treatment of prolonged, continuous hemorrhage. Blood transfusion should be reserved for those instances where normal response mechanisms and iron supplementation are insufficient to sustain an adequate red cell mass or the acuteness of the situation demands an immediate response.
Finch CA, Lenfant G: Oxygen transport in man. N Engl J Med 286:407, 1972.
Ebert RV, Stead EA Jr, Gibson JG: Response of normal subjects to acute blood loss. Arch Intern Med 68:578, 1941.
Theyl RA, Tuohy GF: Hemodynamics and blood volume during operation with ether anesthesia and unreplaced blood loss. Anesthesiology 25:6, 1964.
Howarth S, Sharpey-Schafer EP: Low blood pressure phases following hemorrhage. Lancet 1:19, 1947.
Tovey GH, Lennon GG: Blood volume studies in accidental hemorrhage. J Obstet Gynecol Br Commonw 5:749, 1962.
Lister J, McNeill IF, Marshall VC, et al: Transcapillary refilling after hemorrhage in normal man: basal rates and volumes; effect of norepinephrine. Ann Surg 158:698, 1963.
Adamson J, Hillman RS: Blood volume and plasma protein replacement following acute blood loss in normal man.JAMA 205:609, 1968.
Gann DS, Carlson DE, Brynes GJ: Impaired restitution of blood volume after large hemorrhage. J Trauma 21:598,1981.
Shires GT, Cunningham JN, Barker CRF: Alterations in cellular membrane function during hemorrhagic shock in primates. Ann Surg 176:288, 1972.
Maier RV, Carrico CJ: Developments in the resuscitation of critically ill surgical patients. Adv Surg 19:271, 1986.
Shine KI, Kuhn M, Young LS, Tillisch JH: Aspects of the management of shock. Ann Intern Med 93:723, 1980.
Velanovich VIC: Crystalloid vs colloid fluid resuscitation: a metaanalysis of mortality. Surgery 105:65, 1989.
Lamke LO, Liljedal SO: Plasma volume changes after infusion of various plasma expanders. Resuscitation 5:93,1977.
Thompson WL, Fukishima T, Rutherford RB, Walton RP: Intravascular persistence, tissue storage, and excretion of hydroxyethyl starch.Surg Gynecol Obstet 131:965, 1970.
Hutchison JK, Freedman JO, Richards BA, Burgen ASV: Plasma volume expansion and reactions after infusion of autologous and nonautologous plasma in man. J Lab Clin Med 56:734, 1960.
Torrance J, Jacobs P, Restrepo A, et al: Intraerythrocytic adaptation to anemia. N Engl J Med 283:165, 1970.
Erslev AJ: Erythropoietin. N Engl J Med 324:1339, 1991.
Hillman RS: Characteristics of marrow production and reticulocyte maturation in normal man in response to anemia.J Clin Invest 48:443, 1969.
Hillman RS, Finch CA: Erythropoiesis: Normal and abnormal. Semin Hematol 4:327, 1967.
Hillman RS, Finch CA: Red Cell Manual, 7th ed. Davis, Philadelphia, 1997.
Hillman RS, Henderson PA: Control of marrow production by the level of iron supply. J Clin Invest 48:454, 1969.
Hillman RS: The importance of iron supply in thalassemic erythropoiesis. Ann NY Acad Sci 165:100, 1969.
Erslev AJ, McKenna PJ: Effect of splenectomy on red cell production. Ann Intern Med 67:990, 1967.
Greenburg AG: A physiologic basis for red blood cell transfusion decision. Am J Surg 170:6A(suppl)44S, 1995.
NIH Consensus Conference: Perioperative red blood cell transfusion. JAMA 260:2700, 1988.
Weiskopf RB, Viele MK, Feiner J, et al: Human cardiovascular and metabolic response to acute, severe isovolemic anemia. JAMA 279:217, 1998.
Carson JL: Morbidity risk assessment in the surgically anemic patient. Am J Surg 170:32S, 1995.
Strauss RG, Weiskopf RB, AuBuchon JP: Physiology and practice of red blood cell transfusions for surgical patients, in Hematology 1998, edited by JR McArthur, GP Schechter, SL Schrier, p 454. American Society of Hematology Education Program Book. Miami Beach, 1998.
Reiner AP: Massive transfusion, in Perioperative Transfusion Medicine, edited by BD Spiess, RB Counts, SA Gould, p 351. Williams & Wilkins, Baltimore, 1995.
Carson JL, Duff A, Poses RM, et al: Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet 348:1055, 1996.
Spence RK, Carson JA, Poses R, et al: Elective surgery without transfusion: influence of preoperative hemoglobin level and blood loss on mortality.Am J Surg 159:320, 1990.
Rosengart TK, Helm RE, deBois WJ, et al: Open heart operations without transfusion using a multimodality blood conservation strategy in Jehovah’s witness patients: implications for a “bloodless” surgical technique. J Am Coll Surg 184:618, 1997.
Goldman EB: Legal considerations for allogeneic blood transfusion. Am J Surg 170:27S, 1995.
Alving BM, Spivak JL, DeLoughery TG: Consultative hematology: hemostasis and transfusion issues in surgery and critical care medicine, in Hematology 1998, edited by JR McArthur, GP Schechter, SL Schrier, p 320. American Society of Hematology Education Program Book. Miami Beach, 1998.
Watanabe Y, Fuse K, Naruse Y, et al: Subcutaneous use of erythropoietin in heart surgery. Ann Thorac Surg 54:479, 1992.
Cope W, Gillhespy RO, Richardson RW: Treatment of iron-deficiency anemia: comparisons of methods. Br Med J 2:638, 1956.
Bothwell TH, Charlton RW, Cook JD, Finch CA: Iron Metabolism in Man. Blackwell Scientific, Oxford, 1979.
Henderson PA, Hillman RS: Characteristics of iron dextran utilization in man. Blood 24:357, 1969.
Copyright © 2001 McGraw-Hill
Ernest Beutler, Marshall A. Lichtman, Barry S. Coller, Thomas J. Kipps, and Uri Seligsohn