Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

PART TWO – General Approach to Pediatric Anesthesia

Chapter 12 – Blood Conservation in Infants and Children

  1. Ramez Salem



Adoption of Transfusion Guidelines With Lower Perioperative Hemoglobin Levels, 396



Physiologic Anemia of Infancy, 397



Oxygen Unloading in the Newborn, Infant, and Child, 397



Anemia of Prematurity, 399



Blood Transfusion Therapy, 399



Guidelines for Perioperative Management of Infants and Children With Anemia,400



Guidelines for Transfusion of Critically Ill Infants, 400



Reduction in Blood Loss, 401 



Preoperative Assessment and Preparation, 401



Surgical and Anesthetic Techniques, 401



Positioning, 401



Infiltration With Vasoconstrictors, 402



Tourniquets, 403



Prevention and Treatment of Hypothermia, 404



Minimizing Blood Sampling for Laboratory Testing,404



Pharmacologic Enhancement of Hemostatic Activity,404



Deliberate Hypotension, 405 



Historical Background, 405



Techniques for Inducing Hypotension, 405



Mode of Action of Hypotensive Drugs, 407



Hemodynamic Effects of Hypotensive Drugs, 410



Safety Factors of Induced Hypotension in Children,413



Contraindications to Induced Hypertension, 416



Autologous Blood Transfusion, 416 



Preoperative Autologous Blood Donation,416



Blood Salvage Procedures, 418



Acute Normovolemic Hemodilution, 422



Combined Techniques, 429 



Combined Hemodilution and Deliberate Hypotension,429



Hemodilution, Hypotension, and Hypothermia, 429



Oxygen Therapeutics, 430 



Hemoglobin-Based Oxygen Carriers, 430



Perfluorochemical Solutions, 430



Allosteric Modifiers, 431



Current Status of Oxygen Therapeutics,431



Summary, 431

Merriam-Webster's Collegiate Dictionary defines conservation as “a careful preservation and protection of something; especially: planned management of a natural resource to prevent… destruction…” This definition is essential as applied to blood conservation. Efforts to conserve this natural human resource during operative procedures have assumed escalating importance in recent years. The increased demands for blood and blood products for major surgery; the risks and complications of allogenic blood transfusion; the fear of transfusion-transmitted diseases, especially acquired immunodeficiency syndrome; and the progress made in techniques have all contributed to the development of blood conservation strategies, which have been extended to pediatric surgery. Anesthesiologists should be familiar with the principles and practices of blood conservation. The techniques used to conserve blood in pediatric operative procedures may be classified as (1) adoption of appropriate transfusion guidelines, (2) techniques to reduce blood loss, (3) autologous blood transfusion, (4) combination of techniques, and (5) oxygen therapeutics ( Box 12-1 ).

BOX 12-1 

Blood Conservation Strategies in Pediatric Surgery

Adoption of Appropriate Transfusion Guidelines

Reduction in Blood Loss

Preoperative assessment and preparation

Anesthetic and surgical technique


Infiltration with vasoconstrictors


Prevention and treatment of hypothermia

Minimizing blood sampling for laboratory testing

Deliberate hypotension

Pharmacologic enhancement of hemostatic activity

Autologous Blood Transfusion

Preoperative autologous blood donation

Blood salvage procedures

Acute normovolemic hemodilution

Combination of Techniques


Oxygen Therapeutics



Transfusion practices vary widely despite similar patient and surgical variables, suggesting that attention is being paid more to certain arbitrary criteria for transfusion than to the patients' actual needs. To improve transfusion practices, the American Society of Anesthesiologists Task Force on Blood Component Therapy (1994) concluded that (1) transfusion is rarely indicated when the hemoglobin concentration is greater than 10 g/dL and is almost always indicated when it is less than 6 g/dL; (2) the determination of whether intermediate hemoglobin concentrations (6 to 10 g/dL) justify or require red blood cell transfusion should be based on the patient's risk of developing complications of inadequate oxygenation; (3) the use of a single hemoglobin “trigger” for all patients, and other approaches that fail to consider all important physiological and surgical factors affecting oxygenation, is not recommended; (4) where appropriate, preoperative autologous blood donation (PABD), acute normovolemic hemodilution (ANH), intraoperative and postoperative cell recovery, and measures to decrease blood loss (deliberate hypotension and pharmacologic agents) should be used; and (5) the indications for transfusion of autologous red blood cells may be more liberal than for allogenic red blood cells because of lower risks associated with the former.

Transfusion decisions based on hemoglobin measurements alone are associated with some degree of imprecision and can be misleading because of concomitant administration of fluids. Aspects of oxygen delivery and utilization should be considered when deciding to transfuse a patient; these include blood volume, cardiac output, PaO2, myocardial ischemia, probability of massive blood loss, and factors causing increased oxygen requirements, such as pain, shivering, sepsis, or increased physical activity postoperatively. The decision to transfuse should depend on clinical assessment, aided by hemodynamic profile and indices of oxygen use (mixed venous partial pressure of oxygen [P   O2] and oxygen extraction ratio), if available. Severe chronic anemia (<7 g/dL) is typically characterized by an increase in erythrocyte 2,3-diphosphoglycerate (2,3-DPG), causing a rightward shift of the oxyhemoglobin dissociation curve that benefits oxygen delivery to the tissues (although 2,3-DPG may or may not increase in patients with renal failure). This time-dependent adaptation is not a feature of acute anemia. Consequently, lower hemoglobin levels are tolerated better in chronic than in acute anemia. When anesthetizing patients with low hemoglobin levels, high PaO2, adequate cardiac output, and adequate intravascular volume are essential, whereas factors associated with increased oxygen consumption (e.g., inadequate muscular relaxation, shivering) or leftward shifting of the oxyhemoglobin dissociation curve (e.g., hypocapnia, alkalosis, hypothermia) should be avoided (see Chapter 2 , Respiratory Physiology).

Institutions should be committed to enforce the appropriateness of allogenic transfusions. Empiric or “shotgun” therapy approaches to transfusions can no longer be justified. A quality assessment program to study, monitor, and improve transfusion practices should be implemented, and appropriate blood conservation measures may be introduced. Physician education should be an integral part of the success of such a program.

The etiology of anemia is traditionally considered under three general pathophysiologic categories: (1) decreased production of red blood cells, (2) increased destruction of red blood cells, and (3) blood loss. In infants, this classic approach to anemia is complicated by hemoglobin concentration and oxygen affinity that undergoes physiologic changes during the first few months of life. Although anemia in infants has been defined as a “reduction in hemoglobin concentration,” a more meaningful definition would be “a decrease in the level of hemoglobin content of a unit volume of blood below the level previously established as normal for age and sex.”


The mean hemoglobin concentration in cord blood at full-term birth is 16.8 g/dL, with 95% of all values falling between 13.7 and 20.1 g/dL. Based on these data, cord hemoglobin levels of less than 13 g/dL are considered abnormal ( Blanchette and Zapursky, 1984) . At birth, blood is rapidly transferred from the placenta to the infant, with 25% of the placental transfusion occurring within 15 seconds of birth and 50% occurring by the end of the first minute.

After birth there is a transient increase in hemoglobin concentration as plasma moves extravascularly to compensate for the placental transfusion and an increase in circulating red blood cell volume that occurs at the time of delivery ( Fig. 12-1 ) ( Nathan and Oski, 1987 ). Thereafter, the hemoglobin level decreases rapidly as the proportion of hemoglobin F (HbF) diminishes, reaching a level between 10.5 and 11.5 g/dL in term infants by age 8 to 12 weeks and 7 to 10 g/dL in premature infants by age 6 weeks ( Stockman and Oski, 1978 ). The high hemoglobin and reticulocyte values seen in cord blood reflect an erythropoietic response to the hypoxemic intrauterine environment. In response to the increase in arterial oxygen content after birth, erythropoietin levels decrease markedly, and there are corresponding falls in reticulocyte levels and marrow erythroid activity. Consequently, hematopoiesis virtually ceases after birth. The decrease in erythropoietin production after birth is the major cause of the physiologic anemia of infancy ( Nathan and Oski, 1987 ). Coupled with the shortened red blood cell survival in term (80 to 100 days) and premature (60 to 80 days) infants, progressive fall in hemoglobin concentration occurs during the first 1 to 3 months of life (the life span of normal red blood cells in adults is approximately 120 days). When the hemoglobin concentration falls to a level low enough to affect tissue oxygen delivery, erythropoietin production is stimulated, active marrow erythropoiesis resumes, a reticulocytosis occurs, and the hemoglobin concentration increases. In newborns with congenital cyanotic heart disease, hypoxemia continues to stimulate active erythropoiesis and the postnatal fall in hemoglobin concentration rarely occurs.


FIGURE 12-1  Hemoglobin (Hb) concentration in infants of different degrees of maturation at birth. A, Full-term infants. B, Premature infants with birth weights of 1200 to 2350 g. C, Premature infants with birth weights less than 1200 g.  (From Nathan DG, Oski FA:Hematology of infancy and childhood, 3rd ed. Philadelphia, 1987, WB Saunders, p 29, with permission.)



The umbilical venous blood in the mother has an oxygen content of approximately 13.5 vol% and oxygen saturation of 65%, but in the fetal side of the placenta, oxygen content is about 17 vol% and more than 80% saturated. The difference is caused by a beneficial leftward shift in the fetal oxyhemoglobin dissociation curve. HbF has a relatively low affinity for and a low concentration of 2,3-DPG compared with adult hemoglobin (hemoglobin-A [HbA]). Because both oxygen and 2,3-DPG compete for binding to hemoglobin, the reduced availability of 2,3-DPG causes oxygen to bind more tightly to HbF than to HbA. Before 34 weeks of gestational age, in the human fetus, hemoglobin consists of about 90% of HbF and 10% of HbA. At birth, the concentration of HbF is about 80%; thereafter, HbF decreases and HbA increases rapidly. The switchover is normally completed by 6 months of age (see Chapter 2 , Respiratory Physiology). At birth, P50 (the PO2 at which hemoglobin is 50% saturated with oxygen) is only 18 to 20 mm Hg compared with 27 mm Hg in the adult. During the first 3 months of life, there is a rapid increase in red blood cell HbA and 2,3-DPG, which results in increases in P50 (shifting of the curve to the right) as well as tissue oxygen unloading. At about 10 weeks of age, the increase in P50 reaches the adult level of 27 mm Hg. The oxyhemoglobin dissociation curve shifts farther to the right, and a level as high as 30 mm Hg may be reached by 6 to 11 months of age. Thereafter, P50 remains elevated during childhood and gradually decreases (shifting of the curve to the left), reaching the adult level of approximately 27 mm Hg during the first decade of life ( Fig. 12-2 ). This is accompanied by a gradual decrease in 2,3-DPG levels toward the adult level by 10 years of age ( Motoyama, 1990 ).


FIGURE 12-2  Schematic representation of oxyhemoglobin dissociation curves with different oxygen affinities. In infants older than 3 months with high P50 (30 mm Hg versus 27 mm Hg in adults), tissue oxygen delivery per gram of hemoglobin is increased. In neonates with a lower P50 (20 mm Hg) and a higher oxygen affinity, tissue oxygen unloading at the same tissue PO2 is reduced. Top arrow, Direction of rightward shifting of the oxyhemoglobin dissociation curve (and P50) after birth. By 10 weeks of age, the adult position of the curve is reached. Rightward shifting continues and maximum shifting is observed at 6 to 11 months of age. Lower arrow, Leftward shifting of the curve (and P50) back to the adult level, which is usually completed by approximately 10 years of age.  (From Motoyama EK: Respiratory physiology in infants and children. In Motoyama EK, Davis PJ, editors: Smith's anesthesia for infants and children, 5th ed. St Louis, 1990, CV Mosby, pp 11–76, with permission.)


In the fetus, oxygen unloading occurs across the steep portion of the oxyhemoglobin dissociation curve ( Fig. 12-3 ) ( Smith and Nelson, 1976 ). The leftward shifting of the HbF curve makes it steeper than the HbA curve, and therefore HbF unloads oxygen to the fetal tissues better than does HbA. After birth, however, the presence of HbF puts the newborn at a disadvantage, because it reduces the amount of oxygen that would be otherwise unloaded to the tissues. The decline in hemoglobin levels after birth and the rapid increase in P50 seem to be related to the process of general growth and high plasma levels of inorganic phosphate. Normal children have plasma inorganic phosphate levels that are 50% above the normal adult range. This produces major alterations in red blood cell metabolism and leads to raised levels of red blood cell adenosine triphosphate (ATP) and 2,3-DPG, with a resultant increase in P50. These observations engendered a hypothesis to explain why hemoglobin levels are lower in children than in adults: because infants (older than 3 months) and children have a lower oxygen affinity for hemoglobin (high P50) than younger infants, oxygen unloading at the tissue level is increased. Thus, a lower level of hemoglobin in these infants and children is just as efficient, in terms of tissue oxygen delivery, as a higher hemoglobin level in adults (see Chapter 2 , Respiratory Physiology).


FIGURE 12-3  Oxygen unloading capacities of fetal and adult hemoglobin before and after birth. In the fetus, fetal hemoglobin (HbF) has 15% greater oxygen unloading capacity to the tissues than does adult hemoglobin (HbA). In the newborn, HbF has 36% less oxygen unloading capacity than HbA.  (Modified from Smith CA, Nelson NM: The physiology of the newborn infant. Springfield, IL, 1976, Charles C Thomas, with permission.)




It is evident that the hemoglobin concentration per se, in physiologic as well as other types of anemia, is neither an adequate descriptor of the severity of anemia nor reflective of the adaptive factors that preserve oxygen delivery to the tissues. This led Oski (1973) to introduce the concept of “designation of anemia on a functional basis.” Motoyama (1990) compared hemoglobin requirements for equivalent tissue oxygen delivery ( Table 12-1 ). Functionally, in terms of oxygen unloading capacity, a 10 g/dL hemoglobin level in an adult is equivalent to a level of 8.2 g/dL in infants older than 3 months, whereas a level of 10 g/dL of hemoglobin in preterm infants and in infants younger than 2 months is only as good as a level of 5 to 6 g/dL of hemoglobin in infants older than 3 months and children.

TABLE 12-1   -- Hemoglobin requirement for equivalent tissue oxygen delivery


P50 (mm Hg)

Hemoglobin for Equivalent O2 Delivery (g/dL)










Infant (>3 mo)









Neonate (<2 mo)









From Motoyama EK: Respiratory physiology in infants and children. In Motoyama EK, Davis PJ, editors: Anesthesia of infants and children, 6th ed. St. Louis, Mosby, pp 11–67.



Value of Routine Preoperative Hemoglobin Measurements

The value of routine preoperative hemoglobin determinations in pediatric outpatients has been questioned. In a Canadian study ( Hackmann, Steward and Sheps, 1991 ), several interesting findings emerged, as follows. (1) The prevalence of anemia in their population was remarkably low (0.29%). (2) Anesthesiologists could not reliably predict preoperative anemia in patients presenting for outpatient surgery. (3) In the presence of mild anemia, anesthesia was safely conducted. (4) In view of the costs and patient discomfort, it was concluded that routine preoperative hemoglobin measurements may not be required. In centers where the patient population came from a lower socioeconomic background, the prevalence of anemia may be much higher and more severe than that reported within populations with a higher socioeconomic background.


In preterm infants, the hemoglobin concentration may fall steeply after birth and often reaches 8 g/dL or lower. Anemia of prematurity is attributed mainly to the abrupt increase in oxygen delivery after birth. It is characterized by diminished erythropoietin response (inappropriately low serum concentration of erythropoietin) to decreased oxygen delivery ( Stockman et al., 1984 ). Other factors may contribute to the occurrence of anemia of prematurity, including (1) diagnostic sampling; (2) more rapid destruction of fetal cells and hemodilution due to rapid growth between 30 and 40 weeks' postconception; and (3) nutritional factors, including iron, vitamin E, and folic acid deficiencies, with iron deficiency being the most important.

Iron reserves at birth are quantitatively a direct function of birth weight; the smaller the preterm infant, the greater is the risk of developing iron deficiency. At birth, the majority of body iron is in the hemoglobin fraction. During the postnatal period, iron released from the destruction of red blood cells is either stored or used for tissue growth, with practically zero excretion. Because of the diminished iron stores and the need for expansion of the red blood cell mass with rapid tissue growth, iron deficiency anemia is a common etiology of the late anemia of prematurity. This late anemia occurs at a time when storage iron has been completely utilized in new red blood cell formation and coincides with a doubling of birth weight. In term infants, this occurs by the third to sixth month after birth, whereas in preterm infants, it may occur as soon as 1 to 2 months. It is believed that iron deficiency plays no role in the early or physiologic anemia of prematurity. Other conditions, including infection, renal disease, malignancy, and nutritional deficiencies, may prevent the hematopoietic response to iron and worsen the anemia.

The current opinion is that the addition of iron should be deferred in premature infants until supplementation is really necessary. The reasons are twofold. First, the administration of iron can aggravate the anemia of prematurity. This has been attributed to iron acting as a catalyst in the nonenzymatic auto-oxidation of unsaturated fatty acids and can, in the absence of antioxidants, result in red blood cell lipase peroxidation. Second, large doses of iron may stimulate the proliferation of microorganisms and affect the host's resistance to infection.

Iron supplementation should be provided to preterm infants to prevent the “late anemia” of prematurity. A prudent decision would be to delay iron supplementation until the time of the doubling of birth weight. However, it may be given earlier if the infant has been made iron deficient iatrogenically. Supplementation should start no later than 4 months of age in term infants and no later than 2 months of age in preterm infants. The recommended dosage is 1 mg/kg per day for term infants and 2 mg/kg per day for preterm infants.


The severity of anemia in preterm infants may be limited by (1) optimization of placental transfusion at birth, (2) minimal blood loss, (3) adequate nutritional uptake, (4) clinical and laboratory monitoring, and (5) recombinant human erythropoietin therapy.

Recombinant Human Erythropoietin Therapy

Stimulation of the infant's own erythropoiesis could maintain red blood cell volume and oxygen delivery, thus decreasing the need for transfusion and minimizing the complications associated with anemia. Like the anemia of end-stage renal disease, the anemia of prematurity is associated with a specific deficiency of erythropoietin in which erythroid progenitors remain highly sensitive to erythropoietin. Preterm infants may require relatively larger doses of erythropoietin than adults (in excess of 250 mcg/kg given twice weekly) to stimulate erythropoiesis.


Two expressions have been used as a guide to transfuse critically ill infants ( Holland et al., 1987 ).



Available oxygen. This uses the difference in oxygen content between arterial blood based on a measured PaO2 and an assumed P   O2 of 20 mm Hg. When the available oxygen is calculated from infants of less than 32 weeks' gestational age, a value of less than 7 mL/dL is often associated with clinical signs of anemia, some of which respond to transfusion.



Infants' P   O2 (or central venous PO2). Studies confirmed that, of all variables examined, P   O2 correlated best with plasma erythropoietin levels ( Stockman et al., 1984 ). Thus the decline in P   O2 would seem to be the most sensitive indicator of the presence of anemia, as its value represents the integration of all the variables that determine oxygen supply and demand ( Fig. 12-4 ). When P   O2 is between 35 and 38 mm Hg (normal, ≥38 mm Hg), 41% of erythropoietin levels are above the normal range; for values between 30 and 35 mm Hg, 79% of erythropoietin levels are increased; and at P   O2 values less than 30 mm Hg, erythropoietin levels are uniformly above the normal range.


FIGURE 12-4  Changes in plasma erythropoietin (EP) concentrations in response to declines in central venous oxygen tension in preterm neonates. Arrow, Position of 38 mm Hg, which for the purposes of this figure is the lower limit of normal P   O2Horizontal dashed line, Upper limit of normal for erythropoietin, taking P   O2 ≥38 as normal.  (From Stockman JA III, Graeber JE, Clark DA, et al.: Anemia of prematurity: Determinants of the erythropoietin response. J Pediatr 105:786, 1984, with permission.)







In infants older than 3 months, hemoglobin levels of 8 g/dL or higher may be acceptable.



In infants younger than 2 months (or in preterm infants, 50 to 52 weeks' postconceptual age), a hemoglobin level of 9.5 to 10 g/dL is probably the absolute minimum.



In infants in their first week of life, infants weighing less than 1500 g, and infants with cardiac or pulmonary disease, a preoperative hemoglobin level of 12 g/dL or higher is advisable.



If the hemoglobin levels are lower than these recommended levels and the operation is purely elective, the operation may be postponed for 1 month or longer (if the risk of postponing surgery is small), especially if the anemia is associated with apneic episodes ( Welborn et al., 1991 ). The anemia should be evaluated, and supplemental iron therapy may be given during this time.



If surgery cannot be postponed, anesthesia agent may then be administered with extreme care.



The decision to transfuse intraoperatively should take into consideration the many factors that comprise clinical judgment, including blood volume estimates, preoperative hemoglobin or hematocrit, previous blood transfusion (replacement of HbF in preterm infants), duration of anemia, general condition of the patient, ability to provide adequate tissue oxygenation (cardiopulmonary function and cardiac output), extent of surgical procedure, prob-ability of massive blood loss, and risks versus benefits of transfusion.


Although precise indications for blood transfusion in critically ill infants cannot be given, a decline in available oxygen or central venous PO2 is the most sensitive indicator of the severity of anemia and is very useful in deciding when to transfuse. The following guidelines may be useful.



A cumulative record of blood losses should be kept on all critically ill infants admitted to neonatal units. An infant sufficiently ill to require frequent blood sampling may have such blood losses replaced, especially when 10% of the estimated blood volume has been exceeded. For the infant with low PaO2, there is a lower threshold for early replacement of blood withdrawn.



Infants during the first week of life who weigh less than 1500 g should have a hemoglobin value greater than 12 g/dL. In the presence of cardiac or pulmonary disease resulting in a lower PaO2, the infant hemoglobin level may be maintained in the range of 16 to 17 g/dL.



At several weeks of age, when the clinical status of the preterm infant may have been stabilized, transfusion may or may not be needed at the nadir of the anemia. (a) Infants without compromised cardiopulmonary function and in whom no unusual metabolic needs exist are unlikely to be aided by transfusions when the hemoglobin level is greater than 10 or 11 g/dL. (b) Infants who had been previously transfused with HbA are usually able to tolerate lower levels of hemoglobin because of improved tissue oxygen delivery. (c) Premature infants at the nadir of their anemia when hemoglobin levels may be as low as 7 to 8 g/dL should not receive supplemental red blood cell transfusions unless they manifest clinical signs of tissue hypoxemia.



Other indications for blood transfusion in critically ill infants include improving oxygen delivery in infants with respiratory distress syndrome and stabilizing cardiac dynamics in some forms of acyanotic congenital heart disease. In infants with respiratory distress syndrome, if PaO2 is greater than 50 mm Hg, transfusion of blood containing HbA improves oxygen delivery.

The most common blood component used is packed red blood cells, which have a hematocrit value between 70% and 80%. An average of 1 m L/kg packed red blood cells raises the hematocrit value by 1.5%.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



The importance of preoperative assessment of anemia and hemostasis cannot be overemphasized. Preoperative anemia should be thoroughly evaluated and, if time permits, appropriately treated. Patients, patients' families, or both should be questioned about history of bleeding after previous surgery; family history; severe renal, liver, and other systemic diseases; and history of medications in the past 2 weeks (nonsteroidal anti-inflammatory drugs). The clinical assessment then determines the need for further testing to investigate possible disturbances of hemostasis.


Although blood loss can be controlled by meticulous and expeditious surgical techniques, anesthesiologists should be aware of the anesthetic factors that may increase bleeding. This includes l ight anesthesia, intraoperative hypertension, systemic or regional increase in venous pressure, hyperdynamic circulation (increased cardiac output), and hypercapnia. Thus the anesthetic technique should be planned and executed to avoid increased blood loss.

During light anesthesia, bleeding may increase because of increases in skin and muscle blood flow. The redistribution of cardiac output during anesthesia has been recognized for years and has been reconfirmed during isoflurane anesthesia ( Gelman et al., 1984 ). Coughing, bucking, increased airway resistance, airway obstruction, venous obstruction, inadequate muscular relaxation, positive end-expiratory pressure (PEEP), improper positioning, fluid overload, and congestive heart failure can cause an increase in central venous pressure and increased venous oozing.

Hypercapnia may increase bleeding by augmenting cardiac output and increasing blood pressure. To avoid hypercapnia, it is essential that the anesthesiologist (1) understand the basics of anesthetic equipment, including monitoring devices; (2) detect an abnormal arterial carbon dioxide tension (PaCO2); (3) recognize potential causes of hypercapnia; and (4) correct these problems ( Salem, 1987 ). Because clinical signs are unreliable in detecting hypercapnia ( Cullen and Eger, 1974 ; Don, 1983 ; Nunn, 1993 ), it is essential that ventilation be monitored during anesthesia.


Much of the physiology of posture can be learned from the giraffe ( Warren, 1974 ). Because the head of this animal is far above its heart, the task of supplying it with oxygenated blood calls for a remarkably high blood pressure. At heart level, its blood pressure is much higher than in normal humans, approximately 260/160 mm Hg. Nevertheless, the pressure at which the brain is perfused is the same in the giraffe, humans, and most other animals ( Warren, 1974 ; Salem, 1978 ) ( Fig. 12-5 ).


FIGURE 12-5  Mean arterial pressures at the heart and brain levels in various animals and man. While there is an appreciable difference between mean arterial pressure in the giraffe and in man at the level of the heart, the pressure perfusing the brain is the same in both species.  (From Feldman Z, Narayan RK: Intracranial pressure monitoring: Techniques and pitfalls. In Cooper PR, editor. Head injury. Baltimore, 1993, Williams & Wilkins.)




In a normal subject, in the supine position the arterial (and venous) pressures are the same in various parts of the body. Moving from the supine to the erect position results in considerable changes in arterial pressures. Parts above the heart are perfused at lower pressures, whereas parts below the heart are perfused at higher pressures. Similar changes occur in the veins. In the standing position, the venous pressure is near zero above the heart and is subatmospheric in the cerebral sinuses where venous collapse is prevented ( Hainsworth, 1985 ). Below the heart level, the venous pressure progressively increases, reaching its highest levels in the feet. Vertical tilting produces a gradient of about 2 mm Hg for each inch of vertical height above which the arterial pressure is recorded, and it produces similar increases below the level of the heart. In the adult patient of average height (68 inches), the difference between arterial pressure at the head and that at the feet could be as great as 120 mm Hg. Because of their smaller stature, gravitational effects in arterial and venous pressures are less in children (and even much less in infants) than in adults. When head-up tilting is used during surgery, pressure gradients should be taken into consideration. Either an estimate of the pressure gradient is calculated at the heart and head (brain), or the transducer (in the case of arterial cannulation) is positioned parallel to the site where the perfusion pressure is measured.

With vertical tilting, twice as much blood can accumulate in the legs, and even larger volumes can be accommodated in the abdomen. Tilting also increases the capillary pressure in dependent parts, and, if the tilt is prolonged, increased filtration results in tissue edema, ultimately reducing blood volume. The converse is seen in elevated parts when tissue fluid is effectively reabsorbed.

In awake individuals, tilting initiates increased sympathetic and baroreceptor reflex activity, which leads to an increased production of catecholamines and plasma renin activity, in turn inducing the formation of angiotensin II with aldosterone release and sodium and water retention. These mechanisms result in constriction of the capacitance vessels, thus enhancing the venous return, whereas constriction of the resistance vessels (arterioles) minimizes the decrease in arterial pressure. The effects of these mechanisms become evident 15 to 30 minutes after tilting. In the normal subject, little change occurs in arterial pressure at the heart level during tilting or standing, although heart rate may increase. Anesthetics, ganglionic, and β-adrenergic–blocking drugs interfere with these compensatory mechanisms. Consequently, head-up tilt in the anesthetized patient favors arterial hypotension. Hypotension also tends to occur with the use of muscle relaxants and controlled ventilation.

Although tilting is less effective in infants and children compared with adults, the combination of decreased arterial and venous pressure above the heart and peripheral venous pooling below the heart makes tilting useful in reducing bleeding in pediatric head and neck procedures. It can also be used postoperatively to reduce swelling and edema. Added benefit may be derived in reducing postoperative laryngeal and upper airway edema after airway manipulations, instrumentation, and intubation.

If the prone position is used, as in scoliosis surgery, meticulous attention should be given to positioning so that adverse effects are avoided. The vertebral venous system provides channels into which blood may be diverted from the lower parts of the body if the inferior vena cava is partially or completely obstructed. Any rise in abdominal pressure (increased muscle tone, external pressure on the abdomen, gastric inflation, coughing, bucking, airway obstruction, and increased airway pressure) tends to increase inferior vena cava pressure and causes blood to be diverted into the vertebral venous plexuses, resulting in increased oozing during surgery. Complete relaxation of the diaphragm and the abdominal musculature decreases the intra-abdominal pressure as well as the inferior vena cava pressure and therefore is desirable to reduce bleeding ( Relton, 1975 ).

Probably no modification has greater impact on minimizing blood loss during scoliosis correction than the use of specially designed frames ( Relton and Hall, 1967 ; Relton, 1975 ; Schwentker, 1978 ). When the patient is positioned on the frame, the abdomen is free of pressure, and the pressure on the inferior vena cava is minimized ( Fig. 12-6 ). Also, by avoiding abdominal pressure, the functional residual capacity can be maintained at a near-normal level, helping to prevent atelectasis and hypoxemia.


FIGURE 12-6  Position of patient on Relton-Hall frame.  (From Schwenker EP: Posterior fusion of the spine for scoliosis. Surg Rounds 1:12, 1978, with permission.)





A commonly used method of reducing bleeding involves local infiltration of the skin and subcutaneous tissues with a solution containing a vasoconstrictor drug, usually epinephrine. An undesirable side effect of epinephrine is cardiac arrhythmias. Certain anesthetics such as halothane sensitize the myocardium to epinephrine. The dose of epinephrine needed to produce arrhythmias is lower in halothane-anesthetized patients than in those who are awake. Johnston and associates (1976) determined the median effective dose (ED50) of epinephrine in adult patients; this variable was defined by the appearance of three premature ventricular contractions at any time during or immediately after the subcutaneous injection of epinephrine. The ED50 for epinephrine and halothane was 2.1 mcg/kg; for halothane-lidocaine-epinephrine, it was 3.7 mcg/kg. The ED50 for enflurane was 10.9 mcg/kg, whereas for isoflurane it was 6.7 mcg/kg ( Fig. 12-7 ). Accordingly, for subcutaneous infiltration in adults the maximum doses of epinephrine recommended as unlikely to produce arrhythmias with halothane and isoflurane are 1, 3, and 5.5 mcg/kg, respectively. The “arrhythmogenic” dose of epinephrine during sevoflurane anesthesia seems to be similar to that during isoflurane anesthesia ( Navarro et al., 1994 ). Thiopental lowers the arrhythmogenic threshold when administered with any of the inhalation anesthetics ( Atlee and Roberts, 1986 ; Hayashi et al., 1988 ).


FIGURE 12-7  Submucosal doses of epinephrine produce PVCs in patients during 1.25 MAC inhalation anesthesia. Notice that the enflurane curve is flat.  (Modified from Johnston RR, Eger EI, Wilson C: A comparative interaction of epinephrine with enflurane, isoflurane, and halothane in man. Anesth Analg 55:516, 1976, with permission.)




Children seem to have a higher arrhythmogenic threshold for epinephrine than do adults. Ueda and associates (1983) found that a mean epinephrine dose of 7.8 mcg/kg given with lidocaine was safe during halothane anesthesia for closure of the cleft palate. Karl and others (1983) found that no arrhythmias occurred during various pediatric operations using cutaneous infiltration of 1:100,000 epinephrine (2 to 15 mcg/kg), with a wide range of halothane concentrations. They concluded that at least 10 mcg/kg of subcutaneous epinephrine could be used safely with a normal or lower-than-normal PaCO2. Furthermore, adding lidocaine to the epinephrine solution increases the margin of safety. If a large volume is required, as in scoliosis correction, a 1:500,000 solution may be used.

Infiltration with epinephrine solution to reduce bleeding is used in various minor and major pediatric surgical procedures. One drawback of epinephrine infiltration is that it may produce swelling and distortion of the tissues if an excessive volume is injected. This may interfere with a “precise” repair in certain plastic surgery operations. Close monitoring to detect arrhythmias is essential during and after epinephrine infiltration.


The use of pneumatic tourniquets after exsanguination of the upper and lower extremity decreases blood loss and permits a bloodless operative field. The tourniquet is placed over the arm, thigh, or leg (depending on the site of surgery) over cotton padding. After the limb is raised, an Esmarch bandage is applied tightly to exsanguinate that portion of the limb distal to the tourniquet. The tourniquet is then inflated to a pressure 100 mm Hg above systolic pressure for a lower extremity and 50 mm Hg above systolic pressure for an upper extremity. The size of the cuff is important; the width should be greater than half the limb's diameter. The main concerns associated with the use of tourniquets include hemodynamic, metabolic, and respiratory changes as a result of inflation and deflation; tourniquet pain; and potential injury to nerves and muscles if inflation is prolonged.

After tourniquet inflation, progressive decreases in venous pH and PO2 and increases in venous PCO2, lactate, intracellular enzymes, and potassium occur. Creatine phosphate and nicotinamide adenine dinucleotide stores decrease in muscles within 30 to 50 minutes. When the tourniquet is deflated, products of anaerobic metabolism enter the circulation, causing a transient state of reactive hyperemia and a mixed respiratory and metabolic acidosis. Mixed venous oxygen saturation (SvO2) may fall 20% in 1 minute. The accumulated acid metabolites are buffered by plasma bicarbonate, resulting in the release of CO2 and increase in PaCO2 (and PETCO2).

Time for clearance of the accumulated metabolites after tourniquet deflation depends on duration of inflation, levels of metabolites before deflation, the extremity exsanguinated (upper versus lower and one versus both), efficacy of the buffering capacity, patient's circulatory status, ventilation, and the patient's response to the extra load of metabolites. The time to maximal increase in PETCO2 is 1.5 to 2.5 minutes ( Bourke et al., 1989 ). The maximum increase in PETCO2 is approximately 3 mm Hg after the release of an upper extremity tourniquet and about 9 mm Hg after the release of a lower extremity tourniquet (Dickson et al., 1990 ).

Because of the higher metabolic rate in infants and children, it was assumed that tourniquet hemostasis might result in a greater accumulation of ischemic metabolites and that physiologic compensation might not be adequate. Lynn and associates (1986) found that children tolerated tourniquet release with fewer hemodynamic changes than have been reported in adults. A slight decrease in systolic blood pressure (8 to 10 mm Hg), lasting less than 10 minutes with no change in heart rate, was noted. The potassium levels increased slightly with tourniquet deflation but remained within the normal range. The respiratory acidosis was quickly compensated, but the metabolic acidosis persisted for longer than 10 minutes after tourniquet release. Large increases in lactate were seen with long inflation times (more than 75 minutes) and when bilateral tourniquets were used. The greatest decrease in arterial pH was seen when bilateral tourniquets were deflated simultaneously.

The following measures are recommended to minimize the systemic and metabolic consequences after tourniquet release: (1) attempt to limit inflation times to less than 75 minutes; (2) monitor PETCO2closely before and after the release of tourniquet; (3) if controlled ventilation is used, minute ventilation should be increased by 50% just before and for 5 minutes after tourniquet deflation; and (4) if a Mapleson D circuit is used, an increase in fresh gas flow may be needed to maintain PETCO2 at a near-normal level after tourniquet deflation. These measures are of great importance with long tourniquet inflation times (>75 minutes) and when bilateral tourniquets are deflated simultaneously or within 30 minutes of each other.

A progressive increase in temperature in anesthetized infants and children (0.4° to 1.6°C) can occur during prolonged leg tourniquet inflation (90 minutes), and a greater increase (1.1° to 2.3°C) can occur with bilateral leg tourniquets ( Mostello et al., 1991 ; Bloch et al., 1992 ). This slight hyperthermia may be related to a decrease in the effective heat loss from the skin and an altered distribution of heat within the body. Thus, during prolonged leg tourniquet inflation, attention should be given to the patient's temperature.

The use of tourniquets in patients with sickle cell disease has been discouraged for fear that it may lead to circulatory stasis, acidosis, and hypoxemia, the triad known to induce sickling. Clinical experience in a limited number of patients with sickle cell disease suggests that tourniquets are not associated with harmful effects, provided oxygenation and mild hyperventilation are maintained ( Adu-Gyamfi et al., 1993 ). These patients should not be denied the benefits of tourniquets when indicated, but the usual precautions for tourniquet use should be adhered to.

A dull aching pain may occur 45 minutes after tourniquet inflation in awake children undergoing limb surgery, despite a successful regional anesthetic. The pain becomes unbearable with time but subsides immediately after tourniquet deflation. Attempts to relieve upper extremity tourniquet pain by stellate ganglion block, intercostobrachial nerve block, or intravenous opioids are usually ineffective. Tourniquet pain has been correlated with the sensory level, type of regional anesthetic technique, the local anesthetic, and the dose administered. Prophylaxis includes a high level of spinal or epidural anesthesia (for the lower extremity) and the addition of an opioid to the local anesthetic used for neural blockade. Therapy includes general anesthesia or tourniquet deflation.

Prolonged ischemia results in mitochondrial swelling, myelin degeneration, depletion of glycogen storage, Z-line lysis, and tissue edema. Thromboxane is released with disruption of endothelial integrity. Within 30 minutes of inflation, nerve conduction ceases, reflecting ischemia or direct extrinsic pressure on the nerves by the tourniquet. Paralysis after tourniquet use, although very rare, is a known complication. Direct pressure caused by the cuff is probably the main cause of nerve lesions, although ischemic injury may also play a part. Tourniquet paralysis nearly always resolves spontaneously, although recovery can be slow. If prolonged inflation (>90 minutes) is required, the tourniquet should be deflated periodically every 75 to 90 minutes to minimize the risk of postoperative neurapraxis.

To minimize the risk of complications from excessive inflation pressures, the use of wider tourniquet cuffs and minimal inflation pressures has been suggested. It has also been suggested that with the use of a hypotensive technique, it is possible to decrease the applied tourniquet inflation pressure to 110 to 140 mm Hg, while maintaining a bloodless surgical field ( Tuncali et al., 2003 ). This technique may lead to reduction of complications due to excessive tourniquet inflation.


During hypothermia, platelet dysfunction, impaired coagulation, and enhanced fibrinolytic activity can occur. Valeri and associates (1992) showed that local skin hypothermia produces an increased bleeding time and a marked reduction in thromboxane B2 level at the bleeding site. Conversely, local rewarming produces an increase in shed blood thromboxane B2 level. This hemostatic defect has been attributed to the involvement of platelet glycoprotein receptor (glycoprotein Ib and granule membrane protein 140) alteration. The enzymatic reaction of the coagulation cascade is strongly inhibited by hypothermia, as demonstrated by the dramatic prolongation of prothrombin time and activated partial thromboplastin time ( Rohrer and Natale, 1992 ). The contribution of hypothermia to the hemolytic diathesis may be overlooked, because coagulation testing is performed at 37 ° C rather than at the patient's actual body temperature.

Although a few studies suggested that the indices of platelet activation during cardiopulmonary bypass vary similarly in hypothermic and normothermic patients ( Mazer et al., 1995 ), the evidence is overwhelming that hypothermia contributes to increased bleeding (Rohrer and Natalie, 1992; Valeri et al., 1992 ; Yau et al., 1992 ). When bleeding without an obvious surgical cause is encountered in a hypothermic patient, rewarming may simply correct the hypothermia-induced coagulopathy. Operating under normothermic conditions may also decrease bleeding and the use of blood products and antifibrinolytics ( Rohrer and Natale, 1992 ).


Laboratory testing is an often-overlooked source of blood loss in critically ill patients. Despite the use of micromethods, cumulative blood loss through sampling for laboratory measurements can have a great impact on hemoglobin levels. Much of the transfusion requirement in sick premature infants is a direct consequence of blood removed for laboratory use. The removal of 1 mL of blood from a 1-kg infant is equivalent to removing 70 mL of blood from an average adult. Various approaches to minimizing iatrogenic blood loss are presented in Box 12-2 .

BOX 12-2 

Techniques to Minimize Iatrogenic Blood Loss

Avoidance of Unnecessary Phlebotomy

Heightening staff awareness of iatrogenic blood loss

Charting cumulative diagnostic blood loss

Ordering tests according to need rather than by rigid protocol

Reliance on noninvasive (SaO2 and ETCO2) or continuous “in-line” arterial and mixed venous blood gas/saturation monitors

Multiple tests from a single phlebotomy (Batch ordering)

Limiting Sample Volume

Modifying laboratory procedures regarding sample volume

Use of pediatric blood collection tubes

Use of analyzers requiring smaller samples (micro-samples)

Use of whole blood stat laboratory and bedside (“point of care” technology) analyzers

Decreasing Blood Waste

Use of small dead space tubing sets

Reinfusion of dead space (discard) volume


Studies continue to define the role of antifibrinolytics, desmopressin, and aprotinin for modulating hemostasis in surgical patients, especially during cardiopulmonary bypass.


Desmopressin (DDAVP) is an analogue of the natural hormone vasopressin. Deamination of cysteine in position 1 allows for an increase in the antidiuretic, or V2, effect. Through its V2 effects, DDAVP causes endothelial cells to release von Willebrand factor, tissue-type plasminogen activator, and certain prostaglandins. An increase in the release of von Willebrand factor accounts for the hemostatic activity of DDAVP by promoting platelet adhesion to the vascular endothelium. A single dose of 10 mcg/m2 of DDAVP after the induction of anesthesia could reduce intraoperative bleeding by 30% in patients undergoing spinal fusion with normotensive anesthesia ( Kobrinsky et al., 1987 ); however, some have found that it does not reduce bleeding in patients without a known bleeding diathesis ( Guay et al., 1992 ). The differences in these findings may be attributed to the side effects of DDAVP, including the release of tissue-type plasminogen activator, producing fibrinolysis and its vasodilator effect, which tend to offset the beneficial release of von Willebrand factor. It is also possible that bleeding in certain operations on bony tissues is particularly difficult to control and may not be readily influenced by substances that modify hemostasis, because blood vessels in bone are noncollapsible structures and consequently remain open when the bone is cut. Although randomized trials have shown no benefit from DDAVP on transfusion requirements when used nonselectively, this drug has been associated with reduction in allogenic transfusion in patients receiving aspirin before cardiac surgery and in patients with documented depression of platelet function after cardiopulmonary bypass ( Mongan and Hosking, 1992 ; Dilthey et al., 1993 ).


Aprotinin, a proteinase inhibitor, has been reported to decrease blood loss during cardiac and orthopedic surgery and liver transplantation ( Van Oeveren et al., 1987 ; Alajmo et al., 1989 ; Dietrich et al., 1990 ; Havel et al., 1991 ; Janssens et al., 1994 ). Several mechanisms have been proposed to explain the observed decrease in blood loss with high-dose aprotinin:



It inhibits the fibrinolytic activity both via direct inhibition of plasmin and via inhibition of the kinin-kallikrein system. Decreased production of bradykinin reduces the release of tissue plasminogen activator, resulting in decreased formation of plasmin.



It has a protective effect on platelet function.



It partially inhibits the intrinsic coagulation pathway with preservation of the extrinsic system.

Although the exact hemostatic mechanisms of aprotinin are under investigation and remain to be elucidated, its inhibition of the intrinsic coagulation pathway has been confirmed.

In patients undergoing hip replacement, high-dose aprotinin (2 million kallikrein inactivator units [KIU] followed by an infusion of 500,000 KIU per hour until the end of surgery) resulted in a decrease in blood loss of approximately 25%, allowing almost 50% reduction in transfusion ( Janssens et al., 1994 ). No adverse effects of aprotinin on renal or hepatic function have been reported, and the incidence of deep venous thrombosis was not increased. Aprotinin treatment alone is insufficient to avoid transfusion in most patients. However, when combined with other blood conservation measures, blood transfusions are substantially decreased or virtually eliminated in certain operations. Although aprotinin is expensive, the economic benefit of reducing the requirement for blood transfusion may justify the cost.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



The concept of intentional reduction of the blood pressure to decrease blood loss and thereby decrease the need for allogenic blood transfusion and improve operative conditions for intracranial surgery was first proposed by Cushing in 1917. Gardner (1946) reported on the deliberate decrease in blood pressure in neurosurgical procedures by arteriotomy. The blood removed was kept in heparinized bottles and reinfused at the end of the procedure. The many complications (“irreversible shock,” tissue hypoxemia, acidosis, and overheparinization) indicated that the boundaries of physiologic trespass were broken down, and the technique was abandoned. Griffiths and Gillies (1948) advocated the use of high spinal analgesia to induce hypotension and to produce relatively bloodless conditions for certain operations. The major advancement with normovolemic hypotension was achieved when ganglionic blockade was combined with foot-down tilt. Enderby (1950) described the new method as “controlled circulation with hypotensive drugs and posture to reduce bleeding in surgery.” The enthusiastic initial reception was followed by reports of unexplained morbidity and mortality ( Hampton and Little, 1953 ).

After initial investigations of the use of hexamethonium (C6), one of the series of polymethylene bistrimethyl ammonium salts, attention was drawn to pentolinium, which proved to be superior to other drugs. The knowledge that triethylsulfonium salts, like the quaternary and bisquarternary ammonium ions, possess ganglionic blocking activity culminated in the synthesis of trimethaphan (also known as trimetaphan) ( Magill et al., 1953 ). Because it was short acting, trimethaphan offered a new dimension in the control of blood pressure through continuous infusion. Its short action has been surpassed only by sodium nitroprusside, which was introduced into clinical practice in 1962 by Moraca and associates.

The effects of controlled ventilation and d-tubocurarine on blood pressure had become apparent by 1952. The introduction of halothane achieved a remarkable breakthrough and allowed an easier and more gentle induction of hypotension with and without ganglionic blockade ( Enderby, 1960 ). This milestone in vascular control was responsible for a great improvement in safety ( Enderby, 1985a ). β-Adrenergic–blocking drugs were introduced to treat and prevent tachycardia ( Hellewell and Potts, 1966 ; Salem and Ivankovich, 1970 ). The discovery that labetalol given intravenously was effective in the treatment of severe hypertension prompted its use in hypotensive anesthesia ( Scott et al., 1978 ). Nitroglycerin was introduced as a hypotensive drug by Fahmy in 1978.

The use of controlled hypotension in pediatric surgery was first reported by Anderson and McKissock in 1953 . In 1974, Salem and his associates reported on the use of deliberate hypotension in 137 pediatric patients and concluded that the technique offered definite advantages and played a major role in the success of certain surgical procedures. Thereafter, deliberate hypotension became widely used in various pediatric surgical procedures, including scoliosis surgery, primary excision in burned children, vascular surgery, and neurosurgery ( McNeill et al., 1974 ; Szyfelbein and Ryan, 1974 ; Diaz and Lockhart, 1979 ). Advances in understanding of the physiology and pharmacology of deliberate hypotension, as well as in the application of the newer monitoring techniques, have all contributed to the evolution and safety of the technique.


Hypotensive drugs have been used in modern anesthetic practice to achieve one or more of the following goals ( Salem, 1978 ): (1) reduction in blood loss, (2) facilitation of vessel surgery, and (3) improvement of myocardial performance by reducing the preload and afterload.

Many drugs and techniques have been described for the induction of hypotension ( Adams, 1975 ; Salem, 1978 ; Green, 1985 ). Most of them rely on a combination of physiologic and pharmacologic means for inducing hypotension.

Two basic hypotensive anesthetic techniques currently in use are (1) the conventional technique and (2) deep anesthesia, with or without head-up tilt and PEEP ( Box 12-3 ).

BOX 12-3 

Techniques of Induced Hypotension

Requirements Before Inducing Hypotension

Stable clinical state

Secure airway (tracheal intubation)

Reliable intravenous catheters, basic monitoring

Means of accurate blood pressure measurements

Light anesthesia: inhalation anesthetic (with or without relaxants), or nitrous oxide, propofol, opioid, relaxant

Controlled ventilation

Conventional Technique

Steps of inducing gradual hypotension:



Intravenous administration of a vasodilator or a ganglionic blocker; a β-adrenergic–blocking drug may be given before the hypotensive drug



Tilting (if feasible)



Gradual increase of inhalation anesthetic (rarely used)



PEEP (very rarely needed)

Alternative Technique

Deep anesthesia (halothane or isoflurane) with tilting, PEEP, or both (very rarely needed or used)

Conventional Technique

Hypotension is to be induced only when a steady state of light anesthesia is ensured and airway patency is secured with tracheal intubation. A reliable intravenous catheter and basic monitoring, including accurate blood pressure measurements, should be established before hypotension is initiated. Light anesthesia may be achieved with either an opioid–nitrous oxide – oxygen mixture supplemented with a muscle relaxant or with a volatile anesthetic, while ventilation is controlled. With the patient in the horizontal position, hypotension is then induced gradually with the aid of a hypotensive drug (a comparison of five hypotensive drugs is shown in Table 12-2 ). The drug is administered at least 10 minutes before surgery commences if hypotension is required before the incision is made. If surgery is to be performed with the patient in the prone position, hypotension is induced after the patient is carefully positioned and proper placement of the tracheal tube is verified. It is preferable to give a β-adrenergic–blocking drug before the hypotensive drug is administered. The time of the onset of hypotension depends on the drug given (see Table 12-2 ).

TABLE 12-2   -- Comparison of five hypotensive drugs


Trimethaphan Camphorsulfonate

Sodium Nitroprusside




Onset of action

Rapid onset, usually rapid recovery

Rapid onset, very rapid recovery

Rapid onset, moderately slower recovery

Gradual onset, slower recovery

Rapid onset, rapid recovery


Short acting; may be prolonged with halothane and propranolol

Evanescent action; prolonged with cyanide toxicity

Short acting

Long acting

Very short acting

Method of administration

IV drip (0.2% in D5W solution)

IV drip (0.01% solution)

IV drip (0.01% in D5W or 0.9% NaCl)

IV injection, repeated increments

IV drip followed by constant infusion

Mode of action

Ganglionic blockade, direct effect, α-adrenergic blockade, histamine release?

Direct effect (resistance and capacitance vessels)

Direct effect, capacitance vessels predominantly

α- and β-Adrenoceptor antagonist

β1-Selective adrenoceptor antagonist


Occurs in children (less likely in adults)

Very common

May occur in children (unlikely in adults)

None, usually slight bradycardia

None, usually slight bradycardia

Cardiac output

May remain unchanged, increase, or decrease, depending on posture (venous pooling), changes in heart rate, preload and afterload, anesthetics, other myocardium-depressant drugs, ventilation, intravascular volume status

Slight decrease

Slight decrease

Blood-brain barrier

Minimal or no derangement

Pronounced dysfunction

Probably same as sodium nitroprusside

Probably none

Probably none


Unclear, inhibits plasma cholinesterase but not metabolized by it

Metabolized to cyanide and thiocyanate

Degraded rapidly in the liver

Degraded in the liver

Rapidly metabolized by esterases in cytosol of red blood cells, not by plasma cholinesterases


Unstable; kept refrigerated

Available as powder, unstable when reconstituted, protect from light, use within 12 h

Stable; colorless, absorbed by plastics, use high-density polyethylene drip set




Total IV dose should not exceed 10 mg/kg

Initial dose 0.5 to 1.5 mcg/kg/min, dose usually <8 mcg/kg/min Total dose not to exceed 1.5 mg/kg in 4 hr

10–20 mcg/kg/min

0.2 to 0.4 mg/kg followed by increments of 0.1 to 0.2 mg/kg

500 mcg/kg/min loading dose plus 300 mcg/kg/min constant infusion

Histamine release

Histamine release related to administration rate





Neuromuscular function

A weak nondepolarizing effect in large doses

Probably no effect

A weak nondepolarizing effect

Probably no effect

Slight prolongation of succinylcholineinduced neuromuscular blockade

Rebound hypertension

Does not occur

Occurs in absence of β-blockade

Does not occur

Does not occur

Does not occur

Intracranial pressure

Variable, but may decrease

Increases in early stages, less with hypocapnia

Increases, greater than with trimethaphan

Does not increase

Does not increase



After the effect of the drug (or drugs) is assessed, the patient is then tilted (head-up or foot-down). In some patients a hypotensive response may not be evident in the horizontal position, but subsequent tilting results in a decrease in the arterial pressure as a result of peripheral venous pooling. The patient should not be tilted too quickly. Cerebral autoregulation requires several minutes ( Salem, 1975 ; Patel, 1981 ). Head-up tilt may not be feasible in some operations (hip and back surgery). However, some degree of peripheral pooling may be obtained by lowering the legs while maintaining the operative site horizontally. A further decrease in arterial pressure can be obtained by a gradual increase in the concentration of the volatile anesthetic. This is particularly useful when tilting is not possible.

Deep Anesthesia With or Without Tilting and PEEP

Some anesthesiologists prefer to initiate hypotension by deepening the anesthetic level or relying on posture to maintain the hypotensive state without the use of hypotensive drugs ( Thompson et al., 1978 ;Diaz and Lockhart, 1979 ). Dose-related depression of cardiovascular function by inhalationa l anesthetics has been well documented. The extent of this depression varies among anesthetics ( Calverley et al., 1978 ; Merin, 1981 ). In healthy individuals, both halothane and enflurane have minimal effects on the systemic vascular resistance, and reduced arterial pressure is secondary to a dose-dependent decrease in left ventricular function. Isoflurane decreases the arterial pressure by decreasing the systemic vascular resistance, whereas the cardiac output is well preserved even at twice the minimum alveolar concentration (MAC) because of the associated decrease in afterload ( Lam and Gelb, 1983 ; Van Aken and Cottrell, 1986 ). The intravascular volume status contributes to the decrease in cardiac output during isoflurane anesthesia. Halothane has little effect on heart rate, whereas enflurane produces a dose-related increase in heart rate. Isoflurane also increases heart rate, although the increase is not dose related. The effects of sevoflurane and desflurane seem to be similar to those of isoflurane ( Pagel et al., 1994 ). Because of the impairment of the myocardial contractile function, the use of volatile anesthetics as the sole means of inducing hypotension is not recommended ( Prys-Roberts et al., 1974 ). With high concentrations of volatile anesthetics, blood pressure control may be difficult, and the return of blood pressure to a normal level may be sluggish when the anesthetic is discontinued.

As a result of the decreased cerebrovascular resistance, halothane increases cerebral blood flow in a dose-dependent manner. Intracranial pressure may also increase as a result of increased intracranial blood volume. Cerebral autoregulation is lost with increasing anesthetic concentration. Isoflurane (up to 1 MAC) produces a concentration-related depression of cerebral metabolism, whereas the physiologic relationships between flow and pressure and between flow and metabolism are preserved ( Van Aken and Van Hemelrijck, 1993 ). With higher concentrations, the vasodilator effect predominates; autoregulation is impaired and cerebral blood flow increases ( Van Aken and Cottrell, 1986 ; Madsen et al., 1987 ). Higher concentrations also interfere with the somatosensory evoked potential (SSEP) monitoring (see Chapter 9 , Anesthetic Equipment and Monitoring). Even low concentrations of isoflurane can provoke increases in intracranial pressure in patients with reduced intracranial compliance, and cerebral edema may occur, especially if blood pressure increases ( Grosslight et al., 1985 ). Isoflurane-induced hypotension increases both brain edema and neurologic deficits in dogs with cryogenic brain lesions in contrast to dogs given labetalol ( Bendo et al., 1993 ). A hypotensive technique based on a combination of α- and β-adrenergic receptor–blocking drugs with light isoflurane attenuates the undesirable effects of isoflurane when used as the sole hypotensive drug ( Toivonen et al., 1992 ). Because intracranial pressure increases before the dura is open and cerebral edema may develop, inhalational anesthetics should not be used as the sole means of inducing hypotension in patients with intracranial lesions ( Van Aken and Van Hemelrijck, 1993 ).

Positive End-Expiratory Pressure

Increased mean airway pressure has been used to fine-tune hypotension to the desired level ( Salem, 1978 ; Green, 1985 ). For example, systolic pressure can be decreased rapidly from 80 to 70 mm Hg by adding PEEP (10 cm H2 O), and this change can be quickly reversed by discontinuing PEEP. PEEP, however, may have the following undesirable effects:



It exerts its hypotensive effect by restricting venous return (preload) and thereby reducing cardiac output.



It tends to increase intracranial ( Shapiro and Marshall, 1978 ; Luce et al., 1982 ) and intraocular pressures ( Nimmagadda et al., 1991 ).



Cerebral venous pressure tends to increase, thus reducing perfusion pressure gradient to the brain. Some patients developed critically low jugular bulb oxygen tension, indicating a substantial decrease in cerebral blood flow ( Eckenhoff et al., 1963b ).



Maintaining high alveolar pressure in normal lungs causes increased resting lung volume and physiologic dead space ( Eckenhoff et al., 1963a ).



Excessive PEEP may diminish urine flow ( Berry, 1981 ).



In vertebral surgery, PEEP may increase the inferior vena cava pressure, resulting in a diversion of blood into the vertebral venous plexuses and increased bleeding ( Relton, 1975 ).


Normovolemic hypotension can be produced by either a reduction in cardiac output or a decrease in systemic vascular resistance. Drugs with different modes of action have been used alone or in combination with anesthetics or with other drugs to induce hypotension. These drugs may be classified as (1) ganglionic-blocking drugs (e.g., pentolinium, trimethaphan); (2) nitrovasodilators (e.g., sodium nitroprusside and nitroglycerin); (3) α-adrenergic receptor–blocking drugs (e.g., phentolamine, urapidil, nicergoline); (4) β-adrenergic receptor–blocking drugs (e.g., propranolol, esmolol); (5) drugs with combined α- and β-adrenergic receptor–blocking drugs (e.g., labetalol); (6) calcium channel–blocking drugs (e.g., verapamil, nicardipine); and (7) other vasodilators (e.g., hydralazine, adenosine, prostaglandin E1, fenoldopam).

Vasodilator drugs act primarily on the peripheral circulation. The smaller precapillary arterioles contain relatively large amounts of smooth muscle and thus are the major determinants of resistance. Because systemic vascular resistance is inversely related to arteriolar caliber by approximately the fourth power of the radius, relatively small changes in intraluminal radius have profound effects on systemic vascular resistance ( Longnecker, 1985 ).

Ganglionic-Blocking Drugs

Ganglionic-blocking drugs may compete with acetylcholine for the nicotinic receptors on the postjunctional membrane at the autonomic ganglia. Because most organs are reciprocally inner vated by sympathetic and parasympathetic nerves, the overall effect of autonomic blockade depends on the predominance of one or the other system at the end organ. The arterioles and venules of the skin and splanchnic viscera have predominantly sympathetic vasoconstrictor innervation, and ganglionic blockade produces peripheral vasodilation, increased venous capacitance, and hypotension. In contrast, the iris, ciliary muscle, gastrointestinal tract, urinary bladder, and sweat glands are all under predominantly parasympathetic control. Ganglionic blockade thus pro duces mydriasis, cycloplegia, constipation, urinary retention, and abolition of sweating. Because of these side effects, ganglionic-blocking drugs are no longer used for the treatment of hypertension. However, these side effects are of no consequence after intravenous use during anesthesia (except mydriasis and cycloplegia, which can be misinterpreted in the postoperative neurologic assessment of patients in the early postoperative period).

The hypotensive action of trimethaphan has been attributed to ganglionic blockade, a direct effect on vascular smooth muscle, β-adrenergic blockade, and histamine release ( Taylor, 1985 ). Although histamine release occurs in humans, especially after bolus injections of trimethaphan, it does not play an important role in the hypotensive effect of the drug ( Fahmy and Soter, 1985 ).


Nitric oxide is produced in cardiac and vascular endothelial cells from the amino acid L-arginine in a reaction requiring the constitutive enzyme nitric oxide synthase ( Palmer et al., 1987 ; Rees et al., 1989a, 1989b; Moncada et al., 1991 ). Nitric oxide production is stimulated by increases in intracellular calcium, in response to the interaction of a chemical agent in the blood, such as bradykinin or acetylcholine, with its specific membrane receptor or by increases in shear stress. It diffuses into the underlying vascular smooth muscle, where it stimulates production of cyclic guanosine monophosphate, thus causing vascular relaxation. The vasodilating effects of sodium nitroprusside and nitroglycerin, so-called nitrovasodilators, have been explained by their ability to provide exogenous nitric oxide ( Kruszyna et al., 1987 ). Inhaled nitric oxide has been used as a treatment for pulmonary hypertension ( Rich et al., 1994 ).

Sodium nitroprusside exerts its hypotensive action primarily by decreasing systemic vascular resistance, while the venous effect is minimal, so that cardiac output is maintained. In contrast, nitroglycerin has little effect on arteriolar resistance vessels but exhibits relatively pronounced effects on the venous capacitance vessels, resulting in decreased venous return, decreased ventricular filling pressures, and ultimately reduced cardiac output ( Longnecker, 1985 ).

β-Adrenergic–Blocking Drugs

Treatment with a β-adrenergic–blocking drug prevents the increase in heart rate, cardiac output, plasma renin activity, and catecholamine levels and prevents rebound hypertension after cessation of sodium nitroprusside infusion. Furthermore, the dose requirements of sodium nitroprusside are decreased by approximately 40%.

Propranolol given slowly in small increments up to 60 mcg/kg before or after the administration of a hypotensive drug is effective in preventing the rise in heart rate and in facilitating the control of blood pressure ( Salem et al., 1974 ). It is preferable to give propranolol before rather than after the onset of tachycardia, when a much larger dose may be needed. In this dose range, the action of propranolol is almost exclusively ascribed to blockade of the β-adrenergic receptors. Being a nonselective β-adrenergic–blocking drug, propranolol blocks both β1-(predominantly the heart) and β2-(predominantly blood vessels and bronchial smooth muscles) receptors. Increased airway resistance in normal subjects and the occurrence of bronchospasm in asthmatics after the use of propranolol led to the development of more selective β1-adrenergic–blocking drugs. The currently available β-adrenergic–blocking drugs that can be used as adjuvants to hypotensive anesthesia are listed in Table 12-3 . Some clinicians have found esmolol-induced hypotension to be more effective than sodium nitroprusside in producing better operative conditions ( Blau et al., 1992 ; Boezaart et al., 1995 ). Because these drugs can produce severe myocardial depression (especially with anesthetics), these drugs should be used mostly as adjuvants rather than as the sole hypotensive drug ( Edmondson et al., 1989 ). The advantages of esmolol are its rapid onset, titration of action, short duration, and cardioselectivity ( Sum et al., 1983 ; Menkhaus et al., 1985 ). The drug may be given in a loading dosage of 500 mcg/kg per minute for 2 to 4 minutes and continued via constant infusion at a rate of 300 mcg/kg per minute.

TABLE 12-3   -- Dose, cardioselectivity, and elimination half-life of currently available β-adrenergic–blocking drugs



Cardioselectivity (β1)

Elimination Half-life


0.06 mg/kg


4 hr


0.15 mg/kg


10 hr


0.15 mg/kg


3 to 4 hr


Loading dose 0.5 mg/kg per minute, followed by 300 mcg/kg per minute (constant infusion)


10 min


0.2 to 0.4 mg/kg depending on background anesthetic, additional increments until desired effect

0 (Has α-, β1-, and β2-adrenergic-blocking properties)

3.5 to 4.5 hr




Labetalol acts as a competitive antagonist at both α1- and β-adrenergic receptors. Its pharmacologic properties include selective blockade of α1-adrenergic receptors, blockade of β1- and β2-adrenergic receptors, partial agonist activity at β2-receptors, and inhibition of neural uptake of norepinephrine (cocaine-like effect) ( Hoffman and Lefkowski, 1990) . The potency of labetalol for β-adrenergic blockade is 5-fold to 10-fold that for α-adrenergic blockade. Ten milligrams of labetalol is the equivalent of 2 mg of propranolol at β1-receptors, 0.75 mg of propranolol at β2-receptors, and 2 mg of phentolamine (a pure α1-antagonist) at α-receptors ( Scott et al., 1978 ; Green, 1985 ; Fahmy et al., 1989 ; Goldberg et al., 1990 ). After the administration of labetalol, blood pressure decreases gradually over a 5- to 10-minute period. The hypotensive effect of labetalol is less pronounced when given with intravenous rather than inhalation anesthetics, and therefore the desired level of hypotension cannot always be achieved. Because its elimination half-life is rather long (3 to 6 hours), it may be the preferred hypotensive drug when prolonged hypotension is required. In contrast, its prolonged action may mask the adrenergic response to acute blood loss in the early postoperative period. Labetalol is usually given in an initial dose of 0.2 to 0.4 mg/kg. Incremental doses (half the initial dose) may be repeated after 5 to 10 minutes until the desired hypotension is achieved. It has advantages over sodium nitroprus side, including absence of tachycardia, no increase in cardiac output, no rebound hypertension, no increase in intrapulmonary shunt (   s/   t), and no increase in intracranial pressure even in the presence of a reduced intracranial compliance ( Van Aken et al., 1982 ).

Calcium Channel–Blocking Drugs

Nicardipine and verapamil exert their hypotensive effects primarily by decreasing systemic vascular resistance. Because verapamil also produces myocardial depression and delays atrioventricular conduction, it is not recommended for inducing hypotension. Nicardipine has been successfully used as a hypotensive drug during anesthesia. It vasodilates the peripheral, coronary, and cerebral vessels while maintaining the cardiac output without producing tachycardia. Careful titration of nicardipine infusion (10 to 250 mcg/kg per hour) is mandatory because of “increasing effect” over time and because nicardipine-induced hypotension may be resistant to conventional treatment ( Flamm et al., 1988 ; Bernard et al., 1992 ; Van Aken and Hemelrijck, 1993) . Because of these potential problems, calcium channel–blocking drugs are not recommended for use as hypotensive drugs in pediatric patients.


Adenosine, a metabolic end-product of ATP, is an endogenous vasodilator that has been implicated in local regulation of several vascular beds, including the heart and brain. Although ATP has been used as a hypotensive drug, its effect has been attributed to the arterial adenosine concentration ( Sollevi et al., 1984 ). Intravenous infusion of adenosine causes arterial hypotension that is rapidly achieved, easily controlled, short lasting, and not accompanied by rebound hypertension when the infusion is discontinued ( Lagerkranser et al., 1984 ; Sollevi et al., 1984 ; Crystal et al., 1988 ). The hypotensive effect of adenosine results entirely from a sharp decrease in systemic vascular resistance with minimal effect on the venous vascular bed because of the rapid degradation of adenosine or the decreased sensitivity of the venous vascular bed for adenosine. Because it increases the coronary blood flow and decreases the afterload, adenosine favorably influences the myocardial oxygen supply/demand balance. However, patients with coronary artery disease may develop signs of myocardial ischemia during adenosine-induced hypotension. Increased cardiac output has been noted. Although decreased heart rate occurs (probably because of direct depressive effect on the sinoatrial node) in dogs, this has not been consistent in humans. Adenosine inhibits renin release in the kidney and therefore prevents the activation of the renin-angiotensin system. β-Adrenergic–blocking drugs are not usually needed to control the heart rate. Adenosine-induced hypotension is not associated with cardiovascular, hematologic, central nervous system, renal, or hepatic toxicity. However, adenosine dilates cerebral vessels, increases intracranial pressure, and impairs autoregulation ( Lagerkranser et al., 1984 ; Van Aken et al., 1984 ). Other potential adverse features of adenosine include its paradoxical ability to cause renal vasoconstriction and heart block ( Lagerkranser et al., 1984 ; Crystal et al., 1988 ). Adenosine has not been approved for use as a hypotensive drug in the United States.

Prostaglandin E1

Prostaglandin E1 (PGE1) has a potent vasodilator effect on the pulmonary and systemic vascular beds. Despite its vasodilator properties, PGE1 infusion causes slight hypotension in conscious humans because of the concomitant increase in cardiac output secondary to its positive inotropic effect and the reflex increase in heart rate ( Carlson et al., 1969 ). PGE1 infusion (100 to 150 ng/kg per minute) has been successfully and safely used to induce hypotension during inhalation anesthesia. Resistance was encountered in 2 of 14 patients ( Goto et al., 1982 ). Blood pressure returns to within 15% of normal 15 minutes after the infusion is stopped.

Increase in plasma renin activity occurs during PGE1 infusion, presumably due to the fall in pressure rather than a direct effect on the kidney. PGE1 also causes increases in renal blood flow, urine flow, and sodium excretion ( Goto et al., 1982 ). Although an inhibitory effect on both platelet aggregation and thrombus formation has been described, PGE1 in clinical doses has no discernible effect on platelet aggregation ( Goto and Fujita, 1980 ). Use of PGE1 as a hypotensive drug is not currently recommended pending further investigation.


Fenoldopam is a pure dopamine1 receptor antagonist with selective coronary, renal, mesenteric, and peripheral arteriolar vasodilator action. Mild reflex tachycardia usually occurs with its use. The maximal response is usually achieved in 10 to 20 minutes. Fenoldopam is given via a continuous infusion (0.1 to 0.6 mcg/kg per minute). Unlike other hypotensive drugs, fenoldopam has renal vasodilatation and natriuretic actions that maintain or increase urine flow during hypotension. These properties may constitute a renal protective effect ( Aronson et al., 1990 ).


Resistance to Hypotensive Drugs

After the introduction of hypotensive drugs into anesthetic practice, the problem of failure to achieve or maintain hypotension was soon observed. This phenomenon was assumed to be due to tachyphylaxis, that is, a diminished response despite continued administration of the drug. However, resistance to hypotensive drugs is almost always associated with tachycardia. This is more frequently seen in children than in adults ( Salem et al., 1974 ). Resistance to hypotension is not a unique feature of one drug but has been reported with all ganglionic-blocking drugs, β-adrenergic–blocking drugs, and other direct-acting drugs, including sodium nitroprusside, nitroglycerin, and phentolamine.

Several mechanisms have been postulated to explain tachycardia or resistance to hypotension, or both ( Box 12-4 ). The relative importance of these mechanisms may differ with the various hypotensive drugs. Sodium nitroprusside–induced hypotension is associated with increases in heart rate and cardiac output, activation of the renin-angiotensin system, and release of catecholamines ( Khambatta et al., 1981 ; Knight et al., 1983 ; Fahmy et al., 1984 ). In contrast, ganglionic blockade results in less of an increase in circulating catecholamines and no activation of the renin-angiotensin axis (Knight et al., 1980, 1983 [114] [115]). The increase in heart rate with the use of ganglionic-blocking drugs is probably the result of parasympathetic blockade, which may be more prominent in children because of their increased vagal tone ( Salem et al., 1978 ; Yaster et al., 1986 ). In response to the initial decrease of pressure, reflex tachycardia, mediated through the baroreceptors, occurs with almost all hypotensive drugs. Because the cardiac output in children is rate dependent, increased heart rate results in an increase in cardiac output and counteracts the decrease in blood pressure. As a result of sympathetic activation, renin is released from the juxtaglomerular apparatus in the kidney. This acts on an α2-globulin from the liver to produce the decapeptide angiotensin I, which is converted in the lungs to the octapeptide angiotensin II, a potent vasoconstrictor.

BOX 12-4 

Theories Postulated to Explain Tachycardia and Resistance to Hypotension



Parasympathetic blockade causing tachycardia (ganglionic blockade)



Reflex tachycardia mediated through the baroreceptors in response to the initial fall in pressure (probably all hypotensive drugs)



Stimulation of the sympathetic and renin-angiotensin systems leading to increased plasma renin activity and angiotensin II and catecholamine levels (direct-acting drugs, especially sodium nitroprusside)



Constrictive effect of sympathetic blockade on the α-adrenoceptive blood vessels leading to rise in blood pressure



Inappropriate (excessive) fluid therapy before and during hypotension (causing expansion of blood volume and difficulty in controlling pressure)

Stimulation of the sympathetic and the renin-angiotensin systems may adversely affect the operative course during induced hypotension. Even when hypotension is established, the increased cardiac output can cause bleeding. Rebound hypertension may occur after abrupt termination of the sodium nitroprusside infusion. This is related to the increase in systemic vascular resistance because of the unopposed activation of the sympathetic and renin-angiotensin systems. The consequences of rebound hypertension include wound bleeding, hematoma formation, cerebral edema, cerebrovascular accidents, disrupted cerebral autoregulation, increased myocardial oxygen demand, and pulmonary edema. The varying sensitivities of vessels to catecholamines in different organs may cause major changes in the distribution of blood flow among the various organs, as well as a redistribution of blood flow within organs. Evidence suggests that very high circulating catecholamine and angiotensin II levels have deleterious effects on myocardial and renal tubular cells and may adversely affect arterial and capillary function ( McDonald et al., 1969 ; Giese, 1973 ; Gavras et al., 1975 ).

The key to successful hypotension is control of the heart rate. The techniques advocated to prevent resistance to hypotensive drugs are shown in Box 12-5 . The simplest effective means is the judicious use of a β-adrenergic–blocking drug. Other alternatives include pretreatment with saralasin, an angiotensin II competitive antagonist, and captopril (clonidine), an oral angiotensin-converting enzyme inhibitor (Delaney and Miller, 1980 ; Fahmy and Gavras, 1985 ). Pretreatment with captopril results in lower dosage requirements and prevents rebound hypertension. A 10:1 mixture of trimethaphan (250 mg) and sodium nitroprusside (25 mg) in a solution of 5% dextrose in water has been advocated ( Wildsmith et al., 1983 ; Fahmy, 1985a ). The mixture produces hypotension with smaller doses of sodium nitroprusside and trimethaphan than when either drug is used separately (synergistic effect). When the mixture is used, the dose of sodium nitroprusside is approximately one third to one fifth the amount required when sodium nitroprusside is used alone and blood pressure returns gradually without rebound hypertension.

BOX 12-5 

Techniques to Prevent Resistance to Hypotensive Drugs

Halothane (inactivates the baroreceptor response)

Avoiding fluid overload

Omitting belladonna drugs (or decreasing dosage)

Preoperative sedation and intravenous opioids

Use of a β-adrenergic–blocking drug

Pretreatment with angiotensin II competitive antagonist

Pretreatment with angiotensin-converting enzyme inhibitor

Combining hypotensive drugs

Premedication with clonidine

Clonidine, an α2-adrenoceptor agonist, has been shown to reduce the requirement of isoflurane (by 60%) and sodium nitroprusside (by 45%) and to substantially reduce the need for labetalol during induced hypotension ( Woodcock et al., 1988 ). Through its central effect at sites within the medulla and hypothalamus, clonidine effectively blocks the increased central adrenergic activity concomitant with the use of hypotensive drugs. Other actions include inhibition of renin release in the kidney and a reduction in vasopressin release. The recommended dose is 4 to 8 mcg/kg to be given orally 2 hours before surgery. The main disadvantage associated with premedication with clonidine is a slightly prolonged recovery from anesthesia ( Maroof et al., 1994 ). Clonidine can also be given intravenously in a dose of 2 to 3 mcg/kg as an adjuvant for controlling blood pressure. In addition to the intraoperative hypotensive effect, the analgesic and sedative effects of the drug can last well into the postoperative period.

Effects on Cardiac Output and Regional Blood Flows

The cardiovascular effects of hypotensive drugs may be modified by many factors: the anesthetic and other drugs given, position of the patient, degree of hypotension, intrathoracic pressure, acid-base status, circulating blood volume, age, and change in preload and afterload.

Cardiac Output

Vasodilators alter the cardiac output through changes in stroke volume and heart rate. They decrease afterload by lowering arterial impedance and reduce preload by increasing venous compliance. Depending on the ventricular function as defined by the Frank-Starling mechanism and the relative effect of the drug on the preload and afterload, stroke volume may rise, remain constant, or even decline (Fahmy and Laver, 1976 ). A decrease in afterload shifts the curve to the left ( Fig. 12-8 ). If preload remains constant, stroke volume increases. If the decrease in afterload is associated with a decrease in preload, stroke volume probably remains constant. A decrease in preload without a decrease in afterload probably results in decreased stroke volume.


FIGURE 12-8  Effect of a vasodilator on the Frank-Starling relation. Decreased afterload shifts the curve (solid line) to the left (dashed line). If transmural pressure (preload) remains constant, cardiac output increases; if the decrease in afterload is also associated with decreased preload (transmural pressure), cardiac output remains constant. VC, venous compliance.  (From Fahmy NR, Laver MB: Hemodynamic response to ganglionic blockade with pentolinium during N2O-halothane anesthesia in man. Anesthesiology 44:6, 1976, with permission.)




Changes in cardiac output during deliberate hypotension are a function of heart rate. Changes in heart rate depend mainly on the predominant autonomic tone existing at the time of hypotension. For example, if vagal tone is predominant, heart rate increases via the baroreceptor pathway, as in the case of children. Halothane blunts this reflex increase in heart rate during hypotension by “resetting the baroreceptors,” depressing vasomotor centers, and depressing sinoatrial node activity.

In anesthetized humans with normal cardiac function, sodium nitroprusside–induced hypotension is associated with either increased or unchanged cardiac output. In contrast, nitroglycerin, because of its effect on venous capacitance and venous pressure, decreases ventricular filling pressure and ultimately reduces cardiac output ( Fahmy, 1978 ; Gerson et al., 1982 ; Longnecker, 1985 ). Hypotension secondary to ganglionic blockade has variable effects on cardiac output.

Coronary Circulation

The driving pressure for coronary blood flow is aortic diastolic pressure. During deliberate hypotension, reduced myocardial work secondary to decreased afterload reduces the requirement for coronary blood flow. A decrease in the heart rate × systolic pressure product (an index of myocardial oxygen demand) has been observed during deliberate hypotension in adults and children ( Fahmy and Laver, 1976 ; Salem et al., 1978 ). Unless hypotension is severe, locally mediated changes in vasomotor tone ensure that coronary blood flow remains adequate during deliberate hypotension. This may explain the lack of electrocardiographic evidence of myocardial ischemia during induced hypotension in children ( Salem et al., 1974 ).

Cerebral Circulation

The physiologic response to a change in arterial pressure is a compensatory change in cerebral vascular resistance, that is, autoregulation. In the healthy unanesthetized human, cerebral blood flow remains virtually constant as long as mean systemic blood pressure remains between 50 and 150 mm Hg. Below or above these levels, flow becomes pressure dependent. Cerebral autoregulation is compromised in disease states, by anesthetics, and following head trauma ( Strunin, 1975 ). Because the normal blood pressure in infants and children is lower than that of adults, the blood pressure limits established for adults may not be appropriate for pediatric patients. In general, lower blood pressures are needed to achieve a dry operative field in pediatric patients and the cerebral perfusion pressure–cerebral blood flow relation is likely shifted to the left in infants and young children ( Rogers et al., 1980 ). Chronic hypertension blunts the cerebral autoregulatory response to reduced perfusion pressure ( Strandgaard, 1976 ) (see Chapter 18 , Neurosurgery).

Anesthetics might afford some protection against cerebral metabolic deficits during severe hypotension ( Newberg et al., 1983 ; Newberg and Michenfelder, 1983 ). With reduction in cerebral blood flow, widening of the arteriovenous oxygen difference occurs, resulting in no change in cerebral oxygen consumption. In normotensive individuals, cerebral blood flow changes linearly with changes in PaCO2. At PaCO2 between 20 and 70 mm Hg, a 1 mm Hg change in PaCO2 produces a corresponding 2.6% change in cerebral blood flow. With progressive hypotension, this relationship becomes progressively flatter, so that at a mean arterial blood pressure below 50 mm Hg, cerebral blood flow does not change in response to variations in PaCO2 ( Harper and Glass, 1965 ). However, findings in humans indicate that cerebral vascular response to CO2 continues at moderate levels of induced hypotension ( Salem et al., 1970 ; Eckenhoff et al., 1963a ). Therefore, unless indicated, hypocapnia should be avoided during deliberate hypotension ( Levin et al., 1980 ).

Sodium nitroprusside can abolish cerebral autoregulation and increase cerebral blood flow in both awake and anesthetized patients ( Ivankovich et al., 1976 ). Increases in intracranial pressure are mostly seen during the early stages of sodium nitroprusside infusion and may vary widely from patient to patient ( Stullken and Sokoll, 1975 ; Turner et al., 1977 ; Cottrell et al., 1978a ). Hypocapnia tends to attenuate sodium nitroprusside–induced increases in intracranial pressure. Similar increases in intracranial pressure have been reported during nitroglycerin-induced hypotension ( Rogers et al., 1979 ). In contrast, trimethaphan does not usually result in increased intracranial pressure except when intracranial compression is severe.

Spinal Cord Blood Flow

With the popularity of deliberate hypotension for the operative correction of scoliosis, there has been concern that hypotension may decrease spinal cord blood flow and predispose to spinal cord injury, particularly during instrumentation ( Grundy et al., 1981 ; Jacobs et al., 1982 ). The spinal cord exhibits well-functioning autoregulation between mean pressures of 50 and 150 mm Hg. Autoregulation allows normal spinal cord metabolism in the face of decreased arterial blood pressure. However, hypotension and direct pressure on the spinal cord can combine to impair spinal cord function. In anesthetized cats, neither spinal cord compression nor reduction of aortic pressure (by clamping the thoracic aorta) altered spinal cord function when applied separately ( Brodkey et al., 1972 ). When both were applied simulta neously, spinal cord function, as evidenced by changes in somatosensory evoked potentials, was reversibly altered. Hypotension aggravated the effects of compression, whereas intentional hypertension reversed these effects.

Changes in somatosensory evoked potentials at normotension have also been observed in patients undergoing spinal distraction but have resolved with slight increases of blood pressure. Ponte (1974)reported that in two patients in whom paraplegia developed after hypercorrection of scoliosis, partial recovery of neurologic function occurred when blood was infused to correct hypovolemic hypotension. These reports did not document demonstrable deleterious effects of well-conducted normovolemic hypotension on spinal cord function during scoliosis correction. However, they emphasize that (1) hemorrhagic hypotension could result in a severe reduction in spinal cord blood flow and alteration of spinal cord function; (2) spinal distraction (even without hypotension) may result in altered spinal cord function; (3) changes in somatosensory evoked potentials noted during hypotension may return to normal after increases in blood pressure; and (4) monitoring spinal cord function is essential whenever the spinal cord is potentially at risk for injury or interruption of its blood supply ( Grundy et al., 1981 ). Details for monitoring spinal cord function are addressed in Chapter 18 , Anesthesia for Neurosurgery, and Chapter 21 , Anesthesia for Orthopedic Surgery.

Renal Circulation

Studies in humans show that trimethaphan- and sodium nitroprusside–induced hypotension have similar effects on the kidney (Behnia et al., 1978, 1982 [13] [12]). The findings suggest that renal medullary tissue oxygenation, an index of tissue viability, remains adequate despite a significant reduction in creatinine clearance during the hypotensive period. Renal blood flow is normally autoregulated over a range of changes in mean arterial pressure (80 to 180 mm Hg). This capability is attenuated during anesthesia ( Strunin, 1975 ). Below 60 mm Hg, renal blood flow may decrease to the point where urine flow ceases. Provided that induced hypotension does not decrease the renal blood flow below the critical value for the kidney, it is unlikely that renal damage ensues. Because glomerular filtration rate also is not autoregulated during anesthesia, monitoring the urine output may be useful, especially during prolonged hypotension. Other important factors affecting renal blood flow and urine flow during anesthesia include sympathetic stimulation, exo genous or endogenous catecholamines, renin-angiotensin system, antidiuretic hormone levels, and hypercapnia.

Hepatic and Other Regional Blood Flows

A 40% decrease in arterial pressure by sodium nitroprusside results in a decrease in portal pressure (44%) and portal blood flow (25%) and in an increase in hepatic arterial blood flow (13%) ( Gelman and Ernst, 1978 ). Sodium nitroprusside decreases portal sinusoidal resistance, does not interfere with the ability of the liver to increase hepatic arterial blood flow (in conditions of insufficient portal circulation), and does not lead to hepatic hypoxia.

Dry Operative Field and Deliberate Hypotension

Controversy surrounds the relative importance of arterial pressure and cardiac output (or blood flow at the operative site) in producing a dry operative field. Some authors maintain that a reduction in cardiac output is essential to reduce bleeding and that even when blood pressure is low, bleeding is not necessarily reduced unless there is a concomitant fall in cardiac output ( Enderby, 1985b ). Salem and associates (1978) observed that the onset of a dry operative field during induction of hypotension in the supine position in children was not accomplished unless cardiac output was reduced by 35%. Knight and associates (1980) found a positive correlation between blood loss and left ventricular stroke work index during hypotensive anesthesia for surgical correction of scoliosis. Other investigators reported that blood pressure was the important factor determining blood loss. Amaranath and associates (1975) found correlations between systolic blood pressure and blood loss. Sivarajan and associates (1980)concluded that operative blood loss during induced hypotension is determined by mean arterial pressure, not cardiac output. Their study, however, was conducted on patients undergoing mandibular osteotomies placed in a head-up tilt. This may have resulted in pooling of blood in the dependent areas so that blood flow at the operative site was decreased.

A relatively dry operative field and improved operative conditions are not automatically achieved at a predetermined hypotensive level. The skillful anesthesiologist should ascertain that hypotensive anesthesia has achieved its objective and that improved operative conditions have, in fact, ensued. A lower level of blood pressure or other hemodynamic adjustments (e.g., control of heart rate, positioning) may be needed to improve the operative field.

In orthopedic procedures, where bleeding is mostly of venous origin, blood loss is less with nitroglycerin than with sodium nitroprusside at comparable levels of hypotension ( Fahmy, 1978 ). Lower venous pressure associated with nitroglycerin may be partly responsible for the decreased blood loss. These findings have not yet been confirmed in children ( Yaster et al., 1986 ). The requirement of a relatively bloodless field in children depends on decreased cardiac output, decreased blood flow at the operative site, or both. In contrast, when hypotension is used to facilitate surgery on large vessels, the reduction in vessel tension and not necessarily the decrease in blood flow is required. A hypotensive technique that does not decrease cardiac output may be preferred in these situations.

Both sodium nitroprusside and nitroglycerin infusions lead to prolonged bleeding times in a dose-dependent manner. When the dose of sodium nitroprusside exceeds 3 mcg/kg per minute, a dose-related inhibition of platelet aggregation occurs ( Hines and Barash, 1989 ). The prolonged bleeding time with nitroglycerin seems to result from vasodilation and increased venous capacitance rather than inhibition of platelet aggregation ( Lichtenthal et al., 1985 ). In contrast, trimethaphan has no effect on platelet function ( Hines, 1990 ).


Onset and Degree of Hypotension

Hypotension should be induced slowly over 10 to 15 minutes. Time is needed for the cerebral, coronary, and renal vasculature to dilate in the face of decreased pressure so as to maintain adequate perfusion. If the blood pressure decreases too rapidly, a sharp decrease in mixed venous oxygen saturation (S   O2) and content (C   O2) may occur, reflecting inadequate tissue oxygenation (Salem, 1979) ( Fig. 12-9 ). Cardiac arrest and other complications during the induction of hypotension have been related to a rapid decrease in pressure.


FIGURE 12-9  Effect of speed of onset of hypotension on central venous oxygen content. Top, Changes in arterial and venous oxygen content when maximum hypotension was achieved in 5 minutes in six patients. The venous oxygen content fell within the critical range, indicating a state of circulatory inadequacy. Bottom, In six other patients, hypotension was induced slowly (15 minutes). No remarkable increase in arteriovenous oxygen content difference occurred, and venous oxygen content was above the critical range.  (Modified from Salem MR: Deliberate hypotension is a safe and accepted anesthetic technique. In Eckenhoff JE, editor: Controversy in anesthesiology. Philadelphia, 1975, WB Saunders, p 95.)


Because of individual variations, blood pressure should not be decreased to a “predetermined” level. The desired level of hypotension depends on the age, condition, position of the patient, and on the surgical requirement. In young children, a systolic pressure of 55 or 60 mm Hg in the supine position may be necessary to achieve a relatively bloodless field. The anesthesiologist should look for warning signs, including an excessively dry operative field and dark venous blood, which are indicators to increase the blood pressure ( Salem et al., 1971 ). P   O2 (or central venous or jugular venous PO2) below 30 mm Hg indicates tissue hypoxia, and the blood pressure should be increased. Unusually high P   O2 may be an early sign of cyanide toxicity (if sodium nitroprusside is used). If the blood pressure drifts too low, attempts should be made to raise it by decreasing the degree of head-up tilt, slowing the infusion of the hypotensive drug, lightening the level of anesthesia, and speeding up the intravenous fluids. Vasopressors are best avoided unless uncontrollable hypotension occurs.

Maintenance of Near-Normal PaCO2 and Acid-Base Balance

Unless hypocapnia is required to reduce the intracranial pressure, a near-normal PaCO2 should be maintained. Hypocapnia decreases cardiac output; decreases coronary, cerebral, and spinal cord blood flows; may alter drug action (by altering blood pH); decreases both ionized calcium and serum potassium concentrations; causes leftward shift of the oxyhemoglobin dissociation curve; may increase the oxygen consumption; and may inhibit hypoxic pulmonary vasoconstriction (HPV). If adequate oxygenation and a near-normal PaCO2 are maintained, metabolic acidosis is not a feature of well-managed hypotensive technique.

The redistribution of pulmonary blood during induced hypotension may lead to alteration in alveolar ventilation and perfusion ratios. Eckenhoff and associates (1963b) demonstrated that the alveolar dead space may increase to as much as 80% of the tidal volume in the hypotensive adult patient in the head-up position and with an increased airway pressure. Further studies showed that the increase in alveolar dead space is less than previously thought ( Askrog et al., 1964 ). Data from infants and children suggest that the alveolar dead space does not increase during controlled hypotension even with the head-up tilt ( Salem et al., 1974 ) ( Fig. 12-10 ). Others found that sodium nitroprusside–induced deliberate hypotension caused no change in pulmonary dead space in adequately hydrated patients who were operated on in the prone position.


FIGURE 12-10  Arterial–to–end-tidal PCO2 difference during deliberate hypotension in adults (from Askrog et al., 1964 ) and in children.  (Reprinted with permission from the International Anesthesia Research Society from Salem MR, Wong AY, Bennett EJ: Deliberate hypotension in infants and children. Anesth Analg 53:975, 1974, with permission).





An increased difference between alveolar-to-arterial oxygen tensions (A - aDO2) is observed during induced hypotension ( Stone et al., 1976 ; Casthely et al., 1982 ). This may be explained by (1) increased    s/   t and (2) decreased cardiac output.

Increased    s/   t

Changes in functional residual capacity and closing volume during anesthesia and surgery contribute to airway closure, trapping of gas distal to the closure, and alveolar collapse. This local alveolar hypoxia is normally offset (to a degree) by reflex HPV that directs blood from hypoxic areas of the lung to adequately ventilated alveoli. Blunting or inhibition of this reflex may occur with pulmonary hypertension, inhalation anesthetics, and vasodilators ( Benumof and Wahrenbrock, 1975 ). Although the inhibition of HPV occurs with all vasodilators, it is greater with sodium nitroprusside than with nitroglycerin. Inhibition of HPV is associated with a decrease in pulmonary vascular resistance and pulmonary artery pressure and results in increased    s/   t

Decreased Cardiac Output

Decreased cardiac output is accompanied by increased extraction of oxygen by the tissues resulting in decreased P   O2 and C   O2 ( Kelman et al., 1967 ; Philbin et al., 1970 ). Any portion of blood with decreased C   O2 that passes through hypoventilated or non ventilated areas (low    a/   p) contributes to a greater decrease in PaO2. A decrease in cardiac output during deliberate hypotension can result in a decrease in PaO2, significant only in the presence of regional atelectasis ( Cheney and Colley, 1980 ).

A high FIO2 (>0.9) during deliberate hypotension tends to compensate for the venous admixture contributed by ventilation–perfusion imbalance and is recommended. High FIO2 is also desirable in children because of their increased oxygen demands. Robinson (1967) showed that the lactate-to-pyruvate ratio does not increase during profound hypotension when the PaO2 is kept above 300 mm Hg. Jugular bulb oxygen tension rises significantly when FIO2 is altered from 0.4 to 1.0 during hypotensive anesthesia, although the oxygen delivery would have increased only slightly with an FIO2 of 1.0 ( Salem et al., 1970 ). These findings stress the importance of using high FIO2, monitoring oxygenation, and avoiding profound decreases in cardiac output.


In addition to routine monitoring (electrocardiography, pulse oximetry, capnography), accurate recording of the blood pressure is essential during induced hypotension. Although various methods have been used, direct measurement via arterial cannulation allows continuous blood pressure measurements and arterial blood sampling. A variety of automated noninvasive devices may be used as a backup system for direct arterial pressure measurement. Central venous access permits measurement of central venous blood gases, a fairly accurate estimation of mixed venous blood gases.

Cyanide Toxicity

Sodium nitroprusside has the molecular formula Na2 [Fe(CN)5NO]· 2H2 O. Although cyanide (C N-) released from the sodium nitroprusside molecule is transformed mostly into relatively nontoxic products, C N- can be toxic and sufficient amounts can cause death ( Tinker and Michenfelder, 1976 ; Michenfelder, 1977 ; Verner, 1985 ) ( Fig. 12-11 ). There are four pathways for the disposal of this free C N-.


FIGURE 12-11  Metabolism of sodium nitroprusside.  (Modified from Tinker JH, Michenfelder JD: Sodium nitroprusside: Pharmacology, toxicology and therapeutics. Anesthesiology 45:340, 1976, with permission.)




Binding to Cytochrome Oxidase

Cyanide binds to mitochondrial cytochrome oxidase, inhibiting oxidative phosphorylation. The subsequent anaerobic metabolism leads to acidosis. This binding to cytochrome oxidase is reversible in vitro, but reversibility probably does not contribute to short-term survival in humans.

Conversion to Thiocyanate

Transformation of C N- into thiocyanate is the major metabolic pathway for C N- in humans. It occurs in the liver and kidney, is catalyzed by the enzyme rhodanese, and requires B12a as a cofactor. Added thiosulfate speeds the reaction, which is slowly reversible.

Conversion of B12a to Cyanocobalamin

In the presence of adequate hydroxocobalamin, C N- becomes cyanocobalamin. This is probably not an important pathway in normal humans, but hydroxocobalamin has been suggested for prophylaxis.

Conversion to Cyanmethemoglobin

One of every five C N- ions is normally converted to cyanmethemoglobin. When exposed to light, sodium nitroprusside is rapidly converted to unstable ionic aquapentoferrocyanate, a substance that readily releases free C N-. Sodium nitroprusside solutions should be protected from photolytic decomposition during infusion ( Verner, 1985 ). Although it has been recom mended that sodium nitroprusside solution be discarded 4 hours after reconstitution, evidence substantiates its safety for 24 hours if properly protected from light ( Ikeda et al., 1987 ).

The mechanism for CN- intoxication with sodium nitroprusside overdose is interference with aerobic metabolism, which is the major pathway of high-energy phosphate production ( Tinker and Michenfelder, 1976 ; Michenfelder and Tinker, 1977 ). Free CN- inhibits the reoxidation of reduced cytochrome oxidase by oxygen, and, by crossing cell and mitochondrial membranes, CN- rapidly inhibits the electron transport system. The consequences of CN- intoxication are decreased oxygen utilization, decreased CO2 production via inhibition of the Krebs' cycle, and increased production of anaerobic metabolites. Metabolic acidosis and deterioration in the central nervous and cardiovascular systems ensue. The combination of tissue hypoxia with normal or elevated PvvO2 levels is the hallmark of cytotoxic hypoxia produced by CN- ( Michenfelder, 1977 ).

Detection of Cyanide Toxicity

In the awake patient, the manifestations of cyanide toxicity are fatigue, nausea, vomiting, headache, anorexia, tremors, angina-like syndrome, disorientation, psychotic behavior, muscle spasms, rigidity, convulsions, and cardiovascular collapse. Prolonged intravenous sodium nitroprusside infusion may result in hypothy roidism due to the antithyroid action of thiocyanate ( Palmer and Lasseter, 1975) . The clinical manifestations of C N- intoxication during anesthesia are shown in Box 12-6 ( Jack, 1974 ; Merrifield and Blundell, 1974 ; Davies et al., 1975 ).

BOX 12-6 

Clinical Manifestations of Cyanide Toxicity

Metabolic acidemia

Progressive hypotension with narrow pulse pressure

Refractory hypotension unresponsive to fluids and vasopressors but responsive to thiosulfate

Cardiovascular collapse

Bright venous blood

Increased S   O2 and P   O2

In children and adolescents three abnormal responses have been recognized that suggest impending CN- intoxication: (1) a requirement for high doses of sodium nitroprusside (>10 mcg/kg per minute); (2) tachyphylaxis, which is apparent in 30 to 60 minutes after the start of the infusion; and (3) definite resistance becoming apparent within 5 to 10 minutes after the start of the infusion. The incidence of these abnormal responses may be as high as 30% ( Greiss et al., 1976 ). Tachyphylaxis may or may not be associated with concurrent metabolic acidosis ( Cottrell et al., 1978c ). The severity of acidosis is usually proportional to the CN- level. The progressive hypotension may be responsive to discontinuing the sodium nitroprusside infusion, administering fluid and blood products, or infusing vasopressors. Cardiovascular collapse may ensue and may not respond to cardiopulmonary resuscitation, but a dramatic response may be observed after the administration of sodium thiosulfate.

Early laboratory recognition of CN- intoxication poses difficulties because of the absence of specific tests. The lethal blood CN- level in humans is approximately 500 mcg/dL, whereas lethal blood thiocyanate levels have been reported to be as low as 340 mcg/dL, but this varies with the rate of CN- release as well as the total dose. Measurement of blood levels of CN- or thiocyanate does not reflect the magnitude of CN- released. Consequently, nonspecific tests are relied upon as indicators of CN- toxicity. The most sensitive metabolic indicators of CN- toxicity are blood pH, blood lactate (or lactate/pyruvate), P   O2 or SvO2, sagittal sinus PO2 (reflecting cerebral tissue oxygen tension), cerebral oxygen consumption, and brain lactate (or lactate/pyruvate) ( Michenfelder, 1977 ). Of these, arterial pH and gas tensions, as well as P   O2 or jugular venous PO2, are easily obtained and should be measured when CN- intoxication is suspected. “Bright” venous blood during sodium nitroprusside infusion should alert the anesthesiologist to the possibility of early CN- intoxication.

Prevention of Cyanide Toxicity

Cyanide intoxication associated with the use of sodium nitroprusside is a preventable complication. The total projected dose should not exceed 1.5 mg/kg for short exposures or 0.5 mg/kg per hour for prolonged exposures. Infusion rates exceeding 10 mcg/kg per minute should not be allowed. The rate of sodium nitroprusside infusion should be adjusted to 0.5 to 1 mcg/kg per minute via use of a microdrop or infusion pump, and the infusion rate may be gradually increased as needed. A satisfactory response can be obtained well below the recommended maximum of 10 mcg/kg per minute. The patient's response to the infusion should be ascertained constantly, especially in the first 30 minutes. Frequent (half-hourly) arterial acid-base determinations are recommended during sodium nitroprusside infusion. In addition, C N- antidote therapy should be available. If either a constant response to high doses of sodium nitroprusside (> 10 mcg/kg per minute) or a tachyphylactic response is noted, a β-adrenergic–blocking drug should be administered and the inhaled anesthetic concentration increased. A quick response is usually noted after these measures are instituted, and a rapid decline in the dose requirement usually follows. If resistance is detected (within 5 to 10 minutes), the infusion should be abandoned and a different hypotensive drug given.

Treatment of Cyanide Toxicity

The rational approach in the treatment of C N- toxicity is to prevent the C N- from binding to cytochrome oxidase. Sodium thiosulfate can afford complete protection against C N- and complete detoxification if three times more thiosulfate than C N- is present. In experimental animals, thiosulfate resulted in no noticeable adverse hemodynamic or respiratory effects. Thiosulfate ensures the plentiful supply of sulfhydryl radicals needed to form thiocyanate from C N-. Because thiosulfate is rapidly eliminated by the kidneys, a high level of plasma thiosulfate is best maintained by constant infusion. A bolus injection of 30 mg/kg followed by a continuous infusion of 60 mg/kg per hour appears to be the most effective and safest prophylactic antidote against C N- toxicity and also prevents C N--induced circulatory failure ( Ivankovich et al., 1980 ). Being an osmotic diuretic, thiosulfate can ultimately decrease the plasma volume.

Hydroxocobalamin (vitamin B12a) has been advocated as a treatment for cyanide intoxication ( Posner et al., 1976 ; Cottrell et al., 1978b ). It prevents the increase in the CN- concentration in erythrocytes when given prophylactically with large doses of sodium nitroprusside. Problems associated with the use of vitamin B12a include requirement of a large dose, lack of cardiovascular stability, scarcity, expense, solubility, and proper storage of the powder. The recommended dose is a 50-mg/kg bolus, plus 100 mg/kg per hour. In addition to these specific antidotes, correction of acidosis and fluid replacement are important in the management of CN- intoxication. If the patient is bleeding, blood transfusion may help to “exchange the blood volume” and thus eliminate CN- ( Vesey et al., 1975 ).


At one time, almost all systemic diseases were considered absolute contraindications to induced hypotension ( Box 12-7 ). With this stringent rule, many patients were denied the benefits of the technique. Many heretofore absolute contraindications are now regarded as relative. Because most of the complications associated with deliberate hypotension are related to inexperience and unfamiliarity with the technique, the technique should not be attempted by inexperienced practitioners. Anesthesiologists and surgeons should be familiar with the pharmacology of hypotensive drugs and the physiologic effects of hypotension. Teamwork and cooperation are of great importance in the care of patients who undergo hypotension.

BOX 12-7 

Contraindications to Deliberate Hypotension


Infants (except when there is a definite indication)

Significant reduction in oxygen delivery

Systemic diseases compromising major organ function

Renal, cerebral, or coronary artery stenosis

Children with cardiac shunts

Patients with sickle cell disease

Uncorrected polycythemia

Ganglionic blockers in patients with narrow-angle glaucoma

Age is not a contraindication to induced hypotension. Except in infants, in whom blood pressure may be difficult to measure accurately, the technique need not be withheld ( Salem, 1978 ). Significant reduction in oxygen delivery to the tissues, as in anemia, low fixed cardiac output, severe lung disease, and the presence of severe acute cardiac, cerebral, or renal disease or any combination of these factors, may contraindicate the use of induced hypotension. In the presence of renal artery stenosis, hypotension can cause further decrease in perfusion pressure to the kidney. In children with cardiac shunts, reduction of systemic vascular resistance may increase right-to-left shunting and cause hypoxemia. In patients with sickle cell disease, a decrease in P   O2 (<30 mm Hg) due to decreased cardiac output can trigger a crisis. Uncorrected polycythemia is an additional contraindication, because it may increase sludging and thrombosis.

Diabetes is not a contraindication to induced hypotension if blood sugar levels are controlled perioperatively. β-Adrenergic blockade, through its actions on carbohydrate metabolism, may lower the blood sugar level although not seriously with the small doses given. The pupillary dilation caused by trimethaphan is short lived but may be misinterpreted as cerebral ischemia in the immediate postoperative period.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



Although preoperative autologous blood donation (PABD) was practiced in the 1960s, it did not gain wide acceptance until the 1980s. By 1990, as much as 5% of all blood collected for transfusion was intended for use by the patient-donor from whom it was collected ( Surgenor et al., 1990 ). In some centers, up to 80% of the blood needed for certain surgical procedures has been met by PABD. Contributing to growth in PABD programs were (1) concerns about litigation for blood borne infection; (2) patient fears and concerns about human immunodeficiency virus (HIV) infection; and (3), in some states, hastily conceived legislation. The California Statute CB-37, the Paul Gann Blood Safety Act (1990), requires that in addition to the risks and benefits of blood transfusion, alternatives to allogenic blood must be presented to patients.

Preoperative autologous blood may be donated by patients who are likely to require transfusion during or after surgery. The ideal patient for PABD is one who (1) is healthy enough to undergo elective surgery; (2) is likely to need a transfusion during or after surgery; (3) has 2 or more weeks before surgery; and (4) has a hemoglobin level above 11 g/dL (hematocrit >33%). Some centers accept patients with slightly lower hemoglobin values, especially when there is a strong need for PABD.

Autologous transfusion is ideally suited for children and adolescents because isoimmunization during youth can complicate future transfusion needs. Children weighing less than 50 kg can safely donate blood, although the volume drawn at each donation is reduced in proportion to body weight. Withdrawal of amounts equal to 10% of their estimated blood volume is usually well tolerated. Although PABD has been extended to children as young as 4 years, technical problems and lack of cooperation make young children unlikely candidates for autologous blood donation.

The American Association of Blood Banks (AABB) standards recommend a minimum 4-day interval between phlebotomies and at least 3 days between the last phlebotomy and surgery ( Widman, 1991 ). These minimum intervals are required to allow for synthesis and mobilization of proteins and the return of plasma volume to normal. The commonly used schedule, however, is one donation per week. This permits preoperative acquisition of up to four donations with conventional storage techniques (blood stored as liquid at 1° to 6°C). The shelf-life of red blood cells stored in the liquid state can be prolonged to 42 days when additive solutions such as Adsol (Fenwal Laboratoris, Roundlake, IL) are used. With this longer shelf-life, a weekly donation results in the collection of sufficient amounts of blood before surgery for most patients.

The total volume of blood that the patient can donate preoperatively may be increased by (1) iron therapy and (2) recombinant human erythropoietin therapy to increase the rate of endogenous red blood cell production ( Goodnough et al., 1989 ). Although weekly donations stimulate erythropoiesis, with iron therapy the marrow can double or triple production, and this response can be maintained over several weeks, even with repeated phlebotomies. In patients undergoing weekly phlebotomy without added iron, the hematocrit decreases from 44% to 33% over an 8-week period. In patients who receive iron, it decreases from 44% to 38% during the same period. Because of the side effects of parenteral iron administration, the oral route is preferred. The recommended dose is 5 mg/kg two to three times daily with meals. It is preferable to begin iron therapy 1 month before the first donation and continue therapy for several months after the last donation ( Widman, 1991 ).

The ability of recombinant human erythropoietin therapy to enhance the procurement of autologous blood is now unquestioned. Goodnough and associates (1989) demonstrated that the red blood cell volume donated by the patients who received erythropoietin was 41% greater than the volume donated by the patients who received placebo (961 versus 683 mL). Side effects (hypertension, thrombotic events) observed for long-term treatment in patients with end-stage renal disease are uncommon in surgical patients ( Goodnough et al., 1997 ).

Various regimens for the administration of recombinant erythropoietin have been suggested. Initially, 600 U/kg IV was recommended at each visit for blood collection (twice weekly). This required repeated visits to the blood bank. Kulier and associates (1993) advocated a single weekly dose of erythropoietin (400 U/kg) given subcutaneously once a week starting 4 weeks before surgery. They found that this simple protocol provided a constant and efficient stimulus for erythropoiesis and adequately compensated for the hemoglobin decrease after the weekly donation. To maximize the response, oral iron supplementation is also implemented ( Goodnough et al., 1997 ).

The efficacy of PABD in reducing allogenic blood exposure is well established ( Thomas et al., 1996 ). Nevertheless, some drawbacks exist, including patient inconvenience, delay of the operative procedure, cost efficiency, and the occurrence of adverse reactions. Transient vasovagal reactions vary between 1.5% and 5.5% but could be higher among first-time donors (13%) and females ( Stehling, 1990 ). Donors younger than 17 years, those who weigh less than 45 kg, and those who have experienced reactions during earlier donations are most likely to have a reaction. These reactions consist of light-headedness as a result of transient hypotension and bradycardia; 10% of these patients lose consciousness. Intravenous replacement with crystalloids and monitoring cardiovascular function during blood donation may decrease the incidence of these reactions.

Contraindications to PABD are bacteremia, decrease in oxygen delivery (low cardiac output, severe anemia, and hypoxemia), and very young age of the patient. Clinical judgment is essential when there are relative contraindications or concerns that patients may not tolerate phlebotomies. Establishing an intravenous catheter for infusion of isotonic saline or lactated Ringer's solution (one to two times the volume of blood removed) and close monitoring may be advisable in these patients. Despite the wide acceptance of PABD, controversial issues regarding the extent of testing, disposition of infectious blood, and crossover into the allogenic blood supply remain unresolved ( Silvergleid, 1991 ).

Extent of Testing

Arguments in favor of minimal or no testing of blood intended for autologous blood use include cost savings and simplified bookkeeping. Complete testing, however, makes crossover into the allogenic blood supply possible. Unfortunately, a positive infectious disease test may create a conflict between the patient's right to privacy and the hospital team's right to take additional precautions ( Silvergleid, 1991 ).

Release of Potentially Infectious Units

The rationale for not releasing a potentially infectious unit is to protect health care workers and to prevent the transmission of disease should the blood be unintentionally transfused to the wrong patient.

Crossover Into the Allogenic Blood Supply

Because of logistic problems and concerns about safety, only 15% of centers currently transfer unused autologous blood to the allogenic blood supply. The high cost and the improved safety of allogenic blood account for the low cost-effectiveness of PABD reported recently ( Tretiak et al., 1996 ; Etchason et al., 1995 ).


The salvage and reinfusion of blood during the perioperative period represents the most common form of autologous blood transfusion. Used in conjunction with other techniques in a comprehensive blood conservation program, blood salvage is frequently used during pediatric surgical procedures where moderate blood loss is foreseen. Salvaged blood is a source of immediately available, type-specific, compatible blood without the risks of disease transmission or isoimmunization. The use of salvaged blood greatly reduces or abolishes the incidences of febrile, allergic, graft-versus-host, and hemolytic transfusion reactions. Because salvaged blood usually does not require handling, transport, typing, compatibility, or disease testing, technical errors are virtually eliminated. Salvaged blood is normothermic, and thus hypothermia and the necessity of a blood warmer are avoided. Other advantages may also include psychologic benefit to recipient and parents and avoidance of modification of any immunosuppressive effects of transfusion.

Indications for Blood Salvage

Blood salvage procedures are currently used for some surgical procedures in children, such as orthopedic correction of spinal deformities, correction of congenital cardiac defects, and orthotopic liver transplantation. An indication for the use of blood salvage in children (> 10 kg) includes an anticipated blood loss of 20% or more of their estimated blood volume or a procedure in which greater than 10% of patients are transfused with more than one unit ( Williamson and Taswell, 1991 ; Zauder, 1991 ; Simpson et al., 1993 ).

Blood salvage in infants and small children (<10 kg) is rarely feasible or indicated, because current pediatric blood salvage technology would require a minimum anticipated blood loss equal to 1 to 1.5 times their estimated blood volume. With allogenic blood transfusion unavoidable in these patients, many of the advantages of autologous transfusion are negated. These patients are probably best served through the use of minimal (single) blood donor exposure'that is, the use of mini (50 or 100 mL) blood packs derived from a single donor ( Salem and Podraza, 1992 ).

Blood Salvage Techniques

Clinicians have often raised concerns about possible physical and biochemical alterations of blood following extravasation, collection and processing, and reinfusion that might render salvaged blood inferior to banked blood. That salvaged blood differs significantly from allogenic banked blood or from circulating blood is undisputed ( Table 12-4 ). The diverse circumstances under which shed blood is salvaged and the various techniques that are used result in salvaged blood products with differing characteristics. Blood salvage in the pediatric patient can be accomplished by three different techniques: (1) filtration by a canister device, (2) cell processing by centrifugation and cell washing, and (3) hemoconcentration via hemofiltration.

TABLE 12-4   -- Hematologic comparison of bank and salvaged blood

Index Factor

Processed Salvaged Blood*

Unprocessed Salvaged Blood*

Bank Blood

Normal Blood

Hemoglobin (g/dL)


6.3 ± 1.1

16.0 ± 3.5

12 to 17[†]

Hematocrit (%)


17.0 ± 3.3

46.5 ± 10.5

37 to 50[†]

Red blood cells (/mm3 × 106)


2.0 ± 0.4

5.0 ± 1.1

4.8 to 5.8[†]

Leukocytes (/mm3 × 103)


3.4 ± 1.3

5.3 ± 2.3


Platelets (/mm3 × 103)


384 ± 141


250 to 350

Free hemoglobin (mg/dL)


1000 ± 625

13.3 ± 10.8


Screen filtration pressure (mm Hg)






7.72 ± 0.18


6.63 ± 0.06

7.33 to 7.43

PCO2 (mm Hg)


8.5 ± 4.0

158 ± 68

38 to 50

PO2 (mm Hg)


160 ± 17

50 ± 25

60 to 108

Sodium (mmol/L)

147 ± 2.4

136 ± 5

162 ± 6

135 to 148

Potassium (mmol/L)

1.6 ± 1.5

6.8 ± 1.9

12.6 ± 4.6

3.5 to 5.3

Chloride (mmol/L)

134 ± 3.3

123 ± 5

76 ± 9

98 to 106

CO2 (mmol/L)

5.4 ± 1.6

6 ± 1

11 ± 2

19 to 26

Calcium (ionized) (mmol/L)

2.3 ± 0.5

1.25 ± 0.2

2.25 ± 0.1

2.1 to 2.7

SGOT (units)

9 ±5

128 ± 74

14 ± 4

6 to 21

SGPT (units)

15 ± 9

14 ± 6

11 ± 5

7 to 14

Lactate dehydrogenase (units)

214 ± 103

1605 ± 718

299 ± 146

85 to 190

Creatine phosphokinase (units)

1073 ± 983

96 ± 27

10 to 65


Direct bilirubin (mg/dL)


0.4 ± 0.1

0.06 ± 0.02

0 to 0.2

Total bilirubin (mg/dL)

0.11 ± 0.07

0.7 ± 0.2

0.41 ± 0.13

0.2 to 1.0

Processed salvage data modified from Yawn DH: Properties of salvaged blood. In Taswell HF, Pineda AA, editors: Autologous transfusion and hemotherapy. Boston, 1991 , Blackwell Scientific Publishers, with permission. Filtration data modified from Aaron RK, Beazley RM, Riggle GC: Hematologic integrity after intraoperative allotransfusion. Arch Surg 108:831, 1974. Normal data from Fiereck EA: Appendix. In Tietz NW, editor: Fundamentals of clinical chemistry, 2nd ed. Philadelphia, 1976, WB Saunders, pp 1177–1227, with permission.

Modified Bentley ATS100 used for blood salvage.


Range includes both male and female adults.



Blood salvage by filtration is based on the tenet that the mere extravasation of blood does not render that blood unfit for rein fusion, provided that coagulation, contamination, and significant hemolysis are prevented. Shed blood is altered only by filtration and the addition of anticoagulant, if required. Various filtration systems that are currently available differ slightly depending on the route of administration of anticoagulant, intraoperative or postoperative use, portability, and direct-versus-indirect reinfusion.

Blood is collected in a reservoir that is either divided into two chambers separated by a coarse mesh filter (170 μm) or a single chamber with the filter incorporated into the inlet ( Fig. 12-12 ). A regulated vacuum source is connected to the reservoir, and negative pressure should not exceed -150 mm Hg (and usually kept between -30 and -50 mm Hg) to prevent damage to formed blood elements during aspiration. Defoaming material inside the reservoir chamber or present as a coating on the coarse filter prevents foaming. In the nonheparinized patient, anticoagulant must either be added to the reservoir or mixed with shed blood near the tip of a specially designed dual-channel suction tubing/wand assembly. Heparinized saline, citrate-phosphate-dextrose, and acid-citrate-dextrose formulas have all been used. Citrate anticoagulants are particularly advantageous in filtration systems because of their negligible systemic effects (provided citrated blood is not reinfused too rapidly or in excessive volumes). Anticoagulant requirements may vary depending on the location and extent of bleeding, with citrated solutions used in a ratio of 1 part anticoagulant to 7 parts blood. In the heparinized patient, heparin in a concentration of 20,000 to 40,000 U/L is usually effective ( Zauder, 1991 ). After collection and coarse filtration, salvaged blood can be reinfused directly to the patient. Reinfusion through a microaggregate filter (10 to 50 μm) is recommended by most filtration system manufacturers, although data in favor of the use of these filters are not convincing. If blood cannot be immediately reinfused, it can be stored in the reservoir for up to 6 hours at 20° to 24°C, according to the AABB guidelines (1993) . Blood salvaged by filtration can also be transported to the blood bank for processing (centrifugation and washing) (Pineda and Valbonesi, 1990 ).


FIGURE 12-12  Schematic diagram of a filtration system.  (Courtesy of Boehringer Laboratories, Inc., Norristown, PA.)


The principal advantages of blood salvage by filtration are the simplicity of filter design, ease of operation, and portability. Setup and operation do not require specialized training, nor is a dedicated operator normally necessary. Other than a vacuum source and an anticoagulant solution, the disposable units are entirely self-contained. The collection reservoir can usually follow the patient from operating room to postanesthesia care or intensive care units to minimize costs and maximize salvage yield ( Blevins et al., 1993 ; Davis et al., 1993 ). Because no blood processing (beyond filtration) is required, shed blood drawn into the collection chamber is immediately available for reinfusion. The cost of blood salvage by filtration is generally lower than that associated with cell processing.

Postoperative anemia is a common occurrence after autotransfusion with filtration devices, particularly when salvaged blood is the sole or primary source of replacement red blood cells. Blood salvaged by filtration typically has a low hemoglobin value, ranging between 6 and 9 g/dL ( Yawn, 1991 ). However, the red blood cells that are salvaged have a relatively normal half-life, higher concentrations of 2,3-DPG and adenosine triphosphate (ATP) than banked blood, and normal P50 and pH. Salvaged red blood cells are more resistant to osmotic lysis, because the most frail red blood cells lyse during collection and filtration ( Ray et al., 1986 ; Williamson and Taswell, 1991 ). Platelets, leukocytes, and coagulation factors are often reduced in salvaged blood. The concentration of these components in unprocessed salvaged blood depends on their concentration in the patient's blood, loss in salvage apparatus, and dilution with anticoagulant, tissue fluids, and irrigation solutions ( Yawn, 1991 ). Those platelets present in salvaged blood are considered dysfunctional, having shown signs of activation ( Wilson et al., 1988 ; Yawn, 1991 ; Kongsgaard et al., 1993 ).

The concentration of coagulation factors and the clotting competency of salvaged blood are specifically determined by the location and circumstances of collection. Blood salvaged from the mediastinum, pleural space, and peritoneal cavity has undergone extensive coagulation and clot lysis and is virtually devoid of fibrinogen. Fibrinogen may be significantly higher in blood salvaged from systemically anticoagulated patients. Fibrin degradation products are markedly increased in filtered blood. Although signs of clinical bleeding should be monitored and the need for fresh frozen plasma and platelet concentrates considered, component replacement is rarely necessary ( Yawn, 1991 ; Kongsgaard et al., 1993 ).

As a result of hemolysis, shed blood also can contain high concentrations of free plasma hemoglobin and cellular debris. Under normal circumstances the reticuloendothelial system can remove free plasma hemoglobin in concentrations up to 120 mg/dL. Levels exceeding 1600 mg/dL have been measured in salvaged blood. The filtration technique is not effective for free plasma hemoglobin or cell debris. Regardless of the toxic level, adequate hydration, avoidance of hypotension and acidosis, and maintenance of urinary output with appropriate use of diuretics or dopamine have been shown to prevent this complication.

Cell Processing

Blood salvaging is most commonly accomplished by cell processing. Shed blood is aspirated from the surgical field, mixed with an anticoagulant, and stored in a cardiotomy reservoir. When a sufficient amount of blood has collected, it is pumped into a spinning (5000 rpm) centrifuge bowl, which separates blood components on the basis of density. As the bowl fills, higher-density red blood cells continually displace other constituents of blood, which spill over into a waste container. When the bowl is filled with red blood cells, a wash cycle rinses away all residual contaminants and anticoagulants, suspending the red blood cells to any desired hematocrit (usually 50% to 70%) in normal saline solution ( Stehling and Zauder, 1991 ) ( Fig. 12-13 ). The trauma imposed by suction, centrifugation, washing, and reinfusion appears to have negligible effects on red blood cell survival ( Ray et al., 1986 ). For optimal results and safety, a trained operator whose only responsibility is to “recycle shed blood” should operate these sophisticated systems ( AABB, 1993 ).


FIGURE 12-13  Schematic diagram of a cell-processing system.



The characteristics of salvaged blood following processing (centrifugation and washing) differ significantly from those of nonprocessed blood (see Table 12-4 ). Centrifugation virtually eliminates all the plasma constituents of blood. Washing significantly reduces any residual contaminants, including anticoagulant, free plasma hemoglobin and potassium, fibrin-degradation products and D-dimer fragments, products of platelet activation and lysis, products of complement activation, cellular and tissue debris, and microaggregates. The resulting product is similar to packed red blood cells but suspended in normal saline solution rather than plasma.

Unlike filtration systems, the hematocrit of the infusate following cell processing is unrelated to that of the shed blood. The hematocrit of processed red blood cells depends on the flow rate of salvaged blood into the centrifuge bowl and the rotational speed of the centrifuge and can be further adjusted by varying the volume of saline solution used for resuspension. If negative pressure in the recovery system is low, hemolysis during harvest and processing is between 2% and 10% of all salvageable red blood cells.

Because platelets and plasma constituents, including coagulation factors and plasma proteins, are totally eliminated during processing of salvaged blood, patients transfused with large volumes of processed red blood cells may show a slight reduction of coagulation factors, dilution of platelets, and prolonged prothrombin, activated partial thromboplastin, thrombin, and bleeding times ( Otteson and Frøysaker, 1982 ). Massive autotransfusions of processed blood (usually several blood volumes) often require fresh frozen plasma administration. Various formulas for the replacement of plasma constituents and platelets have been advanced based on the volume of autotransfusion. However, coagulopathy often cannot be predicted based on such formulas. Because coagulopathy can result from the dilution of coagulation factors, fibrinogen, and platelets or a consumption process (disseminated intravascular coagulation), in vitro hemostatic laboratory testing is important in differentiating dilutional from consumptional coagulopathies.


The hemofilter has been used as an adjunct to the extracorporeal circuit during cardiopulmonary bypass in children to facilitate collection of red blood cells ( Tuman et al., 1988 ; Naik et al., 1991 ). The hemofilter is either a hollow fiber filter or a parallel plate with designated membrane pore size. Hemoconcentration is based on a process that imitates physiologic glomerular filtration by applying a hydrostatic pressure gradient across a porous membrane. This efficient method of removing excess circulating blood volume can be achieved without significant alteration in serum electrolyte concentrations or acid-base status. The technique of hemoconcentration is a convective process with plasma and dissolved solutes filtering at the same rate, limited only by the pore size of the device. Transmembrane pressure (gradient) is the driving force, and is calculated by the following formula:

where TMP is transmembrane pressure (mm Hg), PA is arterial (inlet) pressure (mm Hg), PV is venous (outlet) pressure (mm Hg), and PN is absolute value of the suction applied on the ultrafiltrate outlet (mm Hg).

To avoid red blood cell lysis, the transmembrane pressure should remain below 600 mm Hg. Ultrafiltrate flux is determined by several factors, including properties of the membrane, pump flow rate, transmembrane pressure, hematocrit, and plasma protein concentration. Membrane pore size varies among manufacturers (16,000 to 60,000 Da). Depending on pore size, heparin with a molecular weight of less than 20,000 Da can be removed. Anticoagulation status should be closely monitored during intraoperative use and for postbypass concentration of residual oxygenator blood. The hemoconcentrator can be used in a similar manner as a cell processor in conjunction with cardiopulmonary bypass for cardiac ( Naik et al., 1991 ; Friesen et al., 1993 ) or liver transplantation surgery ( Tuman et al., 1988 ), as well as for nonbypass procedures ( Solem et al., 1987 ).

Under normal circumstances, the management of excess hemodilution during extracorporeal circulation is accomplished by pharmacologically induced diuresis. However, certain situations arise when diuresis is impractical or too inefficient. In such situations, the control of plasma water volume during cardiopulmonary bypass can be achieved through hemofiltration much more effectively than through cell processing. By allowing more precise control of blood volume, the hemoconcentrator enables the use of greater hemodilution during cardiopulmonary bypass ( Naik et al., 1991 ).

The main advantage of hemoconcentration over cell processing by centrifugation and washing appears to be the ability to hemoconcentrate blood to any desired hematocrit with preservation of valuable plasma constituents. Better platelet and fibrinogen preservation might avoid coagulopathy associated with the autotransfusion of large amounts of salvaged blood. Electrolytes, which pass through the hemofilter, remain relatively normal, and plasma proteins, which cannot, are concentrated ( Solem et al., 1987 ). Other advantages of the hemoconcentrator are speed, lower cost, and single-pass hemofiltration at rates of 500 mL/min.

Postoperative Blood Salvage

The volume of blood lost after surgery can be significant, but this blood is salvageable by either filtration or cell-processing apparatus. Situated between a vacuum source and the wound drains, a reservoir with an internal 150- to 170-μ m filter collects sanguineous wound drainage. This blood may be continuously reinfused, allowed to collect for a period of time (not longer than 6 hours) before reinfusion without processing through a fine filter (10 to 50 μ m), or processed and reinfused. Anticoagulation is usually not required because this blood is totally defibrinogenated through extensive contact with wound surfaces ( Yawn, 1991 ). The reinfusion of volumes up to 15% of the estimated blood volume has been shown to be safe ( Kongsgaard et al., 1993 ).

Complications and Contraindications

The potential complications of blood salvaging are mostly related to the infusion of salvaged blood. Salvaged blood may be hemo globinemic, thrombocytopenic, leukopenic, hypofibrinogenemic, and depleted or diluted of coagulation factors and plasma proteins. Furthermore, the state of those red blood cells, platelets, and leukocytes present has also been questioned. Because neither filtration nor cell processing can remove all contaminants, some soluble or insoluble foreign substances may be transfused to the patient. Cell washing has been shown to significantly reduce (by ≥ 80%) bacterial cell counts in enteric-contaminated blood. Finally, the concentrations and metabolisms of perioperatively administered drugs are unpredictable following the reinfusion of shed blood.

Various contraindications to the use of blood salvage have been suggested ( Box 12-8 ). Extravasated blood older than 6 hours or excessively hemolyzed blood should not be reinfused. Although cell processing can eliminate most contaminants, shed blood in contact with bowel contents or malignant cells or that aspirated from an infected wound site should not be reinfused except with life-threatening hemorrhage. The collection of shed blood should be interrupted following the application of thrombin, microfibrillar collagen hemostat (Avitene), methylmethacrylate, and irrigation with wound-sterilizing solutions (e.g., Betadine) or antibiotic solutions not meant for parenteral use ( Stehling and Zauder, 1991 ; Williamson and Taswell, 1991 ; AABB, 1993 ).

BOX 12-8 

Potential Contraindications of Blood Salvage

Extravasated blood over 6 hours old

Suspected or confirmed enteric contamination

Suspected or confirmed malignant cell contamination

Sickle cell anemia patients

Hemolyzed blood

Some Jehovah's Witness patients

In the presence of certain hemostatic substances

In the presence of some wound sterilizing substances

Surgical excision of pheochromocytoma tumors

Patients with positive viral antigen markers

The safety of blood salvaging in infants and children with sickle cell disease remains unclear. Although sickle cell disease has been described as a contraindication for the use of blood salvage, the scant literature that is available is inconclusive. In infancy, precautions are probably unnecessary, because erythrocytes with high concentrations of HbF molecules appear to be resistant to sickling. However, as HbF concentrations decline, sickle cell disease may begin to manifest itself ( Karayalcin, 1979 ; Esseltine et al., 1988 ). Further studies are required before sickle cell patients are categorically denied the benefits of this important technique.


Acute isovolemic or normovolemic hemodilution (ANH) entails the withdrawal of a calculated volume of the patient's blood and the simultaneous replacement with a cell-free substitute to maintain a near-normal blood volume. This intentional decrease in hemoglobin concentration (dilutional anemia) is accomplished sometime after anesthetic induction but before the critical phases of surgery are started. The patient's own fresh blood is reinfused near the end of the surgical procedure after the major blood loss has ceased. The rationale for the use of ANH as a method for blood conservation is that if intraoperative blood loss is relatively constant, the loss of blood constituents, especially red blood cells, would be reduced if the blood were diluted. The term normo volemic is commonly used although there are no simple means of predicting the accuracy of the normovolemic status and thus slight hypervolemia or hypovolemia cannot be excluded. Although no standard nomenclature for the degree of hemodilution has been established, a decrease in hematocrit from the normal value of 40% to between 25% and 35% is referred to as moderate or limited hemodilution, whereas a hematocrit below 20% has been referred to as profound or extreme hemodilution.

Although hemodilution was recognized as early as 1882 by Kronecker, who demonstrated that the acute dilution of blood to a hematocrit of 15% was compatible with survival, it was not until the 1960s that intentional hemodilution was used as a method of blood conservation during surgery. Since then, its use has been extended to various surgical procedures, including pediatric surgery ( Lawson et al., 1974 ;Laver et al., 1975 ; Viviani et al., 1978 ; Watzek et al., 1980 ; Wong et al., 1980 ; Fahmy, 1985b ).

ANH can alter oxygen transport via its influences on arterial oxygen–carrying capacity and the rheological properties of blood. This important rheological consequence of hemodilution warrants a discussion of some principles of blood viscosity.

Hemorheology of Hemodilution

The viscosity of a liquid is a measure of its internal friction. Viscosity may be defined as the resistance to flow that depends on the intermolecular forces operating within the liquid. The term internal frictionemphasizes that as a fluid moves within a tube, laminae in the fluid slip on one another and move at different speeds. This results in a velocity gradient in a direction perpendicular to the wall of the tube, termed the shear rate. In the circulation, shear rate bears a positive correlation to blood flow.

In an experiment, a liquid is confined between two closely spaced parallel plates (analogous to playing cards) ( Fig. 12-14 ). The area of each plate is A, and the distance between the plates is Y. If a tangential force is applied along one plate, while the other is kept stationary, the plate moves in the direction of the force and a velocity gradient develops in the fluid. For a given fluid, the velocity gradient (or shear rate) varies with the force applied (F) per plate surface area (or shear stress). The force required to displace one plate relative to its neighbor divided by the surface area of the plate is defined as shear stress (dynes/cm2). The distance that the plate is displaced horizontally per unit time relative to its neighbor is known as the shear rate, which is expressed in units of distance per unit time divided by the thickness of the plate. The shear stress required to achieve a given shear rate varies with the viscosity of the fluid. Hence,


FIGURE 12-14  Relation between shear stress and shear rate when a fluid is sheared between two parallel plates. See text for details.  (From Fahmy NR: Techniques for deliberate hypotension: Haemodilution and hypotension. In Enderby GEH, editor: Hypotensive anaesthesia. Edinburgh, 1985b , Churchill Livingstone, with permission.)




The units of viscosity are expressed in dynes/sec per cm2, or poise. The relation between shear stress and shear rate is the basis for comparing viscosities of different fluids and for studying the effect of hemodilution on whole blood viscosity ( Messmer and Sunder-Plassman, 1974 ; Messmer, 1975 ).

Factors affecting whole blood viscosity include (1) shear rate (rate of flow), (2) red blood cell deformability, (3) plasma proteins, (4) red blood cell aggregation, (5) hematocrit, (6) temperature, and (7) size of vessel.

A Newtonian fluid is characterized by constant viscosity during changes in shear rate. Plasma and saline solution are typical Newtonian fluids; the shear rates increase linearly with increasing shear stress and thus the viscosity remains unaltered. Because blood is a suspension of cells in plasma, it does not exhibit strict Newtonian behavior. Blood behaves like a non-Newtonian fluid when moving slowly and like a Newtonian fluid when flowing rapidly. The viscosity of whole blood increases sharply at low flow rates (as in postcapillary venules) but remains constant at high and moderate rates of blood flow (as in arteries and arterioles). This results from the increased tendency of red blood cells to concentrate in the center of the stream (axial streaming) and from the increased red blood cell deformation occurring at high flow rates (the tendency of red blood cells to assume ellipsoid patterns with their longitudinal axis aligned in the direction of the flow). Plasma proteins, specifically fibrinogen and γ-globulins, exert pronounced effects on red blood cell aggregation. When normal red blood cells are suspended in a medium with the same fluid viscosity as plasma but without the long plasma proteins required for cell bridging, shear-dependent behavior is decreased markedly.

Hematocrit directly influences blood viscosity. The higher the hematocrit, the more friction exists between the layers of blood, resulting in increased viscosity. An increase in hematocrit from 40% to 60% is associated with an approximate doubling of the viscosity, whereas a decrease from 40% to 20% results in an approximate 50% decrease in viscosity ( Messmer, 1975 ) ( Fig. 12-15 ). The most dramatic decrease in blood viscosity during hemodilution is seen in the postcapillary venules, when the hematocrit decreases from the normal level to approximately 25%. Further decrease in hematocrit is not accompanied by further reduction in viscosity ( Fig. 12-16 ).


FIGURE 12-15  Viscosity of whole blood at various hematocrit values, determined at various shear rates (from 11.5 to 230 per second). Hematocrit was varied by the addition of dextran and packed red blood cells. The viscosity of plasma is also shown for comparison. Note that whole blood viscosity increases with hematocrit and that the increments in viscosity are greatest at the lower shear rates.  (From Messmer K, Sunder-Plassman L: Hemodilution. Prog Surg 13:208, 1974.)





FIGURE 12-16  Schematic representation of the effect of hemodilution on whole blood viscosity as related to changes in shear rate in vivo in different vascular compartments. The most pronounced decrease in blood viscosity, and hence resistance to flow, should occur within the postcapillary venules when the hematocrit (HCT) is decreased from its control level (45%) to approximately 30%, that is, limited hemodilution. Further decreases in hematocrit (extreme hemodilution) will reduce viscosity remarkably less.  (Modified from Messmer K, Sunder-Plassman L: Hemodilution. Prog Surg 13:208, 1974.)




Temperature bears an inverse relation to blood viscosity. The decrease in hematocrit required to maintain constant blood viscosity during hypothermia is shown ( Fig. 12-17 ). At 20°C, a reduction in hematocrit from 45% to 25% is needed to restore viscosity to the same value measured at 37°C. After circulatory arrest, however, the shear stress required to reinitiate flow and break up the red blood cell aggregates is likely to be high. Additional rheologic benefit may be gained by a further decrease in hematocrit.


FIGURE 12-17  Illustration of decrease in hematocrit in two patients that must accompany a reduction in body temperature if viscosity is to remain constant. The measurements were made at a low shear rate with an initial hematocrit at 37°C of 45%. Data were obtained with whole blood from two normal adults.  (Modified from Larson L: Changes in flow properties in human blood [thesis]. Bozeman MT, 1973, Montana State University.)




The tendency for increased hematocrit to increase blood viscosity is attenuated when blood flows through tubes of capillary diameter. This is because red blood cells are normally very deformable, and with a diameter similar to that of the capillary, they can squeeze through the vessel lumen in single file with minimal extra force required. The rate at which red blood cells pass through the capillary has little influence on blood viscosity.

Physiologic Responses to Acute Normovolemic Hemodilution

The cardiovascular responses during ANH depend on the degree of hemodilution, the circulating blood volume, the nature of the diluent, and the efficacy of compensatory physiologic mechanisms.

Systemic Responses

The decrease in hemoglobin concentration during ANH leads to a proportional decrease in arterial oxygen content (CaO2). Physiological mechanisms available to compensate for this decrease in CaO2 are (1) increased cardiac output and (2) increased oxygen extraction ratio.

An increase in cardiac output is the main compensatory mechanism during moderate ANH. The increases in cardiac output are primarily due to an augmented stroke volume, although an increased heart rate may assume an important role when basal heart rate is low. The increases in stroke volume are attributable to (1) enhanced myocardial contractility involving activation of the cardiac sympathetic nerves and β-adrenergic receptors, (2) reduction in impedance to left ventricular ejection as a result of decreases in blood viscosity and peripheral vascular resistance, and (3) increased venous return because of reduced peripheral vascular resistance.

The systemic oxygen delivery, the product of cardiac output and CvO2, is reasonably well maintained as long as hematocrit is not reduced below 25% ( Fig. 12-18 ). When oxygen delivery decreases during extreme ANH, oxygen consumption can be maintained if mixed venous oxygen content (CvO2) falls more than CvO2; that is, if oxygen extraction ratio increases resulting in reductions in mixed venous oxygen saturation (SvO2). Increases in oxygen extraction ratio from the normal value of 25% to 50% or higher have been demonstrated during extreme ANH. Such increases in oxygen extraction ratio have been shown to coincide with the onset of cardiac lactate production and cardiac dysfunction. Based on this finding, it has been proposed that an oxygen extraction ratio of 50% may serve as a “trigger” for blood transfusion in severely hemodiluted patients.


FIGURE 12-18  Relationship between hematocrit value and theoretical relative oxygen delivery of blood. Values were calculated from the whole blood viscosity assuming that a hematocrit value of 45% corresponds with a hemoglobin value of 100% and that the velocity of flow is inversely proportional to the viscosity of blood.  (Modified from Hint H: The pharmacology of dextran and clinical background of the clinical use of Rheomacrodex. Acta Anaesthesiol Belg 19:119, 1968, with permission.)




In chronic anemia, a rightward shift of the oxyhemoglobin dissociation curve may facilitate unloading of oxygen at tissue. This mechanism, which apparently is due to an increased 2,3-DPG concentration within the red blood cells, appears to have no important role during ANH.

Regional Responses


During ANH, the heart is the principal organ at risk. This is due to (1) an augmented contractile demand, (2) a low baseline coronary venous PO2 resulting in a limited oxygen extraction reserve, and (3) the tendency for subendocardial ischemia in the left ventricular wall, when tachycardia, aortic hypotension, or both occur in the presence of a dilated coronary vasculature.

In normal hearts, blood flow to both the left and right ventricular myocardium increases in proportion to the decreases in hematocrit during ANH ( Fig. 12-19 ). As long as hematocrit is not reduced below a critical value (approximately 10%), the increases in myocardial blood flow are transmurally uniform and sufficient to maintain myocardial oxygen consumption and oxygen delivery. The adequacy of myocardial oxygen delivery is indicated by the unchanged PO2 of coronary venous blood (a reflection of PO2 within the myocardium) and well-preserved myocardial lactate extraction and uptake ( Fig. 12-20 ), as well as by stability of indices of cardiac performance, including aortic pressure, left atrial pressure, and left ventricular contractility. Although a reduction in blood viscosity has been shown to contribute to the decreased coronary vascular resistance during ANH, significantly diminished reactive hyperemic responses imply that coronary vasodilation via metabolic mechanisms (presumably in response to reduced CaO2) plays a prominent role ( Fig. 12-21 ). Hearts with stenotic coronary arteries (resulting in diminished vasodilator reserve) are less tolerant of ANH; that is, they exhibit signs of global cardiac dysfunction at a higher hematocrit.


FIGURE 12-19  Comparison of effect of graded hemodilution on parameters of O2 delivery and uptake in the right and left ventricles (RV and LV, respectively) in anesthetized dogs. A-V O2 diff, arteriovenous O2 difference. *P < .05, from value at hematocrit of 40%. (From Crystal GJ, Kim S-J, Salem MR: Right and left ventricular O2 uptake during hemodilution and β-adrenergic stimulation. Am J Physiol 265:H1769–H1777, 1993, with permission.)





FIGURE 12-20  Comparison of effect of graded hemodilution on right ventricular (RV) and left ventricular (LV) lactate extraction and uptake in anesthetized dogs (*P<.05 from value at hematocrit of 40%).  (From Crystal GJ, Kim S-J, Salem MR: Right and left ventricular O2 uptake during hemodilution and β-adrenergic stimulation. Am J Physiol 265:H1769–H1777, 1993, with permission.)





FIGURE 12-21  Original tracings demonstrating effect of graded hemodilution on reactive hyperemia after 60-second occlusion of right coronary artery of an anesthetized dog. At A, right coronary artery was occluded; at B, occlusion was released.  (From Crystal GJ, Kim S-J, Salem MR: Right and left ventricular O2 uptake during hemodilution and β-adrenergic stimulation. Am J Physiol 265:H1769–H1777, 1993, with permission.)




Other Organs.

ANH has been shown to increase blood flow throughout the central nervous system, to both normal tissue and to areas with impaired autoregulation. The findings of unimpaired electroencephalographic activity, lack of anaerobic metabolism, and well-maintained cerebral oxygen consumption suggest that, on a global basis, cerebral oxygenation is adequate during moderate ANH.

Studies showed that the renal blood flow may increase or decrease during ANH. Although the hepatic arterial and portal vein blood flow may increase during ANH, there is increased hepatic oxygen extraction. In general, blood flow in the splanchnic organs and kidney changes minimally during ANH, resulting in decreases in regional oxygen delivery. These findings suggest that in these organs, the favorable influence of reduced viscosity on blood flow is antagonized by vasoconstriction, which is presumably mediated by the sympathetic vasoconstrictor nerves. The selective peripheral vasoconstriction during ANH serves two important functions: (1) it combines with an increased cardiac output to support aortic pressure and (2) it ensures that the increased cardiac output is preferentially distributed to the vital organs, namely, the heart and brain.

The effect of hemodilution on pulmonary gas exchange is controversial. In one study, blood gas levels did not change significantly among 47 patients in whom normovolemic hemodilution was conducted before anesthetic induction ( Fahmy, 1985b ). On the other hand, during hemodilution, the intrapulmonary shunt (   s/   t) may increase or decrease, depending on the ventilation–perfusion relationship. When the    /   ratios are low, atelectatic areas may have increased perfusion because of hemodilution despite HPV. As a result, a large    s/   t (between 6% and 15%) may be found. When    /   ratio is normal,    s/   t may actually decrease.

Effects on Coagulation

Theoretically, ANH can influence coagulation, bleeding, or both via three mechanisms: (1) the increase in blood flow can increase oozing; (2) the diluent used can adversely influence coagu-lation; and (3) dilution of fibrinogen, platelets, and other coagu lation factors concomitant with ANH can also impair coagulation. Tuman and others (1987) showed that progressive hemodilution does not result in hypocoagulability as measured by thromboelastography. In fact, the coagulation appears to be stimulated during progressive blood loss. It is likely that surgical stress and tissue trauma (with release of tissue thromboplastin and elevations in catecholamine levels) offset any hypocoagulable tendency resulting from hemodilution and loss of coagulation factors. These offsetting factors are probably responsible for the observed increase in coagulability. Other studies showed that as the hematocrit decreases from 39% to 25%, there are proportionate decreases in fibrinogen, platelets, and factors V and VIII; blood coagulation is not significantly impaired as long as the hematocrit is above 20%.

The slight prolongation of prothrombin time and partial thromboplastin time observed during ANH is not associated with any discernible increase in surgical bleeding or blood loss. Although Kramer and others (1979) found a 30% decrease in platelets and a 50% decrease in fibrinogen levels at a hematocrit of 25% during ANH for major vascular procedures, the values returned to normal levels by the end of the first postoperative day. Because the patient's own fresh blood is reinfused after most of the surgical bleeding has ceased when ANH is terminated, coagulation may be improved and the need for allogenic blood and blood products, including plasma and platelets, is reduced.

Choice of Replacement Fluid

ANH can be tolerated only with adequate circulating blood volume. Crystalloids, colloids, and combinations of both have been used as diluents during hemodilution. Lactated Ringer's solution is the most common crystalloid given. A number of colloids have been used, including albumin (5% solution), dextran (5% dextran solution with a molecular weight of 70,000 Da), and hydroxyethyl starch (hetastarch). Dextran-40 (molecular weight = 40,000 Da) has the advantage of retarding rouleaux formation and sludging, but it may result in allergic reactions and coagulation defects. The maximum recommended doses for dextran-40 and dextran-70 are 2 and 1 g/kg, respectively ( Arfors and Bergquist, 1975 ). When colloids are administered, there is a minimal risk of anaphylaxis and blood coagulation may be compromised, especially if a larger volume of hydroxyethyl starch is given ( Egli et al., 1997 ). The advantages and disadvantages of crystalloids and colloids are compared in Table 12-5.

TABLE 12-5   -- Crystalloid versus colloid as the replacement fluid for hemodilution




Volume required

3 times volume of shed blood

1 to 2 times volume of shed blood

Plasma volume

80% leaves intravascular compartment in 2 hr

Retained longer in the circulation—a better plasma volume expander

Water balance

Positive water balance, more peripheral edema, responds to diuretics

Less peripheral edema

Colloidal osmotic pressure

Reduced, probably not important


Postoperative hematocrit

Higher hematocrit

Lower hematocrit

Coagulation defect


May occur with excessive colloid therapy (dextrans)






When crystalloids are given as the sole replacement fluid, to maintain normovolemia they should be administered in a volume three times that of the blood withdrawn. Less volume is required when colloids are used as the sole diluents. Crystalloids traverse the capillary endothelium so that, within 2 hours, 80% of the administered volume is in the extravascular space ( Gammage, 1987 ). In contrast, colloids are retained longer in the circulation, thereby maintaining the colloid osmotic pressure and plasma volume for several hours. Consequently, albumin and other colloids produce slightly lower hematocrits intraoperatively and postoperatively than does lactated Ringer's solution ( Hallowell et al., 1978 ). Crystalloids dilute serum proteins and lead to a reduction in colloid osmotic pressure, the importance of which is still controversial. Diuresis follows the administration of crystalloids, and there usually is a rapid response to furosemide.

Although peripheral edema may occur after crystalloid therapy, pulmonary edema rarely ensues ( Brinkmeyer et al., 1983 ). When the lungs are normal, pulmonary edema is usually prevented by intrinsic compensatory factors, including the plasma–interstitial oncotic gradient, the high lymphatic capacity of the lung, the physicochemical characteristics of the interstitial space, the integrity of the microvascular membrane, and a low vascular hydrostatic driving pressure. These factors appear to be less effective in patients with underlying cardiopulmonary disease, thus promoting the development of pulmonary edema. If administered in the presence of capillary leak, albumin may cross the pulmonary capillary endothelium, pulling water with it and increasing pulmonary interstitial water. It must be emphasized that any fluid can be administered in excess to produce pulmonary edema. Probably the amount of the fluid administered and the vigilance in monitoring the hemodynamic variables are more important than the choice of fluid.

Clinical Management of Acute Normovolemic Hemodilution

Patients who are to undergo operations in which major blood losses are expected (more than one third of their blood volume) may be considered candidates for ANH. The technique has been extended to pediatric cardiac surgery, spinal surgery for scoliosis, and operations for malignant disease (Wilms tumor, neuroblastoma, teratomas, retroperitoneal ganglioneuroma, liver tumors, and pancreatic tumors). Although age is not considered an absolute contraindication to the use of ANH and has been used in patients weighing 5 kg ( Schaller et al., 1983 ), the use of ANH in small children should be limited to experienced clinicians.

The parents or responsible guardians should be fully informed as to the rationale and methods of hemodilution before surgery. Risks of massive intraoperative blood loss and transfusion of allogenic blood should be explained. A realistic assessment of the benefit-to-risk ratio of the technique must be presented. Accurate preoperative assessment of the cardiovascular, respiratory, and other systems is essential. A history of drug therapy is important, and steps should be taken to correct any coagulation disorders present. Unless their use is necessary, salicylates (such as indomethacin) and other cyclooxygenase inhibitors should be discontinued 1 or 2 weeks before surgery.

After anesthetic induction and tracheal intubation, an arterial cannula is inserted for blood sampling and arterial blood pressure monitoring. Additional large-bore peripheral intravenous catheters may be placed. A central venous (or a pulmonary artery) catheter may be placed, if indicated. Other monitors should include use of a precordial or an esophageal (or both) stethoscope, electrocardiography, pulse oximetry, capnography, and an esophageal or rectal temperature probe. An indwelling catheter is inserted in the urinary bladder to permit urine volume measurements.

Anesthesia is maintained with an inhalational anesthetic and oxygen. Opioids or combinations of inhaled anesthetics and opioids may also be used. Muscular relaxation is maintained with appropriate monitoring of the neuromuscular function. Ventilation is adjusted to maintain PaCO2 between 30 and 40 mm Hg. Arterial pH, blood gas tensions, and hematocrit are measured every 30 minutes during and after hemodilution and during the critical stages of surgery. If central or pulmonary artery catheters are placed, measurements of central venous or mixed venous blood gas tensions may yield valuable information.

After a steady state of anesthesia is achieved, ANH is started. Using a strictly sterile technique, the predetermined volume of blood ( Box 12-9 ) is withdrawn from either the arterial or the central venous catheter and collected into one or more 250-mL citrate-phosphate-dextrose (CPD) blood donor bags. If an adult bag is used in an adolescent, the bag should contain no less than 300 mL and no more than 450 mL to ensure proper blood-to-anticoagulant ratio. The collection of blood from a central venous line is facilitated by placing the bag lower than the patient level. If a peripheral vein is used, withdrawal of blood can be enhanced by cycling the cuff of a noninvasive blood pressure monitor at 2- or 3-minute intervals. The use of kits that contain two blood bags, a Y-type connector set with a Luer-Loc adapter, and a blood recipient identification band simplifies the procedure. The collected blood should be mixed with the anticoagulant in the blood bag, and care should be taken to exclude air bubbles. The volume of blood in the CPD bags is determined by weighing the bags (1 mL=1 g). As blood is being withdrawn, two to three times this amount of lactated Ringer's solution (or an appropriate amount of other diluents) is infused simultaneously. The lactated Ringer's solution should be warmed before its administration to prevent a decrease in the patient's temperature. The blood bags are numbered sequentially, labeled, and kept at room temperature for up to 6 hours to preserve platelet function. If it is expected that the blood will not be transfused within this period, the blood bags should be kept in a small cooler containing wet ice or arrangements should be made with the blood bank for storage up to 24 hours. Although surgery may be started earlier, phases of surgery during which large blood loss tends to occur should not be allowed until ANH is completed. This usually takes about 30 minutes.

BOX 12-9 

Calculation of Volume of Blood Removed for Hemodilution



ERCVT = [HcT × V × bodyweight (kg)]/100 where ERCVT = estimated total red cell volume (mL) and V = estimated blood volume per kilogram of body weight:



90 mL/kg (neonates)



90 mL/kg (infants and children)



65 to 75 mL/kg (teenagers)



ERCV20 = 0.2 × V × bodyweight (kg) where ERCV20 = estimated red blood cell volume at HcT 20%.[*]



RCW = ERCVT-ERCV20 where RCW = red blood cell volume to be withdrawn.



WBW = 3× RCW where WBW = whole blood volume to be withdrawn.[†]

From Schaller RT Jr, Schaller J, Morgan A, et al.: Hemodilution anesthesia: a valuable aid to major cancer surgery in children. Am J Surg 146:79, 1983, with permission of Excerpta Medica, Inc.


*  Hematocrit of 20% is chosen because it is commonly used.

†  The average hematocrit value of the blood withdrawn is assumed to be 33% and thus the total volume withdrawn is three times the red blood cell volume to be withdrawn.

A decision should be made regarding the lowest acceptable hematocrit. Unless hypothermia is used, the lowest safe acceptable hematocrit is 20% in normal healthy children. A higher hematocrit may be chosen depending on the condition of the patient and the anticipated blood loss. Measurements of hematocrit must be accurate during ANH, and it is preferable to do more than one measurement per sample. Heel sticks are unreliable and should not be used for measurements of hematocrit during hemodilution. Conventional laboratory hematocrit measurements, including the microcentrifuge technique or the Coulter method, are accurate during ANH. In contrast, hematocrit values derived by the conductivity of whole blood used in portable compact devices provide artificially low readings in situations where plasma has been replaced by crystalloids.

During the operative procedure, blood loss is initially replaced with an equal volume of lactated Ringer's solution. Third-space losses during surgery are compensated for by additional volumes of lactated Ringer's solution. If the hematocrit decreases below the desired level, allogenic packed red blood cells may be infused to maintain the hematocrit at an acceptable level. The temptation to reinfuse the patient's own blood should be resisted until blood loss has ceased.

The autologous blood is reinfused after most of the blood loss has ceased. Reinfusion is started with the latest collected blood bag. The first obtained blood bag, which is rich in red blood cells, platelets, and coagulation factors, is administered last. The autologous blood is infused through a 170-μm filter, rather than through a 40-μm filter, to avoid trapping of platelets. Fresh autologous blood flows easily through the filter compared with allogenic blood, and the filter remains remarkably free from debris. The hematocrit should be returned to 28% to 30% after surgery. The remaining autologous blood may be refrigerated and can be used within 24 hours.

Furosemide (0.5 to 1 mg/kg) may be administered to promote diuresis and the rapid excretion of excess crystalloids. Large urine volumes with high electrolyte content are expected over the next several hours after ANH is reversed, especially if crystalloids have been given as the sole diluent. Additional doses of furosemide may be needed over the next 2 hours. Blood electrolyte levels may be measured so that hypokalemia resulting from diuresis can be corrected. Postoperatively, the patients should be cared for in an intensive care environment.

Acute Normovolemic Hemodilution and Cardiopulmonary Bypass

ANH before the institution of cardiopulmonary bypass provides a source of autologous blood rich in hemoglobin, platelets, and coagulation factors, for reinfusion after the cessation of cardiopulmonary bypass, and it mitigates the increase in viscosity that accompanies hypothermia. During hypothermic cardiopulmonary bypass to 20 ° C, a decrease in hematocrit from 45% to 25% is necessary to restore viscosity to the same level observed at 37 ° C (see Fig. 12-17 ). After hypothermic circulatory arrest, the shear stress required to reinitiate flow and to break up red blood cell aggregates is likely to be high. An additional rheologic benefit may be given by further decrease in hematocrit.

Although blood can be withdrawn after induction of anesthesia or during the early stage of surgery in patients with normal cardiac function, it is usually done after heparinization and cannulation immediately before instituting cardiopulmonary bypass. The blood is withdrawn from the tubing to the oxygenator or before the blood passes through the roller pump. Alterations in the pump flow rate can be used to compensate for hemodynamic instabilities associated with blood removal. On initiation of cardiopulmonary bypass with a crystalloid (or crystalloid-colloid) prime, a further and sometimes profound hemodilution results depending on the patient's hemoglobin level before bypass.

Current Status of Acute Normovolemic Hemodilution

Prospective, randomized clinical trials showed that the use of ANH reduces allogenic red blood cell transfusion requirements. Some found that ANH decreases allogenic blood use by 25% to 60%, whereas others reported that more than 90% of patients did not require allogenic blood when ANH is combined with other strategies such as PABD or intraoperative blood salvage ( Spahn and Casutt, 2000 ; Matot et al., 2002 ). The use of ANH was also found to be an independent factor reducing allogenic red blood cell transfusion in addition to maintaining normothermia and the use of cell salvage ( Schmied et al., 1998 ). Although the effect of moderate ANH in reducing blood loss is not as great as deliberate hypotension, evidence suggests that the combination of both techniques leads to marked reduction in blood loss and in the need for allogenic blood ( Fahmy, 1985b ). Despite these enthusiastic reports, concerns have been raised as to the efficacy of ANH as a blood conservation measure. Large well-controlled, randomized studies with clearly defined transfusion triggers have been recommended ( Spahn and Casutt, 2000 ).

It is apparent that moderate ANH used as the sole method may contribute only modestly to blood conservation. The efficacy may vary greatly and is certainly enhanced if a lower posthemodilution hematocrit level, in the range from 28% to 20%, is targeted. Although a hematocrit level of 20% has been recommended as the lowest acceptable hematocrit, profound ANH has been advocated in a few centers, especially when combined with hypothermia or when avoiding allogenic blood transfusion is vital. Fontana and others (1995) showed that healthy patients undergoing scoliosis surgery can be safely hemodiluted to an average hemoglobin of 3 g/dL without signs of global hypoxia or impairment of global cardiac performance. However, that study was limited to children/adolescents with normal cardiac function and performed under very controlled conditions. The use of extreme ANH is not recommended unless it is performed in specialized centers.

ANH offers unique advantages. It can be used in situations where PABD is not planned because of urgency of the operative procedure or scheduling conflicts and whenever blood salvage cannot be performed because of unavailability of technical personnel or lack of equipment. Unlike PABD or blood salvage, ANH is performed by the anesthesiologist and therefore may not require the presence of blood bank technicians or perfusionists. The cost of ANH is approximately one third of the cost of PABD. Finally, ANH is the only practical means of providing fresh autologous whole blood for transfusion.

Despite its simplicity and safety, ANH should not be undertaken in certain situations. Inexperience of the team is an absolute contraindication to the use of ANH. The presence of a coexisting disease that may potentially jeopardize tissue oxygen delivery, especially to the heart or brain, is also a contraindication. ANH should not be performed in anemic patients (hemoglobin <11 g/dL) and also should not be used if the anticipated compensatory increase in cardiac output is neither possible nor desirable. The technique is contraindicated in patients with pulmonary disease resulting in impaired arterial oxygenation. Patients with renal disease may not be suitable candidates for ANH because of inability to excrete large amounts of crystalloids. The success of ANH depends on a cooperative effort between surgeon, anesthesiologist, nurses and laboratory personnel, and on good communication between them.

It seems that in the future ANH will continue to be practiced either as the sole blood conservation measure or in combination with other strategies to maximize its efficacy. It may totally replace PABD (Goodnough et al., 1997 ). It may also be used in conjunction with preoperative erythropoietin therapy and with oxygen-carrying blood substitutes ( Spahn and Cassut, 2000) . Because the compensatory increase in cardiac output may be insufficient in itself to restore the oxygen extraction reserve, it has been hypothesized (and confirmed in an animal study) that augmentation of the cardiac output pharmacologically could reverse the increase in oxygen extraction ratio, restore the margin of safety for tissue oxygenation, and extend the limit to which hematocrit can be reduced safely during ANH (Crystal and Salem, 2002 ). The safety of this approach has not yet been confirmed in humans.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier



Because deliberate hypotension can decrease blood loss and because ANH minimizes the need for allogenic blood transfusion, a combination of the techniques has been proposed ( Fahmy, 1985a ). This combined technique has been used primarily for major orthopedic procedures. When hemodilution and hypotension were combined, allogenic blood replacement was decreased by about 80% of the blood replacement required during nor motensive anesthesia compared with 45% when hypotension was used alone ( Fahmy, 1985a ). Observations with the combined use of both techniques indicate that cardiac output tends to decrease after ANH, when the blood pressure decreases below 60 mm Hg. Because decreases in blood pressure can be achieved easily in hemodiluted patients, the dose of hypotensive drugs should be decreased accordingly.

Animal studies of the regional hemodynamic responses to the combination of hemodilution and controlled hypotension reveal that maintenance of oxygen delivery to critical tissue beds may be at risk (Plewes and Farhi, 1985 ; Crystal et al., 1988 ). These animal studies emphasize the importance of preoperative evaluation of patients, vigilance, experience, use of high FIO2, and continuous monitoring of the arterial pressure, blood gases, blood loss, body temperature, and urine output when hypotension and hemodilution are combined ( Fahmy, 1985a ).


The technique of combined ANH, hypotension, and hypothermia was stimulated by the need to carry out major surgical operations (usually associated with massive blood loss) in patients of the Jehovah's Witness faith who refuse to receive blood or blood products ( Laver et al., 1975 ; Lilleaasen et al., 1978 ; Schaller et al., 1983 ; Singler, 1989 ). The rationale of this technique is that because the increased cardiac output may not be sufficient to maintain an adequate oxygen delivery during profound ANH, the protective effects of cooling are used, that is, decreased tissue oxygen requirement and increased fraction of dissolved oxygen in the blood. Anesthetics also decrease the tissue oxygen demand. Moderate hypotension decreases both blood losses and myocardial oxygen demand.

During air breathing at normothermia, the dissolved oxygen in the blood represents 0.3 mL/dL. This value rises to 1.3 to 1.5 mL/dL in normal patients during 100% oxygen administration. This represents about 7% of the CaO2 and 30% of the total oxygen extraction in patients with a normal hemoglobin level at normal temperature ( Singler, 1989 ). As temperature and plasma protein content fall during ANH combined with hypothermia, the fraction of dissolved oxygen increases. With hypothermia to 30° to 31°C and an FIO2 of 1.0, the amount of dissolved oxygen increases to 2 mL/dL; during ANH a hemoglobin of 5 g/dL represents 23% of the total oxygen content (8.7 mL/dL). This dissolved oxygen accounts for more than half of the metabolic requirements ( Fig. 12-22 ). When oxygen consumption decreases by 40% as a result of anesthesia and cooling to 31°C, SvO2 does not change during a reduction in hemoglobin concentration from 15 g/dL to 5 g/dL. When this technique is used properly, the combined effect of increased dissolved oxygen in the blood, increased blood flow, decreased systemic vascular resistance, and decreased oxygen requirements offsets the effect of decreased hemoglobin concentration and results in maintenance of adequate tissue perfusion and adequate oxygen extraction. Because this technique of hemodilution, hypotension, and hypothermia requires extensive experience, it should be used only by experienced clinicians in very specialized centers.


FIGURE 12-22  Curves plotting total arterial oxygen content (CaO2) against PO2 during the awake normal state and during hemodilution. The increase in dissolved oxygen during hemodilution represents approximately 30% of total CaO2. See text for details.  (From Singler RC: Special techniques: Deliberate hypotension, hypothermia, and acute normovolemic hemodilution. In Gregory GA, editor: Pediatric anesthesia. New York, 1989, Churchill Livingstone, with permission.)




Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


The search for a blood substitute that could be stored at room temperature and administered on the battlefield without the need for cross-matching began more than 70 years ago. With time, the focus gradually shifted from the search for a “blood substitute” to the development of “oxygen therapeutics” ( Wahr, 2003 ). There are three types of “oxygen therapeutic” agents: (1) hemoglobin-based solutions, (2) perfluorochemical solutions, and (3) allosteric modifiers of hemoglobin.


Removal of hemoglobin, the oxygen-carrying moiety, from red blood cells initially seemed to offer the following advantages: purification, elimination of ABO incompatibility by removal of the antigens on the red blood cell surface, and extension of the shelf-life to several months or even years. Because hemoglobin is an inherently unstable molecule once it is removed from the red blood cell, the normal tetramer dissociates resulting in dimers easily filtered by the kidney. In addition, the loss of 2,3-DPG markedly reduces the P50. Furthermore, free hemoglobin in the plasma is rapidly engulfed by the reticuloendothelial system, resulting in an intravascular half-life of 24 to 36 hours ( Wahr, 2003 ). In the current generation of solutions, many of the problems have been resolved through chemical or genetic manipulations. The hemoglobin molecules under development are tetramers or large polymers, have a P50 of 28 to 37 mm Hg, have hemoglobin concentrations of 10 to 14 g/dL, and are free of antigens and pathogens. Formation of methemoglobin is still a problem, although the maximal amount of methemoglobin remains less than 10% ( Sprung et al., 2002 ).

The hemoglobin in these solutions is derived from several sources'outdated human volunteer donations, cattle (a potential source of nearly 1 million units per day), and recombinant technology usingEscherichia coli ( Wahr, 2003 ). Regardless of the source, they share common problems: short intravascular half-life, some degree of vasoconstriction particularly in the pulmonary circulation, and interference with spectrophotometric laboratory measurements. The pulmonary vasoconstriction is probably related to the role of hemoglobin in nitric oxide equilibration ( Wahr, 2003 ). Nitric oxide is produced by the vascular endothelium and diffuses both to the smooth muscles, where it exerts its vasodilating effect, and into the vascular lumen, where it is scavenged by hemoglobin inside the red blood cells. Free hemoglobin scavenges nitric oxide much more avidly than does red blood cell hemoglobin, resulting in vasoconstriction. Free hemoglobin in the plasma absorbs light and interferes with spectrophotometric measurements.


Perfluorocarbons are inert substances that have high solubility for all gases, including oxygen and carbon dioxide. Because these liquids are completely immiscible in water, intravenous administration causes a fatal lipid embolus. However, a microemulsion of a perfluorocarbon in normal saline was used to perform exchange transfusion in a rat, which survived breathing 100% oxygen with a hematocrit of zero ( Clark and Gollan, 1966 ).

Fluosol is the only “blood substitute” to have been approved by the U.S. Food and Drug Administration. Because of low demand and major disadvantages, namely low concentration (10%), a short intravascular half-life ( Tremper et al., 1982 ; Gould et al., 1986 ), and instability of the emulsion, commercial production was stopped in the late 1990s. Perfluoroctyl bromide (Oxygent; Alliance Pharmaceuticals, San Diego, CA), a second-generation perfluorocarbon emulsion, has been produced, which contains 90% perflubron (C8F17) by weight or 45% by volume. It is emulsified with lecithin and stable at room temperature for longer than 6 years ( Henry et al., 1994 ).

Perfluorocarbons are not metabolized but rather are cleared by the reticuloendothelial system and eventually are exhaled as vapor. A bloodless animal could survive with a “fluorocrit” of 10%, 20%, and 50% with PaO2 values of 800, 650, and 250 mm Hg, respectively. Because the emulsion is cleared within 24 to 36 hours and because repeat dosing results in hepatosplenomegaly in animals, long-term survival would be problematic. There seems to be a dose-related effect on the platelet count, which decreases by 15% to 25% at 2 to 4 days after administration ( Leese et al., 2000 ). With a maximum dose of 3 g/kg, no bleeding abnormalities have been noted in humans ( Wahr, 2003 ).


Two antilipidemic drugs (clofibrate and benzofibrate) were found to decrease the affinity of hemoglobin for oxygen, thereby enhancing oxygen release to tissues ( Poyart et al., 1994 ; Perutz and Poyart, 1983 ). This in vitro “allosteric” effect of these drugs is inhibited by in vivo serum albumin. Effort to synthesize drugs that would affect allosteric modifications of hemoglobin in vivo culminated in development of a compound known as RSR13 ( Abraham et al., 1992 ; Randad et al., 1991 ). Studies showed dose-dependent rightward shift of the oxyhemoglobin dissociation curve ( Wahr et al., 2001 ). A right shift in P50 of 10 mm Hg was achieved at doses of 75 and 100 mg/kg. The only marked side effect reported is a transient increase in the serum creatinine level of three patients who received RSR13. The rightward shift of the oxyhemoglobin dissociation curve is a double-edged sword; although it enhances the release of oxygen to the tissue, SaO2 decreases if FIO2 is not increased.


Unforeseen adverse side effects have kept commercially available oxygen therapeutics tantalizingly “just out of reach.” However, there are indications that clinical use is very near for a number of these products. Transfusion alternatives eventually become commercially available'the question is “when” rather than “if.” After successful use in adults, the use of oxygen therapeutics will ultimately be extended to the pediatric population.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


A variety of simple and special blood conservation procedures are currently available and applicable to pediatric patients. The anesthesiologist should be familiar with the principles and practices of various blood conservation strategies. A “flawless” anesthetic technique and proper positioning are essential so as to avoid increased blood loss during surgery. Infiltration with vaso constrictors is a simple technique that can be used in many minor and major surgical procedures. The use of pneumatic tourniquets after exsanguination of the upper and lower extremity permits a bloodless operative field.

Advances in knowledge of the physiology of acute normovolemic hemodilution have expanded the application of the technique to a variety of surgical procedures and may in the future replace preoperative autologous blood donation. Similarly, refinements in perioperative blood salvage techniques have led to their use in certain pediatric surgical procedures.

Various hypotensive drugs and techniques are available, but the power to initiate hypotension resides entirely with the anesthesiologist who can, by skillful use of gravity, controlled ventilation, and choice of appropriate drugs, maintain the desired level of blood pressure. Factors that may improve the safety of the technique include careful selection of patients, maintenance of airway patency, avoidance of hypercapnia and hypocapnia, use of high FIO2, gradual onset of hypotension aiming at a level consistent with the patient's condition, proper monitoring, and adequate postoperative care. Some pertinent points should be borne in mind when hypotensive anesthesia is used for pediatric patients. (1) Children respond to some hypotensive drugs with tachycardia. (2) The incidence of failed hypotension may be relatively high unless the heart rate is controlled. (3) Because of the child's smaller stature, tilting may not produce as great a pressure gradient and peripheral venous pooling. (4) The physiologic dead space does not increase in infants and children. (5) Lower arterial pressure may be necessary to achieve the desired bloodless field. (6) Cyanide toxicity is a preventable complication.

There has been an increased tendency to combine blood conservation measures so as to markedly reduce or virtually eliminate allogenic blood transfusion. However, these combined techniques demand the utmost in skill, training, and experience. Great advances have been made in the use of oxygen-carrying blood substitutes, as well as drugs that enhance hemostatic activity, and these approaches are expected to be in common use in the near future.

Copyright © 2008 Elsevier Inc. All rights reserved. -

Motoyama & Davis: Smith's Anesthesia for Infants and Children, 7th ed.

Copyright © 2005 Mosby, An Imprint of Elsevier


American Association of Blood Banks, 1993.   American Association of Blood Banks: Guidelines for blood salvage and reinfusion in surgery and trauma. Bethesda, MD, 1993.

Abraham et al., 1992. Abraham DJ, Wireko RC, Randad RS, et al: Allosteric modifiers of hemoglobin: 2-[4-[[(3,5-disubstituted anilino)carbonyl]methyl]phenoxy]-2-methylpropionic acid derivatives that lower the oxygen affinity of hemoglobin in red cell suspensions, in whole blood, and in vivo in rats.  Biochemistry  1992; 31:9141.

Adams, 1975. Adams AP: Techniques of vascular control for deliberate hypotension during surgery.  Br J Anaesth  1975; 47:777.

Adu-Gyamfi et al., 1993. Adu-Gyamfi Y, Sankarankutty M, Marwa S: Use of a tourniquet in patients with sickle-cell disease.  Can J Anaesth  1993; 40:24.

Alajmo et al., 1989. Alajmo F, Calamai G, Perna AM, et al: High-dose aprotinin: Hemostatic effects in open heart operations.  Ann Thorac Surg  1989; 48:536.

Amaranath et al., 1975. Amaranath L, Cascorbi HF, Singh-Amaranath AV, et al: Relation of anesthesia to total hip replacement and control of operative blood loss.  Anesth Analg  1975; 54:641.

Anderson and McKissock, 1953. Anderson S, McKissock W: Controlled hypotension with Arfonad in neurosurgery with special reference to vascular lesions.  Lancet  1953; 2:754.

Arfors and Bergquist, 1975. Arfors KE, Bergquist D: Microvascular haemostatic plug formation in the rabbit mesentery.  Bibl Haematol  1975; 41:84.

Aronson et al., 1990. Aronson S, Goldberg LI, Roth S, et al: Preservation of renal blood flow during hypotension induced with fenoldopam in dogs.  Can J Anaesth  1990; 37:380.

Askrog et al., 1964. Askrog VF, Pender JW, Eckenhoff JE: Changes in physiological dead space during deliberate hypotension.  Anesthesiology  1964; 25:774.

Atlee and Roberts, 1986. Atlee JL, Roberts FL: Thiopental and epinephrine-induced dysrhythmias in dogs anesthetized with enflurane and isoflurane.  Anesth Analg  1986; 65:437.

Behnia et al., 1982. Behnia R, Martin A, Koushanpour E, et al: Trimethaphan-induced hypotension: Effect on renal function.  Can Anaesth Soc J  1982; 29:581.

Behnia et al., 1978. Behnia R, Raymon F, Cheng SE, et al: Metabolism of sodium nitroprusside in dogs awake and anesthetized with halothane.  Anesthesiology  1978; 48:260.

Bendo et al., 1993. Bendo AA, Kozlowski PB, Capuano C, et al: Cerebral edema formation in dogs following hypotension induced with isoflurane and labetalol.  Acta Anaesthesiol Belg  1993; 44:103-109.

Benumof and Wahrenbrock, 1975. Benumof JL, Wahrenbrock EA: Blunted hypoxic pulmonary constriction by increased lung vascular pressure.  J Appl Physiol  1975; 38:846.

Bernard et al., 1992. Bernard J-M, Passuti N, Pinaud M: Long-term hypotensive technique with nicardipine and nitroprusside during isoflurane anesthesia for spinal surgery.  Anesth Analg  1992; 75:179.

Berry, 1981. Berry AJ: Respiration support and renal function.  Anesthesiology  1981; 55:655.

Biro, 1986. Biro GP: Large venous compliance in carboxyhemoglobinemia and hemodilutional anemia.  Can J Physiol Pharmacol  1986; 64:556.

Blanchette and Zipursky, 1984. Blanchette VS, Zipursky A: Assessment of anemia in newborn infants.  Clin Perinatol  1984; 11:489.

Blau et al., 1992. Blau WS, Kafer ER, Anderson JA: Esmolol is more effective than sodium nitroprusside in reducing blood loss during orthognathic surgery.  Anesth Analg  1992; 75:172.

Blevins et al., 1993. Blevins FT, Shaw B, Valeri CR, et al: Reinfusion of shed blood after orthopaedic procedures in children and adolescents.  J Bone Joint Surg [Am]  1993; 75:363.

Bloch et al., 1992. Bloch EC, Ginsberg B, Binner Jr RA, et al: Limb tourniquets and central temperature in anesthetized children.  Anesth Analg  1992; 74:486.

Boezaart et al., 1995. Boezaart AP, van der Merwe I, Coetzee A: Comparison of sodium nitroprusside- and esmolol-induced controlled hypotension for functional endoscopic sinus surgery.  Can J Anaesth  1995; 42:373.

Bourke et al., 1989. Bourke DL, Silberberg MS, Ortega R, et al: Respiratory responses associated with release of intraoperative tourniquets.  Anesth Analg  1989; 69:541.

Brinkmeyer et al., 1983. Brinkmeyer S, Safar P, Motoyama EK: Superiority of colloid over electrolyte solution for severe normovolemic hemodilution (NVHD) in concurrent treatment of hemorrhage.  Disaster Med  1983; 1:171.

Brodkey et al., 1972. Brodkey JS, Richards DE, Blasingame JP, et al: Reversible spinal cord trauma in cats: Additive effects of direct pressure and ischemia.  J Neurosurg  1972; 37:591.

California statute SB-37, 1990.   California statute SB-37. Paul Gann Blood Safety Act, January 1, 1990.

Calverley et al., 1978. Calverley RK, Smith NT, Jones CW, et al: Ventilatory and cardiovascular effects of enflurane anesthesia during spontaneous ventilation in man.  Anesthesiology  1978; 57:610.

Carlson et al., 1969. Carlson LA, Ekelund LG, Orö L: Circulatory and respiratory effects of different doses of prostaglandin E1 in man.  Acta Physiol Scand  1969; 75:161.

Casthely et al., 1982. Casthely PA, Lear S, Cottrell JE, et al: Intrapulmonary shunting during induced hypotension.  Anesth Analg  1982; 61:231.

Cheney and Colley, 1980. Cheney FW, Colley PS: The effect of cardiac output on arterial blood oxygenation.  Anesthesiology  1980; 52:496.

Clark and Gollan, 1966. Clark Jr L, Gollan F.: Survival of mammals breathing organic liquids equilibrated with oxygen at atmospheric temperature.  Science  1966; 152:1755.

Cottrell et al., 1978a. Cottrell JE, Patel K, Turndorf H, et al: Intracranial pressure changes induced by sodium nitroprusside in patients with intracranial mass lesions.  J Neurosurg  1978; 48:329.

Cottrell et al., 1978b. Cottrell JE, Casthely P, Brodie JD, et al: Prevention of nitroprusside-induced cyanide toxicity with hydroxocobalamin.  N Engl J Med  1978; 298:809.

Cottrell et al., 1978c. Cottrell JE, Patel K, Cathely P, et al: Nitroprusside tachyphylaxis without β acidosis.  Anesthesiology  1978; 9:141.

Crystal et al., 1993. Crystal GJ, Kim S-J, Salem MR: Right and left ventricular O2 uptake during hemodilution and β-adrenergic stimulation.  Am J Physiol  1993; 265:H1769-H1777.

Crystal et al., 1988. Crystal GJ, Rooney MW, Salem MR: Regional hemodynamics and oxygen supply during isovolemic hemodilution alone and in combination with adenosine-induced controlled hypotension.  Anesth Analg  1988; 67:211.

Crystal and Salem, 1996. Crystal GJ, Salem MR: Acute normovolemic hemodilution.   In: Salem MR, ed. Blood conservation in the surgical patient,  Baltimore, MD: Williams & Wilkins; 1996:168-188.

Crystal and Salem, 2002. Crystal GJ, Salem MR: Beta-adrenergic stimulation restores oxygen extraction reserve during acute normovolemic hemodilution.  Anesth Analg  2002; 95:851.

Cullen and Eger, 1974. Cullen DJ, Eger II EI: Cardiovascular effects of carbon dioxide in man.  Anesthesiology  1974; 41:345.

Davies et al., 1975. Davies DW, Greiss L, Kadar D, et al: Sodium nitroprusside in children: Observations on metabolism during normal and abnormal responses.  Can Anaesth Soc J  1975; 22:553.

Davis et al., 1993. Davis RJ, Agnew DK, Shealy CR, et al: Erythrocyte viability in postoperative autotransfusion.  J Pediatr Orthop  1993; 13:781.

Delaney and Miller, 1980. Delaney TJ, Miller Jr ED: Rebound hypertension after sodium nitroprusside prevented by saralasin in rats.  Anesthesiology  1980; 52:154.

Diaz and Lockhart, 1979. Diaz JH, Lockhart CH: Hypotensive anaesthesia for craniectomy in infancy.  Br J Anaesth  1979; 51:233.

Dickson et al., 1990. Dickson M, White M, Kinney W, et al: Extremity tourniquet deflation increases end-tidal PCO2.  Anesth Analg  1990; 70:457.

Dietrich et al., 1990. Dietrich W, Spannagl M, Jochum M, et al: Influence of high-dose aprotinin treatment on blood loss and coagulation patterns in patients undergoing myocardial revascularization.  Anesthesiology  1990; 73:1119.

Dilthey et al., 1993. Dilthey G, Dietrich W, Spannagl M, Richter J: Influence of desmopressin acetate on homologous blood requirements in cardiac surgical patients pretreated with aspirin.  J Cardiothorac Vasc Anesth  1993; 7:425.

Don, 1983. Don H: Hypoxemia and hypercapnia during and after anesthesia.   In: Orkin FK, Cooperman LH, ed. Complications in anesthesiology,  Philadelphia: Lippincott; 1983:183-207.

Eckenhoff et al., 1963a. Eckenhoff JE, Enderby GEH, Larson A, et al: Pulmonary gas exchange during deliberate hypotension.  Br J Anaesth  1963; 35:750.

Eckenhoff et al., 1963b. Eckenhoff JE, Enderby GEH, Larson A, et al: Human cerebral circulation during deliberate hypotension and head-up tilt.  J Appl Physiol  1963; 18:1130.

Edmondson et al., 1989. Edmondson R, Del Valle O, Shah N, et al: Esmolol for potentiation of nitroprusside-induced hypotension: Impact on the cardiovascular, adrenergic, and renin-angiotensin systems in man.  Anesth Analg  1989; 69:202.

Egli et al., 1997. Egli GA, Zollinger A, Seifert B, et al: Effect of progressive haemodilution with hydroxyethyl starch, gelatin and albumin on blood coagulation: An in vitro thromboelastography study.  Br J Anaesth  1997; 78:684.

Enderby, 1950. Enderby GEH: Controlled hypotension with hypotensive drugs and posture to reduce bleeding surgery. Preliminary results with pentamethonium iodide.  Lancet  1950; 1:1145.

Enderby, 1960. Enderby GEH: Halothane and hypotension.  Anaesthesia  1960; 15:25.

Enderby, 1985a. Enderby GEH: Historical review of the practice of deliberate hypotension.   In: Enderby GEH, ed. Hypotensive anaesthesia,  New York: Churchill-Livingstone; 1985:75-91.

Enderby, 1985b. Enderby GEH: Blood pressure and bleeding.   In: Enderby GEH, ed. Hypotensive anaesthesia,  New York: Churchill-Livingstone; 1985:92-98.

Esseltine et al., 1988. Esseltine DW, Baxter MRN, Bevan JC: Sickle cell states and the anaesthetist.  Can J Anaesth  1988; 35:385.

Etchason et al., 1995. Etchason J, Petz L, Keeler E, et al: The cost effectiveness of preoperative blood donations.  N Engl J Med  1995; 332:719.

Fahmy, 1978. Fahmy NR: Nitroglycerine as a hypotensive drug during general anesthesia.  Anesthesiology  1978; 49:17.

Fahmy, 1985a. Fahmy NR: Nitroprusside vs a nitroprusside-trimethaphan mixture for induced hypotension: Hemodynamic effects and cyanide release.  Clin Pharmacol Ther  1985; 37:264.

Fahmy, 1985b. Fahmy NR: Techniques for deliberate hypotension: Haemodilution and hypotension.   In: Enderby GEH, ed. Hypotensive anaesthesia,  New York: Churchill-Livingstone; 1985:164-183.

Fahmy et al., 1989. Fahmy NR, Bottros MR, Charchafieh J, et al: A randomized comparison of labetalol and nitroprusside for induced hypotension.  J Clin Anesth  1989; 1:409.

Fahmy and Gavras, 1985. Fahmy NR, Gavras HP: Impact of captopril on hemodynamic and hormonal effects of nitroprusside.  J Cardiovasc Pharmacol  1985; 7:869.

Fahmy and Laver, 1976. Fahmy NR, Laver MB: Hemodynamic response to ganglionic blockade with pentolinium during N2O-halothane anesthesia in man.  Anesthesiology  1976; 44:6.

Fahmy et al., 1984. Fahmy NR, Mihelakos PT, Battit GE, et al: Propranolol prevents hemodynamic and humoral events after abrupt withdrawal of nitroprusside.  Clin Pharmacol Ther  1984; 36:470.

Fahmy and Soter, 1985. Fahmy NR, Soter NA: Effects of trimethaphan on arterial blood histamine and systemic hemodynamics in humans.  Anesthesiology  1985; 62:562.

Flamm et al., 1988. Flamm ES, Adams Jr HP, Beck DW, et al: Dose-escalation study of intravenous nicardipine in patients with aneurysmal subarachnoid hemorrhage.  J Neurosurg  1988; 68:393.

Fontana et al., 1995. Fontana JL, Welborn L, Mogan PD, et al: Oxygen consumption and cardiovascular function in children during profound intraoperative normovolemic hemodilution.  Anesth Analg  1995; 80:219.

Friesen et al., 1993. Friesen RH, Tornabene MA, Coleman SP: Blood conservation during pediatric cardiac surgery: Ultrafiltration of the extracorporeal circuit volume after cardiopulmonary bypass.  Anesth Analg  1993; 77:702.

Gammage, 1987. Gammage G: Crystalloid versus colloid: Is colloid worth the cost? In Kirby RR, Brown DL, editors: Anesthesia for trauma.  Int Anesthesiol Clin  1987; 25:37-60.

Gardner, 1946. Gardner WJ: The control of bleeding during operation by induced hypotension.  JAMA  1946; 132:572.

Gavras et al., 1975. Gavras H, Kremer D, Brown JJ, et al: Angiotensin and norepinephrine induced myocardial lesions: Experimental and clinical studies in rabbits and man.  Am Heart J  1975; 89:321.

Gelman and Ernst, 1978. Gelman S, Ernst EA: Hepatic circulation during sodium nitroprusside infusion in the dog.  Anesthesiology  1978; 49:182.

Gelman et al., 1984. Gelman S, Fowler K, Smith L: Regional blood flow during isoflurane and halothane anesthesia.  Anesth Analg  1984; 63:557.

Gerson et al., 1982. Gerson JI, Allen FB, Seltzer JL, et al: Arterial and venous dilation by nitroprusside and nitroglycerin: Is there a difference?.  Anesth Analg  1982; 61:256.

Giese, 1973. Giese J: Renin, angiotensin, hypertensive vascular damage: A review.  Am J Med  1973; 55:315-320.

Goldberg et al., 1990. Goldberg ME, McNulty SE, Azad SS, et al: A comparison of labetalol and nitroprusside for inducing hypotension during major surgery.  Anesth Analg  1990; 70:537.

Goodnough et al., 1989. Goodnough LT, Rudnick S, Price TH, et al: Increased preoperative collection of autologous blood with recombinant erythropoietin therapy.  N Engl J Med  1989; 321:1163.

Goodnough et al., 1997. Goodnough LT, Despotis GJ, Parvin CA: Erythropoietin therapy in patients undergoing cardiac operations.  Ann Thorac Surg  1997; 64:1579.

Goodnough et al., 1997. Goodnough LT, Monk TG, Brecher ME: Acute normovolemic hemodilution should replace the preoperative donation of autologous blood as a method of autologous-blood procurement.  Transfusion  1997; 38:396.

Goto and Fujita, 1980. Goto F, Fujita T: Prostaglandins and circulation.  Circ Control  1980; 1:139.

Goto et al., 1982. Goto F, Otani E, Dato S, et al: Prostaglandin E1 as a hypotensive drug during general anesthesia.  Anaesthesia  1982; 37:530.

Gould et al., 1986. Gould SA, Rosen AL, Sehgal LR, et al: Fluosol-DA as a red cell substitute in acute anemia.  N Engl J Med  1986; 314:1653.

Green, 1985. Green DW: Techniques for deliberate hypotension: Pharmacological blockade.   In: Enderby GEH, ed. Hypotensive anaesthesia,  New York: Churchill-Livingstone; 1985:109-133.

Greiss et al., 1976. Greiss L, Tremblay NAG, Davies DW: The toxicity of sodium nitroprusside.  Can Anaesth Soc J  1976; 23:480.

Griffiths and Gillies, 1948. Griffiths HWC, Gillies J: Thoracolumbar splanchnicectomy and sympathectomy: Anaesthetic procedure.  Anaesthesia  1948; 3:134.

Grosslight et al., 1985. Grosslight K, Forster R, Calahan AR, et al: Isoflurane for neuro-anesthesia. Risk factors for increases in intracranial pressure.  Anesthesiology  1985; 63:533.

Grundy et al., 1981. Grundy BL, Nash Jr CL, Brown RH: Arterial pressure manipulation alters spinal cord function during correction of scoliosis.  Anesthesiology  1981; 54:249.

Guay et al., 1992. Guay J, Reinberg C, Poitras B, et al: A trial of desmopressin to reduce blood loss in patients undergoing spinal fusion for idiopathic scoliosis.  Anesth Analg  1992; 75:405.

Hackmann et al., 1991. Hackmann T, Steward DJ, Sheps SB: Anemia in pediatric day-surgery patients: Prevalence and detection.  Anesthesiology  1991; 75:27.

Hainsworth, 1985. Hainsworth R: Arterial blood pressure.   In: Enderby GEH, ed. Hypotensive anaesthesia,  London: Churchill Livingstone; 1985:3-29.

Hallowell et al., 1978. Hallowell P, Bland JHL, Dalton BC, et al: The effect of hemodilution with albumen or Ringer's lactate on water balance and blood use in open-heart surgery.  Ann Thorac Surg  1978; 25:22.

Hampton and Little, 1953. Hampton LJ, Little DM: Complications associated with the use of “controlled hypotension” in anesthesia.  Arch Surg  1953; 67:549.

Harper and Glass, 1965. Harper AM, Glass HI: Effects of alterations in the arterial carbon dioxide tension on the blood flow through the cerebral cortex at normal and low arterial blood pressures.  J Neurol Neurosurg Psychiatry  1965; 28:449.

Havel et al., 1991. Havel M, Teufelsbauer H, Knöbl P, et al: Effect of intraoperative aprotinin administration on postoperative bleeding in patients undergoing cardiopulmonary bypass operation.  J Thorac Cardiovasc Surg  1991; 101:968.

Hayashi et al., 1988. Hayashi Y, Sumikawa K, Tashiro C, et al: Arrhythmogenic threshold of epinephrine during sevoflurane, enflurane, and isoflurane anesthesia in dogs.  Anesthesiology  1988; 69:145.

Hellewell and Potts, 1966. Hellewell J, Potts MW: Propranolol during controlled hypotension.  Br J Anaesth  1966; 38:794.

Henry et al., 1994. Henry CJ, Brewer Jr W, Henderson RA, et al: Pharmacokinetics and tolerance of weekly Oxygent CA infusions in the dog.  Artif Cells Blood Substit Immobil Biotechnol  1994; 22:1155.

Hines, 1990. Hines R: Preservation of platelet function during trimethaphan infusion.  Anesthesiology  1990; 72:834.

Hines and Barash, 1989. Hines R, Barash PG: Infusion of sodium nitroprusside induces platelet dysfunction.  Anesthesiology  1989; 70:611.

Hoffman and Lefkowitz, 1990. Hoffman BB, Lefkowitz RJ: Adrenergic receptor antagonists.   In: Goodman Gilman A, Rall TW, Nies AS, Taylor P, ed. Goodman and Gilman's the pharmacological basis of therapeutics,  8th ed.. New York: McGraw-Hill; 1990.

Holland et al., 1987. Holland BM, Jones JG, Wardrop CAJ: Lessons from the anemia of prematurity.  Hematol/Oncol Clin N Am  1987; 1:355.

Ikeda et al., 1987. Ikeda S, Schweiss JF, Frank PA, et al: In vitro cyanide release from sodium nitroprusside.  Anesthesiology  1987; 66:381.

Ivankovich et al., 1980. Ivankovich AD, Braverman B, Kanuru RP, et al: Cyanide antidotes and methods of their administration in dogs.  Anesthesiology  1980; 52:210.

Ivankovich et al., 1976. Ivankovich AD, Miletich DJ, Albrecht RF, et al: Sodium nitroprusside and cerebral blood flow in the anesthetized and unanesthetized goat.  Anesthesiology  1976; 44:21.

Jack, 1974. Jack RD: Toxicity of sodium nitroprusside.  Br J Anaesth  1974; 46:952.

Jacobs et al., 1982. Jacobs HK, Lieponis JV, Bunch WH, et al: The influence of halothane and nitroprusside on canine spinal cord hemodynamics.  Spine  1982; 7:35.

Janssens et al., 1994. Janssens M, Joris J, David JL, et al: High-dose aprotinin reduces blood loss in patients undergoing total hip replacement surgery.  Anesthesiology  1994; 80:23.

Johnston et al., 1976. Johnston RR, Eger II EI, Wilson C: A comparative interaction of epinephrine with enflurane, isoflurane, and halothane in man.  Anesth Analg  1976; 55:709.

Karl et al., 1983. Karl HW, Swedlow DB, Lee KW, et al: Epinephrine-halothane interactions in children.  Anesthesiology  1983; 58:142.

Karayalcin, 1979. Karayalcin G: Sickle cell anemia in the neonatal period.  South Med J  1979; 72:492.

Kelman et al., 1967. Kelman GR, Nunn JF, Prys-Roberts C, et al: The influence of cardiac output on arterial oxygenation.  Br J Anaesth  1967; 39:450.

Khambatta et al., 1981. Khambatta HJ, Stone JG, Kahn E: Propranolol alters renin release during nitroprusside induced hypotension and prevents hypertension on discontinuation of nitroprusside.  Anesth Analg  1981; 60:569.

Knight et al., 1983. Knight PR, Lane GA, Hensinger RN, et al: Catecholamine and renin-angiotensin response during hypotensive anesthesia induced by sodium nitroprusside or trimethaphan camsylate.  Anesthesiology  1983; 59:248.

Knight et al., 1980. Knight PR, Lane GA, Nicholls MG, et al: Hormonal and hemodynamic changes induced by pentolinium and propranolol during surgical correction of scoliosis.  Anesthesiology  1980; 53:127.

Kobrinsky et al., 1987. Kobrinsky NL, Letts RM, Patel LR, et al: 1-Deamino-8-D-arginine vasopressin (desmopressin) decreases operative blood loss in patients having Harrington rod spinal fusion surgery.  Ann Intern Med  1987; 107:446.

Kongsgaard et al., 1993. Kongsgaard UE, Hovig T, Brosstad F, et al: Platelets in shed mediastinal blood used for postoperative autotransfusion.  Acta Anaesthesiol Scand  1993; 37:265.

Kramer et al., 1979. Kramer AH, Hertzer NR, Beven EG: Intraoperative hemodilution during elective vascular reconstruction.  Surg Gynecol Obstet  1979; 149:831.

Kruszyna et al., 1987. Kruszyna H, Kruszyna R, Smith RP, et al: Red blood cells generate nitric oxide from directly acting, nitrogenous vasodilators.  Toxicol Appl Pharmacol  1987; 91:429.

Kulier et al., 1993. Kulier AH, Gombotz H, Fuchs G, et al: Subcutaneous recombinant human erythropoietin and autologous blood donation before coronary artery bypass surgery.  Anesth Analg  1993; 76:102.

Lagerkranser et al., 1984. Lagerkranser M, Irestedt L, Sollevi A, et al: Central and splanchnic hemodynamics in the dog during controlled hypotension with adenosine.  Anesthesiology  1984; 60:547.

Lam and Gelb, 1983. Lam AM, Gelb AW: Cardiovascular effects of isoflurane-induced hypotension for cerebral aneurysm surgery.  Anesth Analg  1983; 62:742.

Laver et al., 1975. Laver MB, Buckley MJ, Austen WG: Extreme haemodilution with profound hypothermia and circulatory arrest.  Bibl Haematol  1975; 41:225.

Lawson et al., 1974. Lawson NW, Ochsner JL, Mills NL, et al: The use of hemodilution and fresh autologous blood in open-heart surgery.  Anesth Analg  1974; 53:672.

Leese et al., 2000. Leese PT, Noveck RJ, Shorr JS, et al: Randomized safety studies of intravenous perflubron emulsion. I. Effects on coagulation function.  Anesth Analg  2000; 91:804.

Levin et al., 1980. Levin RM, Zadigian ME, Hall SC: The combined effect of hyperventilation and hypotension on cerebral oxygenation in anaesthetized dogs.  Can Anaesth Soc J  1980; 27:264.

Levy et al., 1993. Levy PS, Kim SJ, Eckel PK, et al: Limit to cardiac compensation during acute isovolemic hemodilution: Influence of coronary stenosis.  Am J Physiol  1993; 265:H340.

Lichtenthal et al., 1985. Lichtenthal PR, Rossi EC, Louis G, et al: Dose-related prolongation of the bleeding time by intravenous nitroglycerin.  Anesth Analg  1985; 64:30.

Lilleaason et al., 1978. Lilleaason P, Frøysaker T, Stokke O: Cardiac surgery in extreme haemodilution without donor blood, blood products or artificial macromolecules.  Scand J Thorac Cardiovasc Surg  1978; 12:249.

Longnecker, 1985. Longnecker DE: The microvascular response to hypotensive drugs.   In: Enderby GEH, ed. Hypotensive anaesthesia,  London: Churchill Livingstone; 1985:54-65.

Luce et al., 1982. Luce JM, Huseby JS, Kirk W, et al: Mechanisms by which positive end expiratory pressure increases cerebrospinal fluid pressure in dogs.  J Appl Physiol  1982; 52:231.

Lynn et al., 1986. Lynn AM, Fischer T, Brandford HG, et al: Systemic responses to tourniquet release in children.  Anesth Analg  1986; 65:865.

Madsen et al., 1987. Madsen JB, Cold GE, Hansen ES, et al: Cerebral blood flow and metabolism during isoflurane-induced hypotension in patients subjected to surgery for cerebral aneurysms.  Br J Anaesth  1987; 59:1204.

Magill et al., 1953. Magill IW, Scurr CF, Wyman JB: Controlled hypotension by a thiophanium derivative.  Lancet  1953; I:219.

Maroof et al., 1994. Maroof M, Khan RM, Bhatti TH: Clonidine premedication for induced hypotension with total intravenous anaesthesia for middle ear microsurgery.  Can J Anaesth  1994; 41:164.

Matot et al., 2002. Matot I, Scheinin O, Jurim O, Eid A: Effectiveness of acute normovolemic hemodilution to minimize allogeneic blood transfusion in major liver resection.  Anesthesiology  2002; 97:794.

Mazer et al., 1995. Mazer CD, Hornstein A, Freedman J: Platelet activation in warm and cold heart surgery.  Ann Thorac Surg  1995; 59:1481.

McDonald et al., 1969. McDonald FD, Thiel G, Wilson DR, et al: The prevention of acute renal failure in the rat by long-term saline loading: A possible role of the renin-angiotensin axis.  Proc Soc Exp Biol Med  1969; 131:610.

McNeil et al., 1974. McNeil TW, DeWald RL, Kuo KN, et al: Controlled hypotensive anesthesia in scoliosis surgery.  J Bone Joint Surg [Am]  1974; 56:1167.

Menkhaus et al., 1985. Menkhaus PG, Reves JG, Kissin I, et al: Cardiovascular effects of esmolol in anesthetized humans.  Anesth Analg  1985; 64:327.

Merin, 1981. Merin RG: Are the myocardial, functional, and metabolic effects of isoflurane really different from those of halothane and enflurane?.  Anesthesiology  1981; 55:398.

Merriam-Webster's collegiate dictionary, 2003. Merriam-Webster's collegiate dictionary,  11th ed.. Springfield, MA, Merriam-Webster, Inc., 2003.

Merrifield and Blundell, 1974. Merrifield AJ, Blundell MD: Toxicity of sodium nitroprusside [letter].  Br J Anaesth  1974; 46:324.

Messmer, 1975. Messmer K: Hemodilution.  Surg Clin N Am  1975; 55:659.

Messmer and Sunder-Plassmann, 1974. Messmer K, Sunder-Plassmann L: Hemodilution.  Prog Surg  1974; 13:208.

Michenfelder, 1977. Michenfelder JD: Cyanide release from sodium nitroprusside in the dog.  Anesthesiology  1977; 46:196.

Michenfelder and Tinker, 1977. Michenfelder JD, Tinker JH: Cyanide toxicity and thiosulfate protection during chronic administration of sodium nitroprusside in the dog: Correlation with a human case.  Anesthesiology  1977; 47:441.

Moncada et al., 1991. Moncada S, Palmer RM, Higgs EA: Nitric oxide: Physiology, pathophysiology, and pharmacology.  Pharmacol Rev  1991; 43:109.

Mongan and Hosking, 1992. Mongan PD, Hosking M: The role of desmopressin acetate in patients undergoing coronary artery bypass surgery.  Anesthesiology  1992; 77:38.

Moraca et al., 1962. Moraca PP, Elmars MB, Hale DE, et al: Clinical evaluation of sodium nitroprusside as a hypotensive agent.  Anesthesiology  1962; 23:193.

Mostello et al., 1991. Mostello LA, Casey WF, McGill WA: Does the use of a surgical tourniquet induce fever in infants?.  Anesth Analg  1991; 72:S191.

Motoyama, 1990. Motoyama EK: Respiratory physiology in infants and children.   In: Motoyama EK, Davis PJ, ed. Smith's anesthesia for infants and children,  5th ed.. St Louis: Mosby–Year Book; 1990.

Naik et al., 1991. Naik SK, Knight A, Elliott MJ: A successful modification of ultrafiltration for cardiopulmonary bypass in children.  Perfusion  1991; 6:41.

Nathan and Oski, 1987. Nathan DG, Oski FA: Hematology of infancy and childhood,  3rd ed.. Philadelphia, WB Saunders, 1987.

Navarro et al., 1994. Navarro R, Weiskopf RB, Moore MA, et al: Humans anesthetized with sevoflurane or isoflurane have similar arrhythmic response to epinephrine.  Anesthesiology  1994; 80:545.

Newberg and Michenfelder, 1983. Newberg LA, Michenfelder JD: Cerebral protection by isoflurane during hypoxemia or ischemia.  Anesthesiology  1983; 59:29.

Newberg et al., 1983. Newberg LA, Milde JH, Michenfelder JD: The cerebral metabolic effects of isoflurane at and above concentrations that suppress cortical electrical activity.  Anesthesiology  1983; 59:23.

Nimmagadda et al., 1991. Nimmagadda U, Joseph NJ, Salem MR, et al: Positive end-expiratory pressure increases intraocular pressure.  Crit Care Med  1991; 19:796.

Nunn, 1993. Nunn JF: Applied respiratory physiology,  4th ed.. Boston, Butterworth, 1993.

Oski, 1973. Oski FA: Designation of anemia on a functional basis.  J Pediatr  1973; 83:353.

Ottesen and Frøysaker, 1982. Ottesen S, Frøysaker T: Use of the haemonetic cell saver for autotransfusion in cardiovascular surgery.  Scand J Thorac Cardiovasc Surg  1982; 16:263.

Pagel et al., 1994. Pagel PS, Kersten JR, Hettrick DA, et al: Negative inotropic and lusitropic actions of sevoflurane in chronically instrumented dogs.  Anesth Analg  1994; 78:S332.

Palmer and Lassiter, 1975. Palmer RF, Lassiter KC: Sodium nitroprusside.  N Engl J Med  1975; 292:294.

Palmer et al., 1987. Palmer RM, Ferrige AG, Moncada S: Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor.  Nature  1987; 327:524.

Patel, 1981. Patel H: Experience with the cerebral function monitor during deliberate hypotension.  Br J Anaesth  1981; 53:639.

Perutz and Poyart, 1983. Perutz MF, Poyart C: Bezafibrate lowers oxygen affinity of haemoglobin.  Lancet  1983; 2:881.

Philbin et al., 1970. Philbin DM, Sullivan SF, Bowman FO, et al: Postoperative hypoxemia: Contribution of the cardiac output.  Anesthesiology  1970; 32:136.

Pineda and Valbonesi, 1990. Pineda A, Valbonesi M: Intraoperative blood salvage.  Clin Haematol  1990; 3:385.

Plewes and Farhi, 1985. Plewes JL, Farhi LE: Cardiovascular responses to hemodilution and controlled hypotension in the dog.  Anesthesiology  1985; 62:149.

Ponte, 1974. Ponte A: Postoperative paraplegia due to hypercorrection of scoliosis and drop of blood pressure.  J Bone Joint Surg [Am]  1974; 56:444.

Posner et al., 1976. Posner MA, Tobey RE, McElroy H: Hydroxocobalamin therapy of cyanide intoxication in guinea pigs.  Anesthesiology  1976; 44:157.

Poyart et al., 1994. Poyart C, Marden MC, Kister J: Bezafibrate derivatives as potent effectors of hemoglobin.  Methods Enzymol  1994; 232:496.

Prys-Roberts et al., 1974. Prys-Roberts C, Lloyd JW, Fisher A, et al: Deliberate profound hypotension induced with halothane. Studies of haemodynamics and pulmonary gas exchange.  Br J Anaesth  1974; 46:105.

Randad et al., 1991. Randad RS, Mahran MA, Mehanna AS, et al: Allosteric modifiers of hemoglobin.  J Med Chem  1991; 34:752.

Ray et al., 1986. Ray JM, Flynn JC, Bierman AH: Erythrocyte survival following intraoperative autotransfusion in spinal surgery. An in vivo comparative study and 5 year update.  Spine  1986; 11:879.

Rees et al., 1989a. Rees DD, Palmer RM, Hodson HF, et al: A specific inhibitor of nitric oxide formation from L-arginine attenuates endothelium-dependent relaxation.  Br J Pharmacol  1989; 96:418.

Rees et al., 1989b. Rees DD, Palmer RMJ, Moncada S: Role of endothelium-derived nitric oxide in regulation of blood pressure.  Proc Natl Acad Sci USA  1989; 86:3375.

Relton, 1975. Relton JES: Anesthesia in the original correction of scoliosis.   In: Riseborough EJ, Herndon JH, ed. Scoliosis and other deformities of the axial skeleton,  Boston: Little, Brown and Co; 1975:309-316.

Relton and Hall, 1967. Relton JES, Hall JE: An operation frame for spinal fusion. A new apparatus designed to reduce haemorrhage during operation.  J Bone Joint Surg [Br]  1967; 49:327.

Rich et al., 1994. Rich GF, Lowson SM, Johns RA, et al: Inhaled nitric oxide selectivity decreases pulmonary vascular resistance without impairing oxygenation during one-lung ventilation in patients undergoing cardiac surgery.  Anesthesiology  1994; 80:57.

Robinson, 1967. Robinson JS: Metabolism.  Int Anesthesiol Clin  1967; 5:467-480.Hypotension without hypoxia. In Stetson JB, editor::

Rogers et al., 1979. Rogers MC, Hamburger C, Owen K, et al: Intracranial pressure in the cat during nitroglycerin-induced hypotension.  Anesthesiology  1979; 51:227.

Rogers et al., 1980. Rogers MC, Nugent SK, Traystman RJ: Control of cerebral circulation in the neonate and infant.  Crit Care Med  1980; 8:570.

Rohrer and Natale, 1992. Rohrer MJ, Natale AM: Effect of hypothermia on the coagulation cascade.  Crit Care Med  1992; 20:1402.

Salem, 1975. Salem MR: Deliberate hypotension is a safe and accepted anesthetic technique.   In: Eckenhoff JE, ed. Controversy in anesthesiology,  Philadelphia: WB Saunders; 1975:95-104.

Salem, 1978. Salem MR: Nitroprusside and other short-acting hypotensive agents.  Int Anesthesiol Clin  1978; 16:171-200.Therapeutic uses of ganglionic blocking drugs. In Ivankovich AD, editor::

Salem, 1987. Salem MR: Hypercapnia, hypocapnia, and hypoxemia.  Semin Anesth  1987; 6:202.

Salem and Ivankovich, 1970. Salem MR, Ivankovich AD: The place of beta adrenergic blocking drugs in the deliberate induction of hypotension.  Anesth Analg  1970; 49:427.

Salem et al., 1971. Salem MR, Ivankovich AD, Shaker MH: Safety factors in deliberately induced hypotension.  Middle East J Anesthesiol  1971; 3:107.

Salem et al., 1970. Salem MR, Kim Y, Shaker MH: The effect of alteration of inspired oxygen concentration on jugular-bulb oxygen tension during deliberate hypotension.  Anesthesiology  1970; 33:358.

Salem and Podraza, 1992. Salem MR, Podraza AG: Blood conservation and massive transfusion.  Semin Anesth  1992; 11:339.

Salem et al., 1978. Salem MR, Toyama T, Wong AY, et al: Haemodynamic responses to induced arterial hypotension in children.  Br J Anaesth  1978; 50:489.

Salem et al., 1974. Salem MR, Wong AY, Bennett EJ, et al: Deliberate hypotension in infants and children.  Anesth Analg  1974; 53:975.

Schaller et al., 1983. Schaller Jr RT, Schaller J, Morgan A, et al: Hemodilution anesthesia: A valuable aid to major cancer surgery in children.  Am J Surg  1983; 146:79.

Schmied et al., 1998. Schmied H, Schiferer A, Sessler DI, Meznik C: The effects of red-cell scavenging, hemodilution, and active warming on allogenic blood requirements in patients undergoing hip or knee arthroplasty.  Anesth Analg  1998; 86:387.

Schwentker, 1978. Schwentker EP: Posterior fusion of the spine for scoliosis.  Surg Rounds  1978; 1:12.

Scott et al., 1978. Scott DB, Buckley FP, Littlewood DG, et al: Circulatory effects of labetalol during halothane anaesthesia.  Anaesthesia  1978; 33:145.

Shapiro and Marshall, 1978. Shapiro HM, Marshall LF: Intracranial pressure responses to PEEP in head-injured patients.  J Trauma  1978; 18:254.

Silvergleid, 1991. Silvergleid AJ: Preoperative autologous donation: What have we learned?.  Transfusion  1991; 31:99.

Simpson et al., 1993. Simpson MB, Georgopolulos G, Eilert RE: Intraoperative blood salvage in children and young adults undergoing spinal surgery with predeposited autologous blood: Efficacy and cost effectiveness.  J Pediatr Orthop  1993; 13:777.

Singler, 1989. Singler RC: Special techniques. Deliberate hypotension, hypothermia, and acute normovolemic hemodilution.   In: Gregory GA, ed. Pediatric anesthesia,  New York: Churchill Livingstone; 1989:553-577.

Sivarajan et al., 1980. Sivarajan M, Amory DW, Everett GB, et al: Blood pressure, not cardiac output, determines blood loss during induced hypotension.  Anesth Analg  1980; 59:203.

Smith and Nelson, 1976. Smith CA, Nelson NM: The physiology of the newborn infant,  Springfield, IL, Charles C Thomas, 1976.

Solem et al., 1987. Solem JO, Tengborn L, Steen S, et al: Cell saver versus hemofilter for concentration of oxygenator blood after cardiopulmonary bypass.  Thorac Cardiovasc Surg  1987; 35:42.

Sollevi et al., 1984. Sollevi A, Lagerkranser M, Irestedt L, et al: Controlled hypotension with adenosine in cerebral aneurysm surgery.  Anesthesiology  1984; 61:400.

Spahn and Casutt, 2000. Spahn DR, Casutt M: Eliminating blood transfusions.  Anesthesiology  2000; 93:242.

Sprung et al., 2002. Sprung J, Kindscher JD, Wahr JA, et al: The use of bovine hemoglobin glutamer-250 (Hemopure) in surgical patients: Results of a multicenter, randomized, single-blinded trial.  Anesth Analg  2002; 94:799.

Stehling, 1990. Stehling LC: Autologous transfusion.  Int Anesthesiol Clin  1990; 28:190.

Stehling and Zauder, 1991. Stehling LC, Zauder HL: Autologous blood salvage procedures.  Biotechnology  1991; 19:47.

Stockman et al., 1984. Stockman JA, Graeber JE, Clark DA, et al: Anemia of prematurity: Determinants of the erythropoietin response.  J Pediatr  1984; 105:786.

Stockman and Oski, 1978. Stockman JA, Oski FA: Physiological anemia of infancy and the anemia of prematurity.  Clin Haematol  1978; 7:3.

Stone et al., 1976. Stone JG, Khambatta HJ, Matteo RS: Pulmonary shunting during anesthesia with deliberate hypotension.  Anesthesiology  1976; 45:508.

Strandgaard, 1976. Strandgaard S: Autoregulation of cerebral blood flow in hypertensive patients.  Circulation  1976; 53:720.

Strunin, 1975. Strunin L: Organ perfusion during controlled hypotension.  Br J Anaesth  1975; 47:793.

Stullken and Sokoll, 1975. Stullken EH, Sokoll MD: Intracranial pressure changes during hypotension and subsequent vasopressor therapy in anesthetized cats.  Anesthesiology  1975; 42:425.

Sum et al., 1983. Sum CY, Yacobi A, Katzinel R, et al: Kinetics of esmolol, an ultra-short-acting beta blocker, and of its major metabolite.  Clin Pharmacol Ther  1983; 34:427.

Surgenor et al., 1990. Surgenor DM, Wallace EL, Hao SH, Chapman RH: Collection and transfusion of blood in the United States, 1982–1988.  N Eng J Med  1990; 322:1666-1668.

Szyfelbein and Ryan, 1974. Szyfelbein SK, Ryan JF: Use of controlled hypotension for primary surgical excision in an extensively burned child.  Anesthesiology  1974; 41:501.

Taylor, 1985. Taylor P: Ganglionic stimulating and blocking drugs.   In: Gilman AG, Goodman LS, Rall TW, Murad F, ed. The pharmacological basis of therapeutics,  7th ed.. New York: Macmillan Publishing; 1985:215.

Thomas et al., 1996. Thomas MJ, Gillon J, Desmond MJ: Consensus conference on autologous transfusion: Preoperative autologous donation.  Transfusion  1996; 36:633.

Thompson et al., 1978. Thompson GE, Miller RD, Stevens WC, et al: Hypotensive anesthesia for total hip arthroplasty: A study of blood loss and organ function (brain, heart, liver, and kidney).  Anesthesiology  1978; 48:91.

Tinker and Michenfelder, 1976. Tinker JH, Michenfelder JD: Sodium nitroprusside: Pharmacology, toxicology and therapeutics.  Anesthesiology  1976; 45:340.

Toivonen et al., 1992. Toivonen J, Virtanen H, Kaukinen S: Labetalol attenuates the negative effects of deliberate hypotension induced by isoflurane.  Acta Anesthesiol Scand  1992; 36:84.

Tremper et al., 1982. Tremper KK, Freedman AE, Levine EM: The preoperative treatment of severely anemic patients with perfluorochemical emulsion oxygen transporting fluid, Fluosol-DA.  N Engl J Med  1982; 307:277.

Tretiak et al., 1996. Tretiak R, Laupacis A, Rivière M, et al: Cost of allogenic and autologous blood transfusion in Canada.  Can Med Assoc J  1996; 154:1501.

Tuman et al., 1987. Tuman KJ, Spiess BD, McCarthy RJ, Ivankovich AD: Effect of progressive blood loss and coagulation as measured by thromboelastography.  Anesth Analg  1987; 66:856-963.

Tuman et al., 1988. Tuman KJ, Spiess BD, McCarthy RJ, et al: Effects of continuous arteriovenous hemofiltration on cardiopulmonary abnormalities during anesthesia for orthotopic liver transplantation.  Anesth Analg  1988; 67:363.

Tuncali et al., 2003. Tuncali B, Karci A, Kadir A, et al: Controlled hypotension and minimal inflation pressure: A new approach for pneumatid tourniquet application in upper limb surgery.  Anesth Analg  2003; 97:1529.

Turner et al., 1977. Turner JM, Powell D, Gibson RM, et al: Intracranial pressure changes in neurosurgical patients during hypotension induced with sodium nitroprusside or trimetaphan.  Br J Anaesth  1977; 49:419.

Valeri et al., 1992. Valeri CR, Khabbaz K, Khuri SF: Effect of skin temperature on platelet function in patients undergoing extracorporeal bypass.  J Thorac Cardiovasc Surg  1992; 104:108.

Van Aken and Cottrell, 1986. Van Aken H, Cottrell JE: Hypotensive anesthesia and its effect on the cardiovascular system.   In: Altura BM, Halevey S, ed. Cardiovascular actions of anesthetics and drugs used in anesthesia, 2. 5th ed.. Basel: Karger; 1986:260-279.

Van Aken and Van Hemelrijck, 1993. Van Aken H, Van Hemelrijck J: Deliberate hypotension,  Cleveland, Ohio, Review Course Lectures, International Anesthesia Research Society, 1993.

Van Aken et al., 1982. Van Aken H, Puchstein C, Schweppe ML, et al: Effect of labetalol on intracranial pressure in dogs with and without intracranial hypertension.  Acta Anaesth Scand  1982; 26:615.

Van Aken et al., 1984. Van Aken H, Puchstein C, Anger C, et al: Changes in intracranial pressure and compliance during adenosine triphosphate-induced hypotension in dogs.  Anesth Analg  1984; 63:381.

Van Oeveren et al., 1987. Van Oeveren W, Jansen NJG, Bidstrup BP, et al: Effects of aprotinin on haemostatic mechanisms during cardiopulmonary bypass.  Ann Thorac Surg  1987; 44:640.

Verner, 1985. Verner IR: Techniques for deliberate hypotension: Direct acting vasodilators.   In: Enderby GEH, ed. Hypotensive anaesthesia,  London: Churchill Livingstone; 1985:138-163.

Vesey et al., 1975. Vesey CJ, Cole PV, Simpson PJ: Sodium nitroprusside in anaesthesia.  Br Med J  1975; 3:229.

Viviani et al., 1978. Viviani GR, Sadler JTS, Ingham GK: Autotransfusions in scoliosis surgery.  Clin Orthop Related Res  1978; 135:74.

Wahr, 2003. Wahr JL: Clinical potential of blood substitutes or oxygen therapeutics during cardiac surgery.  Anesthesiol Clin N Am  2003; 21:553.

Wahr et al., 2001. Wahr JL, Gerber M, Venitz J, et al: Allosteric modification of oxygen delivery by hemoglobin.  Anesth Analg  2001; 92:615.

Warren, 1974. Warren JV: The physiology of the giraffe.  Sci Am  1974; 231:96.

Watzek et al., 1980. Watzek G, Watzek C, Draxler V, et al: Experience with ‘isovolaemic’ haemodilution in extensive surgery for oro-facial tumours.  J Oral Maxillofac Surg  1980; 8:131.

Welborn et al., 1991. Welborn LG, Hannallah RS, Luban NLC, et al: Anemia and postoperative apnea in former preterm infants.  Anesthesiology  1991; 74:1003.

Widman, 1991. Widman FK: Standards for blood banks and transfusion services,  14th ed.. Arlington, VA, American Association of Blood Banks, 1991.

Wildsmith et al., 1983. Wildsmith JA, Sinclair CJ, Thorn J, et al: Haemodynamic effects of induced hypotension with a nitroprusside-trimethaphan mixture.  Br J Anaesth  1983; 55:381.

Williamson and Taswell, 1991. Williamson KR, Taswell HF: Intraoperative blood salvage.   In: Taswell HF, Pineda AA, ed. Autologous transfusion and hemotherapy,  Boston: Blackwell Scientific; 1991:122-154.

Wilson et al., 1988. Wilson AJ, Cuddigan BJ, Wyatt AP: Early experience of intraoperative autotransfusion.  J R Soc Med  1988; 81:389.

Wong et al., 1980. Wong KC, Webster LR, Coleman SS, et al: Hemodilution and induced hypotension for insertion of a Harrington rod in a Jehovah's Witness patient.  Clin Orthop Relat Res  1980; 152:237.

Woodcock et al., 1988. Woodcock TE, Millard RK, Dixon J, et al: Clonidine premedication for isoflurane-induced hypotension.  Br J Anaesth  1988; 60:388.

Yaster et al., 1986. Yaster M, Simmons RJ, Tolo VT, et al: A comparison of nitroglycerin and nitroprusside for inducing hypotension in children: A double-blind study.  Anesthesiology  1986; 65:175.

Yau et al., 1992. Yau TM, Carson S, Weisel RD, et al: The effect of warm heart surgery on postoperative bleeding.  J Thorac Cardiovasc Surg  1992; 103:1155.

Yawn, 1991. Yawn DH: Properties of salvaged blood.   In: Taswell HF, Pineda AA, ed. Autologous transfusion and hemotherapy,  Boston: Blackwell Scientific; 1991:194-206.

Zauder, 1991. Zauder HL: Intraoperative and postoperative blood salvage devices.   In: Stehling LC, ed. Perioperative autologous transfusion,  Arlington, VA: American Association of Blood Banks; 1991:25-36.