Potassium is the primary intracellular solute, with approximately one third of cellular energy metabolism devoted to Na⁺/K⁺ exchange. Sodium continuously leaks into cells along its concentration gradient, yet it is rapidly extruded in exchange for potassium. As the cell is exposed to varying osmolarity, water movement occurs, causing cell swelling or shrinkage. Because stable cell volume is critical for survival, complex regulatory mechanisms have evolved to ensure that stability is maintained.^11,¹² The processes by which swollen cells return to normal size are collectively termed regulatory volume decrease processes, and those returning a shrunken cell to normal are termed regulatory volume increase processes (Fig. 8-1). With sudden, brief changes in osmolality, regulatory volume increase or decrease processes are activated after small (1% to 2%) changes in cell volume, returning cell volume to normal primarily through transport of electrolytes. If anisosmotic conditions persist, chronic compensation occurs through the accumulation or loss of small organic molecules termed osmolytes or idiogenic osmoles (e.g., polyols, sorbitol, myoinositol), amino acids and their derivatives (e.g., taurine, alanine, proline), and methylamines (e.g., betaine, glycerylphosphorylcholine).

FIGURE 8-1 Activation of mechanisms regulating cell volume in response to volume perturbations. Volume-regulatory losses and gains of solutes are termed regulatory volume decrease (A) and regulatory volume increase (B), respectively. The course of these decreases and increases varies with the type of cell and experimental conditions. Typically, however, a regulatory volume increase mediated by the uptake of electrolytes or a regulatory volume decrease mediated by the loss of electrolytes and organic osmolytes occurs over a period of minutes. When cells that have undergone a regulatory volume decrease (A) or increase (B) are returned to normotonic conditions, they swell above or shrink below their resting volume. This is due to volume-regulatory accumulation or loss of solutes, which effectively makes the cytoplasm hypertonic or hypotonic, respectively, as compared with normotonic extracellular fluid.

(From McManus ML, Churchwell KB, Strange K. Regulation of cell volume in health and disease. N Engl J Med 1995;333:1260-6. ®Massachusetts Medical Society.)

Like intracellular volume, circulating blood (intravascular) volume is also tightly controlled. Increases in intravascular volume result from increases in sodium and water retention, whereas decreases in intravascular volume result from increases in excretion of sodium and water. As noted earlier, serum osmolality must be maintained within a very narrow range if serum sodium is to be an effective focus of intravascular volume control. Serum osmolality is usually maintained between 280 and 300 mOsm/L. Changes in osmolality as small as 1% trigger compensatory mechanisms.

Serum osmolality is primarily regulated by antidiuretic hormone (ADH), thirst, and renal concentrating ability. Because the indirect aim of osmolar control is actually volume control, these same osmoregulatory mechanisms are also influenced by factors such as blood pressure (BP), cardiac output, and vascular capacitance.^13,¹⁴ In pathologic conditions such as ascites or hemorrhage, intravascular volume preservation takes precedence over osmolality and osmoregulatory mechanisms operate to restore intravascular volume, even at the expense of disrupting physiologic solute balance.

For example, ADH is released from neurons of the supraoptic and paraventricular nuclei in response to osmolar fluctuations in cell size. Solutes that readily permeate cell membranes, such as urea, increase the serum osmolality without triggering the release of ADH. Infusion of solutes with large actual or effective reflection coefficients (σ) at the cell membrane (e.g., sodium, mannitol) elicits a robust ADH release. ADH release begins when the serum osmolality reaches a threshold of approximately 280 mOsm/L. Rapid increases in osmolality lead to a greater release of ADH than do slow increases. Hypovolemia and hypotension diminish the threshold for ADH release and increase the “gain” of the system by exaggerating the rate of increase in serum ADH concentrations (E-Fig. 8-2). Thus, in a volume-depleted or hypotensive child, brisk ADH release occurs in response to plasma osmolalities as low as 260 to 270 mOsm/L. It has been hypothesized that different populations of vasopressin-secreting cells are responsive to osmotic and baroreceptor-mediated input.

E-FIGURE 8-2 A, Relationship between plasma antidiuretic hormone (ADH) concentration and plasma osmolality in normal humans in whom the plasma osmolality was changed by varying the state of hydration. Notice that the osmotic threshold for thirst is a few milliosmols per kilogram higher than that for ADH. B, The influence of hemodynamic status on the osmoregulation of ADH in otherwise healthy humans. The numbers in the center circles refer to the percentage change in volume or pressure; N (in the center circles) refers to the normovolemic normotensive subject. Notice that the hemodynamic status affects both the slope of the relationship between the plasma ADH and osmolality and the osmotic threshold for ADH release.

(A adapted from Robertson GL, Aycinena P, Zerbe RL. Neurogenic disorders of osmoregulation. Am J Med 1982;72:339; B from Rose DB. Clinical physiology of acid-base and electrolyte disorders. 4th ed. New York: McGraw-Hill; 1994. p. 159, 163.)

Intravascular fluid volume, salt and water intake, electrolyte balance, and cardiovascular status are interrelated at several levels.¹⁵ For example, as the veins and arteries become replete with fluid and the systemic BP increases, ADH release wanes as both pressure diuresis and natriuresis commence.¹⁶ The resulting relationship between urine output and arterial pressure is termed the renal function curve, and its intersection with salt and water intake determines the equilibrium point at which arterial BP ultimately stabilizes (Fig. 8-2). Equilibrium (chronic) BP is influenced only by shifts of the renal function or fluid intake curves. Transient changes in arterial pressure secondary to peripheral vascular resistance changes are always resolved by opposing shifts in total body salt and TBW.

FIGURE 8-2 Analysis of arterial pressure regulation by equating the renal output curve with the salt and water intake curve. The equilibrium point describes the level to which the arterial pressure will be regulated. (That portion of the salt and water intake that is lost from the body through nonrenal routes is ignored in this figure.)

(From Guyton AC, Hall JC, editors. Textbook of medical physiology. Philadelphia: WB Saunders; 1996. p. 221-37.)

In response to a decreasing arterial pressure, the renin-angiotensin system is also mobilized. With decreased renal perfusion, juxtaglomerular cells release renin, which in turn converts renin substrate (angiotensinogen) to angiotensin I. Angiotensin I is then rapidly converted to angiotensin II by angiotensin-converting enzyme present in lung endothelium. Angiotensin II supports arterial pressure in three ways: (1) direct vasoconstriction, (2) increased salt and water retention (via renal vasoconstriction and decreased glomerular filtration), and (3) stimulation of aldosterone secretion (Fig. 8-3).

FIGURE 8-3 Physiologic responses to hypotension.

ADH, pressure diuresis, and the renin-angiotensin system permit wide ranges in salt and water intake without large fluctuations in BP or volume status. All serve to support the systemic circulation when threatened and complement the more immediate activity of the sympathetic nervous system. In addition to high-pressure sensors such as aortic arch and carotid sinus baroreceptors, intravascular volume information is provided by low-pressure thoracic sensors. For this reason, effective increases or decreases in intrathoracic blood volume may mimic changes in whole-body volume status and produce natriuresis, diuresis, or fluid retention. Intravascular volume may also be sensed as the stretch of atrial muscle fibers, leading to release of atrial natriuretic peptide.¹⁷ Although its complete physiologic role is uncertain, atrial natriuretic peptide may serve to “fine tune” volume status by causing modest vasodilation, gently increasing the glomerular filtration rate (GFR), and decreasing reabsorption of sodium. The combination of complex autoregulatory mechanisms with complementary actions operating on varying time scales, all responding to different, yet interrelated, effector stimuli, yields an elegant system by which the mature individual may maintain circulation amid a variety of challenges. In this context, it is interesting to observe that successful heart transplant recipients, despite general cardiovascular stability, typically manifest fundamental derangements in body fluid homeostasis.¹⁸

Maturation of Fluid Compartments and Homeostatic Mechanisms

Body Water and Electrolyte Distribution

Much of our understanding of the development of body water compartments is derived from deuterium oxide dilution studies performed in the 1950s.¹⁹ In a series of 21 neonates, TBW was found to be approximately 78 ± 5% of body weight. Subsequent measurements in fewer subjects showed that TBW decreased to approximately 60% in the second 6 months of life with most of the loss being extracellular. A smaller decrease (to about 57%) is observed late in childhood (Fig. 8-4).

FIGURE 8-4 Total body water (blue circles), extracellular water (purple triangles), and intracellular water (orange circles) as percentages of body weight in infants and children, compared with corresponding values for the fetus and adults.

(From Friis-Hansen B. Body water compartments in children: changes during growth and related changes in body composition. Pediatrics 1961;28:169-81.)

The importance of the extracellular compartment, its relationship to the intracellular space, and much of the chemical anatomy of both were first described by Gamble in educational monographs issued during the first part of the 20th century (E-Fig. 8-3).^20,²¹ The chemical compositions of mature body fluid compartments are provided in Table 8-1.

TABLE 8-1 Composition of Body Fluid Compartments

	Extracellular Fluid	Intracellular Fluid
Osmolality (mOsm)	290-310	290-310
Cations (mEq/L)	155	155
Na⁺	138-142	10
K⁺	4.0-4.5	110
Ca²⁺	4.5-5.0	—
Mg²⁺	3	40
Anions (mEq/L)	155	155
Cl⁻	103	—
HCO₃⁻	27	—
HPO₄²⁻	—	10
SO₄²⁻	—	110
PO₄²⁻	3	—
Organic acids	6	—
Protein	16	40

E-FIGURE 8-3 Acid–base composition of body fluids. These diagrams are constructed from average values for the individual factors expressed in terms of acid–base equivalence—that is, as cubic centimeters of one-tenth normal solutions per hundred cubic centimeters of fluid. The base factors are superimposed in the left column and the acid factors are superimposed in the right column for each fluid. The diagrams represent, as is actually the case, a structure composed not of salt but of individually sustained concentrations of ions. The exact acid–base equivalence indicated by the equal height of the two columns is obtained by adjustability of the bicarbonate ion concentration (HCO₃⁻) to any change elsewhere in the structure.

Circulating Blood Volume

The blood volume in neonates was determined to be 82 ± 9 mL/kg using an iodine 121–labeled human serum albumin technique, although substantial variability may result from the degree of placental-fetal transfusion.²² In low-birth-weight (LBW), preterm, or critically ill infants, values as high as 100 mL/kg have been measured.²³ Blood volume increases slightly during the first few months of life, reaching its zenith at 2 months of age (approximately 86 mL/kg), then returns to near 80 mL/kg and finally stabilizes at 70 mL/kg by the end of the first year of life. In general, the ratio of blood volume to weight decreases with growth. The most accurate basis for prediction of blood volume is lean body mass, the consideration of which removes any male/female variation even into adulthood.²⁴ An estimate of the circulating blood volume is presented in Table 8-2.

TABLE 8-2 Estimate of Circulating Blood Volume

Age	Estimated Blood Volume (mL/kg)
Preterm infant	100
Full-term neonate	90
Infant	80
School age	75
Adults	70

Maturation of Homeostatic Mechanisms

Renal development begins at approximately 5 weeks of gestation and continues in a centrifugal pattern until the full complement of nephrons is in place by about the 38th week. In the outermost regions of the renal cortex, postnatal nephron differentiation may continue for several weeks to months. In the early stages of gestation, renal blood flow is approximately one fifth of normal. Initially this is related to structural immaturity, and later it is due to increased renovascular resistance. By 38 weeks of gestation, renal blood flow is approximately one third of normal. High renovascular resistance protects the developing nephron from both pressure and volume overload. The resulting renal contribution to metabolic homeostasis in utero is limited.

As with the pulmonary bed, vascular resistance in the kidney decreases after birth, leading to abrupt increases in renal blood flow and GFR. In utero, despite a low GFR, urine output is brisk, owing to poor reabsorption of salt and water. Plasma renin activity is increased in utero, decreases immediately after birth, and then increases again as excess extracellular water is mobilized and excreted. Aldosterone levels are increased in cord blood and are maintained at this level for the first 3 days of life. The increased aldosterone may be necessary for sodium retention during periods of increased anabolism early in life.

Intrarenal gradients of NaCl and urea are less steep in the immature kidney, and full nephron length has yet to be achieved. Consequently, urine concentrating ability is limited in neonates, with maximum urine osmolality being about half that of the adult (700 to 800 mEq/L versus 1300 to 1400 mEq/L). In part, this also relates to low circulating ADH levels and decreased renal responsiveness to ADH. Although overall ADH production is not impaired, excessive secretion may occur in some disease states. Limited urine-concentrating ability necessitates large urine volumes for elimination of large solute loads.

In the first year of life, renal plasma flow and GFR are approximately one half the adult values of 350 and 70 mL/min/m², respectively.²⁵ Consequently, serum creatinine is increased in term and preterm infants, yet normalizes in the second month of life. Fractional excretion of sodium (FE_Na) is markedly increased in preterm infants, decreases somewhat by term, and stabilizes at adult levels by the second month of life. Although the adult kidney may easily achieve FE_Na values as low as 0.5%, the 34-week-gestation infant is limited to no less than 2%.

These maturational features make it very difficult for the preterm or young infant to handle fluctuations in fluid and solute loads. Both sodium conservation and regulation of extracellular fluid volume are impaired relative to the older child and adult. Limited GFR makes excretion of a fluid challenge difficult. Excessive urinary sodium loss leads to increased maintenance requirements. Hyponatremia is common. Conversely, diminished concentrating ability increases free water losses during excretion of a solute load, whereas the high ratio of surface area to volume produces increased evaporative water loss. Consequently, fluid requirements are relatively high, and dehydration is common. Any errors in the fluid management are poorly tolerated. As a rule, the most severe impairment exists in preterm infants, and the majority of homeostatic mechanisms are fully developed after the first year of life.

Fluid and Electrolyte Requirements

Holliday summarized the evolution of contemporary hydration therapy.²⁰ In 1831, Latta first reported the use of intravenous (IV) fluids in the resuscitation of patients dehydrated by cholera.²⁶ In 1918, growing information on the subject permitted Blackfan and Maxcy to successfully treat nine infants by intraperitoneal injection.²⁷ In 1923, Gamble and associates detailed the anatomy of fluid and electrolyte compartments, introducing the use of milliequivalents to clinical practice.²¹ This paved the way for the development of the “deficit therapy” regimen of Darrow.²⁸

In subsequent decades, various recipes to replace the extracellular and intracellular fluid losses were suggested. For the most part, these failed because of excessive potassium and insufficient sodium content. Hyponatremia was common. When the focus of the treatment shifted to replacing the extracellular fluid deficit, rapid restoration of extracellular fluid volume using solutions with sodium concentrations similar to those in blood became commonplace. This, along with oral rehydration, is the preferred method of treatment today.

The concept of “maintenance fluids” is a complex subject. Although water and salt are required to sustain life, it is fair to say that for an individual child at any particular time, the precise amounts necessary are unknown (and perhaps unknowable). Instead, fluids and electrolytes, like anesthetics, are titrated to effect with general guidelines provided by clinical assessment, basic physiologic principles, and limited published data. The term maintenance fluids is often more limiting than helpful, and in all cases it is less precise than other terms familiar to anesthesiologists, such as minimal alveolar concentration (MAC) or median effective dose (ED₅₀).

Holliday and Segar provided calculations for a first approximation of “the maintenance need for water in parenteral fluid therapy” in 1957.²⁹ Integrating the relevant known physiology at that time, these authors observed that “insensible loss of water and urinary water loss roughly parallel energy metabolism and do not parallel weight.” However, because water utilization parallels energy metabolism, energy metabolism follows surface area, and surface area follows weight, it should be possible to estimate water requirement from weight alone. The authors then proceeded under a series of assumptions to extrapolate from limited data to a “relationship between weight and energy expenditure that might easily be remembered.”

Assuming energy requirements of “hospitalized patients” to be “roughly midway between basal and normal levels,” they constructed a curve of caloric requirement versus weight.²⁹ This curve could be seen as comprising three linear sections: 0 to 10 kg, 10 to 20 kg, and 20 to 70 kg (Fig. 8-5). Viewing the curve in this manner, the authors reasoned that “fortuitously, the average need for water, expressed in milliliters, equals energy expenditure in calories”: 100 mL/kg/day for weights to 10 kg, an additional 50 mL/kg/day for each kilogram from 11 to 20 kg, and 20 mL/kg/day more for each kilogram beyond 20 kg. In anesthetic practice, this formula has been further simplified, with the hourly requirement referred to as the “4-2-1 rule” (4 mL/kg/hr for the first 10 kg of weight, 2 mL/kg/hr for the next 10 kg, and 1 mL/kg/hr for each kilogram thereafter; Table 8-3).

FIGURE 8-5 The upper and lower curves were plotted from data from the study by Talbot.¹²⁶ Weights at the 50th percentile level were selected for converting calories at various ages to calories related to weight. The computed line for the average hospitalized child was derived from the following equations:

1. 0-10 kg: 100 kcal/kg.

2. 10-20 kg: 1000 kcal + 50 kcal/kg for each kg over 10 kg

3. 20 kg and up: 1500 kcal + 20 kcal/kg for each kg over 20 kg

(From Holliday MA, Segar WE. The maintenance need for water in parenteral fluid therapy. Pediatrics 1957;19:823-32.)

TABLE 8-3 Relationship between Weight and Hourly or Daily Maintenance Fluid Requirements of Children as per the 4-2-1 Rule

	MAINTENANCE FLUID REQUIREMENTS
Weight (kg)	Hour	Day
<10	4 mL/kg	100 mL/kg
10-20	40 mL + 2 mL/kg for every kg >10 kg	1000 mL + 50 mL/kg for every kg >10 kg
>20	60 mL + 1 mL/kg for every kg >20 kg	1500 mL + 20 mL/kg for every kg >20 kg

For decades, the simplicity and elegance of the Holliday and Segar formula has made it the starting point for fluid management in healthy children. As recently as 2006, the majority of consultant anesthetists in the United Kingdom administered hyponatremic glucose-containing solutions intraoperatively and postoperatively to children undergoing elective surgery.³⁰ However, the uncritical use of these solutions was never intended, and their blind application in the operating room, or in any clinical situation, is unwarranted. Further, as Holliday and Segar have themselves pointed out, their original approach involved hyponatremic glucose-containing solutions rather than near-isotonic solutions such as 0.9% saline or lactated Ringer solution (LR).^30a In addition, these requirements were assessed at a basal metabolic state and not when the child was acutely ill or under physiologic stress. As the authors cautioned that “understanding of the limitations and of exceptions to the system [is] required. Even more essential is the clinical judgment to modify the system as circumstances dictate.” General water losses for infants and children are summarized in Table 8-4.

TABLE 8-4 Normal Water Losses for Infants and Children

Cause of Loss	Volume of Loss (mL/100 kcal)
Output
Urine	70
Insensible loss
Skin	30
Respiratory tract	15
Hidden intake (from burning 100 calories)	15
Total	100

More recently, Holliday revised the approach to fluid therapy in children that he and Segar enshrined with several caveats.^30b In a related commentary, Holliday pointed out several problems with applying the original 4-2-1 rule to acutely ill children.^30c Namely, dysregulation of ADH is a hallmark of critical illness because ADH secretion is affected by a variety of nonosmotic factors such as pain, stress, mechanical ventilation, and many medications. As a result, the choice of intravenous fluid and the rapidity of deficit replacement must be approached with care. The authors recommended a relatively simple strategy for healthy children undergoing elective surgery (including outpatients) to turn off ADH secretion and prevent perioperative water retention and subsequent hyponatremia. When a child (who is without significant heart or kidney disease) presents with marginal to moderate hypovolemia (e.g., after fasting for surgery), 20 to 40 mL/kg of isotonic fluids should be given during surgery and the postanesthesia care unit stay (as rapidly as 10 to 20 mL/kg/hr). Clinical judgment must always allow for modification of these recommendations if indicated for an individual child.^31,³² If hypovolemia is more severe (e.g., after an extensive bowel preparation), 40 to 80 mL/kg may be necessary during the perioperative period.

Postrecovery IV fluid therapy should consist of an isotonic solution infused at half the rate described in the original 4-2-1 fluid regimen (i.e., 2 mL/kg for the first 10 kg, 1 mL/kg for the next 10 kg, and 0.5 mL/kg for each additional kilogram thereafter). If the child does not or cannot tolerate oral intake after 6 to 12 hours, standard maintenance fluid therapy using hypotonic saline (e.g., 0.45% saline) should be initiated to avoid hypernatremia and fluid overload from prolonged administration of the isotonic solutions. This regimen should limit the ADH response and reduce the risk for postoperative hyponatremia and hypernatremia.³¹^–³³ It remains essential, however, to monitor plasma electrolyte concentrations serially on the ward, regardless of the fluid doctrine followed, to be certain that electrolyte concentrations remain within normal limits.^34,³⁵

Neonatal Fluid Management

In the first few days of life, isotonic losses of salt and water cause the normal neonate to lose 5% to 15% of body weight. Although GFR rises rapidly, urine output is initially low, and renal losses are modest. Day 1 fluid requirements of the wrapped neonate, therefore, are relatively low. Over the next few days of life, losses and requirements increase. In the poorly feeding infant, progression to hypernatremia and dehydration are common. When intake is appropriate, the term infant will regain body weight during the first week of life.

Three distinct phases of fluid and electrolyte homeostasis have been described in LBW ³⁶ and very low–birth-weight (VLBW)³⁷ infants. In the first day of life, there is minimal urine output, and body weight is stable despite low fluid intake. In the second phase, days 2 and 3 of life, diuresis occurs irrespective of the amount of fluid administered. By the fourth and fifth days of life, urine output begins to vary with changes in fluid intake and state of health.

Prematurity increases neonatal fluid requirements significantly. Fluid requirements are therefore estimated and then titrated to the infant’s changing weight, urine output, and serum sodium concentration.

No less important is glucose homeostasis. In the ninth month of gestation, the fetus begins to form glycogen stores at a rate of more than 100 kcal/day. In the unstressed, term infant, hepatic glycogen stores are 5% of body weight. Immediately after birth, glycogenolysis depletes most of these stores within the first 24 to 48 hours. Gluconeogenesis must then proceed to yield glucose at a rate of approximately 4 mg/kg/min.

At birth, fetal serum glucose is 60% to 70% of maternal concentrations. This may decrease within the first hours of life before recovering but should exceed 45 mg/dL to avoid neurologic injury. Symptoms of hypoglycemia may include jitteriness, lethargy, temperature instability, and convulsions. Ten percent dextrose in water (D₁₀W) may be given as a bolus of 2 to 4 mL/kg followed by a continuous infusion (through a pump) providing 4 to 6 mg/kg/min. The serum glucose concentration is then analyzed frequently and the infusion adjusted as necessary to prevent hypoglycemia and hyperglycemia. It is important that the amount of glucose being provided be calculated in milligrams per kilogram per minute to avoid errors during fluid changes and to facilitate the diagnosis of persistent hypoglycemia.

Typical day 1 infant fluid orders call for 70 to 80 mL/kg of D₁₀W. Because D₁₀W contains 10 g of glucose per deciliter, this dosage provides

On day 2, fluids are routinely increased to at least 100 mL/kg/day, and sodium is added at 2 to 3 mEq/dL. After urine output is established, potassium is added at 1 to 2 mEq/dL. The final solution, containing 30 mEq Na⁺ and 10 to 20 mEq K⁺ per liter, approximates the 0.2% saline “maintenance” solution commonly used previously in older children.

In the neonatal ICU, fluid management focuses on provision of adequate nutrition, maintenance of electrolyte balance, and limitation of fluid overload. The last factor is of particular concern because plasma oncotic pressure is reduced in preterm infants and the whole-body protein reflection coefficient is less than that in adults.³⁸ VLBW infants are at particular risk for fluid and electrolyte imbalances.³⁹ Even modest fluid overload may exacerbate pulmonary edema, prolong ductal patency, and more readily produce congestive heart failure. This perspective typically accompanies the infant to the operating room, where the primary considerations are routinely quite the opposite: restoration of circulating blood volume after third space accumulation, maintenance of intravascular volume amid ongoing blood loss, replacement of potentially massive evaporative losses, and maintenance of BP despite anesthetic-induced vasodilatation and increased venous capacitance. During surgery, these concerns must take precedence; yet unnecessary administration of fluid is best avoided.

Intraoperative Fluid Management

Intravenous Access and Fluid Administration Devices

In pediatrics, the first step toward intraoperative fluid management is often the most challenging: that is, gaining IV access. In general, simple procedures in healthy children are successfully approached using a single peripheral IV line. Although preferences vary among anesthesiologists, establishing IV access is most easily accomplished after induction of anesthesia. In young children, anesthesia is often induced by inhalation, and a catheter is inserted by an assistant into a hand or foot vein. In older children, or when IV access is desirable before anesthesia is induced, IV access may be facilitated by the use of topical anesthesia (e.g., EMLA cream, amethocaine, lidocaine infiltration) or sedation or both.

Complex surgeries in sicker children usually require at least two large-bore catheters. In pediatrics, however, “large bore” is a relative term, with 22-gauge catheters typically providing sufficient access in infants. Preferred sites for larger catheters include the antecubital and saphenous veins. In cases in which access to the central circulation is required (as for pressure monitoring, infusion of vasoactive medications, or prolonged access), longer catheters may be placed via the femoral, subclavian, or internal jugular vein (the latter usually via a high, anterior approach).⁴⁰ Although secure access may also be obtained via the external jugular vein, it is often difficult to negotiate the J-wire or catheter tip into the central circulation.⁴¹ Peripherally inserted central catheter (PICC) lines are becoming increasingly used in hospitalized children. Although they represent a long-term means of delivering IV fluids and medications in children who require such treatment interventions, there are limitations to their utility intraoperatively. First, practitioners must maintain strict sterile technique when accessing these lines in order to avoid infection. Second, the lines tend to have smaller diameters and are quite long. Therefore, resistance can be significant, even with larger-bore PICC lines. Consequently, bolus and infusion dosing of drugs can be met with high resistance. These lines are not appropriate for large-volume resuscitation, and care should be taken to secure larger-bore IV access if large fluid shifts or significant blood loss is anticipated.

In selecting the appropriate IV catheter, it is useful to consider the relative effects of catheter length and diameter on solution flow rates. Longer catheters offer more resistance to flow than shorter ones and are therefore less desirable when rapid infusion of a large volume of fluids is necessary (see E-Fig. 51-1 and Figs. 51-1, 51-2). In vitro, catheters that were designed for peripheral venous access had 18% to 164% greater flow rates when compared with the same-gauge catheters designed for central venous use. Under pressure, as might be employed during emergent volume resuscitation, rates differed up to 17-fold.⁴² Although this seems to suggest that short peripheral catheters should be preferred, in vivo data are more complex. In animal models, overall catheter flow rates are less than in vitro rates, and central access presents somewhat less resistance to flow than peripheral access.⁴³ Finally, when weighing the risks and benefits of central versus peripheral access, it is also interesting to consider that central administration of resuscitation medications may provide little practical advantage over peripheral administration.⁴⁴

Intraosseous devices are now commonly used in the initial resuscitation of critically ill or injured children (see Fig. 48-6).^45,⁴⁶ Flow rates via these devices depend less on needle diameter than on resistance in the marrow compartment.⁴⁷ In the operating room, the intraosseous route has been used for both induction and maintenance of anesthesia.⁴⁸^–⁵⁰ However, onset of drug effect is less predictable, and the device is more easily dislodged than an IV catheter. Potential complications include compartment syndrome⁵¹^–⁵³ and, very rarely, damage to the growth plate.^46,⁵⁴ Such devices are probably best considered an emergency or last-resort option.⁴⁹

To prevent accidental volume overload, the amount of IV fluid available to administer to a child at any one time should not exceed the child’s calculated hourly requirement. Particularly in infants, a volumetric chamber should be used to limit the amount of fluid available for infusion. Similarly, a microdrip infusion set limits the rate of fluid administration and permits much greater control. Although a fluid infusion pump provides the most precise mode of regulating the rate of fluid administration (and is therefore very useful in providing supplemental fluids or medications), such devices are impractical on primary access lines because they hinder the ability to administer drugs or fluids rapidly. In addition, the clinician should be mindful that pumps may continue to infuse through dislodged catheters and may give inappropriate reassurance of adequate IV access and fluid administration, providing a false sense of reassurance; administration of a large volume of fluid interstitially will fail to deliver the needed fluid volume (and medications). Moreover, if an identification bracelet of any sort is proximal to the IV insertion site, it may act as a tourniquet, resulting in ischemic digits. Therefore, access to the IV site is important in children, as well as removal of all bracelets that are proximal to ipsilateral IV insertion sites.

In neonates and small infants, when rapid infusion of resuscitation solutions or blood products is anticipated, many practitioners find it helpful to include an in-line stopcock manifold. Additional fluids may be drawn up into syringes and warmed separately; during periods of sudden blood loss, stored syringes may then be inserted into the manifold and a known volume rapidly infused.

Finally, in prolonged surgeries or when volume replacement is great, it is imperative that all IV infusions be adequately warmed. Also, in younger infants and children in whom communication exists between the right and left sides of the circulation (e.g., patency of the foramen ovale), an in-line “bubble” filter is desirable.

Choice and Composition of Intravenous Fluids

In the early 1960s,⁵⁵ simultaneous measurements of plasma and extracellular fluid volumes demonstrated that, during surgery, plasma volume is supported at the expense of the extravascular space. At the same time, it was classically observed that isotonic resuscitation fluids temporarily redistribute from the intravascular spaces to what was originally believed to be a third, nonfunctional space. It is now clear that destruction of the endothelial glycocalyx, as caused by surgical trauma, permits fluid to shift from the intravascular to interstitial space. Therefore, the historical “third space” simply represents reversible expansion of the interstitium. Because of the differences in fluid distribution and renal function in infants compared with older children, it was at first unclear that these findings could be extended to infancy. Thus, fluid restriction remained the standard of care until careful studies specifically demonstrated that fluid and electrolyte requirements are often extremely large in neonates who are undergoing major surgical procedures.⁵⁶^–⁵⁸

Although hypotonic fluids are selected for maintenance hydration throughout the hospital (according to the reasoning outlined previously), isotonic solutions are preferred intraoperatively for several reasons. First, most ongoing volume losses are isotonic, consisting of shed blood and interstitial fluids. Second, large volumes of hypotonic solutions may rapidly diminish serum osmolality, producing very low concentrations of electrolytes (in particular, sodium) and undesirable fluid shifts. Indeed, even large volumes of “isotonic” fluids have been shown to significantly decrease serum osmolality in adult volunteers.⁵⁹ Third, as discussed earlier, the plasma volume expansion that is necessary in response to diminished vascular tone under anesthesia is difficult to achieve even with isotonic fluids. Finally, increases in ADH and other elements of intraoperative physiology result in free water retention in excess of sodium if inadequate amounts of the latter are provided.

The compositions of commonly used IV solutions are presented in Table 8-5. Assuming normal plasma osmolality of 275 to 290 mOsm/L, it is noteworthy that 0.9% NaCl (normal saline, NS) is slightly hypertonic to plasma and that LR is isotonic (273 mOsm/L), although slightly hyponatremic (130 mEq/L). For dextrose-containing solutions, added osmolality is rapidly dissipated as sugar is metabolized, resulting in increased volumes of free water. Therefore, administration of 5% dextrose in water is ultimately equivalent to administration of free water.

TABLE 8-5 Composition of Extracellular Fluid and Common Intravenous Solutions

The controversy regarding the perioperative use of colloid versus crystalloid fluid replacement remains an unresolved subject. It is worth noting that aside from 5% albumin, synthetic colloids are gaining popularity among pediatric practitioners. One reason for this is the development of newer synthetic colloids that have a more favorable side effect profile. Hydroxyethyl starches (HES) are synthetic colloids that are simply modified polysaccharides. Circulating amylases quickly break down natural polysaccharides. HES solutions avoid this problem by substituting hydroxyethyl groups for hydroxyl groups at carbon positions C-2, C-3, and C-6. This results in a soluble molecule that is resistant to hydrolysis. These compounds are characterized by three attributes: average mean molecular weight (MW), molar substitution (MS), and the C2:C6 ratio, which relates to the relative positions of hydroxyethyl groups on the polysaccharide molecule.

HES solutions with a greater MW-to-MS ratio tend to remain in the intravascular space much longer than those with smaller ratios. However, they also are prone to more significant side effects including hypocoagulability. Newer, low MW/low MS solutions have much less effect on hemostatic mechanisms than older, higher MW/higher MS solutions. The exact mechanism of the effect on coagulation remains unclear, although it is thought that the HES compounds interfere with von Willebrand factor, factor VIII, and platelet function. A greater C2:C6 ratio is responsible for the slower degradation of the starch by amylase with fewer side effects.⁶⁰ Newer HES solutions (e.g., HES 130/0.42/6:1) are very safe in children scheduled for surgery who have normal renal function; they maintain hemodynamic stability and produce only mild to moderate changes in acid–base status.⁶¹ Synthetic colloids such as these should be considered in surgical patients who demonstrate the need for aggressive intraoperative fluid resuscitation, such as children with large-volume blood loss or excessive insensible losses. Use of these solutions in cardiac surgery remains controversial given the effects on coagulation factors and platelet function induced by the cardiopulmonary bypass circuit.

The routine intraoperative use of glucose-containing solutions has also been a subject of debate. As a rule, operative stress evokes physiologic responses that increase serum glucose. In practice, therefore, hypoglycemia is seldom a problem in healthy children when glucose is omitted from perioperative IV fluids.^62,⁶³ Indeed, the risk should be particularly small if the period of fasting is limited to less than 10 hours.⁶³ At the same time, rapid administration of dextrose solutions may certainly produce acute hyperglycemia and hyperosmolality.^62,⁶³ Therefore, glucose-containing solutions should not be used to replace fluid deficits, third space losses, or blood losses. However, some populations, such as debilitated infants,⁶⁴ children who are malnourished, neonates and infants < 6 months of age^62,^64a,64b and those undergoing cardiac surgery, have been shown to be at risk for intraoperative hypoglycemia,^65,⁶⁶ and the use of glucose-containing solutions (1% to 2.5% dextrose)^62,^64a ^,67, along with intraoperative glucose monitoring, may be beneficial in these children.

Hyperalimentation

It is now common practice that critically ill children arrive in the operating room with hyperalimentation solutions infusing. Common contents of hyperalimentation solutions are shown in Table 8-6. In general, children require 0.5 to 3.0 mg/kg/day of protein, 6 to 9 mg/kg/min of glucose, and 0.5 to 3 g/kg/day of fat. Children receiving parenteral nutrition preoperatively should continue to receive those infusions separately, and a corresponding volume should be deducted from isotonic operative fluids. Hyperalimentation typically consists of two infusions: Intralipid (Fresenius Kabi, Uppsala, Sweden) and a concentrated glucose/protein solution. It is prudent to discontinue the Intralipid solution during surgery; but if that is not possible, then every effort should be made to avoid accessing any ports in the line to reduce the risk of contaminating the Intralipid. Conversely, the concentrated glucose/protein solution should be continued at the same rate (because circulating insulin concentrations have acclimated accordingly). Because of hyperglycemic responses to the stress of surgery and reduced metabolism due to anesthesia and hypothermia, some practitioners routinely decrease hyperalimentation infusion rates by one third to one half. If the latter practice is followed, clinicians should consider checking serum glucose concentrations at regular intervals to monitor for hypoglycemia. Under no circumstances should concentrated glucose solutions (such as D₁₀ or D₂₀) be abruptly discontinued, because high levels of circulating insulin may cause a precipitous and profound decrease in the serum glucose concentration.

TABLE 8-6 Common Contents of Parenteral Nutrition Solutions *

Carbohydrates

10%, 12.5%, 20%, 25%, 30% Dextrose

Limited to D₁₀ or D_12.5 if through a peripheral catheter

Protein

In the form of amino acids

0.5, 1.0, 1.5, 2.0, 2.5, or 3.0 g/kg/day

Lipids

10%, 20% Lipids

Standard Additives

Phosphorus: 10 mmol/L

Heparin

^*Common contents of parenteral nutrition solutions containing dextrose, protein, lipids, and standard additives such as electrolytes. These values represent standard starting points that may be modified based on individual patient needs.

Concerns regarding the routine use of intraoperative dextrose-containing solutions, in large part from recognition that hyperglycemia may exacerbate neurologic injury after an ischemic or hypoxic event, resulted in most clinicians’ avoiding such solutions for routine cases. If dextrose-containing solutions are used, appropriate monitoring is advised to avoid serum glucose extremes. Many practitioners administer glucose-containing solutions as a separate piggyback infusion using an infusion pump or other rate- or volume-limiting device to avoid accidental bolus administration. Alternatively, evidence indicates that isotonic solutions that contain reduced glucose concentrations (e.g., 1% or 2.5% versus 5%) are safe alternative solutions for intraoperative use that minimize both hyponatremia and hyperglycemia.⁶⁸ In the United States, several Food and Drug Administration (FDA)-approved solutions containing 2.5% dextrose are available but none in lower concentrations. In Europe, 1% dextrose electrolyte solutions are now available.^64a ^,68 Because intraoperative administration of solutions containing 5% dextrose (D₅LR) frequently causes hyperglycemia, prudent anesthesiologists should selectively administer dextrose-containing solutions to those who are at particular risk for intraoperative hypoglycemia (i.e., neonates, chronic malnourished children, and cachectic children). In these instances, it may be sensible to administer solutions with a reduced dextrose concentration.^60,^64a

Fasting Recommendations

The goal of fasting is to minimize the volume of gastric contents and thereby lessen the risk of vomiting and aspiration during induction of anesthesia. In children, as opposed to adults, this is of particular concern because in many institutions induction in children is more often accomplished by inhalation than by IV anesthesia and the period of vulnerability to regurgitation is potentially protracted compared with an IV induction.

At issue is the effectiveness of fasting in lowering a child’s gastric volume and the benefits of this effect when weighed against the added discomfort and risk of dehydration. Numerous studies of gastric volume and pH have convincingly demonstrated that clear liquids are rapidly emptied from the stomach and the stimulated peristalsis actually serves to decrease gastric volume and acidity. Taking this together with the benefits of improved hydration and mental status, it is clear that prolonged nil per os (NPO) status is unwarranted. NPO guidelines currently in use in many institutions are included in Table 4-1.⁶⁹ As described earlier, for the vast majority of children, 20 to 40 mL/kg of LR given intraoperatively will provide adequate fluid deficit replacement.

Assessment of Intravascular Volume

Once the child is anesthetized, many clinical clues to volume status are lost or confounded by operative events. For example, although it is a fairly reliable indicator of volume status in the quietly resting preoperative child, tachycardia may result from any number of factors besides intravascular volume status during surgery. It is the challenge of the anesthesiologist to view the entire clinical picture, consider the possibilities, integrate them into a hypothesis, and then test the hypothesis.

Assessment of intravascular volume begins with knowledge of age-related norms for heart rate and BP (see Tables 2-7 and 2-8). Is the heart rate persistently increased, or does it vary with surgical stimulation? Is the pulse pressure narrow, or, more ominously, is the BP reduced for age? Does it vary with positive-pressure breaths? Are the extremities warm? Is capillary refill brisk? What is the urine output? Are these variables changing? What is the rate of the change? When hypovolemia is suspected, observing the response to a 10- to 20-mL/kg bolus of isotonic crystalloid or colloid may test the hypothesis.

Measurement and continuous monitoring of central venous pressure are often helpful in assessing the status of circulating volume (see Figs. 48-2 to 48-5). In addition to traditional central lines introduced into the superior vena cava or left atrium, animal⁷⁰ and limited clinical⁷¹ data suggest that femoral lines that terminate in the abdominal vena cava may also be useful. In infants and children, mean end-expiratory pressure measurements in the right atrial and inferior vena cava differed by less than 1 mm Hg.⁷¹ Assessment of changes in the contour of the arterial waveform may also be helpful in assessing volume status and the response to volume administration (see Fig. 10-11).

Ongoing Losses and Third-Spacing

During all surgical procedures, fluid loss from the vascular space is primarily the result of three simultaneous physiologic processes. First, whole blood is shed at various rates and must be replaced. Second, capillary leak and surgical trauma result in extravasation of isotonic, protein-containing fluid into interstitial compartments (the so-called third space). Third, anesthetic-induced relaxation of sympathetic tone produces vasodilatation (increased capacitance) and relative hypovolemia (a virtual loss). In very small infants, a fourth source of losses, direct evaporation, must also be carefully considered. These ongoing losses are often difficult to quantitate (or even estimate). Although these losses occur in children of all sizes, the small circulating blood volume of an infant (e.g., for a 5-kg infant, 80 mL/kg × 5 kg = 400 mL) leaves little room for error. Faced with uncertainty, the prudent response is constant vigilance and reliance on general principles.

As a rule, 1 mL of shed blood is replaced with 1 mL of colloid (5% albumin or blood) or about 1.5 mL of isotonic crystalloid such as LR.^67,⁷² Isotonic crystalloid is also used to replenish third space losses. Surgical procedures that involve only mild tissue trauma may entail third space losses of 3 to 4 mL/kg/hr. More extensive surgical procedures involving moderate trauma may require replacement equivalent to 5 to 7 mL/kg/hr to adequately support intravascular volume. In small infants undergoing very large abdominal procedures, the losses may approach 10 mL/kg/hr or more.^56,⁵⁸ In neonates, fluid requirements for emergent abdominal surgery for necrotizing enterocolitis have been estimated at up to 50 mL/kg/hr.⁶⁷ These “losses” result from the vascular compartment and include both evaporation and redistribution of fluid. The latter must be most carefully considered because it is exacerbated by the hemodilution and increased capillary pressures caused by excessive fluid administration.

Although necessary intraoperatively, third space accumulation represents whole-body salt and water overload that will need to be mobilized postoperatively. The price of unchecked fluid administration is generalized anasarca, pulmonary edema, bowel swelling, and laryngotracheal edema. In the healthy child, this relative fluid overload is well tolerated, with most excess fluid excreted over the first 2 postoperative days. In children with impaired pulmonary, cardiac, or renal function, however, such fluid excess may result in clinically important postoperative morbidity.

Postoperative Fluid Management

General Approach

Well-planned postoperative fluid management complements the intraoperative plan and accounts for evolving physiology as the child recovers. Replacement of fluid deficits is completed. Ongoing losses are replaced. The child is repeatedly reassessed, and intake is adjusted until normal fluid and electrolyte homeostasis has returned. To aid in decision making, trends in vital signs are identified, all sources of fluid intake and output are quantitated, urine specific gravity is monitored, daily weights are obtained, and serum electrolytes are measured.

In simple outpatient surgeries, discharge is possible after fluid deficits are replaced. In complex cases, replacement fluids may require hourly readjustment that is based on the prior hour’s intake and output. Rather than reacting to single pieces of data, such as low urine output, one must discern overall patterns. High urine output and low urine specific gravity may indicate overhydration or diabetes insipidus. Oliguria may suggest hypovolemia when it is accompanied by high urine specific gravity and clinical signs of dehydration or low cardiac output when it is accompanied by signs of poor perfusion. In the well-hydrated child, oliguria may represent renal failure if the urine specific gravity is normal (or dilute) but increased concentrations of ADH if the urine is concentrated. A careful physical examination is necessary; in many cases, certainty in diagnosis requires simultaneous measurement of serum and urine electrolytes.

Frequently, losses via surgical or gastric drains are large in both real and relative terms. For example, a neonate with a nasogastric tube may lose more than 100 mL/kg/day (normally 20 to 40 mL/kg/day) in gastric fluid. Therefore, in determining the volume and composition of replacement fluids, it is sometimes helpful to consider the electrolyte content of various losses (Table 8-7).

TABLE 8-7 Composition of Body Fluids

Postoperative Physiology and Hyponatremia

Children retain salt and water postoperatively, in part as a result of neuroendocrine activation by stress, continued capillary leak with third space accumulation, non-osmotic stimulation of ADH (fever, stress, opioid administration) and hypovolemia-induced renin secretion. As outlined earlier, intravascular volume depletion is a potent non-osmotic signal for fluid retention and may override osmotic signals under a variety of clinical circumstances.

At the same time, ongoing fluid and electrolyte losses after surgery via chest tubes, nasogastric suction, weeping incisions, and even continued slow bleeding may be substantial. Postoperatively, children often depend entirely on IV fluids for replacement of these and other losses.

Therefore, unless isotonic, sodium-containing fluids are provided, postoperative children are universally at risk for developing hyponatremia.^73,⁷⁴ In a retrospective review of 24,412 surgical admissions to a large children’s hospital, the incidence of significant postoperative hyponatremia was 0.34%, with a substantial mortality rate (8.4%) in these previously healthy children.⁷⁵ If this measured incidence were extended to the entire population in the United States, 7448 children would present annually with postoperative hyponatremia and 626 would die from an entirely avoidable cause. Mortality rates as great as 40% to 60% have been reported after hyponatremia, although it may only be a surrogate marker for a disease with a poor prognosis rather than the actual cause of the death.^76,⁷⁷

In reviewing the etiology of hyponatremia, two factors stand out: extensive extrarenal loss of electrolyte-containing fluid and IV replacement with hypotonic fluids.⁷⁵ In addition, delay in recognition often plays a major role in associated morbidity. The solution seems a simple one: (1) administration of hypotonic fluids without a specific indication should be avoided postoperatively, (2) ongoing losses should be replaced in a timely fashion, and (3) serum electrolytes should be measured routinely in children exhibiting potential symptoms of hyponatremia (see later discussion).

Postoperative Pulmonary Edema

Children who receive large volumes of fluid intraoperatively are at risk for development of pulmonary edema as operative fluids are mobilized. Usually, fluid mobilization begins to occur on the second postoperative day and continues through day 3 or 4. Although this is less common in children than in the elderly, it occurs occasionally in children with burn injuries ⁷⁸ or in pediatric patients receiving large amounts of fluid during resuscitation from trauma or sepsis. In one review,⁷⁹ 13 patients (11 adults and 2 children) developed postoperative pulmonary edema; all began to exhibit symptoms within 36 hours after surgery and had total net fluid retention in excess of 67 mL/kg postoperatively.

Pathophysiologic States and Their Management

Fluid Overload and Edema

Edema is essentially a “sodium disease,” representing sodium and water overload with excessive fluid residing in the extracellular space. Although intracellular volume changes can sometimes be substantial, prolonged cell swelling represents failure of essential volume regulatory functions and is likely a preterminal event. In fluid-overload states, plasma volume is generally increased unless the balance of Starling forces is disturbed, as in nephrotic syndrome or lymphatic obstruction. Edema formation is opposed by (1) low compliance of the interstitial compartment, (2) increased lymphatic flow, (3) osmotic washout of interstitial proteins, and (4) impedance and elasticity of the proteoglycan gel. The differential diagnosis of fluid overload and edema formation is presented in Table 8-8. Principles of therapy for fluid overload states include the following:

Fluid restriction

Salt restriction

Diuresis, dialysis

Salt-poor albumin for diminished plasma volume

TABLE 8-8 Differential Diagnosis of Fluid Overload and Edema Formation

Condition	Differential Diagnosis
Imbalance of intake and output	Salt poisoning
	Formula dilution errors
	Intravenous infusion errors
	Drugs given as sodium salts
Steroid excess with normal sodium intake	Congenital adrenal hyperplasia
Steroid excess with normal sodium intake	Exogenous steroids
Perceived decreases in effective plasma volume	↓ MAP → baroreceptors → ↑ sympathetic tone, ADH, renin, aldosterone
	Vasodilators
	Congestive heart failure
	Cirrhosis
	Nephrotic syndrome
Impaired sodium excretion	Chronic renal failure
	Acute glomerular disease (↓ GFR with normal tubular function)
	Nonsteroidal antiinflammatory drugs (↓ PGE₂ and RBF)
Water excess	SIADH
	Hypotonic infusion
	Stress (↑ ADH)

ADH, Antidiuretic hormone; GFR, glomerular filtration rate; MAP, mean arterial pressure; PGE₂, prostaglandin E₂; RBF, renal blood flow; SIADH, syndrome of inappropriate antidiuretic hormone secretion.

Dehydration States

Dehydration states are common in children. The extent of dehydration is best assessed by weight, because clinical signs such as tachycardia, capillary refill, and skin elasticity,⁸⁰ although often reliable, may be influenced by factors other than hydration status. A capillary refill time of 1.5 to 3.0 seconds, for example, suggests a fluid deficit of between 50 and 100 mL/kg, yet this sign is extremely dependent on ambient temperature.⁸¹ Similarly, poor skin elasticity reflects significant volume loss, yet elasticity may be well preserved in children with hypernatremic dehydration.⁸⁰ Clinical signs associated with varying levels of dehydration are presented in Table 8-9.

TABLE 8-9 Clinical Signs and Symptoms for Estimation of Severity of Dehydration in Infants

As a first approximation, correction of most dehydration states in older children is most readily achieved with administration of a simple bolus of NS, LR, or PLASMA-LYTE 148 (balanced crystalloid solution). In infants or children with unusual, prolonged, or severe dehydration, however, management must be more precise. A five-point questionnaire to assess the severity of dehydration may help develop an appropriate treatment strategy ⁸²:

1. Does a volume deficit exist and, if so, how great is it?

As noted previously, assessment of volume deficit is best made by weight, yet rough estimates of 5% (mild), 10% (moderate), and 15% (severe) may be made in infants based on clinical signs (see Table 8-9).

2. Does an osmolar disturbance exist? Is it acute or chronic?

An osmolar imbalance is determined by measuring the serum sodium concentration. The majority of clinically encountered dehydration states (~80%) are isotonic (Na⁺ = 130 to 150 mEq/L). These isotonic losses are easily managed by almost any strategy.

Approximately 15% of dehydrated children present with hypertonic dehydration (Na⁺ > 150 mEq/L). These children are at greatest risk and have usually experienced the greatest fluid losses for a given set of clinical signs.⁸³ If the condition is chronic, they may require extensive, slow rehydration over much longer periods.⁸⁴

Five percent of children present with hypotonic dehydration (Na⁺ < 130 mEq/L). For a given fluid deficit, these individuals are often more symptomatic than others, and their requirement for sodium replacement is greatest. Surprisingly, rapid improvement in clinical condition often results after the first fluid bolus.

In general, chronic dehydration states must be repaired slowly and acute dehydration states (<24 hours) may be corrected more rapidly. This is because cell volume equilibration occurs acutely through gain or loss of electrolytes (which are moved rapidly) and chronically through gain or loss of organic osmolytes (which are moved more slowly).¹¹ Reequilibration of brain cell volume during correction of hypertonicity can be very slow, mandating patience in correction of chronic fluid deficits. Similarly, rapid correction of hyponatremic disturbances can be hazardous,⁷⁵ even when seemingly safe isotonic solutions are employed.⁸⁵

3. Does an acid–base abnormality exist?

Quantitation of the child’s acid–base status gives useful, although limited, information as to the severity of dehydration. When evaluating acid–base status, it is important to recall that bicarbonate reabsorption and urine acidification are limited in preterm and young infants, leaving even the normal infant in a state of mild metabolic acidosis (pH, 7.3; serum bicarbonate 20 to 21 mEq/L [normal 22 to 26 mEq/L]). Although slow, spontaneous correction of acid–base status is typically observed on rehydration, rapid fluid boluses in poorly perfused children may result in a transient “reperfusion acidosis” as returning circulation washes the products of anaerobic metabolism out of the tissues. In this setting, or when renal insufficiency exists, blood-buffering capacity is such that children with serum bicarbonate concentrations lower than 8 mEq/L or pH lower than 7.2 may benefit from administration of supplemental base (sodium bicarbonate) (Fig. 8-6).⁸²

FIGURE 8-6 Data from children with metabolic acidosis ¹²⁷ were used to depict the displacement of pH as serum bicarbonate declines. The zone of rapid pH displacement (pH < 7.20) has a slope that is several times greater than the zone of gradual pH displacement (pH ≥ 7.20). As the pH moves through the zone of rapid pH displacement, a further decline of serum bicarbonate, of as little as 1 or 2 mEq/L, produces a highly leveraged further decrease of pH. [H],⁺ Hydrogen ion concentration; [HCO₃⁻], bicarbonate ion concentration; Pco₂, carbon dioxide tension.

(From Kallen RJ. The management of diarrheal dehydration in infants using parenteral fluids. Pediatr Clin North Am 1990;37:265-86.)

Rapid bedside evaluation of acid–base status utilizes the following general relationships: a pH decrease of 0.1 unit accompanies a base excess (BE) of approximately 6 mEq/L or an increase in carbon dioxide tension (Pco₂) of 10 to 12 mm Hg. The total replacement base required is then determined by the following equation:

Clinically, a smaller sodium bicarbonate dose (1 to 2 mEq/kg) is given initially, the response is verified by blood gas analysis, and the remaining doses are titrated to effect.

4. Is renal function impaired?

Initial evaluation includes the timing of the last urine void and recent urine output, measurement of urine specific gravity, and serum levels of blood urea nitrogen and creatinine. If uncertainty persists, measurement of serum and urine electrolytes for comparison and calculation of the FE_Na is indicated (see Chapter 26).

FE_Na values of less than 1% imply prerenal conditions causing renal dysfunction, whereas FE_Na values greater than 2% to 3% suggest renal insufficiency. In prematurity, however, values as high as 9% may be seen in otherwise normal infants.

5. What is the state of potassium balance?

Potassium homeostasis is critical to life, and serum potassium levels are generally maintained within a very narrow range. Nonetheless, serum potassium concentrations do not reflect whole-body stores and substantial potassium depletion may exist in the presence of modest changes in serum potassium concentration (K⁺_serum). Gastrointestinal losses or metabolic acidosis is usually accompanied by a potassium deficit, whereas other dehydration states are not. Rapid fluid boluses or pH correction, or both, may acutely reduce K⁺_serum,⁸⁶ and refractory hypokalemia may occur in children deficient in magnesium.⁸⁷ In all cases, adequate renal function should be present before administration of potassium, and complete repletion should be accomplished over 48 to 72 hours. Once the nature and severity of dehydration have been determined, the clinician may proceed using any of a variety of correction strategies. In one approach, moderate to severe dehydration deficits may be estimated, as in Table 8-9. Fluid and electrolyte repair may then proceed according to a three-phase approach wherein circulating plasma volume, perfusion, and urine output are restored rapidly using isotonic crystalloid or colloid solution and remaining deficits are corrected over 24 hours as follows⁸⁸:

Emergency Phase: 20 to 30 mL/kg isotonic crystalloid/colloid bolus

Repletion Phase 1: 25 to 50 mL/kg over 6 to 8 hours. Anions: Cl⁻ 75%, acetate 25%

Repletion Phase 2: remainder of deficit over 24 hours (isotonic) or 48 hours (hypertonic). Include calcium replacement as necessary.

Hypernatremia and Hyponatremia

As previously detailed, disorders of sodium equilibrium are primarily marked by disturbances of fluid balance and are corrected according to the principles outlined earlier. Serious hypernatremia or hyponatremia is accompanied by neurologic symptoms whose severity is determined by the degree and rate of change of serum sodium concentration (Na⁺_serum).

Hypernatremia

In contrast to its rareness in adults, acute hypernatremia is common in children. A mortality rate greater than 40% for the acute disorder and 10% for the chronic disorder has been quoted for hypernatremia (serum sodium >160 mEq/L).^76,⁸⁹ Mortality and permanent neurologic injury are even more common in infants. Depending on degree and duration, neurologic findings include irritability and coma. Seizures may be a presenting symptom, but they are more commonly encountered after the start of therapy. Children with acute conditions are usually symptomatic, whereas those with chronic conditions (acclimated individuals) are typically asymptomatic. General principles for treatment of hypernatremia are as follows:

In the setting of circulatory collapse, colloid or NS bolus should be administered. Although the assertion is debatable, a colloid bolus provides the theoretical benefit of sustained hemodynamic support with lower fluid load. Saline, in contrast, rapidly reequilibrates, necessitating repeated boluses while adding to the total salt burden.

Once a stable circulation has been restored, fluid deficit should be assessed as accurately as possible and corrected over 48 to 72 hours. Cautious correction of the serum sodium level is required. The serum sodium concentration and osmolality should be continuously reassessed during fluid administration, aiming for a correction of no more than 1 to 2 mOsm/L/hr. After as little as 4 hours of hypernatremia, idiogenic osmoles (e.g., trimethylamines) appear in the brain to prevent cerebral volume depletion. Rapid administration of free water could induce cerebral edema, seizures, and death. Therefore, free water should be administered slowly. Because of possible associated hypoglycemia, some solutions should be glucose-containing, and serum glucose levels should be monitored.

Vigilance for seizures, apnea, and cardiovascular compromise should be maintained, because such complicating factors can be the primary determinants of successful outcome.

Hyponatremia

Hyponatremia is also common in infants and children. Increasing prevalence due to erroneous formula dilution has intermittently been reported.^90,⁹¹ In the practice of anesthesiology, mild hyponatremia is a common postoperative condition after surgery of any severity⁹²^–⁹⁴; in neurosurgical patients, hyponatremia may represent cerebral salt wasting or syndrome of inappropriate antidiuretic hormone secretion (SIADH).⁹⁵ In general, symptomatic patients are acutely hyponatremic and asymptomatic individuals are chronically hyponatremic.⁹⁶ After surgery, acutely hyponatremic children may present with nonspecific symptoms that are often erroneously attributed to other causes. Early central nervous system symptoms include headache, nausea, weakness, and anorexia. Advancing symptoms include mental status changes, confusion, irritability, progressive obtundation, and seizures. Respiratory arrest (or irregularity) is a common manifestation of advanced hyponatremia.

When planning to correct hyponatremia, the presence of symptoms must be considered a medical emergency, whereas asymptomatic children do not require rapid intervention. Chronic hyponatremia must be corrected slowly and by no more than 0.5 mEq/L/hr to avoid neurologic complications that include central pontine myelinolysis.⁹⁷ The best treatment for acute hyponatremia is early recognition and intervention. Because hypoxia exacerbates the neurologic injury, the simple ABCs of resuscitation are attended to first, and the airway is secured in the child who has seizures or respiratory irregularity. Hyponatremic seizures may be quieted by relatively modest (3 to 6 mEq/L) increases of serum sodium.⁹⁸ In several series,^91,^99,¹⁰⁰ such limited, rapid correction of symptomatic hyponatremia with hypertonic saline (514 mEq/L NaCl) was found to be well tolerated. It should be emphasized, however, that complete correction is unnecessary and unwise.¹⁰¹ Initial therapy is aimed at increasing the serum sodium concentration no more than is necessary to stop seizure activity (usually 3 to 5 mEq/L). Further, the correction takes place over several days. Hypertonic saline may be used for the correction until the serum sodium increases to greater than 120 mEq/L. Remembering that TBW may range from 75% in infancy to 60% or less in older children, total sodium deficit is estimated as follows:

For example, in a 25-kg child, to correct a serum sodium concentration of 110 mEq/L to 125 mEq/L using hypertonic saline (514 mEq/L), infuse

Because such calculations involve estimates, frequent measurement of the serum sodium concentration is necessary during correction. As with hypernatremia, much of the morbidity and mortality associated with hyponatremia relates to complicating factors such as seizures and hypoxia that may occur during therapy. Therefore, children undergoing therapy should be cared for in a monitored setting. When overzealous correction has occurred, there may be value in acutely relowering Na⁺_serum using hypotonic fluids,⁹⁶ although such therapy is not without its own hazards.

General principles for treatment of hyponatremia are as follows:

Asymptomatic hyponatremia in and of itself need not be rapidly corrected. Associated cardiovascular compromise due to volume depletion may be addressed by colloid bolus or administration of isotonic saline (1 L/m²/day). Provision of sodium is accompanied by free water restriction.

Symptomatic hyponatremia is a medical emergency and may sometimes reflect irreversible neurologic injury. Correction should be rapid, yet limited, as discussed previously. A dose of 2 to 3 mL/kg of 3% saline (514 mEq/L) may be administered over 20 to 30 minutes to halt seizures.

Subsequent correction is accomplished through calculation of the sodium deficit and provision of sodium so as to slowly correct at a rate not to exceed 0.5 mEq/L/hr or 25 mEq/L (total) in 24 to 48 hours.

If the attendant fluid load is excessive or if oliguria is present, diuretics may be useful.

Disorders of Potassium Homeostasis

Hyperkalemia

Hyperkalemia is occasionally the presenting finding in conditions such as congenital adrenal hyperplasia. More commonly, it results from acute renal insufficiency, massive tissue injury, acidosis, or iatrogenic mishaps. In the operating room, acute hyperkalemia may follow the use of succinylcholine in children with myopathies, burns, upper and lower motor neuron lesions, chronic sepsis, or disuse atrophy and occasionally during massive, rapid transfusion of red blood cells or whole blood (see Chapters 6 and 10).¹⁰² It may occur with rhabdomyolysis or as a late sign in malignant hyperthermia.

Although neurologic status is the main concern for children with abnormal serum sodium levels, cardiac status (rate and rhythm) determines the care of children with hyperkalemia. In children with hyperkalemia, the appearance of peaked T waves is followed by lengthening of the PR interval and widening of the QRS complex until P waves are lost. Finally, the QRS complex merges with its T wave to produce a sinusoidal pattern (Fig. 8-7). Successful treatment traditionally utilizes the following approach:

Emergent therapy is first directed toward antagonism of the cardiac effects of potassium by cautious administration of intravenous calcium (calcium chloride [10 to 30 mg/kg] as 0.1 to 0.3 mL/kg of a 10% solution, or calcium gluconate [30 to 100 mg/kg] as 0.3 to 1 mL/kg of a 10% solution) over 3 to 5 minutes to avoid bradycardia. Calcium does not decrease the serum potassium concentration but rather reestablishes the gradient between the resting membrane potential, which is increased in the presence of hyperkalemia, and the threshold potential, which is determined by the calcium concentration. It also increases the refractory period of the action potential, the net effect being the prevention of spontaneous depolarization.

Serum potassium is then reduced by returning potassium to the intracellular space. This is done by correcting acidosis through administration of sodium bicarbonate (1 to 2 mEq/kg), mild to moderate hyperventilation, and administration of a β-agonist.

To maintain potassium in the intracellular space, glucose and insulin are administered by infusion (0.5 to 1 g/kg glucose with 0.1 U/kg insulin over 30 to 60 minutes).

After stabilization, attention is directed toward removal of the whole-body potassium burden (sodium polystyrene sulfonate [Kayexalate], furosemide, dialysis) and correction of the underlying cause (Fig. 8-8).

FIGURE 8-7 Electrocardiographic changes associated with hyperkalemia.

(From Williams GS, Klenk EL, Winters RW. Acute renal failure in pediatrics. In: Winters RW, editor. The body fluids in pediatrics: medical, surgical, and neonatal disorders of acid-base status, hydration, and oxygenation. Boston: Little, Brown; 1973. p. 523-57.)

FIGURE 8-8 Algorithm for treatment of hyperkalemia. After stabilization, attention is directed toward removal of the whole-body potassium burden (Kayexalate, dialysis) and correction of the underlying cause. D₂₅, 25% dextrose; ECG, electrocardiogram; IV, intravenous; PO, orally; PR, per rectum.

The knowledge that β-adrenergic stimulation modulates the translocation of potassium into the intracellular space ^103,¹⁰⁴ has prompted the consideration of β-agonists in the treatment of acute hyperkalemia.¹⁰⁵^–¹⁰⁸ In children, a single infusion of an IV β-agonist such as salbutamol (5 µg/kg over 15 minutes) effectively reduces serum potassium concentrations within 30 minutes. Because of the rapidity, efficacy, and safety of salbutamol in children, it has become the first-choice treatment for hyperkalemia.¹⁰⁵ In addition to IV therapy, both salbutamol¹⁰⁹ and albuterol¹⁰⁸ by inhalation effectively reduce the serum potassium concentration. It should be noted that at this time, salbutamol is not available in IV formulations in the United States. The inhalation route has the significant advantages of being readily available in emergency departments and not requiring IV access. However, the observation that a paradoxical exacerbation of hyperkalemia sometimes occurs on initiation of treatment,¹⁰⁹ together with concerns regarding the possibility of associated arrhythmias,¹¹⁰ suggests that more experience is required before such therapy can be considered the standard of care. Inhalation of albuterol during such an event in the operating room may speed the reduction in serum potassium while other methods of treatment are instituted.

Hypokalemia

Hypokalemia is most common in children as a complication of diarrhea or persistent vomiting associated with gastroenteritis. Muscle weakness is the most common sign in hypokalemia and has been correlated with the serum potassium concentration.¹¹¹ In the operating room or ICU, hypokalemia may also accompany a wide variety of other conditions, including diabetes, hyperaldosteronism, pyloric stenosis, starvation, renal tubular disease, chronic steroid or diuretic use, and β-agonist therapy. Severe hypokalemia can also be accompanied by electrocardiographic changes, including QT prolongation, diminution of the T wave, and appearance of U waves (see Fig. 8-7).

As noted previously, serum potassium concentrations do not accurately reflect total potassium homeostasis, and low serum concentrations may or may not be associated with significant total body potassium depletion. Indeed, the extracellular fraction of potassium is only a tiny proportion (approximately 3%) of the entire body store. For these reasons, the precise point at which to begin replacement therapy is controversial, and total replacement requirements are impossible to calculate. In general practice, serum potassium values (K⁺_serum) between 2.0 and 2.5 mEq/L are corrected before surgery on the assumption that further decreases may predispose the child to muscular weakness, arrhythmias, and hemodynamic instability.

Potassium replacement is best accomplished orally over an extended period while the underlying cause is evaluated and treated. When IV correction is required, concentrations up to 40 mEq/L should be given slowly (not to exceed 1 mEq/kg/hr) in a monitored setting. Because such solutions often cause phlebitis, large-bore or central catheters are preferred. In the setting of hypochloremia and hypokalemia, chloride deficits must first be replaced, usually by administration of normal saline.

Syndrome of Inappropriate Antidiuretic Hormone Secretion

Many non-osmotic factors are capable of stimulating ADH release, and these can occasionally override osmotic control priorities. When this occurs, clinicians have historically deemed the increased ADH concentrations “inappropriate” because control of the serum osmolarity is lost (see Chapter 26). As detailed earlier, however, intravascular depletion is the most potent stimulus for vasopressin release, and it is hardly inappropriate that defense of circulation takes priority over defense of serum sodium levels. Pain, surgical stress, critical illness, sepsis, pulmonary disease, central nervous system injury, drugs, and a variety of other factors may all stimulate ADH release above and beyond that necessary to maintain osmolar balance.

SIADH is common in children yet is often overlooked. Minor head trauma, for example, may elicit spikes in ADH levels, although infrequently to the extent that it produces serious hyponatremia and seizures.¹¹² Urine output after spinal fusion is often reduced because of increased concentrations of ADH, which usually return to normal within 24 hours without therapy.¹¹³ Infants with bronchiolitis and hyperinflated lungs frequently demonstrate markedly increased plasma ADH concentrations and exhibit fluid retention, weight gain, urinary concentration, and plasma hypoosmolality until their illness begins to resolve.¹¹⁴ Hyponatremia to the point of seizures, however, is only occasionally observed.

The diagnosis of SIADH rests on the identification of impaired urinary dilution in the setting of plasma hypoosmolality. Hyponatremia (Na⁺ < 135 mEq/L), serum osmolality less than 280 mOsm/L, and urine osmolality greater than 100 mOsm/L in the absence of volume depletion, cardiac failure, nephropathy, adrenal insufficiency, or cirrhosis are generally considered sufficient for diagnosis. Therapeutic principles are similar to those for hyponatremia and depend on the following:

Restriction of free water

Repletion of sodium deficits (if present)

Administration of diuretics to offset the effects of vasopressin

Diabetes Insipidus

In the operating room and the ICU, diabetes insipidus is most commonly associated with the care of neurosurgical patients.¹¹⁵^–¹¹⁷ Diabetes insipidus is also caused by neuroendocrine failure in brain death, and management may be necessary if organ donation is requested.^118,¹¹⁹ Diabetes insipidus results from decreased secretion of, or renal insensitivity to, vasopressin (see Chapter 26). Manifestations include massive polyuria, volume contraction, dehydration, and plasma hyperosmolality. Dilute polyuria (<250 mOsm, >2 mL/kg/hr) in the presence of hypernatremia (Na⁺ > 145 mEq/L) with hyperosmolality (>300 mOsm/L) is the hallmark. In central diabetes insipidus, administration of desmopressin concentrates the urine but water deprivation does not. Postoperative diabetes insipidus may initially be difficult to distinguish from mobilization of operative fluids.

Children with craniopharyngioma or a similarly situated pathologic lesion may not manifest vasopressin deficiency early in the disease but become symptomatic preoperatively after steroid administration or intraoperatively during surgical manipulation. Postoperative diabetes insipidus typically begins on the evening after surgery and may resolve in 3 to 5 days if osmoregulatory structures have not been permanently injured. An often-confusing triphasic response may also occur wherein postoperative diabetes insipidus appears to resolve, fluid status normalizes, or SIADH appears and then vasopressin secretion ceases and diabetes insipidus returns. It is hypothesized that this pattern reflects nonspecific vasopressin release from degenerating neurons in the hypothalamic supraoptic and paraventricular nuclei.

Attempts have been made to develop protocols for perioperative management of diabetes insipidus.¹²⁰ Because vasopressin is difficult to titrate to urine output, our practice involves maximal antidiuresis and fluid restriction. In this setting, volume status must be monitored closely, because urine output is no longer a marker of renal perfusion. Children who need close perioperative monitoring for the development of diabetes insipidus include those with preexisting diabetes insipidus as well as those who are undergoing resection of craniopharyngiomas or pituitary lesions or other procedures that involve resection or manipulation of the pituitary stalk.¹²¹

Hyperchloremic Acidosis

Administration of large amounts of NS can lead to excess serum chloride.¹²² The chloride content of NS is 154 mEq/L. This excess of chloride ions can lead to a hyperchloremic acidosis, which has been categorized as a strong-ion acidosis.^123,¹²⁴ This type of acidosis results from an excess of strong anions (e.g., lactate, ketoacids, sulfates) relative to the strong cations. As the strong ion difference increases, acidosis occurs. The magnitude of the acidosis is related to the amount of NS administered as well as the rate of administration. Infusions of 35 mL/kg NS over 2 hours in healthy patients undergoing gynecologic surgery resulted in acidosis. Acidosis did not occur with similar infusions of LR.¹²²