Anatomy

The kidneys are retroperitoneal paired organs located on each side of the vertebral column. A normal adult kidney measures 11 to 12 cm in length, 5 to 7.5 cm in width, and 2.5 to 3.0 cm in thickness. In the adult male, it weighs 125 to 170 g, and in the adult female, it weighs 115 to 155 g. Beneath its fibrous capsule lies the cortex, which contains the glomeruli, the convoluted proximal tubules, the distal tubules, and the early portions of the collecting tubules. The remainder of the tissue, the medulla, contains the pars recta, the loop of Henle, and the middle and distal portions of the collecting duct. The inner medulla borders the renal pelvis, where urine is received from the collecting ducts. The ducts and loops are arranged into cone-shaped bundles called pyramids, which have tips that project into the renal pelvis and form papillae. The pelvis drains into the ureters, which in the adult human descend a distance of 28 to 34 cm to open into the fundus of the bladder. The walls of the pelvis and ureters contain smooth muscles that contract in a peristaltic manner to propel urine to the bladder.

Renal Blood Flow

Despite accounting for only 0.5% of body weight, the kidneys receive about 25% of the cardiac output, with a blood flow of approximately 4 mL/min per gram of kidney tissue. Renal plasma flow (RPF) in women is slightly lower than it is in men, even when normalized for body surface area, averaging 592 ± 153 mL/min per 1.73 m² and 654 ± 163 mL/min per 1.73 m², respectively (Smith, 1943). In children between the ages of 6 months and 1 year, normalized RPF is half that of adults but increases progressively to reach adult levels at about 3 years of age (McCrory, 1972). After the age of 30 years, renal blood flow (RBF) decreases progressively; by the age of 90 years, it is approximately half of the value present at 20 years (Davies and Shock, 1950). This generous supply provides not only for the basal metabolic needs of the kidneys but also for the high demands of ultrafiltration.

The basic arterial supply of the kidneys is a single renal artery that divides into large anterior and posterior branches and subsequently into segmental or interlobar arteries. The latter form the arcuate and interlobular arteries. These blood vessels are end-arteries and therefore predisposed to tissue infarction in the presence of emboli. The arcuate arteries are short, large-caliber vessels that supply blood to the afferent arterioles of the glomeruli at a mean pressure of 45 mm Hg, which is higher than that found in most capillary beds. This high hydraulic pressure and large endothelial pore size lead to enhanced glomerular filtration (Brenner and Beeuwkes, 1978).

Glomerular capillaries have many anastomoses but recombine to form the efferent arteriole. The latter subdivide into an extensive peritubular capillary network. This arrangement allows solute and water to move between the tubular lumen and blood. These networks rejoin to form the venous channels, through which blood exits the kidneys.

Ninety percent of RBF goes to the cortex, which accounts for 75% of the renal weight, whereas the medulla and the rest of the kidneys receive 25% of the RBF. Although cortical blood flow is 5 to 6 mL/g per minute, outer medullary blood flow decreases to 1.3 to 2.3 mL/g per minute, and the flow to the papilla is as low as 0.22 to 0.42 mL/g per minute (Dorkin and Brenner, 1991). The unevenness in the distribution of RBF between the cortex and the medulla is necessary to develop and maintain the medullary gradient of osmotically active solutes that drive the countercurrent exchange/multiplier, which is essential for the elaboration of concentrated urine. Outer medullary blood flow may preferentially supply the loop of Henle, thereby accounting for the striking influence of loop diuretics in that region. Furthermore, papillary blood flow is far greater than the metabolic needs of the renal parenchyma and is well adapted to the countercurrent concentrating mechanism characteristic of this region.

RBF remains almost constant over a range of systolic blood pressures from 80 to 180 mm Hg, a phenomenon known as autoregulation. Consequently, glomerular filtration is also constant over this range of pressures as a result of adaptations in the renal vascular resistance (Selkurt et al., 1949, Gertz et al., 1966). Because the changes in resistance that accompany graded reductions in renal perfusion pressure occur in both denervated and isolated perfused kidneys, autoregulation appears not to depend on extrinsic neural or hormonal factors (Thurau, 1964). According to the “myogenic hypothesis” first proposed by Bayliss (1902), the stimulus for vascular smooth muscle contraction in response to increasing intraluminal pressure is either the transmural pressure itself or the increase in the tension of the vascular wall. An increase in perfusion pressure, which initially distends the vascular wall, is followed by a contraction of the resistance vessels and a return of blood flow to basal levels.

There are only a few studies of autoregulation of RBF in developing animals. Aortic constriction in adult animals reduces renal perfusion by 30% but has minimal effects on RBF and glomerular filtration rate, compared with the significant changes observed in 4- to 5-week-old rats (Yared and Yoskioka, 1989). Furthermore, it has been demonstrated that autoregulation of RBF in young rats occurs at renal perfusion pressures between 70 and 100 mm Hg, compared with pressures of 100 to 130 mm Hg in adult rats (Chevalier and Kaiser, 1985). A similar increase in the pressure set point for autoregulation has been found in dogs (Jose et al., 1975). It appears that autoregulation of RBF occurs in the very young and is sufficient to maintain blood flow constant over a wide range of perfusion pressures that are physiologically adequate for the age. No such human studies are available.

Several substances have been proposed to participate in the autoregulation of RBF, including vasoconstrictor and vasodilator prostaglandins, kinins, adenosine, vasopressin, the renin-angiotensin-aldosterone system, endothelin, and endopeptidases (Herbacznska-Cedro and Vane, 1973, Osswald et al., 1978, Maier et al., 1981, Schnermann et al., 1984). Nitric oxide (NO), previously known as endothelium-derived relaxing factor (EDRF), has also been shown to play an important role in regulating renal vascular tone through its vasodilatory action. Bradykinin, thrombin, histamine, serotonin, and acetylcholine act on endothelial receptors to activate phospholipase C, which in turn results in the formation of inositol triphosphate and diacylglycerol, resulting in the release of intracellular calcium (Marsden and Brenner, 1991, Luscher et al., 1992). This in turn stimulates the synthesis of NO from L-arginine. Other factors that stimulate the formation of NO include hypoxia, calcium ionophores, and mechanical stimuli to the endothelium. NO increases RBF by decreasing efferent arteriolar vascular resistance, while glomerular filtration remains unchanged (Marsden and Brenner, 1991).

Because in mature kidneys, autoregulation is lost at arterial pressures less than 80 mm Hg, the lower physiologic pressures prevailing in the newborn period may be expected to limit this important control mechanism. There is evidence both to support and to refute this conclusion (Kleinman and Lubbe, 1972, Jose et al., 1975).

Renal Physiology

The glomerulus is a specialized capillary cluster arranged in loops that functions as a filtering unit. The capillary walls may be viewed as a basement membrane lined by a single layer of cells on either side. In contact with blood are endothelial cells, which contain many fenestrations; podocytes, with their foot processes, line the other side of the basement membrane.

The route by which water and other solutes are filtered from the blood is not fully understood, but it appears that plasma ultrafiltrate traverses the large fenestrations of the glomerular capillary endothelium and penetrates the basement membrane and the slit pores located between the podocyte foot processes. Filtration of large molecules is greatly influenced by the size and charge of the specific molecule, as well as by the integrity and charge of the glomerular basement membrane. Abnormalities in various structural proteins of the slit-pore diaphragm such as nephrin, podocin, and α-actinin may be responsible for several proteinuric disorders (Mundel and Shankland, 2002). In general, the endothelium and the lamina rara interna of the glomerular basement membrane slow the filtration of circulating polyanions such as albumin (Ryan and Karnovsky, 1976). The lamina rara externa and the slit pores slow the filtration of cationic macromolecules such as lactoperoxidase (Graham and Kellermeyer, 1968). Neutral polymers such as ferritin are not filtered because of their large molecular size and shape (Farauhar et al., 1961). Molecules with a radius of 4.2 nm or more are excluded from the glomerular filtrate. In practical terms, red cells, white cells, platelets, and most proteins are restricted to the circulation.

Glomerular Filtration

Among the main functions performed by the kidneys is the process of glomerular filtration. The glomerulus is primarily responsible for the filtration of plasma. The glomerular filtration rate (GFR) is the product of the filtration rate in a single nephron and the number of such nephrons, which range from 0.7 to 1.4 million in each kidney (Keller et al., 2003). Clearance, which is defined as the volume of plasma cleared of a substance within a given time, provides only an estimate or approximation of GFR.

Although tubular reabsorption and tubular secretion may influence the blood level of numerous medications and endogenously-produced substances such as urea, creatinine, and uric acid, the degree of elimination of such substances depends largely on GFR. Hence, in individuals with renal impairment, estimation or measurement of GFR is crucial in determining the dosage adjustment and choice of medications needed to achieve effectiveness while avoiding toxicity. GFR is also a major factor that affects electrolyte composition and volume of body fluids, as well as acid-base homeostasis.

Glomerular filtration is driven by hydrostatic pressure, which forces water and small solutes across the filtration barrier. In healthy individuals, changes in hydrostatic pressure rarely affect single-nephron GFR because autoregulatory mechanisms sustain or maintain a constant glomerular capillary pressure over a large range of systemic blood pressure (Robertson et al., 1972). Hydrostatic pressure is opposed by the oncotic pressure produced by plasma proteins and the hydrostatic pressure within Bowman’s capsule. Mathematically, this relation can be expressed by the following equation:

SNGFR is the single-nephron glomerular filtration rate; K_f is the glomerular ultrafiltration coefficient; P and p are the average hydraulic and osmotic pressure differences, respectively; and P_UF is the net ultrafiltration pressure. As plasma water is filtered, the proteins within the capillaries become more concentrated, so oncotic pressure increases at the distal end of the glomerular capillary loop and the rate of filtration ceases at the efferent capillary (Blantz, 1977). Under normal conditions, about 20% of the plasma water that enters the glomerular capillary bed is filtered; this quantity is referred to as the filtration fraction

RBF has the greatest influence on GFR. Renal parenchymal disorders interfere with autoregulation of RBF, such that GFR may fall, even with low-normal mean arterial blood pressure (MABP). Still more pronounced changes in GFR may occur with hypotension or hypertension, which may accelerate ischemic or hypertensive injury. Clearance of a molecule may serve as an indicator of GFR only if the assayed molecule is biologically inert and freely permeable across the glomerular capillary, if it remains unchanged after filtration, and if it is neither reabsorbed nor secreted by the tubule. The exogenous-filtration marker inulin (a fructose polymer) has all of these attributes and is the ideal standard for measuring GFR. However, inulin-clearance measurement is rarely used clinically because it is an expensive and cumbersome method. Instead, measurement of an endogenous small molecule such as serum creatinine (molecular weight, 0.113 kDa), which is derived from muscle metabolism at a relatively constant rate and is freely filtered at the glomerulus, is a practical, rapid, and inexpensive means for estimating GFR, thereby aiding clinical decisions. Thus, in the steady state, creatinine production and urinary creatinine excretion are equal even when GFR is reduced.

Serum-creatinine concentrations vary by age and gender. In 1-year-old girls values are 0.35 ± 0.05 mg/dL (mean ± SD) and rise gradually to 0.7 ± 0.02 mg/dL (mean ± SD) by 17 years of age; boys have corresponding mean values that are 0.05 mg/dL higher until 15 years of age and 0.1 mg/dL higher subsequently (Schwartz et al., 1987). Expected creatinine-excretion rates in 24-hour urine collections are often used to validate such collections. Values range from 8 to 14 mg/kg per day in neonates and in infants younger than 1 year of age, with an increase to about 22 ± 7 mg/kg per day (mean ± SD) in preadolescent children of either gender (Hellerstein et al., 2001). Subsequently, creatinine excretion in boys is 27 ± 3.4 mg/kg per day.

In healthy children with proportional height and weight, GFR can be estimated by creatinine clearance (CrCl) as calculated by Schwartz’s formula, which does not rely on measurement of urinary creatinine or timed urine collections:

where height is in centimeters, P_CR is the plasma-creatinine concentration in mg/dL, and k is a constant proportion to muscle mass. The value of k is 0.45 in full-term newborns and until 1 year of age, 0.55 in children 2 years of age and older and in adolescent girls, and 0.70 in adolescent boys (Schwartz et al., 1987). Normal CrCl ranges from 90 to 143 mL/min per 1.73 m², with a mean of 120 mL/min per 1.73 m²

Although more cumbersome, calculation of CrCl based on values obtained in 12- or 24-hour urine collections provide a better estimate of GFR. Once the completeness of such collections is validated based on expected creatinine excretion, CrCl is calculated using the following formula:

U is the urinary concentration of creatinine in mg/dL, V is the total urine volume in mL, min is the time of collection in minutes, and P_CR is the serum concentration of creatinine in mg/dL. To standardize the clearance of children of different sizes, the calculated result is multiplied by 1.73 m² (surface area of a standard man in meters squared) and divided by the surface area of the child in meters squared

In children with impaired renal function, GFR estimates based on creatinine methods may grossly overestimate the true GFR, because tubular and gastrointestinal secretion of creatinine increases disproportionately. Hence, serum creatinine concentrations are less reflective of filtration at the glomerulus. For example, Schwartz’s formulas overestimate GFR by 10% ± 3% when GFR is greater than 50 mL/min per 1.73 m² but by 90% ± 15% when GFR is less than 50 mL/min per 1.73 m². Other limitations of creatinine-based GFR determinations stem from variations of analytical assays, reference values ranging from 0.1 to 0.6 mg/dL in children younger than 9 years of age, diurnal variation in serum creatinine levels resulting from high intake of cooked meat or intense exercise, influence of body mass index, and inaccurate urine collections—all of which make comparisons of GFR difficult over time, especially in growing children (Levey et al., 1988). Using cimetidine to block tubular secretion of creatinine before measuring CrCl in urine collections may improve such measurements (Hellerstein et al., 1998).

Measurement of cystatin-C, a 13-kDa serine proteinase produced at a constant rate by all nucleated cells, is purported to be a superior endogenous marker of filtration, because cystatin-C is less susceptible to variation than is plasma creatinine. A meta-analysis compared the correlation between GFR measured by inulin clearance, radiolabeled methods, nonlabeled iothalamate or iohexol, and either plasma creatinine or cystatin-C concentrations measured nephelometrically (Dharnidharka et al., 2002).The correlation between GFR and cystatin-C was significantly higher compared with plasma creatinine (0.846 versus 0.742, P < 0.001). Thus, cystatin-C measurements are becoming increasingly popular in clinical practice, and reference ranges have been generated in children up to 16 years of age (Table 5-1) (Bokenkamp et al., 1998; Finney et al., 2000; Harmoinen et al., 2000).

Studies in renal transplant donors and in individuals with various renal disorders have shown that plasma-creatinine concentration changes minimally as GFR falls to about 50 mL/min per 1.73 m² (Fig. 5-1) (Shemesh, 1985). This compensation is largely the result of hypertrophy and hyperfiltration of the remaining nephrons. When more than 50% of the nephrons cease to function and “renal reserve” is outstripped, serum creatinine may rise rapidly in a parabolic fashion (Fig. 5-1). Thus, when a more accurate clinical assessment of GFR is desirable for research purposes, radiolabeled methods with an identity exceeding 97% give a better approximation of GFR relative to inulin clearance and may be more useful in aiding clinical decisions. In multicenter investigations conducted in the United States using a uniform method for GFR measurement, ¹²⁵I-iothalamate is often used because this isotope has low radiation exposure and long isotope half-life and can be assayed at a central laboratory (Bajaj et al., 1996). Otherwise, ^99mTc-diethylenetriaminepenta-acetic acid (Tc-DTPA) is commonly used to estimate GFR for routine clinical purposes. In other countries, ⁵¹Cr-ethylenediaminetetra-acetic acid (Cr-EDTA), which delivers a greater radiation dosage, is also popular, as are nonlabeled iothalamate and iohexol methods.

FIGURE 5-1 Relationship of serum creatinine to GFR.

(From Shemesh O, Golbetz H, Kriss JP, et al.: Limitation of creatinine as a filtration marker in glomerulopathic patients, Kidney Int 28:830, 1985.)

Although GFR may fluctuate, the kidneys retain the ability to regulate the rate of solute and water excretion according to changes in intake. This regulation is achieved by changes in tubular reabsorption rates—a phenomenon known as glomerular-tubular balance (Tucker and Blantz, 1977). The end result is preservation of ECF volume and chemical composition. Glomerular-tubular balance can be disturbed by several factors, including volume expansion, loop diuretics, and inappropriate secretion of antidiuretic hormone (ADH).

Overview of Tubular Function

The proximal tubule is the site of reabsorption of large quantities of solute and filtered fluid (Fig. 5-2). Many transporters subserving tubular electrolyte transport have been characterized at the genetic level, and various pathologic disorders have been elucidated (Epstein, 1999). Under physiologic conditions, the proximal convoluted tubule isotonically reabsorbs 50% to 60% of the glomerular filtrate (Berry and Rector, 1991). The initial portion of the proximal convoluted tubule reabsorbs most of the filtered glucose, amino acids, and bicarbonate. Glucose and amino acids are absorbed actively, whereby they are transported against their electrochemical gradient, coupled to sodium (Na⁺). Active Na⁺ transport at the peritubular membrane provides the driving force that ultimately is responsible for other transport processes. The system is driven by sodium, Na⁺, K⁺, (activated) adenosine triphosphatase (Na⁺–, K⁺-ATPase), or the Na⁺ “pump,” which requires the presence of K⁺ in the peritubular fluid and is inhibited by ouabain.Micropuncture studies show that around 50% to 70% of the filtered Na⁺ is reabsorbed in this segment, mostly by a process of active cotransport.

FIGURE 5-2 Sodium and water handling by the nephron. A, Glomerulus. B, Proximal tubule, the major site for the reabsorption of Na⁺ (70%), Cl⁻, K⁺ (80%), HCO₃⁻ (80% to 90%), and water. The reabsorptive process is isomotic, regardless of whether the kidneys are concentrating or diluting urine. C, Thin descending loop of Henle. D, Thick ascending loop of Henle. It is always impermeable to water. The medullary portion is important for the generation of free water. There is active Na⁺, K⁺, and Cl⁻ (20% to 25%) reabsorption, which is responsible for driving the countercurrent multiplier and creating increased medullary tonicity. The cortical thick ascending limb and the early distal tubule (E) are responsible for the reabsorption of the remaining HCO₃⁻, to as well as 5% of the filtered Na⁺ and Cl⁻. These segments are impermeable to water and are unaffected by ADH. In the late distal tubule and the cortical collecting duct (F), aldosterone action controls Na⁺ and K⁺ reabsorption and excretion. The medullary portion of the collecting duct is the major site for ADH-dependent water reabsorption. This segment is permeable to water in the presence of ADH. The vasa recta (G) is important in maintaining a concentrated medullary interstitium.

The major fraction of filtered bicarbonate (HCO₃⁻) is absorbed early in the proximal convoluted tubule. Hydrogen (H⁺) gains access to luminal fluid via an Na⁺/H⁺ electroneutral exchange mechanism and forms carbonic acid. The latter is dehydrated to H₂O and CO₂ under the influence of carbonic anhydrase. CO₂ diffuses into the cell, and HCO₃⁻ is re-formed and ultimately absorbed into the bloodstream. In general, the concentration of HCO₃⁻ is maintained at 26 mmol/L, which is slightly below the renal threshold of approximately 28 mmol/L (Pitts and Lotspeich, 1946).

The renal clearance of glucose is exceedingly low, even after complete maturation of glomerular filtration. The amount filtered increases linearly as plasma glucose increases. Initially, all filtered glucose is reabsorbed until the renal threshold has been exceeded (at around 180 mg/dL), at which point filtered glucose appears in the urine. However, maximal tubular glucose (T_mG) reabsorption is attained at a filtrate glucose concentration of about 350 mg/mL (Pitts, 1974). The reabsorption of glucose in the proximal tubule occurs via a carrier-mediated, Na⁺/glucose cotransport process across the apical membrane, followed by passive facilitated diffusion and active Na⁺ extrusion across the basolateral membrane.

Apart from Na⁺, other solutes reabsorbed in the proximal tubule include K⁺, Ca²⁺, P²⁻, Mg²⁺, and amino acids. These are discussed in detail in other sections of this chapter.

The loop of Henle makes the formation of concentrated urine possible and contributes to the formation of dilute urine (Kokko, 1979). This dual function is achieved through the unique membrane properties of the loop, the postglomerular capillaries, and the hypertonicity of the interstitium. The proximity of the descending and ascending portions of loop allows it to function as a countercurrent multiplier, whereas the capillaries serve as countercurrent exchangers (Fig. 5-2). The descending loop of Henle abstracts water from tubular fluid, increasing the intraluminal concentrations of NaCl and other solutes. However, the intraluminal osmolality remains in equilibrium with the interstitium, where 50% of the osmolality results from urea. In the thin ascending limb of the loop of Henle, there is passive efflux of NaCl and urea into the interstitium. The thick ascending limb of the loop of Henle, by being impermeable to water, contributes to the formation of dilute urine.

The final creation of hypotonic or hypertonic urine depends on the distal tubules and collecting ducts and their interaction with ADH. In the distal convoluted tubule, Na⁺ reabsorption occurs against a steep gradient, largely under the influence of aldosterone. K⁺ is secreted by the distal tubule in association with Na⁺ reabsorption and H⁺ secretion. Moreover, this segment of the nephron acidifies the urine and is the only site of new bicarbonate formation. At the end of the collecting duct, about 1% of the filtered water and about 0.5% of the filtered Na⁺ appear in the final urine.

The Kidneys and Antidiuretic Hormone

ADH plays a pivotal role in water homeostasis by acting on the most distal portion of the nephron. ADH is a cyclic octapeptide that, along with its carrier protein, neurophysin, is synthesized in the supraoptic and paraventricular nuclei of the hypothalamus (Zimmerman and Defendini, 1977). The prohormone migrates along the nerve axons to the posterior pituitary gland, where it is stored as arginine vasopressin. It is released through exocytosis (Douglas, 1973).

Several variables affect ADH secretion. Physiologically, the most important factor is plasma osmolality. A very small rise in plasma osmolality is sufficient to trigger a response from the sensitive osmoreceptors located in and around the hypothalamic nuclei, leading to ADH secretion. Conversely, plasma ADH concentrations are less than 1 pg/mL at a physiologic plasma osmolality of less than 280 mOsm/kg water. The antidiuretic activity of ADH is maximal at plasma osmolality of greater than 295 mOsm/kg water, when plasma ADH exceeds 5 pg/mL (Robertson, 2001). Once plasma osmolality exceeds this limit—thus surpassing the capacity of the ADH system to affect maximal fluid retention—the organism depends on thirst to defend against dehydration. Intracerebral synthesis of angiotensin II largely mediates this thirst response and the oropharyngeal reflex. Atrial natriuretic peptide (ANP) opposes the release of ADH and of angiotensin II. In summary, plasma osmolality and Na⁺ are maintained within a narrow range. The upper limit of this range is determined by the sensitivity of the thirst mechanism located in the hypothalamus, whereas its lower range is affected by ADH release.

Nonosmolar factors also influence ADH secretion and may be the key stimuli of ADH secretion in pathologic disorders, leading to hypovolemia and hypotension. These changes are mediated by low-pressure (located in the left atrium) and high-pressure (located in the carotid sinus) baroreceptors. Experimental studies suggest that this nonosmotic pathway of ADH release is less sensitive than the osmotic pathway and is triggered by a 5% to 10% fall in blood volume, whereas a 1% to 2% increase in ECF osmolality can trigger ADH release.

Nonhypovolemic conditions that stimulate ADH release often result in diminished urine volume, hyponatremia, fractional excretion of uric acid greater than 10%, low serum uric acid level (<4 mg/dL), and urinary sodium greater than 20 mEq/L (Albanese et al., 2001). These conditions result in hyponatremia. Conversely, inhibitors of ADH release or primary or acquired nephropathies may result in the inability to respond to ADH or to conserve water, and these inhibitors are often accompanied by polyuria with Uosm of less than 150 mOsm/kg, dehydration, and hypernatremia.

ADH has a major effect on the medullary thick ascending limb and thereby influences the countercurrent multiplier mechanism and urinary concentration. More directly, ADH binds to V₂ receptors in the basolateral membrane of the collecting duct, causing the activation of adenylate cyclase and the formation of cyclic 3′,5′-adenosine monophosphate (cAMP) (Dorisa and Valtin, 1976; Schwartz et al., 1974). This results in insertion of aquaporin-2 water channels in apical membranes and in the activation of apical Na⁺ channels, which causes water conservation (Andreoli, 2001). These effects are counterbalanced by prostaglandin E₂ (PGE₂) and the calcium-sensing receptor in cells of the medullary thick ascending limb that mediate saluresis and diuresis.

Polyuric syndromes can be separated on the basis of urine osmolality and generally consist of water diuresis, solute diuresis, or a mixed water-solute diuresis with typical Uosm of less than 150 mOsm/kg, 300 to 500 mOsm/kg, and 150 to 300 mOsm/kg, respectively (Oster et al., 1997). The etiology of polyuria may be facilitated by obtaining a urinalysis; a measurement of urine pH; and measurements of electrolytes, creatinine, osmolality, glucose, urea nitrogen, and bicarbonate, preferably in a timed urine collection together with the corresponding serum values. Such assessment may serve to prevent dehydration, acid-base disturbances, hypokalemia, or hypernatremia, which often accompany such polyuric disorders (Table 5-2) (Oster et al., 1997). Proper correction of acute hypernatremia is needed to prevent brain demyelination. Normal saline infusion may be the agent of choice in polyuric conditions associated with solute diuresis, whereas ADH and electrolyte-free fluid administration may be appropriate in cases of “pure” water diuresis. The recommended rate of correction of hypernatremia is about 10 mEq/L per 24 hours, amounting to a fall in plasma osmolality of about 20 mOsm/kg H₂O per day (Adrogue and Madias, 2000b).

Renin-Angiotensin-Aldosterone System

The renin-angiotensin-aldosterone axis plays a key role in control of vascular tone, Na⁺ and K⁺ homeostasis, and, ultimately, circulatory volume and cardiovascular and renal function. Renin is an enzyme with a molecular weight of 40 kDa that is synthesized and stored in the juxtaglomerular apparatus surrounding the afferent arterioles of the glomeruli (Davis and Freeman, 1976). The primary stimuli for renal renin release are reductions in renal-perfusion pressure, Na⁺ restriction, and Na⁺ loss as detected by the specialized macula densa cells located in the distal tubule. Mechanical (stretch of the afferent glomerular arterioles), neural (sympathetic nervous system), and hormonal (PGE₂ and prostacyclin) stimuli act in an integrated fashion to regulate the rate of renin secretion (Fig. 5-3).

FIGURE 5-3 Effects of decreased intravascular volume on the renin-angiotensin-aldosterone system.

Once released into the circulation, renin cleaves the leucine-valine bond of angiotensinogen, forming angiotensin I. Angiotensin-converting enzyme that is present in the lungs, as well as in the kidneys, large caliber vessels, and other tissues, cleaves the carboxyl terminal (histidine-leucine dipeptide) from angiotensin I to form the biologically active angiotensin II (Ng and Vane, 1967).

Angiotensin II has numerous important hemodynamic functions that are mediated largely by binding to angiotensin-II T1-receptors in endothelial cells, tubular epithelial cells, and smooth muscle (Box 5-1) (Burnier and Brunner, 2000). It plays a key role in regulating blood volume and long-term blood pressure through stimulation of several tubular transporters of Na⁺-conversation that are mainly located in the proximal tubule, as well as through its effects in enhancing aldosterone secretion and Na⁺ reabsorption in the distal tubule. As a potent direct smooth-muscle vasoconstrictor and as an enhancer of ADH and sympathetic nervous system activity, angiotensin II also participates in short-term blood-pressure regulation in disorders associated with volume depletion or circulatory depression. Research has uncovered multiple nonhemodynamic functions that are primarily mediated by binding to T1 receptors of angiotensin II, which are particularly important in the pathophysiology of progressive renal injury (Hall et al., 1999).

A rise in plasma aldosterone concentration stimulates urinary K⁺ secretion, thus allowing maintenance of K⁺ balance. Aldosterone also increases the excretion of ammonium (NH⁴⁺) and magnesium (Mg²⁺) and increases the absorption of Na⁺ in the distal tubule, both by increasing the permeability of the apical membrane and by increasing the activity of Na⁺, K⁺-adenosine triphosphatase (ATPase) (Marver and Kokko, 1983). The net effect is to generate more negative potential in the lumen, a driving force for increased K⁺ secretion. In addition, aldosterone enhances reabsorption of sodium in the cortical collecting duct through activation of the epithelial sodium-specific channel, ENaC (Greger, 2000). In performing these functions, aldosterone plays a key role in regulating fluid and electrolyte balance. Long-term aldosterone administration to healthy volunteers increases the ECF volume. Clinical edema does not occur, however, because after several days the kidneys “escape” from the Na⁺-retaining effect while maintaining the K⁺-secretory effect (August et al., 1958).

The Kidneys and Atrial Natriuretic Peptide

ANP is secreted by atrial monocytes in response to local stretching of the atrial wall in cases of hypervolemia (e.g., congestive heart failure or renal failure) and ultimately results in the reduction of intravascular volume and systemic blood pressure (Brenner et al., 1990). In the kidneys, ANP acts in the medullary collecting duct to inhibit sodium reabsorption during ECF expansion. ANP induces hyperfiltration, natriuresis, and suppression of renin release, and it inhibits receptor-mediated aldosterone biosynthesis (Greger, 2000). In the cardiovascular system, it diminishes cardiac output and stroke volume and reduces peripheral vascular resistance. Some of these effects are mediated through the influence of ANP on vagal and sympathetic nerve activity.

Body Fluid Compartments

The internal environment of the body consists of fluids contained within compartments. Water accounts for 50% to 80% of the human body by weight. The variation in water content depends on tissue type: adipose tissue contains only 10% water, whereas muscle contains 75% water. Total body water (TBW) decreases with age, mainly as a result of loss of water in ECF. For clinical purposes, TBW is estimated at 60% of body weight in infants older than age 6 months, as well as in children and adolescents. This value is very inaccurate for low–birth-weight premature infants in whom TBW comprises as much as 80% of total body weight (Friis-Hensen, 1971; Kagan et al., 1972). In term infants younger than 6 months of age, TBW may be approximated as 75% of total body weight (Hill, 1990). Newer formulas that consider the height (cm) and weight (kg), but not the degree of adiposity or the child’s surface area, have improved the estimation of TBW, particularly in healthy children between 3 months and 13 years of age (Fig. 5-4) (Mellits and Cheek, 1970; Morgenstern, 2002). TBW can be determined as follows:

FIGURE 5-4 Total body water (TBW) plotted against the parameter (Ht × Wt) for children from 3 months to 13 years of age. The 10th, 50th, and 90th percentile curves, generated from the equations in the text, are shown. The curves for both males and females are presented.

(From Morgenstern BZ, Mahoney DW, Warady BA: Estimating total body water in children on the basis of height and weight: a reevaluation of the formulas of Mellits and Cheek, J Am Soc Nephrol 13:1884-1888, 2002.)

Parenteral and Oral Fluids and Electrolytes

Except for the first 3 postnatal days when full-term neonates require only 40 to 60 mL/kg fluid per day, in general, 100 mL of water is needed for each 100 kcal expended. Notably, an additional 15 mL of water is generated endogenously for each 100 kcal used (water of oxidation), which is also available for body functions. In preterm infants, fluid intake may be gradually increased to 150 mL/kg per day, whereas 100 to 125 mL/kg per day generally suffices for infants weighing less than 10 kg. The fluid requirement decreases to 50 mL/kg per day for those weighing 11 to 20 kg and to 20 mL/kg per day for those with body weights above 20 kg. These fluid volumes are sufficient to allow excretion of dietary solute load, as well as to replace insensible fluid loss through the skin, lungs, and intestines (Table 5-5) (Winters, 1982). It should be noted that energy expenditure and, therefore, fluid intake may be significantly increased with stress (Table 5-6) (Holliday et al., 1994).

TABLE 5-5 Normal Losses and Maintenance Requirements for Fluid, Electrolytes, and Dextrose in Infants and Children

Adapted from Winters RW: Principles of pediatric fluid therapy, ed 2. Boston, 1982, Little, Brown.

The high fractional excretion of Na⁺ (FE_Na+) in premature infants can lead to negative Na⁺ balance, hyponatremia, neurologic disturbances, and poor growth unless an Na⁺ intake of 3 to 5 mmol/kg per day is given; in full-term infants and older children, 2 to 3 mmol/kg per day is sufficient (Drukker et al., 1980). Premature infants have a lower renal threshold for bicarbonate. In addition, several functional and anatomic factors combine to limit tubular excretion of weak organic acids (Avner et al., 1990). Consequently, premature infants may need small supplements of base. Sodium bicarbonate at 1 to 2 mmol/kg per day is generally recommended for the very small premature infant. Clinically important disturbances in acid-base status are unusual in full-term neonates unless they consume excessive amounts of protein.

Assessment of Dehydration

Assessment of the extent and type of dehydration is important for formulating a therapeutic strategy. Table 5-7 provides guidelines for the clinical assessment of the severity of dehydration in children (Ellis and Avner, 1985). Laboratory measurements should include hematocrit, blood gases, glucose, calcium, blood urea nitrogen, and albumin, as well as serum and urinary creatinine, osmolality, and electrolytes. A urinalysis should be done to detect cellular elements and to measure specific gravity. Urinary osmolality and specific gravity are of minimal value in assessing dehydration in premature infants with tubular immaturity and in infants with reduced urinary concentrating ability caused by low protein intake. In general, however, such data, together with a careful medical history, physical examination, and assessment of fluid input and loss, aid in the diagnosis of dehydration and guide adjustment of the amount and composition of fluid administration during various phases of therapy.

Treatment of Dehydration

Severely dehydrated infants and children should be cared for in the intensive care unit with constant monitoring of central venous pressure and serial measurements of the initial laboratory studies. Box 5-2 shows a stepwise approach to the treatment of isotonic, hypotonic, and hypertonic dehydration (Ellis and Avner, 1985). Particular attention should be given to hypernatremic dehydration and brain injury. The initial measures for fluid resuscitation are to stabilize the vital signs by administering crystalloid or colloid solutions and to correct severe acid-base imbalance, hypoglycemia, and other metabolic disturbances. The objective of subsequent measures is to assess further the kind of dehydration and to plan a time course for administration of the appropriate fluid volume and chemical composition needed to correct previous deficits and to replace ongoing losses.

The composition of select parenteral and oral rehydration solutions are shown in Tables 5-8 and 5-9. In most infants and children receiving parenteral solutions for brief periods, the normal fluid and electrolyte needs can be easily satisfied. The caloric needs, however, are not readily met. It is customary to provide 5% dextrose in parenteral solutions. Although this concentration provides only a fraction of the optimal number of calories (20% of total kilocalories needed by infants younger than 1 year of age), it is sufficient to prevent ketosis. In less mature neonates, higher infusion rates of 5% dextrose generally suffice to maintain blood glucose concentrations between 50 to 90 mg/dL (Roy and Sinclair, 1975; Winters, 1982).

Provided that infants and children are less than 10% dehydrated and have minimal electrolyte abnormalities, a good level of consciousness, adequate bowel sounds, and absence of signs of hypovolemia, oral rehydration may be used to replace deficits and maintain fluid volume. Commercially available preparations, such as Pedialyte RS (Ross) with a Na⁺ content of 45 mmol/L may be used. In children with diarrhea in developing countries, the World Health Organization (WHO) has recommended the use of an inexpensive and effective oral rehydration solution consisting of 90 mmol/L Na⁺ and 111 mmol/L of glucose (total osmolarity, 311 mOsmol/L). However, glucose-based solutions with a lower osmolality may further optimize fluid and glucose-sodium coupled absorption in the small intestine (Hahn et al., 2001).

Perioperative Fluid Management of Premature and Full-Term Neonates

Guidelines for the intraoperative fluid and electrolyte management of premature and full-term neonates are largely based on available knowledge of renal physiology rather than on data obtained from clinical investigations. The physiology of the healthy neonate is influenced by the short tubular length and is characterized by immature reabsorption mechanisms, an activated renin-angiotensin-aldosterone system, and low circulating ADH concentrations (Avner et al., 1990; El-Dahr and Chevalier, 1990). Thus, healthy preterm neonates weighing less than 1300 g, or of fewer than 32 weeks’ gestation, have FE_Na+ rates that range from 8.2% to 2.1% between 28 and 32 weeks’ gestation, with further gradual decrease to less than 1% at term; such rates may increase to 15% with stress (Arant, 1978; Delgado et al., 2003). The high FE_Na+ in preterm infants is ascribed to decreased Na⁺ reabsorption in the proximal tubule, together with hyporesponsiveness of the distal tubule to aldosterone (Sulyok et al., 1979). When combined with a negative Na⁺ balance that results from inadequate Na⁺ supplementation as well as decreased sensitivity of the collecting duct to ADH, up to one third of such infants develop significant hyponatremia (Na⁺ < 130 mEq/L), often manifesting with neurologic disturbances during the first 6 weeks of life (Roy and Sinclair, 1975).

Both premature and term neonates have a limited capacity to excrete K⁺, possibly because of distal tubular insensitivity to aldosterone. Hence, baseline reference plasma K⁺ concentrations range from 3.9 to 5.9 mEq/L. Moreover, both preterm and term neonates are capable of producing maximally diluted urine while concentrating capacity is limited. Yet, hyponatremia may develop after administration of large volumes of hypotonic fluids, because fluid excretion may be limited mainly because of low GFR. Stress may cause profound reduction in GFR in premature and term neonates through release of various extrarenal vasoactive and hormonal substances that modify the response of “immature kidneys,” thereby further disturbing fluid and electrolyte homeostasis. The higher body content of water and the higher metabolic rate, as well as a propensity to metabolic acidosis and hypocalcemia in premature newborns, are other important factors in deciding the volume and composition of intraoperative fluids.

Such considerations support the avoidance of boluses of hypotonic fluids while keeping in mind the lower age- and size-appropriate circulatory pressures that may serve as the goal of fluid management. In the absence of the expected physiologic fluid loss, which may range from 5% to 15% of body weight during the first 3 days of postnatal life, fluid volume during this time period may be limited to 60 mL/kg per day, whereas blood pressure support may be sustained with small infusions (5 mL/kg) of 5% albumin or other blood products as needed. Beyond 3 days of life, maintenance fluid volume is gradually increased to 150 mL/kg per day. Deficits beyond the expected physiologic losses and ongoing losses and allowance for surgical trauma may be replaced by a similar fluid composition, but the volume replacement may be more gradual or less rapid than outlined for older infants and children (Table 5-10). Na⁺ bicarbonate and calcium may be supplemented, while K⁺ should be limited. Also, a higher glucose concentration is generally desirable in premature infants. A recommended fluid composition is 0.75 normal saline with 20 mEq sodium bicarbonate/L (total Na⁺ = 135 mmol/L) in 5% to 10% dextrose, as well as 20 mEq/L KCl if plasma K⁺ falls below 3.5 mEq/L. Close attention to change in body weight and urine output and serial measurements of plasma electrolytes are essential in guiding the perioperative fluid management of the sick premature and full-term neonate.

Fluid Management of Children Undergoing Renal Transplantation

The key goal of intraoperative management is to expand the circulatory volume and to maintain systemic blood pressure between the 90th and 95th percentiles for age, gender, and height percentile, so as to allow for adequate perfusion of the renal allograft (Update on the 1987 Task Force Report on High Blood Pressure in Children and Adolescents, 1996). An adult kidney may sequester up to 250 mL of blood, and in infants nearly 50% of the cardiac output may be directed toward perfusion of the allograft. To ensure adequate perfusion of the allograft, the anesthesiologist actually needs to maximize the circulatory volume of the recipient, mainly with crystalloid or packed cytomegalovirus-safe, leukocyte-poor red blood cells if hemoglobin is below 9 g/dL, while closely monitoring the central venous pressure (CVP) and systemic MABP during vessel anastomosis. Near the completion of the vascular anastomoses, 20% mannitol (0.5 to 1.0 g/kg) and intravenous furosemide, 1 mg/kg, may be given before the cross-clamps are released. Before cross-clamp release, CVP should be maintained at 8 to 12 cm H₂O, and the systolic blood pressure and MABP should be kept above 120 mm and 70 mm Hg, respectively. If the MABP is inadequate to achieve good renal perfusion of the adult kidneys, a constant dopamine infusion of up to 5 mcg/kg per minute may be started. Intraoperative blood gases may be monitored frequently, because clamping of the aorta and accumulation of lactic acid can result in metabolic acidosis and vasoconstriction. The critical goal is to obtain immediate allograft function; hypotension after cross-clamp release in an infant with inadequate circulatory volume and an underperfused allograft is a potential catastrophe.

Intravenous furosemide (1 to 2 mg/kg per dose), 25% salt-poor albumin (0.5 g/kg per dose), 20% mannitol (0.3 to 0.5 g/kg per dose), or 0.9 normal saline solution (10 mL/kg bolus) may be given to help promote urine output in the immediate postoperative period.

The volume and composition of intravenous fluid administered during the first 48 postoperative hours are essential to ensure continued renal function. The urine output often exceeds 5 mL/kg per hour. Thus, insensible losses are quantitatively less important. In children with such high urine output, the concentration of dextrose to 1% is routinely reduced to prevent hyperglycemia, which may compound osmotic diuresis because of high preoperative blood urea nitrogen concentrations. Urine output is replaced on a milliliter-for-milliliter basis. In infants and children with body weight below 30 kg, the CVP should be maintained in the range of 5 to 10 cm H₂O, and the MABP should be at greater than 70 mm Hg. One fluid solution that may be used during the first 24 to 48 hours consists of 1% dextrose, 0.45 mEq/L NaCl solution, and 10 to 20 mEq sodium bicarbonate per liter. During the first 24 to 36 hours, additional fluid boluses of 10 mL/kg of normal saline or a 5% albumin solution may be given if CVP falls below 5 cm H₂O, with the goal of maintaining a urine output above certain arbitrary limits (5, 4, and 3 mL/kg per hour for body weight <10 kg, <20 kg, and <30 kg, respectively). In conjunction with such fluid boluses, intravenous furosemide (1 mg/kg) is also administered because renal allografts tend to be dependent on diuretics in the early perioperative setting.

Serum electrolytes (Na⁺, K⁺, Cl⁻, HCO₃⁻, Ca²⁺, P²⁻, Mg²⁺) may be monitored at 8- to 12-hour intervals during the first 2 postoperative days. Potassium chloride is given separately as needed when the plasma K⁺ falls below 3.5 mEq/L. In infants and young children, close monitoring of fluid balance and cardiovascular examination are essential to prevent electrolyte imbalance and fluid overload, which may result in severe hypertension or pulmonary edema, or, in reduction of intravascular volume and acute tubular necrosis (ATN). A bladder catheter inserted intraoperatively is necessary for accurate measurement of urine volume. Unless there are specific urologic indications, the catheter is removed after 4 to 5 days.

Immediate measures must be undertaken to improve postoperative oliguria. Beside the most easily correctable causes of oliguria, such as hypovolemia or a malfunctioning catheter, other potential causes include vascular bleeding or occlusion, ATN caused by prolonged cold ischemia storage, hyperacute rejection, or urinary extravasation or obstruction. In patients with oxalosis, precipitation of calcium oxalate crystals in the graft may cause acute allograft failure. Children with delayed graft function, congestive heart failure, or marked electrolyte abnormalities may require removal of fluid by hemodialysis or peritoneal dialysis. Fluid removal should be performed cautiously to avoid allograft hypoperfusion.

Hyponatremia (Plasma NA⁺ <135 mmol/L)

In infants and children, hyponatremia occurs much more often than hypernatremia. Although premature and full-term infants are capable of producing hypotonic urine, large, hypotonic fluid loads cannot be excreted, especially during the first 6 weeks of postnatal life. This is most evident during the first week of life, when only 10% to 50%, rather than 80%, of an intravenous challenge of 5% dextrose in water is excreted within 4 hours. The major factor limiting the response to a fluid challenge, especially in preterm infants during the first 5 weeks of postnatal life, is the physiologically low GFR and low urinary flow rate (Svenningsen and Aronson, 1974; Leake et al., 1976). Moreover, high urinary Na⁺ excretion and negative Na⁺ balance may contribute to the hyponatremia found in about one third of low–birth-weight premature infants (Engelke et al., 1978). This Na⁺ wasting has been attributed to deficient proximal and distal tubular reabsorption of Na⁺ in such infants (Sulyok et al., 1979).

A positive water balance, rather than Na⁺ wasting, has also been implicated in the hyponatremia of healthy premature infants (Rees et al., 1984; Sulyok et al., 1985). Low serum albumin concentrations and reduced plasma oncotic pressure may also bring about fluid retention and “late hyponatremia” in preterm infants (Menon et al., 1986). Thus, independent of the specific pathologic process, preterm infants are at high risk of developing hyponatremia. Therefore, electrolytes should be monitored frequently during the first 4 to 6 weeks of life, especially in infants of fewer than 34 weeks’ gestation. Other causes of hyponatremia in neonates are shown in Table 5-11. In older infants and children, hyponatremia may occur with dehydration, edema-forming states, and syndrome of inappropriate secretion of antidiuretic hormone (SIADH). Such conditions may be differentiated clinically at the bedside. In addition, the simple laboratory studies described in the section on dehydration may reveal urinary hypotonicity, suggesting water intoxication or the dilutional hyponatremia associated with renal failure.

TABLE 5-11 Causes of Hyponatremia

FE_Na+, Excreted fraction of filtered sodium; SIADH, syndrome of inappropriate antidiuretic hormone; TmG, tubular maximum for glucose reabsorption.

Surgery and anesthesia stimulate ADH release. This may persist for 2 or more days and may result in acute hypotonic hyponatremia, particularly when hypotonic fluids are also administered. Even isotonic fluid administration may not prevent hyponatremia caused by ADH release, because such individuals often excrete urine with sodium and potassium concentrations higher than plasma. This is partly because of high ANP concentrations that may coexist in this setting. The syndrome of SIADH is discussed in a later section of this chapter.

Several causes of hyponatremia require special emphasis. Pseudohyponatremia is associated with normal plasma osmolality and occurs in the setting of severe hyperlipidemia, hyperproteinemia, or disorders in which solutes other than sodium such as glucose, mannitol, or sorbitol result in high plasma osmolality. Administration of thiazide diuretics to individuals with cardiac failure may impair distal tubular dilution of urine and the ability to excrete a water load. Finally, large amounts of hypotonic fluid intake during prolonged exercise can cause acute symptomatic hyponatremia and noncardiogenic pulmonary edema, particularly under high ambient temperatures when sweat sodium and chloride are high or when nonsteroidal antiinflammatory agents are administered.

In hyponatremic patients who require parenteral fluids, the decision to correct the plasma Na⁺ level may be based both on clinical symptoms and signs and on the rapidity with which the disorder developed. Children are especially prone to neurologic symptoms (Laureno and Karp, 1997; Lauriat and Berl, 1997, Albanese et al., 2001). In clinical practice, the most important prophylactic measure for preventing hyponatremia is to avoid the infusion of hypotonic fluids. Common symptoms and signs of hyponatremia include headache, fatigue, nausea, and vomiting; seizures and respiratory arrest are more severe and sometimes delayed manifestations. Guidelines for replacing such sodium deficits may be based on the following formula:

For example, the amount of 3% saline solution needed to raise the plasma Na⁺ to 125 mmol/L in a 10-kg infant with a plasma Na⁺ concentration of 115 mmol/L (assuming that TBW is 65% of body weight) may be calculated as follows:

In asymptomatic children, the Na⁺ deficit of 65 mmol may be replaced by 422 mL of normal saline (154 mmol/L) given over 24 hours; symptomatic children may receive 127 mL of 3% saline solution (513 mmol/L) given at a rate of about 5 mL/kg per hour, that is, 2.5 mmol/kg per hour

With chronic hyponatremia (over 48 hours in duration), adaptive increases in neuronal osmolytes (glutamine, taurine, phosphocreatinine, myoinositol) diminish cellular uptake of water, thereby preventing brain edema (Gullans and Verbalis, 1993). Thus, in contrast to acute hyponatremia, in which brain edema combined with noncardiogenic pulmonary edema and hypoxia can disrupt neuronal function, such events are uncommon with chronic hyponatremia. Edema-forming disorders such as nephrosis, liver failure, or congestive heart failure are commonly associated with chronic hyponatremia in children. Although physiologic mechanisms stimulate both sodium and fluid retention aimed at preventing hypovolemia, hypovolemic stimuli for ADH release result in free-fluid retention and hyponatremia. Treatment includes the possible correction of the primary disorder, the elimination of diuretics and other offending agents, and limitation of electrolyte-free water intake. Because most individuals with chronic hyponatremia are asymptomatic, and because slow recovery of brain osmolytes coupled with iatrogenic correction of chronic hyponatremia can result in fatal or serious pontine and extrapontine myelinolysis, the rate of correction of serum sodium should be slow (about 0.3 mEq/L per hour) (Ayus et al., 1987; Adrogue and Madias, 2000a).

Hypernatremia (Plasma NA⁺ >145 mmol/L)

The causes of hypernatremia in infants and children are listed in Box 5-4. Hypernatremia commonly results from excessive water loss and inadequate water intake and occurs most often in individuals who are unable to communicate or satisfy their own thirst by accessing water. Thus, infants and debilitated individuals of any age are particularly susceptible to this disorder. A primary lack of thirst sensation is a rare cause of this disorder in children. The condition has been caused by improper mixing of formulas and is also increasingly reported with inadequate breastfeeding (Manganaro et al., 2001; Oddie et al., 2001). It is also increasingly recognized as a complication in hospitalized individuals who are very ill and have edema in association with renal failure, heart failure, hypotension, or liver failure, resulting in impaired sodium excretion and sodium overload (Kahn, 1999; Adrogue and Madias, 2000b). Administration of isotonic saline to maintain systemic blood pressure, together with associated hyperglycemia, may promote hypernatremia in these settings. Also, the deliberate use of hypertonic saline for the treatment of brain edema has occasionally resulted in severe hypernatremia (Peterson et al., 2000).

Premature infants and full-term neonates are also prone to hypernatremia because, in addition to an inability to excrete a water load, they are unable to excrete a large solute load (Aperia et al., 1975a, 1975b, 1977). The renal response to an Na⁺ load improves gradually, so that by the end of the first year of life, Na⁺ excretion reaches maximal levels of 16 mmol/hr per 1.73 m² (Aperia et al., 1975b). The limited ability to excrete an Na⁺ load appears to result from a reduced GFR and tubular inability to significantly increase the FE_Na+ because of the effect of aldosterone in increasing distal tubular Na⁺ reabsorption.

Signs and symptoms associated with hypernatremia in infants include muscle weakness, hyperpnea, apnea, bradycardia, restlessness, a high-pitched cry, lethargy, insomnia or coma, and muscular hypertonicity (Finberg and Harrison, 1955). Older children may exhibit thirst, lethargy, confusion, muscle irritability, rhabdomyolysis, respiratory arrest, seizures, or coma. Tachycardia and hypotension are symptoms of hypovolemia, which is an ominous sign suggestive of extreme dehydration. Because of hypertonicity, fluid shift from the intracellular compartment may result in brain shrinkage, subarachnoid hemorrhage, and permanent brain injury when chronic adaptive solute gain fails to maintain cell volume. Even with such correction, the ensuing neuronal hyperosmolality may predispose a patient to cerebral edema and serious neurologic consequences when rehydration with hypotonic fluid is used aggressively.

The initial laboratory investigation of hypernatremia is similar to that noted in the section on dehydration. In the absence of dehydration, the treatment of hypernatremia depends largely on the underlying disorder. The judicious administration of insulin and avoidance of colloids may be needed in patients with hyperosmolar nonketotic hyperglycemic coma. On the other hand, the administration of free water, together with appropriate replacement of arginine vasopressin, is useful for the treatment of central diabetes insipidus. Surgery may be used to treat several endocrinopathic conditions, whereas thiazides and a diet low in osmotic activity may be of benefit in nephrogenic diabetes insipidus.

In children with acute hypernatremia resulting from pure water loss, such as those with nephrogenic diabetes insipidus, a hypotonic fluid may be given at a rate that decreases plasma osmolality by no more than 2 mOsmol/kg per hour. The rate can be calculated using the following formula:

For example, to calculate the water deficit in a 10-kg infant with diabetes insipidus, a current serum-Na⁺ concentration of 160 mmol/L, and a desired plasma-Na⁺ concentration of 140 mmol/L, the following applies:

In children with hypotonic sodium loss, this fluid deficit, together with ongoing fluid losses that occurring during replacement of such a deficit, may be infused as 1/8 or ¼ saline with 1% to 2% dextrose

With chronic hypernatremia, correction of plasma osmolality may occur at a rate of less than 1 mOsmol/kg per hour, or a reduction in serum sodium of less than 0.5 mEq/L. Hypotonic fluids given at a low rate are the most suitable for this purpose. In individuals with accidental sodium loading, furosemide combined with adequate replacement of urine volume with 5% dextrose in water, is usually effective unless renal failure is present, in which case dialysis may be of benefit.

Potassium Homeostasis

Total body K⁺ content correlates with body weight and height and depends on muscle mass (Pierson et al., 1974; Patrick, 1977). In a healthy 20-year-old adult, total body K⁺ approximates 58 mmol/kg, a value that decreases progressively with age as muscle mass decreases and body fat increases. In children, the total body K⁺ is 38 mmol/kg or less (Pierson et al., 1974). More than 90% of body K⁺ is intracellular, and most of that is in muscle tissue. Of the extracellular K⁺, only 1.4% is contained within the ECF, whereas the remaining 8.6% is contained in the bone matrix.

The daily need for K⁺ is about 2 mmol/100 kcal of expended energy. The daily intake from a standard Western diet is estimated to be 0.75 to 1.5 mmol/kg of body weight. Typically, approximately 90% of the K⁺ ingested each day is eliminated in the urine. Less than 15% is eliminated in the stool, while a negligible amount is lost through the skin. The amount of K⁺ eliminated in stool increases, however, with a significant degree of renal failure, reaching 34% of dietary intake at a GFR of less than 5 mL/min per 1.73 m² (Hayes et al., 1967).

The renal tubular mechanisms involved in K⁺ homeostasis have been extensively reviewed (Halperin and Kamel, 1998). Nearly 85% of the filtered K⁺ is reabsorbed in the proximal tubule and the ascending thick limb of the loop of Henle. The amount of K⁺ in the final urine depends on the amount of intake and on the tubular secretion of K⁺.

Two key hormones lower ECF K⁺ acutely through redistribution in various tissues. Insulin causes Na⁺ to enter and H⁺ to exit cells through the electroneutral Na⁺/H⁺ exchanger, and epinephrine (and β₂-adrenergic agonists) activates the Na⁺, K⁺-ATPase, which exports three Na⁺ ions for each two K⁺ ions that enter the cell. Although epinephrine possesses both α- and β-adrenergic properties, it first causes a hyperkalemic response (during the first 1 to 3 minutes) and then a sustained decrease in plasma K⁺ concentration through trapping of K⁺ in cells by maintenance of an intracellular net negative charge that is essential in determining the electronegative resting membrane potential. By contrast, α-adrenergic agonists raise plasma K⁺ concentration by modifying muscle K⁺ uptake (Rosa et al., 1980).

Chronic K⁺ homeostasis is largely regulated by secretion of K⁺ by principal cells that predominate in the renal cortical collecting duct and are located in smaller numbers in the connecting tubule. The main mechanisms of K⁺ secretion are depicted in Figure 5-5. Secretion is aided by the following factors:

FIGURE 5-5 Major factors that regulate K⁺ secretion in principal cells. Sodium is reabsorbed across the luminal membrane through ENaC Na⁺ channels, with the resultant cellular depolarization increasing the electrical driving force for K⁺ secretion through ROMK K⁺ channels. 1, Elevation of peritubular [K⁺] (oval arrowheads) increases the density of luminal ENaC and ROMK channels, which promote both K⁺ secretion by increasing the electrical driving force and K⁺ permeability, respectively. Increases in peritubular [K⁺] also activate the Na⁺,K⁺-ATPase pump in the basolateral membrane and stimulate aldosterone release. 2, Aldosterone (diamond arrowheads) increases the density of ENaC (but not ROMK) channels and activates the Na⁺,K⁺-ATPase pump, both of which increase the driving force for K⁺ secretion. The surface area of the BLM containing the Na⁺, K⁺-ATPase pump undergoes amplification during prolonged exposure to either increased peritubular [K⁺] or aldosterone. 3, Kaliuretic factors, including K⁺ itself, have been proposed to somehow directly increase K⁺ secretion. For example, high luminal [K⁺] may directly increase the activity of ROMK channels.

(From Gennari FJ, Segal AS: Hyperkalemia: An adaptive response in chronic renal insufficiency, Kidney Int 62:1-9, 2002.)

Understanding such mechanisms provides physiologic explanations for plasma K⁺ alterations and for therapeutic rationale. An example is the individual with hypovolemia and hyperkalemia. The kidneys guard against extracellular volume contraction by raising angiotensin II levels. The latter increases bicarbonate reabsorption in the proximal and distal tubules such that the cortical collecting duct lumen becomes less electronegative and favors NaCl retention but reduced K⁺ secretion despite elevated aldosterone levels. In this clinical setting, low urinary flow rate also contributes to hyperkalemia. This is in contrast to individuals with euvolemia or hypervolemia, in whom a high plasma K⁺ concentration inhibits proximal tubular bicarbonate reabsorption while it directly stimulates aldosterone release, thereby promoting K⁺ secretion (along with Cl⁻ and HCO₃⁻) in the cortical collecting duct. In this setting Na⁺ is reabsorbed through the important effect of aldosterone in activating the specific ENaC at the apical membrane of principal cells (Halperin and Kamel, 1998). When ENaC is blocked by trimethoprim, amiloride, or triamterene, kaliuresis is inhibited.

Hypokalemia (Plasma K+ <3.5 mmol/L or <3.5 mEq/L)

Although hypokalemia usually implies total body K⁺ depletion, it can also be caused by transcellular shifts of K⁺ without extrarenal losses. Several classifications of hypokalemia have been devised according to whether the condition is acute or chronic, with or without K⁺ shift, renal or extrarenal. One classification is shown in Box 5-5. Because the etiology of renal K⁺ wasting is extensive, it may be facilitated by further subclassification on the basis of systemic blood pressure (Box 5-6).

Manifestations of Potassium Depletion

Hypokalemia results in multiple biochemical and neurophysiologic disturbances (Box 5-7) (Weiner and Wingo, 1997). Chronic hypokalemia and moderate degrees of acute K⁺ depletion (5% to 10% of total body K⁺) are generally well tolerated. More profound deficits result in clinical manifestations that are independent of the underlying cause of hypokalemia.

Box 5-7 Pathophysiologic Consequences of Hypokalemia

Neuromuscular

Peripheral nerves: Paresthesias

Skeletal muscle: Fatigue, weakness, cramps, flaccid paralysis, rhabdomyolysis, myoglobinuria

Smooth muscle: Paralytic ileus, increased vascular pressor resistance

Renal

Concentration defect: Polyuria, nocturia

Sodium retention

Increased ammonia production, enhanced bicarbonate reabsorption

Reduced renal blood flow, decreased GFR

Predisposition to urinary tract infection

Interstitial fibrosis

Cyst formation

Metabolic

Metabolic alkalosis

Impaired hepatic glycogen storage

Impaired protein metabolism

Insulin resistance

Increased plasma ammonia

Growth retardation

Hormonal

Impaired growth hormone release

Impaired insulin secretion

Decreased aldosterone secretion

Increased renin release

Increased synthesis of prostaglandins

Gill JR, Santos F, Chan JCM: Disorders of potassium metabolism. In Chan JCM, Gill JR, editors: Kidney electrolyte disorders, New York, 1990, Churchill Livingstone.

Biochemical consequences of hypokalemia include impairment in insulin release and insulin end-organ sensitivity, thereby increasing the risk of hyperglycemia or precipitation of frank diabetes mellitus (Rowe et al., 1980). Metabolic alkalosis results from direct stimulation of proximal tubular bicarbonate reabsorption and ammonia genesis by increased proton secretion via H⁺, K⁺-ATPase located in the collecting duct and by decreasing citrate excretion. Inadequate ADH response and increased synthesis of angiotensin II in the central nervous system may cause polyuria. Hypokalemia may raise plasma ammonia levels in patients with reduced hepatic function. Also, chronic renal K⁺ wasting may predispose to formation of renal cysts and interstitial fibrosis (Torres et al., 1990).

Many of the symptoms of acute hypokalemia relate to disturbed neuromuscular functions. Skeletal muscle weakness caused by cell hyperpolarization is the earliest manifestation of K⁺ depletion. Plasma K⁺ concentrations below 3.0 mEq/L lower the resting cell membrane potential and thereby increase the voltage needed to reach the threshold and initiate an action potential (Fig. 5-6). The symptoms include restless legs syndrome, fatigue, muscle cramps, paralysis, and rhabdomyolysis. Frank muscle necrosis may occur with serum K⁺ concentrations below 2.0 mEq/L (Knochel, 1982). Cardiac manifestations of hypokalemia include abnormalities in rhythm as a result of slowed repolarization (Helfant, 1986). The presence of electrocardiographic changes helps to exclude spurious hypokalemia and may aid the decision to manage the hypokalemia urgently (see Fig. 5-6). Abnormalities include depression of the ST segment, lower T-wave voltage, and appearance of U waves. Patients receiving cardiac glycosides are especially at risk of developing such abnormalities. Hypokalemia also impairs the cardiovascular responses to norepinephrine and angiotensin II.

FIGURE 5-6 Typical action potential profile. Normal threshold is influenced by Ca²⁺ concentrations, whereas K⁺ concentrations affect the resting potential.

(From Leaf A, Cotran RS: Renal pathophysiology, ed 2, New York, 1980, Oxford University Press.)

K⁺ depletion may lead to several functional and structural abnormalities in the kidneys, including reduced RBF and GFR, renal hypertrophy, tubuloepithelial dilation, vacuolization, and sclerosis (Relman and Schwartz, 1956). Functionally, patients develop urinary acidification and concentration defects and polyuria.

Hyperkalemia (Plasma K⁺ >5.5 mEq/L in Infants and Children, or >6 mmol/L in Neonates)

Conditions causing hyperkalemia are listed in Box 5-8. Hyperkalemia may result from a surprisingly small increase in total body K⁺ or may result rapidly from transcellular shift. Normally, the kidneys provide the crucial defense against slight elevations in serum K⁺ level. Thus, with a few exceptions, hyperkalemia commonly occurs in conditions characterized by decreased urine flow rates or marked reduction in GFR. In the absence of renal insufficiency, drugs are responsible for most cases of hyperkalemia. Several agents may cause hyperkalemia by increasing the K⁺ load, by facilitating transcellular K⁺ efflux, or by impairing renal excretion of K⁺ (Table 5-12) (Perazella, 2000).

TABLE 5-12 Medications That Can Cause Hyperkalemia and Their Mechanisms of Action

ACE, Angiotensin-converting enzyme.

Modified from Perazella MA: Drug-induced hyperkalemia: old culprits and new offenders, Am J Med 109:307–314, 2000.

Manifestations of Hyperkalemia

The clinical manifestations of hyperkalemia relate to interference with the electrophysiologic activities of muscle. Under the influence of hyperkalemia, the ratio of intracellular to extracellular K⁺ is decreased, resulting in delayed depolarization, hastened repolarization, and slow conduction velocity (Knochel, 1982). The most important effect involves the heart. Diagnostic electrocardiographic alterations include tenting or symmetric peaking of the T wave in the precordial leads and depression of the ST wave (Fig. 5-7). In severe hyperkalemia, there is widening of the QRS complex, lengthening of the PR interval, first-degree or second-degree heart block, disappearance of the P wave, and, finally, atrial standstill. Ventricular fibrillation or asystole follows the development of a sinusoid wave. Although the magnitudes of hyperkalemia and of cardiotoxicity correlate well, arrhythmias may develop with even mild hyperkalemia when other metabolic abnormalities such as hyponatremia, acidosis, or calcium disorders coexist.

FIGURE 5-7 The relationship between plasma potassium concentration and electrocardiographic changes.

(From Winters RW: The body fluids in pediatrics, Boston, 1973, Little, Brown.)

Hyperkalemia affects electrical activities in noncardiac muscle as well. Such manifestations as paresthesias, weakness, and flaccid paralysis that spare the head and trunk are not rare.