Fluid Compartments

TBW is divided between 2 main compartments: intracellular fluid (ICF) and extracellular fluid (ECF). In the fetus and newborn, the ECF volume is larger than the ICF volume (see Web Fig. 52-1). The normal postnatal diuresis causes an immediate decrease in the ECF volume. This is followed by continued expansion of the ICF volume, which results from cellular growth. By 1 yr of age, the ratio of the ICF volume to the ECF volume approaches adult levels. The ECF volume is approximately 20-25% of body weight, and the ICF volume is approximately 30-40% of body weight, close to twice the ECF volume (Web Fig. 52-2). With puberty, the increased muscle mass of males causes them to have a higher ICF volume than females. There is no significant difference in the ECF volume between postpubertal females and males.

Web Figure 52-1 Total body water, intracellular fluid, and extracellular fluid as a percentage of body weight and a function of age.

(From Winters RW: Water and electrolyte regulation. In Winters RW, editor: The body fluids in pediatrics, Boston, 1973, Little, Brown.)

Body Content and Physiologic Function

Sodium is the dominant cation of the ECF (see Web Fig. 52-3), and it is the principal determinant of extracellular osmolality. Sodium is therefore necessary for the maintenance of intravascular volume. Less than 3% of sodium is intracellular. More than 40% of total body sodium is in bone; the remainder is in the interstitial and intravascular spaces. The low intracellular sodium concentration, approximately 10 mEq/L, is maintained by Na⁺,K⁺-ATPase, which exchanges intracellular sodium for extracellular potassium.

Intake

A child’s diet determines the amount of sodium ingested—a predominantly cultural determination in older children. An occasional child has salt craving because of an underlying salt-wasting renal disease or adrenal insufficiency. Children in the USA tend to have very high sodium intakes because their diets include a large amount of “junk” food or fast food. Infants receive sodium from breast milk (≈7 mEq/L) and formula (7-13 mEq/L, for 20 calorie/oz formula). Lower recommended sodium intake limits for adults and children are being considered by the U.S. Department of Health and Human Services. In 2005-2006, the average daily sodium intake in those aged ≥2 yr was 3436 mg/day. Sodium intake is now recommended not to exceed 2500 mg/day.

Sodium is readily absorbed throughout the gastrointestinal tract. Mineralocorticoids increase sodium transport into the body, although this effect has limited clinical significance. The presence of glucose enhances sodium absorption owing to the presence of a cotransport system. This is the rationale for including sodium and glucose in oral rehydration solutions (Chapter 332).

Excretion

Sodium excretion occurs in stool and sweat, but the kidney regulates sodium balance and is the principal site of sodium excretion. There is some sodium loss in stool, but it is minimal unless diarrhea is present. Normally, sweat has 5-40 mEq/L of sodium. Sweat sodium concentration is increased in children with cystic fibrosis, aldosterone deficiency, or pseudohypoaldosteronism. The higher sweat losses in these conditions may cause or contribute to sodium depletion.

Sodium is unique among electrolytes because water balance, not sodium balance, usually determines its concentration. When the sodium concentration increases, the resultant higher plasma osmolality causes increased thirst and increased secretion of ADH, which leads to renal conservation of water. Both of these mechanisms increase the water content of the body, and the sodium concentration returns to normal. During hyponatremia, the decrease in plasma osmolality stops ADH secretion, and consequent renal water excretion leads to an increase in the sodium concentration. Even though water balance is usually regulated by osmolality, volume depletion does stimulate thirst, ADH secretion, and renal conservation of water. Volume depletion takes precedence over osmolality; volume depletion stimulates ADH secretion, even if a patient has hyponatremia.

The excretion of sodium by the kidney is not regulated by the plasma osmolality. The patient’s effective plasma volume determines the amount of sodium in the urine. This is mediated by a variety of regulatory systems, including the renin-angiotensin-aldosterone system and intrarenal mechanisms. In hyponatremia or hypernatremia, the underlying pathophysiology determines the amount of urinary sodium, not the serum sodium concentration.

Etiology and Pathophysiology

There are 3 basic mechanisms of hypernatremia (Table 52-1). Sodium intoxication is frequently iatrogenic in a hospital setting as a result of correction of metabolic acidosis with sodium bicarbonate. Baking soda, a putative home remedy for upset stomach, is another source of sodium bicarbonate; the hypernatremia is accompanied by a profound metabolic alkalosis. In hyperaldosteronism, there is renal retention of sodium and resultant hypertension; the hypernatremia is usually mild.

The classic causes of hypernatremia from a water deficit are nephrogenic and central diabetes insipidus (Chapters 524 and 552). Hypernatremia develops in diabetes insipidus only if the patient does not have access to water or cannot drink adequately because of immaturity, neurologic impairment, emesis, or anorexia. Infants are at high risk because of their inability to control their own water intake. Central diabetes insipidus and the genetic forms of nephrogenic diabetes insipidus typically cause massive urinary water losses and very dilute urine. The water losses are less dramatic, and the urine often has the same osmolality as plasma when nephrogenic diabetes insipidus is secondary to disease (obstructive uropathy, renal dysplasia, sickle cell disease).

The other causes of a water deficit are also secondary to an imbalance between losses and intake. Newborns, especially if premature, have high insensible water losses. Losses are further increased if the infant is placed under a radiant warmer or with the use of phototherapy for hyperbilirubinemia. The renal concentrating mechanisms are not optimal at birth, providing an additional source of water loss. Ineffective breast-feeding, often in a primiparous mother, can cause severe hypernatremic dehydration. Adipsia, the absence of thirst, is usually secondary to damage to the hypothalamus, such as from trauma, tumor, hydrocephalus, or histiocytosis. Primary adipsia is rare.

When hypernatremia occurs in conditions with deficits of sodium and water, the water deficit exceeds the sodium deficit. This occurs only if the patient is unable to ingest adequate water. Diarrhea results in depletion of both sodium and water. Because diarrhea is hypotonic—typical sodium concentration of 35-65 mEq/L—water losses exceed sodium losses, potentially leading to hypernatremia. Most children with gastroenteritis do not have hypernatremia because they drink enough hypotonic fluid to compensate for stool water losses (Chapter 332). Fluids such as water, juice, and formula are more hypotonic than the stool losses, allowing correction of the water deficit and potentially even causing hyponatremia. Hypernatremia is most likely to occur in the child with diarrhea who has inadequate intake because of emesis, lack of access to water, or anorexia.

Osmotic agents, including mannitol, and glucose in diabetes mellitus, cause excessive renal losses of water and sodium. Because the urine is hypotonic—sodium concentration of approximately 50 mEq/L—during an osmotic diuresis, water loss exceeds sodium loss, and hypernatremia may occur if water intake is inadequate. Certain chronic kidney diseases, such as renal dysplasia and obstructive uropathy, are associated with tubular dysfunction, leading to excessive losses of water and sodium. Many children with such diseases have disproportionate water loss and are at risk for hypernatremic dehydration, especially if gastroenteritis supervenes. Similar mechanisms occur during the polyuric phase of acute tubular necrosis and after relief of urinary obstruction (postobstructive diuresis). Patients with either condition may have an osmotic diuresis from urinary losses of urea and an inability to conserve water because of tubular dysfunction.

Clinical Manifestations

Most children with hypernatremia are dehydrated and show the typical clinical signs and symptoms (Chapter 54). Children with hypernatremic dehydration tend to have better preservation of intravascular volume because of the shift of water from the intracellular space to the extracellular space. This shift maintains blood pressure and urine output and allows hypernatremic infants to be less symptomatic initially and potentially to become more dehydrated before medical attention is sought. Breast-fed infants with hypernatremia are often profoundly dehydrated, with failure to thrive. Probably because of intracellular water loss, the pinched abdominal skin of a dehydrated, hypernatremic infant has a “doughy” feel.

Hypernatremia, even without dehydration, causes central nervous system (CNS) symptoms that tend to parallel the degree of sodium elevation and the acuity of the increase. Patients are irritable, restless, weak, and lethargic. Some infants have a high-pitched cry and hyperpnea. Alert patients are very thirsty, even though nausea may be present. Hypernatremia may cause fever, although many patients have an underlying process that contributes to the fever. Hypernatremia is associated with hyperglycemia and mild hypocalcemia; the mechanisms are unknown. Beyond the sequelae of dehydration, there is no clear direct effect of hypernatremia on other organs or tissues, except the brain.

Brain hemorrhage is the most devastating consequence of hypernatremia. As the extracellular osmolality increases, water moves out of brain cells, leading to a decrease in brain volume. This decrease can result in tearing of intracerebral veins and bridging blood vessels as the brain moves away from the skull and the meninges. Patients may have subarachnoid, subdural, and parenchymal hemorrhages. Seizures and coma are possible sequelae of the hemorrhage, although seizures are more common during correction of hypernatremia. The cerebrospinal fluid (CSF) protein is often elevated in infants with significant hypernatremia, probably owing to leakage from damaged blood vessels. Neonates, especially if premature, seem especially vulnerable to hypernatremia and excessive sodium intake. There is an association between rapid or hyperosmolar sodium bicarbonate administration and the development of intraventricular hemorrhages in neonates. Even though central pontine myelinolysis (CPM) is classically associated with overly rapid correction of hyponatremia, both CPM and extrapontine myelinolysis can occur in children with hypernatremia. Thrombotic complications occur in severe hypernatremic dehydration; they include stroke, dural sinus thrombosis, peripheral thrombosis, and renal vein thrombosis. This is secondary to dehydration and possibly hypercoagulability associated with hypernatremia.

Diagnosis

The etiology of hypernatremia is usually apparent from the history. Hypernatremia resulting from water loss occurs only if the patient does not have access to water or is unable to drink. In the absence of dehydration, it is important to ask about sodium intake. Children with excess sodium intake do not have signs of dehydration, unless another process is present. Severe sodium intoxication causes signs of volume overload, such as pulmonary edema and weight gain. Salt poisoning is associated with an elevated fractional excretion of sodium, whereas hypernatremic dehydration causes a low fractional excretion of sodium. In hyperaldosteronism, hypernatremia is usually mild or absent and is associated with edema, hypertension, hypokalemia, and metabolic alkalosis.

When there is isolated water loss, the signs of volume depletion are usually less severe initially because much of the loss is from the intracellular space. When pure water loss causes signs of dehydration, the hypernatremia and water deficit are usually severe. In the child with renal water loss, either central or nephrogenic diabetes insipidus, the urine is inappropriately dilute and urine volume is not low. The urine is maximally concentrated and urine volume is low if the losses are extrarenal or due to inadequate intake. With extrarenal causes of loss of water, the urine osmolality should be >1,000 mOsm/kg. When diabetes insipidus is suspected, the evaluation may include measurement of ADH and a water deprivation test, including a trial of desmopressin acetate (synthetic ADH analog) to differentiate between nephrogenic diabetes insipidus and central diabetes insipidus (Chapter 552.1). A water deprivation test is unnecessary if the patient has simultaneous documentation of hypernatremia and poorly concentrated urine (osmolality lower than that of plasma). In children with central diabetes insipidus, administration of desmopressin acetate increases the urine osmolality above the plasma osmolality, although maximum osmolality does not occur immediately because of the decreased osmolality of the renal medulla due to the chronic lack of ADH. In children with nephrogenic diabetes insipidus, there is no response to desmopressin acetate.

With combined sodium and water deficits, analysis of the urine differentiates between renal and nonrenal etiologies. When the losses are extrarenal, the kidney responds to volume depletion with low urine volume, concentrated urine, and sodium retention (urine sodium <20 mEq/L, fractional excretion of sodium <1%). With renal causes, the urine volume is not appropriately low, the urine is not maximally concentrated, and the urine sodium may be inappropriately elevated.

Treatment

As hypernatremia develops, the brain generates idiogenic osmoles to increase the intracellular osmolality and prevent the loss of brain water. This mechanism is not instantaneous and is most prominent when hypernatremia has developed gradually. If the serum sodium concentration is lowered rapidly, there is movement of water from the serum into the brain cells to equalize the osmolality in the 2 compartments (Fig. 52-1). The resultant brain swelling manifests as seizures or coma.

Figure 52-1 Mechanism of brain edema during correction of hypernatremia. A rapid decrease of the serum concentration during treatment of hypernatremia causes movement of water into brain cells, leading to cerebral edema. The presence of idiogenic osmoles in brain cells is responsible for the osmotic gradient.

Because of the associated dangers, hypernatremia should not be corrected rapidly. The goal is to decrease the serum sodium by <12 mEq/L every 24 hr, a rate of 0.5 mEq/L/hr. The most important component of correcting moderate or severe hypernatremia is frequent monitoring of the serum sodium value so that fluid therapy can be adjusted to provide adequate correction, neither too slow nor too fast. If a child has seizures as a result of brain edema secondary to rapid correction, administration of hypotonic fluid should be stopped. An infusion of 3% saline can acutely increase the serum sodium, reversing the cerebral edema.

In the child with hypernatremic dehydration, as in any child with dehydration, the first priority is restoration of intravascular volume with isotonic fluid (Chapter 54). Normal saline is preferable to lactated Ringer solution because the lower sodium concentration of the latter can cause the serum sodium to decrease too rapidly, especially if multiple fluid boluses are given. Repeated boluses of normal saline (10-20 mL/kg) may be required to treat hypotension, tachycardia, and signs of poor perfusion (peripheral pulses, capillary refill time) (Chapters 54 and 64).

The sodium concentration of the deficit replacement fluid, the rate of fluid administration, and the presence of continued water losses determine the rate of decrease of the sodium concentration. The following formula is often cited for calculating the water deficit:

This calculation is equivalent to 3-4 mL of water per kg for each 1 mEq that the current sodium level exceeds 145 mEq. The utility of such formulas has never been proven in clinical practice. Most patients with hypernatremic dehydration do well with a fluid sodium concentration of approximately half-normal saline, but with a fluid rate that is only 20-30% greater than maintenance fluid. Use of this concentration prevents excessive delivery of free water and too rapid a decrease in the serum sodium level. Patients with pure water loss may require a more hypotonic fluid (0.2 normal saline). Excessive water and sodium losses may also need to be replaced. If signs or symptoms of volume depletion develop, the patient receives additional boluses of isotonic saline. Monitoring of the rate of decrease of the serum sodium concentration permits adjustment in the rate and sodium concentration of the fluid that the patient is receiving, avoiding overly rapid correction of the hypernatremia. Many patients with mild to moderate hypernatremic dehydration due to gastroenteritis can be managed with oral rehydration (Chapter 332).

Acute, severe hypernatremia, usually secondary to sodium administration, can be corrected more rapidly because idiogenic osmoles have not had time to accumulate. This fact balances the high morbidity and mortality rates associated with hypernatremia with the dangers of overly rapid correction. When hypernatremia is severe and is due to sodium intoxication, it may be impossible to administer enough water to correct the hypernatremia rapidly without worsening the volume overload. In this situation, peritoneal dialysis allows for removal of the excess sodium. This requires dialysis fluid with a high glucose concentration and a low sodium concentration. In less severe cases, the addition of a loop diuretic increases the removal of excess sodium and water, decreasing the risk of volume overload. With sodium overload, hypernatremia is corrected with sodium-free intravenous fluid (5% dextrose in water [D5W]).

Hyperglycemia from hypernatremia is not usually a problem and is not treated with insulin because the acute decrease in glucose may precipitate cerebral edema by lowering plasma osmolality. Rarely, the glucose concentration of intravenous fluids must be reduced (from D5W to D2.5W). The secondary hypocalcemia is treated as needed.

It is important to address the underlying cause of the hypernatremia, if possible. The child with central diabetes insipidus should receive desmopressin acetate. Because this treatment reduces renal excretion of water, excessive intake of water must consequently be avoided to prevent both overly rapid correction of the hypernatremia and the development of hyponatremia. Over the long term, reduced sodium intake and the use of medications can somewhat ameliorate the water losses in nephrogenic diabetes insipidus (Chapter 524). The daily water intake of a child who is receiving tube feeding may need to be increased to compensate for high losses. The patient with significant ongoing losses, such as through diarrhea, may need supplemental water and electrolytes (Chapter 53). Sodium intake is reduced if it contributed to the hypernatremia.

Etiology and Pathophysiology

The causes of hyponatremia are listed in Table 52-2. Pseudohyponatremia is a laboratory artifact that is present when the plasma contains very high concentrations of protein (multiple myeloma, intravenous immunoglobulin infusion) or lipid (hypertriglyceridemia, hypercholesterolemia). It does not occur when a direct ion-selective electrode determines the sodium concentration in undiluted plasma, a technique that is used by the instruments used for measuring arterial blood gases. In true hyponatremia, the measured osmolality is low, whereas it is normal in pseudohyponatremia. Hyperosmolality, as may occur with hyperglycemia, causes a low serum sodium concentration because water moves down its osmotic gradient from the intracellular space into the extracellular space, diluting the sodium concentration. However, because the manifestations of hyponatremia are due to the low plasma osmolality, patients with hyponatremia resulting from hyperosmolality do not have symptoms of hyponatremia. When the etiology of the hyperosmolality resolves, such as hyperglycemia in diabetes mellitus, water moves back into the cells and the sodium concentration rises to its “true” value. Mannitol or sucrose, a component of intravenous immunoglobulin preparations, may cause hyponatremia due to hyperosmolality.

Classification of hyponatremia is based on the patient’s volume status. In hypovolemic hyponatremia, the child has lost sodium from the body. The water balance may be positive or negative, but sodium loss has been higher than water loss. The pathogenesis of the hyponatremia is usually a combination of sodium loss and water retention to compensate for the volume depletion. The patient has a pathologic increase in fluid loss, and this fluid contains sodium. Most fluid that is lost has a lower sodium concentration than that of plasma. Viral diarrhea fluid has, on average, a sodium concentration of 50 mEq/L. Replacing diarrhea fluid, which has a sodium concentration of 50 mEq/L, with formula, which has only approximately 10 mEq/L of sodium, reduces the sodium concentration. Intravascular volume depletion interferes with renal water excretion, the body’s usual mechanism for preventing hyponatremia. The volume depletion stimulates ADH synthesis, resulting in renal water retention. Volume depletion also decreases the GFR and enhances water resorption in the proximal tubule, thereby reducing water delivery to the collecting duct.

Diarrhea due to gastroenteritis is the most common cause of hypovolemic hyponatremia in children. Emesis causes hyponatremia if the patient takes in hypotonic fluid, either intravenously or enterally, despite the emesis. Most patients with emesis have either a normal sodium concentration or hypernatremia. Burns may cause massive losses of isotonic fluid and resultant volume depletion. Hyponatremia develops if the patient receives hypotonic fluid. Losses of sodium from sweat are especially high in children with cystic fibrosis, aldosterone deficiency, or pseudohypoaldosteronism, although high losses can occur simply in a hot climate. Third space losses are isotonic and can cause significant volume depletion, leading to ADH production and water retention, which can cause hyponatremia if the patient receives hypotonic fluid. In diseases that cause volume depletion through extrarenal sodium loss, the urine sodium level should be low (<10 mEq/L) as part of the renal response to maintain the intravascular volume. The only exceptions are diseases that cause both extrarenal and renal sodium losses: adrenal insufficiency and pseudohypoaldosteronism.

Renal sodium loss may occur in a variety of situations. In some situations, the urine sodium concentration is >140 mEq/L; thus, hyponatremia may occur without any fluid intake. In many cases, the urine sodium level is less than the serum concentration; thus, the intake of hypotonic fluid is necessary for hyponatremia to develop. In diseases associated with urinary sodium loss, the urine sodium level is >20 mEq/L despite volume depletion. This may not be true if the urinary sodium loss is no longer occurring, as is frequently the case if diuretics are discontinued. Because loop diuretics prevent generation of a maximally hypertonic renal medulla, the patient can neither maximally dilute nor concentrate the urine. The inability to maximally retain water provides some protection against severe hyponatremia. The patient receiving thiazide diuretics can concentrate the urine and is at higher risk for severe hyponatremia. Osmotic agents, such as glucose during diabetic ketoacidosis, cause loss of both water and sodium. Urea accumulates during renal failure and then acts as an osmotic diuretic after relief of urinary tract obstruction and during the polyuric phase of acute tubular necrosis. Transient tubular damage in these conditions further impairs sodium conservation. The serum sodium concentration in these conditions depends on the sodium concentration of the fluid used to replace the losses. Hyponatremia develops when the fluid is hypotonic relative to the urinary losses.

Renal salt wasting occurs in hereditary kidney diseases, such as juvenile nephronophthisis and autosomal recessive polycystic kidney disease. Obstructive uropathy, most commonly a consequence of posterior urethral valves, produces salt wasting, but patients with the disease may also have hypernatremia as a result of impaired ability to concentrate urine and high water loss. Acquired tubulointerstitial nephritis, usually secondary to either medications or infections, may cause salt wasting, along with other evidence of tubular dysfunction. CNS injury may produce cerebral salt wasting, which is theoretically due to the production of a natriuretic peptide that causes renal salt wasting. In type II renal tubular acidosis (RTA), usually associated with Fanconi syndrome (Chapter 523), there is increased excretion of sodium and bicarbonate in the urine. Patients with Fanconi syndrome also have glycosuria, aminoaciduria, and hypophosphatemia due to renal phosphate wasting.

Aldosterone is necessary for renal sodium retention and for the excretion of potassium and acid. In congenital adrenal hyperplasia due to 21-hydroxylase deficiency, the absence of aldosterone produces hyponatremia, hyperkalemia, and metabolic acidosis. In pseudohypoaldosteronism, aldosterone levels are elevated, but there is no response because of either a defective sodium channel or a lack of aldosterone receptors. A lack of tubular response to aldosterone may occur in children with urinary tract obstruction, especially during an acute urinary tract infection.

In hypervolemic hyponatremia, there is an excess of TBW and sodium, although the increase in water is greater than the increase in sodium. In most of the conditions that cause hypervolemic hyponatremia, there is a decrease in the effective blood volume, due to third space fluid loss, vasodilation, or poor cardiac output. The regulatory systems sense a decrease in effective blood volume and attempt to retain water and sodium to correct the problem. ADH causes renal water retention, and the kidney, under the influence of aldosterone and other intrarenal mechanisms, retains sodium. The patient’s sodium concentration decreases because water intake exceeds sodium intake and ADH prevents the normal loss of excess water.

In these disorders, there is a low urine sodium concentration (<10 mEq/L) and an excess of both TBW and sodium. The only exception is in patients with renal failure and hyponatremia. These patients have an expanded intravascular volume, and hyponatremia can therefore appropriately suppress ADH production. Water cannot be excreted because very little urine is being made. Serum sodium is diluted through ingestion of water. Because of renal dysfunction, the urine sodium concentration may be elevated, but urine volume is so low that urine sodium excretion has not kept up with sodium intake, leading to sodium overload. The urine sodium concentration in renal failure varies. In patients with acute glomerulonephritis, because it does not affect the tubules, the urine sodium level is usually low, whereas in patients with acute tubular necrosis, it is elevated because of tubular dysfunction.

Patients with hyponatremia and no evidence of volume overload or volume depletion have euvolemic hyponatremia. These patients typically have an excess of TBW and a slight decrease in total body sodium. Some of these patients have an increase in weight, implying that they are volume-overloaded. Nevertheless, from a clinical standpoint, they usually appear normal or have subtle signs of fluid overload.

In SIADH, the secretion of ADH is not inhibited by either low serum osmolality or expanded intravascular volume (Chapter 553). The result is that the child with SIADH is unable to excrete water. This results in dilution of the serum sodium and hyponatremia. The expansion of the extracellular volume due to the retained water causes a mild increase in intravascular volume. The kidney increases sodium excretion in an effort to decrease intravascular volume to normal; thus, the patients has a mild decrease in body sodium. SIADH most commonly occurs with disorders of the CNS (infection, hemorrhage, trauma, tumor, thrombosis), but lung disease (infection, asthma, positive pressure ventilation) and malignant tumors (producing ADH) are other potential causes. A variety of medications may cause SIADH, including recreational use of 3,4-methylenedioxymethylamphetamine (MDMA, or “Ecstasy”), opiates, antiepileptic drugs (carbamazepine, oxcarbamazepine, valproate), tricyclic antidepressants, vincristine, cytoxan, and selective serotonin reuptake inhibitors. The diagnosis of SIADH is one of exclusion, because other causes of hyponatremia must be eliminated (Table 52-3). Because SIADH is a state of intravascular volume expansion, low serum uric acid and BUN levels are supportive of the diagnosis.

A rare gain-of-function mutation in the renal ADH receptor causes nephrogenic syndrome of inappropriate antidiuresis. Patients with this X-linked disorder appear to have SIADH but have undetectable levels of ADH.

Hyponatremia in hospitalized patients is frequently due to inappropriate production of ADH and administration of hypotonic intravenous fluids. Causes of inappropriate ADH production include stress, medications such as narcotics or anesthetics, nausea, and respiratory illness. The synthetic analog of ADH, desmopressin acetate, causes water retention and may cause hyponatremia if fluid intake is not appropriately limited. The main uses of desmopressin acetate in children are for the management of central diabetes insipidus and of nocturnal enuresis.

Excess water ingestion can produce hyponatremia. In these cases, the sodium concentration decreases as a result of dilution. This decrease suppresses ADH secretion, and there is a marked water diuresis by the kidney. Hyponatremia develops only because the intake of water exceeds the kidney’s ability to eliminate water. This condition is more likely to occur in infants because their lower GFR limits their ability to excrete water. In some situations, the water intoxication causes acute hyponatremia and is due to a massive acute water load. Examples of causes of this water load include infant swimming lessons, inappropriate use of hypotonic intravenous fluids, water enemas, and forced water intake as a form of child abuse. Chronic hyponatremia occurs in children who receive water, but limited sodium and protein. The minimum urine osmolality is approximately 50 mOsm/kg, so the kidney can excrete 1 L of water only if there is enough solute ingested to produce 50 mOsm for urinary excretion. Because sodium and urea (a breakdown product of protein) are the principal urinary solutes, a lack of intake of sodium and protein prevents adequate water excretion. This occurs with the use of diluted formula or other inappropriate diets. Subsistence on beer, a poor source of sodium and protein, causes hyponatremia due to the inability to excrete the high water load (“beer potomania”).

The pathogenesis of the hyponatremia in glucocorticoid deficiency or hypothyroidism is incompletely understood. There is an inappropriate retention of water by the kidney, but the precise mechanisms are not clearly elucidated.

Clinical Manifestations

Hyponatremia causes a decrease in the osmolality of the extracellular space. Because the intracellular space then has a higher osmolality, water moves from the extracellular space to the intracellular space to maintain osmotic equilibrium. The increase in intracellular water causes cells to swell. Although cell swelling is not problematic in most tissues, it is dangerous for the brain, which is confined by the skull. As brain cells swell, there is an increase in intracranial pressure, which impairs cerebral blood flow. Acute, severe hyponatremia can cause brainstem herniation and apnea; respiratory support is often necessary. Brain cell swelling is responsible for most of the symptoms of hyponatremia. Neurologic symptoms of hyponatremia include anorexia, nausea, emesis, malaise, lethargy, confusion, agitation, headache, seizures, coma, and decreased reflexes. Patients may have hypothermia and Cheyne-Stokes respirations. Hyponatremia can cause muscle cramps and weakness; rhabdomyolysis can occur with water intoxication.

The symptoms of hyponatremia are mostly due to the decrease in extracellular osmolality and the resulting movement of water down its osmotic gradient into the intracellular space. Brain swelling can be significantly obviated if the hyponatremia develops gradually, because brain cells adapt to the decreased extracellular osmolality by reducing intracellular osmolality. This reduction is achieved by extrusion of the main intracellular ions (potassium and chloride) and a variety of small organic molecules. This process explains why the range of symptoms in hyponatremia is related to both the serum sodium level and its rate of decrease. A patient with chronic hyponatremia may have only subtle neurologic abnormalities with a serum sodium level of 110 mEq/L, but another patient may have seizures because of an acute decline in serum sodium level from 140 to 125 mEq/L.

Diagnosis

The history usually points to a likely etiology of the hyponatremia. Most patients with hyponatremia have a history of volume depletion. Diarrhea and diuretic use are very common causes of hyponatremia in children. A history of polyuria, perhaps with enuresis, and/or salt craving is present in children with primary kidney diseases or absence of aldosterone effect. Children may have signs or symptoms suggesting a diagnosis of hypothyroidism or adrenal insufficiency (Chapters 559 and 569). Brain injury raises the possibility of SIADH or cerebral salt wasting, with the caveat that SIADH is much more likely. Liver disease, nephrotic syndrome, renal failure, or congestive heart failure may be acute or chronic. The history should include a review of the patient’s intake, both intravenous and enteral, with careful attention to the amounts of water, sodium, and protein.

The traditional 1st step in the diagnostic process is determination of the plasma osmolality. This is done because some patients with a low serum sodium value do not have low osmolality. The clinical effects of hyponatremia are secondary to the associated low osmolality. Without a low osmolality, there is no movement of water into the intracellular space.

A patient with hyponatremia can have a low, normal, or high osmolality. A normal osmolality in combination with hyponatremia occurs in pseudohyponatremia. Children with elevation of serum glucose concentration or of another effective osmole (mannitol) have a high plasma osmolality and hyponatremia. The presence of a low osmolality indicates “true” hyponatremia. Patients with low osmolality are at risk for neurologic symptoms and require further evaluation to determine the etiology of the hyponatremia.

In some situations, true hyponatremia is present despite a normal or elevated plasma osmolality. The presence of an ineffective osmole, most commonly urea, increases the plasma osmolality, but because the osmole has the same concentration in the intracellular space, it does not cause fluid to move into the extracellular space. There is no dilution of the serum sodium by water, and the sodium concentration remains unchanged if the ineffective osmole is eliminated. Most importantly, the ineffective osmole does not protect the brain from edema due to hyponatremia. Hence, a patient may have symptoms of hyponatremia despite having a normal or increased osmolality because of uremia.

In patients with true hyponatremia, the next step in the diagnostic process is to clinically evaluate the volume status. Patients with hyponatremia can be hypovolemic, hypervolemic, or euvolemic. The diagnosis of volume depletion relies on the usual findings with dehydration (Chapter 54), although subtle volume depletion may not be clinically apparent. In a patient with subtle volume depletion, a fluid bolus results in a decrease in the urine osmolality and an increase in the serum sodium concentration. Children with hypervolemia are edematous on physical examination. They may have ascites, pulmonary edema, pleural effusion, or hypertension.

Hypovolemic hyponatremia can have renal or nonrenal causes. The urine sodium concentration is very useful in differentiating between renal and nonrenal causes. When the losses are nonrenal and the kidney is working properly, there is renal retention of sodium, a normal homeostatic response to volume depletion. Thus, the urinary sodium concentration is low, typically <10 mEq/L, although sodium conservation in neonates is less avid. When the kidney is the cause of the sodium loss, the urine sodium concentration is >20 mEq/L, reflecting the defect in renal sodium retention. The interpretation of the urine sodium level is challenging with diuretic therapy because it is high when diuretics are being used but low after the diuretic effect is gone. This becomes an issue only when diuretic use is surreptitious. The urine sodium concentration is not useful if a metabolic alkalosis is present; the urine chloride concentration must be used instead (Chapter 52.7).

Differentiating among the nonrenal causes of hypovolemic hyponatremia is usually facilitated by the history. Although the renal causes are more challenging to distinguish, a high serum potassium concentration is associated with disorders in which the sodium wasting is due to absence of or ineffective aldosterone.

In the patient with hypervolemic hyponatremia, the urine sodium concentration is a helpful parameter. It is usually <10 mEq/L, except in the patient with renal failure.

Treatment

The management of hyponatremia is based on the pathophysiology of the specific etiology. The management of all causes requires judicious monitoring and avoidance of an overly quick normalization of the serum sodium concentration. A patient with severe symptoms (seizures), no matter the etiology, should be given a bolus of hypertonic saline to produce a small, rapid increase in serum sodium. Hypoxia worsens cerebral edema, and hyponatremia may cause hypoxia. Hence, pulse oximetry should be monitored, and hypoxia aggressively corrected.

With all causes of hyponatremia, it is important to avoid “overly rapid” correction. The reason is that rapid correction of hyponatremia may cause central pontine myelinolysis (CPM). This syndrome, which occurs within several days of rapid correction of hyponatremia, produces neurologic symptoms, including confusion, agitation, flaccid or spastic quadriparesis, and death. There are usually characteristic pathologic and radiologic changes in the brain, especially in the pons.

CPM is more common in patients who are treated for chronic hyponatremia than in those treated for acute hyponatremia. Presumably, this difference is based on the adaptation of brain cells to the hyponatremia. The reduced intracellular osmolality that is an adaptive mechanism for chronic hyponatremia makes brain cells susceptible to dehydration during rapid correction of the hyponatremia, and this may be the mechanism of CPM. Even though CPM is rare in pediatric patients, it is advisable to avoid correcting the serum sodium concentration by >12 mEq/L/24 hr or > 18 mEq/L/48 hr. Desmopressin is a potential option if the serum sodium level is increasing too rapidly. This guideline does not apply to acute hyponatremia, as may occur with water intoxication, because the hyponatremia is more often symptomatic and there has not been time for the adaptive decrease in brain osmolality to occur. The consequences of brain edema in acute hyponatremia exceed the small risk of CPM.

Patients with hyponatremia can have severe neurologic symptoms, such as seizures and coma. The seizures associated with hyponatremia generally are poorly responsive to anticonvulsants. The child with hyponatremia and severe symptoms needs to receive treatment that will quickly reduce cerebral edema. This goal is best accomplished by increasing the extracellular osmolality so that water moves down its osmolar gradient from the intracellular space to the extracellular space.

Intravenous hypertonic saline rapidly increases serum sodium, and the effect on serum osmolality leads to a decrease in brain edema. Each mL/kg of 3% sodium chloride increases the serum sodium by approximately 1 mEq/L. A child with active symptoms often improves after receiving 4-6 mL/kg of 3% sodium chloride.

The child with hypovolemic hyponatremia has a deficiency in sodium and may have a deficiency in water. The cornerstone of therapy is to replace the sodium deficit and any water deficit that is present. The 1st step in treating any dehydrated patient is to restore the intravascular volume with isotonic saline. Ultimately, complete restoration of intravascular volume suppresses ADH production, thereby permitting excretion of the excess water. Chapter 54 discusses the management of hyponatremic dehydration.

The management of hypervolemic hyponatremia is difficult. Patients with this disorder have an excess of both water and sodium. Administration of sodium leads to worsening volume overload and edema. In addition, the patients are retaining water and sodium because of their ineffective intravascular volume or renal insufficiency. The cornerstone of therapy is water and sodium restriction, because the patients have volume overload. Diuretics may help by causing excretion of both sodium and water. Vasopressin antagonists (tolvaptan), by blocking the action of ADH and causing a water diuresis, are effective in correcting the hypervolemic hyponatremia due to heart failure or cirrhosis.

Some patients with low albumin due to nephrotic syndrome have a better response to diuretics after an infusion of 25% albumin; the sodium concentration often normalizes as a result of expansion of the intravascular volume. A child with congestive heart failure may have an increase in renal water and sodium excretion if there is an improvement in cardiac output. This improvement will “turn off” the regulatory hormones that are causing renal water (ADH) and sodium (aldosterone) retention. The patient with renal failure cannot respond to any of these therapies except fluid restriction. Insensible fluid losses eventually result in an increase in the sodium concentration as long as insensible and urinary losses are greater than intake. A more definitive approach in children with renal failure is to perform dialysis, which removes water and sodium.

In isovolumic hyponatremia, there is usually an excess of water and a mild sodium deficit. Therapy is directed at eliminating the excess water. The child with acute excessive water intake loses water in the urine because ADH production is turned off as a result of the low plasma osmolality. Children may correct their hyponatremia spontaneously over 3-6 hr. For acute, symptomatic hyponatremia due to water intoxication, hypertonic saline may be needed to reverse cerebral edema. For chronic hyponatremia due to poor solute intake, the child needs to receive an appropriate formula, and excess water intake should be eliminated.

Children with iatrogenic hyponatremia due to the administration of hypotonic intravenous fluids should receive 3% saline if they are symptomatic. Subsequent management is dictated by the patient’s volume status. The hypovolemic child should receive isotonic intravenous fluids. The child with nonphysiologic stimuli for ADH production should undergo fluid restriction. Prevention of this iatrogenic complication requires judicious use of intravenous fluids (Chapter 53).

Specific hormone replacement is the cornerstone of therapy for the hyponatremia of hypothyroidism or cortisol deficiency. Correction of the underlying defect permits appropriate elimination of the excess water.

SIADH is a condition of excess water, with limited ability of the kidney to excrete water. The mainstay of its therapy is fluid restriction. Furosemide is effective in the patient with SIADH and severe hyponatremia. Even in a patient with SIADH, furosemide causes an increase in water and sodium excretion. The loss of sodium is somewhat counterproductive, but this sodium can be replaced with hypertonic saline. Because the patient has a net loss of water and the urinary losses of sodium have been replaced, there is an increase in the sodium concentration, but no significant increase in blood pressure. Vasopressin antagonists (conivaptan, tolvaptan), which block the action of ADH and cause a water diuresis, are effective at correcting euvolemic hyponatremia, but overly rapid correction is a potential complication.

Treatment of chronic SIADH is challenging. Fluid restriction in children is difficult for nutritional and behavioral reasons. Other options are long-term furosemide therapy with sodium supplementation, an oral vasopressin antagonist (tolvaptan), or oral urea.

Body Content and Physiologic Function

The intracellular concentration of potassium, approximately 150 mEq/L, is much higher than the plasma concentration (see Web Fig. 52-3). The majority of body potassium is contained in muscle. As muscle mass increases, there is an increase in body potassium. There is thus an increase in body potassium during puberty, and it is more significant in males. The majority of extracellular potassium is in bone; <1% of total body potassium is in plasma.

Because most potassium is intracellular, the plasma concentration does not always reflect the total body potassium content. A variety of conditions alter the distribution of potassium between the intracellular and extracellular compartments. The Na⁺,K⁺-ATPase maintains the high intracellular potassium concentration by pumping sodium out of the cell and potassium into the cell. This activity balances the normal leak of potassium out of cells via potassium channels that is driven by the favorable chemical gradient. Insulin increases potassium movement into cells by activating the Na⁺,K⁺-ATPase. Hyperkalemia stimulates insulin secretion, which helps mitigate the hyperkalemia. Acid-base status affects potassium distribution, probably via potassium channels and the Na⁺,K⁺-ATPase. A decrease in pH drives potassium extracellularly; an increase in pH has the opposite effect. β-Adrenergic agonists stimulate the Na⁺,K⁺-ATPase, increasing cellular uptake of potassium. This increase is protective, in that hyperkalemia stimulates adrenal release of catecholamines. α-Adrenergic agonists and exercise cause a net movement of potassium out of the intracellular space. An increase in plasma osmolality, as with mannitol infusion, leads to water movement out of the cells, and potassium follows as a result of solvent drag. The serum potassium concentration increases by approximately 0.6 mEq/L with each 10-mOsm rise in plasma osmolality.

The high intracellular concentration of potassium, the principal intracellular cation, is maintained via the Na⁺,K⁺-ATPase. The resulting chemical gradient is used to produce the resting membrane potential of cells. Potassium is necessary for the electrical responsiveness of nerve and muscle cells and for the contractility of cardiac, skeletal, and smooth muscle. The changes in membrane polarization that occur during muscle contraction or nerve conduction make these cells susceptible to changes in serum potassium levels. The ratio of intracellular to extracellular potassium determines the threshold for a cell to generate an action potential and the rate of cellular repolarization. The intracellular potassium concentration affects cellular enzymes. Potassium is necessary for maintaining cell volume because of its important contribution to intracellular osmolality.

Intake

Potassium is plentiful in food. Dietary consumption varies considerably, even though 1-2 mEq/kg is the recommended intake. The intestines normally absorb approximately 90% of ingested potassium. Most absorption occurs in the small intestine, whereas the colon exchanges body potassium for luminal sodium. Regulation of intestinal losses normally has a minimal role in maintaining potassium homeostasis, although renal failure, aldosterone, and glucocorticoids increase colonic secretion of potassium. The increase in intestinal losses in the setting of renal failure and hyperkalemia, which stimulates aldosterone production, is clinically significant, helping to protect against hyperkalemia.

Excretion

There is some loss of potassium in sweat, but it is normally minimal. The colon has the ability to eliminate some potassium. In addition, after an acute potassium load, much of the potassium, >40%, moves intracellularly, through the actions of epinephrine and insulin, which are produced in response to hyperkalemia. This process provides transient protection from hyperkalemia, but most ingested potassium is eventually excreted in the urine. The kidneys principally regulate long-term potassium balance, and they alter excretion in response to a variety of signals. Potassium is freely filtered at the glomerulus, but 90% is resorbed before the distal tubule and collecting duct, the principal sites of potassium regulation. The distal tubule and the collecting duct have the ability to absorb and secrete potassium. It is the amount of tubular secretion that regulates the amount of potassium that appears in the urine. The plasma potassium concentration directly influences secretion in the distal nephron. As the potassium concentration increases, secretion increases.

The principal hormone regulating potassium secretion is aldosterone, which is released by the adrenal cortex in response to increased plasma potassium. Its main site of action is the cortical collecting duct, where aldosterone stimulates sodium movement from the tubule into the cells. This movement creates a negative charge in the tubular lumen, facilitating potassium excretion. In addition, the increased intracellular sodium stimulates the basolateral Na⁺,K⁺-ATPase, causing more potassium to move into the cells lining the cortical collecting duct. Glucocorticoids, ADH, a high urinary flow rate, and high sodium delivery to the distal nephron also increase urinary potassium excretion. Potassium excretion is decreased by insulin, catecholamines, and urinary ammonia. Whereas ADH increases potassium secretion, it also causes water resorption, decreasing urinary flow. The net effect is that ADH has little overall impact on potassium balance. Alkalosis causes potassium to move into cells, including the cells lining the collecting duct. This movement increases potassium secretion, and because acidosis has the opposite effect, it decreases potassium secretion.

The kidney can dramatically vary potassium excretion in response to changes in intake. Normally, approximately 10-15% of the filtered load is excreted. In an adult, excretion of potassium can vary from 5-1,000 mEq/day.

Etiology and Pathophysiology

Three basic mechanisms cause hyperkalemia (Table 52-4). In the individual patient, the etiology is sometimes multifactorial.

Fictitious or spurious hyperkalemia is very common in children because of the difficulties in obtaining blood specimens. This laboratory result is usually due to hemolysis during a heelstick or phlebotomy, but it can be the result of prolonged tourniquet application or fist clenching, either of which causes local potassium release from muscle.

The serum potassium level is normally 0.4 mEq/L higher than the plasma value, secondary to potassium release from cells during clot formation. This phenomenon is exaggerated with thrombocytosis because of potassium release from platelets. For every 100,000/m³ increase in the platelet count, the serum potassium level rises by approximately 0.15 mEq/L. This phenomenon also occurs with the marked white blood cell count elevations sometimes seen with leukemia. Elevated white blood cell counts, typically >200,000/m³, can cause a dramatic elevation in the serum potassium concentration. Analysis of a plasma sample usually provides an accurate result. It is important to analyze the sample promptly to avoid potassium release from cells, which occurs if the sample is stored in the cold, or cellular uptake of potassium and spurious hypokalemia, which occurs with storage of the sample at room temperature.

Because of the kidney’s ability to excrete potassium, it is unusual for excessive intake, by itself, to cause hyperkalemia. This condition can occur in a patient who is receiving large quantities of intravenous or oral potassium for excessive losses that are no longer present. Frequent or rapid blood transfusions can acutely increase the potassium level because of the potassium content of blood, which is variably elevated. Increased intake may precipitate hyperkalemia if there is an underlying defect in potassium excretion.

The intracellular space has a very high potassium concentration, so a shift of potassium from the intracellular space to the extracellular space can have a significant effect on the plasma potassium level. This shift occurs with metabolic acidosis, but the effect is minimal with an organic acid (lactic acidosis, ketoacidosis). A respiratory acidosis has less impact than a metabolic acidosis. Cell destruction, as seen with rhabdomyolysis, tumor lysis syndrome, tissue necrosis, or hemolysis, releases potassium into the extracellular milieu. The potassium released from red blood cells in internal bleeding, such as hematomas, is resorbed and enters the extracellular space.

Normal doses of succinylcholine or β-blockers and fluoride or digitalis intoxication all cause a shift of potassium out of the intracellular compartment. Succinylcholine should not be used during anesthesia in patients at risk for hyperkalemia. β-Blockers prevent the normal cellular uptake of potassium mediated by binding of β-agonists to the β₂-adrenergic receptors. Potassium release from muscle cells occurs during exercise, and levels can increase by 1-2 mEq/L with high activity. With an increased plasma osmolality, water moves from the intracellular space and potassium follows. This process occurs with hyperglycemia, although in nondiabetic patients, the resultant increase in insulin causes potassium to move intracellularly. In diabetic ketoacidosis, the absence of insulin causes potassium to leave the intracellular space, and the problem is compounded by the hyperosmolality. The effect of hyperosmolality causes a transcellular shift of potassium into the extracellular space after mannitol or hypertonic saline infusions. Malignant hyperthermia, which is triggered by some inhaled anesthetics, causes muscle release of potassium (Chapter 603.2). Hyperkalemic periodic paralysis is an autosomal dominant disorder caused by a mutated sodium channel. It results in episodic cellular release of potassium and attacks of paralysis (Chapter 603.1).

The kidneys excrete most of the daily potassium intake, so a decrease in kidney function can cause hyperkalemia. Newborn infants in general, and especially premature infants, have decreased kidney function at birth and thus are at increased risk for hyperkalemia despite an absence of intrinsic renal disease. Neonates also have decreased expression of potassium channels, further limiting potassium excretion.

A wide range of primary adrenal disorders, both hereditary and acquired, can cause decreased production of aldosterone, with secondary hyperkalemia (Chapters 569 and 570). Patients with these disorders typically have metabolic acidosis and salt wasting with hyponatremia. Children with more subtle adrenal insufficiency may have electrolyte problems only during acute illnesses. The most common form of congenital adrenal hyperplasia, 21-hydroxylase deficiency, typically manifests in male infants as hyperkalemia, metabolic acidosis, hyponatremia, and volume depletion. Females with this disorder usually are diagnosed as newborns because of their ambiguous genitalia; treatment prevents the development of electrolyte problems.

Renin, via angiotensin II, stimulates aldosterone production. A deficiency in renin, a result of kidney damage, can lead to decreased aldosterone production. Hyporeninemia occurs in many kidney diseases, with some of the more common pediatric causes listed in Table 52-4. These patients typically have hyperkalemia and a metabolic acidosis, without hyponatremia. Some of these patients have impaired renal function, partially accounting for the hyperkalemia, but the impairment in potassium excretion is more extreme than expected for the degree of renal insufficiency.

A variety of renal tubular disorders impair renal excretion of potassium. Children with pseudohypoaldosteronism type 1 have hyperkalemia, metabolic acidosis, and salt wasting leading to hyponatremia and volume depletion; aldosterone values are elevated. In the autosomal recessive variant, there is a defect in the renal sodium channel that is normally activated by aldosterone. Patients with this variant have severe symptoms, beginning in infancy. Patients with the autosomal dominant form have a defect in the aldosterone receptor, and the disease is milder, often remitting in adulthood. Pseudohypoaldosteronism type 2, also called Gordon syndrome, is an autosomal dominant disorder characterized by hypertension due to salt retention and impaired excretion of potassium and acid, leading to hyperkalemia and metabolic acidosis. Activating mutations in either WNK1 or WNK4, both serine-threonine kinases located in the distal nephron, cause Gordon syndrome. In Bartter syndrome due to mutations in the potassium channel ROMK (type 2 Bartter syndrome), there can be transient hyperkalemia in neonates, but hypokalemia subsequently develops (Chapter 525).

Acquired renal tubular dysfunction, with an impaired ability to excrete potassium, occurs in a number of conditions. These disorders, all characterized by tubulointerstitial disease, are often associated with impaired acid secretion and a secondary metabolic acidosis. In some affected children, the metabolic acidosis is the dominant feature, although a high potassium intake may unmask the defect in potassium handling. The tubular dysfunction can cause renal salt wasting, potentially leading to hyponatremia. Because of the tubulointerstitial damage, these conditions may also cause hyperkalemia as a result of hyporeninemic hypoaldosteronism.

The risk of hyperkalemia resulting from medications is greatest in patients with underlying renal insufficiency. The predominant mechanism of medication-induced hyperkalemia is impaired renal excretion, although ACE inhibitors may worsen hyperkalemia in anuric patients, probably by inhibiting gastrointestinal potassium loss, which is normally upregulated in renal insufficiency. The hyperkalemia caused by trimethoprim generally occurs only at the very high doses used to treat Pneumocystis jiroveci pneumonia in patients with AIDS. Potassium-sparing diuretics may easily cause hyperkalemia, especially because they are often used in patients who are receiving oral potassium supplements. The oral contraceptive Yasmin-28 contains drospirenone, which blocks the action of aldosterone.

Clinical Manifestations

The most important effects of hyperkalemia are due to the role of potassium in membrane polarization. The cardiac conduction system is usually the dominant concern. Changes in the electrocardiogram (ECG) begin with peaking of the T waves. This is followed, as the potassium level increases, by ST-segment depression, an increased PR interval, flattening of the P wave, and widening of the QRS complex. This process can eventually progress to ventricular fibrillation. Asystole may also occur. Some patients have paresthesias, fasciculations, weakness, and even an ascending paralysis, but cardiac toxicity usually precedes these clinical symptoms, emphasizing the danger of assuming that an absence of symptoms implies an absence of danger. Chronic hyperkalemia is generally better tolerated than acute hyperkalemia.

Diagnosis

The etiology of hyperkalemia is often readily apparent. Spurious hyperkalemia is very common in children, so obtaining a second potassium measurement is often appropriate. If there is a significant elevation of the white blood cell or platelet count, the second measurement should be performed on a plasma sample that is evaluated promptly. The history should initially focus on potassium intake, risk factors for transcellular shifts of potassium, medications that cause hyperkalemia, and the presence of signs of renal insufficiency, such as oliguria and edema. Initial laboratory evaluation should include creatinine, BUN, and assessment of the acid-base status. Many etiologies of hyperkalemia cause a metabolic acidosis; a metabolic acidosis worsens hyperkalemia through the transcellular shift of potassium out of cells. Renal insufficiency is a common cause of the combination of metabolic acidosis and hyperkalemia. This association is also seen in diseases associated with aldosterone insufficiency or aldosterone resistance. Children with absence of or ineffective aldosterone often have hyponatremia and volume depletion because of salt wasting. Genetic diseases, such as congenital adrenal hyperplasia and pseudohypoaldosteronism, usually manifest in infancy and should be strongly considered in the infant with hyperkalemia and metabolic acidosis, especially if hyponatremia is present. It is important to consider the various etiologies of a transcellular shift of potassium. In some of these disorders, the potassium level continues to increase, despite the elimination of all potassium intake, especially when there is concurrent renal insufficiency. This increase is potentially seen in tumor lysis syndrome, hemolysis, rhabdomyolysis, and other causes of cell death. All of these entities can cause concomitant hyperphosphatemia and hyperuricemia. Rhabdomyolysis produces an elevated creatinine phosphokinase (CPK) value and hypocalcemia, whereas children with hemolysis have hemoglobinuria and a decreasing hematocrit. For the child with diabetes, an elevated blood glucose value suggests a transcellular shift of potassium.

When there is no clear etiology of hyperkalemia, the diagnostic approach should focus on differentiating decreased potassium excretion from the other etiologies. Measuring urinary potassium assesses renal excretion of potassium. The transtubular potassium gradient (TTKG) is a useful method to evaluate the renal response to hyperkalemia, as follows:

where [K]_urine is urine potassium concentration and [K]_plasma is plasma potassium concentration. For the result to be valid, the urine osmolality must be greater than the serum osmolality. The TTKG normally varies widely, ranging from 5-15. The TTKG should be >10 in the setting of hyperkalemia, assuming normal renal excretion of potassium. A TTKG <8 during hyperkalemia suggests a defect in renal potassium excretion, which is usually due to lack of aldosterone or an inability to respond to aldosterone. Measurement of aldosterone is useful for differentiating these possible mechanisms. Patients with a lack of aldosterone respond to fludrocortisone, an oral mineralocorticoid, by increasing urinary potassium and decreasing serum potassium. An appropriate TTKG with normal kidney function argues for a nonrenal cause of hyperkalemia.

Treatment

The plasma potassium level, the ECG, and the risk of the problem worsening determine the aggressiveness of the therapeutic approach. High serum potassium levels and the presence of ECG changes require vigorous treatment. An additional source of concern is the patient in whom plasma potassium levels are rising despite minimal intake. This situation can happen if there is cellular release of potassium (tumor lysis syndrome), especially in the setting of diminished excretion (renal failure).

The 1st action in a child with a concerning elevation of plasma potassium is to stop all sources of additional potassium (oral, intravenous) (Chapter 529). Washed red blood cells can be used for patients who require blood transfusions. If the potassium level is >6.0-6.5 mEq/L, an ECG should be obtained to help assess the urgency of the situation. Peak T waves are the first sign of hyperkalemia followed by a prolonged PR interval and, when most severe, a prolonged QRS complex. Life-threatening ventricular arrhythmias may also develop. The treatment of hyperkalemia has 2 basic goals: (1) to stabilize the heart to prevent life-threatening arrhythmias and (2) to remove potassium from the body. The treatments that acutely prevent arrhythmias all have the advantage of working quickly (within minutes) but do not remove potassium from the body. Calcium stabilizes the cell membrane of heart cells, preventing arrhythmias. It is given intravenously over a few minutes, and its action is almost immediate. Calcium should be given over 30 min in a patient receiving digitalis, because otherwise the calcium may cause arrhythmias. Bicarbonate causes potassium to move intracellularly, lowering the plasma potassium level. It is most efficacious in a patient with a metabolic acidosis. Insulin causes potassium to move intracellularly but must be given with glucose to avoid hypoglycemia. The combination of insulin and glucose works within 30 min. Nebulized albuterol, by stimulation of β₁-receptors, leads to rapid intracellular movement of potassium. This has the advantage of not requiring an intravenous route of administration, allowing it to be given concurrently with the other measures.

It is critical to begin measures that remove potassium from the body. In patients who are not anuric, a loop diuretic increases renal excretion of potassium. A high dose may be required in a patient with significant renal insufficiency. Sodium polystyrene sulfonate (Kayexalate) is an exchange resin that is given either rectally or orally. Sodium in the resin is exchanged for body potassium, and the potassium-containing resin is then excreted from the body. Some patients require dialysis for acute potassium removal. Dialysis is often necessary if the patient has either severe renal failure or an especially high rate of endogenous potassium release, as is sometimes present with tumor lysis syndrome or rhabdomyolysis. Hemodialysis rapidly lowers plasma potassium levels. Peritoneal dialysis is not nearly as quick or reliable, but it is usually adequate as long as the acute problem can be managed with medications and the endogenous release of potassium is not high.

Long-term management of hyperkalemia includes reducing intake via dietary changes and eliminating or reducing medications that cause hyperkalemia (Chapter 529). Some patients require medications to increase potassium excretion, such as sodium polystyrene sulfonate and loop or thiazide diuretics. Some infants with chronic renal failure may need to start dialysis to allow adequate caloric intake without hyperkalemia. It is unusual for an older child to require dialysis principally to control chronic hyperkalemia. The disorders that are due to a deficiency in aldosterone respond to replacement therapy with fludrocortisone.

Etiology and Pathophysiology

There are 4 basic mechanisms of hypokalemia (Table 52-5). Spurious hypokalemia occurs in patients with leukemia and very elevated white blood cell counts if plasma for analysis is left at room temperature, permitting the white blood cells to take up potassium from the plasma. With a transcellular shift, there is no change in total body potassium, although there may be concomitant potassium depletion resulting from other factors. Decreased intake, extrarenal losses, and renal losses are all associated with total body potassium depletion.

Because the intracellular potassium concentration is much higher than the plasma level, a significant amount of potassium can move into cells without markedly changing the intracellular potassium concentration. Alkalemia is one of the more common causes of a transcellular shift. The effect is much greater with a metabolic alkalosis than with a respiratory alkalosis. The impact of exogenous insulin on potassium movement into the cells is substantial in patients with diabetic ketoacidosis. Endogenous insulin may be the cause when a patient is given a bolus of glucose. Both endogenous (epinephrine in stress) and exogenous (albuterol) β-adrenergic agonists stimulate cellular uptake of potassium. Theophylline overdose, barium intoxication, administration of cesium chloride (a homeopathic cancer remedy), and toluene intoxication from paint or glue sniffing can cause a transcellular shift hypokalemia, often with severe clinical manifestations. Children with hypokalemic periodic paralysis, a rare autosomal dominant disorder, have acute cellular uptake of potassium (Chapter 603). Thyrotoxic periodic paralysis, which is more common in Asians, is an unusual initial manifestation of hyperthyroidism. Affected patients have dramatic hypokalemia as a result of a transcellular shift of potassium. Hypokalemia can occur during refeeding syndrome (Chapter 330.08).

Inadequate potassium intake occurs in anorexia nervosa; accompanying bulimia and laxative or diuretic abuse exacerbates the potassium deficiency. Sweat losses of potassium can be significant during vigorous exercise in a hot climate. Associated volume depletion and hyperaldosteronism increase renal losses of potassium (discussed later). Diarrheal fluid has a high concentration of potassium, and hypokalemia due to diarrhea is usually associated with a metabolic acidosis resulting from stool losses of bicarbonate. In contrast, a normal acid-base balance or a mild metabolic alkalosis is seen with laxative abuse. Intake of sodium polystyrene sulfonate or ingestion of clay due to pica increases stool losses of potassium.

Urinary potassium wasting may be accompanied by a metabolic acidosis (proximal or distal RTA). In diabetic ketoacidosis, although it is often associated with normal plasma potassium caused by transcellular shifts, there is significant total body potassium depletion from urinary losses due to the osmotic diuresis, and the potassium level may decrease dramatically with insulin therapy (Chapter 583). Both the polyuric phase of acute tubular necrosis and postobstructive diuresis cause transient, highly variable potassium wasting and may be associated with a metabolic acidosis. Tubular damage, which occurs either directly from medications or secondary to interstitial nephritis, is often accompanied by other tubular losses of nutrients, including magnesium, sodium, and water. Such tubular damage may cause a secondary RTA with a metabolic acidosis. Isolated magnesium deficiency causes renal potassium wasting. Penicillin is an anion that is excreted in the urine, resulting in increased potassium excretion because the penicillin anion must be accompanied by a cation. Hypokalemia from penicillin therapy occurs only with the sodium salt of penicillin, not with the potassium salt.

Urinary potassium wasting is often accompanied by a metabolic alkalosis. This condition is usually associated with increased aldosterone, which increases urinary potassium and acid losses, contributing to the hypokalemia and the metabolic alkalosis. Other mechanisms often contribute to both the potassium losses and the metabolic alkalosis. With emesis or nasogastric suction, there is gastric loss of potassium, but this is fairly minimal, given the low potassium content of gastric fluid (≈10 mEq/L). More important is the gastric loss of hydrochloric acid (HCl), leading to a metabolic alkalosis and a state of volume depletion. The kidney compensates for the metabolic alkalosis by excreting bicarbonate in the urine, but there is obligate loss of potassium and sodium with the bicarbonate. The volume depletion raises aldosterone levels, further increasing urinary potassium losses and preventing correction of the metabolic alkalosis and hypokalemia until the volume depletion is corrected. Urinary chloride is low as a response to the volume depletion. Because the volume depletion is secondary to chloride loss, this is a state of chloride deficiency. There were cases of chloride deficiency resulting from infant formula deficient in chloride, which caused a metabolic alkalosis with hypokalemia and low urine chloride levels. Current infant formula is not deficient in chloride. A similar mechanism occurs in cystic fibrosis because of chloride loss in sweat. In congenital chloride-losing diarrhea, an autosomal recessive disorder, there is high stool loss of chloride, leading to metabolic alkalosis, an unusual sequela of diarrhea. Because of stool potassium losses, chloride deficiency, and metabolic alkalosis, patients with this disorder have hypokalemia. During respiratory acidosis, there is renal compensation, with retention of bicarbonate and excretion of chloride. After the respiratory acidosis is corrected, the patients have chloride deficiency and post-hypercapnic alkalosis with secondary hypokalemia. Patients with chloride deficiency, metabolic alkalosis, and hypokalemia have a urinary chloride level of <10 mEq/L. Loop and thiazide diuretics lead to hypokalemia, metabolic alkalosis, and chloride deficiency. During treatment, these patients have high urine chloride levels resulting from the effect of the diuretic. However, after the diuretics are discontinued, there is residual chloride deficiency, the urinary chloride level is appropriately low, and neither the hypokalemia nor the alkalosis resolves until the chloride deficiency is corrected.

The combination of metabolic alkalosis, hypokalemia, a high urine chloride level, and normal blood pressure is characteristic of Bartter syndrome, Gitelman syndrome, and current diuretic use. Patients with any of these conditions have high urinary losses of potassium and chloride, despite a state of relative volume depletion with secondary hyperaldosteronism. Bartter and Gitelman syndromes are autosomal recessive disorders caused by defects in tubular transporters (Chapter 525). Bartter syndrome is usually associated with hypercalciuria, and often with nephrocalcinosis, whereas children with Gitelman syndrome have low urinary calcium losses but hypomagnesemia due to urinary magnesium losses. Some patients with Bartter syndrome have hypomagnesemia.

Some patients with hypoparathyroidism and hypocalcemia due to an activating mutation of the calcium-sensing receptor (autosomal dominant hypoparathyroidism) have hypokalemia, hypomagnesemia, and metabolic alkalosis. The reason is that activation of the calcium-sensing receptor in the loop of Henle impairs tubular resorption of sodium and chloride, causing volume depletion and secondary hyperaldosteronism EAST syndrome, an autosomal recessive disorder due to mutations in the gene for a potassium channel present in the kidney, inner ear and brain, consists of epilepsy, ataxia, sensorineural hearing loss, and tubulopathy (hypokalemia, metabolic alkalosis, hypomagnesemia, and hypocalciuria).

In the presence of high aldosterone levels, there is urinary loss of potassium, hypokalemia, metabolic alkalosis, and an elevated urinary chloride level. Also ,renal retention of sodium leads to hypertension. Primary hyperaldosteronism caused by adenoma or hyperplasia is much less common in children than in adults (Chapter 572). Glucocorticoid-remediable aldosteronism, an autosomal dominant disorder that leads to high levels of aldosterone, is often diagnosed in childhood, although hypokalemia is not always present.

Increased aldosterone levels may be secondary to increased renin production. Renal artery stenosis leads to hypertension from increased renin and secondary hyperaldosteronism. The increased aldosterone can cause hypokalemia and metabolic alkalosis, although most patients have normal electrolyte levels. Renin-producing tumors, which are extremely rare, can cause hypokalemia.

A variety of disorders causes hypertension and hypokalemia without increased aldosterone levels. Some are due to increased levels of mineralocorticoids other than aldosterone. Such increases occur in 2 forms of congenital adrenal hyperplasia (Chapter 570). In 11β-hydroxylase deficiency, which is associated with virilization, the value of 11-deoxycorticosterone (DOC) is elevated, causing variable hypertension and hypokalemia. A similar mechanism, increased DOC, occurs in 17α-hydroxylase deficiency, but patients with this disorder are more uniformly hypertensive and hypokalemic, and they have a defect in sex hormone production. Cushing syndrome, frequently associated with hypertension, less commonly causes metabolic alkalosis and hypokalemia. This is secondary to the mineralocorticoid activity of cortisol. In 11β-hydroxysteroid dehydrogenase deficiency, an autosomal recessive disorder, the enzymatic defect prevents the conversion of cortisol to cortisone in the kidney. Because cortisol binds to and activates the aldosterone receptor, children with this deficiency have all the features of excessive mineralocorticoids, including hypertension, hypokalemia, and metabolic alkalosis. Patients with this disorder, which is also called apparent mineralocorticoid excess, respond to spironolactone therapy, which blocks the mineralocorticoid receptor. An acquired form of 11β-hydroxysteroid dehydrogenase deficiency occurs from the ingestion of substances that inhibit this enzyme. A classic example is glycyrrhizic acid, which is found in natural licorice. Liddle syndrome is an autosomal dominant disorder that results from an activating mutation of the distal nephron sodium channel that is normally upregulated by aldosterone. Patients have the characteristics of hyperaldosteronism—hypertension, hypokalemia, and alkalosis—but low serum aldosterone levels. These patients respond to the potassium-sparing diuretics (triamterene and amiloride) that inhibit this sodium channel (Chapter 525.3).

Clinical Manifestations

The heart and skeletal muscle are especially vulnerable to hypokalemia. ECG changes include a flattened T wave, a depressed ST segment, and the appearance of a U wave, which is located between the T wave (if still visible) and the P wave. Ventricular fibrillation and torsades de pointes may occur, although usually only in the context of underlying heart disease. Hypokalemia makes the heart especially susceptible to digitalis-induced arrhythmias, such as supraventricular tachycardia, ventricular tachycardia, and heart block (Chapter 429).

The clinical consequences of hypokalemia in skeletal muscle include muscle weakness and cramps. Paralysis is a possible complication, generally only at potassium levels <2.5 mEq/L. It usually starts in the legs and moves to the arms. Respiratory paralysis may require mechanical ventilation. Some patients have rhabdomyolysis; the risk increases with exercise. Hypokalemia slows gastrointestinal motility. This effect manifests as constipation; with potassium levels <2.5 mEq/L, an ileus may occur. Hypokalemia impairs bladder function, potentially leading to urinary retention.

Hypokalemia causes polyuria and polydipsia by 2 mechanisms, primary polydipsia and impairment of urinary concentrating ability, which produces nephrogenic diabetes insipidus. Hypokalemia stimulates renal ammonia production, an effect that is clinically significant if hepatic failure is present, because the liver cannot metabolize the ammonia. Hypokalemia may therefore worsen hepatic encephalopathy.

Chronic hypokalemia may cause kidney damage, including interstitial nephritis and renal cysts. In children, chronic hypokalemia, like Bartter syndrome, leads to poor linear growth.

Diagnosis

Most causes of hypokalemia are readily apparent from the history. It is important to review the child’s diet, gastrointestinal losses, and medications. Both emesis and diuretic use can be surreptitious. The presence of hypertension suggests excess mineralocorticoids. Concomitant electrolyte abnormalities are useful clues. The combination of hypokalemia and metabolic acidosis is characteristic of diarrhea and of distal and proximal RTA. A concurrent metabolic alkalosis is characteristic of emesis or nasogastric losses, aldosterone excess, use of diuretics, and Bartter and Gitelman syndromes. An approach to persistent hypokalemia is shown in Figure 52-2.

Figure 52-2 Diagnostic algorithm to evaluate persistent hypokalemia. *Spurious hypokalemia must be excluded. **Hypokalemia is uncommon in uncomplicated edematous disorders and in conditions associated with excessive glucocorticosteroids. Conditions associated with high circulating levels of glucocorticosteroids often have normal renin activity. 17-OHP, 17-hydroxyprogesterone; ACTH, adrenocorticotropic hormone; AME, apparent mineralocorticoid excess; BP, blood pressure; Cl⁻, chloride; DOC, 11-deoxycorticosterone; DR, direct renin assay; GI, gastrointestinal; FH-II, familial hyperaldosteronism type II; GR, glucocorticoid receptor; GRA (FH-I), glucocorticoid remediable aldosteronism (familial hyperaldosteronism type I); K⁺, potassium; MR, mineralocorticoid receptor; PA-I, pseudoaldosteronism type I; PA-II, pseudoaldosteronism type II; PRA, plasma renin activity; TTKG, transtubular potassium gradient.

(From Shoemaker LR, Eaton BV, Buchino JJ: A three-year-old with persistent hypokalemia, J Pediatr 151:696–699, 2007.)

If a clear etiology is not apparent, the measurement of urinary potassium distinguishes between renal and extrarenal losses. The kidneys should conserve potassium in the presence of extrarenal losses. Urinary potassium losses can be assessed with a 24-hr urine collection, a spot potassium-creatinine ratio, a fractional excretion of potassium, or calculation of the TTKG, which is the most widely used approach in children:

where [K]_urine = urine potassium concentration and [K]_plasma = plasma potassium concentration.

The urine osmolality must be greater than the serum osmolality for the result of this calculation to be valid. A TTKG >4 in the presence of hypokalemia suggests excessive urinary losses of potassium. The urinary potassium excretion value can be misleading if the stimulus for renal loss, such as a diuretic, is no longer present.

Treatment

Factors that influence the treatment of hypokalemia include the potassium level, clinical symptoms, renal function, the presence of transcellular shifts of potassium, ongoing losses, and the patient’s ability to tolerate oral potassium. Severe, symptomatic hypokalemia requires aggressive treatment. Supplementation is more cautious if renal function is decreased because of the kidney’s limited ability to excrete excessive potassium. The plasma potassium level does not always provide an accurate estimation of the total body potassium deficit because there may be shifts of potassium from the intracellular space to the plasma. Clinically, such shifts occur most commonly with metabolic acidosis and the insulin deficiency of diabetic ketoacidosis; the plasma potassium measurement underestimates the degree of total body potassium depletion. When these problems are corrected, potassium moves into the intracellular space, so more potassium supplementation is required to correct the hypokalemia. Likewise, the presence of a transcellular shift of potassium into the cells indicates that the total body potassium depletion is less severe. In an isolated transcellular shift, as occurs in hypokalemic periodic paralysis, potassium supplementation should be used cautiously, given the risk of hyperkalemia when the transcellular shift resolves. This caution is especially required in thyrotoxic periodic paralysis, which responds dramatically to propranolol, with correction of weakness and hypokalemia. Patients who have ongoing losses of potassium need correction of the deficit and replacement of the ongoing losses.

Because of the risk of hyperkalemia, intravenous potassium should be used very cautiously. Oral potassium is safer, albeit not as rapid in urgent situations. Liquid preparations are bitter tasting; microencapsulated or wax matrix formulations are less irritating than tablets to the gastric mucosa (oral dose: 2-4 mEq/kg/day with a maximum of 120-240 mEq/day in divided doses). The dose of intravenous potassium is 0.5-1 mEq/kg, usually given over 1 hr. The adult maximum dose is 40 mEq. Conservative dosing is generally preferred. Potassium chloride is the usual choice for supplementation, although the presence of concurrent electrolyte abnormalities may dictate other options. Patients with acidosis and hypokalemia can receive potassium acetate or potassium citrate. If hypophosphatemia is present, then some of the potassium deficit can be replaced with potassium phosphate. It is sometimes possible to decrease ongoing losses of potassium. For patients with excessive urinary losses, potassium-sparing diuretics are effective, but they need to be used cautiously in patients with renal insufficiency. If hypokalemia, metabolic alkalosis, and volume depletion are present (with gastric losses), then restoration of intravascular volume with adequate sodium chloride will decrease urinary potassium losses. Correction of concurrent hypomagnesemia is important because hypomagnesemia may cause hypokalemia. Disease-specific therapy is effective in many of the genetic tubular disorders.

Body Content and Physiologic Function

Magnesium is the 4th most common cation in the body and the 3rd most common intracellular cation (see Web Fig. 52-3). Between 50% and 60% of body magnesium is in bone, where it serves as a reservoir because 30% is exchangeable, allowing movement to the extracellular space. Most intracellular magnesium is bound to proteins; only approximately 25% is exchangeable. Because cells with higher metabolic rates have higher magnesium concentrations, most intracellular magnesium is present in muscle and liver.

The normal plasma magnesium concentration is 1.5-2.3 mg/dL (1.2-1.9 mEq/L; 0.62-0.94 mmol/L), with some variation among clinical laboratories. In the USA, serum magnesium is reported as mg/dL (Web Table 52-1). Infants have slightly higher plasma magnesium concentrations than older children and adults. Only 1% of body magnesium is extracellular (60% ionized; 15% complexed; 25% protein bound).

Magnesium is a necessary cofactor for hundreds of enzymes. It is important for membrane stabilization and nerve conduction. Adenosine triphosphate (ATP) and guanosine triphosphate need associated magnesium when they are used by ATPases, cyclases, and kinases.

Intake

Between 30% and 50% of dietary magnesium is absorbed. Good dietary sources include green vegetables, cereals, nuts, meats, and hard water, although many foods contain magnesium. Human milk contains approximately 35 mg/L of magnesium; formula contains 40-70 mg/L. The small intestine is the major site of magnesium absorption, but the regulation of magnesium absorption is poorly understood. There is passive absorption, which permits high absorption in the presence of excessive intake. It probably occurs via a paracellular mechanism. Absorption is diminished in the presence of substances that complex with magnesium (free fatty acids, fiber, phytate, phosphate, oxalate); increased intestinal motility and calcium also decrease magnesium absorption. Vitamin D and parathyroid hormone (PTH) may enhance absorption, although this effect is limited. Intestinal absorption does increase when intake is decreased, possibly via a saturable active transport system. If there is no oral intake of magnesium, obligatory secretory losses prevent the complete elimination of intestinal losses.

Excretion

Renal excretion is the principal regulator of magnesium balance. There is no defined hormonal regulatory system, although PTH may increase tubular resorption. Approximately 15% of resorption occurs in the proximal tubule, and 70% in the thick ascending limb (TAL) of the loop of Henle. Proximal resorption may be higher in neonates. High serum magnesium levels inhibit resorption in the TAL, suggesting that active transport is involved. Approximately 5-10% of filtered magnesium is resorbed in the distal tubule. Hypomagnesemia increases absorption in the TAL and the distal tubule.

Etiology and Pathophysiology

Gastrointestinal and renal losses are the major causes of hypomagnesemia (Web Table 52-2). Diarrheal fluid contains up to 200 mg/L of magnesium; gastric contents have only approximately 15 mg/L, but high losses can cause depletion. Steatorrhea causes magnesium loss as a result of the formation of magnesium-lipid salts; restriction of dietary fat can decrease losses.

Hypomagnesemia with secondary hypocalcemia, a rare autosomal recessive disorder, is due to decreased intestinal absorption of magnesium and renal magnesium wasting. Patients with this disorder have mutations in a gene (TRPM6) that is expressed in intestine and kidney. TRPM6 codes for a transient receptor potential cation channel. The patients have seizures, tetany, tremor, or restlessness at 2-8 wk of life as a result of severe hypomagnesemia (0.2-0.8 mg/dL) and secondary hypocalcemia.

Renal losses may occur because of medications that are direct tubular toxins. Amphotericin frequently causes significant magnesium wasting and is typically associated with other tubular defects (especially potassium wasting). Cisplatin produces dramatic renal magnesium losses. Diuretics affect tubular handling of magnesium. Loop diuretics cause a mild increase in magnesium excretion, and thiazide diuretics have even less effect. Potassium-sparing diuretics reduce magnesium losses. Osmotic agents, such as mannitol, glucose in diabetes mellitus, and urea in the recovery phase of acute tubular necrosis, increase urinary magnesium losses. Intravenous fluid, by expanding the intravascular volume, decreases renal resorption of sodium and water, thereby impairing magnesium resorption. Hypercalcemia inhibits magnesium resorption in the loop of Henle, although this inhibition does not occur in hypercalcemia due to familial hypercalcemic hypocalciuria or lithium.

A number of rare genetic diseases cause renal magnesium loss. Gitelman and Bartter syndromes, both autosomal recessive disorders, are the most common entities (Chapter 525). Gitelman syndrome, which is due to a defect in the thiazide-sensitive Na⁺-Cl⁻ cotransporter in the distal tubule, is usually associated with hypomagnesemia. Hypomagnesemia occurs in a minority of patients with Bartter syndrome, which can be caused by mutations in at least 5 different genes. In both disorders, there is hypokalemic metabolic alkalosis. Typically, hypomagnesemia is not severe and is asymptomatic, although tetany due to hypomagnesemia occasionally occurs.

Familial hypomagnesemia with hypercalciuria and nephrocalcinosis (Michelis-Castrillo syndrome), an autosomal recessive disorder, is due to mutations in the gene for claudin 16 (paracellin-1), which is located in the tight junctions of the TAL of the loop of Henle. Patients with the disease have severe renal wasting of magnesium and calcium with secondary hypomagnesemia and nephrocalcinosis; serum calcium levels are normal. Chronic renal failure frequently occurs during childhood. Other features include kidney stones, urinary tract infections, hematuria, increased PTH levels, tetany, seizures, incomplete distal RTA, hyperuricemia, polyuria, and polydipsia. Patients with familial hypomagnesemia with hypercalciuria, nephrocalcinosis, and severe ocular involvement have mutations in the gene for claudin 19.

Autosomal recessive renal magnesium wasting with normocalciuria is caused by mutations in the epidermal growth factor gene. Clinical manifestations include seizures, mild to moderate psychomotor retardation, and brisk tendon reflexes.

Autosomal dominant renal magnesium wasting is usually due to a dominant-negative mutation in the gene encoding the Na⁺,K⁺-ATPase γ subunit and is associated with hypomagnesemia, increased urinary magnesium losses, hypocalciuria, and normocalcemia. There are families without genetic linkage to this locus. Patients may present with seizures; most are asymptomatic, despite serum magnesium levels of 0.8-1.5 mg/dL. Renal cysts and diabetes syndrome, which is caused by mutations in the gene for hepatocyte nuclear factor-1β, is associated with hypomagnesemia, despite the frequent presence of renal insufficiency. The hypomagnesemia is usually mild but may cause symptomatic hypocalcemia.

EAST syndrome is due to mutations in a potassium channel, and patients with this autosomal recessive disorder have hypokalemia, metabolic alkalosis, and hypomagnesemia. Autosomal dominant hypoparathyroidism is caused by an activating mutation in the calcium-sensing receptor, which also senses magnesium levels in the kidney (Chapter 565). The mutated receptor inappropriately perceives that magnesium and calcium levels are elevated, leading to urinary wasting of both cations. Hypomagnesemia, if present, is usually mild. A mutation in a mitochondrially encoded transfer RNA is associated with hypomagnesemia, hypertension, and hypercholesterolemia. Hypomagnesemia is occasionally present in children with other mitochondrial disorders.

Poor intake is an unusual cause of hypomagnesemia, although it can be seen in children who are hospitalized and receive only intravenous fluids without magnesium. In hungry bone syndrome, which most frequently occurs after parathyroidectomy in patients with hyperparathyroidism, magnesium moves into bone as a result of accelerated bone formation. These patients usually have hypocalcemia and hypophosphatemia via the same mechanism. A similar mechanism can occur during the refeeding phase of protein-calorie malnutrition in children, with high magnesium use during cell growth depleting the patient’s limited reserves. Insulin therapy stimulates uptake of magnesium by cells, and in diabetic ketoacidosis, in which total body magnesium is low because of osmotic losses, hypomagnesemia frequently occurs. In pancreatitis, there is saponification of magnesium and calcium in necrotic fat, causing both hypomagnesemia and hypocalcemia.

Transient hypomagnesemia in newborns, which is sometimes idiopathic, is more commonly seen in infants of diabetic mothers, presumably as a result of maternal depletion from osmotic losses. Other maternal diseases that cause magnesium losses predispose infants to hypomagnesemia. Hypomagnesemia is more common in infants with intrauterine growth restriction. Hypomagnesemia may develop in newborn infants who require exchange transfusions because of magnesium removal by the citrate in banked blood (Chapter 100).

Clinical Manifestations

Hypomagnesemia causes secondary hypocalcemia by impairing the release of PTH by the parathyroid gland and through blunting of the tissue response to PTH. Thus, hypomagnesemia is part of the differential diagnosis of hypocalcemia (Chapter 565). It usually occurs only at magnesium levels <0.7 mg/dL. The dominant manifestations of hypomagnesemia are due to hypocalcemia: tetany, presence of Chvostek and Trousseau signs, and seizures. However, with severe hypomagnesemia, these same signs and symptoms may be present despite normocalcemia. Persistent hypocalcemia due to hypomagnesemia is a rare cause of rickets.

Many causes of hypomagnesemia also result in hypokalemia. Hypomagnesemia may produce renal potassium wasting and hypokalemia that corrects only with magnesium therapy. ECG changes with hypomagnesemia include flattening of the T wave and lengthening of the ST segment. Arrhythmias may occur, almost always in the setting of underlying heart disease.

Diagnosis

The etiology of hypomagnesemia is often readily apparent from the clinical situation. The child should be assessed for gastrointestinal disease, adequate intake, and kidney disease, with close attention paid to medications that may cause renal magnesium wasting. When the diagnosis is uncertain, an evaluation of urinary magnesium losses distinguishes between renal and nonrenal causes. The fractional excretion of magnesium (FE_Mg) is calculated via the following formula:

where U_Mg is urinary magnesium concentration, P_Cr is plasma creatinine concentration, P_Mg is plasma magnesium concentration, and U_Cr is urinary magnesium concentration. The plasma magnesium concentration is multiplied by 0.7 because approximately 30% is bound to albumin and not filtered at the glomerulus.

The fractional excretion of magnesium does not vary with age, but it does change according to the serum magnesium concentration. The fractional excretion ranges from 1% to 8% in children with normal magnesium levels. In the presence of hypomagnesemia due to extrarenal causes, it should be low because of renal conservation, typically <2%. The fractional excretion of magnesium is inappropriately elevated in the setting of renal magnesium wasting; values are usually >4% and frequently are >10%. The measurement should not be made during a magnesium infusion, because the acute increase in serum magnesium increases urinary magnesium. Other approaches for evaluating urinary magnesium losses include calculation of 24-hr urinary magnesium losses and of the ratio of urine magnesium with urine creatinine, both of which vary with age.

The genetic causes of renal magnesium loss are distinguished on the basis of the measurement of other serum and urinary electrolytes. Children with Gitelman and Bartter syndromes have hypokalemia and metabolic alkalosis.

Treatment

Severe hypomagnesemia is treated with parenteral magnesium. Magnesium sulfate is given at a dose of 25-50 mg/kg (0.05-0.1 mL/kg of a 50% solution; 2.5-5.0 mg/kg of elemental magnesium). It is administered as a slow intravenous infusion, although it may be given intramuscularly in neonates. The rate of intravenous infusion should be slowed if a patient experiences diaphoresis, flushing, or a warm sensation. The dose is often repeated every 6 hr (every 8-12 hr in neonates), for a total of 2-3 doses, before the plasma magnesium concentration is rechecked. Lower doses are used in children with renal insufficiency.

Long-term therapy is usually given orally. Preparations include magnesium gluconate (5.4 mg elemental magnesium/100 mg), magnesium oxide (60 mg elemental magnesium/100 mg), and magnesium sulfate (10 mg elemental magnesium/100 mg). There are sustained-released preparations, such as Slow-Mag (60 mg elemental magnesium/tablet) and Mag-Tab SR (84 mg elemental magnesium/tablet). Oral magnesium dosing should be divided to decrease cathartic side effects. Alternatives to oral magnesium are intramuscular injections and nighttime nasogastric infusion, both designed to minimize diarrhea. Magnesium supplementation must be used cautiously in the context of renal insufficiency.

Etiology and Pathophysiology

There is no feedback mechanism to prevent magnesium absorption from the gastrointestinal tract. Magnesium is present in high amounts in certain laxatives, enemas, cathartics used to treat drug overdoses, and antacids. It is also usually present in total parenteral nutrition, and neonates may receive high amounts transplacentally if maternal levels are elevated. Usually the kidneys excrete excessive magnesium, but this ability is diminished in patients with chronic renal failure. In addition, neonates and young infants are vulnerable to excessive magnesium ingestion because of their reduced GFR. Most pediatric cases not related to maternal hypermagnesemia occur in infants as a result of excessive use of antacids or laxatives. Mild hypermagnesemia may occur in chronic renal failure, familial hypocalciuric hypercalcemia, diabetic ketoacidosis, lithium ingestion, milk-alkali syndrome, and tumor lysis syndrome. The hypermagnesemia in diabetic ketoacidosis occurs despite significant intracellular magnesium depletion due to urinary losses; hypomagnesemia often occurs after insulin treatment.

Clinical Manifestations

Symptoms usually do not appear until the plasma magnesium level is >4.5 mg/dL. Hypermagnesemia inhibits acetylcholine release at the neuromuscular junction, producing hypotonia, hyporeflexia, and weakness; paralysis occurs at high concentrations. The neuromuscular effects may be exacerbated by aminoglycoside antibiotics. Direct CNS depression causes lethargy and sleepiness; infants have a poor suck. Elevated magnesium values are associated with hypotension because of vascular dilation, which also causes flushing. Hypotension can be profound at higher concentrations from a direct effect on cardiac function. ECG changes include prolonged PR, QRS, and QT intervals. Severe hypermagnesemia (>15 mg/dL) causes complete heart block and cardiac arrest. Other manifestations of hypermagnesemia include nausea, vomiting, and hypocalcemia.

Diagnosis

Except for the case of the neonate with transplacental exposure, a high index of suspicion and a good history are necessary to make the diagnosis of hypermagnesemia. Prevention is essential; magnesium-containing compounds should be used judiciously in children with renal insufficiency.

Treatment

Most patients with normal renal function rapidly clear excess magnesium. Intravenous hydration and loop diuretics can accelerate this process. In severe cases, especially in patients with underlying renal insufficiency, dialysis may be necessary. Hemodialysis works faster than peritoneal dialysis. Exchange transfusion is another option in newborn infants. Supportive care includes monitoring of cardiorespiratory status, provision of fluids, monitoring of electrolyte levels, and the use of pressors for hypotension. In acute emergencies, especially in the context of severe neurologic or cardiac manifestations, 100 mg/kg of intravenous calcium gluconate is transiently effective.

Body Content and Physiologic Function

Most phosphorus is in bone or is intracellular, with <1% in plasma. At a physiologic pH, there are monovalent and divalent forms of phosphate because the pK of these forms is 6.8. Approximately 80% is divalent, and the remainder is monovalent at a pH of 7.4. A small percentage of plasma phosphate, approximately 15%, is protein bound. The remainder can be filtered by the glomerulus, with most existing as free phosphate and a small percentage complexed with calcium, magnesium, or sodium. Phosphate is the most plentiful intracellular anion, although the majority is part of a larger compound (ATP).

More than that of any other electrolyte, the phosphorus concentration varies with age (Table 52-6). The teleologic explanation for the high concentration during childhood is the need for phosphorus to facilitate growth. There is diurnal variation in the plasma phosphorus concentration, with the peak during sleep.

Table 52-6 SERUM PHOSPHORUS LEVELS DURING CHILDHOOD

Phosphorus, as a component of ATP and other trinucleotides, is critical for cellular energy metabolism. It is necessary for cell signaling and nucleic acid synthesis, and it is a component of cell membranes and other structures. Along with calcium, phosphorus is necessary for skeletal mineralization. There is a significant need for a net positive phosphorus balance during growth, with the growing skeleton especially vulnerable to deficiency.

Intake

Phosphorus is readily available in food. Milk and milk products are the best sources of phosphorus; high concentrations are present in meat and fish. Vegetables have more phosphorus than fruits and grains. Gastrointestinal absorption of phosphorus is fairly proportional to intake, with approximately 65% of intake being absorbed, including a small amount that is secreted. Absorption, almost exclusively in the small intestine, occurs via a paracellular diffusive process and a vitamin D–regulated transcellular pathway. However, the impact of the change in phosphorus absorption caused by vitamin D is relatively small compared with the effect of variations in phosphorus intake.

Excretion

Despite the wide variation in phosphorus absorption dictated by oral intake, excretion matches intake, except for the needs for growth. The kidney regulates phosphorus balance, which is determined by intrarenal mechanisms and hormonal actions on the nephron.

Approximately 90% of plasma phosphate is filtered at the glomerulus, although there is some variation based on plasma phosphate and calcium concentrations. There is no significant secretion of phosphate along the nephron. Resorption of phosphate occurs mostly in the proximal tubule, although a small amount can be resorbed in the distal tubule. Normally, approximately 85% of the filtered load is resorbed. A sodium-phosphate cotransporter mediates the uptake of phosphate into the cells of the proximal tubule.

The dietary phosphorus determines the amount of phosphate resorbed by the nephron. There are both acute and chronic changes in phosphate resorption that are based on intake. Many of these changes appear to be mediated by intrarenal mechanisms that are independent of regulatory hormones. PTH, which is secreted in response to a low plasma calcium level, decreases resorption of phosphate, increasing the urinary phosphate level. This process appears to have a minimal effect during normal physiologic variation in PTH levels. However, it does have an impact in the setting of pathologic changes in PTH synthesis.

Low plasma phosphorus stimulates the 1α-hydroxylase in the kidney that converts 25-hydroxyvitamin D to 1,25-dihydroxyvitamin D (calcitriol). Calcitriol increases intestinal absorption of phosphorus and is necessary for maximal renal resorption of phosphate. The effect of a change in calcitriol on urinary phosphate is significant only when the level of calcitriol was initially low, arguing against a role for calcitriol in nonpathologic conditions.

A humoral mediator called phosphatonin inhibits renal resorption of phosphorus, causing both phosphaturia and hypophosphatemia in a variety of pathologic conditions. Phosphatonin also inhibits synthesis of calcitriol in the kidney by decreasing 1α-hydroxylase activity. Fibroblast growth factor-23 (FGF-23) has been identified as the phosphatonin that causes autosomal dominant hypophosphatemic rickets. Other putative phosphatonins include secreted frizzled-related protein 4, FGF-7, and matrix extracellular phosphoglycoprotein. The role of phosphatonins in normal physiology is not clear.

Etiology and Pathophysiology

A variety of mechanisms cause hypophosphatemia (Table 52-7). A transcellular shift of phosphorus into cells occurs with processes that stimulate cellular usage of phosphorus (glycolysis). Usually, this shift causes only a minor, transient decrease in plasma phosphorus, but if intracellular phosphorus deficiency is present, the plasma phosphorus level can decrease significantly, producing symptoms of acute hypophosphatemia. Glucose infusion stimulates insulin release, leading to entry of glucose and phosphorus into the cells. Phosphorus is then used during glycolysis and other metabolic processes. A similar phenomenon can occur during the treatment of diabetic ketoacidosis, and patients with this disorder are typically phosphorus-depleted owing to urinary phosphorus losses. Refeeding of patients with protein-calorie malnutrition causes anabolism, which leads to significant cellular demand for phosphorus. The increased phosphorus uptake for incorporation into newly synthesized compounds containing phosphorus leads to hypophosphatemia, which can be severe and symptomatic. Refeeding hypophosphatemia occurs frequently during treatment of severe anorexia nervosa. It can occur during treatment of children with malnutrition due to any cause, such as cystic fibrosis, Crohn disease, burns, neglect, chronic infection, or famine. Hypophosphatemia usually occurs within the 1st 5 days of refeeding and is prevented by a gradual increase in nutrition with appropriate phosphorus supplementation (Chapter 43). Total parenteral nutrition without adequate phosphorus can cause hypophosphatemia.

Phosphorus moves into the intracellular space during a respiratory alkalosis and during recovery from a respiratory acidosis. An acute decrease in the carbon dioxide concentration, by raising the intracellular pH, stimulates glycolysis, leading to intracellular use of phosphorus and hypophosphatemia. Because a metabolic alkalosis has less effect on the intracellular pH (carbon dioxide diffuses across cell membranes much faster than bicarbonate), there is minimal transcellular phosphorus movement with a metabolic alkalosis.

Tumors that grow rapidly, such as leukemia and lymphoma, may use large amounts of phosphorus, leading to hypophosphatemia. A similar phenomenon may occur during the hematopoietic reconstitution that follows bone marrow transplantation. In hungry bone syndrome, there is avid bone uptake of phosphorus, along with calcium and magnesium, which can produce plasma deficiency of all 3 ions. Hungry bone syndrome is most common after parathyroidectomy for hyperparathyroidism because the stimulus for bone dissolution is acutely removed, but bone synthesis continues.

Nutritional phosphorus deficiency is unusual because most foods contain phosphorus. However, infants are especially susceptible because of their high demand for phosphorus to support growth, especially of the skeleton. Very low birthweight infants have particularly rapid skeletal growth, and phosphorus deficiency and rickets may develop if they are fed human milk or formula for term infants. There is also a relative deficiency of calcium. The provision of additional calcium and phosphorus, using breast milk fortifier or special premature infant formula, prevents this complication. Phosphorus deficiency, sometimes with concomitant calcium and vitamin D deficiencies, occurs in infants who are not given enough milk or who receive a milk substitute that is nutritionally inadequate.

Antacids containing aluminum hydroxide, such as Maalox and Mylanta, bind dietary phosphorus and secreted phosphorus, preventing absorption. This process can cause phosphorus deficiency and rickets in growing children. A similar mechanism causes hypophosphatemia in patients who are overtreated for hyperphosphatemia with phosphorus binders. In children with kidney failure, the addition of dialysis to phosphorus binders increases the risk of iatrogenic hypophosphatemia in these normally hyperphosphatemic patients. This complication, which is more common in infants, can worsen renal osteodystrophy.

Excessive renal losses of phosphorus occur in a variety of inherited and acquired disorders. Because PTH inhibits the resorption of phosphorus in the proximal tubule, hyperparathyroidism causes hypophosphatemia (Chapter 567). The dominant clinical manifestation, however, is hypercalcemia, and the hypophosphatemia is usually asymptomatic. The phosphorus level in hyperparathyroidism is not extremely low, and there is no continued loss of phosphorus because a new steady state is achieved at the lower plasma phosphorus level. Renal excretion, therefore, does not exceed intake over the long term. There are occasional malignancies that produce PTH-related peptide, which has the same actions as PTH and causes hypophosphatemia and hypercalcemia.

A variety of diseases cause renal phosphate wasting, hypophosphatemia, and rickets due to excess phosphatonin (Chapter 48). These disorders include X-linked hypophosphatemic rickets, tumor-induced osteomalacia, autosomal dominant hypophosphatemic rickets, and autosomal recessive hypophosphatemic rickets. Heterozygous mutations in a phosphate transporter or a regulator of proximal tubule phosphate transport cause hypophosphatemia, osteoporosis and nephrolithiasis (hypophosphatemic nephrolithiasis/osteoporosis type 1 or 2).

Fanconi syndrome is a generalized defect in the proximal tubule leading to urinary wasting of bicarbonate, phosphorus, amino acids, uric acid, and glucose (Chapter 523). The clinical sequelae are due to the metabolic acidosis and hypophosphatemia. In children, an underlying genetic disease, most commonly cystinosis, often causes Fanconi syndrome, but it can be secondary to a variety of toxins and acquired diseases. Some patients have incomplete Fanconi syndrome, and phosphorus wasting may be one of the manifestations.

Dent disease, an X-linked disorder, can cause renal phosphorus wasting and hypophosphatemia, although the latter is not present in most cases. Other possible manifestations of Dent disease include tubular proteinuria, hypercalciuria, nephrolithiasis, rickets, and chronic renal failure. Dent disease may be secondary to mutations in a gene that encodes a chloride channel or the OCRL1 gene, which may also cause Lowe syndrome (Chapter 523.1). Hypophosphatemic rickets with hypercalciuria is a rare disorder, principally described in kindreds from the Middle East. Mutations in a sodium-phosphate cotransporter cause hypophosphatemia in this disorder, and complications may include nephrolithiasis and osteoporosis; the disorder is autosomal dominant.

Metabolic acidosis inhibits resorption of phosphorus in the proximal tubule. In addition, metabolic acidosis causes a transcellular shift of phosphorus out of cells because of intracellular catabolism. This released phosphorus is subsequently lost in the urine, leading to significant phosphorus depletion, even though the plasma phosphorus level may be normal. This classically occurs in diabetic ketoacidosis in which renal phosphorus loss is further increased by the osmotic diuresis. With correction of the metabolic acidosis and the administration of insulin, both of which cause a transcellular movement of phosphorus into the cells, there is a marked decrease in the plasma phosphorus level.

Volume expansion from any cause, such as hyperaldosteronism or SIADH, inhibits resorption of phosphorus in the proximal tubule. This effect also occurs with high rates of intravenous fluids. Thiazide and loop diuretics can increase renal phosphorus excretion, but the increase is seldom clinically significant. Glycosuria and glucocorticoids inhibit renal conservation of phosphorus. Hypophosphatemia is common after kidney transplantation as a result of urinary phosphorus losses. Possible explanations include pre-existing secondary hyperparathyroidism from chronic renal failure, glucocorticoid therapy, and upregulation of phosphatonins before transplantation. The hypophosphatemia usually resolves in a few months.

Both acquired and genetic causes of vitamin D deficiency are associated with hypophosphatemia (Chapter 48). The pathogenesis is multifactorial. Vitamin D deficiency, by impairing intestinal calcium absorption, causes secondary hyperparathyroidism that leads to increased urinary phosphorus wasting. An absence of vitamin D decreases intestinal absorption of phosphorus and directly decreases renal resorption of phosphorus. The dominant clinical manifestation is rickets, although some patients have muscle weakness that may be related to phosphorus deficiency.

Alcoholism is the most common cause of severe hypophosphatemia in adults. Fortunately, many of the risk factors that predispose adult alcoholics to hypophosphatemia are not usually present in adolescents (malnutrition, antacid abuse, recurrent episodes of diabetic ketoacidosis). Hypophosphatemia often occurs in sepsis, but the mechanism is not clear. Aggressive, protracted hemodialysis, as might be used for the treatment of methanol or ethylene glycol ingestion, can cause hypophosphatemia.

Clinical Manifestations

There are acute and chronic manifestations of hypophosphatemia. Rickets occurs in children with long-term phosphorus deficiency. The clinical features of rickets are described in Chapter 48.

Severe hypophosphatemia, typically at levels <1-1.5 mg/dL, may affect every organ in the body because phosphorus has a critical role in maintaining adequate cellular energy. Phosphorus is a component of ATP and is necessary for glycolysis. With inadequate phosphorus, red blood cell 2,3-diphosphoglycerate levels decrease, impairing release of oxygen to the tissues. Severe hypophosphatemia can cause hemolysis and dysfunction of white blood cells. Chronic hypophosphatemia causes proximal muscle weakness and atrophy. In the intensive care unit, phosphorus deficiency may slow weaning from mechanical ventilation or cause acute respiratory failure. Rhabdomyolysis is the most common complication of acute hypophosphatemia, usually in the setting of an acute transcellular shift of phosphorus into cells in a child with chronic phosphorus depletion (anorexia nervosa). The rhabdomyolysis is actually somewhat protective, in that there is cellular release of phosphorus. Other manifestations of severe hypophosphatemia include cardiac dysfunction and neurologic symptoms, such as tremor, paresthesia, ataxia, seizures, delirium, and coma.

Diagnosis

The history and basic laboratory evaluation often suggest the etiology of hypophosphatemia. The history should investigate nutrition, medications, and familial disease. Hypophosphatemia and rickets in an otherwise healthy young child suggests a genetic defect in renal phosphorus conservation, Fanconi syndrome, inappropriate use of antacids, poor nutrition, vitamin D deficiency, or a genetic defect in vitamin D metabolism. The patient with Fanconi syndrome usually has metabolic acidosis, glycosuria, aminoaciduria, and a low plasma uric acid level. Measurement of 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D, calcium, and PTH differentiates among the various vitamin D deficiency disorders and primary renal phosphate wasting (Chapter 48). Hyperparathyroidism is easily distinguished by the presence of elevated plasma PTH and calcium values.

Treatment

The plasma phosphorus level, the presence of symptoms, the likelihood of chronic depletion, and the presence of ongoing losses dictate the approach to therapy. Mild hypophosphatemia does not require treatment unless the clinical situation suggests that chronic phosphorus depletion is present or that losses are ongoing. Oral phosphorus can cause diarrhea, so the doses should be divided. Intravenous therapy is effective in patients who have severe deficiency or who cannot tolerate oral medications. Intravenous phosphorus is available as either sodium phosphate or potassium phosphate, with the choice usually based on the patient’s plasma potassium level. Starting doses are 0.08-0.16 mmol/kg over 6 hr. The oral preparations of phosphorus are available with various ratios of sodium and potassium. This is an important consideration because some patients may not tolerate the potassium load, whereas supplemental potassium may be helpful in some diseases, such as Fanconi syndrome and malnutrition. Oral maintenance dosages are 2-3 mmol/kg/day in divided doses.

Increasing dietary phosphorus is the only intervention needed in infants with inadequate intake. Other patients may also benefit from increased dietary phosphorus, usually from dairy products. Phosphorus-binding antacids should be discontinued in patients with hypophosphatemia. Certain diseases require specific therapy (Chapter 48).

Etiology and Pathophysiology

Renal insufficiency is the most common cause of hyperphosphatemia, with the severity proportional to the degree of kidney impairment (Chapter 529). This occurs because gastrointestinal absorption of the large dietary intake of phosphorus is unregulated, and the kidneys normally excrete this phosphorus. As renal function deteriorates, increased excretion of phosphorus is able to compensate. When kidney function is <30% of normal, hyperphosphatemia usually develops, although the time of its development may vary considerably according to dietary phosphorus absorption. Many of the other causes of hyperphosphatemia are more likely to develop in the setting of renal insufficiency (Table 52-8).

Cellular content of phosphorus is high relative to plasma phosphorus, and cell lysis can release substantial phosphorus. This is the etiology of hyperphosphatemia in tumor lysis syndrome, rhabdomyolysis, and acute hemolysis. These disorders cause concomitant potassium release and the risk of hyperkalemia. Additional features of tumor lysis and rhabdomyolysis are hyperuricemia and hypocalcemia, whereas indirect hyperbilirubinemia and elevated lactate dehydrogenase values are often present with hemolysis. An elevated CPK level is suggestive of rhabdomyolysis. During lactic acidosis or diabetic ketoacidosis, usage of phosphorus by cells decreases, and phosphorus shifts into the extracellular space. This problem reverses when the underlying problem is corrected, and especially with diabetic ketoacidosis, patients subsequently become hypophosphatemic as a result of previous renal phosphorus loss.

Excessive intake of phosphorus is especially dangerous in children with renal insufficiency. Neonates are at risk because renal function is normally reduced during the 1st few months of life. In addition, they may erroneously be given doses of phosphorus that are meant for an older child or adult. In infants fed cow’s milk, which has higher phosphorus content than breast milk or formula, hyperphosphatemia may develop. Fleet Enema has a high amount of phosphorus that can be absorbed, especially in the patient with an ileus. Infants and children with Hirschsprung disease are especially vulnerable. There is often associated hypernatremia owing to sodium absorption and water loss from diarrhea. Sodium phosphorus laxatives may cause hyperphosphatemia if the dose is excessive or if renal insufficiency is present. Hyperphosphatemia occurs in children who receive overly aggressive treatment for hypophosphatemia. Vitamin D intoxication causes excessive gastrointestinal absorption of both calcium and phosphorus, and the suppression of PTH by hypercalcemia decreases renal phosphorus excretion.

The absence of PTH in hypoparathyroidism or PTH responsiveness in pseudohypoparathyroidism causes hyperphosphatemia because of increased resorption of phosphorus in the proximal tubule of the kidney (Chapters 565 and 566). The associated hypocalcemia is responsible for the clinical symptoms. The hyperphosphatemia in hyperthyroidism or acromegaly is usually minor. It is secondary to increased resorption of phosphorus in the proximal tubule due to the actions of thyroxine or growth hormone. Excessive thyroxine can also cause bone resorption, which may contribute to the hyperphosphatemia and cause hypercalcemia. Patients with familial tumoral calcinosis, a rare autosomal recessive disorder, have hyperphosphatemia due to decreased renal phosphate excretion and heterotopic calcifications. The disease may be secondary to mutations in the genes for a glycosyltransferase, the phosphatonin FGF-23, or the gene for Klotho, which encodes the co-receptor for FGF-23.

Clinical Manifestations

The principal clinical consequences of hyperphosphatemia are hypocalcemia and systemic calcification. The hypocalcemia is probably due to tissue deposition of calcium-phosphorus salt, inhibition of 1,25-dihydroxyvitamin D production, and decreased bone resorption. Symptomatic hypocalcemia is most likely to occur when the phosphorus level increases rapidly or when diseases predisposing to hypocalcemia are present (chronic renal failure, rhabdomyolysis). Systemic calcification occurs because the solubility of phosphorus and calcium in the plasma is exceeded. This is believed to happen when plasma calcium × plasma phosphorus, both measured in mg/dL, is more than 70. Clinically, this condition is often apparent in the conjunctiva, where it manifests as a foreign body feeling, erythema, and injection. More ominous manifestations are hypoxia from pulmonary calcification and renal failure from nephrocalcinosis.

Diagnosis

Plasma creatinine and BUN levels should be assessed in any patient with hyperphosphatemia. The history should focus on intake of phosphorus and the presence of chronic diseases that may cause hyperphosphatemia. Measurement of potassium, uric acid, calcium, lactate dehydrogenase, bilirubin, and CPK may be indicated if rhabdomyolysis, tumor lysis, or hemolysis is suspected. With mild hyperphosphatemia and significant hypocalcemia, measurement of the serum PTH level distinguishes between hypoparathyroidism and pseudohypoparathyroidism.

Treatment

The treatment of acute hyperphosphatemia depends on its severity and etiology. Mild hyperphosphatemia in a patient with reasonable renal function spontaneously resolves; the resolution can be accelerated by dietary phosphorus restriction. If kidney function is not impaired, then intravenous fluids can enhance renal phosphorus excretion. For more significant hyperphosphatemia or a situation such as tumor lysis or rhabdomyolysis, in which endogenous phosphorus generation is likely to continue, addition of an oral phosphorus binder prevents absorption of dietary phosphorus and can remove phosphorus from the body by binding what is normally secreted and absorbed by the gastrointestinal tract. Phosphorus binders are most effective when given with food. Binders containing aluminum hydroxide are especially efficient, but calcium carbonate is an effective alternative and may be preferred if there is a need to treat concomitant hypocalcemia. Preservation of renal function, for example with high urine flow in rhabdomyolysis or tumor lysis, is an important adjunct because it will permit continued excretion of phosphorus. If the hyperphosphatemia is not responding to conservative management, especially if renal insufficiency is supervening, then dialysis may be necessary to increase phosphorus removal.

Dietary phosphorus restriction is necessary for diseases causing chronic hyperphosphatemia. However, such diets are often difficult to follow, given the abundance of phosphorus in a variety of foods. Dietary restriction is often sufficient in conditions such as hypoparathyroidism and mild renal insufficiency. For more problematic hyperphosphatemia, such as with moderate renal insufficiency and end-stage renal disease, phosphorus binders are usually necessary. They include calcium carbonate, calcium acetate, and sevelamer hydrochloride. Aluminum-containing phosphorus binders are no longer used in chronic renal insufficiency because of the risk of aluminum toxicity. Dialysis directly removes phosphorus from the blood in patients with end-stage renal disease, but it is only an adjunct to dietary restriction and phosphorus binders, in that elimination of phosphorus by dialysis is not efficient enough to keep up with normal dietary intake.

Introduction and Terminology

Close regulation of pH is necessary for cellular enzymes and other metabolic processes, which function optimally at normal pH. Chronic, mild derangements in acid-base status may interfere with normal growth and development, whereas acute, severe changes in pH can be fatal. Control of acid-base balance depends on the kidneys, the lungs, and intracellular and extracellular buffers.

A normal pH is 7.35-7.45. There is an inverse relationship between the pH and the hydrogen ion concentration. At a pH of 7.40, the hydrogen ion concentration is 40 nmol/L. A normal serum sodium concentration, 140 mEq/L, is 1 million times higher. Maintaining a normal pH is necessary because hydrogen ions are highly reactive and are especially likely to combine with proteins, altering their function.

An acid is a substance that releases (“donates”) a hydrogen ion (H⁺). A base is a substance that accepts a hydrogen ion. An acid (HA) can dissociate into a hydrogen ion and a conjugate base (A⁻), as follows:

A strong acid is highly dissociated, so in this reaction, there is little HA. A weak acid is poorly dissociated; not all of the hydrogen ions are released from HA. A⁻ acts as a base when the reaction moves to the left. These reactions are in equilibrium. When HA is added to the system, there is dissociation of some HA until the concentrations of H⁺ and A⁻ increase enough that a new equilibrium is reached. Addition of hydrogen ions causes a decrease in A⁻ and an increase in HA. Addition of A⁻ causes a decrease in hydrogen ions and an increase in HA.

Buffers are substances that attenuate the change in pH that occurs when acids or bases are added to the body. Given the extremely low concentration of hydrogen ions in the body at physiologic pH, without buffers, a small amount of hydrogen ions could cause a dramatic decline in the pH. Buffers prevent the decrease in pH by binding the added hydrogen ions, as follows:

The increase in hydrogen ion concentration drives this reaction to the right. Similarly, when base is added to the body, buffers prevent the pH from increasing by releasing hydrogen ions, as follows:

The best buffers are weak acids and bases. This is because a buffer works best when it is 50% dissociated (half HA and half A⁻). The pH at which a buffer is 50% dissociated is its pK. The best physiologic buffers have a pK close to 7.40. The concentration of a buffer and its pK determine the buffer’s effectiveness (buffering capacity). When the pH is lower than the pK of a buffer, there is more HA than A⁻. When the pH is higher than the pK, there is more A⁻ than HA.

PHYSIOLOGIC BUFFERS. The bicarbonate and nonbicarbonate buffers protect the body against major changes in pH. The bicarbonate buffer system is routinely monitored clinically. The bicarbonate buffer system is based on the relationship between carbon dioxide (CO₂) and bicarbonate (HCO₃⁻):

Carbon dioxide acts as an acid in that, after combining with water, it releases a hydrogen ion; bicarbonate acts as its conjugate base in that it accepts a hydrogen ion. The pK of this reaction is 6.1. The Henderson-Hasselbalch equation expresses the relationship among pH, pK, and the concentrations of an acid and its conjugate base. This relationship is valid for any buffer. The Henderson-Hasselbalch equation for bicarbonate and carbon dioxide is as follows:

The Henderson-Hasselbalch equation for the bicarbonate buffer system has 3 variables: pH, [HCO₃⁻], and [CO₂]. Thus, if any 2 of these variables are known, it is possible to calculate the 3rd. When one is using the Henderson-Hasselbalch equation, it is important that carbon dioxide and bicarbonate have the same units. Carbon dioxide is reported clinically as mm Hg and must be multiplied by its solubility constant, 0.03 mmol/L/ mm Hg, before the Henderson-Hasselbalch equation can be used. Mathematical manipulation of the Henderson-Hasselbalch equation produces the following relationship:

At a normal hydrogen ion concentration of 40 nmol (pH 7.40), the PCO₂, which is expressed as mm Hg in this equation, is 40 when the bicarbonate concentration is 24 mEq/L. This equation emphasizes that the hydrogen ion concentration, and hence pH, can be determined by the ratio of PCO₂ and the bicarbonate concentration.

The bicarbonate buffer system is very effective as a result of the high concentration of bicarbonate in the body (24 mEq/L) and the fact that it is an open system. The remaining body buffers are in a closed system. The bicarbonate buffer system is an open system because the lungs increase carbon dioxide excretion when the blood carbon dioxide concentration increases. When acid is added to the body, the following reaction occurs:

In a closed system, the CO₂ would increase. The higher CO₂ concentration would lead to an increase in the reverse reaction:

This would increase the concentration of hydrogen ions, limiting the buffering capacity of bicarbonate. However, because the lungs excrete the excess carbon dioxide, the reverse reaction does not increase; this fact enhances the buffering capacity of bicarbonate. The same principle holds with the addition of base, because the lungs decrease carbon dioxide excretion and prevent the level of carbon dioxide from falling. The lack of change in carbon dioxide concentration dramatically increases the buffering capacity of bicarbonate.

The nonbicarbonate buffers include proteins, phosphate, and bone. Protein buffers consist of extracellular proteins, mostly albumin and intracellular proteins, including hemoglobin. Proteins are effective buffers, largely due to the presence of the amino acid histidine, which has a side chain that can bind or release hydrogen ions. The pK of histidine varies slightly, depending on its position in the protein molecule, but its average pK is approximately 6.5. This is close enough to a normal pH (7.4) to make histidine an effective buffer. Hemoglobin and albumin have 34 and 16 histidine molecules, respectively.

Phosphate can bind up to 3 hydrogen molecules, so it can exist as PO₄³⁻, HPO₄²⁻, H₂PO₄¹⁻, or H₃PO₄. However, at a physiologic pH, most phosphate exists as either HPO₄²⁻ or H₂PO₄¹⁻. H₂PO₄¹⁻ is an acid, and HPO₄²⁻ is its conjugate base:

The pK of this reaction is 6.8, making phosphate an effective buffer. The concentration of phosphate in the extracellular space is relatively low, limiting the overall buffering capacity of phosphate; it is less important than albumin. However, phosphate is found at a much higher concentration in the urine, where it is an important buffer. In the intracellular space, most phosphate is covalently bound to organic molecules (ATP), but it still serves as an effective buffer.

Bone is an important buffer. Bone is basic—it is composed of compounds such as sodium bicarbonate and calcium carbonate—and thus, dissolution of bone releases base. This release can buffer an acid load, although at the expense of bone density, if it occurs over an extended period. In contrast, bone formation, by consuming base, helps buffer excess base.

Clinically, we measure the extracellular pH, but it is the intracellular pH that affects cell function. Measurement of the intracellular pH is unnecessary because changes in the intracellular pH parallel the changes in the extracellular pH. However, the change in the intracellular pH tends to be less than the change in the extracellular pH because of the greater buffering capacity in the intracellular space.

Renal Mechanisms

The kidneys regulate the serum bicarbonate concentration by modifying acid excretion in the urine. This requires a 2-step process. First, the renal tubules resorb the bicarbonate that is filtered at the glomerulus. Second, there is tubular secretion of hydrogen ions. The urinary excretion of hydrogen ions generates bicarbonate that neutralizes endogenous acid production. The tubular actions necessary for renal acid excretion occur throughout the nephron (Web Fig. 52-4).

Web Figure 52-4 Tubular sites involved in acid-base balance. The proximal tubule is the site where most filtered bicarbonate is reclaimed, even though other sites along the nephron, especially the thick ascending limb of the loop of Henle, resorb some of the filtered bicarbonate. The collecting duct is the principal location for the hydrogen ion secretion that acidifies the urine. The proximal tubule generates the ammonia that serves as a urinary buffer in the collecting duct.

The resorption of filtered bicarbonate is a necessary 1st step in renal regulation of the acid-base balance. A normal adult has a GFR of approximately 180 L/24 hr. This fluid enters Bowman space with a bicarbonate concentration that is essentially identical to the plasma concentration, normally 24 mEq/L. Multiplying 180 L by 24 mEq/L indicates that >4,000 mEq of bicarbonate enters Bowman space each day. This bicarbonate, if not reclaimed along the nephron, would be lost in the urine and would cause a profound metabolic acidosis.

The proximal tubule reclaims approximately 85% of the filtered bicarbonate. The final 15% is reclaimed beyond the proximal tubule, mostly in the ascending limb of the loop of Henle (Web Fig. 52-5). Bicarbonate molecules are not transported from the tubular fluid into the cells of the proximal tubule. Rather, hydrogen ions are secreted into the tubular fluid, leading to conversion of filtered bicarbonate into carbon dioxide and water. The secretion of hydrogen ions by the cells of the proximal tubule is coupled to generation of intracellular bicarbonate, which is transported across the basolateral membrane of the proximal tubule cell and enters the capillaries. The bicarbonate produced in the cell replaces the bicarbonate filtered at the glomerulus.

Web Figure 52-5 Resorption of filtered bicarbonate in the proximal tubule. The Na⁺,K⁺-ATPase (1) excretes sodium across the basolateral cell membrane, maintaining a low intracellular sodium concentration. The low intracellular sodium concentration provides the energy for the Na⁺,H⁺ antiporter (2), which exchanges sodium from the tubular lumen for intracellular hydrogen ions. The hydrogen ions that are secreted into the tubular lumen then combine with filtered bicarbonate to generate carbonic acid. CO₂ and water are produced from carbonic acid (H₂CO₃). This reaction is catalyzed by luminal carbonic anhydrase (3). CO₂ diffuses into the cell and combines with OH⁻ ions to generate bicarbonate. This reaction is catalyzed by an intracellular carbonic anhydrase (4). The dissociation of water generates an OH⁻ ion and an H⁺ ion. The Na⁺,H⁺ antiporter (2) secretes the hydrogen ions. Bicarbonate ions cross the basolateral membrane and enter the blood via the 3HCO₃⁻/1Na⁺ cotransporter (5). The energy for the 3HCO₃⁻/1Na⁺ cotransporter comes from the negatively charged cell interior, which makes it electrically favorable to transport a net negative charge (i.e., 3 bicarbonates and only 1 sodium) out of the cell.

Increased bicarbonate resorption by the cells of the proximal tubule—the result of increased hydrogen ion secretion—occurs in a variety of clinical situations. Volume depletion increases bicarbonate resorption. This is partially mediated by activation of the renin-angiotensin system; angiotensin II increases bicarbonate resorption. Increased bicarbonate resorption in the proximal tubule is one of the mechanisms that accounts for the metabolic alkalosis that may occur in some patients with volume depletion. Other stimuli that increase bicarbonate resorption include hypokalemia and an increased PCO₂. This partially explains the observations that hypokalemia causes a metabolic alkalosis and that a respiratory acidosis leads to a compensatory increase in serum bicarbonate concentration.

Stimuli that decrease bicarbonate resorption in the proximal tubule may cause a decrease in the serum bicarbonate concentration. A decrease in the PCO₂ (respiratory alkalosis) decreases proximal tubule bicarbonate resorption, partially mediating the decrease in serum bicarbonate concentration that compensates for a respiratory alkalosis. PTH decreases proximal tubule bicarbonate resorption; hyperparathyroidism may cause a mild metabolic acidosis. A variety of medications and diseases cause a metabolic acidosis by impairing bicarbonate resorption in the proximal tubule. Examples are the medication acetazolamide, which directly inhibits carbonic anhydrase, and the many disorders that cause proximal RTA (Chapter 523).

After reclaiming filtered bicarbonate, the kidneys perform the 2nd step in renal acid-base handling, the excretion of the acid created by endogenous acid production. Excretion of acid occurs mostly in the collecting duct, with a small role for the distal tubule.

Along with secretion of hydrogen ions by the tubular cells lining the collecting duct, adequate excretion of endogenous acid requires the presence of urinary buffers. The hydrogen pumps in the collecting duct cannot lower the urine pH below 4.5. The hydrogen ion concentration at pH 4.5 is <0.04 mEq/L; it would require >25 L of water with a pH of 4.5 to excrete 1 mEq of hydrogen ions. A 10-kg child, with an endogenous acid production of 20 mEq of hydrogen ions each day, would need to have a daily urinary output of >500 L without the presence of urinary buffers. As in the blood, buffers in the urine attenuate the decrease in pH that occurs with the addition of hydrogen ions. The 2 principal urinary buffers are phosphate and ammonia.

Urinary phosphate is proportional to dietary intake. Whereas most of the phosphate filtered at the glomerulus is resorbed in the proximal tubule, the urinary phosphate concentration is usually much greater than the serum phosphate concentration. This arrangement allows phosphate to serve as an effective buffer via the following reaction:

The pK of this reaction is 6.8, making phosphate an effective buffer as the urinary pH decreases from 7.0 to 5.0 within the collecting duct. Although phosphate is an effective buffer, its buffering capacity is limited by its concentration; there is no mechanism for increasing urinary phosphate excretion in response to changes in acid-base status.

In contrast, ammonia production can be modified, allowing for regulation of acid excretion. The buffering capacity of ammonia is based on the reaction of ammonia with hydrogen ions to form ammonium:

The cells of the proximal tubule are the source of the excreted ammonia, mostly through metabolism of glutamine via the following reactions:

The metabolism of glutamine generates 2 ammonium ions. In addition, the metabolism of α-ketoglutarate generates 2 bicarbonate molecules. The ammonium ions are secreted into the lumen of the proximal tubule, whereas the bicarbonate molecules exit the proximal tubule cells via the basolateral Na⁺,3HCO₃⁻ cotransporter (see Web Fig. 52-4). This arrangement would seem to accomplish the goal of excreting hydrogen ions (as NH₄⁺) and regenerating bicarbonate molecules. However, the ammonium ions secreted in the proximal tubule do not remain within the tubular lumen. Cells of the TAL of the loop of Henle resorb the ammonium ions. The result is that there is a high medullary interstitial concentration of ammonia, but the tubular fluid entering the collecting duct does not have significant amounts of ammonium ions. Moreover, the hydrogen ions that were secreted with ammonia, as ammonium ions, in the proximal tubule enter the bloodstream, canceling the effect of the bicarbonate generated in the proximal tubule. The excretion of ammonium ions, and hence of hydrogen ions, depends on the cells of the collecting duct.

The cells of the collecting duct secrete hydrogen ions and regenerate bicarbonate, which is returned to the bloodstream (Web Fig. 52-6). This bicarbonate neutralizes endogenous acid production. Phosphate and ammonia buffer the hydrogen ions secreted by the collecting duct. Ammonia is an effective buffer because of the high concentrations in the medullary interstitium and because the cells of the collecting duct are permeable to ammonia but not to ammonium. As ammonia diffuses into the lumen of the collecting duct, the low urine pH causes almost all of the ammonia to be converted into ammonium. This process maintains a low luminal ammonia concentration. Because the luminal pH is lower than the pH in the medullary interstitium, there is a higher concentration of ammonia within the medullary interstitium than in the tubular lumen, favoring movement of ammonia into the tubular lumen. Even though the concentration of ammonium in the tubular lumen is higher than in the interstitium, the cells of the collecting duct are impermeable to ammonium, preventing back-diffusion of ammonium out of the tubular lumen and permitting ammonia to be an effective buffer.

Web Figure 52-6 Secretion of hydrogen ions in the collecting duct. The dissociation of water generates an OH⁻ ion and an H⁺ ion. The H⁺-ATPase (1) secretes hydrogen ions into the tubular lumen. Bicarbonate is formed when an OH⁻ ion combines with CO₂ in a reaction mediated by carbonic anhydrase (2). Bicarbonate ions cross the basolateral membrane and enter the blood via the HCO₃⁻/Cl⁻ exchanger (3). The hydrogen ions in the tubular lumen are buffered by phosphate and ammonia (NH₃). NH₃ can diffuse from the peritubular fluid into the tubular lumen, but ammonium (NH₄⁺) cannot pass through the cells of the collecting duct.

The kidneys adjust hydrogen ion excretion according to physiologic needs. There is variation in endogenous acid production, largely due to diet and to pathophysiologic stresses, such as diarrheal losses of bicarbonate, which increase the need for acid excretion. Hydrogen excretion is increased by upregulation of hydrogen ion secretion in the collecting duct, causing the pH of the urine to decrease. This response is fairly prompt, occurring within hours of an acid load, but it is limited by the buffering capacity of the urine; the hydrogen pumps in the collecting duct cannot lower the pH to <4.5. A more significant increase in acid excretion requires upregulation of ammonia production by the proximal tubule so that more ammonia is available to serve as a buffer in the tubular lumen of the collecting duct. This response to a low serum pH reaches its maximum within 5-6 days; ammonia excretion can increase approximately 10-fold over the baseline value.

Acid excretion by the collecting duct increases in a number of different clinical situations. The extracellular pH is the most important regulator of renal acid excretion. A decrease in the extracellular pH from either a respiratory or a metabolic acidosis causes an increase in renal acid excretion. Aldosterone stimulates hydrogen ion excretion in the collecting duct, causing an increase in the serum bicarbonate concentration. This explains the metabolic alkalosis that occurs with primary hyperaldosteronism or secondary hyperaldosteronism due to volume depletion. Hypokalemia increases acid secretion, by both stimulating ammonia production in the proximal tubule and increasing hydrogen ion secretion in the collecting duct. Hypokalemia therefore tends to produce a metabolic alkalosis. Hyperkalemia has the opposite effects, which may cause a metabolic acidosis.

In patients with an increased pH, the kidney has 2 principal mechanisms for correcting the problem. First, less bicarbonate is resorbed in the proximal tubule, leading to an increase in urinary bicarbonate losses. Second, in a limited number of specialized cells, the process for secretion of hydrogen ions by the collecting duct (see Web Fig. 52-6) can be reversed, leading to secretion of bicarbonate into the tubular lumen and secretion of hydrogen ions into the peritubular fluid, where they enter the bloodstream.

Terminology

Acidemia is a pH below normal (<7.35), and alkalemia is a pH above normal (>7.45). An acidosis is a pathologic process that causes an increase in the hydrogen ion concentration, and an alkalosis is a pathologic process that causes a decrease in the hydrogen ion concentration. Whereas acidemia is always accompanied by an acidosis, a patient can have an acidosis and a low, normal, or high pH. For example, a patient may have a mild metabolic acidosis but a simultaneous, severe respiratory alkalosis; the net result may be alkalemia. Acidemia and alkalemia indicate the pH abnormality; acidosis and alkalosis indicate the pathologic process that is taking place.

A simple acid-base disorder is a single primary disturbance. During a simple metabolic disorder, there is respiratory compensation. With a metabolic acidosis, the decrease in the pH increases the ventilatory drive, causing a decrease in the PCO₂. The decrease in the carbon dioxide concentration leads to an increase in the pH. This appropriate respiratory compensation is expected with a primary metabolic acidosis. Despite the decrease in the carbon dioxide concentration, appropriate respiratory compensation is not a respiratory alkalosis, even though it is sometimes erroneously called a compensatory respiratory alkalosis. A low PCO₂ can be due either to a primary respiratory alkalosis or to appropriate respiratory compensation for a metabolic acidosis. Appropriate respiratory compensation also occurs with a primary metabolic alkalosis, although in this case the carbon dioxide concentration increases to attenuate the increase in the pH. The respiratory compensation for a metabolic process happens quickly and is complete within 12-24 hr; it cannot overcompensate for or normalize the pH.

During a primary respiratory process, there is metabolic compensation, mediated by the kidneys. The kidneys respond to a respiratory acidosis by increasing hydrogen ion excretion, thereby increasing bicarbonate generation and raising the serum bicarbonate concentration. The kidneys increase bicarbonate excretion to compensate for a respiratory alkalosis; the serum bicarbonate concentration decreases. Unlike respiratory compensation, which occurs rapidly, it takes 3-4 days for the kidneys to complete appropriate metabolic compensation. There is, however, a small and rapid compensatory change in the bicarbonate concentration during a primary respiratory process. The expected appropriate metabolic compensation for a respiratory disorder depends on whether the process is acute or chronic.

A mixed acid-base disorder is present when there is more than 1 primary acid-base disturbance. An infant with bronchopulmonary dysplasia may have a respiratory acidosis from chronic lung disease and a metabolic alkalosis from the furosemide used to treat the chronic lung disease. More dramatically, a child with pneumonia and sepsis may have severe acidemia due to a combined metabolic acidosis caused by lactic acid and respiratory acidosis caused by ventilatory failure.

There are formulas for calculating the appropriate metabolic or respiratory compensation for the 6 primary simple acid-base disorders (Table 52-9). The appropriate compensation is expected in a simple disorder; it is not optional. If a patient does not have the appropriate compensation, then a mixed acid-base disorder is present. A patient has a primary metabolic acidosis with a serum bicarbonate concentration of 10m Eq/L. The expected respiratory compensation is a carbon dioxide concentration of 23 mm Hg ± 2 (1.5 × 10 + 8 ± 2 = 23 ± 2; see Table 52-9). If the patient’s carbon dioxide concentration is >25 mm Hg, a concurrent respiratory acidosis is present; the carbon dioxide concentration is higher than expected. A patient may have a respiratory acidosis despite a carbon dioxide level below the “normal” value of 35-45 mm Hg. In this example, a carbon dioxide concentration <21 mm Hg indicates a concurrent respiratory alkalosis; the carbon dioxide concentration is lower than expected.

Table 52-9 APPROPRIATE COMPENSATION DURING SIMPLE ACID-BASE DISORDERS

Diagnosis

A systematic evaluation of an arterial blood gas sample, combined with the clinical history, can usually explain the patient’s acid-base disturbance. Assessment of an arterial blood gas sample requires knowledge of normal values (Table 52-10). In most cases, this is accomplished via a 3-step process (Fig. 52-3):

Table 52-10 NORMAL VALUES OF ARTERIAL BLOOD GASES

Figure 52-3 Three-step process for interpreting acid-base disturbances. In step 1, determine whether the pH is low (acidemia) or high (alkalemia). In step 2, establish an explanation for the acidemia or alkalemia. In step 3, calculate the expected compensation (see Table 52-9) and determine whether a mixed disturbance is present. Met. alk., metabolic alkalosis; Met. Acid., metabolic acidosis; Resp. Alk., respiratory alkalosis; Resp. Acid., respiratory acidosis.

Most patients with an acid-base disturbance have an abnormal pH, although there are 2 exceptions. The 1st exception is in the patient with a mixed disorder, wherein the 2 processes have opposite effects on pH (a metabolic acidosis and a respiratory alkalosis) and cause changes in the hydrogen ion concentration that are comparable in magnitude, albeit opposite. The 2nd exception is in the patient with a simple chronic respiratory alkalosis; in some instances, the appropriate metabolic compensation is enough to normalize the pH. In both of these situations, the presence of an acid-base disturbance is deduced because of the abnormal carbon dioxide and/or bicarbonate levels. Determining the acid-base disturbance in these situations requires proceeding to the 3rd step of this process.

The 2nd step requires inspection of the serum bicarbonate and carbon dioxide concentrations to determine a cause of the abnormal pH (see Fig. 52-3). In most cases, there is only 1 obvious explanation for the abnormal pH. In some mixed disorders, however, there may be 2 possibilities (a high PCO₂ and a low [HCO₃⁻] in a patient with acidemia). In such cases, the patient has 2 causes for abnormal pH (a metabolic acidosis and a respiratory acidosis, in this instance), and it is unnecessary to proceed to the 3rd step.

The 3rd step requires determining whether the patient’s compensation is appropriate. It is assumed that the primary disorder was diagnosed in the 2nd step, and the expected compensation is calculated (see Table 52-9). If the compensation is appropriate, then a simple acid-base disorder is present. If the compensation is not appropriate, then a mixed disorder is present. The identity of the 2nd disorder is determined by deciding whether the compensation is too little or too much compared with what was expected (see Fig. 52-3).

The history is always useful in evaluating and diagnosing patients with acid-base disturbances. It is especially helpful in a respiratory process. The expected metabolic compensation for a respiratory process changes according to whether the process is acute or chronic, which can be deduced only from the history. The metabolic compensation for an acute respiratory acidosis is less than that for a chronic respiratory acidosis. In a patient with a respiratory acidosis, a small increase in the bicarbonate concentration would be consistent with a simple acute respiratory acidosis or a mixed disorder (a chronic respiratory acidosis and a metabolic acidosis). Only the history can differentiate among the possibilities. Knowledge of the length of the respiratory process and the presence or absence of a risk factor for a metabolic acidosis (diarrhea) allows the correct conclusion to be reached.

Etiology and Pathophysiology

There are many causes of a metabolic acidosis (Table 52-11), which occur via 3 basic mechanisms:

Diarrhea, the most common cause of metabolic acidosis in children, causes a loss of bicarbonate from the body. The amount of bicarbonate lost in the stool depends on the volume of diarrhea and the bicarbonate concentration of the stool, which tends to increase with more severe diarrhea. The kidneys attempt to balance the losses by increasing acid secretion, but metabolic acidosis occurs when this compensation is inadequate. Diarrhea often causes volume depletion as a result of losses of sodium and water, potentially exacerbating the acidosis by causing shock and a lactic acidosis. In addition, diarrheal losses of potassium lead to hypokalemia. Moreover, the volume depletion causes increased production of aldosterone. This increase stimulates renal retention of sodium, helping to maintain intravascular volume, but also leads to increased urinary losses of potassium, exacerbating the hypokalemia.

There are 3 forms of renal tubular acidosis (RTA): distal (type I), proximal (type II), and hyperkalemic (type IV) [Chapter 523]. In distal RTA, children may have accompanying hypokalemia, hypercalciuria, nephrolithiasis, and nephrocalcinosis. Failure to thrive due to chronic metabolic acidosis is the most common presenting complaint. Patients with distal RTA cannot acidify their urine and, thus, have a urine pH >5.5 despite a metabolic acidosis.

Proximal RTA is rarely present in isolation. In most patients, proximal RTA is part of Fanconi syndrome, a generalized dysfunction of the proximal tubule. The dysfunction leads to glycosuria, aminoaciduria, and excessive urinary losses of phosphate and uric acid. The presence of a low serum uric acid level, glycosuria, and aminoaciduria is helpful diagnostically. Chronic hypophosphatemia leads to rickets in children (Chapter 48). Rickets and/or failure to thrive may be the presenting complaint. The ability to acidify the urine is intact in proximal RTA; thus, untreated patients have a urine pH <5.5. However, bicarbonate therapy increases bicarbonate losses in the urine, and the urine pH increases.

In hyperkalemic RTA, renal excretion of acid and potassium is impaired. Hyperkalemic RTA is due to either an absence of aldosterone or an inability of the kidney to respond to aldosterone. In severe aldosterone deficiency, as occurs with congenital adrenal hyperplasia due to 21α-hydroxylase deficiency, the hyperkalemia and metabolic acidosis are accompanied by hyponatremia and volume depletion from renal salt wasting. Incomplete aldosterone deficiency causes less severe electrolyte disturbances; children may have isolated hyperkalemic RTA, hyperkalemia without acidosis, or isolated hyponatremia. Patients may have aldosterone deficiency due to decreased renin production by the kidney; renin normally stimulates aldosterone synthesis. Children with hyporeninemic hypoaldosteronism usually have either isolated hyperkalemia or hyperkalemic RTA. The manifestations of aldosterone resistance depend on the severity of the resistance. In the autosomal recessive form of pseudohypoaldosteronism type I, which is due to an absence of the sodium channel that normally responds to aldosterone, there is often severe salt wasting and hyponatremia. In contrast, the aldosterone resistance in kidney transplant recipients usually produces either isolated hyperkalemia or hyperkalemic RTA; hyponatremia is unusual. Similarly, the medications that cause hyperkalemic RTA do not cause hyponatremia. Pseudohypoaldosteronism type II, an autosomal recessive disorder also known as Gordon syndrome, is a unique cause of hyperkalemic RTA because the genetic defect causes volume expansion and hypertension.

Children with abnormal urinary tracts, usually secondary to congenital malformations, may require diversion of urine through intestinal segments. Ureterosigmoidostomy, anastomosis of a ureter to the sigmoid colon, almost always produces a metabolic acidosis and hypokalemia. Consequently, ileal conduits are now the more commonly used procedure, although there is still a risk of a metabolic acidosis.

The appropriate metabolic compensation for a chronic respiratory alkalosis is a decrease in renal acid excretion. The resultant decrease in the serum bicarbonate concentration lessens the alkalemia caused by the respiratory alkalosis. If the respiratory alkalosis resolves quickly, the patient continues to have a decreased serum bicarbonate concentration, causing acidemia due to a metabolic acidosis. This resolves over 1-2 days via increased acid excretion by the kidneys.

Lactic acidosis most commonly occurs when inadequate oxygen delivery to the tissues leads to anaerobic metabolism and excess production of lactic acid. Lactic acidosis may be secondary to shock, severe anemia, or hypoxemia. When the underlying cause of the lactic acidosis is alleviated, the liver is able to metabolize the accumulated lactate into bicarbonate, correcting the metabolic acidosis. There is normally some tissue production of lactate that is metabolized by the liver. In children with severe liver dysfunction, impairment of lactate metabolism may produce a lactic acidosis. Rarely, a metabolically active malignancy grows so fast that its blood supply becomes inadequate, with resultant anaerobic metabolism and lactic acidosis. Patients who have short bowel syndrome due to small bowel resection can have bacterial overgrowth. In these patients, excessive bacterial metabolism of glucose into D-lactic acid can cause a lactic acidosis. Lactic acidosis occurs in a variety of inborn errors of metabolism, especially those affecting mitochondrial oxidation (Chapter 81.4). Finally, medications can cause lactic acidosis. Nucleoside reverse transcriptase inhibitors that are used to treat HIV infection inhibit mitochondrial replication; lactic acidosis is a rare complication, although elevated serum lactate concentrations without acidosis are quite common. Metformin, commonly used for treating type 2 diabetes mellitus, is most likely to cause a lactic acidosis in patients with renal insufficiency. High dosages and prolonged use of propofol can cause lactic acidosis.

In insulin-dependent diabetes mellitus, inadequate insulin leads to hyperglycemia and diabetic ketoacidosis (Chapter 583). Production of acetoacetic acid and β-hydroxybutyric acid causes the metabolic acidosis. Administration of insulin corrects the underlying metabolic problem and permits conversion of acetoacetate and β-hydroxybutyrate into bicarbonate, which helps correct the metabolic acidosis. However, in some patients, urinary losses of acetoacetate and β-hydroxybutyrate may be substantial, preventing rapid regeneration of bicarbonate. In these patients, full correction of the metabolic acidosis requires renal regeneration of bicarbonate, a slower process. The hyperglycemia causes an osmotic diuresis, usually producing volume depletion, along with substantial losses of potassium, sodium, and phosphate.

In starvation ketoacidosis, the lack of glucose leads to keto acid production, which in turn can produce a metabolic acidosis, although it is usually mild as a result of increased acid secretion by the kidney. In alcoholic ketoacidosis, which is much less common in children than in adults, the acidosis usually follows a combination of an alcoholic binge with vomiting and poor intake of food. The acidosis is potentially more severe than with isolated starvation, and the blood glucose level may be low, normal, or high. Hypoglycemia and acidosis also suggest an inborn error of metabolism.

Renal failure causes a metabolic acidosis because of the need for the kidneys to excrete the acid produced by normal metabolism. With mild or moderate renal insufficiency, the remaining nephrons are usually able to compensate by increasing acid excretion. When the GFR is <20-30% of normal, the compensation is inadequate and a metabolic acidosis develops. In some children, especially those with chronic renal failure due to tubular damage, the acidosis develops at a higher GFR because of a concurrent defect in acid secretion by the distal tubule (a distal RTA).

A variety of toxic ingestions (Chapter 58) can cause a metabolic acidosis. Salicylate intoxication is now much less common because aspirin is no longer recommended for fever control in children. Acute salicylate intoxication occurs after a large overdose. Chronic salicylate intoxication is possible with gradual buildup of the drug. Especially in adults, respiratory alkalosis may be the dominant acid-base disturbance. In children, the metabolic acidosis is usually the more significant finding. Other symptoms of salicylate intoxication are fever, seizures, lethargy, and coma. Hyperventilation may be particularly marked. Tinnitus, vertigo, and hearing impairment are more likely with chronic salicylate intoxication.

Ethylene glycol, a component of antifreeze, is converted in the liver to glyoxylic and oxalic acids, causing a severe metabolic acidosis. Excessive oxalate excretion causes calcium oxalate crystals to appear in the urine, and calcium oxalate precipitation in the kidney tubules can cause renal failure. The toxicity of methanol ingestion also depends on liver metabolism; formic acid is the toxic end product that causes the metabolic acidosis and other sequelae, which include damage to the optic nerve and CNS. Symptoms may include nausea, emesis, visual impairment, and altered mental status. Toluene inhalation and paraldehyde ingestion are other potential causes of a metabolic acidosis.

Many inborn errors of metabolism cause a metabolic acidosis (Chapters 78–81). The metabolic acidosis may be due to excessive production of keto acids, lactic acid, and/or other organic anions. Some patients have accompanying hypoglycemia or hyperammonemia. In most patients, the acidosis occurs episodically, only during acute decompensations, which may be precipitated by ingestion of specific dietary substrates, the stress of a mild illness, or poor compliance with dietary or medical therapy. In a few inborn errors of metabolism, patients have a chronic metabolic acidosis.

Clinical Manifestations

The underlying disorder usually produces most of the signs and symptoms in children with a mild or moderate metabolic acidosis. The clinical manifestations of the acidosis are related to the degree of acidemia; patients with appropriate respiratory compensation and less severe acidemia have fewer manifestations than those with a concomitant respiratory acidosis. At a serum pH <7.20, there may be impaired cardiac contractility and an increased risk of arrhythmias, especially if underlying heart disease or other predisposing electrolyte disorders are present. With acidemia, there may be a decrease in the cardiovascular response to catecholamines, potentially exacerbating hypotension in children with volume depletion or shock. Acidemia causes vasoconstriction of the pulmonary vasculature, which is especially problematic in newborn infants with persistent pulmonary hypertension (Chapter 95.7).

The normal respiratory response to metabolic acidosis—compensatory hyperventilation—may be subtle with mild metabolic acidosis, but it causes discernible increased respiratory effort with worsening acidemia. The acute metabolic effects of acidemia include insulin resistance, increased protein degradation, and reduced ATP synthesis. Chronic metabolic acidosis causes failure to thrive in children. Acidemia causes potassium to move from the intracellular space to the extracellular space, thereby increasing the serum potassium concentration. Severe acidemia impairs brain metabolism, eventually resulting in lethargy and coma.

Diagnosis

The etiology of a metabolic acidosis is often apparent from the history and physical examination. Acutely, diarrhea and shock are common causes of a metabolic acidosis. Shock, which causes a lactic acidosis, is usually apparent on physical examination and can be secondary to dehydration, acute blood loss, sepsis, or heart disease. Failure to thrive suggests a chronic metabolic acidosis, as happens with renal insufficiency or RTA. New onset of polyuria occurs in children with undiagnosed diabetes mellitus and diabetic ketoacidosis. Metabolic acidosis with seizures and/or a depressed sensorium, especially in an infant, warrants consideration of an inborn error of metabolism. Meningitis and sepsis with lactic acidosis are more common explanations for metabolic acidosis with neurologic signs and symptoms. Identification of a toxic ingestion, such as of ethylene glycol or methanol, is especially important because of the potentially excellent response to specific therapy. A variety of medications can cause a metabolic acidosis; they may be prescribed or accidentally ingested. Hepatomegaly and metabolic acidosis may occur in children with sepsis, congenital or acquired heart disease, hepatic failure, or inborn errors of metabolism.

Basic laboratory tests in a child with a metabolic acidosis should include measurements of BUN, serum creatinine, serum glucose, urinalysis, and serum electrolytes. Elevated BUN and creatinine values are present in renal insufficiency, whereas an elevated BUN:creatinine ratio (>20:1) supports a diagnosis of prerenal azotemia and the possibility of poor perfusion with lactic acidosis. Metabolic acidosis, hyperglycemia, glycosuria, and ketonuria support a diagnosis of diabetic ketoacidosis. Starvation causes ketosis, but the metabolic acidosis, if present, is usually mild (HCO₃⁻ >18). In most children with ketosis due to poor intake and metabolic acidosis, there is a concomitant disorder, such as gastroenteritis with diarrhea, that explains the metabolic acidosis. Alternatively, metabolic acidosis with or without ketosis occurs in inborn errors of metabolism; patients with these disorders may have hyperglycemia, normoglycemia, or hypoglycemia. Adrenal insufficiency may cause metabolic acidosis and hypoglycemia. Metabolic acidosis with hypoglycemia also occurs with liver failure. Metabolic acidosis, normoglycemia, and glycosuria occur in children when type II RTA is part of Fanconi syndrome; the defect in resorption of glucose by the proximal tubule of the kidney causes the glycosuria.

The serum potassium level is often abnormal in children with a metabolic acidosis. Even though a metabolic acidosis causes potassium to move from the intracellular space to the extracellular space, many patients with a metabolic acidosis have a low serum potassium level owing to excessive body losses of potassium. With diarrhea, there are high stool losses of potassium and often secondary renal losses of potassium, whereas in type I or type II RTA, there are increased urinary losses of potassium. In diabetic ketoacidosis, urinary losses of potassium are high, but the shift of potassium out of cells due to lack of insulin and metabolic acidosis is especially significant. Consequently, the initial serum potassium level can be low, normal, or high, even though total body potassium is almost always decreased. The serum potassium level is usually increased in patients with acidosis due to renal insufficiency; urinary potassium excretion is impaired. The combination of metabolic acidosis, hyperkalemia, and hyponatremia occurs in patients with severe aldosterone deficiency (adrenogenital syndrome) or aldosterone resistance. Patients with less severe, type IV RTA often have only hyperkalemia and metabolic acidosis. Very ill children with metabolic acidosis may have an elevated serum potassium value as a result of a combination of renal insufficiency, tissue breakdown, and a shift of potassium from the intracellular space to the extracellular space secondary to the metabolic acidosis.

The plasma anion gap is useful for evaluating patients with a metabolic acidosis. It divides patients into 2 diagnostic groups, those with normal anion gap and those with increased anion gap. The following formula determines the anion gap:

A normal anion gap is 4-11, although there is variation among laboratories. The number of serum anions must equal the number of serum cations to maintain electrical neutrality (Fig. 52-4). The anion gap is the difference between the measured cation (sodium) and the measured anions (chloride + bicarbonate). The anion gap is also the difference between the unmeasured cations (potassium, magnesium, calcium) and the unmeasured anions (albumin, phosphate, urate, sulfate). An increased anion gap occurs when there is an increase in unmeasured anions. With a lactic acidosis, there is endogenous production of lactic acid, which is composed of positively charged hydrogen ions and negatively charged lactate anions. The hydrogen ions are largely buffered by serum bicarbonate, resulting in a decrease in the bicarbonate concentration. The hydrogen ions that are not buffered by bicarbonate cause the serum pH to decrease. The lactate anions remain, causing the increase in the anion gap.

Figure 52-4 The anion gap, which is the difference between the sodium concentration and the combined concentrations of chloride and bicarbonate (vertical line). In both a gap and a non-gap metabolic acidosis, there is a decrease in the bicarbonate concentration. There is an increase in unmeasured anions (UA) in patients with a gap metabolic acidosis. In a non-gap metabolic acidosis, there is an increase in the serum chloride concentration. UC, unmeasured cations.

An increase in unmeasured anions, along with hydrogen ion generation, is present in all causes of an increased gap metabolic acidosis (see Table 52-11). In diabetic ketoacidosis, the keto acids β-hydroxybutyrate and acetoacetate are the unmeasured anions. In renal failure, there is retention of unmeasured anions, including phosphate, urate, and sulfate. The increase in unmeasured anions in renal failure is usually less than the decrease in the bicarbonate concentration. Renal failure is thus a mix of an increased gap and a normal gap metabolic acidosis. The normal gap metabolic acidosis is especially prominent in children with renal failure due to tubular damage, as occurs with renal dysplasia or obstructive uropathy, because these patients have a concurrent RTA. The unmeasured anions in toxic ingestions vary: formate in methanol intoxication, glycolate in ethylene glycol intoxication, and lactate and keto acids in salicylate intoxication. In inborn errors of metabolism, the unmeasured anions depend on the specific etiology and may include keto acids, lactate, and other organic anions. In a few inborn errors of metabolism, the acidosis occurs without generation of unmeasured anions; thus, the anion gap is normal.

A normal anion gap metabolic acidosis occurs when there is a decrease in the bicarbonate concentration without an increase in the unmeasured anions. With diarrhea, there is a loss of bicarbonate in the stool, causing a decrease in the serum pH and bicarbonate concentration; the serum chloride concentration increases to maintain electrical neutrality (see Fig. 52-4). Hyperchloremic metabolic acidosis is an alternative term for a normal anion gap metabolic acidosis. Calculation of the anion gap is more precise than using the chloride concentration to differentiate between a normal and an increased gap metabolic acidosis, in that the anion gap directly determines the presence of unmeasured anions. Electrical neutrality dictates that the chloride concentration increases or decreases according to the serum sodium concentration, making the chloride concentration a less reliable predictor of unmeasured anions than the more direct measure, calculation of the anion gap.

Approximately 11 mEq of the anion gap is normally secondary to albumin. A 1-g/dL decrease in the albumin concentration decreases the anion gap by roughly 2.5 mEq/L. Similarly, an increase in unmeasured cations, such as calcium, potassium, and magnesium, decreases the anion gap. Conversely, a decrease in unmeasured cations is a very unusual cause of an increased anion gap. Because of these variables, the broad range of a normal anion gap, and other variables, the presence of a normal or an increased anion gap is not always reliable in differentiating among the causes of a metabolic acidosis, especially when the metabolic acidosis is mild. In some patients there is more than one explanation for the metabolic acidosis, such as the child with diarrhea and lactic acidosis due to poor perfusion. The anion gap should not be interpreted in dogmatic isolation; consideration of other laboratory abnormalities and the clinical history improves its diagnostic utility.

Treatment

The most effective therapeutic approach for patients with a metabolic acidosis is repair of the underlying disorder, if possible. The administration of insulin in diabetic ketoacidosis and the restoration of adequate perfusion with intravenous fluids in lactic acidosis due to hypovolemia or shock eventually result in normalization of the acid-base balance. In other diseases, the use of bicarbonate therapy is indicated because the underlying disorder is irreparable. Children with metabolic acidosis due to RTA or chronic renal failure require long-term base therapy. Patients with acute renal failure and metabolic acidosis need base therapy until their kidneys’ ability to excrete hydrogen normalizes. In other disorders, the cause of the metabolic acidosis eventually resolves, but base therapy is necessary during the acute illness. In salicylate poisoning, alkali administration increases renal clearance of salicylate and decreases the amount of salicylate in brain cells. Short-term base therapy is often necessary in other poisonings (ethylene glycol, methanol) and inborn errors of metabolism (pyruvate carboxylase deficiency, propionic acidemia). Some inborn errors of metabolism require long-term base therapy.

The use of base therapy in diabetic ketoacidosis and lactic acidosis is controversial; there is little evidence that it improves patient outcome, and it has a variety of potential side effects. The risks of giving sodium bicarbonate include the possibility of causing hypernatremia or volume overload. Furthermore, the patient may have overcorrection of the metabolic acidosis once the underlying disorder resolves, because metabolism of lactate or keto acids generates bicarbonate. The rapid change from acidemia to alkalemia can cause a variety of problems, including hypokalemia and hypophosphatemia. Bicarbonate therapy increases the generation of carbon dioxide, which can accumulate in patients with respiratory failure. Because carbon dioxide readily diffuses into cells, the administration of bicarbonate can lower the intracellular pH, potentially worsening cell function. Base therapy is usually reserved for children with severe acute lactic acidosis and severe diabetic ketoacidosis.

Oral base therapy is given to children with chronic metabolic acidosis. Sodium bicarbonate tablets are available for older children. Younger children generally take citrate solutions; the liver generates bicarbonate from citrate. Citrate solutions are available as sodium citrate, potassium citrate, and a 1 : 1 mix of sodium citrate and potassium citrate. The patient’s potassium needs dictate the choice. Children with type I or type II RTA may have hypokalemia and may benefit from potassium supplements, whereas most children with chronic renal failure cannot tolerate additional potassium.

Oral or intravenous base can be used in acute metabolic acidosis; intravenous therapy is generally used when a rapid response is necessary. Sodium bicarbonate may be given as a bolus, usually at a dose of 1 mEq/kg, in an emergency situation. Another approach is to add sodium bicarbonate or sodium acetate to the patient’s intravenous fluids, remembering to remove an equal amount of sodium chloride from the solution to avoid giving an excessive sodium load. Careful monitoring is mandatory so that the dose of base can be titrated appropriately. Tris-hydroxymethyl aminomethane (THAM) is an option in patients with a metabolic acidosis and a respiratory acidosis, because it neutralizes acids without releasing CO₂. THAM also diffuses into cells and therefore provides intracellular buffering.

Hemodialysis is another option for correcting a metabolic acidosis, and it is an appropriate choice in patients with renal insufficiency, especially if significant uremia or hyperkalemia is also present. Hemodialysis is advantageous for correcting the metabolic acidosis due to methanol or ethylene glycol intoxication, because hemodialysis removes the offending toxin. In addition, these patients often have a severe metabolic acidosis that does not respond easily to intravenous bicarbonate therapy. Peritoneal dialysis is another option for correcting the metabolic acidosis due to renal insufficiency, although, because it relies on lactate as the source of base, it may not correct the metabolic acidosis in patients with concomitant renal failure and lactic acidosis.

Many causes of metabolic acidosis require specific therapy. Administration of a glucocorticoid and a mineralocorticoid is necessary in patients with adrenal insufficiency. Patients with diabetic ketoacidosis require insulin therapy, whereas patients with lactic acidosis respond to measures that alleviate tissue hypoxia. Along with correction of acidosis, patients with methanol or ethylene glycol ingestion should receive an agent that prevents the breakdown of the toxic substance to its toxic metabolites. Fomepizole has supplanted ethanol as the treatment of choice. These agents work by inhibiting alcohol dehydrogenase, the enzyme that performs the 1st step in the metabolism of ethylene glycol or methanol. There are a variety of disease-specific therapies for patients with a metabolic acidosis due to an inborn error of metabolism.

Etiology and Pathophysiology

The kidneys normally respond promptly to a metabolic alkalosis by increasing base excretion. Two processes are therefore usually present to produce a metabolic alkalosis. The 1st process is the generation of the metabolic alkalosis, which requires the addition of base to the body. The 2nd process is the maintenance of the metabolic alkalosis, which requires impairment in the kidney’s ability to excrete base.

The etiologies of a metabolic alkalosis are divided into 2 categories on the basis of urinary chloride level (Table 52-12). The alkalosis in patients with a low urinary chloride level is maintained by volume depletion; thus, volume repletion is necessary for correction of the alkalosis. The volume depletion in these patients is due to losses of sodium and potassium, but the loss of chloride is usually greater than the losses of sodium and potassium combined. Because chloride losses are the dominant cause of the volume depletion, these patients require chloride to correct the volume depletion and metabolic alkalosis; they are said to have chloride-responsive metabolic alkalosis. In contrast, the alkalosis in a patient with an elevated urinary chloride concentration does not respond to volume repletion and so is termed chloride-resistant metabolic alkalosis.

Emesis or nasogastric suction results in loss of gastric fluid, which has a high content of HCl. Generation of hydrogen ions by the gastric mucosa causes simultaneous release of bicarbonate into the bloodstream. Normally, the hydrogen ions in gastric fluid are reclaimed in the small intestine (by neutralizing secreted bicarbonate). Thus, there is no net loss of acid. With loss of gastric fluid, this does not occur, and a metabolic alkalosis develops. This period is the generation phase of the metabolic alkalosis.

The maintenance phase of the metabolic alkalosis from gastric losses is due to the volume depletion (“chloride depletion” from gastric loss of HCl). Volume depletion interferes with urinary loss of bicarbonate, the normal renal response to a metabolic alkalosis. During volume depletion, several mechanisms prevent renal bicarbonate loss. First, there is a reduction in the GFR, so less bicarbonate is filtered. Second, volume depletion increases resorption of sodium and bicarbonate in the proximal tubule, limiting the amount of bicarbonate that can be excreted in the urine. This effect is mediated by angiotensin II and by adrenergic stimulation of the kidney, which are both increased in response to volume depletion. Third, the increase in aldosterone during volume depletion increases bicarbonate resorption and hydrogen ion secretion in the collecting duct.

In addition to volume depletion, gastric losses are usually associated with hypokalemia as a result of both gastric loss of potassium and, most importantly, increased urinary potassium losses. The increased urinary losses of potassium are mediated by aldosterone, through volume depletion, and by the increase in intracellular potassium secondary to the metabolic alkalosis, which causes potassium to move into the cells of the kidney, causing increased potassium excretion. Hypokalemia contributes to the maintenance of the metabolic alkalosis by decreasing bicarbonate loss. Hypokalemia increases hydrogen ion secretion in the distal nephron and stimulates ammonia production in the proximal tubule. Ammonia production enhances renal excretion of hydrogen ions.

A metabolic alkalosis can develop in patients receiving loop or thiazide diuretics. Diuretic use leads to volume depletion, which increases angiotensin II, aldosterone, and adrenergic stimulation of the kidney. Diuretics increase the delivery of sodium to the distal nephron, further enhancing acid excretion. Moreover, these diuretics cause hypokalemia, which increases acid excretion by the kidney. The increase in renal acid excretion generates the metabolic alkalosis, and the decrease in bicarbonate loss maintains it. In addition, patients who are receiving diuretics have a “contraction alkalosis.” Diuretic use causes fluid loss without bicarbonate; thus, the remaining body bicarbonate is contained in a smaller total body fluid compartment. The bicarbonate concentration increases, helping to generate the metabolic alkalosis.

Diuretics are often used in patients with edema, such as those with nephrotic syndrome, heart failure, or liver failure. In many of these patients, metabolic alkalosis resulting from diuretic use develops despite the continued presence of edema. This is because the effective intravascular volume is low, and it is the effective intravascular volume that stimulates the compensatory mechanisms that cause and maintain a metabolic alkalosis. Many of these patients have a decreased effective intravascular volume before they begin diuretic therapy, increasing the likelihood of diuretic-induced metabolic alkalosis.

Diuretic use increases chloride excretion in the urine. Consequently, while a patient is receiving diuretics, the urine chloride level is typically high (>20 mEq/L). After the diuretic effect has worn off, the urinary chloride level is low (<15 mEq/L) owing to appropriate renal chloride retention in response to volume depletion. Thus, categorization of diuretics on the basis of urinary chloride level depends on the timing of the measurement. However, the metabolic alkalosis from diuretics is clearly chloride responsive; it is corrected after adequate volume repletion. This is the rationale for including this process among the chloride-responsive causes of a metabolic alkalosis.

Most patients with diarrhea have a metabolic acidosis due to stool losses of bicarbonate. In chloride-losing diarrhea, an autosomal recessive disorder, there is a defect in the normal intestinal exchange of bicarbonate for chloride, causing excessive stool losses of chloride (Chapter 330). In addition, stool losses of hydrogen ions and potassium cause metabolic alkalosis and hypokalemia, both of which are exacerbated by increased renal hydrogen and potassium losses due to volume depletion. Treatment is with oral supplements of potassium and sodium chloride. Use of a gastric proton pump inhibitor, by decreasing gastric HCl production, reduces both the volume of diarrhea and the need for electrolyte supplementation.

An infant formula with extremely low chloride content has led to chloride deficiency and volume depletion. The infants fed this formula, which is no longer available, had a metabolic alkalosis and hypokalemia. Cystic fibrosis can rarely cause metabolic alkalosis, hypokalemia, and hyponatremia due to excessive losses of sodium chloride in sweat (Chapter 395). The volume depletion causes the metabolic alkalosis and hypokalemia through increased urinary losses, whereas the hyponatremia, a less common finding, is secondary to sodium loss combined with renal water conservation in an effort to protect the intravascular volume (“appropriate” ADH production).

A post-hypercapnic metabolic alkalosis occurs after the correction of a chronic respiratory acidosis. This is typically seen in patients with chronic lung disease who are started on mechanical ventilation. During chronic respiratory acidosis, appropriate renal compensation leads to an increase in the serum bicarbonate concentration. This elevated bicarbonate concentration, because it is still present after acute correction of the respiratory acidosis, causes a metabolic alkalosis. The metabolic alkalosis persists because the patient with a chronic respiratory acidosis is intravascularly depleted because of the chloride loss that occurred during the initial metabolic compensation for the primary respiratory acidosis. In addition, many children with a chronic respiratory acidosis receive diuretics, which further decrease the intravascular volume. The metabolic alkalosis responds to correction of the intravascular volume deficit.

The chloride-resistant causes of metabolic alkalosis can be subdivided according to blood pressure status. Patients with hypertension either have increased aldosterone levels or act as if they do. Aldosterone levels are elevated in children with adrenal adenomas or hyperplasia. Aldosterone causes renal retention of sodium, with resultant hypertension. Metabolic alkalosis and hypokalemia result from aldosterone-mediated renal excretion of hydrogen ions and potassium. The urinary chloride level is not low because these patients are volume-overloaded, not volume-depleted. The volume expansion and hypertension allow normal excretion of sodium and chloride despite the presence of aldosterone. This is known as the mineralocorticoid escape phenomenon.

In glucocorticoid-remediable aldosteronism, an autosomal dominant disorder, there is excess production of aldosterone owing to the presence of an aldosterone synthase gene that is regulated by adrenocorticotropic hormone (ACTH) (Chapter 570.8). Glucocorticoids effectively treat this disorder by inhibiting ACTH production by the pituitary, downregulating the inappropriate aldosterone production. Renovascular disease and renin-secreting tumors both cause excessive renin, leading to an increase in aldosterone, although hypokalemia and metabolic alkalosis are less common findings than hypertension. In 2 forms of congenital adrenal hyperplasia, 11β-hydroxylase deficiency and 17α-hydroxylase deficiency, there is excessive production of the mineralocorticoid 11-deoxycorticosterone (Chapters 570.2 and 570.4). Hypertension, hypokalemia, and metabolic alkalosis are more likely in 17α-hydroxylase deficiency than in 11β-hydroxylase deficiency. These disorders respond to glucocorticoids because the excess production of 11-deoxycorticosterone is under the control of ACTH.

Cushing syndrome frequently causes hypertension. Cortisol has some mineralocorticoid activity, and high levels can produce hypokalemia and metabolic alkalosis in patients with Cushing syndrome.

Cortisol can bind to the mineralocorticoid receptors in the kidney and function as a mineralocorticoid. This binding normally does not occur because 11β-hydroxysteroid dehydrogenase in the kidney converts cortisol to cortisone, which does not bind to the mineralocorticoid receptor. In 11β-hydroxysteroid dehydrogenase deficiency, also called apparent mineralocorticoid excess, however, cortisol is not converted in the kidney to cortisone. Cortisol is therefore available to bind to the mineralocorticoid receptor in the kidney and act as a mineralocorticoid. Patients with this deficiency, despite low levels of aldosterone, are hypertensive and hypokalemic, and they have a metabolic alkalosis. The same phenomenon can occur with excessive intake of natural licorice, a component of which, glycyrrhizic acid, inhibits 11β-hydroxysteroid dehydrogenase. The autosomal dominant disorder Liddle syndrome is secondary to an activating mutation of the sodium channel in the distal nephron (Chapter 525.3). Upregulation of this sodium channel is 1 of the principal actions of aldosterone. Because this sodium channel is continuously open, children with Liddle syndrome have the features of hyperaldosteronism, including hypertension, hypokalemia, and metabolic alkalosis, but low serum levels of aldosterone.

Bartter syndrome and Gitelman syndrome are autosomal recessive disorders associated with normal blood pressure, elevations of urinary chloride, metabolic alkalosis, and hypokalemia (Chapter 525). In Bartter syndrome, patients have a defect in sodium and chloride resorption in the loop of Henle. This leads to excessive urinary losses of sodium and chloride, and as in patients receiving loop diuretics, volume depletion and secondary hyperaldosteronism occur, causing hypokalemia and metabolic alkalosis. Gitelman syndrome is usually milder than Bartter syndrome. Patients have renal sodium and chloride wasting with volume depletion due mutations in the gene encoding the thiazide-sensitive sodium-chloride transporter in the distal tubule. As in patients receiving a thiazide diuretic, affected patients have volume depletion and secondary hyperaldosteronism with hypokalemia and metabolic alkalosis. Children with Gitelman syndrome have hypocalciuria and hypomagnesemia. Some patients with autosomal dominant hypoparathyroidism have hypokalemia and metabolic alkalosis due to impaired sodium and chloride resorption in the loop of Henle. EAST syndrome (epilepsy, ataxia, sensorineural deafness and tubulopathy) causes hypokalemia, metabolic alkalosis and hypokalemia.

Excessive base intake can cause a metabolic alkalosis. Affected patients do not have a low urine chloride level, unless there is associated volume depletion. In the absence of volume depletion, excess base is rapidly corrected via renal excretion of bicarbonate. Rarely, massive base intake can cause a metabolic alkalosis by overwhelming the kidney’s ability to excrete bicarbonate. This may occur in infants who are given baking soda as a “home remedy” for colic or stomach upset. Each teaspoon of baking soda has 42 mEq of sodium bicarbonate. Infants have increased vulnerability because of a lower GFR, limiting the rate of compensatory renal bicarbonate excretion. A metabolic alkalosis may also occur in patients who receive a large amount of sodium bicarbonate during cardiopulmonary resuscitation. Blood products are anticoagulated with citrate, which is converted into bicarbonate by the liver. Patients who receive large amounts of blood products may have a metabolic alkalosis. Iatrogenic metabolic alkalosis can occur as a result of acetate in total parenteral nutrition. Aggressive use of bicarbonate therapy in a child with a lactic acidosis or diabetic ketoacidosis may cause a metabolic alkalosis. This event is especially likely in a patient in whom the underlying cause of the lactic acidosis is successfully corrected (restoration of intravascular volume in a patient with severe dehydration). Once the cause of the lactic acidosis resolves, lactate can be converted by the liver into bicarbonate, which when combined with infused bicarbonate can create a metabolic alkalosis. A similar phenomenon can occur in a child with diabetic ketoacidosis because the administration of insulin allows keto acids to be metabolized, producing bicarbonate. However, this phenomenon rarely occurs because of judicious use of bicarbonate therapy in diabetic ketoacidosis and because there are usually significant pretreatment losses of keto acids in the urine, preventing massive regeneration of bicarbonate. Base administration is most likely to cause a metabolic alkalosis in patients who have an impaired ability to excrete bicarbonate in the urine. This impairment occurs in patients with concurrent volume depletion or renal insufficiency.

Clinical Manifestations

The symptoms in patients with a metabolic alkalosis are often related to the underlying disease and associated electrolyte disturbances. Children with chloride-responsive causes of metabolic alkalosis often have symptoms related to volume depletion, such as thirst and lethargy. In contrast, children with chloride-unresponsive causes may have symptoms related to hypertension.

Alkalemia causes potassium to shift into the intracellular space, producing a decrease in the extracellular potassium concentration. Alkalemia leads to increased urinary losses of potassium. Increased potassium losses are present in many of the conditions that cause a metabolic alkalosis. Therefore, most patients with a metabolic alkalosis have hypokalemia, and their symptoms may be related to the hypokalemia (Chapter 52.4).

The symptoms of a metabolic alkalosis are due to the associated alkalemia. The magnitude of the alkalemia is related to the severity of the metabolic alkalosis and the presence of concurrent respiratory acid-base disturbances. During alkalemia, the ionized calcium concentration decreases as a result of increased binding of calcium to albumin. The decrease in the ionized calcium concentration may cause symptoms of tetany (carpopedal spasm).

Arrhythmias are a potential complication of a metabolic alkalosis, and the risk for arrhythmia increases if there is concomitant hypokalemia. Alkalemia increases the risk of digoxin toxicity, and antiarrhythmic medications are less effective in the presence of alkalemia. In addition, alkalemia may decrease cardiac output. A metabolic alkalosis causes a compensatory increase in the PCO₂ by decreasing ventilation. In patients with underlying lung disease, the decrease in ventilatory drive can cause hypoxia. In patients with normal lungs, the hypoventilation seen in severe metabolic alkalosis can cause hypoxia.

Diagnosis

Measurement of the urinary chloride concentration is the most helpful test in differentiating among the causes of a metabolic alkalosis. The urine chloride level is low in patients with a metabolic alkalosis resulting from volume depletion, unless there is a defect in renal handling of chloride. The urine chloride level is superior to the urine sodium level in assessment of volume status in patients with a metabolic alkalosis, because the normal renal response to a metabolic alkalosis is to excrete bicarbonate. Because bicarbonate is negatively charged, it can be excreted only with a cation, usually sodium and potassium. Hence, a patient with a metabolic alkalosis may excrete sodium in the urine despite the presence of volume depletion, which normally causes avid sodium retention. The urine chloride level is usually a good indicator of volume status, and it differentiates among the chloride-resistant and chloride-responsive causes of a metabolic alkalosis.

Diuretics and gastric losses are the most common causes of metabolic alkalosis and are usually readily apparent from the patient history. Occasionally, metabolic alkalosis, usually with hypokalemia, may be a clue to the presence of bulimia or surreptitious diuretic use (Chapter 26). Patients with bulimia have a low urine chloride level, indicating that they have volume depletion as a result of an extrarenal etiology, but there is no alternative explanation for their volume depletion. Surreptitious diuretic use may be diagnosed by obtaining a urine toxicology screen for diuretics. The urine chloride level is increased while a patient is using diuretics but is low when the patient stops taking them. Rarely, children with mild Bartter syndrome or Gitelman syndrome are misdiagnosed as having bulimia or abusing diuretics. The urine chloride value is always elevated in Bartter syndrome and Gitelman syndrome, and the urine toxicology screen for diuretics has a negative result. Metabolic alkalosis with hypokalemia is occasionally the initial manifestation of cystic fibrosis. An elevated sweat chloride finding is diagnostic.

Patients with a metabolic alkalosis and a high urinary chloride level are subdivided according to blood pressure status. Children with normal blood pressure may have Bartter syndrome or Gitelman syndrome. Excess base administration is another diagnostic possibility, but it is usually apparent from the history. In patients with sodium bicarbonate ingestion (baking soda), which may be unreported by the parent, the metabolic alkalosis usually occurs with significant hypernatremia. In addition, unless volume depletion is superimposed, the metabolic alkalosis from base ingestion resolves itself once the source of base is eliminated.

Measuring serum concentrations of renin and aldosterone differentiates children with a metabolic alkalosis, a high urinary chloride level, and elevated blood pressure. Both renin and aldosterone are elevated in children with either renovascular disease or a renin-secreting tumor. Aldosterone is high and renin is low in patients with adrenal adenomas or hyperplasia and glucocorticoid-remediable aldosteronism. Renin and aldosterone are low in children with Cushing syndrome, Liddle syndrome, licorice ingestion, 17α-hydroxylase deficiency, 11β-hydroxylase deficiency, and 11β-hydroxysteroid dehydrogenase deficiency. An elevated 24-hr urine cortisol value is diagnostic of Cushing syndrome, which is suspected from the presence of the other classic features of this disease (Chapter 571). Elevations of 11-Deoxycorticosterone values are seen in 17α-hydroxylase deficiency and 11β-hydroxylase deficiency.

Treatment

The approach to treatment of metabolic alkalosis depends on the severity of the alkalosis and the underlying etiology. In children with a mild metabolic alkalosis ([HCO₃⁻] <32), intervention is often unnecessary, although this depends on the specific circumstances. In a child with congenital heart disease who is receiving a stable dose of a loop diuretic, a mild alkalosis does not require treatment. In contrast, intervention may be appropriate in a child with a worsening mild metabolic alkalosis due to nasogastric suction. The presence of a concurrent respiratory acid-base disturbance also influences therapeutic decision-making. A patient with a concurrent respiratory acidosis should have some increase in bicarbonate owing to metabolic compensation; thus, the severity of the pH elevation is more important than the bicarbonate concentration. In contrast, a patient with a respiratory alkalosis and a metabolic alkalosis is at risk for severe alkalemia; treatment may be indicated, even if the increase in bicarbonate value is only mild.

Intervention is usually necessary in children with moderate or severe metabolic alkalosis. The most effective approach is to address the underlying etiology. In some children, nasogastric suction may be decreased or discontinued. Alternatively, the addition of a gastric proton pump inhibitor reduces gastric secretion and losses of HCl. Diuretics are an important cause of metabolic alkalosis, and if a change is tolerated, they should be eliminated or the dose reduced. Adequate potassium supplementation or the addition of a potassium-sparing diuretic is also helpful in a child with a metabolic alkalosis due to diuretics. Potassium-sparing diuretics not only decrease renal potassium losses but, by blocking the action of aldosterone, also decrease hydrogen ion secretion in the distal nephron, increasing urinary bicarbonate excretion. Many children cannot tolerate discontinuation of diuretic therapy; thus, potassium supplementation and potassium-sparing diuretics are the principal therapeutic approach. Arginine HCl may also be used to treat chloride-responsive metabolic acidosis if sodium or potassium salts are not appropriate. Arginine HCl may raise the serum potassium levels during administration. Rarely, in cases of severe metabolic alkalosis, acetazolamide is an option. A carbonic anhydrase inhibitor, acetazolamide decreases resorption of bicarbonate in the proximal tubule, causing significant bicarbonate loss in the urine. The patient receiving this drug must be monitored closely, because acetazolamide produces major losses of potassium in the urine and increases fluid losses, potentially necessitating a reduction in dosage of other diuretics.

Most children with a metabolic alkalosis have one of the chloride-responsive etiologies. In these situations, administration of sufficient sodium chloride and potassium chloride to correct the volume deficit and the potassium deficit is necessary to correct the metabolic alkalosis. This approach may not be an option in the child who has volume depletion due to diuretics, because volume repletion may be contraindicated. Adequate replacement of gastric losses of sodium and potassium in a child with a nasogastric tube can minimize or prevent the development of the metabolic alkalosis. With adequate intravascular volume and a normal serum potassium concentration, the kidney is able to excrete the excess bicarbonate within a couple of days.

In children with the chloride-resistant causes of a metabolic alkalosis that are associated with hypertension, volume repletion is contraindicated because it would exacerbate the hypertension and would not repair the metabolic alkalosis. Ideally, treatment focuses on eliminating the excess aldosterone effect. Adrenal adenomas can be resected, licorice intake can be eliminated, and renovascular disease can be repaired. Glucocorticoid-remediable aldosteronism, 17α-hydroxylase deficiency, and 11β-hydroxylase deficiency respond to the administration of glucocorticoids. The mineralocorticoid effect of cortisol in 11β-hydroxysteroid dehydrogenase deficiency can be decreased with the use of spironolactone, which blocks the mineralocorticoid receptor. In contrast, the metabolic alkalosis in children with Liddle syndrome does not respond to spironolactone; however, either triamterene or amiloride is effective therapy because both agents block the sodium channel that is constitutively active in Liddle syndrome.

In children with Bartter syndrome and Gitelman syndrome, therapy includes oral potassium supplementation and potassium-sparing diuretics. Children with Gitelman syndrome often require magnesium supplementation, whereas children with severe Bartter syndrome often benefit from indomethacin.

Etiology and Pathophysiology

The causes of a respiratory acidosis are either pulmonary or nonpulmonary (Table 52-13). CNS disorders can decrease the activity of the central respiratory center, reducing ventilatory drive. A variety of medications and illicit drugs suppress the respiratory center. The signals from the respiratory center need to be transmitted to the respiratory muscles via the nervous system. Respiratory muscle failure can be secondary to disruption of the signal from the CNS in the spinal cord, the phrenic nerve, or the neuromuscular junction. Disorders directly affecting the muscles of respiration can prevent adequate ventilation, causing a respiratory acidosis.

Mild or moderate lung disease often causes a respiratory alkalosis as a result of hyperventilation secondary to hypoxia or stimulation of lung mechanoreceptors or chemoreceptors. Only more severe lung disease causes a respiratory acidosis. Upper airway diseases, by impairing air entry into the lungs, may decrease ventilation, producing a respiratory acidosis.

Increased production of carbon dioxide is never the sole cause of a respiratory acidosis, but it can increase the severity of the disease in a patient with decreased ventilation of carbon dioxide. Increased production of carbon dioxide occurs in patients with fever, hyperthyroidism, excess caloric intake, and high levels of physical activity. Increased respiratory muscle work also increases carbon dioxide production.

Clinical Manifestations

Patients with a respiratory acidosis are often tachypneic in an effort to correct the inadequate ventilation. Exceptions include patients with a respiratory acidosis resulting from CNS depression and patients who are on the verge of complete respiratory failure secondary to fatigue of the respiratory muscles.

The symptoms of respiratory acidosis are related to the severity of the hypercarbia. Acute respiratory acidosis is usually more symptomatic than chronic respiratory acidosis. Symptoms are also increased by concurrent hypoxia or metabolic acidosis. In a patient breathing room air, hypoxia is always present if a respiratory acidosis is present. The potential CNS manifestations of respiratory acidosis include anxiety, dizziness, headache, confusion, asterixis, myoclonic jerks, hallucinations, psychosis, coma, and seizures.

Acidemia, no matter the etiology, affects the cardiovascular system. An arterial pH <7.20 impairs cardiac contractility and the normal response to catecholamines, in both the heart and the peripheral vasculature. Hypercapnia causes vasodilation, most dramatically in the cerebral vasculature, but hypercapnia produces vasoconstriction of the pulmonary circulation. Respiratory acidosis increases the risk of cardiac arrhythmias, especially in a child with underlying cardiac disease.

Diagnosis

The history and physical findings often point to a clear etiology. For the obtunded patient with poor respiratory effort, evaluation of the CNS is often indicated. This may include imaging studies (CT or MRI) and, potentially, a lumbar puncture for CSF analysis. A toxicology screen for illicit drugs may also be appropriate. A response to naloxone is both diagnostic and therapeutic. In many of the diseases affecting the respiratory muscles, there is evidence of weakness in other muscles. Stridor is a clue that the child may have upper airway disease. Along with a physical examination, a chest radiograph is often helpful in diagnosing pulmonary disease.

In many patients, respiratory acidosis may be multifactorial. A child with bronchopulmonary dysplasia, an intrinsic lung disease, may worsen because of respiratory muscle dysfunction due to severe hypokalemia resulting from long-term diuretic therapy. Conversely, a child with muscular dystrophy, a muscle disease, may worsen because of aspiration pneumonia.

For a patient with respiratory acidosis, calculation of the gradient between the alveolar oxygen concentration and the arterial oxygen concentration, the A-aO₂ gradient, is useful for distinguishing between poor respiratory effort and intrinsic lung disease. The A-aO₂ gradient is increased if the hypoxemia is due to intrinsic lung disease (Chapter 365).

Treatment

Respiratory acidosis is best managed by treatment of the underlying etiology. In some instances, the response is very rapid, such as after the administration of naloxone to a patient with a narcotic overdose. In contrast, in the child with pneumonia, a number of days of antibiotic therapy may be required before the respiratory status improves. In many children with a chronic respiratory acidosis, there is no curative therapy, although an acute respiratory illness superimposed on a chronic respiratory condition is usually reversible.

All patients with an acute respiratory acidosis are hypoxic and therefore need to receive supplemental oxygen. Mechanical ventilation is necessary in some children with a respiratory acidosis. Children with a significant respiratory acidosis due to a CNS disease usually require mechanical ventilation because such a disorder is unlikely to respond quickly to therapy. In addition, hypercarbia causes cerebral vasodilation, and the increase in intracranial pressure can be dangerous in a child with an underlying CNS disease. Readily reversible CNS depression, such as from a narcotic overdose, may not require mechanical ventilation. Decisions on mechanical ventilation for other patients depend on a number of factors. Patients with severe hypercarbia—PCO₂ > 75 mm Hg—usually require mechanical ventilation (Chapters 65 and 366). The threshold for intubation is lower if there is concomitant metabolic acidosis, a slowly responsive underlying disease, or hypoxia that responds poorly to oxygen, or if the patient appears to be tiring and respiratory arrest seems likely.

In patients with a chronic respiratory acidosis, the respiratory drive is often less responsive to hypercarbia and more responsive to hypoxia. Hence, with chronic respiratory acidosis, excessive use of oxygen can blunt the respiratory drive and therefore increase the PCO₂. In these patients, oxygen must be used cautiously.

When possible, it is best to avoid mechanical ventilation in a patient with a chronic respiratory acidosis because extubation is often difficult. However, an acute illness may necessitate mechanical ventilation in a child with a chronic respiratory acidosis. When intubation is necessary, the PCO₂ should be lowered only to the patient’s normal baseline, and this should be done gradually. These patients normally have an elevated serum bicarbonate concentration as a result of metabolic compensation for their respiratory acidosis. A rapid lowering of the PCO₂ can cause a severe metabolic alkalosis, potentially leading to complications, including cardiac arrhythmias, decreased cardiac output, and decreased cerebral blood flow. In addition, prolonged mechanical ventilation at a normal PCO₂ causes the metabolic compensation to resolve. When the patient is subsequently extubated, the patient will no longer benefit from metabolic compensation, causing a more severe acidemia because of the respiratory acidosis.

Etiology and Pathophysiology

A variety of stimuli can increase the ventilatory drive and cause a respiratory alkalosis (Table 52-14). Arterial hypoxemia or tissue hypoxia stimulates peripheral chemoreceptors to signal the central respiratory center in the medulla to increase ventilation. The resultant greater respiratory effort increases the oxygen content of the blood but depresses the PCO₂. The effect of hypoxemia on ventilation begins when the oxygen saturation decreases to approximately 90% (PO₂ = 60 mm Hg), and hyperventilation increases as hypoxemia worsens. Acute hypoxia is a more potent stimulus for hyperventilation than chronic hypoxia; thus, chronic hypoxia, as occurs in cyanotic heart disease, causes a much less severe respiratory alkalosis than an equivalent degree of acute hypoxia. There are many causes of hypoxemia or tissue hypoxia, including primary lung disease, severe anemia, and carbon monoxide poisoning.

The lungs contain chemoreceptors and mechanoreceptors that respond to irritants and stretching and send signals to the respiratory center to increase ventilation. Aspiration or pneumonia may stimulate the chemoreceptors, whereas pulmonary edema may stimulate the mechanoreceptors. Most of the diseases that activate these receptors may also cause hypoxemia and can, therefore, potentially lead to hyperventilation via 2 mechanisms. Patients with primary lung disease may initially have a respiratory alkalosis, but worsening of the disease, combined with respiratory muscle fatigue, often causes respiratory failure and the development of a respiratory acidosis.

Hyperventilation in the absence of lung disease occurs with direct stimulation of the central respiratory center. This occurs with CNS diseases, such as meningitis, hemorrhage, and trauma. Central hyperventilation due to lesions, such as infarcts or tumors near the central respiratory center in the midbrain, increases the rate and depth of the respiratory effort. This respiratory pattern portends a poor prognosis because these midbrain lesions are frequently fatal. Systemic processes may cause centrally mediated hyperventilation. Although the exact mechanisms are not clear, liver disease causes a respiratory alkalosis that is usually proportional to the degree of liver failure. Pregnancy causes a chronic respiratory alkalosis, probably mediated by progesterone acting on the respiratory centers. Salicylates, although often causing a concurrent metabolic acidosis, directly stimulate the respiratory center to produce a respiratory alkalosis. The respiratory alkalosis during sepsis is probably due to cytokine release.

Hyperventilation may be secondary to an underlying disease that causes pain, stress, or anxiety. In psychogenic hyperventilation, there is no disease process accounting for the hyperventilation. This disorder may occur in a child who has had an emotionally stressful experience. Alternatively, it may be part of a panic disorder, especially if there are repeated episodes of hyperventilation. In such a patient, the symptoms of acute alkalemia increase anxiety, potentially perpetuating the hyperventilation.

A respiratory alkalosis is quite common in children receiving mechanical ventilation because the respiratory center is not controlling ventilation. In addition, these children may have a decreased metabolic rate and hence less carbon dioxide production because of sedation and paralytic medications. Normally, decreased carbon dioxide production and the resultant hypocapnia decrease ventilation, but this physiologic response cannot occur in a child who cannot reduce the ventilatory effort.

Clinical Manifestations

The disease process that is causing the respiratory alkalosis is usually more concerning than the clinical manifestations of the respiratory alkalosis. Chronic respiratory alkalosis is usually asymptomatic because metabolic compensation decreases the magnitude of the alkalemia.

Acute respiratory alkalosis may cause chest tightness, palpitations, lightheadedness, circumoral numbness, and paresthesias of the extremities. Less common manifestations include tetany, seizures, muscle cramps, and syncope. The lightheadedness and syncope are probably due to the reduction in cerebral blood flow that is caused by hypocapnia. The reduction in cerebral blood flow is the rationale for using hyperventilation to treat children with increased intracranial pressure. The paresthesias, tetany, and seizures may be partially related to the reduction in ionized calcium that occurs because alkalemia causes more calcium to bind to albumin. A respiratory alkalosis also causes a mild reduction in the serum potassium level. Patients with psychogenic hyperventilation tend to be most symptomatic as a result of the respiratory alkalosis, and these symptoms, along with a sensation of breathlessness, exacerbate the hyperventilation.

Diagnosis

In many patients, hyperventilation producing a respiratory alkalosis is not clinically detectable, even with careful observation of the patient’s respiratory effort. Metabolic compensation for a respiratory alkalosis causes a low serum bicarbonate concentration. When hyperventilation is not appreciated and only serum electrolytes are evaluated, there is often a presumptive diagnosis of a metabolic acidosis. If a respiratory alkalosis is suspected, only a blood gas determination can make the diagnosis.

Hyperventilation does not always indicate a primary respiratory disorder. In some patients, the hyperventilation is appropriate respiratory compensation for a metabolic acidosis. With a primary metabolic acidosis, acidemia is present and the serum bicarbonate level is usually quite low if there is clinically detectable hyperventilation. In contrast, the serum bicarbonate level never goes below 17 mEq/L as part of the metabolic compensation for acute respiratory alkalosis, and simple acute respiratory alkalosis causes alkalemia.

The etiology of a respiratory alkalosis is often apparent from the physical examination or history, and it may consist of lung disease, neurologic disease, or cyanotic heart disease. Hypoxemia is a common cause of hyperventilation, and it is important to diagnose because it suggests a significant underlying disease that requires expeditious treatment. Hypoxemia may be detected on physical examination (cyanosis) or by pulse oximetry. However, normal pulse oximetry values do not completely eliminate hypoxemia as the etiology of the hyperventilation. There are 2 reasons why pulse oximetry is not adequate for eliminating hypoxemia as a cause of a respiratory alkalosis. First, pulse oximetry is not very sensitive at detecting a mildly low pO₂. Second, the hyperventilation during a respiratory alkalosis causes the pO₂ to increase, possibly to a level that is not identified as abnormal by pulse oximetry. Only an arterial blood gas measurement can completely eliminate hypoxia as an explanation for a respiratory alkalosis. Along with hypoxemia, it is important to consider processes that cause tissue hypoxia without necessarily causing hypoxemia. Examples are carbon monoxide poisoning, severe anemia, and congestive heart failure.

Lung disease without hypoxemia may cause hyperventilation. Although lung disease is often apparent by history or physical examination, a chest radiograph may detect more subtle disease. The patient with a pulmonary embolism may have benign chest radiograph findings, normal pO₂, and isolated respiratory alkalosis, although hypoxia may eventually occur. Diagnosis of a pulmonary embolism requires a high index of suspicion and should be considered in children without another explanation for respiratory alkalosis, especially if risk factors are present, such as prolonged bed rest and a hypercoagulable state (e.g., nephrotic syndrome or lupus anticoagulant).

Treatment

There is seldom a need for specific treatment of respiratory alkalosis. Rather, treatment focuses on the underlying disease. Mechanical ventilator settings are adjusted to correct iatrogenic respiratory alkalosis, unless the hyperventilation has a therapeutic purpose (e.g., treatment of increased intracranial pressure).

For the patient with hyperventilation secondary to anxiety, efforts should be undertaken to reassure the child, usually enlisting the parents. Along with reassurance, patients with psychogenic hyperventilation may benefit from benzodiazepines. During an acute episode of psychogenic hyperventilation, rebreathing into a paper bag increases the patient’s PCO₂. Using a paper bag, instead of a plastic bag, allows adequate oxygenation but permits the carbon dioxide concentration in the bag to increase. The resultant increase in the patient’s PCO₂ decreases the symptoms of the respiratory alkalosis that tend to perpetuate the hyperventilation. Rebreathing should be performed only once other causes of hyperventilation have been eliminated; pulse oximetry during the rebreathing is prudent.