Body Water Distribution in Children

Body water content changes significantly with age in children.^1,2 Total body water (TBW) is high in the fetus and preterm infant. During early fetal life, TBW represents 90% of total body weight, with 65% being in the extracellular fluid (ECF) compartment. By term, ECF and intracellular fluid (ICF) volume has fallen to 45% and 30% of TBW, respectively (Figure 113-1). The preterm infant has a relative expansion of both TBW and ECF volume expansion, and a diuresis in the first few days of postnatal life is a common finding. Fractional excretion of sodium is inversely correlated with age in the preterm, who is susceptible to both sodium loss and sodium and volume overload.³ In addition, glomerular filtration rate is lower than in the term infant, and the large surface area–to–body weight ratio leads to considerable evaporative losses.^4–7 Further discussion of fluid and electrolyte physiology in the preterm infant is beyond the scope of this chapter.

Figure 113-1 Changes in total body water (TBW) with age. Intracellular water is represented by the difference between the two diagonal lines.

(Adapted from Kooh SW Metcoff J. Physiologic considerations in fluid and electrolyte therapy with particular reference to diarrheal dehydration in children. J Pediatr 1963;62:107–31, with permission.)

Significant changes occur in TBW over the first year of life, from 75% of body weight at birth to 65% at 6 months and 60% at 1 year (Table 113-1). Some of this is accounted for by an increase in body fat. By puberty, TBW is approximately 60% of body weight in males, with a slightly lower percentage in females. Extracellular fluid volume decreases over the first year of life to 30% of TBW and decreases with age thereafter, reaching adult values early in childhood. The relatively high ECF volume in infancy is largely due to the larger interstitial lymph space. In contrast, the ICF volume remains relatively constant during childhood.

Fluid Homeostasis in Children

To achieve normal fluid homeostasis, fluid intake must balance losses. The latter consist of urine output plus insensible losses (evaporative from the skin surface and respiratory tract), with the addition of fluid loss in the stool, which in the absence of diarrhea should be minimal. Insensible losses are mainly in the form of electrolyte-free water (EFW) from the respiratory tract (15 mL/100 kcal/d). This loss is eliminated during positive-pressure ventilation. Sweat contains mainly water with a small amount of sodium, except in situations where sweat glands contain excessive amounts of sodium, such as in patients with cystic fibrosis. Evaporative losses also increase with elevations in body temperature; during thermal stress, water losses may increase to as much as 25 mL/100 kcal/d (Table 113-2).

Obligate water excretion in the urine is dependent upon solute load and the ability to concentrate and dilute urine. The average osmolar excretion in newborn infants receiving infant formula is 16 to 20 mOsm/kg/d.² Infants are somewhat disadvantaged compared to the older child and adult in that they cannot maximally dilute (infant 200 mOsm/L versus adult 80 mOsm/L) and concentrate urine (infant 800 mOsm/L versus adult 1200 mOsm/L). In addition, the infant’s high metabolic rate and the solute load from enteral feeding formula means they require more water excretion per unit solute amount. High solute load and limited urine concentrating ability makes them prone to significant ECF contraction (dehydration) when there are excessive amounts of water loss. Typically this occurs in gastroenteritis, where reduced oral intake is combined with excessive water and electrolyte loss in the stool.

Urine is the major source of electrolyte loss in the body except when there are fluid losses from the gastrointestinal tract. The commonly used values for sodium (Na) and potassium (K) requirements in parenteral fluids in children are 2 to 3 mmol/kg/d and 1 to 2 mmol/kg/d. This assumes that these are the amounts of cations needed for normal homeostasis. However, in critically ill children, urinary Na and K concentration may be much higher.

In the normal healthy individual, water intake is regulated by thirst stimulated via osmoreceptors in the hypothalamus. Infants and small children are unable to regulate their intake because they do not have access to water for the same reasons that apply in older children or adults in coma or with reduced levels of consciousness. When oral intake is replaced by parenteral fluids in children, the amount of fluid (i.e., water) given depends on body weight and energy expenditure. In 1957, Holliday⁸ published a formula that linked body weight to energy expenditure (Table 113-3). An allowance of 100 mL/100 kcal/d was made for insensible water loss, with 66.7 mL/100 kcal/d to replace urine output. Factoring in water of oxidation of 16.7 mL/100 kcal/d leaves a total of 100 mL/100 kcal/d for replacement of normal losses. The estimates for Na (3 mmol/100 kcal/d) and K (2 mmol/100 kcal/day) in maintenance fluids were calculated from the sodium and potassium concentration of cow’s milk and breast milk.

This paper by Holliday became the standard reference text for parenteral fluid administration in pediatrics. Although convenient and simple to use, the assumptions made about daily requirements for sodium, potassium, and EFW mandate the use of hypotonic intravenous (IV) solutions, which has been almost universal practice in pediatric medicine for almost 50 years (Table 113-4). However, nonphysiologic stimuli for antidiuretic hormone (ADH) secretion, which inhibits excretion of EFW (e.g., pain, anxiety, narcotics, positive-pressure ventilation), are common in critically ill patients. It is therefore not surprising that mild degrees of hyponatremia are a common finding in pediatric patients receiving parenteral fluid therapy. In a study by Gerigk⁹ of 103 children admitted to the hospital with acute medical illnesses, the median plasma Na value was 136 mmol/L, with plasma ADH levels that were higher than would be expected for that degree of hyponatremia. In 31 control patients (elective surgical admissions), median serum Na levels were 139 mmol/L, with lower ADH levels. We have made a similar observation in patients with hospital-acquired hyponatremia, who received twice as much EFW compared with a control group.¹⁰ The nonphysiologic secretion of ADH has been reported in association with many acute medical illnesses including meningitis, bronchiolitis, encephalitis, traumatic brain injury, and gastroenteritis.^11–22 An increasing number of publications are now recommending the use of isotonic or near-isotonic fluids for standard maintenance in pediatrics to avoid administration of EFW, which is potentially hazardous in situations where ADH secretion is not inhibited.^10,23–26 Hypotonic fluids should be reserved for patients with a demonstrated need for EFW (serum Na⁺ > 145 mmol/L).

Perioperative Fluid Management

Standard practice in perioperative fluid management has been to replace intravascular volume loss with blood or colloid solutions and to use electrolyte solutions to provide for ongoing fluid requirements, replacement of losses from exposed serosal surfaces in open body cavities in thoracic and abdominal surgery, and losses from third-space fluid sequestration (Table 113-5). Extra fluid is also frequently administered to treat hypotension due to the vasodilating effects of anesthetic agents. The preferred electrolyte solution used by most anesthesiologists for intraoperative fluid administration is now Ringer’s lactate or isotonic saline because of concerns about the development of postoperative fluid retention and hyponatremia associated with elevated ADH levels.^27–29 The potential for this is increased when hypotonic dextrose/saline solutions are used.^28,30–32 This inability to excrete a sodium-free water load is amply illustrated in scoliosis surgery, where patients seem to be particularly at risk for the development of hyponatremia postoperatively.³³ Two nonrandomized studies have shown that the degree of hyponatremia is less when isotonic or near-isotonic solutions are used.^34,35 In a nonrandomized trial, Burrows³⁵ compared Ringer’s lactate with 0.2% sodium chloride (NaCl) infusion in a group of children following scoliosis surgery. He found that the postoperative plasma Na level fell in both groups, but that the reduction was marked in those patients receiving the hypotonic fluid. Although at first glance, the explanation for this is EFW retention due to nonphysiologic stimulation of ADH secretion, it does not explain the reduction in plasma Na seen with Ringer’s lactate.

Further insights to explain this observation come from the study by Steele,³⁶ where plasma and urine Na were measured in adult patients undergoing elective surgery, all of whom received Ringer’s lactate as their perioperative fluid. They found that the urine Na concentration was consistently above 150 mmol/L and as high as 350 mmol/L in some instances. This was associated with a significant positive water balance and a fall in the plasma Na, a process they termed postoperative desalination. In a similar study of children undergoing elective surgery, all of whom received Ringer’s lactate, we found similar levels of urinary Na loss (unpublished observations). We think that this desalination process is consistent with the kidney’s attempts to deal with a volume overload situation after the vasodilating effects of anesthetic agents are no longer present, but ADH is still being actively secreted. In this situation, it would be unwise to prescribe hypotonic fluids in the postoperative period and impose an extra burden of more EFW to be excreted by the kidney.

Further evidence has now emerged that supports the use of isotonic rather than hypotonic fluid in the perioperative period. A prospective observational study in patients admitted to the intensive care unit (ICU) postoperatively documented an increased risk of development of hyponatremia associated with use of hypotonic saline; water retention and increased sodium excretion are to blame.³⁷ Two recent prospective randomized trials have compared the use of isotonic with hypotonic saline. Both have shown that the incidence of hyponatremia was significantly reduced with isotonic saline, and in neither study did the patients develop hypernatemia.^38,39

Disorders of Sodium Homeostasis

Sodium is the principal cation of the ECF compartment. Movement of Na into the ICF compartment is reversed by activation of the Na⁺/K⁺-ATPase pump. Sodium is absorbed in the proximal tubule under the influence of aldosterone. The serum Na reflects the osmolality and the ECF water volume, which is tightly regulated by ADH secretion.

Acute hyponatremia, defined as a fall in plasma Na to less than 130 mmol/L within 48 hrs, leads to rapid movement of water from the ECF to the ICF compartment and can cause cerebral edema, with catastrophic outcomes reported in children.^30,40,41 Clinical findings are those of raised intracranial pressure (nausea, vomiting, headache), frequently undiagnosed until the onset of seizures. This is usually followed by apnea, indicating that brainstem coning has occurred. Symptomatic hyponatremia rarely occurs below a serum Na level of 125 mmol/L, but when it does, it constitutes a medical emergency. The primary objective is to raise serum Na to above this level to prevent brainstem herniation. This can be most effectively achieved with the use of hypertonic saline.⁴² Once this threshold has been reached, the serum Na can be allowed to correct by fluid restriction with or without the use of furosemide. IV mannitol has also been used successfully in the emergency treatment of acute symptomatic hyponatremia.⁴³

Chronic hyponatremia is a common finding in patients with heart failure and renal failure and is associated with increased TBW and salt retention. It is not associated with cerebral edema, but correction of chronic hyponatremia with isotonic or hypertonic saline has been associated with central pontine demyelination.^44–46

Hypernatremia

Hypernatremia is defined as a serum Na greater than 145 mmol/L and is caused by either water deficit or salt gain (Table 113-7). The former is seen in infants with severe gastroenteritis with a loss of water in excess of sodium, sometimes compounded by increased solute intake from incorrect mixing of infant formula. The absence of ADH secretion causing diabetes insipidus is seen in patients with pituitary tumors, traumatic brain injury, and central nervous system (CNS) infections.^47–50 Water loss in critically ill children may also be associated with the use of loop diuretics or mannitol. Hypernatremia secondary to salt gain is seen with the excessive use of isotonic or hypertonic saline solutions or with the administration of IV bicarbonate.

TABLE 113-7 Causes of Hypernatremia

A rise in serum Na is associated with movement of water from the ICF to the ECF compartment and development of a hyperosmolar state. Brain cells adapt with an increase in electrolytes and “ideogenic” osmoles (inositol, taurine), which tends to mitigate the fluid shift with partial restoration of intracellular osmolality and brain cell volume.^51–53 Levels of Na over 155 mmol/L are frequently associated with abnormal CNS findings, and there is an increased risk of subdural hemorrhage and infarction in infants with hypernatremic dehydration and serum Na levels higher than 160 mmol/L.^54–57 There is also the added danger of development of brain edema during the attempt to correct these hyperosmolar states rapidly, using solutions that are hypo-osmolar compared to the ICF compartment.^58–63 Published recommendations suggest that the rate of correction of serum Na should be less than 0.5 mmol/L/h using the following formula for correction, which estimates the effect of 1 L of any infusate on serum Na:

In severe hypernatremia (serum Na > 170 mmol/L), it is recommended that the maximum Na concentration not be corrected to below 150 mmol/L in the first 48 to 72 hrs.⁶⁴

The epidemiology of hypernatremia in children has changed recently from gastroenteritis with dehydration as the principal cause to one of a hospital-acquired problem in association with either excess salt administration or a free-water deficit. In a study by Moritz of children with a serum Na above 150 mmol/L, the problem was hospital acquired in 60%, and the mortality was 11%.⁶⁵ In a similar series of adult patients, the ICU mortality rate for patients with plasma Na levels above 150 mmol/L was 30%.⁶⁶

Management of Acute Water and Sodium Deficits in Children

Two major problems of acute water and electrolyte deficits are worthy of specific mention because of the potential for serious adverse outcomes associated with both.