Definition and Classification

A metabolic acidosis is a process that, if unopposed, would cause acidemia (a high hydrogen ion concentration, or low pH, of the blood) by reducing the extracellular bicarbonate concentration. The extracellular bicarbonate concentration may be reduced by either addition of acid and consequent consumption of bicarbonate, or by primary loss of bicarbonate.

An adult eating a normal diet generates 16,000 to 20,000 mmol of acid a day.³ Almost all of that acid is in the form of carbonic acid, resulting from CO₂ and water generation in the metabolism of carbohydrates and fats. Individuals with normal ventilatory capacity eliminate this prodigious acid load through the lungs, thus the term volatile acid. The remainder of the daily acid load, about 1 mmol/kg body weight per day, derives from metabolism of phosphate- and sulfate-rich protein (yielding phosphoric and sulfuric acid). These nonvolatile or fixed acids are buffered, primarily by extracellular bicarbonate under normal circumstances. The kidneys are responsible for regenerating the consumed bicarbonate by secreting hydrogen ions (protons) in the distal nephron. These secreted protons must be buffered in the tubule lumen in order to allow elimination of the daily fixed acid load within the physiologic constraint of the minimum urinary pH. The urinary buffers are composed of the filtered sodium salts of the phosphoric acid and ammonia, which is synthesized in the proximal tubule and acidified in the collecting duct to form ammonium (NH₄⁺). Under conditions of acid loading, the normal kidney reabsorbs all the filtered bicarbonate in the proximal tubule. Urinary net acid excretion therefore comprises phosphoric acid (so-called titratable acidity, because it is quantified by titrating the urine with alkali to pH 7.40) and ammonium, less any excreted bicarbonate.⁴

Many factors modify the kidney’s capacity to regulate acid-base balance. For example, renal ammoniagenesis is stimulated by acidemia, and inhibited by alkalemia, and thus participates in a homeostatic feedback loop.¹ Hyperkalemia inhibits and hypokalemia stimulates renal ammoniagenesis. Hypokalemia further stimulates acid secretion by activating the Na⁺-H⁺ exchanger in the proximal tubule and the H⁺/K⁺-ATPase in the collecting duct. Finally, aldosterone stimulates both proton and K⁺ secretion in the collecting duct. For these reasons, hypokalemia tends to perpetuate a metabolic alkalosis, and hyperkalemia a metabolic acidosis.¹

Metabolic acidosis can be caused by excessive production of fixed acid, decreased renal secretion of fixed acid, or loss of bicarbonate, either through the kidney or through the intestine.⁴ The net effect of any of these processes is a reduction in the blood bicarbonate concentration. The plasma anion gap helps to distinguish among the various causes of metabolic acidosis. Of course, because of charge neutrality, the sum of the concentration of all cations in the plasma is equal to the sum of all the anions. By convention, however, the anion gap is defined as the difference between the plasma sodium concentration and the sum of the bicarbonate and chloride concentrations. It represents the concentration of anions that are normally unmeasured by a basic metabolic chemistry panel.⁵ The anion gap normally is about 8 mmol/L, but it varies widely according to the methods employed by the clinical chemistry laboratory.⁶ The anion gap is composed mainly of albumin, along with phosphates, sulfates, and organic anions.

There are two important pitfalls in the interpretation of the anion gap. First, because the anion gap is proportional to the plasma albumin concentration, hypoalbuminemia (common in critically ill patients) will lower the “baseline” anion gap (by approximately 2.5 mmol/L for each g/dL decline in the albumin concentration).⁷ Thus, profound hypoalbuminemia may falsely lower the anion gap, and thus mask a high anion gap acidosis. Second, alkalemia increases the anion gap by causing lactate generation and by titrating plasma buffers, most notably albumin.⁸ (Thus, in respiratory alkalosis, the bicarbonate concentration will be low in compensation, and the anion gap may be elevated, giving a false impression of a high anion gap metabolic acidosis by inspection of the electrolytes alone.)

If bicarbonate is lost (e.g., through diarrhea), or hydrochloric acid is gained (e.g., renal tubular acidosis or administration of unbuffered amino acid solutions⁹), the bicarbonate concentration falls with a commensurate increase in the plasma chloride concentration; thus the anion gap is unchanged. If, on the other hand, bicarbonate is lost in buffering an organic acid such as lactic acid or a ketoacid, the decrement in the bicarbonate concentration is more or less matched by an increase in the anion gap. These processes are illustrated in Figure 57.1.

Box 57.1 lists the causes of hyperchloremic metabolic acidosis. Two diagnoses are of particular interest in the critical care arena. First is the posthypocapnic metabolic acidosis, in which bicarbonate falls in compensation for a chronic respiratory alkalosis. When “normal” ventilation is restored, the pH falls until bicarbonate can be retained, giving the appearance of a hyperchloremic metabolic acidosis. This emphasizes the importance of observation over time in the analysis of acid-base status. The second entity of interest is a so-called dilutional hyperchloremic acidosis. This is seen in patients who are rapidly resuscitated with large volumes of isotonic saline solution. The acidosis traditionally has been attributed to dilution of blood bicarbonate. Analysis based on physical-chemistry principles may better explain the phenomenon (see later).¹⁰

The differential diagnosis of high anion gap metabolic acidosis is limited (Box 57.2). The most common cause in critically ill patients is a lactic acidosis. The causes of lactic acidosis are numerous. As shown in Box 57.3, they are divided into type A (imbalance between tissue oxygen demand and supply) and type B (impaired oxygen utilization).⁵ Diabetic ketoacidosis (DKA) (Chapter 58) and intoxications (Chapter 68) are discussed elsewhere. Two causes of high anion gap acidosis recently added to the differential diagnosis, and of particular relevance to intensivists, are pyroglutamic acidosis and intoxication with propylene glycol.

Pyroglutamic acid is a metabolic intermediate in the γ-glutamyl cycle, one product of which is glutathione. Pyroglutamic acidosis may be congenital (caused by one of several enzyme deficiencies) or acquired.⁸ The acquired syndrome may be caused by acetaminophen (which depletes glutathione, leading to uninhibited pyroglutamic acid synthesis), β-lactam antibiotics, or glycine deficiency. The acidosis may be profound and the anion gap greater than 30 mmol/L.¹¹ Definitive diagnosis is made by urinary screen for organic acids. In practice, however, circumstantial evidence suggests the acquired syndrome and the diagnosis is supported by a favorable response to appropriate intervention.

Propylene glycol is a solvent for medications, many of which are commonly infused intravenously in critically ill patients, such as lorazepam, nitroglycerin, etomidate, and phenytoin. Propylene glycol is metabolized by alcohol dehydrogenase to lactic acid. High anion gap acidosis has been associated with high- and even low-dose infusions, particularly of lorazepam.7,12 Thus, development of a high anion gap acidosis in a critically ill patient should prompt a search for a source of propylene glycol, because withdrawal of the agent will promptly alleviate the acidosis.

Consequences of Acidemia

It has been generally accepted that severe acidemia (pH < 7.20) is associated with a variety of deleterious effects. Of particular concern are the cardiovascular effects, including pressure-resistant arterial vasodilation, venoconstriction, diminished myocardial contractility, and impaired hepatic and renal perfusion.¹³ (Some controversy exists as to which of these effects are directly caused by acidemia.⁴) A predisposition to malignant arrhythmias has been reported in vitro and in animal models. Finally, numerous metabolic derangements have been attributed to the effect of acidemia on key enzymes in metabolic pathways, resulting in sympathetic hyperactivity with diminished catecholamine responsiveness; insulin resistance and suppressed glycolysis; and reduced hepatic lactic acid uptake and metabolism.¹⁴

Diagnosis of Acid-Base Disorders

Acid-base disorders are revealed most commonly through the basic metabolic chemistry panel, when the plasma bicarbonate concentration is noted to be outside the normal range. If the bicarbonate is low, and if the anion gap is clearly elevated on that sample, a diagnosis of high anion gap metabolic acidosis can be made with some confidence, keeping in mind the pitfalls in the interpretation of the anion gap mentioned earlier.⁷

If the bicarbonate is low and the anion gap normal, two possibilities exist: either a hyperchloremic metabolic acidosis or a respiratory alkalosis with metabolic compensation. These two entities can be distinguished by examination of the blood pH and blood gases, a low pH being diagnostic of the former.

If the bicarbonate concentration is high, again there are two alternative diagnoses, requiring blood pH measurement for their differentiation: either a metabolic alkalosis or metabolic compensation for a respiratory acidosis.

Once the primary disturbance has been identified, the astute clinician, recognizing the possibility of a mixed disturbance, is obligated to ask, “Is that all there is?” This question can be answered only by an understanding of the rules of normal compensation for simple acid-base disorders (see Table 57.1).² Knowing at least the expected direction of compensation will allow the clinician to diagnose the most obvious mixed disturbances. For example, if the pH is low, the bicarbonate is low, and the PCO₂ is above 40 mm Hg, there is clearly a mixed metabolic and respiratory acidosis. Similarly, if the pH is high, the bicarbonate is high, and the PCO₂ is below 40 mm Hg, the diagnosis is a mixed respiratory and metabolic alkalosis. More subtle mixed disorders can be diagnosed only by understanding not only the expected direction, but the expected magnitude of compensation. This will allow one to conclude, for example, whether the hyperventilation in a patient with metabolic acidosis is appropriate (expected compensation), inadequate (a separate respiratory acidosis), or excessive (a separate respiratory alkalosis).

The preceding method permits the diagnosis of simple and dual acid-base disorders. Triple acid-base disorders can be diagnosed only by comparing the change in the anion gap with the change in the plasma bicarbonate concentration. Most simply conceived, the fall in the bicarbonate should equal the rise in the anion gap (see Fig. 57.1). If the rise in the anion gap exceeds the fall in the bicarbonate, a metabolic alkalosis is said to be present in addition to the high anion gap acidosis. Conversely, if the fall in the bicarbonate exceeds the rise in the anion gap, mixed hyperchloremic and high anion gap acidoses are said to coexist. Although this analysis is useful in the case of large discrepancies, in more subtle cases it is confounded by theoretical and practical considerations.^5,15

The classical approach to acid-base disorders described earlier has been challenged recently by proponents of a physical-chemistry approach described originally by Stewart.¹⁶ According to this method, the pH of the blood depends on the ionization of water by the difference in the concentration of so-called strong ions (the strong ion difference, or SID). The SID offers a quantitative approach to measuring the degree of acidosis in hyperchloremic metabolic acidosis. Although there is evidence that this approach may offer prognostic capabilities in patients with severe sepsis and septic shock with hyperchloremic metabolic acidosis, the complexity of the equations for calculating the SID may make this method cumbersome in clinical settings.¹⁷ The main utility of this construct in the critical care setting seems to be its explanation of a hyperchloremic metabolic acidosis in patients who receive large volumes of isotonic saline.

Treatment of Metabolic Acidosis

Treatment of metabolic acidosis is aimed at reversing the adverse consequences of acidemia. Treatment of hyperchloremic metabolic acidosis is straightforward. In cases of acute metabolic acidosis, treatment depends on successful therapy of the underlying cause (e.g., diarrhea) and correction of the bicarbonate deficit, usually in the form of sodium bicarbonate. One can estimate the bicarbonate deficit as follows:

The difficulty in accurately estimating this value arises from two factors: First, the apparent volume of distribution of bicarbonate varies more than twofold from 50% of body weight to 100% of body weight and is inversely proportional to the initial bicarbonate concentration.¹⁸ Second, there are often many simultaneous processes in a critically ill patient that tend to ameliorate or exacerbate the metabolic acidosis, such as vomiting, shock, and liver failure. In order to avoid overshoot alkalemia, it is prudent to estimate the volume of distribution to be 50% of the body weight¹⁴ and to target an increase in the bicarbonate concentration of no more than 8 mmol/L over 12 to 24 hours.

Sodium bicarbonate generally is considered to be the alkalinizing agent of choice for severe acidemia. Alternative alkalinizing agents such as citrate, acetate, and lactate, which under normal circumstances are oxidized in the liver to bicarbonate, should not be used to treat acidemia in patients with suspected or confirmed hepatic impairment or circulatory compromise. The sodium bicarbonate should be administered as a continuous infusion, the concentration of which should be guided by the patient’s serum sodium concentration. Bolus injection of undiluted ampules of sodium bicarbonate (1000 mmol/L) should be used with great restraint and only in patients with the most severe acidemia because of the risk of hyperosmolality. Large volumes of any bicarbonate solution can lead to volume overload, a reduction in the ionized calcium concentration (see “Hypocalcemia”), and increased generation of CO₂. This last effect will tend to cause a respiratory acidosis in patients with ventilatory insufficiency. Plasma electrolytes and blood gases must be monitored frequently to guide adjustments in the composition of the solution and its rate of infusion.

Tris(hydroxymethyl)aminomethane, or THAM, is an amino alcohol that buffers without generating CO₂. It has the advantage, therefore, of avoiding a superimposed respiratory acidosis. It has been used successfully in animals and humans with various metabolic acidoses.13,19 It is eliminated by the kidney, and thus should be used with caution in the setting of renal insufficiency. Risks include hyperkalemia, hypoglycemia, and hepatic necrosis (in neonates).⁷

The treatment of choice for lactic acidosis is reversal of the underlying cause of the acidosis (see Box 57.3). Pending resolution of the underlying disorder, however, the intensivist is often confronted with an unstable patient who is profoundly acidemic. Treatment at this stage is controversial.^5,13 The debate has focused on the potentially deleterious effects of bicarbonate administration in lactic acidosis.²⁰ In addition to the effects mentioned earlier, bicarbonate in animal models of lactic acidosis has been associated with increased lactate generation, reduction in intracellular pH, increased venous PCO₂, and reduction in cardiac output. (This last effect correlates well with the reduction in ionized calcium concentration.¹³) Studies in humans likewise show no improvement in cardiac output, morbidity rate, or mortality rate with bicarbonate.¹³ Continuous venovenous hemodialysis (e.g., CVVHD) may be a promising tool for treating lactic acidosis, because it provides large amounts of bicarbonate without the risks of volume overload or hypocalcemia. There are several reported cases of successful treatment of metformin-associated lactic acidosis using continuous hemodialysis.^21,22 Because of its superior short-term clearance compared with continuous renal replacement therapy, however, conventional hemodialysis remains the preferred treatment for metformin intoxication.²³ Treatment of DKA and the acidoses associated with other intoxications are discussed in Chapters 58 and 68, respectively.

Definition and Classification

Metabolic alkalosis is a process leading to accumulation of extracellular bicarbonate that, if unopposed, will result in an increase in the plasma pH (alkalemia). It can be caused either by a gain of bicarbonate or a loss of fixed acid from the ECF. The causes of metabolic alkalosis have been described.²⁴ In its pure form, it is accompanied by hypoventilation (CO₂ retention).²

From a pathophysiologic perspective, metabolic alkalosis is divided into those factors that generate the alkalosis and those factors that maintain or perpetuate it.24,25 Metabolic alkalosis is generated by addition of bicarbonate to the blood. This can occur either by loss of acid from the body or by addition of exogenous alkali. Loss of acid may be from the stomach (e.g., vomiting or nasogastric suction) or kidney. Renal acid loss is enhanced by a high rate of sodium delivery to the distal nephron, high circulating mineralocorticoid levels, potassium depletion, and high rates of ammoniagenesis.

Because of the kidney’s prodigious ability to excrete bicarbonate, however, addition of bicarbonate to the blood is not sufficient to cause a sustained metabolic alkalosis. Some mechanism(s) to maintain the alkalosis must prevail. The most common mechanism contributing to the maintenance of metabolic alkalosis is volume depletion, either absolute or relative (e.g., congestive heart failure), which (1) reduces glomerular filtration, (2) enhances tubular bicarbonate reabsorption, and (3) causes secondary hyperaldosteronism, further enhancing urinary acidification. Another common perpetuating factor is potassium depletion, which stimulates proton secretion at several sites along the nephron.24–26

Patients with metabolic alkalosis and signs of volume expansion—especially hypertension—usually have excess mineralocorticoid as the explanation for the metabolic alkalosis. Aldosterone and glucocorticoids (other than dexamethasone) stimulate renal loss of acid and potassium, and thereby generate and maintain the alkalosis.

Most cases of clinically significant metabolic alkalosis are maintained by loss of chloride or potassium. Although total body sodium (and hence, volume) derangements are not directly responsible for the generation and maintenance of the metabolic alkalosis, potassium and chloride depletion are commonly seen in settings of volume depletion or excess. Therefore, from a clinical standpoint, it is useful to approach the patient with metabolic alkalosis centering on the history and physical examination, with special attention to the ECF volume status, followed by sequential analysis of blood chemistries.²⁶ Causes of metabolic alkalosis are shown in Box 57.4. One entity unique to critically ill patients is posthypercapnic metabolic alkalosis. This syndrome is caused by abrupt treatment (usually with tracheal intubation and mechanical ventilation) of a chronic respiratory acidosis. The renal bicarbonate retention that compensated for the chronic respiratory acidosis persists (because of volume depletion) after restoration of a normal PCO₂, resulting in the high pH and high plasma bicarbonate characteristic of metabolic alkalosis. The key to the diagnosis is the history and sequential analysis of blood chemistries.²⁴

Clinical Consequences

Alkalemia in critically ill patients is associated with increased mortality rate.²⁷ Patients with combined metabolic and respiratory alkalosis have a higher mortality rate than those with respiratory alkalosis alone, and mortality rate in alkalemia is roughly proportional to the pH.²⁷ Although no causal relationship between alkalemia and mortality rate has been established, the pathophysiology of alkalemia is far from benign.²⁵

First, metabolic alkalosis suppresses ventilation, causing CO₂ retention and relative hypoxemia.²⁸ Second, alkalemia acutely increases hemoglobin’s oxygen affinity (Bohr effect). Third, respiratory alkalosis causes vasoconstriction, particularly in the cerebral circulation.²⁵ All these processes tend to decrease tissue oxygen delivery.²⁹ (Note that chronic alkalemia inhibits 2,3-diphosphoglycerate synthesis, allowing normalization of the oxyhemoglobin desaturation curve, mitigating tissue hypoxia to some extent.) These alterations in tissue oxygen delivery could be responsible at least in part for some of the clinical manifestations of metabolic alkalosis.

Because alkalemia causes a decrease in ionized calcium concentration (see discussion under “Calcium Homeostasis”), many of the neuromuscular manifestations of metabolic alkalosis overlap with those of hypocalcemia, including paresthesias, tetany, and a predisposition to seizures.¹ The acutely diminished tissue oxygen delivery to the brain may contribute to initial confusion and obtundation seen with metabolic alkalosis.

Metabolic alkalosis often is accompanied by hypokalemia and hypomagnesemia. Thus, there is an association between alkalosis and arrythmias,²⁴ but an independent effect of the alkalosis on cardiac arrythmogenesis has not been established.

Increases in blood lactate concentration may occur in patients with metabolic alkalosis due to upregulation of phosphofructokinase and thus glycolysis, and because of tissue hypoxia (see earlier).⁸ With severe metabolic alkalosis (arterial pH above 7.55), the tissue hypoxia may be so marked that compensatory hypoventilation will be overridden by hypoxic drive, resulting in a normal to low arterial PCO₂ and elevated blood lactate levels (so-called “lactic alkalosis”).³⁰

Treatment

Treatment of metabolic alkalosis entails correcting the factor(s) responsible for its maintenance and, if possible, correcting the factor that generated the alkalosis. Once the underlying diagnosis is clear (see Box 57.4), therapy is usually straightforward. If the metabolic alkalosis is maintained by chloride depletion and ECF volume contraction, the intravascular volume should be restored to normal, usually with intravenous isotonic saline.^24,25 Potassium should be given, as KCl, to replace any deficits (see “Disorders of Potassium Homeostasis”), because potassium depletion perpetuates the metabolic alkalosis. If nasogastric suction cannot be stopped, acid loss can be reduced by the use of H₂ blockers and proton pump inhibitors.

Treating patients with metabolic alkalosis in the setting of volume overload and diminished effective circulating volume (e.g., congestive heart failure, hepatic cirrhosis) is more challenging, because saline infusion is contraindicated. Unless hyperkalemia is present, chloride should be replenished with KCl supplementation. In rare cases of concurent hyperkalemia, acetazolamide (a carbonic anydrase inhibitor) may be of benefit as it produces a bicarbonate diuresis. Acetazolamide should be avoided in patients with hypokalemia, because the alkaline diuresis will cause renal potassium wasting.²⁴ Another potential complication of acetazolamide administration, particularly in patients with impending ventilatory failure, is worsening of hypercapnia owing to inhibition of red blood cell carbonic anhydrase and impaired CO₂ transport.²⁵ Hydrochloric acid infusion, as a 0.1 to 0.25 N solution, has been used with success in patients with severe metabolic alkalosis (pH > 7.55 and systemic instability such as encephalopathy or cardiac arrhythmia²⁴) refractory to conventional measures.^25,31 Correction of the metabolic disturbances has been reported with infusion of 0.25 N HCl at 100 mL/hour over about 12 hours.³¹ Extreme care must be taken to ensure that the infusion catheter is properly positioned within the vena cava, because the solution is highly caustic. Plasma chemistries must be monitored frequently in order to avoid overcorrection. If renal function is severely impaired or medical therapy is not possible, hemodialysis against a low-bicarbonate bath may be used.²⁴

In states of primary mineralocorticoid excess, an aldosterone antagonist such as spironolactone should be used until the underlying abnormality can be corrected. Other potassium-sparing diuretics, such as amiloride and triamterene, are useful as well and are essential in managing the rare patient with Liddle syndrome.

Regulation of Internal Potassium Balance

Internal potassium balance serves to protect against changes in Ke; potassium tends to move out of cells during potassium depletion and into cells following potassium intake. This process tends to prevent drastic alterations of Ke : Ki.36,37 The factors that influence internal potassium balance include hormones, acid-base status, plasma tonicity, exercise, and cell integrity (Box 57.5).

The direction and magnitude of an acid-base-related change in P_K depend on the nature and the duration of the disturbance. The most consistent and pronounced relationship between changes in pH and P_K occurs in acute mineral (hyperchloremic) acidosis, where there is a strong inverse relationship between these two variables.38–40 Interestingly, hypokalemia is seen with prolonged mineral acidosis in patients with normal renal function and reflects increased renal potassium excretion.³⁹ Unlike mineral acidoses, however, even severe acute organic (high anion gap) acidoses are not usually associated with hyperkalemia.^41–44 Indeed, organic acidoses, such as lactic acidosis, actually tend to cause cellular potassium uptake.⁴¹ Nonetheless, factors coincident with the acidosis may alter P_K. For example, mesenteric ischemia may result in both lactic acidosis (from anaerobic metabolism) and hyperkalemia. Even the hyperkalemia so commonly seen in patients with DKA does not result from the acidemia; rather, it appears to be a consequence of the characteristic insulin deficiency and hyperglycemia (see discussion of hypertonicity in the next paragraph).⁴⁴ Respiratory disturbances typically alter P_K less than metabolic disturbances. Alkaloses, respiratory or metabolic, have less effect on P_K than their corresponding acidoses.³⁸ Bicarbonate administration, which was once thought to reduce the P_K by stimulating cellular potassium uptake,⁴⁵ is now known to have very little if any immediate effect on internal potassium balance,^46,47 except perhaps in patients with preexisting severe metabolic acidosis.⁴¹ It is clear, however, that longstanding alkalemia causes urinary potassium losses that may over time result in profound potassium depletion.⁴⁸

Hypertonicity, as seen with hypertonic fluid administration⁴⁹ or diabetic hyperglycemic states,⁵⁰ leads to hyperkalemia, probably as a result of potassium efflux from cells by way of solvent drag. Fatal hyperkalemia has been attributed to this phenomenon in diabetic patients with end-stage renal disease (ESRD).⁵¹

Exercise causes a transient shift of potassium out of cells. Clinically significant hyperkalemia may result from exercise52,53 (and clinically misleading local venous hyperkalemia results from fist clenching during phlebotomy⁵⁴).

Regulation of External Potassium Balance

In contrast to the prodigious capacity of the kidney to excrete potassium,⁵⁵ renal potassium conservation is imperfect and explains why significant potassium depletion and hypokalemia may result from dietary potassium deficiency alone.⁵⁶

Normally, 90% to 95% of dietary potassium is eliminated through the kidney, and only about 5% to 10% through the intestine. It is the kidney that is almost entirely responsible for matching potassium output to potassium intake in order to maintain total body potassium constant.⁵⁷ The majority of potassium excreted by the kidney derives from potassium secretion in the distal nephron (connecting tubule and collecting duct).⁵⁷ Virtually all regulation of potassium excretion takes place at this site in the nephron, under the influence of two principal factors: the rate of flow and sodium delivery through that part of the nephron, and the effect of aldosterone.⁵⁷ Potassium secretion is directly proportional to flow rate and sodium delivery through the distal nephron, explaining in part why diuretic use often is accompanied by hypokalemia.

Metabolic acidosis with acidemia results in inhibition of renal potassium secretion.⁴¹ In contrast, metabolic alkalosis and bicarbonate delivery to the distal nephron stimulate kaliuresis by increasing the electrochemical “driving force” for potassium secretion.⁵⁷ Other anions that are poorly reabsorbed in the distal nephron (e.g., synthetic penicillins) have a similar effect to stimulate potassium secretion.⁵⁸

It is well established that aldosterone participates in a homeostatic feedback loop with P_K such that increases in P_K stimulate adrenal aldosterone production, which in turn reduces P_K primarily by stimulating renal potassium excretion.⁵⁷ Hypokalemia is a prominent feature of primary aldosteronism (Conn syndrome) because the high circulating aldosterone levels are accompanied by volume expansion and thus a high rate of sodium delivery to the distal nephron. When circulating aldosterone levels are high due to volume depletion (secondary aldosteronism), the increase in distal potassium secretion is offset by a decrease in distal nephron flow, thus mitigating renal potassium loss. Indeed, it is only when patients with secondary hyperaldosteronism (e.g., in congestive heart failure or hepatic cirrhosis) are treated with diuretic drugs that distal nephron flow is increased and hypokalemia may ensue.

Magnesium deficiency is associated with renal potassium wasting and may result in severe potassium depletion.⁵⁹ Because magnesium, like calcium, acts to stabilize excitable membranes, the deleterious effects of hypokalemia on the myocardium are magnified by concurrent hypomagnesemia (see “Clinical Manifestations” later).⁶⁰

The effect of dexamethasone (a pure glucocorticoid) to enhance renal potassium excretion appears to result entirely from hemodynamic changes that cause an increase in glomerular filtration rate and distal flow rate. All other glucocorticoids tend to further stimulate potassium secretion in proportion to their mineralocorticoid activity.⁵⁷

Acute Hyperkalemia (Box 57.6)

Acute Hypokalemia

Hypokalemia that develops over hours is virtually always the result of redistribution of potassium from the extracellular to the intracellular space. The causes of acute hypokalemia are summarized in Box 57.7. Selected causes are discussed as follows.

Chronic Hyperkalemia

Chronic Hypokalemia

Chronic hypokalemia is virtually always the result of altered external balance: insufficient potassium intake, excessive potassium losses, or a combination of the two. Losses most often are either GI or renal.

Clinical Manifestations of Hyperkalemia

Clinical Manifestations of Hypokalemia

Although less immediately life threatening than hyperkalemia, hypokalemia has many detrimental effects in critically ill patients. Along with cardiac and neuromuscular manifestations are many more subtle effects.

Evaluation of Acute Hyperkalemia

When P_K rises abruptly or if P_K is high (greater than 6.5 mmol/L) on initial presentation of the patient, the first step is to obtain an ECG to look for electrophysiologic evidence of hyperkalemia. In the presence of such signs, treatment for hyperkalemia should begin urgently (see “Treatment of Potassium Imbalance”). At the same time, an unclotted blood sample should be obtained, using meticulous phlebotomy technique, for another set of electrolytes, glucose, blood urea nitrogen (BUN) and creatinine, and complete blood count (CBC). Urine should be tested for heme pigments to exclude acute rhabdomyolysis or hemolysis. The patient’s list of medications and diet should be reviewed promptly, looking for exogenous sources of potassium and drugs that may impair potassium tolerance (see Box 57.6).

Evaluation of Chronic Hyperkalemia

Figure 57.2 outlines an approach to the patient with hyperkalemia lasting for days. Failure to stimulate cortisol release with a cosyntropin stimulation test (see Chapter 59) supports a diagnosis of Addison’s disease. Absent that diagnosis, the patient is likely to have either selective aldosterone deficiency or tubular unresponsiveness to aldosterone. Assessment of the renin-angiotensin-aldosterone axis is most simply done by measuring plasma renin activity (PRA) and aldosterone levels in the basal and diuretic/posture-stimulated state.¹⁰⁸ Tubular unresponsiveness to aldosterone is assessed by measuring the potassium secretory effect of 9a-fludrocortisone (9a-F, Florinef). One index of the driving force for potassium secretion that may be useful in this regard is the transtubular potassium gradient (TTKG), calculated as follows:

where P_osm and U_osm are plasma and urine osmolalities, respectively.¹⁵⁰ A spot urine potassium : creatinine ratio may be used instead of the TTKG.

Evaluation of Acute Hypokalemia

Hypokalemia accompanied by serious cardiac or neuromuscular manifestations is an emergency. Likewise, urgent therapy is indicated for profound hypokalemia (P_K less than 2.0 mmol/L) even in the absence of clinical complications. In addition, moderate hypokalemia (P_K less than 3.0 mmol/L) in patients taking digitalis,¹⁴³ and perhaps with acute myocardial ischemia,¹⁴⁵ should be treated urgently because of the risk of ventricular arrhythmias. In all these situations, it is imperative that the blood specimen be obtained and handled properly, especially in patients with leukocytosis, because rapid administration of potassium to a patient with pseudohypokalemia may cause severe hyperkalemia.

The remainder of the evaluation of acute hypokalemia derives mainly from the patient’s history, with an emphasis on treatments causing cellular potassium uptake (e.g., insulin, β-adrenergic agonists) or a rapid increase in tissue anabolism, and a history of periodic paralysis. Patients who are hypokalemic upon presentation should be evaluated as though their hypokalemia were acute.

Evaluation of Chronic Hypokalemia

Once acute hypokalemia and transient potassium redistribution have been excluded, one should next determine whether the kidney is responding appropriately to the potassium deficit or whether it is contributing to the problem. This is best done by measuring the 24-hour urinary excretion of potassium during potassium repletion (Fig. 57.3). Potassium excretion less than 20 mmol per day suggests appropriate renal potassium conservation and points to extrarenal (lower GI or skin) potassium losses, recovery from diuretic-induced hypokalemia, or chronically potassium-deficient diet. Excretion of greater than 20 mmol per day is evidence of inadequate renal potassium conservation indicating a renal cause of the hypokalemia. Renal potassium losses associated with normal systemic blood pressure are most commonly seen with the use of thiazide or loop diuretics and are accompanied by a metabolic alkalosis. Other causes of hypokalemia with metabolic alkalosis in a normotensive patient include gastric fluid loss and Bartter and Gitelman syndromes. These are separable most often by history, but if not, the urinary chloride measurement will be helpful, being low with gastric fluid losses. Renal hypokalemia may accompany a renal tubular acidosis, in which case the plasma bicarbonate will be low.

Mineralocorticoid excess may be the cause if the renal potassium loss is associated with systemic hypertension, and the renin-aldosterone axis should be studied with basal and saline-suppressed blood hormone measurements. High PRA and aldosterone levels suggest renal artery stenosis, malignant hypertension, or rarely a renin-secreting tumor. Low PRA and high aldosterone levels indicate primary aldosteronism. When both PRA and aldosterone levels are low, one should suspect the syndrome of apparent mineralocorticoid excess, Cushing syndrome, or rarely Liddle syndrome. Note that Cushing syndrome due to ectopic adrenocorticotropic hormone (ACTH) secretion often is not accompanied by typical cushingoid features.¹⁵¹

Treatment of Acute Hyperkalemia

In considering when hyperkalemia constitutes an emergency, two points should be kept in mind. First, the electrophysiologic effects of hyperkalemia are directly proportional to both the absolute P_K and its rate of rise.¹³⁵ Second, although the ECG manifestations of hyperkalemia are generally progressive and proportional to the P_K, ventricular fibrillation may be the first ECG disturbance of hyperkalemia;¹³⁸ conversely, a normal ECG may be seen with extreme hyperkalemia.¹³⁷ Thus, it is apparent that neither the ECG nor the P_K alone is an adequate index of the urgency of hyperkalemia, and that the clinical context must be considered when assessing a hyperkalemic patient. Because most patients manifest hyperkalemic ECG changes at P_K greater than 6.7 mmol/L,¹³⁶ hyperkalemia should be treated emergently for (1) P_K greater than 6.5 mmol/L and (2) ECG manifestations of hyperkalemia regardless of the P_K.¹⁵²

Therapy of acute or severe hyperkalemia is directed at preventing or ameliorating its untoward electrophysiologic effects on the myocardium. The goals of therapy, in chronologic order, are as follows (Table 57.2):

Treatment of Chronic Hyperkalemia

As established previously, chronic hyperkalemia always implies deficient renal potassium excretion. It follows that the therapy of chronic hyperkalemia is primarily directed toward stimulating renal potassium excretion while limiting potassium intake. For all adults with chronic hyperkalemia, daily potassium intake should be restricted to 60 mmol. All drugs known to impair either internal or external potassium balance should be eliminated if possible. Finally, all patients with chronic hyperkalemia should be evaluated for occult urinary tract obstruction.

Further therapy of the persistently hyperkalemic patient should be guided by the diagnostic evaluation outlined in Figure 57.2. In cases of mineralocorticoid unresponsiveness or when mineralocorticoid treatment is complicated by fluid overload, a thiazide or loop diuretic can be added to the regimen. This will restore normal volume status and enhance renal tubular potassium secretion in many mineralocorticoid-resistant patients. It is crucial to avoid diuretic-induced volume depletion, however, because this will exacerbate the renal potassium secretory defect.

Patients who fail to respond to the previously mentioned measures with an increase in TTKG or urine potassium : creatinine ratio and a decrease in P_K may be given sodium bicarbonate, which will stimulate renal potassium secretion. This is especially appropriate for patients whose chronic hyperkalemia is accompanied by a renal tubular acidosis (type IV RTA). The usual dose is 1 to 2 mmol bicarbonate per kg body weight per day in three or four divided doses.¹⁵²

Treatment of Acute Hypokalemia

A low P_K almost always indicates a large total body potassium deficit. In fact, P_K decreases by approximately 0.3 mmol/L for each decrement of 100 mmol total body potassium.¹⁶⁶ But if potassium is replenished too quickly, the homeostatic mechanisms that defend P_K will be overwhelmed and P_K will rise abruptly. The rate of rise of P_K with potassium administration can be greatly altered by factors that affect internal potassium balance. For example, during treatment of DKA with insulin, cellular uptake of potassium may be massive, obligating enormous replacement doses of potassium. Conversely, insulin deficiency markedly impairs tolerance to a potassium load.⁶¹

See “Evaluation of Acute Hypokalemia” for definitions of urgent hypokalemia. Limited information exists on which to base a rational prescription of KCl in an emergency.^61,167,168

Based on the available literature, we can estimate that nondiabetic patients with normal renal function should respond well to a 1- to 2-hour infusion of KCl at 0.6 mmol/kg/hour given intravenously in saline. In patients with renal failure of any degree, the infusion rate should be halved (0.3 mmol/kg/hour). Patients with diabetes mellitus not being treated for DKA or hyperglycemia should receive no more than 0.2 mmol/kg/hour, or about 0.1 mmol/kg/hour in the setting of renal failure. For severe hypokalemia, the ECG should be monitored continuously and the infusion stopped immediately if signs of hyperkalemia develop. The maximum increase in P_K is seen at the end of the infusion, and about 50% of the increase is lost over the next 2 to 3 hours when a new steady state is achieved. Thus, P_K should be measured at the end of the infusion. If the patient is still dangerously hypokalemic at this point, additional potassium may be given. If at the end of the infusion P_K is in an acceptable range, the measurement should be repeated 2 to 3 hours later when disposal of potassium load is complete in order to determine the need for further treatment.¹⁵²

Hypokalemia in the setting of aggressive “refeeding,” and especially in the treatment of severe DKA, should be treated initially as described earlier. Frequent monitoring of P_K with rapid laboratory turnaround time is critical for proper management.

Hypokalemia that is not life threatening is best treated with oral potassium replacement. It is important to recognize that GI absorption of an oral dose of KCl elixir is essentially complete. Dangerous hyperkalemia can occur in entirely normal individuals following KCl ingestion.¹⁶⁹ The maximum increase in P_K is seen 1.5 to 2 hours after an oral potassium load. Thus, a sensible oral dose of KCl in moderate hypokalemia should probably not exceed the hourly intravenous doses proposed earlier. There is no reason to give a simultaneous oral and intravenous potassium dose; serious hyperkalemia may ensue.

Treatment of Chronic Hypokalemia

The treatment of chronic hypokalemia depends entirely on identifying and, if possible, remediating the cause (see Fig. 57.3). When the cause of the excessive potassium loss cannot be treated specifically, maintenance potassium supplementation is needed.

Nonosmotic Vasopressin Release

It is important to recognize that plasma osmolality is not the only determinant of AVP synthesis and release. Low arterial blood pressure and low effective arterial volume powerfully stimulate AVP release.¹⁷³ This baroreceptor-mediated AVP release is teleologic, because water retention is an important component in the defense against hypovolemia. So primal is this circulatory defense that the baroreceptor stimulation predominates over any osmolal effect on AVP release.¹⁷³ Thus, a volume-contracted or hypotensive individual will have high circulating AVP levels even if his plasma osmolality is low. In addition, circulating AVP levels rise with pain, stress, nausea, hypoxia, hypercapnia, and a variety of medications, most notably epinephrine and high doses of narcotic analgesics.¹⁷³

Epidemology and Clinical Manifestations

Hyponatremia (plasma sodium concentration < 135 mmol/L) is one of the most common electrolyte disorders, found in approximately 3% of hospitalized patients and as many as 30% of patients in ICUs.¹⁷⁰

The clinical manifestations of hyponatremia are largely attributed to intracellular volume expansion (cellular edema), which occurs only when hyponatremia is associated with hypotonicity. Intracellular volume expansion is of greatest consequence in the brain, where it is translated into increased intracranial pressure because of the rigid calvarium.¹⁷⁴

Pathophysiology

The pathophysiology of hypotonic hyponatremia has important implications for its management. Most cells—especially brain cells—have adaptive mechanisms for mitigating tonicity-related volume changes.¹⁷⁴ Cell volume peaks 1 to 2 hours after the onset of acute hypotonicity. Thereafter, solute and water are lost from cells, and cell volume returns toward normal. After several days of sustained hypotonicity, cell volume is restored nearly to normal.¹⁷⁴

The morbidity and mortality risks associated with hypotonic hyponatremia are influenced by several factors, including the magnitude and rate of development of the hyponatremia, the patient’s age and gender, and the nature and severity of any underlying diseases.¹⁷⁴ The very young and very old, women, and alcoholics appear to be at particular risk.¹⁷⁵ Cell-volume adaptation to hypotonicity may be deficient in premenopausal women, who suffer more frequent and more severe neurologic consequences than men with equivalent degrees of hypotonicity.¹⁷⁶

Neurologic symptoms usually do not occur until the plasma sodium concentration falls below 125 mmol/L, at which point the patient may complain of anorexia, nausea, and malaise. Between 120 and 110 mmol/L, headache, lethargy, confusion, agitation, and obtundation may be seen. More severe symptoms (seizures, coma) may occur with levels below 110 mmol/L.¹⁷⁷ Focal neurologic findings are unusual but do occur, and transtentorial cerebral herniation has been described in severe cases, especially in young women following surgery.¹⁷⁶ In that setting, hypoxemia is common and often is associated with noncardiogenic pulmonary edema.¹⁷⁸ Hypoxia appears to exacerbate the cerebral damage in hyponatremia.¹⁷⁹

Although symptoms generally resolve with correction of the hypotonicity, permanent neurologic deficits may occur, particularly in acute severe hypotonicity, when the brain’s volume-regulatory defenses may be overwhelmed.¹⁷⁶ Profound hypotonicity that develops in less than 24 hours may be associated with residual neurologic deficits and has a 50% mortality rate in some populations.¹⁷⁶ In contrast, when hypotonicity develops more gradually, symptoms are both less common and less severe. Indeed, patients with chronic hyponatremia, even in the range of 115 to 120 mmol/L, may be completely asymptomatic.¹⁷⁴

Differential Diagnosis

Hyponatremia may coexist with a normal, high, or low plasma osmolality. Thus, the diagnostic algorithm for hyponatremia (see Fig. 57.2) begins with an assessment of the plasma osmolality (P_osm). This may be estimated by the following formula:

where P_gluc is the plasma glucose concentration and BUN is blood urea nitrogen concentration, both in mg/dL. If there is a suspicion that an unmeasured, osmotically effective solute may be implicated (e.g., mannitol or glycerol), the P_osm should be measured directly.

Isotonic hyponatremia (also known as factitious hyponatremia or pseudohyponatremia) is a laboratory artifact seen with analytic techniques that measure the mass of sodium per unit volume of plasma sampled.¹⁸⁰ It is seen in the presence of marked hypertriglyceridemia or paraproteinemia, when the measurement method involves a predilution step. Direct potentiometry (which uses an ion-selective electrode in undiluted plasma) avoids this problem.¹⁸⁰

Hypertonic hyponatremia results from the presence in ECF of abnormal amounts of osmotically effective solutes other than sodium (e.g., glucose, mannitol, or glycerol). The osmotic pressure exerted by the nonsodium solute leads to redistribution of water from the intracellular to the ECF compartment, resulting in cellular dehydration and hyponatremia. The hyponatremia is real (not pseudo-), but it is accompanied by hypertonicity and a decrease in cellular volume.

Hypotonic hyponatremia is almost always caused by an inability of the kidney to excrete sufficient electrolyte-free water to match water intake. This may occur either because the normal diluting capacity of the kidney is overwhelmed by excessive water intake or because the diluting capacity of the kidney is impaired. These alternatives usually can be distinguished by measuring the urine osmolality. A urine osmolality less than 100 mOsm/kg in a patient with hypotonic hyponatremia points to excessive water intake as the cause (Fig. 57.5). It is a prodigious feat for an individual eating a normal diet to overwhelm the normal diluting capacity of the kidney. Estimates are that one can ingest (and excrete) about 20 L of water a day without affecting the plasma osmolality appreciably.¹⁷³ Thus, patients who develop hyponatremia from so-called psychogenic or primary polydipsia—usually patients with obsessive-compulsive disorder or psychosis—typically have concurrent urinary diluting defects, either in association with the underlying mental illness or perhaps as a side effect of psychotropic or anticonvulsant medications.¹⁸¹

Not all patients with hypotonic hyponatremia and a dilute urine have primary polydipsia. The patient may be ingesting a diet so deficient in protein and salt that he excretes very little solute in the urine. In that situation (called beer potomania for obvious reasons,¹⁸² although the syndrome has been seen in other patients with very low daily solute intake¹⁸³) the low daily solute excretion limits the total amount of water that can be eliminated even with a maximally dilute urine (i.e., maximum urine volume = solute excretion ÷ minimum U_osm). This might reduce the maximum water excretion to only 3 to 4 L/day, a quantity easily exceeded by an enthusiastic beer drinker.

A urine osmolality above 100 mOsm/kg in the face of hypotonic hyponatremia signifies impaired urinary diluting capacity. The concentrated urine usually reflects a high circulating AVP level. Because circulating AVP is affected by systemic hemodynamics as well as osmolality, assessment of the patient’s ECF volume status and hemodynamics is crucial at this juncture. Hypotonic hyponatremia may be associated with normal, decreased, or increased extracellular volume.

Management and Complications

Hyponatremia per se requires treatment only when it is associated with hypotonicity. Hypertonic hyponatremia responds to the treatment of the underlying disorder, most commonly a hyperosmolar hyperglycemic state (see Chapter 58).

The therapy of hypotonic hyponatremia must be tailored to (1) the patient’s signs and symptoms, and (2) the duration of the disorder.¹⁷⁵ Severe hyponatremia (plasma sodium concentration < 115 mmol/L) can be life threatening, especially if it develops rapidly.^177,197 The therapy of symptomatic hyponatremia, irrespective of cause, is directed at raising ECF tonicity to shift water out of the intracellular space, thereby ameliorating cerebral edema. The rate of correction, however, must be carefully regulated. Overly rapid correction, particularly in patients with chronic hyponatremia, in whom cell volume adaptations may be complete, can produce osmotic demyelination syndrome.^198,199 Osmotic demyelination syndrome is associated with a variety of sometimes irreversible neurologic deficits (e.g., dysarthria, dysphagia, behavioral disturbances, ataxia, quadriplegia, coma), which typically develop 3 to 10 days after treatment.²⁰⁰ Additional risk factors for osmotic demyelination include hypokalemia, malnutrition, alcoholism, advanced age, female gender, and the postoperative state, particularly after orthotopic liver transplantation.^201–203

For patients with chronic hyponatremia (>48 hours duration) or hyponatremia of unknown duration, the plasma sodium concentration should be raised by a maximum of 0.5 mmol/L/hour, 8 to 10 mmol/L in the first 24 hours¹⁹⁹ and 20 mmol/L over the first 48 hours. Care should be taken to avoid overcorrecting the plasma sodium concentration.^187,200 In grave situations (plasma sodium concentration < 105 mmol/L or in the presence of seizure or coma), initial therapy can be more aggressive (targeting a change in the plasma sodium concentration of 1 to 2 mmol/L/hour for the first few hours), but the recommended daily target should not be exceeded.^187,200

Correction of severe symptomatic hypotonic hyponatremia, regardless of cause, should be accomplished with hypertonic (3%) saline (sodium concentration 513 mmol/L). The volume of 3% saline required can be estimated by the following formula:

For example, in a 70-kg man with a plasma sodium concentration of 105 mmol/L and total body water (TBW) of 42 L (60% of body weight), the amount of sodium needed to raise the plasma sodium concentration by 10 mmol/L is 10 × 42, or 420 mmol. Therefore, 420 ÷ 513 or 0.82 L of 3% saline would be required in the first 24 hours, or 34 mL/hour. It is important to recognize that the calculation provides only a very rough guideline. The plasma sodium concentration must be monitored frequently during treatment to adjust the rate of correction. If the rate of correction begins to exceed the target rate, the hypertonic saline infusion should be stopped; rarely, it may be necessary to administer water (enterally or intravenously) or even desmopressin²⁰⁴ in order to prevent overly rapid correction or overcorrection. Rapid extracellular volume expansion with hypertonic saline may precipitate pulmonary edema, particularly in patients with underlying heart disease. Thus, patients receiving 3% saline should be assessed frequently for evidence of volume overload. One may administer a loop diuretic if necessary, recognizing that this will enhance electrolyte-free water clearance and accelerate the correction. Rarely, administration of isotonic (normal) saline to patients with SIADH paradoxically may lower the plasma sodium concentration if the urine osmolality remains high—a process that has been called desalination.²⁰⁵

The treatment of chronic asymptomatic hypotonicity should be directed at correcting the pathophysiologic mechanisms involved in generating the hypotonic state. Because euvolemic hyponatremia represents pure water excess, treatment depends on restricting water intake to less than the daily water output. Patients with SIADH excrete little or no electrolyte-free water in the urine. Therefore, if water intake is limited to less than the amount of insensible water losses (approximately 10 mL/kg/day), the plasma sodium concentration will slowly rise. Patients with the reset osmostat variant of SIADH characteristically do not develop progressive hypotonicity, and therapy is rarely required.

If the cause of SIADH cannot be corrected and if water restriction is poorly tolerated or ineffective, a specific vasopressin (V₂) receptor antagonist (VRA) can be used.²⁰⁶ Conivaptan (parenteral) and tolvaptan (oral) were the first VRAs to be approved by the U.S. Food and Drug Administration for clinical use. These agents have changed the management of patients with SIADH.²⁰⁷ Tolvaptan reliably increases P_Na concentration, but hyponatremia recurs within 1 week after the medication is discontinued.²⁰⁸ For patients admitted with hyponatremia related to decompensated congestive heart failure, tolvaptan along with diuretic therapy improves most signs and symptoms of heart failure within 1 week,²⁰⁹ although no effect on 24-month mortality rate or heart failure morbidity has been demonstrated.²¹⁰ Before the introduction of VRAs, demeclocycline (a tetracycline antibiotic that increases electrolyte-free water excretion by inhibiting vasopressin-mediated water reabsorption in the collecting duct) was commonly used to treat patients with SIADH. Demeclocycline is contraindicated in patients with renal disease, hepatic cirrhosis, or congestive heart failure because drug-related renal insufficiency has been described in these situations.²⁰⁶ Neither VRAs nor demeclocycline are indicated for the treatment of acute, severe hyponatremia.

Therapy of hypovolemic hyponatremia should be directed at restoring intravascular volume with intravenous isotonic saline while identifying and correcting the cause of the excessive solute loss. Volume repletion readily elicits a water diuresis by increasing the delivery of fluid to the renal diluting segments and suppressing vasopressin release. As with all categories of hypotonic hyponatremia, the rate of correction must be carefully controlled.

The treatment of diuretic-induced hyponatremia is straightforward: Withdrawing the offending drug, liberalizing salt intake, and replenishing body potassium stores usually correct the disorder. Severe symptomatic hyponatremia in this setting should be treated with hypertonic saline as detailed earlier. Patients must be watched carefully after correction of the hyponatremia because relapse may occur for up to a week.¹⁹³

Resolution of the hyponatremia associated with any of the pathologic edematous disorders ultimately depends on effective treatment of the underlying disease. Regardless of the specific therapy of the underlying disorder, the mainstay of therapy for the hyponatremic edematous patient remains salt and water restriction. Diuretics are often a double-edged sword in the hyponatremic edematous patient: They may be needed to treat pulmonary vascular congestion, peripheral edema, and ascites but if used to excess can produce further decrements in effective arterial blood volume and exacerbate water retention. Strategies directed at increasing effective arterial blood volume (e.g., afterload reduction with angiotensin-converting enzyme inhibitors211,212) have had some success in increasing electrolyte-free water excretion and ameliorating hyponatremia in patients with congestive heart failure. VRAs are likely to facilitate treatment of hypervolemic hyponatremia (see earlier).²⁰⁷

Epidemiology and Clinical Manifestations

Hypernatremia is common in critically ill patients, being present on admission in about 9% of patients and developing during the course of the ICU stay in another 6%.²¹³ It is associated with a significantly higher mortality rate than is seen in patients without hypernatremia.²¹³

Sustained hypernatremia develops in patients whose water output exceeds their input. Water ingestion can defend against the development of hypernatremia even when water losses are prodigious. For that reason, hypernatremia upon presentation to the hospital occurs most commonly in patients who are incapacitated: those who have impaired thirst sensation, who cannot access water, or who cannot express their need for water (e.g., infants and patients with neurologic impairments). Similar predispositions prevail among critically ill patients. Thus, the development of hypernatremia in hospitalized patients is considered to be iatrogenic, reflecting an incomplete understanding of the factors that lead to hypernatremia.²¹³ The increased mortality rate seen in patients with hypernatremia probably is due to their underlying vulnerabilities rather than an effect of the hypernatremia itself.²¹⁴

The clinical manifestations of hypernatremia are proportional to the magnitude and rate of rise of the plasma sodium concentration and are attributable to intracellular volume contraction. To counteract cellular volume contraction, cells begin to adapt within minutes by allowing the influx of electrolytes, thus mitigating cell shrinkage. When hypernatremia lasts more than a few hours, brain cells generate new organic osmolytes. This leads to further water movement back into brain cells, restoring cell volume nearly to normal after about 3 days.²¹⁵ Thus, chronic progressive hypernatremia is associated with fewer and milder symptoms than acute severe hypernatremia. Most often, patients with longstanding hypernatremia present with weakness, lethargy, and confusion. Seizure and coma may supervene. Acute severe hypernatremia in infants and small children is associated with intracranial bleeding,²¹⁶ presumably caused by brain shrinkage and traction on the penetrating vessels. There is some controversy, however, as to whether the hypernatremia in that situation is the cause or the effect of the intracranial hemorrhage.²¹⁷

Differential Diagnosis

The plasma sodium concentration reflects the ratio of body sodium content to total body water. Thus, hypernatremia (plasma sodium concentration > 145 mmol/L) can result from loss of pure water alone, loss of hyponatric* fluid, or a gain of sodium or hypernatric fluid. It is important to distinguish among these paths to hypernatremia because they have diagnostic and therapeutic implications (Fig. 57.6).

Treatment

The initial treatment of the hypernatremic patient depends on his or her volume status. For patients with pure water losses (euvolemic), therapy has two goals: (1) reduction or replacement of ongoing water losses, and (2) replacement of the existing water deficit.

If the water losses are urinary (see Fig. 57.6) and due to CDI, ADH should be administered. Several formulations are available (Table 57.3). In the acute (postsurgical or posttraumatic) setting, L-arginine vasopressin may be used either subcutaneously or intravenously, although the latter route may be associated with hypertension and coronary spasm and should therefore be used with extreme caution.²²³ The advantage of vasopressin in this setting is its short half-life, which allows the physician to repeatedly assess the need for continued hormone replacement, especially when the disorder may be self-limited. Desmopressin (DDAVP) is a synthetic analog of vasopressin that has no vasoconstrictor properties, thus avoiding the risks of hypertension and myocardial ischemia. The treatment of the urinary water losses associated with NDI are best treated with thiazide diuretics with or without COX inhibitors. Because most of these agents are orally administered, treatment of NDI in the critically ill patient often consists of urinary water replacement until he or she is able to take medications by mouth.

The current body water deficit can be estimated by the following formula:

where TBW is total body water in liters (estimated as about 0.5 × lean body weight [kg] in women and 0.6 × lean body weight in men). For example, a 60-kg woman presenting with a plasma sodium concentration of 160 mmol/L is estimated to have a total body water deficit of 30 (1 – 0.875), or 3.75 L. This formula provides only a rough estimate of the water deficit.

The rate of water replacement should be proportional to the rapidity with which the hypernatremia developed.²¹⁵ Thus, if the hypernatremia had developed over only a few hours (such as in postsurgical or posttraumatic DI), it can be corrected just as quickly. On the other hand, hypernatremia of more than a day’s duration, or of unknown duration, must be correctly slowly in order to avoid cerebral edema. In general, one should aim to correct half the water deficit in the first 24 hours and the remainder over the next 24 to 48 hours.

Water is best administered enterally, as tap water. If that route is unavailable, 5% dextrose in water (D₅W) may be used, with the understanding that the capacity to metabolize glucose is limited to about 15 g/hour in a critically ill adult.²²⁴ Thus, even in nondiabetic patients, the administration of more than 300 mL/hour of D₅W is likely to result in hyperglycemia, which may be relatively resistant to insulin administration. Hyperglycemia will exacerbate urinary water losses by causing an osmotic diuresis. Half-normal (0.45%) saline may be a good alternative, as long as one recognizes that only half the administered volume is electrolyte-free water and that the sodium load may cause unwanted volume expansion.

Regardless of the degree of hypernatremia, normal (.9%) saline should be given intravenously to patients who present with obvious volume depletion, manifested by hypotension, tachycardia, and evidence of impaired tissue perfusion. This is consistent with the first principles of emergency and critical care, prioritizing the adequacy of the circulation. Only after the extracellular volume deficits have been largely corrected may the physician direct his or her attention to the total body water deficit (see earlier).

Patients with hypervolemic hypernatremia need reduction in their extracellular and intravascular volume before their water deficit can be corrected. Failure to do so will exacerbate the volume overload. For patients with adequate renal function, this may be accomplished with the use of diuretic drugs. Loop diuretics tend to cause the excretion of an isotonic urine. Replacement of that urine volume with pure water will allow correction of the hypervolemia and the hypernatremia simultaneously.

Because of the imprecision of the estimation formulas and the failure of the foregoing analysis to take account of other fluids and electrolytes both administered and lost, it is crucial that the plasma electrolytes be monitored frequently during the correction of hypernatremia, especially in view of the dire consequences of overly rapid correction.