Insulin

The hypokalemic effect of insulin has been known since the early twentieth century. The impact of insulin on plasma K ⁺ and plasma glucose is separable at multiple levels, suggesting independent mechanisms. For example, despite impaired glucose uptake, peripheral K ⁺ uptake is not impaired in humans with type 2 diabetes. Notably, the hypokalemic effect of insulin is not renal-dependent. Insulin and K ⁺ appear to form a feedback loop of sorts, in that increases in plasma K ⁺ have a marked stimulatory effect on insulin levels. Insulin-stimulated K ⁺ uptake, measured in rats using a K ⁺ clamp technique, is rapidly reduced by 2 days of K ⁺ depletion, prior to a modest drop in plasma K ⁺ , and in the absence of a change in plasma K ⁺ in rats subject to a lesser K ⁺ restriction for 14 days. Insulin-mediated K ⁺ uptake is thus modulated by the factors that serve to preserve plasma K ⁺ in the setting of K ⁺ deprivation (see also “ Control of Potassium Secretion: Effect of Potassium Intake ”). Inhibition of basal insulin secretion in normal subjects by somatostatin infusion increases serum K ⁺ by up to 0.5 mmol/L in the absence of a change in urinary excretion, emphasizing the crucial role of circulating insulin in the regulation of plasma K ⁺ . Clinically, inhibition of insulin secretion by the somatostatin agonist octreotide can cause significant hyperkalemia in anephric patients and patients with normal renal function.

Insulin stimulates the uptake of K ⁺ by several tissues, most prominently liver, skeletal muscle, cardiac muscle, and fat. It does so by activating several K ⁺ transport pathways, with particularly well-documented effects on the Na ⁺ -K ⁺ -ATPase. Insulin activates Na ⁺ -H ⁺ exchange and/or Na ⁺ -K ⁺ -2Cl ⁻ cotransport in several tissues; although the ensuing increase in intracellular Na ⁺ was postulated to have a secondary activating effect on Na ⁺ -K ⁺ -ATPase, it is clear that this is not the primary mechanism in most cell types. Insulin induces translocation of the Na ⁺ -K ⁺ -ATPase α ₂ -subunit to the plasma membrane of skeletal muscle cells, with a lesser effect on the α ₁ -subunit. This translocation is dependent on the activity of phosphatidylinositol-3-kinase (PI3K), which itself also binds to a proline-rich motif in the N terminus of the α-subunit. The activation of PI3K by insulin thus induces phosphatase enzymes to dephosphorylate a specific serine residue adjacent to the PI3K binding domain. Trafficking of Na ⁺ -K ⁺ -ATPase to the cell surface also appears to require the phosphorylation of an adjacent tyrosine residue, perhaps catalyzed by the tyrosine kinase activity of the insulin receptor itself. Finally, the serum- and glucocorticoid-regulated kinase-1 (SGK1) plays a critical role in insulin-stimulated K ⁺ uptake, presumably via the known stimulatory effects of this kinase on Na ⁺ -K ⁺ -ATPase activity and/or Na ⁺ -K ⁺ -2Cl ⁻ cotransport. The hypokalemic effect of insulin plus glucose is blunted in SGK1 knockout mice, with a marked reduction in hepatic insulin-stimulated K ⁺ uptake.

Sympathetic Nervous System

The sympathetic nervous system plays a prominent role in regulating the balance between extracellular and intracellular K ⁺ . Again, as is the case for insulin, the effect of catecholamines on plasma K ⁺ has been known since the 1930s ; however, a complicating issue is the differential effect of stimulating α- and β-adrenergic receptors ( Table 18.2 ). Uptake of K ⁺ by liver and muscle, with resultant hypokalemia, is stimulated via β ₂ -receptors. The hypokalemic effect of catecholamines appears to be largely independent of changes in circulating insulin and has been reported in nephrectomized animals. The cellular mechanisms whereby catecholamines induce K ⁺ uptake in muscle include an activation of the Na ⁺ -K ⁺ -ATPase, likely via increases in cyclic adenosine monophosphate (cAMP). However, β-adrenergic receptors in skeletal muscle also activate the inwardly directed Na ⁺ -K ⁺ -2Cl ⁻ cotransporter NKCC1, which may account for as much as one third of the uptake response to catecholamines.

In contrast to β-adrenergic stimulation, α-adrenergic agonists impair the ability to buffer increases in K ⁺ induced via intravenous loading or by exercise ; the cellular mechanisms whereby this occurs are not known. It is thought that β-adrenergic stimulation increases K ⁺ uptake during exercise to avoid hyperkalemia, whereas α-adrenergic mechanisms help blunt the ensuing postexercise nadir. The clinical consequences of the sympathetic control of extrarenal K ⁺ homeostasis are reviewed elsewhere in this chapter.

Acid-Base Status

The association between changes in pH and plasma K ⁺ was observed in 1934. It has long been thought that acute disturbances in acid-base equilibrium result in changes in serum K ⁺ , such that alkalemia shifts K ⁺ into cells, whereas acidemia is associated with K ⁺ release. It is thought that this effective K ⁺ -H ⁺ exchange serves to help maintain extracellular pH. Rather limited data exist for the durable concept that a change of 0.1 unit in serum pH will result in a 0.6-mmol/L change in serum K ⁺ in the opposite direction. However, despite the complexities of changes in K ⁺ homeostasis associated with various acid-base disorders, a few general observations can be made. The induction of metabolic acidosis by the infusion of mineral acids (NH ₄ ⁺ -Cl ⁻ or H ⁺ -Cl ⁻ ) consistently increases serum K ⁺ , whereas organic acidosis generally fails to increase serum K ⁺ . Notably, another study failed to detect an increase in serum K ⁺ in normal human subjects with acute acidosis secondary to duodenal NH ₄ ⁺ -Cl ⁻ infusion, in which a modest acidosis was accompanied by an increase in circulating insulin. However, as noted by Adrogué and Madias, the concomitant infusion of 350 mL of 5% dextrose in water (D ₅ W) in these fasting subjects may have served to increase circulating insulin, thus blunting the potential hyperkalemic response to NH ₄ ⁺ -Cl ⁻ . Clinically, use of the oral phosphate binder sevelamer hydrochloride in patients with end-stage kidney disease (ESKD) is associated with acidosis due to effective gastrointestinal absorption of H ⁺ -Cl ⁻ ; in hemodialysis patients, this acidosis has been associated with an increase in serum K ⁺ , which is ameliorated by an increase in dialysis bicarbonate concentration. Of note, hyperkalemia is not an expected complication of sevelamer carbonate, which has supplanted sevelamer hydrochloride as a phosphate binder. Metabolic alkalosis induced by sodium bicarbonate infusion usually results in a modest reduction in serum K ⁺ . Respiratory alkalosis reduces plasma K ⁺ by a magnitude comparable to that of metabolic alkalosis. Finally, acute respiratory acidosis increases plasma K ⁺ ; the absolute increase is smaller than that induced by metabolic acidosis secondary to inorganic acids. Again, however, some studies have failed to show a change in serum K ⁺ following acute respiratory acidosis.

Aldosterone

Aldosterone is well established as an important regulatory factor in K ⁺ excretion, and increases in plasma K ⁺ are an important stimulus for aldosterone secretion (see also “ Regulation of Renal Renin and Adrenal Aldosterone ”). However, an increasingly dominant theme is that aldosterone plays a permissive and synergistic role in K ⁺ homeostasis. This is reflected clinically in the frequent absence of hyperkalemia or hypokalemia in disorders associated with a deficiency or an overabundance of circulating aldosterone, respectively (see “ Hyperaldosteronism ” and “ Hypoaldosteronism ”). Regardless, it is clear that aldosterone and downstream effectors of this hormone have clinically relevant effects on plasma K ⁺ levels and that the ability to excrete K ⁺ is modulated by systemic aldosterone levels (see Figure 18.4 ).

Aldosterone has no effect on the density of apical SK channels in the CCD or CNT; rather, the hormone induces a marked increase in the density of apical Na ⁺ channels, thus increasing the driving force for apical K ⁺ excretion. The apical, amiloride-sensitive ENaC is comprised of three subunits, α-, β-, and γ-, that assemble together to traffic synergistically to the cell membrane and mediate Na ⁺ transport. Aldosterone activates ENaC channel complexes by multiple mechanisms. First, it induces transcription of the α-ENaC subunit, increasing the availability for co-assembly with the more abundant β- and γ-subunits. Second, aldosterone and dietary Na ⁺ -Cl ⁻ restriction stimulate a significant redistribution of ENaC subunits in the CNT and early CCD, from a largely cytoplasmic location during dietary Na ⁺ -Cl ⁻ excess to a purely apical distribution after aldosterone or Na ⁺ -Cl ⁻ restriction. Third, aldosterone induces the expression of a serine-threonine kinase called SGK-1 (serum and glucocorticoid-induced kinase-1) ; coexpression of SGK-1 with ENaC subunits results in increased expression at the plasma membrane. SGK-1 modulates membrane expression of ENaC by interfering with regulated endocytosis of its channel subunits. Specifically, the kinase interferes with interactions between ENaC subunits and the ubiquitin ligase Nedd4-2. The so-called PPxY domains in the C termini of all three ENaC subunits bind to WW domains of Nedd4-2 ; these PPxY domains are deleted, truncated, or mutated in patients with Liddle’s syndrome (see later, “ Liddle’s Syndrome ”), leading to a gain of function in channel activity. Nedd4-2 ubiquitinates ENaC subunits, thus inducing removal of channel subunits from the cell membrane. followed by degradation in lysosomes and the proteasome. A PPxY domain in SGK-1 also binds to Nedd4-2, which is a phosphorylation substrate for the kinase; phosphorylation of Nedd4-2 by SGK-1 abrogates the inhibitory effect of this ubiquitin ligase on ENaC subunits ( Figure 18.5 ).

Coordinated regulation of the epithelial Na ⁺ channel (ENaC) by the aldosterone-induced SGK kinase and ubiquitin ligase Nedd4-2. Nedd4-2 binds via its WW domains to ENaC subunits via their PPXY domains (denoted PY here), ubiquitinating the channel subunits and targeting them for removal from the cell membrane and destruction in the proteasome. Aldosterone induces the SGK kinase, which phosphorylates and inactivates Nedd4-2, thus increasing surface expression of ENaC channels. Mutations that cause Liddle’s syndrome affect the interaction between ENaC and Nedd4-2.

(From Snyder PM, Olson DR, Thomas BC: Serum and glucocorticoid-regulated kinase modulates Nedd4-2-mediated inhibition of the epithelial Na ⁺ channel. J Biol Chem 277:5-8, 2002.)

The importance of SGK-1 in K ⁺ and Na ⁺ homeostasis is illustrated by the phenotype of SGK-1 knockout mice. On a normal diet, homozygous SGK-1 ^−/− mice exhibit normal blood pressure and a normal plasma K ⁺ , with only a mild elevation of circulating aldosterone. However, dietary Na ⁺ -Cl ⁻ restriction of these mice results in relative Na ⁺ wasting and hypotension, marked weight loss, and a drop in the glomerular filtration rate (GFR), despite considerable increases in circulating aldosterone. In addition, dietary K ⁺ loading over 6 days leads to a 1.5-mmol/L increase in serum K ⁺ , also accompanied by a considerable increase in circulating aldosterone (~fivefold greater than that of wild-type littermate controls). This hyperkalemia occurs despite evident increases in apical ROMK expression, compared to the normokalemic littermate controls. The amiloride-sensitive, lumen-negative potential difference generated by ENaC is reduced in these SGK-1 knockout mice, resulting in a decreased driving force for distal K ⁺ secretion and the observed susceptibility to hyperkalemia.

Another mechanism whereby aldosterone activates ENaC involves proteolytic cleavage of the channel proteins by serine proteases. A channel-activating protease that increases channel activity of ENaC was initially identified in Xenopus laevis A6 cells. The mammalian ortholog, denoted CAP1 (channel-activating protease-1), or prostasin, is an aldosterone-induced protein in principal cells. Urinary excretion of CAP1 is increased in hyperaldosteronism, with a reduction after adrenalectomy. CAP1 is membrane-associated via a glycosylphosphatidylinositol (GPI) linkage ; mammalian principal cells also express two transmembrane proteases, denoted CAP2 and CAP3, with homology to CAP1. These and other proteases (e.g., furin, plasmin) activate ENaC by excising extracellular inhibitory domains from the α- and γ-subunits, increasing the open probability of channels at the plasma membrane. This proteolytic cleavage of ENaC appears to activate the channel by removing the so-called self-inhibitory effect of external Na ⁺ ; in the case of furin-mediated proteolysis of α-ENaC, this appears to involve the removal of an inhibitory domain from within the extracellular loop. Extracellular Na ⁺ appears to interact with a specific acidic cleft in the cleaved extracellular loop of α-ENaC, causing inhibition of the channel. Since SGK-1 increases channel expression at the cell surface, one would expect synergistic activation by coexpressed CAP1-3 and SGK, and this is indeed the case. Therefore, aldosterone activates ENaC by at least three separate synergistic mechanisms—induction of α-ENaC, induction of SGK-1 and repression of Nedd4-2, and induction of channel-activating proteases. Clinically, the inhibition of channel-activating proteases by the protease inhibitor nafamostat causes hyperkalemia due to inhibition of ENaC activity.

Effect of Potassium Intake

Changes in K ⁺ intake strongly modulate K ⁺ channel activity in the CNT and CCD (secretory capacity), in addition to H ⁺ -K ⁺ -ATPase activity in the OMCD (reabsorptive capacity). Increased dietary K ⁺ rapidly increases the activity of SK channels in the CCD and CNT, along with a modest increase in Na ⁺ channel (ENaC) activity ; this is associated with an increase in apical expression of the ROMK channel protein. The increase in ENaC and SK channel density in the CCD occurs within hours of intake of a high-K ⁺ diet, with a minimal associated increase in circulating aldosterone. BK channels in the CNT and CCD are also activated by dietary K ⁺ loading. Trafficking of BK subunits is thus affected by dietary K ⁺ , with largely intracellular distribution of α-subunits in K ⁺ -restricted rats and prominent apical expression in K ⁺ -loaded rats. Again, aldosterone does not contribute to the regulation of BK channel activity or expression in response to a high-K ⁺ diet.

A complex synergistic mix of signaling pathways regulates K ⁺ channel activity in response to changes in dietary K ⁺ (see also Chapter 6 ). In particular, the WNK ( w ith n o l ysine) kinases play a critical role in modulating distal K ⁺ secretion. WNK1 and WNK4 were initially identified as the causative genes for FHHt (see also “ Hyperkalemia: and Potassium Excretion ”). ROMK expression at the membrane of Xenopus oocytes is reduced by coexpression of WNK4; FHHt-associated mutations increase this effect, suggesting a direction inhibition of SK channels in FHHt. Transcription of the WNK1 gene generates several different isoforms; the predominant intrarenal WNK1 isoform is generated by a distal nephron transcriptional site that bypasses the N-terminal exons that encode the kinase domain, yielding a kinase-deficient short isoform (“WNK1-S”). Full-length WNK1 (WNK1-L) inhibits ROMK by inducing endocytosis of the channel protein; the shorter, kinase-deficient WNK1-S isoform inhibits this effect of WNK1-L. The ratio of WNK1-S to WNK1-L transcripts is reduced by K ⁺ restriction (greater endocytosis of ROMK) and increased by K ⁺ loading (reduced endocytosis of ROMK), suggesting that this ratio between WNK1-S and WNK1-L functions as a molecular switch to regulate distal K ⁺ secretion. The BK channel is also regulated by the WNK kinases.

The membrane trafficking of ROMK is also modulated by tyrosine phosphorylation of the channel protein, such that tyrosine phosphorylation stimulates endocytosis, and tyrosine dephosphorylation induces exocytosis. Intrarenal activity of the cytoplasmic tyrosine kinases c-src and c-yes is inversely related to dietary K ⁺ intake, with a decrease under high K ⁺ conditions and a marked increase after several days of K ⁺ restriction. Several studies have implicated the intrarenal generation of superoxide anions in the activation of cytoplasmic tyrosine kinases by K ⁺ depletion. Potential candidates for the upstream hormonal signals include angiotensin II (Ang II) and growth factors such as insulin-like growth factor-1 (IGF-1). In particular, Ang II inhibits ROMK activity in K ⁺ -restricted rats, but not rats on a normal K ⁺ diet. This inhibition by Ang II involves downstream activation of superoxide production and c-src activity, such that the induction of Ang II by a low-K ⁺ diet appears to play a major role in reducing distal tubular K ⁺ secretion.

Excitable Tissues: Muscle and Heart

Hypokalemia is a well-described risk factor for ventricular and atrial arrhythmias. For example, in patients undergoing cardiac surgery, a serum K ⁺ of less than 3.5 mmol/L is a predictor of serious intraoperative arrhythmia, perioperative arrhythmia, and postoperative atrial fibrillation. Moderate hypokalemia does not, however, appear to increase the risk of serious arrhythmia during exercise stress testing. Electrocardiographic changes in hypokalemia include broad flat T waves, ST depression, and QT prolongation; these are most marked when serum K ⁺ is less than 2.7 mmol/L. Hypokalemia, often accompanied by hypomagnesemia, is an important cause of the long QT syndrome (LQTS) and torsades de pointes, either alone or in combination with drug toxicity, or with LQTS-associated mutations in cardiac K ⁺ and Na ⁺ channels. Hypokalemia accelerates the clathrin-dependent internalization and degradation of the cardiac hERG (human ether-à-go-go–related gene) K ⁺ channel protein. hERG encodes pore-forming subunits of the cardiac rapidly activating delayed rectifier K ⁺ channel (I _Kr ); I _Kr is largely responsible for potassium efflux during phases 2 and 3 of the cardiac action potential. Loss-of-function mutations in hERG reduce I _Kr and cause type II LQTS ; downregulation of hERG and I _Kr by hypokalemia provides an elegant explanation for the association with LQTS and torsades de pointes.

In accordance with the Nernst equation, the resting membrane potential is related to the ratio of the intracellular to extracellular potassium concentration. In skeletal muscle, a reduction in plasma K ⁺ will increase this ratio and therefore hyperpolarize the cell membrane (i.e., make the resting potential more electronegative); this impairs the ability of the muscle to depolarize and contract, leading to weakness. However, in some human cardiac cells, particularly Purkinje fibers in the conducting system, hypokalemia results in a paradoxic depolarization ; this paradoxic depolarization plays an important role in the genesis of hypokalemic cardiac arrhythmias. Resting membrane potential in excitable cells is largely determined by a large family of K2P1 K ⁺ channels, so-named due to the presence of two pore-forming (P) loop domains in each subunit. Hypokalemia causes K2P1 channels, which are normally selective for potassium, to transport sodium into cells suddenly, causing the paradoxic depolarization. Notably, rodent cardiac cells respond to hypokalemia with a Nernst equation–predicted hyperpolarization and, unlike human cardiomyocytes, do not express the K2P1 channel TWIK-1; genetic manipulation indicates that TWIK-1 expression confers this paradoxic depolarization behavior on human and mouse cardiomyocytes.

In skeletal muscle, hypokalemia causes hyperpolarization, thus impairing the capacity to depolarize and contract. Weakness and paralysis is therefore a not infrequent consequence of hypokalemia of diverse causes. On a historical note, the realization in 1946 that K ⁺ replacement reversed the hypokalemic diaphragmatic paralysis induced by treatment of diabetic ketoacidosis (DKA) was a milestone in diabetes care. Pathologically, muscle biopsies in hypokalemic myopathy demonstrate phagocytosis of degenerating muscle fibers, fiber regeneration, and atrophy of type 2 fibers. Most patients with significant myopathy will have elevations in creatine kinase, and hypokalemia of diverse causes predisposes to rhabdomyolysis, with acute renal failure.

Renal Consequences

Hypokalemia causes a host of structural and functional changes in the kidney, which are reviewed in detail elsewhere. In humans, the renal pathology includes a relatively specific proximal tubular vacuolization, interstitial nephritis, and renal cysts. Hypokalemic nephropathy can cause ESKD, mostly in patients with long-standing hypokalemia due to eating disorders and/or laxative abuse ; acute renal failure with proximal tubular vasculopathy has also been described. In animal models, hypokalemia increases susceptibility to acute renal failure induced by ischemia, gentamicin, and amphotericin. Potassium restriction in rats induces cortical AT-II and medullary endothelin-1 expression, with an ischemic pattern of renal injury.

The prominent functional changes in renal physiology that are induced by hypokalemia include Na ⁺ -Cl ⁻ retention, polyuria, phosphaturia, hypocitraturia, and increased ammoniagenesis. K ⁺ depletion in rats causes proximal tubular hyperabsorption of Na ⁺ -Cl ⁻ in association with an upregulation of Ang II, AT ₁ receptor, and α ₂ -adrenergic receptor in this nephron segment. NHE3, the dominant apical Na ⁺ entry site in the proximal tubule, is massively (>700%) upregulated in K ⁺ -deficient rats, which is consistent with the observed hyperabsorption of Na ⁺ -Cl ⁻ and bicarbonate. Polyuria in hypokalemia is due to polydipsia and to a vasopressin-resistant defect in urine-concentrating ability. This renal concentrating defect is multifactorial, with evidence for a reduced hydroosmotic response to vasopressin in the collecting duct and for decreased Na ⁺ -Cl ⁻ absorption by the TAL. K ⁺ restriction has been shown to result in a rapid, reversible decrease in the expression of aquaporin-2 in the collecting duct, beginning in the CCD and extending to the medullary collecting duct within the first 24 hours. In the TAL, the marked reductions seen during K ⁺ restriction in the apical K ⁺ channel ROMK and apical Na ⁺ -K ⁺ -2Cl ⁻ cotransporter NKCC2 reduce Na ⁺ -Cl ⁻ absorption, thus inhibiting countercurrent multiplication and the driving force for water absorption by the collecting duct.

Cardiovascular Consequences

A large body of experimental and epidemiologic evidence has implicated hypokalemia and/or reduced dietary K ⁺ in the genesis or worsening of hypertension, heart failure, and stroke. K ⁺ depletion in young rats induces hypertension, with a salt sensitivity that persists after K ⁺ levels are normalized; presumably this salt sensitivity is due to the significant tubulointerstitial injury induced by K ⁺ restriction. Short-term K ⁺ restriction in healthy humans and patients with essential hypertension also induces Na ⁺ -Cl ⁻ retention and hypertension, and abundant epidemiologic data have linked dietary K ⁺ deficiency and/or hypokalemia with hypertension. Correction of hypokalemia is particularly important in hypertensive patients treated with diuretics; blood pressure in this setting is improved with the establishment of normokalemia, and the cardiovascular benefits of diuretic agents are blunted by hypokalemia. Hypokalemia reduces insulin secretion; this mechanism may play an important role in thiazide-associated diabetes. Finally, K ⁺ depletion may play important roles in the pathophysiology and progression of heart failure.

Excitable Tissues: Muscle and Heart

Hyperkalemia constitutes a medical emergency, primarily due to its effect on the heart. Hyperkalemia depolarizes cardiac myocytes, reducing the membrane potential from −90 mV to approximately −80 mV. This brings the membrane potential closer to the threshold for generation of an action potential; mild and/or rapid-onset hyperkalemia will initially increase cardiac excitability, since a lesser depolarizing stimulus is required to generate an action potential. Mild increases in extracellular K ⁺ also affect the repolarization phase of the cardiac action potential, via increases in I _Kr ; as discussed earlier (see “ Consequences of Hypokalemia ”), I _Kr is highly sensitive to changes in extracellular K ⁺ . This effect on repolarization is thought to underlie the early signs of hyperkalemia, including ST-T segment depression, peaked T waves, and Q-T interval shortening. Persistent and increasing depolarization inactivates cardiac sodium channels, thus reducing the rate of phase 0 of the action potential (V _max ); the decrease in V _max results in a reduction in myocardial conduction, with progressive prolongation of the P wave, PR interval, and QRS complex. Severe hyperkalemia results in loss of the P wave and a progressive widening of the QRS complex; fusion with T waves causes a sine wave sinoventricular rhythm.

Cardiac arrhythmias associated with hyperkalemia include sinus bradycardia, sinus arrest, slow idioventricular rhythms, ventricular tachycardia, ventricular fibrillation, and asystole. The differential diagnosis and treatment of a wide-complex tachycardia in hyperkalemia can be particularly problematic; moreover, hyperkalemia potentiates the blocking effect of lidocaine on the cardiac Na ⁺ channel, such that use of this agent may precipitate asystole or ventricular fibrillation in this setting. Hyperkalemia can also cause a type I Brugada pattern in the electrocardiogram, with a pseudo–right bundle branch block (RBBB) and persistent coved ST segment elevation in at least two precordial leads. This hyperkalemic Brugada sign occurs in critically ill patients with significant hyperkalemia (serum K ⁺ > 7 mmol/L) and can be differentiated from genetic Brugada syndrome by an absence of P waves, marked QRS widening, and an abnormal QRS axis.

Classically, the electrocardiographic manifestations in hyperkalemia progress as shown in Table 18.3 . However, these changes are notoriously insensitive, such that only 55% of patients with serum K ⁺ more than 6.8 mmol/L in one case series manifested peaked T waves. There is large interpatient variability in the absolute potassium level, leading to electrocardiographic changes and cardiac toxicity of hyperkalemia. Relevant variables include the rapidity of the onset of hyperkalemia and the presence or absence of concomitant hypocalcemia, acidemia, and/or hyponatremia. Hemodialysis patients and patients with chronic renal failure in particular may not demonstrate electrocardiographic changes. Care should also be taken to adequately distinguish the symmetrically peaked, church steeple T waves induced by hyperkalemia from T wave changes of other causes. The ratio of precordial T wave to R wave amplitude (T/R ratio) may be a more specific sign of hyperkalemia than T wave tenting.

Hyperkalemia can also rarely present with ascending paralysis, termed secondary hyperkalemic paralysis to differentiate it from familial hyperkalemic periodic paralysis (HYPP). This presentation of hyperkalemia can mimic Guillain-Barré syndrome and may include diaphragmatic paralysis and respiratory failure. Hyperkalemia from a diversity of causes can cause paralysis, as reviewed by Evers and colleagues. The mechanism is not entirely clear; however, nerve conduction studies in one case suggested a neurogenic mechanism, rather than a direct effect on muscle excitability.

In contrast to secondary hyperkalemic paralysis, HYPP is a primary myopathy. Patients with HYPP develop myopathic weakness during hyperkalemia induced by increased K ⁺ intake or rest after heavy exercise. The hyperkalemic trigger in HYPP serves to differentiate this syndrome from hypokalemic periodic paralysis (HOKP); a further distinguishing feature is the presence of myotonia in HYPP. Depolarization of skeletal muscle by hyperkalemia unmasks an inactivation defect in a tetrodotoxin-sensitive Na ⁺ channel in patients with HYPP, and autosomal dominant mutations in the SCN4A gene encoding this channel cause most forms of the disease. Mild muscle depolarization (5 to 10 mV) in HYPP results in a persistent inward Na ⁺ current through the mutant channel; the normal, allelic SCN4 channels quickly recover from inactivation and can then be reactivated, resulting in myotonia. When muscle depolarization is more marked (20 to 30 mV), all the Na ⁺ channels are inactivated, rendering the muscle inexcitable and causing weakness ( Figure 18.8 ). Related disorders due to mutations within the large SCN4A channel protein include HOKP type II, paramyotonia congenita, and K ⁺ -aggravated myopathy. American quarter horses have a high incidence (4.4%) of HYPP due to a mutation in equine SCN4A traced to the sire named Impressive (see Figure 18.8 ). Finally, loss-of-function mutations in the muscle-specific K ⁺ channel subunit MinK-related peptide 2 (MiRP2) have also been shown to cause HYPP; MiRP2 and the associated Kv3.4 K ⁺ channel play a role in setting the resting membrane potential of skeletal muscle.

Hyperkalemic periodic paralysis (HYPP) due to mutations in the voltage-gated Na ⁺ channel of skeletal muscle. A, This disorder is particularly common in thoroughbred quarter horses; an affected horse is shown during a paralytic attack, triggered by rest after heavy exercise. B, Mechanistic explanation for muscle paralysis in HYPP.

(From Lehmann-Horn F, Jurkat-Rott K: Voltage-gated ion channels and hereditary disease. Physiol Rev 79:1317–1372, 1999; A courtesy Dr. Eric Hoffman).

Renal Consequences

Hyperkalemia has a significant effect on the ability to excrete an acid urine due to interference with the urinary excretion of ammonium (NH ₄ ⁺ ). Potassium loading in humans results in modest reduction in urinary NH ₄ ⁺ excretion and an impaired response to acid loading. In rats, chronic potassium loading leads to hyperkalemia and a metabolic acidosis due to a 40% reduction in urinary NH ₄ ⁺ excretion. Proximal tubular ammonia generation falls, but without a significant effect on proximal tubular secretion of NH ₄ ⁺ . The TAL absorbs NH ₄ ⁺ from the tubular lumen, followed by countercurrent multiplication and ultimately excretion from the medullary interstitium ; hyperkalemia appears to inhibit renal acid excretion by competing with NH ₄ ⁺ for reabsorption by the TAL.

The NH ₄ ⁺ ion has the same ionic radius as K ⁺ and can be transported in lieu of K ⁺ by NKCC2, the apical Na ⁺ -K ⁺ -NH ₄ ⁺ -2Cl ⁻ cotransporter of the TAL; NH ₄ ⁺ exits the TAL via the basolateral Na ⁺ -H ⁺ exchanger NHE4. As is the case for other cations, countercurrent multiplication of NH ₄ ⁺ by the TAL greatly increases the concentration of NH ₄ ⁺ -NH ₃ available for secretion in the collecting duct. The NH ₄ ⁺ produced by the proximal tubule in response to acidosis is thus reabsorbed across the TAL, concentrated by countercurrent multiplication in the medullary interstitium, and secreted in the collecting duct. The capacity of the TAL to reabsorb NH ₄ ⁺ is increased during acidosis due to an induction of NKCC2 and NHE4 expression. Hyperkalemia induces acidosis in rats by reducing the NH ₄ ⁺ between the vasa recta (surrogate for interstitial fluid) and collecting duct due to interference with absorption of NH ₄ ⁺ by the TAL.

Clinically, patients with hyperkalemic acidosis due to hyporeninemic hypoaldosteronism demonstrate an increase in urinary NH ₄ ⁺ excretion in response to normalization of plasma K ⁺ with cation exchange resins, indicating a significant role for hyperkalemia in generation of the acidosis.

Nonrenal Potassium Loss

The loss of K ⁺ from skin is typically low, with the exception of extremes in physical exertion. Direct gastric loss of K ⁺ due to vomiting or nasogastric suctioning is also typically minimal; however, the ensuing hypochloremic alkalosis results in persistent kaliuresis due to secondary hyperaldosteronism and bicarbonaturia. Intestinal loss of K ⁺ due to diarrhea is a quantitatively important cause of hypokalemia, given the worldwide prevalence of diarrheal disease, and may be associated with acute complications, such as myopathy and flaccid paralysis. The presence of a non–anion gap metabolic acidosis with a negative urinary anion gap (consistent with an intact ability to increase NH ₄ ⁺ excretion) should strongly suggest diarrhea as a cause of hypokalemia. Polyethylene glycol–based bowel preparation regimens for colonoscopy can also lead to hypokalemia in older patients. Noninfectious gastrointestinal processes such as celiac disease, ileostomy, and chronic laxative abuse can present with acute hypokalemic syndromes or with chronic complications, such as ESKD.

Three reports initially identified a novel association between colonic pseudo-obstruction (Ogilvie’s syndrome) and hypokalemia due to secretory diarrhea with an abnormally high K ⁺ content. In one patient with concomitant ESKD, immunohistochemistry revealed massive upregulation of the apical BK channel throughout the surface crypt axes ; colonic BK channels may play a significant role in intestinal K ⁺ secretion in a variety of pathologies, including ESKD. Several hypotheses for the association between Ogilvie’s syndrome and enhanced intestinal K ⁺ secretion have been postulated, including active stimulation by catecholamines induced by colonic pseudo-obstruction ; BK channels appear to mediate adrenaline-induced colonic K ⁺ secretion.

Increased fecal loss of K ⁺ may play a broader role in hypokalemia associated with diarrhea. Recruitment of colonic BK channels along intestinal crypts, similar to that seen in Ogilvie’s syndrome, has thus been demonstrated as a consistent feature of colonic biopsies in ulcerative colitis. Direct enhancement of intestinal K ⁺ excretion has also been demonstrated in a hypokalemic patient with Crohn’s disease following treatment with budesonide.

Renal Potassium Loss

Hyperaldosteronism

Increases in circulating aldosterone (hyperaldosteronism) may be primary or secondary. Increased levels of circulating renin in secondary forms of hyperaldosteronism lead to increased levels of Ang II, and thus aldosterone, and can be associated with hypokalemia; causes include renal artery stenosis, Page kidney (renal compression by a subcapsular mass or hematoma, with hyperreninemia), a paraneoplastic process, or renin-secreting renal tumors. The incidence of hypokalemia in renal artery stenosis is thought to be less than 20%. An unusual presentation of renal artery stenosis and renal ischemia is what has been termed the hyponatremic hypertensive syndrome, in which concurrent hypokalemia may be profound.

Primary hyperaldosteronism may be genetic or acquired. Hypertension and hypokalemia, generally attributed to increases in circulating 11-deoxycorticosterone, are seen in patients with congenital adrenal hyperplasia due to defects in steroid 11β-hydroxylase or steroid 17α-hydroxylase ; deficient 11β-hydroxylase results in virilization and other signs of androgen excess, whereas reduced sex steroids in 17α-hydroxylase deficiency result in hypogonadism. The two major forms of isolated PA are denoted familial hyperaldosteronism type I (FH-I, also known as glucocorticoid-remediable hyperaldosteronism [GRA]) and familial hyperaldosteronism type II (FH-II), in which aldosterone production is not repressible by exogenous glucocorticoids. Patients with FH-II are clinically indistinguishable from sporadic forms of PA due to bilateral adrenal hyperplasia; a gene has been localized to chromosome 7p22 by linkage analysis, but has yet to be characterized. A third form of familial hyperaldosteronism (FH-III) was initially described in 2008, with hyporeninemia, hyperaldosteronism resistant to dexamethasone, and very high levels of 18-oxocortisol and 18-hydroxycortisol. FH-III is due to somatic mutations in the adrenal K ⁺ channel KCNJ5, which endow the channel with a novel Na ⁺ conductance and activate adrenal glomerulosa proliferation and aldosterone release ; somatic mutations in KCNJ5 are also found in spontaneous adrenal adenomas (see later).

Patients with FH-I (GRA) are generally hypertensive, typically presenting at an early age; the severity of hypertension is variable, however, such that some affected individuals are normotensive. Aldosterone levels are modestly elevated and regulated solely by ACTH. The diagnosis can be biochemically confirmed by a dexamethasone suppression test, with a suppression of aldosterone to less than 4 ng/dL consistent with the diagnosis. Patients also have high levels of abnormal hybrid 18-hydroxylated steroids, generated by transformation of steroids typically formed in the zona fasciculata by aldosterone synthase, an enzyme that is normally expressed in the zona glomerulosa. FH-I has been shown to be caused by a chimeric gene duplication between the homologous 11β-hydroxylase gene ( CYP11B1 ) and aldosterone synthase gene ( CYP11B2 ), fusing the ACTH-responsive 11β-hydroxylase promoter to the coding region of aldosterone synthase. This chimeric gene is thus unde the control of ACTH and is expressed in a glucocorticoid-repressible fashion. Ectopic expression of the hybrid CYP11B1-CYP11B1 gene in the zona fasciculata has been reported in a single case in which adrenal tissue became available for molecular analysis. Direct genetic testing for the hybrid CYP11B1-CYP11B2 has largely supplanted biochemical screening for FH-I; genetic testing for FH-I should be pursued in patients with PA and a family history of PA and/or of strokes at a young age, or in younger patients with PA (age < 20 years).

Although the initial patients reported with FH-I were hypokalemic, most are normokalemic, albeit perhaps with a propensity to develop hypokalemia while on thiazide diuretics. Patients with FH-I are able to increase K ⁺ excretion appropriately in response to K ⁺ loading or fludrocortisone, but fail to increase plasma aldosterone in response to hyperkalemia. This may reflect the ectopic expression of the chimeric aldosterone synthase in the adrenal fasciculata, which likely lacks the appropriate constellation of ion channels to respond to increases in extracellular K ⁺ with an increase in aldosterone secretion.

Acquired causes of PA include aldosterone-producing adenomas (APAs; 35% of cases), primary (or unilateral) adrenal hyperplasia (PAH; 2% of cases), idiopathic hyperaldosteronism (IHA) due to bilateral adrenal hyperplasia (60% of cases), and adrenal carcinoma(<1% of cases). A rare case involving paraneoplastic overexpression of aldosterone synthase in lymphoma has also been described.

The molecular characterization of adrenal adenomas with whole-genome sequencing and related techniques has been remarkably fruitful. In particular, acquired mutations in the adrenal K ⁺ channel KCNJ5 can be detected in about 40% of aldosterone-producing adrenal adenomas. As in FH-III (see earlier), these somatic mutations endow the channel with a novel Na ⁺ conductance, leading to adrenal glomerulosa cell depolarization, Ca ²⁺ influx, and aldosterone release. Clinically, patients with adrenal KCNJ5 mutations have a higher preoperative aldosterone level and higher lateralization index in adrenal vein sampling, without affecting the surgical response to adrenalectomy. Less frequently, somatic mutations in adrenal adenomas can be detected in the calcium channel CACNAID or in a subunit (ATP2B3) of the Ca ²⁺ -ATPase pump, predicted also to lead to increased Ca ²⁺ influx and aldosterone release. Acquired mutations in the ATP1A α ₁ -subunit of Na+-K+-ATPase are in turn thought to generate chronic depolarization, leading also to exaggerated aldosterone release.

Increasing use of the plasma aldosterone concentration (PAC)/plasma renin activity (PRA) ratio in hypertension clinics has led to reports of a much higher incidence of PA than previously appreciated, with incidence rates in hypertension ranging from 0% to 72% ; however, the prevalence was 3.2% in a large multicenter study of patients with mild to moderate hypertension without hypokalemia. Regardless, the PAC/PRA ratio is a screening tool, which typically must be confirmed by aldosterone suppression testing, which measures PAC or aldosterone secretion after loading with salt or intravenous saline. After controlling hypertension and hypokalemia, oral salt loading over 3 days is followed by measurement of 24-hour urine aldosterone, sodium, and creatinine excretion. The 24-hour sodium excretion should exceed 200 mmol/day for adequate suppression, and a urinary aldosterone of more than 33 nmol/day (12 µm/day) is consistent with PA. Alternatively, in the saline infusion test, recumbent patients are infused with 2 L of isotonic saline over 4 hours, followed by measurement of PAC. In patients without PA, the measured PAC after saline infusion should decrease to less than 139 pmol/L. The measured PAC in patients with PA usually does not decrease to less than 277 pmol/L; indeterminate values between 139 and 277 pmol/L can been seen in patients with IHA. It should be noted, however, that patients with high-probability factors (hypokalemia, hypertension, high PAC/PRA ratio, abnormalities) may not necessarily require confirmatory testing, proceeding instead directly to adrenal venous sampling.

Since surgery can be curative in APA, adequate differentiation of APA from IHA is critical; this requires adrenal imaging and adrenal venous sampling ( Figure 18.9 ). Contemporary reports and recommendations have thus emphasized the continued importance of adrenal vein sampling in subtype differentiation. Laparoscopic adrenalectomy has increasingly been the preferred surgical management for APA or PAH. Mineralocorticoid receptor antagonists are indicated for medical therapy of PA, with carefully monitored use of glucocorticoid to suppress ACTH in some patients with FH-I.

Diagnostic algorithm for patients with primary hyperaldosteronism. Adrenal adenoma (APA) must be distinguished from glucocorticoid remediable hyperaldosteronism (FH-I [GRA]), primary or unilateral adrenal hyperplasia (PAH), and idiopathic hyperaldosteronism (IHA). This requires computed tomography (CT), adrenal venous sampling (AVS), and the relevant diagnostic biochemical and hormonal assays (see text).

(From Young WF Jr: Adrenalectomy for primary aldosteronism. Ann Intern Med 138:157-159, 2003.)

The true incidence of hypokalemia in patients with acquired forms of PA remains difficult to evaluate due to a variety of factors. First, historically, patients have only been screened for hyperaldosteronism when hypokalemia is present, so even case series from clinics with such a referral pattern may suffer from a selection bias; other series have concentrated on hypertensive patients, also with a selection bias. Second, the incidence of hypokalemia is higher in adrenal adenomas than in IHA, likely due to higher average levels of aldosterone. Third, since increased kaliuresis in hyperaldosteronism can be induced by dietary Na ⁺ -Cl ⁻ loading or diuretics, dietary factors and/or medications may play a role in the incidence of hypokalemia at presentation. Regardless, it is clear that hypokalemia is not a universal feature of PA; this is perhaps not unexpected, since aldosterone does not appear to affect the hypokalemic response of H ⁺ -K ⁺ -ATPase, the major reabsorptive pathway for K ⁺ in the distal nephron (see also Chapter 6 ). A related issue is whether PA is underdiagnosed when hypokalemia is used as a criterion for further investigation; the utility of the PAC/PRA ratio in screening for hyperaldosteronism is an active issue in hypertension research.

Finally, hypokalemia may also occur with systemic increases in glucocorticoids. In bona fide Cushing’s syndrome caused by increases in pituitary ACTH, the incidence of hypokalemia is only 10%, whereas it is 57% to 100% in patients with ectopic ACTH, despite a similar incidence of hypertension. Ectopic ACTH expression is associated primarily with neuroendocrine malignancies, most commonly bronchial carcinoid tumors, small lung cancer, and other neuroendocrine tumors. Indirect evidence suggests that the activity of renal 11β-hydroxysteroid dehydrogenase-2 (11βHSD-2) is reduced in patients with ectopic ACTH compared with Cushing’s syndrome, resulting in a syndrome of apparent mineralocorticoid excess (see later). Whether this reflects a greater degree of saturation of the enzyme by circulating cortisol or direct inhibition of 11βHSD-2 by ACTH is not entirely clear, and there is evidence for both mechanisms ; however, indirect indices of 11βHSD-2 activity in patients with ectopic ACTH expression correlate with hypokalemia and other measures of mineralocorticoid activity. Similar mechanisms likely underlie the severe hypokalemia reported in patients with familial glucocorticoid resistance, in which loss-of-function mutations in the glucocorticoid receptor result in marked hypercortisolism without cushingoid features, accompanied by very high ACTH levels.

Familial Hypokalemic Alkalosis

Bartter’s Syndrome.

Bartter’s and Gitelman’s syndromes are the two major variants of familial hypokalemic alkalosis; Gitelman’s syndrome is a much more common cause of hypokalemia than Bartter’s syndrome. Whereas a clinical subdivision of these syndromes has been used in the past, a genetic classification is increasingly in use, due in part to phenotypic overlap. Patients with classic Bartter’s syndrome (BS) typically suffer from polyuria and polydipsia and manifest a hypokalemic, hypochloremic alkalosis. They may have an increase in urinary calcium (Ca ²⁺ ) excretion, and 20% are hypomagnesemic. Other features include marked elevation of plasma Ang II, plasma aldosterone, and plasma renin levels. Patients with antenatal BS present earlier in life with a severe systemic disorder characterized by marked electrolyte wasting, polyhydramnios, and significant hypercalciuria with nephrocalcinosis. Prostaglandin synthesis and excretion is significantly increased and may account for many of the systemic symptoms. Decreasing prostaglandin synthesis by COX inhibition can improve polyuria in patients with BS by reducing the amplifying inhibition of urine-concentrating mechanisms by prostaglandins. Indomethacin also increases plasma K ⁺ and decreases plasma renin activity, but does not correct the basic tubular defect; it does, however, appear to help increase the growth of BS patients. Of interest, COX-2 immunoreactivity is increased in the TAL and macula densa of patients with BS, and studies have indicated a clinical benefit of COX-2 inhibitors.

Early studies in BS have suggested that these patients have a defect in the function of the TAL. Many of the clinical features are mimicked by the administration of loop diuretics, to which at least a subset of patients with antenatal BS do not respond. The apical Na ⁺ -K ⁺ -2Cl ⁻ cotransporter (NKCC2, SLC12A1) of the mammalian TAL ( Figure 18.10 ) was thus an early candidate gene. In 1996, disease-associated mutations were found in the human NKCC2 gene in four kindreds with antenatal BS ; in the genetic classification of BS, these patients are considered to have BS type I. Although the functional consequences of disease-associated NKCC2 mutations have not been comprehensively studied, the first and subsequent reports included patients with frameshift mutations and premature stop codons that predicted the absence of a functional NKCC2 protein.

Bartter’s syndrome and the thick ascending limb. Bartter’s syndrome can result from loss-of-function mutations in the Na ⁺ -K ⁺ -2Cl ⁻ cotransporter NKCC2, the K ⁺ channel subunit ROMK, or the Cl ⁻ channel subunits CLC-NKB and Barttin (Bartter’s syndrome types I to IV, respectively). Gain-of-function mutations in the calcium-sensing receptor (CaSR) can also cause a Bartter’s syndrome phenotype (type V); the CaSR has an inhibitory effect on salt transport by the thick ascending limb, targeting several transport pathways. ROMK encodes the low-conductance, 30-pS K ⁺ channel in the apical membrane and also appears to function as a critical subunit of the higher conductance 70-pS channel. The loss of K ⁺ channel activity in Bartter’s syndrome type II leads to reduced apical K ⁺ recycling and reduced Na ⁺ -K ⁺ -2Cl ⁻ cotransport. Decreased apical K ⁺ channels also lead to a decrease in the lumen-positive potential difference, which drives paracellular Na ⁺ , Ca ²⁺ , and Mg ²⁺ transport.

BS is a genetically heterogeneous disease. Given the role of apical K ⁺ permeability in the TAL, encoded at least in part by ROMK, this K ⁺ channel was another early candidate gene. K ⁺ recycling via the Na ⁺ -K ⁺ -2Cl ⁻ cotransporter and apical K ⁺ channels generates a lumen-positive potential difference in the TAL, which drives the paracellular transport of Na ⁺ and other cations (see also Figure 18.10 ). Multiple disease-associated mutations in ROMK have been reported in patients with BS type II, most of whom exhibited the antenatal phenotype. Finally, mutations in BS type III have been reported in the chloride channel CLC-NKB, which is expressed at the basolateral membrane of at least the TAL and DCT. Patients with mutations in CLC-NKB typically have the classic Bartter’s phenotype, with a relative absence of nephrocalcinosis. In a significant fraction of patients with BS, the NKCC2, ROMK, and CLC-NKB genes are not involved. For example, a subset of patients with associated sensorineural deafness exhibit linkage to chromosome 1p31 ; the gene for this syndrome, denoted Barttin, is an obligatory subunit for the CLC-NKB chloride channel. The occurrence of deafness in these patients suggests that Barttin functions in the regulation or function of Cl ⁻ channels in the inner ear. Notably, the CLC-NKB gene is immediately adjacent to that for another epithelial Cl ⁻ channel, denoted CLC-NKA; digenic inactivation was described in two siblings with deafness and BS, suggesting that CLC-NKA plays an important role in Barttin-dependent Cl ⁻ transport in the inner ear.

Patients with activating mutations in the CaSR have been described with autosomal dominant hypocalcemia and hypokalemic alkalosis. The CaSR is heavily expressed at the basolateral membrane of the TAL, where it is thought to play an important inhibitory role in regulating the transcellular transport of both Na ⁺ -Cl ⁻ and Ca ²⁺ . For example, activation of the basolateral CaSR in the TAL is known to reduce apical K ⁺ channel activity, which would induce a Bartter-like syndrome (see Figure 18.10 ). Coexpression of NKCC2 with a Bartter’s syndrome gain-of-function mutant in the CaSR reveals reduced phosphorylation and reduced activity of NKCC2, dependent on the generation of inhibitory arachidonic acid–derived metabolites known to inhibit TAL function. Genetic activation of the CaSR by these mutations is also expected to increase urinary Ca ²⁺ excretion by inhibiting generation of the lumen-positive potential difference that drives paracellular Ca ²⁺ transport in the TAL. In addition, the set point of the CaSR response to Ca ²⁺ in the parathyroid is shifted to the left, inhibiting parathyroid hormone (PTH) secretion by this gland. It is very likely that the positional cloning of other BS genes will have a considerable impact on the mechanistic understanding of the TAL.

Despite the reasonable correlation between the disease gene involved and the associated subtype of familial alkalosis, there is significant phenotypic overlap and phenotypic variability in hereditary hypokalemic alkalosis. For example, patients with mutations in CLC-NKB usually exhibit classic BS, but can present with a more severe antenatal phenotype, or even with a phenotype similar to Gitelman’s syndrome. With respect to BS due to mutations in NKCC2, a number of patients have been described with variant presentations, including an absence of hypokalemia. Two brothers were described with a late onset of mild BS; these patients were found to be compound heterozygotes for a mutant form of NKCC2 that exhibits partial function, with a loss-of-function mutation on the other NKCC2 allele.

BS type II is particularly relevant to K ⁺ homeostasis, given that ROMK is the SK secretory channel of the CNT and CCD (see Chapter 6 , “ K ⁺ Secretion by the Distal Convoluted Tubule, Connecting Tubule, and Cortical Collecting Duct ”). Patients with BS type II typically have slightly higher serum K ⁺ than the other genetic forms of BS ; patients with severe (9.0 mmol/L), transient, neonatal hyperkalemia have also been described. It is likely that this reflects a transient developmental deficit in the other K ⁺ channels involved in distal K ⁺ secretion, including the apical maxi-K channel responsible for flow-dependent K ⁺ secretion in the distal nephron. Distal K ⁺ secretion in ROMK knockout mice is primarily mediated by maxi-K–BK channel activity, such that developmental deficits in this channel would lead to hyperkalemia in BS type II. The mammalian TAL has two major apical K ⁺ conductances, the 30-pS channel corresponding to ROMK and a 70-pS channel ; both are thought to play a role in transepithelial salt transport by the TAL. ROMK is evidently a subunit of the 70-pS channel, given the absence of this conductance in TAL segments of ROMK knockout mice. The identity of the other putative subunit of this 70-pS channel is not as yet known; one would assume that deficiencies in this gene would also be a cause of BS.

Finally, BS must be clinically differentiated from the various causes of pseudo–Bartter’s syndrome; these commonly include laxative abuse, furosemide abuse, and bulimia (see “ Clinical Approach to Hypokalemia ”). Other reported causes include gentamicin nephrotoxicity, Sjögren’s syndrome, and cystic fibrosis (CF). Fixed loss of Na ⁺ -Cl ⁻ in sweat is likely the dominant predisposing factor for hypokalemic alkalosis in patients with CF; patients with this presentation generally respond promptly to intravenous fluids and electrolyte replacement. However, the cystic fibrosis transmembrane conductance regularor (CFTR) protein coassociates with ROMK in the TAL and confers sensitivity to ATP and glibenclamide to apical K ⁺ channels in this nephron segment. Lu and associates have proposed that this interaction serves to modulate the response of ROMK to cAMP and vasopressin, such that K ⁺ excretion in CFTR deficiency would not be appropriately reduced during water diuresis, therefore predisposing these patients to the development of hypokalemic alkalosis.

Gitelman’s Syndrome.

A major advance in the understanding of hereditary alkaloses was the realization that a subset of patients exhibit marked hypocalciuria, rather than the hypercalciuria typically seen in BS; patients in this hypocalciuric subset are universally hypomagnesemic. Such patients are now clinically classified as suffering from Gitelman’s syndrome (GS). Although plasma renin activity may be increased, renal prostaglandin excretion is not elevated in these hypocalciuric patients, another distinguishing feature between BS and GS. GS is a milder disorder than BS; however, patients do report significant morbidity, mostly related to muscular symptoms and fatigue. The QT interval is frequently prolonged in GS, suggesting an increased risk of cardiac arrhythmia ; however, a more exhaustive cardiac evaluation of a large group of patients failed to detect significant abnormalities of cardiac structure or rhythm. However, presyncope and/or ventricular tachycardia has been observed in at least two patients with GS, one with concomitant long QT syndrome due to a mutation in the cardiac KCNQ1 K ⁺ channel.

The hypocalciuria detected in GS was an expected consequence of inactivating the thiazide-sensitive Na ⁺ -Cl ⁻ cotransporter NCC (SLC12A2), and loss-of-function mutations in the human gene have been reported. Many of these mutations lead to a defect in cellular trafficking when introduced into the human NCC protein. GS is genetically homogeneous, except for the occasional patient with mutations in CLC-NKB and an overlapping phenotype. However, genetic analysis is not generally available, given the significant number of exons in SLC12A2 and the absence of hot spot mutations in this disorder. One diagnostic alternative is to assess the physiologic response to thiazides; patients with GS have a blunted excretion of chloride after the administration of hydrochlorothiazide.

The NCC protein has been localized to the apical membrane of epithelial cells in the DCT and connecting segment. A mouse strain with targeted deletion of the Slc12a2 gene encoding NCC exhibits hypocalciuria and hypomagnesemia, with a mild alkalosis and marked increase in circulating aldosterone. These knockout mice exhibit marked morphologic defects in the early DCT, with a reduction in absolute number of DCT cells and changes in ultrastructural appearance. That GS is a disorder of cellular development and/or cellular apoptosis should perhaps not be a surprise, given the observation that thiazide treatment promotes marked apoptosis of this nephron segment. This cellular deficit leads to downregulation of the DCT magnesium channel TRPM6, resulting in the magnesium wasting and hypomagnesemia seen in GS. The downstream CNT tubules are hypertrophied in NCC-deficient mice, reminiscent of the hypertrophic DCT and CNT segments seen in furosemide-treated animals. These CNT cells also exhibit an increased expression of ENaC at their apical membranes versus littermate controls ; this is likely due to activation of SGK1-dependent trafficking of ENaC by the increase in circulating aldosterone (see “ Control of Potassium Secretion: Aldosterone ”).

Hypokalemia does not occur in NCC ^−/− mice on a standard rodent diet but emerges on a K ⁺ -restricted diet; plasma K ⁺ of these mice is about 1 mmol/L lower than K ⁺ -restricted littermate controls. Several mechanisms account for the hypokalemia seen in GS and NCC ^−/− mice. The distal delivery of Na ⁺ and fluid is decreased in NCC ^−/− mice, at least those on a normal diet; however, the increased circulating aldosterone and CNT hypertrophy likely compensate, leading to increased kaliuresis. As discussed earlier for thiazides, decreased luminal Ca ²⁺ in NCC deficiency may augment baseline ENaC activity, further exacerbating the kaliuresis. Of particular interest, NCC-deficient mice develop considerable polydipsia and polyuria on a K ⁺ -restricted diet ; this is reminiscent perhaps of the poly­dipsia that has been implicated in thiazide-associated hyponatremia.

Hypocalciuria in GS is not accompanied by changes in plasma calcium, phosphate, vitamin D, or PTH levels, suggesting a direct effect on renal calcium transport. The late DCT is morphologically intact in NCC-deficient mice, with preserved expression of the epithelial calcium channel (ECAC1, or TRPV5) and the basolateral Na ⁺ -Ca ²⁺ exchanger. Furthermore, the hypocalciuric effect of thiazides persists in mice deficient in TRPV5, arguing against the putative effects of this drug on distal Ca ²⁺ absorption. Rather, several lines of evidence have argued that the hypocalciuria of GS and thiazide treatment is due to increased absorption of Na ⁺ by the proximal tubule, with secondary increases in proximal Ca ²⁺ absorption. Regardless, reminiscent of the clinical effect of thiazides on bone, there are clear differences in bone density between affected and unaffected members of specific Gitelman kindreds. Thus, homozygous patients have much higher bone densities than unaffected wild-type family members, whereas heterozygotes have intermediate values for bone density and calcium excretion. An interesting association has repeatedly been described between chondrocalcinosis, the abnormal deposition of calcium pyrophosphate dihydrate (CPPD) in joint cartilage, and GS. Patients have also been reported with ocular choroidal calcification.

Finally, as in BS, there have been reports of acquired tubular defects that mimic those of GS. These include patients with hypokalemic alkalosis, hypomagnesemia, and hypocalciuria after chemotherapy with cisplatin. Patients have also been described with acquired GS due to Sjögren’s syndrome and tubulointerstitial nephritis, with a documented absence of coding sequence mutations in NCC.

Renal Tubular Acidosis

Renal tubular acidosis (RTA) and related tubular defects can be associated with hypokalemia. Proximal RTA is characterized by a reduction in proximal bicarbonate absorption, with a reduced plasma bicarbonate concentration. Isolated proximal RTA is rare; genetic causes include loss of function by mutations in the basolateral Na ⁺ -HCO ₃ ⁻ transporter. More commonly, proximal RTA occurs in the context of multiple proximal tubular transport defects, encompassing Fanconi’s syndrome (FS). The cardinal features of FS include hyperaminoaciduria, glycosuria with a normal plasma glucose concentration, and phosphate wasting; associated defects include the proximal RTA, hypouricemia, hypercalciuria, hypokalemia, salt wasting, and increased excretion of low-molecular-weight proteins. FS is usually drug-associated; important causes include aristolochic acid, ifosfamide, and the acyclic nucleoside phosphonates (e.g., tenofovir, cidofovir, adefovir). Prior to treatment with bicarbonate, patients with a proximal RTA will typically demonstrate mild hypokalemia, due primarily to baseline hyperaldosteronism ; however, prior to treatment, patients have been described with profound hypokalemia on presentation. Regardless, treatment with oral sodium bicarbonate will markedly increase distal tubular Na ⁺ and HCO ₃ ⁻ delivery, causing a marked increase in renal potassium wasting. Patients will often require mixed base replacement with oral citrate and bicarbonate in addition to aggressive K ⁺ -Cl ⁻ supplementation.

Hypokalemia is also associated with distal RTA, the so-called type 1 RTA. Hypokalemic distal RTA is most commonly due to a secretory defect, with reduced H ⁺ -ATPase activity and decreased ability to acidify the urine. For example, hereditary defects in subunits of H ⁺ -ATPase are associated with profound hypokalemia in addition to acidosis and hypercalciuria. Pathophysiology of the associated hypokalemia is multifactorial due to the loss of electrogenic H ⁺ secretion (with enhanced K ⁺ secretion to maintain electroneutrality in the distal nephron), loss of H ⁺ -K ⁺ -ATPase activity, and increases in aldosterone. Sjögren’s syndrome is perhaps the most common cause of hypokalemic distal RTA in adults; the associated hypokalemia can be truly profound, often resulting in marked weakness and respiratory arrest.

Magnesium Deficiency

Magnesium deficiency results in refractory hypokalemia, particularly if the plasma Mg ²⁺ level is less than 0.5 mmol/L ; hypomagnesemic patients are thus refractory to K ⁺ replacement in the absence of Mg ²⁺ repletion. Magnesium deficiency is also a common concomitant of hypokalemia, in part because associated tubular disorders (e.g., aminoglycoside nephrotoxicity) may cause kaliuresis and magnesium wasting. Plasma Mg ²⁺ levels must therefore be checked on a routine basis in hypokalemia.

Several mechanisms appear to contribute to the effect of magnesium depletion on plasma K ⁺ . Magnesium depletion has inhibitory effects on muscle Na ⁺ -K ⁺ -ATPase activity, resulting in significant efflux from muscle and a secondary kaliuresis. Distal K ⁺ secretion also appears to be enhanced due to a reduction in the normal physiologic inward rectification of ROMK secretory K ⁺ channels, with a subsequent increase in outward conductance. ROMK and other Kir channels are inward-rectifying—that is, K ⁺ flows inward more readily than outward; even though outward conductance is usually less than inward conductance, K ⁺ efflux predominates in the CNT and CCD since the membrane potential is more positive than the equilibrium potential for K ⁺ . Intracellular Mg ²⁺ plays a key role in inward rectification, binding and blocking the pore of the channel from the cytoplasmic side. The hypomagnesemia-associated reduction in cytoplasmic Mg ²⁺ in principal cells reduces inward rectification of ROMK, increasing outward conductance and increasing K ⁺ secretion; this has been confirmed in vivo. Finally, it has been suggested that the repletion of intracellular K ⁺ is impaired in hypomagnesemia, even in normokalemic patients. Decreased intracellular Mg ²⁺ enhances K ⁺ efflux from the cytoplasm of cardiac and perhaps skeletal myocytes, likely due to reduced intracellular blockade of inward rectifying K ⁺ channels (increased efflux) and inhibition of Na ⁺ -K ⁺ -ATPase (decreased influx); plasma K ⁺ levels thus remain normal at the expense of intracellular K ⁺ . This phenomenon is particularly important in patients with cardiac disease who are taking both diuretics and digoxin. In such patients, hypokalemia and arrhythmias will respond to a correction of the magnesium deficiency and potassium supplementation.