The ability of insulin to shift potassium intracellularly has been known for over 70 years⁹ and has been used therapeutically for the treatment of hyperkalemia. Pancreatectomized dogs tolerate exogenous potassium loads poorly.¹⁰ This is reversed by the exogenous replacement of insulin.¹¹,¹² The partial inhibition of endogenous insulin in dogs by somatostatin infusion results in a twofold rise in serum potassium compared to controls.⁶ If physiologic doses of insulin were added to the somatostatin infusion, potassium tolerance returned to normal. In healthy volunteers, somatostatin infusion in the postabsorptive state led to a 50% decline in the plasma insulin concentration and a 0.5 to 0.7 mEq per liter rise in serum potassium that was reversed by a physiologic infusion of exogenous insulin.⁶ A similar phenomenon was observed in maturity-onset diabetic patients who have normal or increased fasting plasma insulin levels, but not in insulin-deficient juvenile diabetic patients.¹³

The primary sites of insulin-mediated potassium uptake include muscle and the liver, and to a lesser degree, adipose tissue.¹⁴,¹⁵ In normal volunteers on variable insulin doses, the liver is the primary site of potassium uptake during the first hour.¹⁵ However, during the second hour, despite a continued decrease in serum potassium, there is net release of potassium from the portal and splanchnic bed, indicating a shift of potassium uptake to the peripheral tissue, especially muscle.¹⁵

At the cellular level, insulin interacts with specific receptors on the plasma membrane,¹⁶ increasing the activity
of the Na-K-ATPase pump in the skeletal and heart muscle, epithelial cells of the kidney and bladder, as well as liver and fat cells.¹⁷ This results in a series of intracellular events leading to hyperpolarization of cell membranes.¹⁷ The time course for this interaction is consistent with both an increase in enzyme activity as well as the rapid recruitment of Na-K-ATPase pumps to the cellular membrane. In contrast, chronic stimulation by insulin probably increases the total number of available pump sites. This occurs through the regulation of the Na-K-ATPase pump at the transcriptional and posttranscriptional levels by inducing the synthesis of new α and β subunits.¹ McDonough and Youn,¹⁸ using a potassium clamp, have recently shown that after 10 days of potassium deprivation in rats Na-K-ATPase activity decreased by more than 50% and insulin-mediated potassium shift decreased by 94%, whereas in rats deprived of potassium for only 2 days the number of pumps did not decrease, but insulin-mediated potassium shift decreased by 80%. This would indicate that insulin resistance precedes a decrease in the number of pump expression during hypokalemia. The molecular mechanism underlying this response, however, remains poorly understood.¹⁹ Several in vitro studies, including one study in humans, have shown that insulin-driven potassium uptake by both muscle and the liver is independent of glucose uptake.¹⁵,²⁰

D’Silva,²¹ beginning in 1934, first observed a biphasic response of plasma potassium to epinephrine injection. Plasma potassium rose during the first 1 to 3 minutes, but
with continued infusion, fell and remained lower than baseline. Other investigators have shown increased potassium tolerance in animals infused with pharmacologic doses of epinephrine²²,²³ despite a pancreatectomy or nephrectomy.²⁴ Brown and coworkers²⁵ have shown that the infusion of stress-level doses of epinephrine resulted in a decrease in serum potassium by 0.4 to 0.6 mEq per liter. Because epinephrine inhibits the renal excretion of potassium,²⁶,²⁷ the decline in potassium concentration is entirely accounted for by enhanced cellular potassium uptake.

Specific receptors are involved in the cellular disposal of potassium by catecholamines. Alpha stimulation in humans by phenylephrine²⁸ significantly impairs cellular potassium tolerance, which is reversed by the α-antagonist phentolamine. This phenomenon may explain the initial rise in serum potassium after the infusion of catecholamine.²⁶,²⁷ β₂-blockade impairs the catecholamine-induced shift of potassium into extrarenal tissues²⁹,³⁰ and causes hyperkalemia despite an increase in renal excretion of this ion. In normal volunteers who exercise while taking β-adrenergic blocking agents, the serum potassium level is raised 2- to 2.5-fold higher than during similar exercise performed without a β blockade.³,³¹ The effect of nonspecific β-blockers such as propranolol on serum potassium is mimicked by specific β₂-blockers³² but not β₁-blockers. Although an important role for catecholamine-stimulated uptake of potassium by muscle has been demonstrated, the role of the liver remains controversial. The effect of potassium on catecholamine levels is less clear.

At the cellular level, epinephrine binds to the β₂-receptor resulting in the stimulation of adenyl cyclase and the conversion of adenosine triphosphate to cyclic 3′,5′-adenosine mono- phosphate (cAMP). It is postulated that cAMP then activates protein kinase A, which then phosphorylates the Na-K-ATPase pump, increasing its activity and promoting potassium influx into the cell and Na⁺ efflux.³¹ Binding catecholamines to the α receptor decreases cellular potassium uptake by inhibiting adenylate cyclase activity and decreasing Na-K-ATPase pump activity.³² In addition, activation of the α-1 receptor alters cytoplasmic calcium, thereby increasing intracellular calcium concentration and opening calcium-activated potassium channels, which allow potassium to exit the cell.³² Interestingly, the effect of insulin and epinephrine on plasma potassium is additive, which confirms a separate mechanisms of action.³¹ In insulin-induced hypoglycemia, hypokalemia is therefore due to the combined effect of both insulin and the hypoglycemia-induced rise in catecholamines.³¹

Mineralocorticoids play a major role in external potassium homeostasis by increasing its excretion by the kidney,³³ colon,³⁴ salivary,³⁵ and sweat glands.³⁶ However, aldosterone’s role in internal potassium homeostasis is less clear.³⁷,³⁸ Anephric rats adapted to high potassium intake handle an acute potassium load more efficiently than do nonadapted rats. This adaptation is lost by prior adrenalectomy and restored by exogenous mineralocorticoid replacement.³⁹ However, Spital and Sterns⁴⁰,⁴¹ observed that during the 20 hours of fasting before a nephrectomy and acute potassium loading, these rats became potassium depleted owing to marked kaliuresis resulting from high serum potassium coupled with a high aldosterone level. In adrenalectomized dogs, Young and Jackson⁴² have shown that plasma potassium concentration at any exchangeable potassium level was a function of aldosterone replacement dose. High-dose aldosterone in anephric rabbits delays death due to hyperkalemia.⁴³ Similarly, baseline potassium was significantly higher in hormonally deficient adrenalectomized rats despite negative potassium balance compared to exogenously replaced controls, thus supporting a defect in the cellular uptake of potassium.⁵,⁴⁴ This impairment was corrected by either aldosterone or epinephrine replacement. In rats, aldosterone has been shown to increase Na-K-ATPase pump activity by inducing the synthesis of new α- and β-subunits in heart and vascular smooth muscle.¹ This effect presumably represents the action of aldosterone on Na-K-ATPase pump gene expression and supports a role for aldosterone in cellular potassium homeostasis. In anephric humans treated with deoxycorticosterone acetate (DOCA), spironolactone, or placebo for 3 days, the baseline potassium was similar; however, the DOCA-treated subjects showed greater tolerance to acute potassium load than did the other two groups.⁴⁵ In a study of 15 patients on hemodialysis that were treated with 0.05 to 2.0 mg per day of fludrocortisone acetate, the serum K⁺ decreased significantly.⁴⁶ Interestingly, the effect of exogenous mineralocorticoid was more pronounced in patients with a low compared to a high plasma aldosterone concentration. Low dose spironolactone (25 mg per day) was associated with an increase in a mean serum K⁺ concentration of 0.3 mEq per liter over 4 weeks of therapy in 15 chronic hemodialysis patients.⁴⁷ In the largest study to date, serum potassium in 50 hemodialysis patients treated with 25 mg per day of spironolactone increased from baseline 4.96 to 5.16 in 2 weeks and remained stable for 6 months.⁴⁸ Very low dose spironolactone (25 mg thrice weekly), however, did not increase serum K⁺ in hemodialysis patients,⁴⁹ whereas a very high dose (300 mg per day) induced a significant rise in plasma potassium (0.5 mEq per liter) and caused hyperkalemia after 3 weeks of therapy in nine chronically hemodialyzed end-stage renal disease (ESRD) patients (three were anephric).⁵⁰ In summary, these studies support a small but significant role for aldosterone in internal potassium homeostasis in anephric animals and ESRD patients.

The role of acid-base balance on the internal distribution of potassium⁵¹ is based on the concept that during the development of acute acidemia, the hydrogen ion enters the cell in exchange for potassium and that the reverse occurs during the development of alkalemia.⁵¹,⁵²,⁵³ This dynamic interrelationship has been simplified clinically to a general rule that
for each 0.1 U change in serum pH, the serum potassium changes in the opposite direction by 0.6 mEq per liter. However, the relationship between serum potassium and serum pH is much more complex and depends on the type and severity of the acid-base disorder, the anion accompanying hydrogen, the duration of acidosis, changes in plasma bicarbonate concentration independent of changes in pH and the extent of intracellular buffering, and renal adaptation as well as hormonal changes in response to the disorder.⁵⁴ In addition, in clinical settings, there are often other physiologic and pathophysiologic processes that may be present, which would affect both transcellular as well as the renal and extrarenal handling of potassium. The following generalizations should therefore be used with caution.

1. On the whole, acidosis is accompanied by a greater change in serum potassium than is alkalosis.⁵⁵

2. Mineral acidosis (Fig. 72.3) causes the greatest shift (0.24 to 1.7 mEq per liter for each 0.1 U in pH change), whereas organic acidosis has a much smaller effect.⁵³,⁵⁶,⁵⁷ Mild mineral acidosis (a decrease in serum bicarbonate by 5 mEq per liter and an increase in hydrogen ion concentration by 0.45 nmol per liter), however, does not result in a significant change in serum potassium.⁵⁸

3. Acute respiratory alkalosis paradoxically results in a small but significant rise in serum potassium (+0.30 mEq per liter with a drop in pCO₂ of 16 to 22.5 mm Hg). The rise was primarily due to stimulation of α-adrenergic receptors by catecholamine.⁵⁹ Chronic respiratory alkalosis, however, results in sustained hypokalemia due to a renal loss of potassium.⁶⁰

4. The amounts of potassium shifted into the cell in metabolic and chronic respiratory alkalosis are approximately similar (0.1 to 0.4 mEq per liter for each 0.1 U of pH change).

5. Acute respiratory acidosis resulting in a decrease in pH to 7.24 had no effect on serum K⁺.⁶¹

6. Changes in serum bicarbonate, independent of serum pH, have an inverse effect on the serum potassium concentration.

7. In chronic acidosis and alkalosis, the final serum K⁺ is a function of the effect of acid-base disturbance on the renal handling of potassium, as well as on the transcellular distribution of this ion. In dogs with ammonium chloride-induced acidosis, Magner and associates⁶² noted a fall in serum potassium below baseline by days 3 to 5, owing to severe kaliuresis.

The acute hyperkalemic effect of a sudden rise in plasma osmolality is probably caused by the shift of potassium-rich intracellular fluid by solvent drag.⁶³ Clinically, this phenomenon is most commonly observed in hyperosmolar diabetic patients (Fig. 72.4), with or without ketoacidosis⁶⁴,⁶⁵,⁶⁶,⁶⁷ when insulin deficiency augments the rise in potassium. Although chronic hyperkalemia in diabetic patients is multifactorial, a sudden rise in plasma osmolality seems to play a contributory role. The infusion of hypertonic mannitol in healthy humans⁶⁸ or hypertonic saline⁶⁹ or hypertonic contrast media⁷⁰ in patients with chronic kidney disease results in a modest rise in serum potassium (0.4 to 0.6 mEq per liter). Hyperkalemia can be severe, especially in diabetic patients with little or no
renal function facing sudden hyperglycemia.⁷¹ These clinical observations support an independent role of sudden osmolar shifts in the regulation of serum potassium.

Feedback or Feedforward Control of Potassium Homeostasis. It is well known that an increase in potassium concentration directly stimulates renal potassium excretion through an increase in potassium secretion in the collecting duct. This is accomplished by the direct stimulation of Na-K-ATPase, an increased tubular flow, and an increase in aldosterone. However, as Rabinowitz et al.⁷² first noted an increase in renal potassium excretion after meals in sheep was independent of change in serum potassium and aldosterone. In normal human subjects, urinary potassium excretion increased significantly 20 minutes after the ingestion of potassium salts before any change in serum potassium. Kaliuresis was more robust if potassium is ingested with meals rather than without meals or given intravenously. These and other observations support a role for a direct gut-kidney axis in potassium homeostasis favoring a feedforward rather than a feedback homeostatic mechanism (Fig. 72.5). The specific gut sensor and the gut-kidney loop remains speculative at this point. For a more detailed discussion, readers are referred to two recent reviews of this topic.⁷³,⁷⁴

Patients with a marked decrease in GFR are able to excrete the ingested dietary potassium load and maintain near normal potassium balance. This adaptive process is reflected by an increase in the fractional excretion of potassium (FE_K) modulated by an increase in secretory rate per functioning nephron. However, this adaptive response is limited and a sudden increase in potassium intake may result in life-threatening hyperkalemia. The quantitative aspects as well as the anatomic and functional characteristics of this adaptive response are briefly reviewed herein.

In conscious dogs with a 10% remnant kidney, Schultze and coworkers¹¹⁹ showed that potassium excretion by the remnant kidney increased fourfold by 18 hours and approached 85% of the control value by the 7th day. Kunau and Whinnery¹²⁰ and Wilson and Sonnenberg¹⁰³ reported similar data in rats. In experiments by Schultze and associates,¹¹⁹ animals with a remnant kidney manifested an exaggerated kaliuresis following a potassium load. In contrast to these data and independent of previous potassium intake, dogs with 25% remnant kidney were only able to excrete 30% to 37% of the load in 5 hours compared with 70% to 90% in the control animals.¹⁰² There is no easy resolution to the differences in these two studies.¹⁰²,¹¹⁹

Gonick and colleagues⁷⁶ documented that human subjects with CKD were able to excrete only 20% of an oral potassium load in 6 hours compared with 46% in normal controls. Similar data were reported by Perez and colleagues¹⁰⁰ in patients with tubulointerstitial disease. Kahn and colleagues⁷⁷ demonstrated in 10 patients with stable chronic kidney disease renal adaptation to increased dietary potassium. In summary, it can be concluded that residual renal tissue is able to maintain external potassium homeostasis in the postabsorptive state. However, the initial phase of this adaptation is impaired when an acute potassium load is administered.

The nephron sites involved in this adaptation have been studied using a variety of techniques in both rats and rabbits and appear to include both the distal convoluted tubule and the collecting duct.¹⁰³,¹¹⁹,¹²⁰,¹²¹,¹²²,¹²³ The discrepancies reported in the literature most likely owe to interspecies and intraspecies differences as well as the anatomic definition of different distal tubular segments.

The mechanisms involved in this renal adaptation have been partially defined. In both humans¹¹⁵ and rodents,¹²⁴ aldosterone has been shown to play an important role in the adaptive ability of the diseased kidney to maintain a normal rate of potassium excretion. This renal adaptation has been shown to be independent of dietary sodium intake.¹²⁵ Schultze and coworkers¹¹⁹ argued that aldosterone is not important in the renal potassium adaptation that occurs following a reduction in renal mass, because uremic dogs maintained on constant aldosterone replacement maintained normal rates of potassium excretion. However, the replacement dose of aldosterone in this study was in the high pharmacologic range. Serum potassium concentration itself plays an important role in augmenting urinary potassium excretion.⁸¹ Bourgoignie and colleagues¹⁰² found a direct relationship between serum potassium and both the absolute and fractional potassium excretion (EE_k). The slope of the curve relating serum potassium to the absolute rate of urinary potassium excretion was much steeper in normal dogs than in dogs with a remnant kidney. However, the slope of the curve relating serum potassium to the FE_K was similar in the control and uremic dogs.

Microperfusion studies by Fine and associates¹²² indicate that adaptation is an inherent characteristic of the renal tubular cells of uremic animals and, once learned, it can be retained in vitro, at least for short periods of time. Schon and associates¹⁰³ showed that augmented potassium excretion is associated with an increase in Na-K-ATPase in the outer medulla in animals subjected to a three-quarter nephrectomy. This increase is quite specific to this enzyme and occurs only in the kidney¹⁰³ and the colon.¹²⁶ Muto and colleagues¹²⁷ demonstrated that an increase in peritubular [K⁺] increased renal potassium excretion by also enhancing K⁺ conductance (ROMK) and Na⁺ conductance (ENaC) in principal cells (Fig. 72.6). Other mechanisms may include a higher rate of potassium delivery and an increase in tubular flow rate in the distal nephron.¹²⁰

Patients with renal failure secrete more potassium in the stool than do normal controls.¹¹⁵,¹²⁸,¹²⁹ Net colonic secretion of potassium is increased significantly above control levels in rats with renal insufficiency.¹²⁸ This increase is associated with an increase in Na-K-ATPase activity in colonic mucosa and is functionally similar to the increase seen with the administration of DOCA, glucocorticoids, or high dietary potassium.¹³⁰ Although the rise in fecal potassium concentration is significant, the absolute amount of K⁺ lost through this route in patients with mild-to-moderate CKD is small and contributes only minimally to the external K⁺ homeostasis. In patients with advanced renal insufficiency (GFR < 5 to 10 mL per minute), however, up to 30% to 40% of the ingested potassium load may be excreted in the stool.¹²⁹

Several studies demonstrate that some patients with severe kidney disease have significant bicarbonate wasting.¹³⁵,¹⁴⁰,¹⁴¹,¹⁴²,¹⁴³,¹⁴⁴ In an early study by Schwartz and coworkers,¹³⁵ three out of four patients with renal failure had significant bicarbonaturia, which disappeared only after the fall of serum bicarbonate to below 20 mEq per liter. In a more detailed study in 17 uremic patients (serum creatinine of 5.6 to 18.9 mg per deciliter), the majority had significant bicarbonate wasting (fractional excretion of HCO₃ of 0% to 17.56%) despite the presence of metabolic acidosis (serum HCO₃ of 16 to 23 mEq per liter). After NH₄Cl loading, serum bicarbonate decreased to below 14 mEq per liter, and bicarbonaturia disappeared in all but four patients.¹⁴⁴ Interestingly, the bicarbonate wasting in these four patients also disappeared with the institution of a low-sodium diet.¹⁴⁴ These two studies support the presence of a diminished maximal tubular reabsorption (T_m) for bicarbonate in the majority of patients with renal failure. Further, they demonstrate that the low T_m is partly responsive to volume status.

Arruda and colleagues¹⁴³ and Wong and associates,¹⁴⁵ working with a remnant kidney model in dogs with variable levels of volume expansion and serum bicarbonate, noted that the ratio of absolute bicarbonate to sodium reabsorption was increased in CKD. In addition, Wong and associates,¹⁴⁵ using a micropuncture method, showed that this ratio was also higher at the beginning of the distal tubule, indicating avid bicarbonate absorption by the proximal tubule of the remnant kidney. Although absolute absorption was higher, the absolute amount of bicarbonate delivered to the distal tubule was also higher, reflecting the marked increase in filtered load per nephron owing to an increase in single nephron GFR.¹⁴⁵ In summary, the whole kidney T_m for bicarbonate is, in general, diminished in CKD despite an absolute increase in bicarbonate resorption at the single nephron. The discrepancy in these findings may reflect the variation in the experimental designs and the role of nonvolume regulators in bicarbonate handling by the kidney.

The excretion of titratable acids chiefly reflects the amount of urinary phosphate and the urinary pH. Most CKD patients are able to maximally acidify their urine,¹³⁵,¹⁴⁶ and urine-serum PCO₂, as a measure of hydrogen pump activity in the distal tubule, is normal.¹⁴⁷ The amount of titratable acids in these patients is normal.¹³⁷,¹³⁸,¹³⁹,¹⁴⁰,¹⁴¹,¹⁴²,¹⁴³,¹⁴⁴,¹⁴⁵,¹⁴⁶,¹⁴⁷,¹⁴⁸,¹⁴⁹,¹⁵⁰ This is primarily owing to an increase in the fractional excretion of phosphate initiated by secondary hyperparathyroidism. It should be noted, however, that urinary phosphate does decrease with severe renal failure. This reflects both a decrease in dietary phosphate as well as the effect of phosphate binders commonly used in these patients.

Although bicarbonaturia may contribute to metabolic acidosis, the major abnormality is a decrease in renal excretion of ammonium. Ammonium is primarily produced by the deamination of amino acids, chiefly glutamine, in the proximal tubule and, to a much lesser extent, in the loop of Henle and the distal convoluted tubule.¹³⁰,¹³¹ This is reviewed in detail in Chapter 7, Renal Acid-Base Transport and will not be reviewed here. In CKD, fractional renal ammonium excretion initially increases by severalfold, thereby resulting in the maintenance of a normal absolute excretion rate.¹⁵¹ However, as the GFR decreases below 20 mL per minute, despite a maximal increase in fractional excretion of ammonium, the absolute excretory rate decreases significantly. Thus, progressive metabolic acidosis results. This decrease in the rate of ammonium excretion also reflects a decreased ability of the kidney to trap ammonia in the collecting duct.¹⁴⁶ Warnock¹³⁹ has suggested that the decrease in ammonia trapping in the remnant kidney model may be secondary to excess delivery of bicarbonate to the collecting duct, thereby resulting in an unfavorable environment for the diffusion and trapping of ammonia.

The role of aldosterone in ammonium excretion is complex. Aldosterone increases the rate of Na⁺-dependent and Na⁺-independent H⁺ secretion in the cortical and medullary collecting duct.¹⁵²,¹⁵³ Hypoaldosteronism is associated with a decrease in the rate of H⁺ secretion, whereas the ability to maintain a steep H⁺ gradient between urine and plasma, as measured by urinary pH and urine minus blood PCO₂ in alkaline urine, is not affected.¹⁵⁴,¹⁵⁵ The decrease in the rate of H⁺ secretion is associated with a decrease in the availability of ammonium buffer in the urine that is not augmented appropriately in response to sodium sulfate infusion.¹⁵⁶,¹⁵⁷ Hypoaldosteronism is universally associated with a decreased potassium excretion and hyperkalemia. Hyperkalemia decreases renal ammonium excretion significantly. A decrease in accumulation of ammonium in the renal interstitium despite normal production by the proximal tubule underlies this effect.¹⁵⁸ In the syndrome of hyperkalemic renal tubular acidosis, this mechanism probably plays the major role in the production of hyperchloremic acidosis seen early in the course of renal failure (Fig. 72.7).¹⁵⁹ Reversal of hyperkalemia with sodium binding resin,¹⁶⁰ mineralocorticoids,¹⁶¹ or low-potassium diet¹⁶² ameliorates the metabolic acidosis by increasing ammonium secretion.

In summary, the metabolic acidosis develops universally in all patients with CKD as GFR decreases to below 20 mL per minute. The pathogenesis of this disorder is complex and reflects renal defect in both resorption as well as the generation of bicarbonate. The major mechanism, however, is in a decrease in absolute ammonia excretion despite the presence of acidosis.

At present, no single abnormality can explain the low PRA seen in 80% of these patients.¹⁷¹,¹⁸²,¹⁸³ Evidence has been presented in support of a defect in one or more physiologic regulators of renin secretion including volume, autonomic nervous system, serum potassium concentration, and prostaglandins.

Oh and colleagues,¹⁶⁹ Perez and colleagues,¹⁸⁵ and others¹⁸⁶,¹⁸⁷ have demonstrated that long-term sodium and volume depletion in these patients is associated with a significant increase in the PRA. However, comparable data in normal controls with the same degree of volume depletion were not provided. In the report of Oh and colleagues,¹⁶⁹ after 3 to 6 weeks of salt depletion, the PRA rose into the normal range, but plasma aldosterone remained subnormal. In the study by Chan and coworkers, 8 of the 12 patients with hyporreninism responded to 2 weeks of furosemide with an increase in PRA without a similar response in plasma aldosterone.¹⁸⁷ In a study of four patients with acute postinfectious glomerulonephritis,¹⁵⁵ plasma renin and aldosterone concentrations were low during the acute phase, but returned to normal following recovery from acute nephritis. Interestingly, in two patients, the renin and aldosterone levels remained low during the acute phase despite an excellent response to diuretics. These two patients, however, responded appropriately to
physiologic doses of fludrocortisone. This study,¹⁵⁵

Stay updated, free articles. Join our Telegram channel

Join