American Journal of Kidney Diseases

IF The National Kidney Foundation The Official Journal of

VOL XVIII, NO 4, OCTOBER 1991

IN-DEPTH REVIEW

Extrarenal Potassium Tolerance in Chronic Renal Failure: Implications for the Treatment of Acute Hyperkalemia Mohamed M. Salem, MD, Robert M. Rosa, MD, and Daniel C. Batlle, MD • The role of extrarenal potassium homeostasis is well recognized as a major mechanism for the acute defense against the development of hyperkalemia. The purpose of this report is to examine whether or not the various mechanisms of extrarenal potassium regulation are intact in patients with end-stage renal disease (ESRD). The available data suggest that with the development of ESRD and the uremic syndrome there is impaired extrarenal potassium metabolism that is related to a defect in the Na,K-adenosine triphosphatase (ATPase). The responsiveness of uremic patients to the various effector systems that regulate extrarenal potassium handling Is discussed. Insulin is well pOSitioned to play an important role in the regulation of plasma potassium concentration in patients with impaired renal function. The role of basal Insulin may be even more important than previously appreciated, since somatostatin infusion causes a much greater increase in the fasting plasma potassium in rats with renal failure than in controls. Furthermore, stimulation of endogenous insulin by oral glucose results in a greater intracellular translocation of potassium in uremic rats than in controls. Under at least two common physiologic circumstances, feeding and vigorous exercise, endogenous catecholamines might also act to defend against acute Increments in extracellular potassium concentration. However, it is important to appreciate that the response to P2-adrenoreceptor-mediated internal potassium disposal is heterogeneous as judged by the variable responses to epinephrine infusion. Based on the evidence presented in this report, a regimen for the treatment of life-threatening hyperkalemia is outlined. Interpretation of the available data demonstrate that bicarbonate should not be relied on as the sole initial treatment for severe hyperkalemia, since the magnitude of the effect of bicarbonate on potassium is variable and may be delayed. The initial treatment for lifethreatening hyperkalemia should always include insulin plus glucose, as the hypokalemic response to insulin is both prompt and predictable. Combined treatment with .B2-agonists and insulin is also effective and may help prevent insulininduced hypoglycemia. © 1991 by the National Kidney Foundation,lnc. INDEX WORDS: Hyperkalemia; bicarbonate; end-stage renal disease; extrarenal potassium homeostasis.

T wo

MAJOR SYSTEMS maintain plasma potassium within the narrow normal range: a renal system excreting the daily potassium load, and an extrarenal system involving mechanisms that act to translocate potassium intracellularly. The integrity of both of these systems is required to maintain normokalemia. The extrarenal system is critical to the defense against acute hyperkalemia, because renal excretion of a potassium load is a relatively slow process. During the first 3 to 6 hours following the ingestion of a potassium load, only about halfofthe administered load appears in urine. I Thus, an average daily potassium load (l00 mmol/d [mEq/d)), if given acutely, would result in frank hyperkalemia were

it not for extrarenal potassium homeostatic mechanisms that rapidly translocate administered potassium into the intracellular compartment. The extrarenal mechanisms that govern internal potassium distribution include hormones such as insulin, epinephrine, and, possibly, aldosterone. In addition, acid base status and

From the Department ofMedicine and the Division ofNephrology, Northwestern University Medical School and Veterans Administration Lakeside Medical Center, Chicago, lL. Address reprint requests to Daniel C. Batlle, MD, Northwestern University, 303 E Chicago Ave, Chicago, lL 60611. © 1991 by the National Kidney Foundation, 1nc. 0272-6386/91/1804-0001$3.00/0

American Journal of Kidney Diseases, Vol XVIII, No 4 (October), 1991: pp 421-440

421

422

SALEM, ROSA, AND BATLLE

plasma osmolality influence internal potassium balance. Also, the gastrointestinal tract, which normally contributes little to potassium excretion, assumes a greater role in the maintenance of potassium homeostasis in individuals with chronic renal failure. The purpose of this report is to examine whether or not the various mechanisms of extrarenal (internal) potassium regulation are intact in patients with end-stage renal disease (ESRD). The integrity of this system has implications for the management of severe hyperkalemia in such patients, which have not been well emphasized. The relative efficacy of the various modalities usually used to treat severe hyperkalemia will also be reviewed. POTASSIUM ADAPTATION

Animals fed a high potassium diet can survive acute oral loads that produce lethal hyperkalemia in nonadapted animals. This phenomenon is referred to as potassium tolerance, a term introduced by Thatcher and Radike 2 in 1947. Alexander et aJ3 first provided evidence for the existence of extrarenal potassium adaptation. Animals fed or challenged chronically with high potassium intake not only were able to excrete the load more readily, but also were able to translocate the potassium load intracellulary more promptly. This was demonstrated in a series of experiments showing that the increase in plasma potassium caused by an acute potassium load was less pronounced in potassium-adapted rats than in control animals. After both groups of animals had undergone nephrectomy, challenge with an acute potassium load resulted in an increase in plasma potassium concentration that was higher in the nonadapted animals as compared with the adapted ones. This response could be abolished by prior adrenalectomy, a finding that was originally interpreted as evidence for an extrarenal effect of aldosterone on internal potassium distribution. However, subsequent work clearly established that potassium intolerance following adrenalectomy was largely due to a lack of epinephrine production by the adrenal glands. 4 While the immediate mechanisms for defense against acute hyperkalemia rely on extrarenal mechanisms, a renal adaptive response also takes place. The kidneys efficiently alter their rate of potassium excretion in response to variations in

dietary potassium intake. The ability of the kidneys to excrete and dispose of a potassium load is enhanced in animals 5 and humans6 chronically challenged with increased potassium intake. These observations support the concept of renal adaptation as a mechanism that contributes to potassium tolerance. Renal adaptation for potassium is a segment specific process in that it is confined to discrete segments within the nephron. Major adaptive changes occur in the late distal tubule and cortical collecting tubule. Morphological changes at the level of the collecting tubules are found in potassium-adapted animals. 7 These changes include increased basolateral membrane surface area in this nephron segment that may be cell-specific. Namely, the surface area of cells that secrete potassium (principal cells) increases, while the surface area of cells that secrete acid (intercalated cells) does not. 8 ,9 The factors that influence renal potassium excretion are multiple. While a significant increase in plasma potassium has been difficult to demonstrate in most animal studies after chronic potassium loads, a recent study in humans has shown that plasma potassium increases, albeit transiently, following an oral load. 6 Previous studies in the rat failed to show an increase in plasma potassium after large dietary potassium loads, so that an increase in plasma potassium, a presumed signal for renal adaptation, remained elusive. An increase in plasma potassium would be expected to result in enhanced peritubular renal cell uptake of potassium, as a consequence of a direct stimulatory effect on Na,K-adenosine triphosphatase (ATPase) activity. An increase in cell potassium should, in turn, facilitate its passive secretion into the tubular urine. Aldosterone production by the adrenal gland increases in response to increased plasma potassium concentration. The cortical collecting duct appears to be the primary mineralocorticoid target segment for aldosterone action and potassium secretion. Aldosterone augments potassium secretion by stimulating Na,K-ATPase in the basolateral membrane, by increasing the transepithelial voltage gradient, and possibly by altering the potassium conductance of the luminal membrane. 10-12 Although the functional renal adaptive changes can be demonstrated in adrenalectomized animals, a full adaptive response requires the presence of aldosterone. 13

HYPERKALEMIA AND CHRONIC RENAL FAILURE

The mechanism of potassium adaptation in renal insufficiency appears to be similar to the process that occurs in subjects with intact renal function given a high potassium diet. 14-19 Adaptation was demonstrated as early as 24 hours following partial renal ablation in both rats l4 and dogs. 15 Micropuncture studies in rats with a remnant kidney suggest that net potassium secretion occurs between the end of the distal tubule and the end of collecting tubule.20.21 Thus, the site of adaptive cellular changes in renal insufficiency, as in animals with normal kidneys that are potassium-adapted, is the collecting tubule. In the renal ablation model of renal insufficiency there is increased potassium excretion by the remaining nephrons, such that a sixfold increase in potassium secretion was observed in isolated perfused collecting tubules from uremic rabbits maintained on a high potassium diet. 22 The renal adaptation to an increased potassium load that has been shown to occur in animals following unilateral nephrectomy is associated with an increase in both the number and the activity of the sodium potassium pump in renal cells. 23 ,24 This increased activity is not a nonspecific event secondary to the hypertrophic changes taking place since, Na,K-A TPase site density is increased out of proportion to the increase in basolateral membrane surface area. 23 The renal medullary structures may also contribute to renal potassium adaptation. The fate of filtered potassium in the juxtamedullary nephrons, which are not accessible to micropuncture, may be different from that in the more superficial nephrons. The intrinsic capacity for potassium secretion is increased in deeper nephrons. Increased delivery of potassium to the juxtamedullary nephrons may also contribute to the increased excretion of potassium by these nephrons. Additionally, potassium recycling allows for the excretion of any potassium that might have diffused back from the cortical collecting duct into the interstitial medulla, thus increasing the efficiency of the kidney to excrete potassium. 25 Potential clinical correlates are that patients with preferential papillary damage, such as occurs in sickle cell anemia, may be unable to excrete a potassium load normally26 and that the occurrence of frank hyperkalemia has been described in patients with sickle cell hemoglobinopathy. 27 Likewise, obstructive nephropathy, a condition

423

with preferential damage to the deep nephrons, is associated with hyperkalemia disproportionate to any decline in renal function. 28 The extracellular potassium concentration is maintained in the normal range until late in the course of renal insufficiency, because renal adaptive changes permit the excretion of the normal daily load and, therefore, the maintenance of normokalemia despite reduced glomerular filtration rate (GFR).29,30 As indicated above, the adaptive response that prevents the development of hyperkalemia despite reduced GFR is an increase in the fractional excretion ofpotassium. 28 The renal adaptive changes in potassium excretion in response to a reduction in nephron mass are dependent largely on intact aldosterone secretion and adequate sodium delivery. If aldosterone production is deficient, hyperkalemia is likely to occur as is seen in patients with generalized or selective aldosterone deficiency.28,29,31 Aldosterone levels are often elevated in patients with renal insufficiency, probably as an adaptive response to reduced renal mass, which helps prevent hyperkalemia. Eventually, the mechanisms for potassium adaptation are overwhelmed, and the absolute quantity of potassium excreted is less than the intake of potassium despite an increase in fractional potassium excretion greater than 100%. With a normal potassium intake, potassium retention does not occur until a GFR of 5 to 10 mL/min is reached. However, if potassium intake is increased, especially acutely, hyperkalemia may ensue despite a relatively preserved GFR (ie, 10 to 40 mL/min). Normally, only 10% of an ingested potassium load is excreted in the stool. However, in the presence of renal failure, adaptive changes in the colonic mucosa permit an increase in the amount of potassium excreted by the colon. Potassium excretion by the colon is an active process that is markedly enhanced in patients with ESRD. 32 Indeed, rectal excretion of potassium in hemodialysis patients may be 200% greater than in normal individuals33 and increased activity of Na,K-A TPase in the colonic mucosa of patients on dialysis has been reported. 34 Aldosterone receptors are present in colonic mucosa, and a role for mineralocorticoids in this adaptive response has also been suggested.35 The clinical relevance of these observations is not clear at present, but they suggest that interference with colonic po-

424

SALEM, ROSA, AND BATLLE 4.0 3.5

~

3.0

EO

2.5

.z -' ~

2.0

'" '"'" 0::

1.5

..,

E

1.0 0.5 0

30

60 Time,

90

120

minutes

Fig 1. Time-related changes in plasma potassium concentration in uremic and control rats after an oral load of potassium or potassium plus glucose. Plasma potassium levels were significantly higher following an oral potassium load compared with potassium plus glucose. Uremic animals had a significantly greater elevation in plasma potassium in either case compared with control animals. Control + K (0); control + K + glucose (e); uremic animals + K (6); uremic animals + K + glucose (A). (Reprinted with permission.")

tassium excretion in conditions such as constipation may further predispose to hyperkalemia in patients with ESRD. EXTRARENAL POTASSIUM TOLERANCE IN CHRONIC RENAL FAILURE AND UREMIA

Na,K-ATPase Activity Na,K-A TPase is an ubiquitous enzyme vital for the regulation of intracellular potassium concentration and, consequently, the plasma potassium level. Marked inhibition of the activity of this enzyme is known to cause hyperkalemia, even in subjects with normal kidney function, in the setting of severe digitalis intoxication. 36 ,37 In view of the critical role of this enzyme in potassium homeostasis, a careful examination of the activity of this pump in uremia is essential. Since increased activity of renal Na,K-ATPase plays a major role in the kidney's ability to increase potassium excretion under conditions of reduced renal mass, one might anticipate that the activity of this enzyme would be also increased in skeletal muscle in the setting of chronic renal insufficiency, thereby facilitating extrarenal potassium disposal. However, the bulk of the available evidence suggests that the activity of this enzyme is reduced in skeletal muscle in rats with advanced renal insufficiency. A recent study demonstrated a 50% reduction of basal Na,K-

ATPase activity in skeletal muscle tissue from rats made uremic by three-fourths nephrectomy.38 The functional correlate of the defect in muscle Na,K-ATPase activity was the finding that at 30 and 60 minutes after the administration of a potassium load, the magnitude of the increment in plasma potassium was significantly greater in uremic rats than in control rats (Fig 1). It must be appreciated that at least two molecular forms of the catalytic a-subunit of Na,K-ATPase (al and a2) exist. 39 Reduced total Na,K-ATPase activity may be secondary to impaired activity of the a l-isoform of the enzyme, whereas the a2isoform, which has a high affinity for ouabain, may be intact. 38 Some of the factors that may be responsible for impairment of the pump function in uremia are listed in Table 1. In humans, investigation of erythrocytes from uremic patients have reported both a decrease in Na,K-ATPase activity and a reduced number of pump sites. 40-42 The intracellular sodium concentration has also been found to be increased in erythrocytes from uremic patients,43 consistent with impaired Na,KATPase function. Muscle cells from uremic patients show a decrease in total exchangeable potassium, as well as decreased muscle K content and concentration. 44 Indeed, some uremic symptoms, such as those related to peripheral neuropathy and uremic encephalopathy, have been ascribed to defective pump function. 45 The abnormal protein metabolism in uremia may be causally linked to the decreased number of pumps.46 Further, there is some evidence that fluid overload may result in elaboration of a humoral factor that may alter Na and/or K transport in uremic cells. 41 ,47 While the mechanisms involved in the modification of pump activity in uremia need further study, it seems reasonable to postulate that inhibition ofNa,K-ATPase could result in a defect Table 1. Possible Causes of Impaired Na, K-ATPase in Uremia Decreased pump sites due to decreased rate of synthesis Downregulation of the pump by total body potassium deficit Circulating inhibitor in uremia Fluid overload Abnormal protein metabolism in uremia

425

HYPERKALEMIA AND CHRONIC RENAL FAILURE

in extrarenal potassium disposal that would enhance the predisposition to hyperkalemia in patients with advanced renal failure. A Na-K pump defect in uremia would imply that patients with chronic renal failure are less able to handle an external potassium load not only because oflack of kidney function, but also because of impairment of extrarenal handling of potassium. The described pump defect seems correctable by dialysis. 43 It follows that inadequately dialyzed patients might be more susceptible to hyperkalemia than well-dialyzed patients. However, correction of the pump defect in dialysis patients takes approximately 8 to 12 weeks when studied in erythrocytes. 43 Since the mature human erythrocyte is enucleate and unable to synthesize new pump sites, correction of the pump defect most likely occurred in the nucleated precursor cells during the course of erythropoiesis. The long lag period between initiation of dialysis and the correction of the pump defect may reflect the time required for the appearance in the peripheral circulation of a pure population of cells produced at a time when uremia was corrected by dialysis. Because the leukocyte and the muscle cell are nucleated, it may be possible to correct the abnormality in the number of pump sites in these tissues after only a few days of dialysis. 43 Other Factors That May Influence Extrarenal Potassium Handling

In addition to alterations in the activity of the Na,K-A TPase, ESRD patients are exposed to other factors that may result in either enhancement or impairment of extrarenal potassium tolerance (Table 2). Some hormonal changes associated with uremia may facilitate the defense against hyperkalemia. For example, immunoreactive insulin levels are usually increased. 48 ,49 While end-organ resistance to the glucose translocating effect of insulin is well described,48-50 the potassium-lowering effect of insulin has been shown to be intact in humans with uremia. 51 Chronic hyperinsulinemia may thus afford some protection against hyperkalemia. The plasma aldosterone level is also frequently elevated in patients with advanced chronic renal insufficiency.52,53 This elevation may facilitate some potassium translocation intracellulary and may

Table 2. Possible Factors Influencing Potassium Tolerance in Renal Failure Factors Enhancing Potassium Tolerance

Factors Impairing Potassium Tolerance

Renal adaptation Increased potassium excretion by colon Hyperinsulinemia High aldosterone levels High catecholamine levels Extrarenal adaptation

Defective sodium potassium pump Resistance to catecholamine action PTH excess

also increase fecal potassium content through its action on the colonic mucosa. A hormone that might impair extrarenal potassium disposal in patients with chronic renal insufficiency is parathyroid hormone (PTH). Although PTH does not have a direct effect on Na,K-ATPase, high circulating levels of PTH lessen extrarenal disposal of potassium in uremic rats. 54 This impairment may be a consequence of the enhancement of potassium efflux from cells produced by increasing intracellular calcium. Potassium tolerance in partially nephrectomized rats with chronic renal failure is improved by parathyroidectomy and by the administration of the calcium channel blocker verapamil. 55,56 PTH may also impair potassium tolerance by other mechanisms. An effect for PTH on release of insulin from the pancreas has been reported. Insulin release in response to potassium infusion was found to be impaired in uremia, and could be corrected by prior parathyroidectomy. 57 In contrast to these findings that indicate that hyperparathyroidism impairs potassium tolerance, clinical observations have been made that plasma potassium concentration increases after parathyroidectomy.58,59 However, hyperkalemia following parathyroidectomy in dialysis patients may reflect confounding elements in the early postoperative period, rather than any specific effect of PTH removal. This issue clearly needs to be examined under controlled conditions. In rats, extrare:nal potassium tolerance could be improved by the administration of calcium channel blockers. 56 This effect may be the consequence of a diminution in calcium-mediated net potassium efflux from cells, since increased intracellular calcium enhances potassium permeability

426

in vitro, in human red blood cells (the Gardos effect),6o and in other cells. 61

Evaluation of Extrarenal Potassium Tolerance Extrarenal potassium tolerance in ESRD patients has not been examined extensively. In one study, Fernandez et al62 examined extrarenal p0tassium handling in 10 patients with ESRD. Their response to a standard potassium infusion was compared with that of eight healthy control subjects. The absolute amount of potassium translocated into the intracellular fluid compartment was similar in both groups. However, when expressed as a percent of the retained amount, the patients translocated less potassium into the cells (21 % in ESRD patients v 51 % in controls). These investigators concluded that extrarenal potassium homeostasis in ESRD patients is impaired. However, it should be noted that the baseline potassium in the group with renal failure was higher than the control group. Since cellular potassium uptake is decreased at higher plasma potassium concentrations,63 the possibility that the higher plasma potassium concentration in the group with ESRD influenced the results cannot be ruled out. In addition, the study of Fernandez et al included patients who would be expected to have impaired extrarenal potassium regulatory mechanisms, such as diabetics and patients with hyporeninemic hypoaldosteronism, which clearly obscures the interpretation of the results. Other studies reporting impaired extrarenal disposal of an oral potassium load in patients with chronic renal failure might also be criticized because of a higher potassium level or a greater degree of acidosis in the patients compared with the controls. 64•65 However, impaired potassium tolerance has been demonstrated recently in uremic rats under similar conditions of basal plasma potassium, blood pH, and plasma bicarbonate concentration. 38 Contrariwise, some studies have failed to demonstrate an exaggerated increase in serum potassium in patients with chronic renal failure. 1.66 Perez et al66 administered an oral potassium load to patients with moderate degrees of renal impairment (creatinine clearance of approximately 0.53 mLls [32 mL/minD and monitored the plasma potassium response, as well as the urinary potassium excretion. Although renal failure patients retained more of the potassium

SALEM. ROSA, AND BATLLE

load, they translocated more potassium into the intracellular compartment, which was interpreted as an indication of enhanced potassium tolerance. Whether extrarenal potassium tolerance, strictly defined as the kalemic response to potassium loading, is normal, impaired, or enhanced in patients with moderate renal insufficiency and in those with ESRD remains to be determined. The available data suggest that with the development of ESRD and the uremic syndrome there is impaired extrarenal potassium tolerance that is related, in part, to a defect in Na,K-ATPase. This is not to say that the detrimental effects of hyperkalemia on organ function (eg, the heart) are worsened in uremia. In fact, some "tolerance" to hyperkalemia is said to exist in uremia, and cases of severe hyperkalemia without electrocardiographic (EKG) changes have been documented. 67 Before embarking on an extensive discussion of the individual hormones that regulate extrarenal potassium metabolism, it is important to recognize that major alteration in hormonal regulation and actions may occur in uremia. Uremia may be associated with decreased synthesis of certain hormones (eg, gonadal hormones), abnormal feedback regulation (eg, prolactin) and altered effects on membrane function and hormone receptors. Since the kidney takes part in the metabolism of several hormones, particularly of the polypeptide type (eg, insulin and PTH), elevated levels secondary to impaired renal clearance are usually present. While these elevated levels may be expected to prolong the action of these hormones, interference at the level of the receptor may be present. The difficulty arises when one attempts to equate high hormone levels with hyperfunction. In uremia, elevated hormone levels alone do not necessarily signify hyperfunction. Moreover, uremia may be associated with selective impairment of some but not all actions of the hormone. The prime example of this selective impairment is insulin where its effect on carbohydrate metabolism is impaired while that on potassium transport remains intact. Receptor number, affinity, or postreceptor events following hormone binding may also be altered in uremia. In the following sections, the various hormones that influence extrarenal potassium disposal are discussed.

427

HYPERKALEMIA AND CHRONIC RENAL FAILURE

Adrenergic-Mediated Internal Potassium Disposal The effect of catecholamines on internal potassium distribution was noted as early as 1934 by D'Silva, who first reported that a rapid injection of adrenaline in cats caused an immediate short-lived increase followed by a prolonged de~ crease in the serum potassium concentration.68 More recently it has been established that administration of the {3-blocker propranolol69 or the a-agonist phenylephrine70 will impair the extrarenal disposition of an acute load of potassium. Thus, {3-adrenergic stimulation enhances and aadrenergic stimulation impairs extrarenal potassium disposal. The increased cellular uptake of potassium promoted by epinephrine is mediated by stimulation of the {3rreceptor. 71,72 Whether physiologic elevations of extracellular potassium can stimulate secretion of endogenous adrenomedullary epinephrine remains unresolved although supraphysiologic levels of potassium' can cause the release of catecholamines from isolated chromaffin cells73 and perfused adrenal glands. 74 However, there are at least two common physiologic circumstances in which endogenous catecholamines might act to defend against acute increments in extracellular potassium concentration: feeding and vigorous exercise. Since feeding is associated with stimulation ofthe sympathetic nervous system,75 enhanced {3-adrenergic-mediated extrarenal potassium disposal may help to limit elevations of the plasma potassium in the immediate postprandial period. The high circulating levels of catecholamines that occur during vigorous exertion may attenuate the acute hyperkalemia of exercise, as suggested by the finding that exercise-induced hyperkalemia is exaggerated by concomitant {3-blockade. 76 In a clinical context, the contribution of the adrenergic nervous system to potassium homeostasis has been studied in patients with renal failure. It has been well documented that the administration of drugs that possess {3rantagonist · 77 or t h ose that are nonselective {3prope rtles blockers78 can produce significant hyperkalemia in patients on dialysis. However, the increase in plasma potassium during exercise in patients on selective .B-blockers is similar to that induced by exercise a1one.79 It is important to appreciate that the response

to .Bradrenoreceptor-mediated internal potassium disposal is heterogeneous, as judged by the variable responses to epinephrine infusion (Fig 2). Allon et al 80 also observed that some patients with hyperkalemia and ESRD failed to respond to nebulized albuterol, a selective {3radrenergic receptor stimulant. This observation is consistent with previous studies showing that in patients with ESRD epinephrine infusion does not increase the heart rate normally either because of reduced end-organ response, or because of receptor occupancy by a circulating inhibitor. 81 ,82 In favor of the circulating inhibitor hypothesis is the finding by Souchet et al 83 of unexpectedly high density of {3radrenoreceptors in mononuclear cells from patients with chronic renal failure despite markedly elevated plasma catecholamine levels. From the limited data available, it may be estimated that resistance to catecholamine action is seen in approximately 40% of patients with ESRD (see Table 5). This has obvious implications for the treatment of hyperkalemia (see below).

Insulin-Mediated Internal Potassium Disposal Insulin-receptor interaction. The metabolic effects of insulin are initiated by the interaction of the insulin molecule with a highly specific re-

! Epinephrin

B

0

15

30

Time, min

Fig 2. Effect of epinephrine infUSion on plasma potassium in (A) patients with ESRD (n = 10) and (8) normal control subjects (n = 6). Plasma potassium levels decreased significantly at 15 and 30 minutes of epinephrine infusion in both groups. However, four patients with ESRD were resistant to the hypokalemic effect of epinephrine. (Reprinted with permission. 78 )

428

ceptor on the plasma membrane. This binding is saturable, of very high affinity, and highly specific. The insulin receptor is a membrane protein of high molecular weight made up of four disulphide-linked subunits. The binding of insulin induces autophosphorylation of the receptor. 84-86 The stimulation of glucose transport, amino acid transport, and ion transport can be considered as more or less immediate consequences of changes in membrane configuration induced by interaction with membrane receptors. However, the effect of insulin on intracellular metabolism requires the generation of a second messenger that carries the information into the cell. The nature of this second messenger remains elusiveY Insulin-induced potassium uptake. Both splanchnic (liver) and peripheral tissue (muscle) potassium uptake is enhanced by insulin infusion. Due to its anatomic position, the liver is exposed to higher insulin concentrations than other tissues. Insulin concentration in the portal blood is three to 10 times higher than that of the systemic circulation. In addition, approximately 50% of secreted insulin may be metabolized in one pass in the liver before being diluted in the systemic circulation. This differential exposure of the liver to insulin cannot be mimicked by insulin administered parenterally. Not surprisingly, hepatic uptake accounts for 70% of potassium translocated intraceUularly in the first hour after a potassium load. 88 Patients with hepatic cirrhosis tolerate a potassium load less well than patients with normal liver function, possibly because of decreased potassium uptake by the liver.89 Peripheral tissue uptake (mainly muscle and adipose tissue) is responsible for uptake of the remaining 30% in the first hour and assumes a more important role thereafter. The effect of insulin on plasma potassium is thought to be primarily the consequence of direct stimulation of the activity of the Na,K-ATPase by a mechanism independent of cyclic adenosine monophosphate (cAMP).90,91 However, the exact mechanism is unknown. Because the insulin effect on cellular potassium uptake is rapid, it is unlikely to be mediated through the synthesis of new protein. Also, inhibitors of protein synthesis do not interfere with insulin's action on cellular potassium uptake. 92 The binding of insulin to its receptor results in an alteration in the cellular membrane lipid matrix and fluidity, with a subsequent effect

SALEM, ROSA, AND BATLLE

on membrane permeability.93 Additionally, insulin stimulates the Na+ /H+ exchanger resulting in cellular H+ efflux and sodium influx into the cells. 90,94 The expected increase in intracellular sodium may stimulate Na,K-ATPase, causing potassium entry from the extracellular space. The action of insulin on potassium translocation into the cellular space can be dissociated from its effect on plasma glucose. The latter is usually impaired in uremia. 51 Skeletal muscle uptake of glucose in response to insulin is reduced and hyperinsulinemia in the face of noromoglycemia is well documented in uremic patients. This has been attributed to a post-receptor-binding defect, which has not been fully defined. 95 In contrast, cellular potassium uptake in response to insulin is similar in both uremics and healthy subjects. 51 In a recent study in a rat model of chronic renal insufficiency, Goecke et al 38 concluded that there was enhanced sensitivity to the action of insulin on extrarenal potassium disposal in uremia. These investigators demonstrated that stimulation of endogenous insulin by oral glucose resulted in a greater intracellular translocation of potassium in uremic rats than in controls (Fig 1). Interestingly, in vitro studies using soleus muscle derived from the same animals showed that the addition of insulin caused a greater stimulation of ouabain-sensitive 86Rb uptake in uremic muscle than in normal tissue. These results are even more impressive if one considers that under baseline conditions ouabain-sensitive Rb uptake, a marker of Na,K-ATPase activity, is depressed in muscle from rats with chronic renal failure. 38 ,96 Moreover, Tuck et al97 showed that uremic dogs demonstrated an enhanced insulin response to a potassium load compared with controls. However, other reports showed decreased insulin secretion in response to insulinogenic stimuli in uremia. 98-1OO In one study, the insulin response to a potassium load in pancreatic islet cells isolated from chronic renal failure animals was lower than controls. 57 In this study,57 the altered insulin response to potassium was attributed to hyperparathyroidism, since cells from parathyroidectomized animals demonstrated improved insulin response. On the other hand, Allegra et al 100 in a study in uremic patients could improve glucose tolerance by concomitant administration of aminophylline, suggesting that the

429

HYPERKALEMIA AND CHRONIC RENAL FAILURE

defect in insulin response to insulinogenic stimuli may rest in the adenyl cyclase cAMP system. Insulin is well positioned to play an important role in the regulation of plasma potassium concentration in patients with impaired renal function. Insulin secretion is stimulated by feeding, hyperglycemia, and increased plasma potassium concentration. While an increase of plasma potassium concentration of I mmol/L has been shown to stimulate insulin secretion in both humans and dogs, 19, 101 more modest elevations of potassium do not appear to elevate the insulin level. I02 However, since the portal to systemic blood venous concentration ratio of insulin is 3: 1, a small increase in portal circulating insulin would not necessarily be detected in the systemic circulation, and Hiatt et al 103 have confirmed the effect of smaller increments in plasma potassium on portal insulin level. Basal insulin secretion has also been demon-. strated to play an important role in extrarenal potassium homeostasis, When the insulin secretion was reduced by somatostatin infusion in human volunteers the potassium concentration increased significantly and this effect was reversed by restoration of the basal insulin levels. 104 An independent role for basal insulin secretion on the regulation of the plasma potassium concentration in renal failure can be deduced from the recent finding that inhibition of basal insulin secretion through fasting resulted in hyperkalemia in nondiabetic patients with ESRD,105 In this study, significant hyperkalemia (serum potassium > 5,7 mmol/L) was corrected by food ingestion alone. The role of basal insulin on internal potassium disposal in the setting of renal failure may be even more important than previously appreciated, since somatostatin infusion caused a much greater increase in the fasting plasma potassium in rats with renal failure than in controls38 (Fig 3).

Mineralocorticoid-Dependent Potassium Disposal Whether mineralocorticoids have a physiologic role in extrarenal potassium homeostasis when the extracellular potassium concentration is within the normal range has not been cle~rly established. However, deviations of plasma potassium from normal clearly affect production of

1.2

~ '".-:E

•

1.0

0-

~

., .,.,E

"'1

ii:

0.8 0.6 0.4 0.2

0

30

60

90

Time, minutes

Fig 3. Effect of somatostatin infusion on plasma potassium in uremic and control rats (n = 3). Uremic animals (e) exhibited a significantly greater increase in plasma potassium when basal insulin levels were inhibited by insulin compared with control animals (0) (P < 0.0001). (Reprinted with permission. H )

mineralocorticoids and thereby suggest a regulatory role. Hyperkalemia stimulates, while hypokalemia suppresses, mineralocorticoid production by the zona glomerulosa of the adrenals. Several studies have shown that plasma aldosterone levels are elevated in patients with ESRD. 52,53 The elevated plasma aldosterone level that occurs in chronic renal failure clearly participates in the adaptive increase in potassium disposal by the kidneys, which helps to prevent hyperkalemia. Indeed, selective hypoaldosteronism is by far the most common cause of hyperkalemia in patients with moderate degrees of renal insufficiency. 29 There is also some evidence that, in addition to their potent kaliuretic effect, mineralocorticoids playa role in extrarenal potassium homeostasis. 106,107 Mineralocorticoids restored potassium levels toward normal in glucocorticoid-replaced, adrenalectomized rats that received a potassium load following bilateral nephrectomy.71 The hyperkalemia that occurs in patients with hyporeninemic hypoaldosteronism has been corrected with mineralocorticoid replacement alone without necessarily producing detectable increases in renal or fecal potassium excretion. 108 In anephric dialysis patients, the addition of exogenous mineralocorticoid over a 60hour period prior to the introduction of an acute oral potassium load decreased the rate of increase in serum potassium compared with a control period. Furthermore, blocking the activity of circulating levels of aldosterone resulted in a higher rate of increase of plasma potassium after the

430

acute potassium load. The volume of distribution of potassium during deoxycorticosterone acetate (DOCA) administration was significantly greater than in the spironolactone and control periods. 109 Moreover, in anuric dialysis patients, hyperkalemia has been corrected by mineralocorticoid administration. 110 Despite these observations, the relative contribution of aldosterone to extrarenal potassium homeostasis remains unclear and the tissue site for the mineralocorticoid action is not established. Receptors for the mineralocorticoids are known to exist in the colon, III mononuclear leukocytes,112 and kidney.ll3 However, we know of no study showing aldosterone receptors in skeletal muscle, the presumed target of its cell potassium-translocating action.

Acid-Base Status and Potassium In 1934, Fenn and Cobb l14 described an inverse relationship between the plasma potassium concentration and the blood pH. Although changes in blood pH may influence renal potassium handling, the effect discussed here relates to potassium shifts between the intracellular and extracellular compartments. In 1956, Burnell et aI, liS in what later became a rather established axiom, suggested that the kalemic response to blood pH changes is predictable, with an increase or a decrease in plasma potassium of approximately 0.6 mmolfL for each 0.1 U of pH change. It is often not appreciated that this axiom originated as a qualitative observation in a small number of patients with differing acid-base disturbances. 115 ,116 Burnell's study included only five patients with complex acid-base disorders observed over an extended period of time. Significant renal losses of potassium, as well as correction of alterations other than acid base status, may have contributed to the decrease in plasma potassium noted by these investigators. A wide variability in the kalemic response to pH changes is apparent from a critical review of the literature. 117 Metabolic acid-base disorders have a more pronounced effect on plasma potassium concentration than do respiratory acidbase disorders. 118 The smaller effect of respiratory disorders compared with metabolic disorders also illustrates the limited effect of pH per se on the plasma potassium. Within the category of metabolic acidosis, a striking discrepancy exists between the effect of mineral acid induced acidosis

SALEM, ROSA, AND BATLLE

and that produced by organic acids. 119 While mineral acid administration consistently leads to increments in plasma potassium concentration, organic acidosis does not elicit a significant plasma potassium increment. 1l9,120 As an explanation for this difference, it has been suggested that in acidosis due to retention of mineral acid, the entry of hydrogen ions into the cells is incompletely accompanied by its anion. Electroneutrality is maintained in part by the translocation of potassium out of the cell and, thus, its concentration in plasma increases. In contrast, organic acids are more diffusible across cell membranes by a process of non ionic diffusion of the undissociated form. Clinical examples of the lack of effect of organic acidosis on the plasma potassium include lactic acidosis and diabetic ketoacidosis. Hyperkalemia is not a prominent feature of lactic acidosis, despite marked reduction in the blood pH. 121 ,122 Likewise, diabetic ketoacidosis is often not associated with hyperkalemia, despite the presence of insulinopenia. However, total body potassium deficit is present in this condition. These observations can now be extended to patients with ESRD. In a group of dialysis patients who performed brief intense exercise, the increased serum potassium induced by exercise decreased to the basal value after 1 minute, despite the persistence of significant hyperlactatemia and metabolic acidosis.1 23 Finally, it should be appreciated that the decrease in plasma potassium concentration associated with metabolic alkalosis is considerably less than the increase in plasma potassium concentration caused by the infusion of acid. 124 Fraley and Adler l25 suggested that the bicarbonate level may be more crucial to the potassium alteration than the pH change itself. In both human and animal studies, plasma bicarbonate concentration, but not blood pH, was found to correlate with plasma potassium concentration. In rats studied under isohydric conditions, increasing plasma bicarbonate level was associated with a decrease in plasma potassium concentration, independent of the change in pH. 126 The human studyl25 included 14 patients with chronic renal failure and hyperkalemia. Patients received variable amounts of sodium bicarbonate solution in dextrose and the effect on plasma potassium concentration was examined. Two groups of pa-

431

HYPERKALEMIA AND CHRONIC RENAL FAILURE Table 3. Studies Evaluating the Effect of Bicarbonate Infusion on Plasma Potassium No. of Patients

Dose (mmol/L)

Total Dose (mmoIfL)

K, (mmoIfL)

Kiln (mmol/L)

ilK (mmol/L)

3

?

?

7.5 ± 0.91

5.93 ± 1.01

-2.23 ± 0.17

17to19h

108 to 408

7.25 ± 0.61

4.68 ± 0.73

-2.58 ± 0.22

2.5 to 12 h

?

6.4 ± 0.4

4.8 ± 0.3

-1 .6

89 to 178

6.3 ± 0.3

4.8 ± 0.2

-1.6 ± 1

240

5.66 ± 0.19

5.83 ± 0.23

+0.17

30 min

ESRD

120

5.39

5.6

+0.21

60 min

ESRD

?

4.8 ± 0.3

4.4 ± 0.3

-0.4

180 min

ESRD

Study

Type

Burnell. 1956115

Prospective (no controls) Retrospective

4

220 ± 70

Prospective

5

89 to 133 Over4t06h

9

89 to 178 Over4t06h 8.4% in water 4 mmol/min 1.4% in water 2 mmol/min 1 mmol/kg BW over 2 h

Schwarz. 1959"· Fraley. 1977'"

Blumberg. 1989137

Prospective

10 8

Gutierrez. 1989'"

Prospective

12

Duration of Treatment

Comments

At the end of 5 periods 3.5 to 4 h each Same

t

Abbreviations: K,. initial potassium; K.... final potassium concentration; ilK. decrement achieved over the specified period of time . • In this subgroup. pH increased with bicarbonate infusion. t In this subgrouP. pH remained unchanged.

tients could be identified retrospectively (Table 3). In one group, an increase in plasma bicarbonate concentration was associated with an increased blood pH. In the other group, blood pH did not change appreciably because Peo2 was normalized (ie, the compensatory hyperventilation was corrected). Yet, plasma potassium decreased in both groups as plasma bicarbonate increased. It must be appreciated that the reported changes in potassium occurred after 24 hours of bicarbonate infusion, at a time when plasma bicarbonate had markedly increased from approximately 14 to 24 mmol/L. An increase in plasma bicarbonate of this magnitude obviously cannot be readily achieved over a brief period of time (see Implications for Treatment of Hyperkalemia). Levels of hormones that may affect potassium translocation may be influenced by changing blood pH. However, this cannot fully explain the alterations in plasma potassium induced by pH changes. 127ol30 Reviewing the sequence of events that follow the infusion of an acid load may help clarify some of the aspects of the relationship between acid-base disorders and the changes in plasma potassium. Within seconds of its infusion, an acid load is initially buffered by the plasma buffers. Since the bicarbonate buffer is the most abundant in the plasma and constitutes an open buffer system physiologically, it is the most im-

portant. Hydrogen ions will also diffuse into the red blood corpuscles where they are mainly buffered by hemoglobin. Over the next 30 minutes or so the acid load diffuses into the interstitial space, which, due to its relatively large volume, contributes more to the buffering of an acid load than the plasma. 131 Distribution into the intracellular compartment then follows. 132.133 Depending on the permeability of the cell membrane to the accompanying anion, the acid may enter the cell as a molecule or only the hydrogen ion will diffuse unaccompanied by its anion. It is the latter process that will result in the exchange of hydrogen for potassium. Potassium, which neutralizes the negative charge on the intracellular proteins, may then be released and can diffuse out of the cell. If this scenario is correct, it means that the change in plasma potassium with acidosis will depend on several factors. One is the amount and type of acid infused and, eventually, the amount of acid that is neutralized inside the cells ( - 50% of an acid load is neutralized intracellularly.132.133 The status of acid-base balance before the infusion of the acid load and the presence or absence of total body potassium deficit will further determine the magnitude of the change in plasma potassium. This may explain why the ratio of increase in serum potassium to any given decrease in the pH is inconsistent from one study to another,

SALEM, ROSA, AND BATLLE

432

and tends to be higher later in the course of an acute acid load when the buffer reserve is being exhausted. THERAPEUTIC IMPLICATIONS

Although hyperkalemia is a frequently encountered, potentially life-threatening condition, there have been no well-controlled clinical studies evaluating the various therapies often used. Personal approaches based on soft data and "timehonored" regimens are commonplace. While careful clinical investigations are needed to define the optimal approach, we believe that the following review of the literature may lead to reasonable conclusions as to the effectiveness of the various therapeutic modalities used to lower the plasma potassium concentration (see Tables 3-5). Since severe hyperkalemia most often occurs in the setting of chronic renal insufficiency, it is important to know whether the mechanisms of internal potassium disposal are intact or not in this setting. Acute treatment of hyperkalemia should be followed by identifying the potential risk factors in each patient and designing a plan that may avert recurrence of the problem.

Treatment of Severe Hyperkalemia In the presence of EKG changes, the intravenous administration of calcium can be life-saving, as it rapidly counteracts the toxic effect of hyperkalemia on the myocardium. This approach is universally accepted. However, plasma potassium is not lowered by calcium administration. Lowering the plasma potassium concentration must be achieved by other measures, including dialysis, elevation of insulin levels in the blood, induction of metabolic alkalosis (bicarbonate infusion), administration of potassium exchange resins, and administration of ,82-adrenergic stimulants. The order in which these drugs are used, as well as their comparative efficacy in frankly hyperkalemic patients, has not been criticallyexamined in controlled studies. The use of bicarbonate has enjoyed widespread popularity for many years. In a recent survey addressed to the directors of nephrology training programs in the United States, bicarbonate was recommended by most responders as the first line of therapy. 134 This approach has been the subject of some debate. In our opinion, alkali therapy should not be relied on solely as first-line therapy

for severe hyperkalemia occurring in the context of renal failure. 135,136 Interpretation of the results of the available clinical studies that used bicarbonate therapy is confounded by several factors (Table 3). Most of these studies were the result of qualitative, uncontrolled observations in a small number of patients, many of whom suffered from complex acid-base disturbances. As it is often the case in retrospective studies, no uniform protocol was followed. The dosage was not always documented, but it ranged from 89 to 408 mmol of bicarbonate. However, most important is the observation that the hypokalemic effect ofbicarbonate infusion required many hours (Table 3). In the study by Fraley and Adler, plasma potassium had decreased by only 0.6 mmol/L after 6 hours of bicarbonate infusion and by as much as 1.6 mmol/L after 24 hours. In a recent controlled study, Blumberg et al 137 showed that bicarbonate infusion failed to produce a hypokalemic effect over a period of 1 hour in patients with ESRD (Fig 4). These results have been confirmed in a preliminary report by Gutierrez et all38 in a group of dialysis patients who received a bicarbonate infusion over a 3-hour period. These two controlled studies would appear to demonstrate that bicarbonate infusion does not have any major effect on plasma potassium concentration in patients with ESRD. However, it should be noted that the patients in the study of Blumberg et al had very mild hyperkalemia, while those in the study of Gutierrez et al were actually normokalemic and received a very small dose of bicarbonate. Whether bicarbonate is more effective in patients with more severe degrees of hyperkalemia is unknown. Until this issue is examined more rigorously, it is inadvisable to rely solely on this therapeutic modality when a rapid decrease in plasma potassium concentration is desired, since the onset of this effect may be delayed and its magnitude may vary. Obviously, bicarbonate should be administered promptly for the correction of severe acidosis if present. The risk of volume overload and the potential to produce a hyperosmolar state mandates close monitoring of the patients' electrolytes, particularly since hyperosmolarity, per se, may result in a shift of potassium into the extracellular space. 139,140 Table 4 lists the studies that have evaluated the efficacy of insulin in the treatment of hyperkalemia. The effect of insulin on plasma potas-

433

HYPERKALEMIA AND CHRONIC RENAL FAILURE ~

Plasma Potassium, mmol/L

o

-0.5

-1 .0

-1.5

l' P < 0.01 I

o

I

10

I

20

I

I

30 40

I

I

50 60

Duration of Treatment, min Fig 4. Changes in plasma potassium (mmol/L) during intravenous infusion of bicarbonate 8.5% (_), epinephrine (0), or insulin plus glucose (e), and during hemodialysis (.6.). Note the lack of effect of bicarbonate infusion on plasma potassium concentration. (Reprinted with permission. 137)

sium concentration was demonstrated in all patients studied, in contradistinction to the effect of i3z-stimulation on plasma potassium, which may be impaired in some patients with chronic renal insufficiency (see below). A decrease of plasma potassium concentration of approximately 1 mmol/L in response to insulin administration has been reported consistently. This decrease in potassium was seen within the first 20 to 30 minutes and lasted for up to 2 hours. The higher the initial potassium, the greater was the reported decline. Also, the higher the insulin dose the more remarkable was the effect and the decrement of plasma potassium was often accompanied by reversal of the EKG changes of hyperkalemia. When insulin was compared with the other treatments for hyperkalemia, Blumberg et al l37 found that insulin was the most rapidly act-

ing agent short of hemodialysis (Fig 4). These observations further support the use of insulin, together with glucose to prevent hypoglycemia, as the most efficacious way to treat hyperkalemia in ESRD until dialysis is available. Since a prompt hypokalemic response is dependent on achieving high plasma insulin levels, the drug should be administered as an intravenous bolus at a dose of 5 to IOU of regular insulin. Hypoglycemia can occur in patients receiving glucose and insulin. Transient hyperglycemia usually follows the administration of glucose and insulin at 15 minutes, which usually resolves by 30 minutes. A progressive decline in plasmaglucose concentration to frankly hypoglycemic levels at 60 minutes may sometimes occur. In one study, hypoglycemia occurred in 75% of the patients following insulin and glucose. 141 In the study by Blumberg et al, five of 10 hemodialysis patients developed hypoglycemia. This propensity toward hypoglycemia with insulin can be attributed to several factors, such as malnutrition, absent renal gluconeogenesis, prolonged half-life of insulin in renal failure patients, and possibly impaired response to counterregulatory hormones, particularly epinephrine. Lens et al l42 avoided hypoglycemia by administering 40 g of glucose together with the insulin. It is therefore important that the blood sugar be monitored carefully. To prevent hypoglycemia, the glucose bolus can be followed by a dextrose 10% drip for several hours. A group of investigators from Barcelona first demonstrated that albuterol (a i3z-agonist) administered intravenously was effective in the treatment of hyperkalemia in dialysis patients. '42,'43 Subsequently, Allon et al 80 successfully employed the nebulized form of albuterol, a selective i32-adrenergic stimulant. Montoliu et al '43 also demonstrated reversal of severe EKG changes of hyperkalemia, accompanied by a decrease in serum potassium from 8.7 to 5.9 in one patient treated only with intravenous albuterol. Table 5 summarizes the studies that have examined i3-adrenergic-mediated extrarenal potassium disposal. The impression that emerges from these studies is that the responses to i3-adrenergic stimulation are heterogeneous in patients with renal failure. Some patients have blunted responses, whereas many display a significant decrease in plasma potassium in response to epi-

434

SALEM, ROSA, AND BATLLE Table 4. Studies Examining the Effect of Glucose and Insulin on Plasma Potassium No. of Patients

Study

Type

De Fronzo, 1980""

Prospective

19 (healthy subjects)

Alvestrand, 198461

Prospective

12 (CRF) 18 (healthy controls)

Lens, 1989'42

Prospective

10 (CRF)

Blumberg, 1989'37

Prospective

10 (ESRD)

Allon, 1990'41

Prospective

12 (ESRD)

K. Dose

(mmol/L)

Kftn (mmol/L)

D.K (mmol/L)

?

?

-0.62 ± 0.1 -1.09 ± 0.1 -1 .49 ± 0.11

120 min

Splanchnic uptake 70% in first hour, then declines

?

?

-0.82 ± 0.12 -0.93 ± 0.07

120 min

6.7 ± 0.2

5.7 ± 0.2

-1 ± 0.1

60 min

Steepest decline during first 20 min Reversal of EKG changes

4.7 ± 0.22

-0.92

60 min

0.25 to 0.5 mU/kg/min 1 mU/kg/min 5-10 mU/kg/ min 42.6 mU/rrr/ min

Insulin (lOU) IV bolus + 40g glucose over 15 min 100 U of insulin in 500 mLof 20% glucose at a dose of 5 mU/kg/min lOU insulin as IV bolus + 50 mL of D50 over 5 min

5.62 ± 0.33

5.48 ± 0.21 mmol/L

-0.65 ± 0.09

Time

Comments

The greater the initial K, the greater the decline

SignifICant decrease noted at 15 min

Abbreviations as in Table 3.

nephrine infusion or adrenergic stimulation with specific ,82-agonists. Since this blunted response was seen both with epinephrine (which has stimulatory effect on both the a- and ,8-adrenergic receptors) and with the selective ,82 stimulant albuterol, it is unlikely that the absence of the decrease in plasma potassium is due to stimulation or enhanced activity of the a-receptors. The effect of ,82-agonists on plasma potassium concentration is rapid in onset and persists for up to 2 hours. Although albuterol given either intravenously or via a nebulizer reduces serum potassium, it should be noted that the dose used for nebulization is roughly 10 times higher than that used for the treatment of asthma (20 mg v 2.5 mg). When smaller doses similar to those given to patients with acute asthma were used (180 IJ.g twice by inhalation) only a minimal response was noted. 144 Although no serious side effects have been reported with this relatively large dose, two potential problems may be associated with the use of ,8z-adrenergic stimulants. One is the risk oftachyarrhythmias. The other is the risk of precipitation of angina in patients with coronary artery disease. Therefore, it may be prudent to avoid the use of these agents in pa-

tients with known coronary artery disease. Another disadvantage to the use of these agents is that some patients with ESRD are resistant to the hypokalemic effect of epinephrine and ,82 stimulants (Fig 2). Additionally, patients receiving nonselective ,8-blockers will not respond to this therapy. Combined treatment with albuterol and insulin plus glucose has been shown to produce a substantially greater decrease in potassium than that following each drug administered separately141,142 (Fig 5). Such an additive effect is consistent with experimental studies in rat soleus muscle, where tissue potassium uptake was higher following concomitant exposure to both agents. 145 While both drugs stimulate Na,KATPase, this effect is achieved by different pathways. The effect of ,8-adrenergic stimulation is mediated by a cAMP-dependent mechanism, while that of insulin is not completely understood (see above). Interestingly, when albuterol is administered with insulin, the late hypoglycemic effect that occurs when glucose and insulin are given could be minimized. 141 Following combined treatment with albuterol and insulin plus glucose, transient hyperglycemia occurred in a

HYPERKALEMIA AND CHRONIC RENAL FAILURE

435

Table 5. Studies Examining the Effect of Catecholamines on Plasma Potassium

Study Yang, 198678

Montoliu, 1987'43

Allon , 1989"" Blumberg, 1989'37 Lens, 1989'"

Allon, 1990'"

Type

No. of Patients

Prospective

10

Prospective

6 (healthy) controls 20

K.

Kiln (mmoIfL)