Kidney Iniernational, Vol. 11(1977), pp. 491—504

Potassium homeostasis in renal failureC. VAN YPERSELE DE STRIH0u

Renal Laboratory, Department of Medicine, Cliniques UniversitairesSt-Pierre, Louvain, Belgium

The critical importance of serum potassium con-centration in both health and disease has been dem-onstrated more than a century ago when it was shownthat rapid i.v. injection of potassium salts causedtoxic and even fatal effects on the heart of the experi-mental animal [I]. The role played by potassium inthe excitability of both nerve and muscle cells is nowwell established.

Maintenance of a normal serum potassium concen-tration depends to a large extent on renal function, asmore than 80% of the ingested potassium is excretedin the urine. Surprisingly, however, serum potassiumlevel is kept within or slightly above normal limitsduring most of the progressive course of renal failure.Maintenance of potassium equilibrium despite a fall-ing glomerular filtration reflects a progressive adapta-tion of the mechanisms regulating serum potassiumconcentration.

In the present paper we shall review the differentfactors involved in the defense of potassium homeo-stasis in renal failure: the adaptation of renal ex-cretory mechanisms, the role of gastrointestinal ex-cretion, the cellular tolerance to potassium load, theresulting changes in potassium balance, and finally,the complications which might alter potassium home-ostasis in renal failure.

Renaladaptation

It has long been recognized that severe renalinsufficiency is associated with increased concen-trations of serum potassium resulting in electro-cardiographic and neurologic disorders, sometimessevere enough to be lethal [2, 3]. Contrary to urea,however, whose rise more or less parallels thedecline in renal function, serum potassium risesabove normal limits only at the ultimate stage ofrenal failure [2—6]. Maintenance of a normal urinarypotassium excretion in the face of a falling glomeru-lar filtration [7—9] implies that a progressively larger

© 1977. by the International Society of Nephrology.

491

fraction of the filtered potassium appears in the urine(Fig. I).

It was initially thought that potassium excretiondepends upon glomerular filtration, hyperkalemiaappearing only when dietary intake exceeds the fil-tered load of potassium [4]. Leaf and Camara [8], incareful studies of four patients with advanced renalfailure, established that urinary potassium output ex-ceeds occasionally the filtered load of potassium, afinding taken to indicate that urinary potassium is, atleast in part, secreted by the tubular cells. Sub-sequently, it became clear that most of the urinarypotassium is derived from tubular secretion and notfrom glomerular filtration [10]. The adaptation pre-venting a rise of serum potassium concentration inrenal failure relies, thus, more on an enhanced tubu-lar secretion than on a decreased tubular reabsorp-tion of filtered potassium.

The intimate nature of the transport of potassiumby the nephron has been further defined during thelast decade. The present state of our knowledge isdiscussed in detail by Wright in this symposium [II].It has been shown that potassium reabsorption andsecretion vary from one nephron segment to anotherand depend on a variety of factors such as peritubularuptake of potassium mediated by the membrane so-dium-potassium activated adenosine triphosphatases(Na-K-ATPases), urine flow within the distal neph-ron, the transtubular potential difference generatedby sodium transport and potassium pumps locatedon the luminal side of tubular cells.

Localization and nature of enhanced renal excretionof potassium. Most of our knowledge on potassiumexcretion by the failing kidney has been obtained inthe experimental animal and, especially, in the sub-totally nephrectomized dog or rat. In the dog, with aremnant half kidney, Schultze et al [12] demonstratethat potassium excretion per nephron increases four-fold within 24 hr and six-fold within seven days afterremoval of the heterolateral kidney. After a week, 24-hr potassium excretion by the remnant kidneyreaches 87% of that achieved during the control pe-

492 van Ypersele de Strihou

asC0

xE

80C-

nod by the two kidneys (Fig. 2). Dogs maintainedconstantly on large doses of 9-cr-fluorohydrocorti-sone and adrenalectomized animals given a minimalmaintenance dose of DOCA undergo the same adap-tive changes after 3/4 nephrectomy. This observa-tion suggests that regulated changes in mineralo-corticoid hormone activity are not essential foradaptation to occur. Similarly, a persistent hyper-kalemia is not a prerequisite since serum potassiumincreases only by 0.8 mEq/liter 24 hr after nephrec-tomy and returns to control values by day seven.Modulations of distal sodium delivery after nephrec-tomy do not play a crucial role either: adaptation isthe same whether the animal is given a low or a highsalt intake, a finding subsequently confirmed in ratsby Espinel [13]; moreover, urinary potassium excre-tion is not modified by a reduction of distal sodiumdelivery obtained by renal artery constriction of theremnant kidney. These studies further demonstratethat increased potassium excretion per nephron isassociated not with a reduced but rather with anaugmented acid secretion per nephron: overall so-diuni cation exchange appears thus enhanced in theremnant kidney. It might have been anticipated thatthe failing kidney challenged by an acute potassiumload does not respond with as brisk a kaliuresis as anormal kidney. This is obviously not the case in theremnant kidney: hourly excretion of potassium after

a 25 mEq oral load is virtually the same as it was inthe control period.

In order to localize the sites of enhanced potassiumsecretion, Bank and Aynedjian [141 carry out micro-puncture and clearance experiments in rats in whichnephron reduction has been produced by 3/4 ne-phrectomy. Just as in the dog model, equilibriumbetween potassium intake and output is achievedwithin 24 hr after nephron reduction and maintainedfor up to two weeks. Micropunctures demonstratethat both 24 hr and seven days after nephrectomy,fractional reabsorption of potassium along the distaltubule is closely comparable in control and 3/4 ne-phrectomized rats: in neither case is there evidencefor tubular secretion of potassium. Approximately 4and 7% of the filtered potassium enter the collectingduct in the control and in the 3/4 nephrectomizedgroup, respectively, whereas final potassium excre-tion reaches 7 and 59% of the filtered load in the twogroups, respectively. Most of the adaptive increase inpotassium secretion occurs, thus, along the collectingduct. The reabsorptive pattern for potassium alongthe distal tubule remains unaffected by stimuli, suchas suppression of dietary potassium intake or potas-sium chloride infusion, resulting in a marked fall orrise in urinary potassium output. Although the inter-pretation of these results for the understanding ofdistal tubular handling of potassium may be debated[II], these data confirm that the collecting duct plays

200

160

120

80

40

Fig. 1. Relationship between renal filnction (expressed hi serumcreatinine concentration) and fractional excretion of potassium (cx—

pressed as percentage of filtered potassium eliminated in the urine).1 he heavy line represents the mean relationship between the twoparameters (percent of potassium excretion 8.6 X serum ereati-nine in mg per 100 ml ). The fine lines delineate the area coveredby the standard deviation of These data document alsopotassium secretion in severe renal failure, potassium excretionexceeding the hltered load when serum creatinine exceeds ¶0 mgper 100 ml (Reproduced from Platt 191.)

10

Serum creatinine, mg/lOO ml

Fig. 2. Adaptation in potassium excretion associated with nephronreduction in the dog. Potassium excretion is measured Iirst withboth intact (hatched bar) and remnant (R) kidneys contributing torenal function. The intact kidney is then removed and studies arerepeated. Potassium excretion by the remnant kidney increasesfour-fold within 18 hr and six-fold within seven days after nephrcc-tomy. Serum potassium (represented by closed circles) increasestransiently after the operation and returns to control values a weeklater. (Reproduced from Schultze et al [121.)

18 dogs)

Control Poflnephrectomy

40-

11

Intactkidney

5 t

4g0-

3

Potassium homeostasis in renal Jàilure 493

a central role in modulating urinary potassium excre-tion by the remnant kidney.

The mechanism responsible for collecting duct ad-aptation remains elusive. The micropuncture data areobtained from rats given large amounts of aldoste-rone. They confirm, thus, the conclusions of Schultzeet al [12] that modulation of aldosterone secretiondoes not play a fundamental role in the adaptiveprocess. Distal delivery of water and sodium to thecollecting duct is increased 24 hr after nephron reduc-tion and remains so, albeit to a lesser extent, twoweeks later. Although this augmented delivery of so-dium and water might enhance potassium secretionin the collecting duct, it is certainly not the solefactor: chronic 3/4 nephrectomized rats given a po-tassium-free diet do not exhibit collecting duct potas-sium secretion, despite similar increases in distaltubular flow.

Schultze et al [12] already pointed out the sim-ilarity between the adaptation occurring in remnantkidneys excreting a normal daily potassium intakeand that developing in normal kidneys challengedwith a sustained high potassium intake. This latterphenomenon is extensively discussed by Silva,Brown, and Epstein [15] in this symposium. Silva,Hayslett, and Epstein [161 had demonstrated thatkidneys adapted to an increased potassium intakehave an augmented Na-K-ATPase activity. This en-zyme is thought to be an essential element of theactive transport of sodium out of and potassium intothe cell interior. The hypothesis was therefore pro-posed that the increased potassium secretion noted inkidneys adapted to a high potassium intake resultedfrom augmented Na-K-ATPase activity. Schon,Silva, and Hayslett [17] explore the possibility thatnephron adaptation in the remnant kidney resultsfrom a similar change in enzymatic activity. Theystudy rats which have undergone, at least two weeksearlier, 3/4 nephrectomy. If the animals are main-tained on a normal potassium intake, their fractionalpotassium excretion increases markedly with a paral-lel augmentation of Na-K-ATPase activity in boththe outer medulla and the cortex. By contrast, if therats are given a reduced potassium intake, neitherfractional excretion of potassium nor cortical andmedullary Na-K-ATPase activity increase. The rise inNa-K-ATPase appears quite specific, as it is not ac-companied by any change in magnesium activatedadenosine triphosphatase (Mg-ATPase) or in 5'-nu-cleotidase activity. These data suggest that in renalfailure, just as in chronic potassium-loading, renaladaptation results from an enhanced Na-K-ATPaseactivity produced by an augmented urinary potas-sium load.

Finkelstein and Hayslett [18] further investigatethe role of the collecting ducts' secretion and aug-mented Na-K-ATPase in the adaptive process. Theycompare the renal handling of potassium in a groupof 2/3 nephrectomized rats and in a group of ratswith unilateral nephrectomy and papillectomy of theremnant kidney. Glomerular filtration is reduced by35% in both groups. Two weeks after the operationthe serum potassium concentration is significantlyhigher in the papillectomized rats than in the sub-totally nephrectomized and in the control animals.Challenged with an i.v. potassium load, the kidneysof the subtotally nephrectomized animals increasefractional excretion of potassium more than the pa-pillectomized and control rats. The ability to excretepotassium is correlated with the Na-K-ATPase con-tent of the outer medulla: enzyme activity is aug-mented by 25% in the subtotally nephrectomized butunmodified in the papillectomized group.

Other factors, however, could influence potassiumhomeostasis in renal failure. Natriuretic substanceshave been isolated in renal failure patients [19, 20],and it has been recently shown that they exert anaction on the collecting duct [21]. It will be of interestin the future to evaluate whether they intervene alsoin the adaptive increase in potassium secretion ofrenal disease.

The reported experimental observations demon-strate that subtotal nephrectomy is accompaniedwithin 24 hr by an augmented rate of potassiumexcretion per nephron and by an enhanced kaliuresisin response to a potassium load. This adaptivechange is located in the most distal parts of the neph-ron. Hyperkalemia, modulation ofaldosteronesecre-tion and distal sodium delivery do not play a criticalrole. By contrast, an increased activity of Na-K-ATP-ase in the medulla, induced perhaps by an aug-mented urinary potassium load, appears to be ofcritical importance.

The relevance of these data to the understanding ofthe adaptation developing in human renal diseaseremains to be determined. The model of subtotalnephrectomy may not be representative of humanchronic renal disease. It is characterized by an aug-mented single nephron glomerular filtration rate witha resulting increase in the distal delivery of sodiumand water. By contrast, in experimental glomerulo-nephritis [22, 23] and in toxic nephritis [24], singlenephron glomerular filtration rate is much more vari-able, ranging from strikingly reduced to elevated val-ues with parallel changes in distal filtrate delivery.The remaining nephron after subtotal nephrectomyundergoes an homogenous hypertrophy that isstrikingly different from that reported in glomerulo-

494 van Ypersele de Strihou

nephritis [22, 23], pyelonephritis [25], toxic renal dis-ease [24], and human renal disease [26].

It is likely that renal adaptation varies, to someextent, from one model to another. This is illustratedby Finkelstein and Hayslett's observation [18] thatthe renal handling of potassium and the changes inNa-K-ATPase are different if papillectomy is super-imposed on subtotal nephrectomy. Similarly, it hasbeen recently reported that in experimental gb-meruloncphritis Na-K-ATPase activity does not in-crease [27].

Role of the renin-angiotensin-aldosterone system.Aldosterone is a potent stimulus of tubular po-tassium secretion. The possibility therefore has beenconsidered that an augmented aldosterone secretionmight contribute to the maintenance of potassiumbalance in chronic renal failure.

As pointed out earlier, experimental evidence ob-tained with models of remnant kidney does not sug-gest that aldosterone plays such a role [12, 14]. Fur-thermore, aldosterone production by adrenal tissuemeasured in vitro is the same in subtotally nephrecto-

I

Sodium intake

mized and control rats [28]. In human renal disease,increased aldosterone production does not seem to bea critical element of the adaptive process. In 1963,Cope and Pearson [29] report an increased aldoste-rone secretion rate in 6 out of 12 uremic patients.Their results, however, rely on several assumptionsabout aldosterone metabolisni and are therefore dif-ficult to interpret. More recently, Schrier and Regal[30] study nine patients with stable renal failure(mean creatinine clearance, 28 mI/mm). Aldosteroneexcretion rate is normal if salt intake is adequate andincreases of sodium is restricted (Fig. 3). Similarly,Weidman, Maxwell, and Lupu [31] and Weidman etal [32] provide evidence that in patients with terminalrenal failure, plasma aldosterone concentration re-mains normal as long as plasma renin activity andserum potassium concentrations are within normallimits. Increased serum potassium and/or plasma re-nm levels, by contrast, are associated with elevatedplasma aldosteronc levels [3 1—33].

Although increased plasma aldosterone levels maycontribute to potassium homeostasis in renal failure,these observations suggest that in human disease, justas in the experimental animal, increased aldosteroneproduction is not indispensable for the maintenanceof potassium equilibrium in uremia. The demonstra-tion that spironolactone given to patients with mildrenal failure [30] reduces potassium secretion sug-gests, however, that normal levels of aldosterone arerequired to avoid hyperkalemia. This conclusion issupported by the observation of sustained hyper-kalemia when hypoaldosteronism complicates thecourse of renal failure [34].

Gastrointestinal excretion n/potassium

In normal subjects, potassium balance is main-tained mainly by the kidney. The amount of potas-sium lost in the feces represents a small proportion ofthe dietary intake and is independant of its magni-tude.

The digestive tract remains, however, a potentiallylarge source of potassium loss, as illustrated by thepotassium deficits produced by protracted diarrhea.The possibility, therefore, has been considered that inchronic renal failure, fecal excretion of potassiummight contribute significantly to potassium home-ostasis.

Hayes et al [35] perform balance studies in 21severely u remic patients (creatinine clearance, below30 mI/mm), eight of whom are maintained on inter-mittent hemodialysis. In control subjects on a normaldaily potassium intake (47 to 100 mEq/day), fecalpotassium excretion is less than 20% of the intake(mean, 12%). By contrast, in patients whose creati-

Normal Low

0

Fig. 3. Fileci of prolonged sodium restriction on aldosterone excre-tion rate in nine patients with chronic renal disease (mean GFR. 28nil nun 1. Aldosterone excretion was not increased in any of thesestudies during the normal sodium intake (2 mEg/kg of body wt perday for at least one week) but increased in each patient duringsodium restriction (10 to 21 mEg/day for 8 to 25 days). (Repro-duced from Schrier and Regal [28].)

Potassium h omeostasis in renal Jilure 495

nine clearance is below 5 mI/mm, stool potassiumoutput exceeds 20% of the intake (mean, 34%; range 7to 76%). Intermediate values (mean, 18%; range, 15to 24%) are noted when creatinine clearance rangesfrom 10 to 30 mI/mm (Fig. 4). No difference is ob-served between patients receiving a low or a normalsodium intake. Fecal potassium output is related tostool wet weight, although none of the patients havediarrhea. These abnormalities are related to renalfailure per se, as they disappear after renal trans-plantation. Similar observations of increased stoolpotassium have been reported in patients with renalfailure [30, 36] although not constantly [37].

The understanding of the mechanisms responsiblefor enhanced excretion of potassium by the gut ishampered by the fact that concentration of elec-trolytes in digestive fluids remains difficult to mea-sure. Wrong et al [38] have developed an ingeniousmethod to obtain samples of the extracellular phaseof stool water. Cellophane capsules containing dcx-tran are ingested by patients and collected sub-sequently from their feces, filled with fluid whosecomposition reflects that of stool water. Wilson et al[39] utilize this method in 30 patients with acute orchronic renal failure. Fecal dialysate composition dif-fers markedly from control values: sodium concen-tration falls from a mean of 19 mEq/liter in normalsubjects to 5 mEq/liter in uremics, whereas potas-sium concentration rises from 77 to 125 mEq/literwith a resulting fall in the sodium to potassium ratio.

Whereas Wilson et al [39] do not provide estimatesof glomerular filtration rate (GFR), it appears quiteclearly from the data of Hayes et al [35] (Fig. 4) thatfecal potassium output increases markedly only whencreatinine clearance drops below 5 ml/min. Even atthat stage the vicariant potassium excretion is notvery important (an average of 23 mEq/day for thepatients of Hayes et al and an average of 9 mEq/dayfor the patients of Wilson et al). This amount might,however, become critical at the very final stages ofrenal failure when potassium intake is severely cur-tailed, as might be the case in patients on a Giordano-Giovannetti diet.

The experimental models used to study potassiummetabolism in renal failure do not have the markedlylowered glomerular filtrations at which changes ingut potassium excretion are observed in humans. Itis, therefore, not surprising that in dogs [12] whoseGFR has been reduced to 25% of normal, 93% of theintake is excreted in the urine, just as in the controlperiod. Similarly, in rats with a GFR averaging 40%of normal [14], 92% of the potassium intake is elimi-nated by the kidney, versus 96% in the control ani-m al s.

60.9 a

.5

E2 2 40

0.-c—'a

ss 20

PATIENTS Dietary Na

Low

* Intake

High

CRF (dialysis)

CRF (consv. Px)

Normal function

• A

a

o

0 5 10 15 20 25 " 80 100

Creatinine clearance, rn//rn in

Fig. 4.Re/ationship between renal Junction (expressed by creatinineclearance) and stool potassium excretion (expressed as percentage n-c/al/v potassium intake) in 23 patients with normal and reduced renalfunction (66 balance periods). (CRE = chronic renal Jailure). Potas-sium intake ranged from 47 to 100 mEq/day. It is noteworthy that,independently of the sodium intake, stool potassium increasesmarkedly when creatinine clearance falls below 5 mI/mm. (Repro-duced from Hayes et al [35].)

The mechanisms responsible for increased stoolpotassium have not been completely elucidated. Therelationship noted between fecal potassium outputand stool weight [35] suggests that potassium reab-sorption from the gut is decreased as a consequenceof a diminished intestinal water reabsorption. Alter-natively, the possibility has been considered that thehigh stool potassium results from hyperaldosteron-ism. The decreased sodium to potassium ratio foundin uremic stool dialysate [39] is analogous to thatobserved in normal subjects given mineralocorticoidsor patients with primary hyperaldosteronism [40].Moreover, it is known that aldosterone stimulatespotassium secretion along the colon [41, 42].

This hypothesis remains open for discussion. Aspointed out earlier, there is no evidence that renalfailure is associated per se with an increased aldoste-rone production. Hyperaldosteronism, is not unusualhowever, in renal failure as a result of an aug-mented renin level, an elevated serum potassium con-centration, or a decreased extracellular volume [30,32]. None of the studies reporting increased stoolpotassium provide information on aldosterone levels.Furthermore, the results of spironolactone adminis-tration on stool electrolytes remain equivocal. Hayeset al [35] report no effect of 100mg of spironolactonegiven to three uremic subjects, a dose that might beinsufficient to inhibit the peripheral effects of aldoste-rone. Wilson et al [39], on the other hand, note areversal of the sodium : potassium of stool dialysate in

8C

.

£

- . if.

0. A 0

0 0SAo 0.0

I I I I I It I I I

496 van Ypersele de Strihou

two patients given 300 mg of spironolactone. Resultshowever, are striking in only one case and debatablein the other.

Finally, by analogy with the adaptive mechanismsdeveloping in animals fed a high potassium diet, itmay be suggested that the increased stool potassiumobserved in renal failure results from an augmentedlevel of Na-K-ATPase in the colonic mucosa. It hasbeen demonstrated that chronic increases in potas-sium intake in normal individuals result in aug-mented fecal output of potassium and that this phe-nomenon is associated with a raised level of Na-K-ATPasc in the colon [431.

('el/u/ar tolerance to potassium loads

When given to normal subjects, a potassium load,to a large extent, is taken up initially by the cells andreleased slowly thereafter as renal excretion increases.

In animals adapted to a high potassium intake, thiscellular uptake is enhanced so that the changes inserum potassium produced by an acute potassiumload are minimized [15, 44]. It might have been antic-ipated that the adaptive increment in potassium ex-cretion observed both in the uremic patients and inexperimental renal failure would be associated with aparallel enhancement of the extrarenal uptake of p0-t a ssi Um.

In uremic subjects, there is no evidence for such anextrarenal adaptation. Winkler, Hoff, and Smith [7}challenge five uremic subjects with 100 mEq of potas-sium. The resulting increase in serum potassium islarger than that noted in a normal subject. Keith andOsterberg [45] obtain similar results in ten patientswith renal failure who were given 5 g of potassiumbicarbonate. The final concentrations of serum potas-sium are substantially higher in uremic than in con-trol subjects. In these early studies, neither basal po-tassium intake nor the degree of renal failure areaccurately quantitated. More recently, however,Gonick et al [46] have confirmed these findings: cellu-lar potassium tolerance is not increased but rather isdecreased in renal failure.

This conclusion applies also to the animal withexperimental renal insufficiency. Schon, Silva, andHayslett [17] inject an i.v. load of potassium to sub-totally nephrectomized rats. The rise in the serumpotassium (+2.8 mEq/liter) is significantly largerthan that observed in rats adapted to a high potas-sium intake (+0.9 mEq/liter) but not significantlydifferent from that noted in normal rats (+ 1.8mEq/Iiter). These results are confirmed by Finkel-stein and l-Iayslett [18].

These data demonstrate a significant difference be-tween the adaptative mechanisms associated withchronic potassium loads and those observed in renal

failure. Extrarenal adaptation occurs only in the for-mer situation, whereas renal adaptation can be dem-onstrated in both models.

Body potassium stores in renal disease

The inability of the kidney to eliminate adequatelypotassium may lead to an overall increase in bodypotassium stores. On the other hand, the dietary re-strictions associated with increasingly severe renalfailure, the loss of appetite, nausea, vomiting, anddiarrhea may induce weight loss, malnutrition, anddecreases of body potassium stores.

Several approaches have been used to ascertainpotassium stores in patients with renal failure: deter-mination of exchangeable potassium, measurementof total body potassium with a whole body scintilla-tion counter, and tissue analysis. Elsewhere in thissymposium, Patrick [47] reviews these methods, theirlimitations, and the pitfalls encountered in the inter-pretation of their results.

I) Exchangeable potassium (Ke). In patients withadvanced chronic renal failure, variable results havebeen obtained. Adesman et al [48] report a decreasedKe per kg of body wt in three out of five uremicpatients, a finding confirmed by Spergel et al [491 inan additional six patients. Patrick et al [50] find areduced Ke in six of nine patients with terminaluremia, some of whom suffered from diarrhea, vom-iting or having received diuretics. Similar observa-tions are made by Bilbrey et al [51] in nine patientsmaintained on a low protein diet. By contrast, Ber-lyne, van Laethem, and Ben An [361 select carefully13 severely uremic patients (mean creatinine clear-ance, 4.1 mI/mm) without gastrointestinal distur-bances, diuretic therapy, or cation exchange resins. Inthis group, Ke is within normal limits.

Similar discrepancies are noted in patients treatedby chronic hemodialysis. Seedat [52] observes a 20 to35% decrease of Ke in five out of seven patientsdialyzed for 14 to 42 hr per week with a solutioncontaining I mEq/liter of potassium. Ke falls pro-gressively with the duration of the treatment. By con-trast, Ram and Chisholm [53] note that Ke is initiallylow in seven patients but returns towards normalover the subsequent years, provided dialysis and pro-tein intake are adequate. Rettori et al [54] exploreseven male patients dialyzed with a solution contain-ing either I mEq/liter or no potassium. Ke is lowerthan in control subjects, whether expressed per kg ofbody wt, per kg of lean body mass, or per liter ofintracellular water.

Although the impression may be gained that thesedata demonstrate potassium depletion in renal fail-ure, several factors render their evaluation difficult.First, the patient groups are very different from one

Potassium homeosiasis in rena/failure 497

another. Some studies [50] include debilitated pa-tients with digestive problems, whereas others [36]exclude them. In patients treated by chronic hemo-dialysis, dialysate potassium content, dialyzer effi-ciency, and duration of dialysis vary from one studyto another. Second, as emphasized by Patrick [47],normal values of exchangeable potassium are notwell defined. These values are usually derived fromthe relationship existing between Ke and body wt orlean body mass (as evaluated from total body water)[55]. This relationship varies with sex and age and islikely to be distorted in renal failure, a conditioncommonly associated with increased extracellularfluid volume, even in the absence of edema [56]. Eventhe use of intracellular water as a reference standard[54] is questionable as it relies on the identity ofsulfate space and extracellular water, an assumptionwhich has not been validated in severe uremia. Thelast difficulty in the assessment of Ke in uremic pa-tients stems from the observation that the 24 hr,allowed in most studies for radioactive potassium toequilibrate with native potassium, may be in-sufficient. Boddy et al [57] measure simultaneouslytotal body potassium with a whole body counter andKe after equilibration periods of 24, 48, and 64 hr in12 patients with stable chronic renal failure. Theyobtain increasing values of Ke, expressed as a per-centage of total body potassium measured by wholebody counting: 60.7, 83.6, and 85.9% at 24,48, and 64hr, respectively. They conclude that most studies us-ing potassium 42 (42K) and a 24-hr equilibrationperiod underevaluate exchangeable potassium.

2) Total body potassium. In order to obviate theproblems inherent to delayed equilibration of radio-active potassium, other investigators have resorted tothe measurement of the natural isotope potassium 40(40K) by whole body counting techniques.

In stable chronic renal failure, total body potas-sium remains within values predicted on the basis ofsex, age, height, and weight in 13 out of 15 uremicpatients (GFR ranging from 5 to 40 mI/mm) main-tained on a low protein diet [57] and in eight severelyuremic patients (mean GFR, 8 mI/mm) studied byLitteri et al [58]. In dialyzed patients, Morgan et al[59] find no evidence of a total body potassium reduc-tion in 21 subjects dialyzed against a bath containing1.5 mEq/liter, whereas Johny et al [60] report signifi-cantly reduced values in 5 out of 1 5 patients dialyzedagainst a dialysate containing I to 1.4 mEq/liter.Boddy et al [57] publish similar results: 4 of 18 pa-tients dialyzed with a bath containing 1 mEq/liter ofpotassium have a total body potassium below pre-dicted values.

These data suggest that occasionally whole bodypotassium is reduced. Whether this decrease repre-

sents a true deficiency or reflects merely a decreasedcellular mass resulting from malnutrition remains tobe established.

3) Cellular potassium content: a) Muscle ce/I analy-sis. Bittar et al [61] report a normal cellular po-tassium in two out of three muscle biopsies ob-tained from severely uremic patients. Bergstrom[62] observes a normal muscle cell potassium con-tent in 13 out of 15 uremic patients (serum creati-nine level, above 6 mg/lOO ml), an observationconfirmed by Graham, Lawson, and Linton [63] inmuscle biopsies obtained from 22 patients with renalinsufficiency (GFR ranging from 3 to 20 mI/mm).

In these studies intracellular potassium is expressedas mEq per liter of intracellular water, a parametercalculated as the difference between measured totalwater and calculated extracellular water. The latterquantity is assumed to equal chloride space. Sincechloride is supposed to be passively distributed acrossthe polarized cell membrane, this assumption is validonly if a normal potential difference is maintainedacross the cell membrane. Cunningham et al [641have demonstrated that this assumption is not validin severely ill patients, including uremic subjects: un-der these circumstances, the transmembrane poten-tial difference falls. Bilbrey et al [51] confirm thisfinding in patients with renal failure and use themeasured potential difference to calculate chloridedistribution across the cell membrane and, thus, theextracellular water content. They report that in ninepatients, muscle cell potassium, expressed either perliter of intracellular water or per 100 g of fat-freesolids, is lower than normal. Interestingly, thesevalues return to normal after initiation of dialysis.

These studies are based on biopsy material. But-kus, Alfrey, and Miller [65], by contrast, analyzetissue samples obtained at autopsy in 24 patients whohad been treated by hemodialysis. Cellular potassiumcontent of muscle or heart is significantly below con-trol values whether potassium concentration is ex-pressed per kilo of fat-free dry solids or per g ofnitrogen. Potassium content is more than two SDbelow mean control levels in almost two-thirds of thepatients. The influence of dialysate composition oncellular potassium stores is evidenced by the fact thatmuscle potassium concentration is significantly lowerin subjects dialyzed with a potassium-free dialysatethan in those treated with a solution containing 2 to2.6 mEq/liter.

b) Leukocyle composition. Patrick et al [50] andPatrick and Jones [66] find that leukocyte potassiumcontent, expressed as mmoles per kg, is more than 1SD below control mean values in 9 out of 16 patientssuffering from chronic renal failure.

This abnormality is fully corrected by chronic

498 van Ypersele de Sirihou

hemodialysis. It is of interest to note that the samegroup has recently reported [67] a decreased sodiumefflux in uremic leukocytes, an abnormality also par-tially corrected by hemodialysis. This finding is com-patible with the hypothesis that the low leukocytepotassium content reflects an abnormality of themembrane ionic pump, rather than a true deficit cor-rectable by an increased potassium intake.

c) Red blood cell composition. The erythrocyte is ahighly specialized cell which may not be representa-tive of cells at large. It is nevertheless noteworthy thatKramer, Gospodinov, and Kruck [68] observe alowered red blood cell potassium content in 2 out of13 patients. They also note, in several of their sub-jects, an increased red blood cell sodium contentassociated with a decreased sodium efflux, an abnor-mality already reported by Welt, Sachs, andMcManus [69]. This disorder has been ascribed to areduced activity of the membrane ouabain-sensitiveadenosine triphosphatases [70, 711 resulting from aninhibitory factor present in uremic plasma. Indeed,changes in red blood cell membrane enzymes canbe reproduced in normal red blood cells incubated inuremic plasma [721. They disappear after hemodialy-sis [69] or renal transplantation [73]. The hypothesisthat uremic toxins alter not only sodium hut alsopotassium transport by the red blood cell, andreduce occasionally red blood cell potassium content,remains to be tested.

Johny et al [60] suggest that erythrocyte potassiumis a useful index of the direction and the magnitude ofchanges in total body potassium. In 12 dialyzed pa-tients, they observe a significant correlation betweenred blood cell potassium and either total body potas-sium or dialysis-induced changes in potassium bal-ance, From their table, however, it appears that theirconclusions rest on a small number of patients: redblood cell potassium is reduced in only 4 of their 12patients, and total body potassium is decreased inonly three of them. Rettori et al [541, by contrast, donot observe lowered red blood cell potassium contentin their dialyzed patients and find no correlation be-tween exchangeable potassium and red blood cellpotassium content.

4) Conclusion. Evaluation of potassium stores isclearly a difficult task. As pointed out by Patrick [47]in this symposium, reduction of body potassium maybe encountered in three different conditions [74],each of which occurs in uremic subjects. The first,termed "potassium deficiency," refers to a cellularloss of potassium, readily corrected by potassiumsupplements. In the second, called "potassium deple-tion," lowered cellular potassium stores cannot becorrected by potassium supplements. In this state,

disturbed metabolic processes impair the cells' abilityto maintain a normal potassium content. The lastcondition, "pseudo-depletion," is observed when to-tal muscle mass is reduced: cellular potassium con-centration is normal, and potassium supplements donot return total body potassium to normal.

Nausea, vomiting, and steroid therapy in terminalrenal failure are probably associated, in some pa-tients, with true potassium deficiency [50]. The cureof digestive symptoms by appropriate dialysis and anormal dietary potassium readily return cellular po-tassium to normal. Potassium deficiency may alsodevelop when a protracted low potassium intake iscombined with prolonged hemodialysis with low po-tassium bath [651. The incidence of this condition inrenal failure with or without dialysis cannot be eval-uated for lack of balance data during potassium sup-plementation.

The existence of potassium depletion in terminalrenal failure is suggested by the observations made onred and white blood cells [67, 70, 711. The low cellularpotassium content might reflect the reduced mem-brane Na-K-ATPase activity engendered by the accu-mulation of toxic molecules in uremic serum. Theselatter molecules might play a similar role in musclecells, reducing the transmembrane potential differ-ence [511 and decreasing cellular potassium content.Hemodialysis could improve cellular potassiumstores not only by allowing an augmented dietaryintake of' potassium hut also by enhancing cellularpotassium uptake [50, 51, 69].

Finally, pseudo-depletion is probably present in alarge number of patients maintained for prolongedperiods of time on a low protein diet with a resultingmalnutrition.

These different types of lowered potassium reservesdo probably coexist very often. The intricacies ofclinical settings combining inappropriate intake andloss of potassium, impaired cellular potassium up-take and decreased muscle mass, taken together withthe uncertainties of the diverse methods assessingpotassium stores, easily account for the apparentlycontradictory results reported in the literature. Atany rate, all studies agree that serum potassium is notcorrelated with potassium stores: this fact underlinesonce more how little serum potassium reflects theactual state of total body potassium.

The clinical consequences of reduced body potas-sium reserves have not been clearly defined. Alteredglucose tolerance [49] and disturbed cardiac function[651 are likely complications of potassium deficiency.Pseudo-depletion on the other hand, is not expectedto lead to such disorders, whereas the effect of potas-sium depletion remains to be demonstrated.

Potassium homeostasis in renal failure 499

Factors interJring with the adapt ative mechanismsregulating serum potassium in renal Jailure

The development of hyperkalemia during the ulti-mate phase of renal failure signifies that despite itsadaption, renal function fails to maintain a normalpotassium excretion.

Several events may precipitate hyperkalemia be-fore these final limits are reached. We shall considerthem briefly.

A Iterations of the intra-extracellular distribution ofpotassium: I) Acid-base disturbances. Burnell et al.[751 have demonstrated that in man, changes inacid-base balance alter serum potassium concentra-tions independently of total body potassium: acidosisincreases, whereas alkalosis decreases serum potas-sium, every 0.1 U change in extracellular pH elicitingan inverse 0.6 niEq/liter change in serum potassiumconcentration. These alterations in serum potassiumconcentration depend solely on extracellular pHmodifications [76]. They result from a redistributionof potassium across the cellular membrane. Thisphenomenon is readily observed in patients withrenal failure [77] who are prone to develop metabolicacidosis. Correction of acidosis may return to normalan alarmingly high serum potassium concentration[78, 79].

2) Arginine. A similar redistribution of potassiumbetween extracellular and intracellular spaces hasbeen recently invoked to account for the hyperka-lemia elicited in uremic patients by the infusion ofarginine hydrochloride [80]. Arginine, just as othercationic amino acids, penetrates cells and displacespotassium [81] towards the extracellular space. Theresulting hyperkalemia is mild if renal function isadequate but potentially lethal in severe renal failure.

3) Serum hyperosmolality. Acute increases in serumosmolality may also provoke a redistribution of po-tassium across cell membranes. Seldin and Tarail [82]have demonstrated that sudden augmentation of se-rum glucose is accompanied by an outward move-ment of cellular potassium and water. Makoff et al[83, 84] report similar observations in dogs whoseextracellular osmotic pressure has been augmentedby mannitol. In isolated muscle cells, elevation ofbath medium glucose content reduces intracellularpotassium [85]. In diabetic patients with a failingrenal function, sudden hyperglycemia may lead tosevere hyperkalemia [86].

Alterations of renal adaptation: 1) Aldosterone. Asalready pointed out, there is no evidence that in-creased aldosterone secretion is part of the adaptionrequired to maintain potassium homeostasis inchronic renal failure.

Recently, however, it has become evident that anormal plasma level of aldosterone is indispensablefor the maintenance of a normal serum potassiumlevel in patients with chronic renal disease. In 1968,Gerstein et al [34], reported the occurrence of severehyperkalemia, sodium wasting, and an inability toacidify the urines in a patient with non-oliguric renalfailure. Hypoaldosteronism is documented. Replace-ment therapy with mineralocorticoids corrects theelectrolyte and acid-base disorders. Several similarpatients have been reported since, all of whom havein common sustained hyperkalemia despite a normalpotassium intake and a moderate degree of renalfailure [87—92]. Interestingly, most of these patientsare above 50 years in age, and a large proportion ofthem suffer from diabetes [87]. Although plasma re-nm levels were initially reported to be normal, itbecame rapidly apparent that hypoaldosteronism iscommonly associated with hyporeninemia [87—90].

The observation that replacement therapy withmineralocorticoids promptly increases kaliuresis andreturns serum potassium to normal suggests stronglythat hyperkalemia is caused by the deficient aldoste-rone secretion. Several factors have been consideredto account for hypoaldosteronism. A selective defi-ciency in the response of aldosterone secretion toserum potassium concentrations seems unlikely. Al-though potassium is a direct stimulus of aldosteronesecretion [93—95], there is good evidence that aldoste-ronc responsiveness to a variety of other more potentstimuli is also decreased or abolished: adrenocorti-cotropic hormone [87—91, 96], angiotensin II [34,87—89, 91, 92], catecholamines [87, 88], or a salt-freediet [34, 86, 87, 89—91].

Alterations in the biosynthesis of aldosterone isalso improbable. Although such a defect has beenrecently documented in two cases of hypoaldosteron-ism associated with renal disease [96], it is absent inmost others [87, 88, 90]. Furthermore, such enzymedefects are usually associated with elevated ratherthan lowered renin levels [97].

The most likely cause of hypoaldosteronism seemsto be the lowered serum renin level with a resultantdecrease in angiotensin II generation, a situation alsoencountered in anephric patients without measurableplasma renin levels [31, 98, 99]. In the case of hypo-aldosteronism, response of plasma renin levels toseveral stimuli, such as upright posture [87, 89, 901,catecholamine infusion [87, 88], furosemide injection[89—91], or salt-free diet [87—91] is blunted or absent.

The cause of the lowered plasma renin level has notbeen completely elucidated. Hyperkalemia, a mildsuppressor of renin secretion [100, 101], does notappear to play a significant role, since correction of

500 van Ypersele de Strihou

hyperkalemia with a cation exchange resin [88, 90]does not elevate the plasma renin level. The alteredrelease of renin may result from specific alterations ofthe j uxtaglomerular apparatus developing during thecourse of renal disease. This hypothesis is compatiblewith the observations [102, 1031 that the plasma reninlevel and its responsiveness to a low salt diet varyaccording to the type of renal disease, being lower inpyelonephritis and polycystic kidney disease than inglomerulonephritis. Another explanation has beenproposed more recently by Dc Leiva et al [961. Theyobserve two diabetic patients with mild renal failureand hyporeninemic hypoaldosteronism and note thatthe lowered plasma renin levels are associated with anelevated level of the renin precursor prorenin. Theypropose that, in these cases, hyporeninemia resultsfrom an impaired conversion of prorenin into renin.Whether this attractive hypothesis applies to all casesof hyporeninemic hypoaldosteronism remains to bedetermined. It is noteworthy, however, that increasedlevels of prorenin are not unusual in diabetic nephro-pathy and chronic renal disease [104], a finding com-patible with the hypothesis that a defective transfor-mation of prorenin is the first sign of hyporeninemichypoaldosteronism.

As an alternative explanation for hyporeninemia,Oh et al [91] have suggested that it is not a pathologicfeature but rather reflects a physiologic suppressionof the enzyme by an augmented extracellular volume.In support of their hypothesis, they draw attention tothe fact that most reported cases of hyporeninemichypoaldosteronism are hypertensive and that, in twoof their own patients, continued administration offurosemide increases both plasma renin and aldoste-rone. The effect of these maneuvers on serum potas-sium is not reported. This explanation appears diffi-cult to reconcile with the general observation thatpatients with hypoaldosteronism have a tendency tobe salt wasters and that sustained salt restrictionleads usually to a further increase in serum potas-5! U Ill.

2) Insulin. The role of insulin in potassium home-ostasis has been recently recognized [105—109]. In-creases in serum potassium trigger the release of in-sulin by the pancreas [108—1101, whereas in turn,insulin stimulates cellular potassium uptake[111—113]. A deficiency of this feedback mechanismin the presence of an impaired renal excretion ofpotassium might produce hyperkalemia. This hy-pothesis could account for the susceptibility to hy-perkalemia of diabetic patients receiving potassium-sparing diurectics [114—115], aldosterone antagonist[116], or suffering from hypoaldosteronism [86] andrenal failure (VAN YPERSELF. i)E STRII-lou, VAN-DENBROUCKE, unpublished data).

3) Tubular resistance of renal tubules to miner-alocorticoid action. Hypoaldosteronism or its associa-tion with diabetes probably accounts for several casesof unexplained hyperkalemia observed in non-oh-guric renal disease [117]. In some other patients, hy-perkalemia results apparently from a tubular unre-sponsiveness to mineralocorticoids. Luke, Allison,and Davidson [118] describe hyperkalemia in a pa-tient with renal failure due to amyloidosis. The al-tered renal excretion of potassium is not corrected bymineralocorticoid therapy. Similarly, Popovizer et al[119] observed a patient with salt-wasting neph-ropathy and hyperkalemia, despite adequate aldoste-rone secretion. Resistance of the tubules to the pe-ripheral effects of mineralocorticoids is furtherdocumented by the failure of DOCA injections toaugment urinary potassium output.

4) Spironolactone. potassium-sparing diuretics. In-hibition of distal potassium secretion as a result ei-ther of aldosterone antagonists or of potassium-spar-ing diuretics may dramatically impair potassiumexcretion in renal failure and produce severe hy-perkalemia [34, 116, 120].

Increased loads oJ potassium. As stated earlier, thetolerance to exogenous potassium loads is decreasedin renal failure as a result of a reduced extrarenaluptake of potassium and an impaired potassium ex-cretion. Sudden increases in potassium load, there-fore, may lead to hyperkalemia and, occasionally,cardiac arrest.

1) Exogenous potassium loads. Potassium supple-ments prescribed in association with diuretics, potas-sium salts used as substitutes for sodium chloride insodium-free diets, consumption of low sodium (highpotassium) milk [121] may all precipitate lethal hy-perkalemia.

2) Endogenous potassium loads. Catabolism asso-ciated with trauma, infections, or high fever mayrelease massive amounts of cellular potassium andproduce potassium intoxication. The catabolism of100 g of muscle protein liberates 44 mEq of potas-sium [122]. A massive release of cellular potassium isalso the cause of the hyperkalemia encountered occa-sionally after succinyichohine administration to pa-tients with neuromuscular disorders [123].

Conclusions

Potassium homeostasis is remarkably maintainedthroughout the evolution of renal disease up to theultimate stages of renal insufficiency. The kidney re-mains the main organ regulating potassium balanceand serum potassium concentrations, despite lossesof up to 80 to 90% of its mass.

Potassium excretion is maintained through an aug-

Potassium homeostasis in renal failure 501

mented tubular secretion, localized, at least in someexperimental models, in the most distal parts of thenephron where it is associated with an augmentedactivity of Na-K-ATPase. Enhanced tubular secre-tion of potassium is certainly improved by increasedlevels of aldosterone resulting from hyperkalemia,dehydration, or increased renin levels, although anaugmented rate of aldosterone secretion is neither aprerequisite for potassium homeostasis to be main-tained nor a necessary element of renal failure.

In severe renal failure, gastrointestinal excretion ofpotassium may increase and contribute to the mainte-nance of potassium balance. Its mechanism is not yetclear, and more data will be necessary to evaluate therole of aldosterone, colonic enzyme activity, etc.

Adaptation of renal and digestive excretion of po-tassium are not accompanied by a parallel improve-ment in cellular uptake. On the contrary, cellulartolerance to potassium loads is decreased. As a result,increases in potassium intake may produce lethal lev-els of serum potassium.

Overall balance of potassium in renal failure variesmarkedly from patient to patient and is difficult toassess by currently available methods. Decreaseddietary intake and digestive intolerance to potassium,increased loss due to dialysis with low potassiumbaths may lead to potassium deficiency that can berepaired by appropriate maneuvers. Uremic serummight also contain toxic molecules capable of inhib-iting electrolyte transport by the cellular membrane.Consequently, cellular potassium might drop and re-sult in a potassium depletion resistant to an aug-mented potassium intake. Finally, malnutrition mayso reduce muscle cell mass that total body potassiumfalls and produces pseudo-depletion.

The delicate mechanisms maintaining potassiumhomeostasis in renal failure may be disrupted byseveral factors and lead to hyperkalemia. Thegradient between intracellular and extracellular po-tassium is modified by acidosis, cationic amino acids,and sudden increases in plasma osmolality. Absenceof aldosterone, resistance of renal tubules to mm-eralocorticoids, inhibition of tubular transport ofpotassium by spironolactone and potassium-sparingdiuretics, may jeopardize the adaptive increases intubular potassium secretion. Lack of insulin asso-ciated with an impaired renal excretion of potassiummay diminish the cellular tolerance to potassiumloads and result in brisk hyperkalemia. Finally, anincreased endogenous or exogenous load of potas-sium may precipitate potassium intoxication.

Reprint requests to Dr. C'. ran Ypersele de Strihou, Renal Labora-/0ev, Department 01 Medicine, Cliniques Universitaires St- Pierre,Louvain, Belgium.

References

I. BLAKEJ: Ed/n Med Surgf5l:330, 18392. KEITH NM, BURCHELI, HB:Clinical intoxication with potas-

sium: Its occurrence in severe renal insufficiency. Ani J MedSd 217:1—12, 1949

3. KEITH NM, KING HE. OSTERBERG AE: Serum concentrationand renal clearance of potassium in severe renal insufficiencyin man. Arch Intern Med 71:675—701, 1943

4. ELKINTON JR, TARAII, R, PETERS JP: Transfers of potassiumin renal insufficiency. J C/in Invest 28:378—388, 1949

5. SCIIWARTZ WB, POI.AK A: Electrolyte disorders in chronicrenal disease. J C'hrnnic Dis 11:319—339, 1960

6. KI.EEMAN CR, OKUN R, HEILER RJ: The renal regulation ofsodium and potassium in patients with chronic renal failure(CRF) and the effect of diuretics on the excretion of theseions. Ann NY Acad Sd 139:520—539, 1966

7. WINKLER AW, HOFF HE, SMITH PK: The toxicity of orallyadministered potassium salts in renal insufficiency. J C/inInvest 20: 119—126, 1941

8. LEAF A, CAMARA AA: Renal tubular secretion of potassiumin man. J C/in Invest 28:1526—1533, 1949

9. PIATT R: Sodium and potassium excretion in chronic renalfailure. C/in Sci 9:367—377, 1950

10. BERLINER RW: Renal mechanisms for potassium excretion.Harvey Lect 55:141—171, 1961

II. WRIGHT ES: Sites and mechanisms of potassium transportalong the renal tubule. Kidney tnt 11:415—432, 1977, thisissue

12, SCHULTZE RG, TAGGART DD, SHAPIRO H, PENNELL JP,CAGI.AR S., BRICKFR NS: On the adaptation in potassiumexcretion associated with nephron reduction in the dog. IC/in Invest 50:1061—1068, 1971

13. ESPINEL CH: Effect of proportional reduction of sodiumintake on the adaptive increase in glomerular filtrationrate/nephron and potassium and phosphate excretion inchronic renal failure in the rat. C/in Sci 49:193—200, 1975

14. BANK N, AYNEDJIAN HS: A micropuncture study of potas-sium excretion by the remnant kidney. I C/in Invest52:1480—1490, 1973

IS. SILVA P, BROWN RS, EPSTEIN FH: Adaptation to potassium.Kidnei' hit 11:466—475, 1977, this issue

16. SILVA P, HAVSLFTT JP, EPSTEIN FH: The role of Na-K-activated adenosine triphosphatase in potassium adaptation.I C'lin invest 52:2665—2671, 1973

17. SCHON HA, SILVA P. HAYSLETT JP: Mechanism of potassiumexcretion in renal insufficiency. Am I Physiol 227:1323—1330,1974

18. FINKELSTFIN FO, HAYSLETT JP: Role of medullary structuresin the functional adaptation of renal insufficiency. Kidney Int6:419—425. 1974

19. BOURGOIGNIE JJ, HWANG KH, ESI'INEL C, KI.AHR 5,BRICKFR NS: A natriuretic factor in the serum of patientswith chronic uremia J C/in Invest 51:1514—1527, 1972

20. BOURGOIGNIE JJ, HWANG KH, IPAKCHI E, BRICKER NS: Thepresence of a natriuretic factor in urine of patients withchronic uremia: The absence of the factor in nephroticuremic patients, J Cl/n Invest 53:1559—1567, 1974

21. FINE, LG, B0URGOIGNIE JJ, HWANG KH, BRICKER NS: Onthe influence of the natriuretic factors from patients withchronic uremia on the bioelectric properties and sodiumtransport of the isolated mammalian collecting tubule. J C/inInvest 58:590—597, 1976

22. ALLISoN, MEM, WII.SCN CB, GOTTSCIiAI.K CW: Pathophy-siology of experimental glomerulonephritis in rats. J C/inInvest 53:1402—1423, 1974

23. LUBOWITZ H, MAZUMDAR DC, KAWAMURA J, Cisooi' iT,

502 van Ypersele de Strihou

WEISSER F, ROLE 0, BRICKER NS: Experimental glomerulo-nephritis in the rat: Structural and functional observations.Kidney In! 5:356—364, 1974

24. KRAMP RA, MACDOWELL M, GO1TSCI-IALK CW, OLIVER

JR: A study by microdissection and micropuncture of thestructure and the function of the kidneys and the nephrons ofrats with chronic renal damage. Kidney mt 5:147—176, 1974

25. LUBOwITz H, PURKERSON ML, SUGITA M, BRICKER NS:

GFR per nephron and per kidney in chronically diseased(pyelonephritic) kidney of the rat. Am i Physio/ 217:853—857,1969

26. OLIvER J: Architecture of the Kidney in Chronic Bright's Dis-ease. New York, Paul B. Hoeber, Inc., 1939

27. MALONE 0, Luisowrrz H: The effect of filtered sodium loadper nephron on renal Na-K-ATPase activity (abstr.) KidneyIn! 6:72, 1974

28. HAYSLEFr JP, Bo'u JE, EPSTEIN FH: Aldosterone produc-tion in chronic renal failure. Proc Soc Exp Biol Med130:912—914, 1969

29. Cop!: CL, PEARSON J: Aldosterone secretion in severe renalfailure. C/in Sd 25:33 1—341, 1963

30. SCILRIER RW, REGAl. FM: Influence of aldosterone on so-dium, water and potassium metabolism in chronic renal dis-ease. Kidney In! 1:156—168, 1972

31. WEIDMANN P, MAXWELL MH, Li:pu AN: Plasma aldoste-

rone in terminal renal failure. Ann Intern Med 78:13—18, 197332. WEIDMANN P. MAXWEI.L MH, Dc LIMA J, HIRSCH D,

FRANKLIN SS: Control of aldnsterone responsiveness in ter-minal renal failure. Kidney In! 7:351—359, 1975

33. COOKE CR, RU1Z-MAZA F, KOWARSKI A, MIGEON Ci,WALKER WG: Regulation of plasma aldosterone concentra-tion in anephric man and renal transplant recipients. Kidneymt 3:160—166, 1973

34. GERSTEIN AR, KLEEMAN CR, GolD EM, FRANKLIN SS,MAXWELL MH, GONICK HC, FEFFER ML, STEINMAN TI:Aldosterone deficiency in chronic renal failure. Nephron5:90—105, 1968

35. HAYES CP, MCLEOD ME, ROBINSON RR, STEAD EA: Anextrarenal mechanism for the maintenance of potassium bal-ance in severe chronic renal failure. Trans Assoc Am Physi-cians 80:207—216, 1967

36. BI:RI.YNF GM, VAN LAFTI-IEM L,BFN AR! J: Exchangeable

potassium and renal potassium handling in advanced chronicrenal failure in man. Nephron 8:264—269, l97l

37. BUDDY K, KING PC, LINDSAY RM, BRIGGS JD, WINChESTER

iF, KENNEDY AC: Total body potassium in non-dialysed anddialysed patients with chronic renal failure. Br Med il:77l—775, 1972

38. WRONG 0, METCALFE-GIBSON A, MORRISON RBI, No ST,HowARD AV: In viva dialysis of faeces as a method of stoolanalysis: I. Technique and results in normal sub.jects. Clin Sd28:357—375, 1965

39, WILSON OR, ING TS. METCALFE-GIBSON A. WRONG OM:The chemical composition of faeces in uraemia, as revealedby in-viva faecal dialysis. C/in Sci 35:197—209, 1968

40. CIIARRON RC, LI:ME CE, WII SON OR, ING TS, WRONG OM:The effect of adrenal steroids on stool composition, as re-vealed by in viva dialysis of faeces. C/in Sd 37:151—167, 1969

41. SHIElDs R, MULHOII.AND AT, ELMSLIE RG: Action of al-

dosterone upon the intestinal transport of potassium, sodiumand water. Gut 7:686—696, 1966

42. SHIELDS R, MItEs JB, GILBERTSON C: Absorption and secre-tion of water and electrolytes by the intact colon in a patientwith primary hyperaldosteronism. Br Mcdi 1:93—96, 1968

43. SI! VA P, CHARNEY AN, EPSTEIN FH: Potassium adaptation

and Na-K-ATPase activity in mucosa of colon. Am J Physiol229:1576—1579, 1975

44. ALEXANDER EA, LEVINSKY NG: An extrarenal mechanism ofpotassium adaptation. J C/in Invest 47:740—748, 1968

45. Kcirtl NM, OSTERBERG AE: The tolerance for potassium insevere renal insufficiency: A study of ten cases. J C/in Invest26:773—783, 1947

46. GONICK [IC, MAXWELl. MH, CUTLER RE, DOWLING iT,KI.EEMAN CR: Potassium excretion in renal disease, J C/inInvest 40:1044, 1961

47. PATRICK J: The assessment of body potassium stores. Kidneylot 11:476—490, 1977. this issue

48. ADESMAN J, GOlDBERG M, CASTI.EMAN I., FRIEDMAN IS:Simultaneous measurement of body sodium and potassiumusing Na22 and K42. Metabolism 9:561—569, 1960

49. SPERGEI. G, BI.EICHER Si, GOLDBERG M, ADESMAN J, Got.D-NER MG: The effect of potassium on the impaired glucose

tolerance in chronic uremia, Metabolism 16:581—585. 196750. PATRICK J. JONES NF, BRADI'ORD B, GAUNT J: Leucocyte

potassium in uraemia: Comparisons with erythrocyte potas-sium and total exchangeable potassium. C/in Sd 43:669—678,1972

51. BILBRITY GL, CARTER NW, WhITE MG, SCIIIIIING iF, KN0-CIIEI JP: Potassium deficiency in chronic renal failure. Kid-ncr in! 4:423—430, 1973

52. SEEDAT YK: Exchangeable potassIum study in patients un-dergoing chronic haemodialysis. Br Med J 2:344—345, 1969

53. RAM MD, CHISHOLM GD: Blood-potassium in patients un-dergoing hemodialysis. Lancet 1:260, 1969

54. RETTORI V GRAI, T, MASSRY SG, VIIIAMIt MF: Exchange-able potassium content and distribution in normal subjectsand uraemic patients on chronic haemodialysls. Cli,i Sci42:673—684, 972

55. JONEs NE: Potassium deficiency: A reappraisal, in Proc ConJRoyal College 0/ Physicians of London, edited by SIATERiOU, 1970, pp. 203—214

56. COLES GA: Body composition in chronic renal failure. Q JMed 41:25—47, 1972

57. BODDY, K, KING PC, LINDSAY RM, WINChESTER i, KIN-

NEDY AC: Exchangeable and total body potassium in pa-tients with chronic renal failure. Br Mcdi 1:140—142, 1972

58. LIT1ERI JM, ASAD SN, CASEI.NOVA R, ELLIS KJ, COHN SO:Creatinine excretion and total body potassium in renal faIl-ure. C/in Nephro/ 4:58—61, 1975

59. MORGAN AG, BURKIN5IIAW I., ROBINSON PJA, ROSEN SM:Potassium balance and acid-base changes in patients under-going regular haemodialysis therapy. Br Med J 1:779—783,1970

60. JOFINY Ky, LAWRENCE JR, O'HALLORAN MW, WELLBY ML,WORTIIII:Y BW: Studies on total body, serum and eryth-rocyte potassium in patients on maintenance haemodialysis:The value of erythrocyte potassium as a measure of bodypotassium. Nephron 7:230—240, l970

61. BITTAR FE, WATT MF, PAERAS VR, PARRISH AE: The pH ofmuscle in Laennec's cirrhosis and uremia. C/in Sci23:265—276, 1962

62. BERGSTROM J: Muscle electrolytes in man. Scandi C/in I.abInvest l4(suppl. 68): I —110, I 962

63. GRAIIAM JA, LAWSON OH, LINTON AL: Muscle biopsy waterand electrolyte contents in chronic renal failure. C/in Sci38:583—591, 1970

64. CUNNINGHAM iN, CARTER NO, RECTOR FC JR, SEIDIN DW:

Resting transmembrane potential difference of skeletalmuscle in normal subjects and severely ill patients. J C/inInvest 50:49—59, 1971

Potassium homeostasis in renal failure 503

65. BUTKUS DE, ALFREY AC, MILLER NL: Tissue potassium in

chronic dialysis patients. Nephron 13:314—324, 1974

66. PATRICK J, JONES NF: Cell sodium, potassium and water inuraemia and the effects of regular dialysis as studied in theleucocyte. C/in Sci 46:583—590, 1974

67. EOMONDSON RPS, HIt: ON PJ, JONES NF, PATRICK i,THOMAS RD: Leucocyte sodium transport in uraemia. C/inSd 49:213—216, 1975

68. KRAMER I-IJ, GOSPODLNOV D, KRUCK F: Functional andmetabolic studies on red blood cell sodium transport inchronic uremia. Nephron 16:344—358, 1976

69. WELT LG, SACLIs JR, MCMANUS Ti: An ion transport defectin erythrocytes from uremic patients. Trans Assoc A,n Physi-

cian,s 77:169—181, 1964

70. WELT LG, SMITH EKM, DUNN Mi: Membrane transportdefect: The sick cell. Trans A,ssoc Am Physicians 80:217—226,1967

71. COL,E CH: Decreased ouabain-sensitive adenosine tn-phosphatase activity in the erythrocyte membrane of patientswith chronic renal disease. C/in Sci 45:775—784, 1973

72. COLE CH, BALEE JW, WELT LG: Induction of ouabain-sensitive ATPase defect by uremic plasma. Trans Assoc AOlPhesicians 31:213—220,1968

73. Coii CH, MALETZ R: Changes in erythrocyte membraneouabain-sensitive adenosine triphosphatase after renal trans-plantation. (I/n Sd 48:239—242, 1975

74. NAGANT LiE DEUXCUAISNES C, COLLET RA, BUSSET R,MACU RS: Exchangeable potassium in wasting, amyotrophy,heart disease and cirrhosis of the liver. Lance! 1:681—687,1961

75. BURNEIL iM, VILIAMI! MF, UYEN0 BT, SCRIBNER BI-l:Effect in humans of extracellular p1-I change on relationshipbetween serum potassium concentration and intracellular po-tassium. J (7/n Invest 35:935—939, 1956

76. SIMMONS DH, AVEDON M: Acid-base alterations and plasmapotassium concentration. Am J Phvsiol 197:319—326, 1959

77. SCRIBNER BH, BURNELL JM: Interpretation of the serumpotassium concentration. Metabolism 5:468—479, 1956

78. STANBURY SW, MAHI.ER RF: Salt-wasting renal disease.Metabolic observations on a patient with "salt-losing ne-phritis." Q J Med 28:425—447, 1959

79. LATLIEM W: Hyperchloremic acidosis in chronic pyelone-phritis. N EnglJ Med 258:1031—1036, 1958

80. HERTZ P. RIChARDSON IA: Arginine-induced hyperkalemiain renal failure patients. Arch intern Med 130:778—780, 1972

81. DICKERMAN 1-lW, WALKER WG: Effect of cationic aminoacid infusion on potassium metabolism in viva. Am J Phy,siol206:403—408, 1964

82. SELDIN DW, TARAIL R: Effect of hypertonic solutions onmetabolism and excretion of electrolytes. J C/in invest159:160—174, 1949

83. MAKOFF DL, DA SII,vA JA, ROSENBAUM BJ, LEVY SE, MAX-WELL. MH: Hypertonic expansion: 1. Acid-base and elec-trolyte changes. Am ,/ Phy.siol 218:1201—1207, 1970

84. MAKOFF DL, DA SLLVA JA, ROSENBAUM Bi: On the mecha-nism of hyperkalaemia due to hyperosmotic expansion withsaline or mannitol. C/in Sci 41:383—393, 1971

85. ADlER S: An extrarenal action of aldosterone on mammalianskeletal muscle. Am J Phvsiol 218:616—621, 1970

86. GOI.DFARB 5, Cox M, SINGER I, GOLDBERG M: Acute hy-perkalemia induced by hyperglycemia: Hormonal mecha-nisms. Ann intern Med 84:426—432, 1976

87. WEIDMANN P. MAXWELL MU, ROWE P, WLNER R, MASSRYSG: Role of the renin angiotensin-aldosterone system in the

regulation of plasma potassium in chronic renal disease.

Nephron 15:35—49, 1975

88. WEIDMANN P, RELNHART R, MAxwEll. MU, ROWE P. Co-BURN 1W, MASSRY SG: Syndrome of hyporeninemic hypoal-dosteronism and hyperkalemia in renal disease. J C/in En-docrinol Metab 36:965—977, 1973

89. PEREZ. G, SIEGEL. L. SCHRELNER GE: Selective hypoaldoste-

ronism with hyperkalemia. Ann Intern Med 76:757—763, 1972

90. SCILAMBELAN M, STOCKIGT JR. BIGLIERI EG: Isolated hy-poaldosteronism in adults: A renin-deficiency syndrome. NEngi J Med 287:573—578, 1972

91. Oti MS, CARROLL HH, CIEMMONS iE. VAGNUCCI AH, LEVI-SON SP, WHANG ESM: A mechanism for hyporeninemichypoaldosteronism in chronic renal disease. Metabolism23:1 157—1166, 1974

92. VAGNUCCL AH: Selective aldosterone deficiency in chronic

pyelonephritis. Nephron 7:524—537, 1970

93. BAYARD F, COOKE CR, TLL.L.ER Di, BEITINS IZ, KOWARSKIA, WALKER GT, MLGEON Ci: The regulation of aldosteronesecretion in anephric man. J C/in Invest 50:1585—1595, 1971

94. WEIDMANN P, HORTON R, MAXWEL.!. MU, FRANKLIN SS,

FICUMAN M: Dynamic studies of aldosterone in anephricman. Kidnet In! 4:289—298, 1973

95. ØL.GAARD K: Plasma aldosterone in anephric and non-neph-rectomized dialysis patients in relation to changes in plasma

potassium without change in total potassium balance. ActaMed Scand 198:213—218, 1975

96. DELELVA A, CHRLSTL.LEB AR, MELBY JC,GRAHAM CA, DAY

RP, LUETSCIIER iA, ZAGER PG: Big renin and biosynthetic

defect of aldosterone in diabetes mellitus. N EngI J Med295:639—643, 1976

97. DAVID R, GOlAN 5,DRUCKER W: Familial aldosterone defi-

ciency: Enzyme defect diagnosis and clinical course. Pediat-

ric.s 41:403—412, 1968

98. WILLIAMS GH, BAILEY GL, HAMPERS CL, LAULER DP, MER-RILL. iP, UNDERWOOD RH, BL.ALR-WEST iR, COGHLAN iP,DENTON DA, SCOGGLNS BA, WRIGHT RD: Studies on themetabolism of aldostenone in chronic renal failure andanephric man. Kidney In! 4:280—288, 1973

99. GOODWIN Ti, JAMES VHT, PEART WS: The control of al-dosterone secretion in nephrectomized man. C/in Sci47:235—248, 1974

100. BRUNNER HR. BAER L, SEALEY iE, LEDINGUAM JGG, LA-RACI! iH: The influence of potassium administration and ofpotassium deprivation on plasma renin in normal and hyper-tensive subjects. J C/in Invest 49:2128—2138, 1970

101. VANDER Al: Direct effects of potassium on renin secretion

and renal function. Am J Physiol 2 19:455—459, 1970102. REUBI F, HODEER I: L'activité rénine au cours des néphno-

pathies parenchymateuses avec et sans hypertension, in Ac-tual Nbphrol (Paris). Flammarion Press., 1968, pp. 221—231

103. KLELNKNECHT D, MAXWELL MU: Etude statistique des vari-

ations de l'activité rénine plasmatique dans les néphropathiesbilatérales: Résultats préliminaires. A ctual Nbphrol (Paris),Flammarion Press, 1970, pp. 63—68

04. DAY RP, LUETSCHER JA, GONZALES CM: Occurrence of bigrenin in human plasma, amniotic fluid and kidney extracts. J

Endocrinol Metab 40:1078—1084, 197505. SANTEUSANLO F, FAIOONA GF, KNOCHEL iP, UNGER RH:

Evidence for a role ofendogenous insulin and glucagon in the

regulation of potassium homeostasis. J Lab C/in Med81:809—817, 1973

106. PETTIT GW, VICK RL: Contribution of pancreatic insulin toextrarenal potassium homeostasis: A two-compartmentmodel. A,'n J Phy,siol 226:319—324, 1974

504 van Ypersele de Sir//iou

107. PETTLT OW, VICK RL, SWANDER AM: Plasma K and in-sulin: Changes during KCI infusion in normal andnephrectomized dogs. Am J Physiol 228:107—109, 975

108. HENQUIN JC, LAMIIFRT AL: Cationic environment and dy-namics of insulin secretion: II. Effect of a high concentrationof potassium. Diabetes 23:933—942, 1974

09. TAJESHI K, KAJINUMA H, IDE T: Effect of intrapancreatic

injection of potassium and calcium on insulin and glucagonsecretion in dogs. Diabetes 23:55—60, 1974

110. Hixn N, DAvIDsON MB, BoNossis G: The effect of potas-sium chloride infusion on insulin secretion in vivo. HormMetab Rex 4:64—68, 1972

Ill, ANDRES R, BALTZAN MA, CADER U, ZIERIER KL: Effect ofinsulin on carbohydrate metabolism and on potassium inthe forearm of man. J C/in Invest 41:108—115, 1962

112. BURTON SD, Isiiios T: Effect of insulin on potassium andglucose movement in perfused rat liver. Am J Physiol209:ll45—l161, 1965

113. KE5iLNS Pi, IIAXI-LE ii, LAMBOTTE L, L,AMROTIE C: Theeffect of insulin on the uptake of potassium and phosphate bythe isolated perfused canine liver. Metabolism 12:941—950,1963

114. MCNAY JL, ORAN E: Possible predisposition of diabetic pa-tients to hyperkalemia following administration of the potas-

sium retaining diuretic, Amiloride (M K 870). Metabolism19:58—70, l970

115. WALKER BR, CAPUZZI DM, ALEXANDER F, FAMILIAR RU,Hopt's RC: Hyperkalernia after triarnterene in diabetic pa-tients. (i/n Pharmacol iher 13:643—651, 1972

116. 1-IERSIAN F, RADO J: Fatal hyperkalemic paralysis associatedwith spironolactone. Arch Neurol 15:74—77, 1966

117. CARR0I.t. HH, FARBER Si: Hyperkaleniia and hypcrchlo-remic acidosis in chronic pyelonephritis. Metabolism13:808—817, 1964

118. LUKE RU, AllisoN M, DAVIDSON J: I-lyperkalemia andrenal tubular acidosis due to renal amyloidosis. A no InternMccl 70:121 1—1217, 1969

119. Poi'ovLzEis MM, KATZ FH, PINGGERA WF, ROBINETTE J,HALGRIMSON CU, BLTKUS DE: Hyperkalemia in salt-wastingnephropathy. Arch Intern Med 132:203—208, 1973

120. MANNING RT, BELIRLE FC: Use of spironolactone in renaledema: Effectiveness and association with hyperkalemia.JAMA 176:769—771, 1961

121. ALEXANDER S. ROBERGE U: Low sodium milk in potassium

depletion. Lancet 1:1263—I 264, 1966122. SIRAUSS MV, RAISZ LU: in Clinical Management oJ Renal

failure. Springfield, Illinois, Thomas, 1956123. GRONIiRT FA, TIIEYE RA: Pathophysiology of hyperkalemia

induced by succinylcholine. Anesthesiologc' 43:89—99, 975