Kidney International Vol 51 1997 pp 591602NEPHROLOGY FORUMHyper - PDF document

Kidney International, Vol. 51 (1997), pp. 591—602NEPHROLOGY FORUMHyperkalemic hyperchloremic metabolic acidosis:Pathophysiologic insightsPrincipal discussant: THOMAS D. DUBOSE, JR.Universityof Texas Medical School-Houston, Houston, Texas, USAA 28-year-old woman with acquired immunodeficiency syndrome man-ifest as HIV (CD4 T-cells )(cerebral toxoplasmosis and Pneumocystis carinil pneumonia) was admittedto Hermann Hospital-University of Texas, Houston, with severe dyspneaand chest pain. An initial chest radiograph revealed bilateral pleuraleffusions and massive cardiomegaly. Echocardiography demonstrated alarge pericardial effusion, and a pericardiotomy with window was per-formed on the first hospital day. Although the patient improved subjec-tively, she remained febrile, and on the second hospital day pentamidinewas administered on the assumption that the Pneumocystis carinii pneu-monia had recurred.Laboratory values on admission were: BUN, 27 mg/dl; creatinine, 1.2mg/dl; sodium, 138 mEq/liter; potassium, 4.1 mEq/liter; chloride, 107mEq/liter; and bicarbonate, 23 mEq/liter. Three days after initiation ofpentamidine, the chemistries were as follows: BUN, 72 mg/dl; serumcreatinine, 2.7 mg/dl; sodium, 139 mEq/liter; potassium, 6.8 mEq/liter;chloride, 111 mEq/liter; bicarbonate, 13 mEq/liter, and serum osmolality,307 mOsm/kg H20; the Nephrology Service was consulted. The followingurinary chemistries were obtained at that time: sodium, 40 mEq/liter;potassium, 49 mEq/liter; chloride, 32 mEq/liter; creatinine, 79 mg/dl; andosmolality, 419 mOsm/kg H20. The calculated transtubular potassiumgradient (TTKG) was 6.0, and the urine net charge (UNC) was +57mEq/liter. Despite severe hyperkalemia and metabolic acidosis, the elec-trocardiogram revealed a normal sinus rhythm. Kayexalate was givenrectally and the potassium declined to 5.9 mEq/liter in 12 hours. Penta-midine administration was discontinued and bicarbonate was adminis-tered intravenously. The serum potassium declined to 3.5 mEq/liter within24 hours.The Nephrology Forum is funded in part by grants from Amgen,Incorporated; Merck & Co., Incorporated; Marion Merrell Dow, Incor-porated; Dialysis Clinic, Incorporated; and R & D Laboratories.© 1997 by the International Society of NephrologyDiscussionDR. THOMAS D. DUBOSE, JR. (Director, Division of RenalDiseases and Hypertension, and Professor of Internal Medicine andIntegrative Biology, University of Texas Medical School, Houston,Texas, USA): This case represents a problem encountered com-monly by internists and nephrologists: hyperchioremic metabolicacidosis with hyperkalemia and renal insufficiency in associationwith drug nephrotoxicity. Hyperkalemic hyperchloremic meta-bolic acidosis invariably indicates an abnormality in potassium,ammonium, and hydrogen ion secretion which, while most evidentin patients with renal insufficiency, is not always the result of areduction in renal mass. Indeed, the decrease in whole-kidneypotassium and ammonium excretion is usually out of proportionto the degree of renal insufficiency (as defined by a decrease inglomerular filtration rate) and often can be attributed to ageneralized defect in renal tubular function in the cortical and/ormedullary collecting ducts. In this discussion, I will review thephysiology and pathophysiology of potassium and ammoniumexcretion, and I will focus on the interrelationship betweenpotassium and ammonium transport as it pertains to the patho-physiology of the clinical syndrome of hyperkalemic hyperchior-emic metabolic acidosis.Renal potassium secretionThe renal contribution to potassium balance is a homeostati-cally regulated process accomplished through regulation of potas-sium secretion across the apical membrane of principal cells in thecortical collecting duct. The quantity of potassium secretiondepends on the apical membrane potassium conductance and theelectrochemical driving force across the apical membrane. Potas-sium conductance is defined as the sum of the collective potassiumchannels in the apical membrane. Variables that regulate potas-sium secretion in this nephron segment include factors that affectthe potassium conductance (for example, urine flow and sodiumdelivery) and factors that regulate the electrochemical drivingforce (for example, potassium balance, transepithelial potentialdifference, the pH and bicarbonate concentration of tubule fluid,systemic acid-base balance, and aldosterone) [1]. The electro-chemical driving force for potassium secretion is maintained bythe basolateral Na-K-ATPase (that is, the "sodium pump").Sodium absorption through apical sodium-selective channelsmaintains the negative potential difference in the tubular lumen.The sodium pump generates a negative lumen potential and ahigh intracellular potassium concentration, which together aug-ment potassium secretion. An early effect of aldosterone isEditorsJORDAN J. COHENJOHN T. HARRINGTONNIcoLAos E. MADIASManaging EditorCHERYL J. ZUSMANTufts University School of MedicineCase presentation591 592Nephrology Fonim: Hyperkalemic hyperchloremic metabolic acidosisincreased activity of the apical sodium-selective channel; withtime, aldosterone also increases the activity of the basolateralNa '-K' -ATPase. As expected for a homeostatically regulatedprocess, potassium secretion varies in parallel with systemicpotassium balance (that is, potassium secretion increases dramat-ically in acute and chronic hyperkalemia, and decreases in hypo-kalemia) [1].Renal potassium excretion is primarily the result of regulationof tubular potassium secretion. It therefore follows that a clinicalestimate of potassium transfer into the cortical collecting duetcould be helpful in the recognition of hyperkalemia of renalorigin. The transtubular potassium gradient (TFKG) [21 hasemerged as a clinically useful tool for estimating the potassiumconcentration "gradient" between the peritubular capillary andthe tubular lumen at the level of the cortical collecting duct(CCD). A low TTKG in the hyperkalemie patient ()that the collecting tubule is not responding appropriately to theprevailing hyperkalemia and that potassium secretion is impaired.The formula defining the YI'KG follows:[K] I [K]TTKG =______UmIPmwhere [KI and [K4} represent potassium concentration in theur

ine and plasma, respectively, and Uosm and Osmrepresenturine and plasma osmolality respectively. In this expression, theurine-to-plasma potassium concentration ratio is corrected forwater abstraction in the more distal segments of the collectingduet. The 171KG computation assumes no significant alteration inpotassium content between the CCD and final urine; that CCDtubular fluid osmolality is approximately the same as plasmaosmolality; that "osmoles" are not extracted between CCD andfinal urine; and that plasma [K] approximates peritubular fluid[KJ [3]. Under certain clinical conditions, some or none of theseassumptions might be entirely correct [1]. Particularly problematicis the effect on the YFKG of a dilute urine or of high urine flowrates (polyuria). In these situations, the 171KG underestimatespotassium secretory capacity in the hyperkalemie patient. Withconsideration for these potential pitfalls, the 171KG appears to bea useful clinical tool for estimating potassium secretory ability, butit might be no more useful than is the fractional excretion ofpotassium (FEK+). Because the hyperkalemia of mineraloeorti-eoid deficiency should respond to mineralocorticoid replacement,a patient with hypoaldosteronism would be expected to exhibit anincrease in the TTKG after fludroeortisone administration forseveral hours; the 171KG would not be expected to be altered ina patient with resistance to mineralocortieoid. Thus, an increase in im;&#xplie;&#xs000;the TIKO to 8 after mineraloeortieoid administration suggestsaldosterone deficiency; failure of the 171KG to increase suggeststubular insensitivity to mineraloeortieoid (pseudohypoaldosteron-ism, voltage defect) [41.Therenal causes of hyperkalemia, whichmay be evidenced by an inappropriately low 171KG in thepresence of hyperkalemia, are reviewed in Table 1.Renal ammonium production and transportSeveral nephron segments possess ammoniagenie enzymes, butthe majority of ammonium excreted in the urine is derived fromthe metabolism of glutamine in proximal tubule cells [5—7].Ammonium production in the proximal tubule is regulatedthrough the pivotal enzymes glutaminase and phosphoenolpyru-vate carboxykinase. In chronic acidosis, the activities of bothenzymes and the abundance of their respective messenger RNAsincrease [8, 9]. At physiologic pH, two ammonium ions and thedivalent anion alpha ketoglutarate (ultimately metabolized to twobicarbonate ions) are the major products of glutamine metabo-lism. Figure 1 depicts the nephron segments responsible forammonium transport and its regulation. Ammonium is preferen-tially secreted into the proximal tubular lumen across the apicalmembrane [10—13]. Most of the ammonium is secreted in the firstportion of the proximal convoluted tubule [10]. Direct ammoniumsecretion occurs via substitution of ammonium for hydrogen ionon the apical membrane Na/H exchanger [10, 11]. In the S3segment of the proximal straight tubule, ammonium secretion isincreased by the presence of an acid disequilibrium pH [14].As fluid leaves the proximal tubule and enters the loop ofHenle, a number of processes lead to ammonium and ammoniaefflux and result in a high medullary interstitial ammoniumconcentration. When fluid delivered out of the proximal tubule isalkalinized in the thin descending limb by water abstraction [15],a milieu favorable for ammonia efflux by non-ionic diffusion iscreated (Fig. 1). Direct ammonium transport across the medullarythick ascending limb of Henle's loop (mTALH) apical membrane(by substitution of ammonium for potassium on the Na-2Cl-Keotransporter) is the major mechanism for absorption and isresponsible for generation of high medullary ammonium concen-trations [16]. Evidence for this mechanism is the observation thatammonium absorption is attenuated by luminal application offurosemide and is competitively inhibited by increasing concen-trations of potassium [171. In addition, ammonium also might beabsorbed across the apical membrane of the mTALH throughbarium-sensitive and -insensitive potassium channels. However,the physiologic significance of the Na-2Cl-K cotransporter forammonium absorption is better understood [18]. The pathway forammonium exit across the basolateral membrane is not clearlydefined but likely is achieved through substitution for potassiumon the basolateral potassium conductance [17]. A highly selectivepotassium channel that is sensitive to ATP (ROMK-II) has beencloned and localized to the mTALH [19] and could represent achannel that is responsible for ammonium uptake across eithermembrane of the mTALH.Ammonia can re-enter the proximal straight tubule from theinterstitium [141, thus leading to countercurrent multiplication,where the "single effect" involves selective addition of ammoniumby the proximal tubule and active ammonium absorption in thethick ascending limb of Henle's loop. The countercurrent systemin the loop then multiplies the effect. The net result of this systemis an axial gradient for ammonium. Thus, medullary concentra-tions of ammonium exceed cortical concentrations severalfold[20]. As a consequence, the concentrations of ammonium andammonia in the medullary interstitium exceed the concentrationTable 1. Causes of hyperkalemia attributed to decreased potassiumexcretionDecrease in distal flow rateDecrease in sodium deliveryDecrease in secretion of renin or aldosteroneRenal secretory defectsPrimarySecundaryDrugs/toxinsInterstitial diseases Nephrology Forum: Hyperkalemic hyperchioremic metabolic acidosis593Fig. 1.Nephronsegments responsible forammonium excretion. Ammonium excretion isregulated in response to changes in systemicacid-base and potassium balance. Segmentalcontributions include: proximal convolutedtubule, proximal straight tubule, thindescending limb, thick ascending limb, and4medullaiycollecting duct. GLN =glutamine.prevailing in the inner medullary collecting duct (IMCD) lumen.Thus, a concentration gradient favorable for entry into themedullary collecting duct is created. Ammoriium concentrationsin the inner medullary interstitium reach the greatest amplifica-tion over cortical levels during chronic metabolic acidosis [21].The high concentration of ammonium in the inner medulla can beobliterated by medullary washout and selective medullary dest

ruc-tion (for example, in chronic tubulointerstitial diseases).Ammonium is secreted from the medullary interstitium into themedullary collecting ducts by a combination of ammonia diffusionand active hydrogen ion secretion (H-ATPase and HKtATPase). These processes result in high concentrations of ammo-nium in final urine. In addition, ammonium entry into theterminal IMCD (tIMCD) cell on the basolateral membrane is alsoaccomplished by competition of ammonium for potassium on thesodium pump (Na-K-ATPase) [22].Arole for the Na-K-ATPase in ammonium transport in other collecting duct segmentshas not been established, however. The final step in ammoniumsecretion into the lumen of the tIMCD has not been elucidatedunambiguously. It is generally assumed that ammonia diffusesacross the apical membrane of the tIMCD in parallel withhydrogen ion secretion. An additional possibility, as displayed inFigure 2, is through an ROMK channel in the apical membrane incompetition with potassium. Pathways for ammonia/ammoniumtransport in the medulla are summarized in Figure 2.Metabolic acidosis increases ammonium production throughstimulation of glutaminase and phosphoenolpyruvate carboxyki-nase primarily in the proximal tubule [7, 8, 23]. Moreover, inchronic metabolic acidosis, net ammonium addition to the prox-imal tubule is increased through augmentation of early proximaltubule ammonium secretion [71.Therefore,ammonium deliveryout of the late proximal tubule is increased dramatically. Inresponse to the increase in delivery, ammonium absorption by thethick ascending limb of Henle's loop is augmented, thus increas-ing inner medullary interstitial concentrations of ammonia [23].As a consequence of inner medullary accumulation of ammoniaand the increase in the acid disequilibrium pH in the innermedullary collecting duct, ammonium addition to the medullarycollecting duct increases [21].The kidney's response to chronic metabolic acidosis is anincrease in ammonium production and excretion. Therefore, in apatient with chronic metabolic acidosis of non-renal origin(whether of high or normal anion gap varieties), an increase inammonium excretion ensues. Renal tubular disorders, such as therenal tubular acidoses (proximal, classical distal, and the general-ized distal defect with hyperkalemia), are associated with aninappropriately low ammonium excretion rate when the degree ofsystemic acidosis is taken into consideration. In the absence of alaboratory determination of ammonium concentration in theurine, the urine anion gap (or the urine net charge) has beenwidely accepted as an indirect, clinical means of estimating theresponse of urinary ammonium excretion to metabolic acidosis.This calculation is based on the relative constancy of the unmea-sured cations, such as magnesium and calcium (excluding ammo-nium), and unmeasured anions such as phosphate, sulfate, andorganic anions present in the urine. In a person ingesting anaverage North American diet, the unmeasured anions exceed theunmeasured cations by approximately 80 mEq/day; thus on a dailybasis, the sum of urinary Na + K + NH4 =Cl+ 80 [24].Urinary ammonium concentrations usually are estimated from aspot urine by calculating the urine net charge (UNC) thus:UNC =[Nat+ K] —2HC03GLNNH +Na (H)fH20NH4NF13 594Nephrology Forum: Ilyperkalemic hyperchloremic metabolic acidosisFig. 2. Relationship between ammoniumtransport in the medulla,y thick ascending limb ofHenle's loop (mTALH) and the terminal innermedulla,y collecting duct (tIMCD).where u indicates urinary concentration. Hyperchioremic meta-bolic acidosis due to gastrointestinal losses can be differentiated inthe laboratory from a renal defect in ammonium excretion, asurinary ammonium excretion is typically low in renal tubularabnormalities and high in patients with extrarenal bicarbonateloss (for example, diarrhea). As illustrated by the case presenta-tion, the urine net charge of +57 predicts that little, if any,ammonium is present in the urine, signifying that the metabolicacidosis is of renal origin. The calculation of the urine net chargefrom a spot urine is derived from consideration of total dailyexcretion of unmeasured anions and cations. Thus it is moreaccurate when the daily urine volume is approximately one liter.The higher the urine output, the more unreliable the urine netcharge will be in predicting the urinary [NH4]. In polyuric States,the calculation of the urinary osmolal gap may be more precise[24, 25]. The utility of the urine net charge in estimating urinaryammonium excretion in metabolic acidosis is also affected ad-versely by ketonuria or the presence of drug anions in the urine.Calculation of [NH4] in the urine using the urinary osmolal gapshould not be affected adversely by ketoacids, hippurate, or druganions in the urine. Since this estimation has not been validated ina wide variety of clinical circumstances and has a number ofpotential pitfalls, the most reliable method for detecting a lowammonium excretion ratio is by direct measurement of urinary[NH4] in the clinical laboratory. If the urine sample is diluted1:100, and if appropriate standards are run, the determination ofthe urinary [NH4] will be accurate.Relationship between potassium and ammoniumproduction and transportHyperkalemia should be regarded as an important determinantof the renal response to changes in acid-base balance. Potassiumstatus can affect distal nephron acidification both by direct andindirect mechanisms. First, the level of potassium in systemicblood is an important determinant of aldosterone secretion, whichin turn is an important determinant of distal hydrogen ionsecretion. Further, studies in our laboratory over the last five or soyears have established a critical role for potassium in ammoniumsynthesis in the proximal tubule and in ammonium excretion[26—28]. Chronic potassium deficiency increased, while chronichyperkalemia suppressed, ammonium production [26]. Thesechanges in ammonium production also affect medullaiy interstitialammonium concentration and buffer availability [27]. Moreover,the effects of potassium balance on ammonium transport inproximal tubule and the thick ascending limb of Henle's loop havebeen demonstrated [29]. Hyperkalemia had no effect on ammo-nium transport in the superfi

cial proximal tubule but markedlyimpaired ammonium absorption in the thick ascending limb,reducing inner medullary concentrations of total ammonia anddecreasing secretion of ammonia into the IMCD [26, 29]. Themechanism for impaired absorption of ammonium in the TALH iscompetition between potassium and ammonium for the potassiumsecretory site on the Na-2Cl-K transporter, and perhaps for theapical (and basolateral) potassium channel as well [17, 18].Hyperkalemia also might decrease entry of ammonium into themedullary collecting duct through competition of ammonium andpotassium for the potassium-secretory site on the basolateralmembrane sodium pump [18].In summary, hyperkalemia can dramatically affect amnioniumproduction and excretion. Chronic hyperkalemia decreases am-monium production in the proximal tubule and whole kidney,inhibits absorption of ammonium in the mTALH, reduces mcd-ullary interstitial concentrations of ammonium and ammonia, anddecreases entry of ammonium and ammonia into the medullarycollecting duct. The potential for development of a hyperchior-ernie metabolic acidosis is greatly augmented when renal insuffi-ciency coexists with the hyperkalemia, or in the presence ofaldosterone deficiency or resistance. Such a cascade of eventshelps to explain, in part, the hyperchloremic metabolic acidosisand reduction in net acid excretion characteristic of severalexperimental models of hyperkalemic-hyperchloremic metabolicacidosis, including obstructive nephropathy, selective aldosteronemTALHtIMCD Nephrology Forum: Hyperkalemic hyperchioremic metabolic acidosis595Table 2. Animal models of hyperkalemic hyperchloremic metabolicacidosisChronic amiloride administration: voltage defectPost-obstructed kidney: pump defectSelective aldosterone deficiency:Pump/voltage/NH4 deficiencyChronic dietary hyperkalemia: NH3/NH4 deficiencydeficiency, chronic amiloride administration, and chronic dietaryhyperkalemia in the rat [26, 30, 31].Acidification defects associated with distal nephron dysfunctionExperimental models. Experimental models of hyperkalemic hy-perchloremic metabolic acidosis have done much to elucidate themechanisms responsible for this acid-base alteration (Table 2).Chronic amiloride administration. Studies in our laboratory thatemployed microelectrodes to measure disequilibrium pH as anindex of proton secretion, as well as PCO2 in the papillarycollecting duct in rats after administration of amiloride for fourdays, revealed that this model of hyperkalemia and metabolicacidosis is associated with a reduction in proton secretion. Im-paired potassium secretion in the CCT, the result of such a"voltage" defect, resulted in hyperkalemia [32, 33].Unilateral ureteral obstruction (UUO). Another experimentalmodel of distal renal tubular acidosis associated with hyperkale-mia is that secondary to unilateral ureteral obstruction. Whilefindings in the postobstructed kidney are similar in many respectsto the arniloride model discussed previously [30, 33], several linesof evidence clearly indicate that this disorder results from anon-voltage-mediated rate defect rather than a "voltage" defect.In the postobstructed model, the animal is unable to acidify theurine after an acid challenge, ammonium excretion decreases, theacid disequilibrium pH is obliterated, and the papillary PCO2 ismarkedly reduced [30]; all these sequelae are compatible with ahydrogen ion pump defect. The decrease in ammonium entry intothe inner medullary collecting duct appears to be the conse-quence, in part, of the defect in hydrogen ion secretion. Thedecrease in bicarbonate transport (JtCO2) observed in vitro inperfused collecting ducts from rabbits subjected to ureteral ob-struction appears first in medullary segments [34]. After UUO,H-ATPase activity is reduced to a greater extent in medullarythan in cortical segments [35]. With immunocytochemical tech-niques, interruption in the cellular distribution of the 31 kDsubunit of H-ATPase has been reported in rat intercalated cells[36]. This alteration suggests that the H-ATPase fails to insertinto the apical membrane ("gaps" or "discontinuity") [36]. Insummary, biochemical, in-vitro microperfusion, and in-vivo mi-cropuncture studies support the view that the proton pump isimpaired in unilateral ureteral obstruction.Selective aldosterone deficiency. Selective aldosterone deficiencyin the rat impairs hydrogen secretion by the inner medullarycollecting duct [31]. This study also demonstrated impairedammonium transfer from the interstitium into the IMCD lumen.As a result, ammonium excretion was reduced dramatically.Moreover, as a result of impaired ammonium production, ammo-nium delivery to the loop of Henle also was reduced. Thus, thereduction in ammonium transfer to the IMCD could be explainedby a decrease in inner medullary ammonium accumulation. Sincepapillary PCO2 was reduced during bicarbonate loading, the rateof proton secretion also was clearly compromised. Findings inexperimental animals and in patients with adrenal insufficiency[37—39] provide evidence that mineralocorticoid deficiency cancause acidosis and impairment of renal acidification, even in theabsence of renal disease or glucocorticoid deficiency. The poten-tial for systemic metabolic acidosis in such a setting can beamplified greatly, however, in individuals with renal insufficiencyand a decrease in functioning renal mass. In patients with selectivehypoaldosteronism and chronic renal insufficiency, mineralocorti-coid administration increases renal acid excretion directly byincreasing renal hydrogen ion secretion, and indirectly by correct-ing hyperkalernia, the latter allowing ammoniurn production andexcretion to increase [23].Chronic dietary potassium loading. The importance of hyperka-lemia in the development of metabolic acidosis due to mineralo-corticoid deficiency has been verified. In patients with adrenalinsufficiency, correction of hyperkalemia with cation exchangeresins significantly increases net acid excretion (ammonium excre-tion) and corrects the metabolic acidosis, even in the absence ofaldosterone administration [40, 41]. Using micropuncture studiesin immature rats, we have shown that whole-kidney ammoniumexcretion falls significantly in vivo in a model of chronic dietarypotassium lo

ading [261. This decrease in excretion was associatedwith a marked reduction in whole-kidney ammonium production,which occurred despite coexistent chronic metabolic acidosis.Nevertheless, chronic hyperkalemia did not affect net secretion ofammonium by the superficial proximal convoluted tubule.Chronic hyperkalemia impairs accumulation of ammonium in theinner medulla and significantly compromises the transfer ofammoniurn into the IMCD [27], at least in part by inhibition ofactive ammonium absorption by the medullary TALH. Therefore,hyperkalemia can lead to metabolic acidosis as a result of adecrease in ammonium excretion. The tendency for metabolicacidosis to develop as a consequence of hyperkalemia mightdepend on the degree of remaining functional renal mass and theintegrity of the renin-aldosterone system [33, 40].Clinical disorders. The coexistence of hyperkalemia and hyper-chlorernic metabolic acidosis suggests generalized distal tubuledysfunction. Generalized distal nephron dysfunction manifests asa hyperchioremic hyperkalemic metabolic acidosis, in which uri-nary ammonium excretion is invariably depressed and renalglomerular function is often compromised. Although hyperchlor-ernie metabolic acidosis and hyperkalernia occur with regularity inadvanced renal insufficiency, patients identified and studied sim-�ply because of severe hyperkalemia (5.5 mEq/liter), for example,those with diabetic nephropathy and/or tubulointerstitial disease,have hyperkalemia that is disproportionate to the reduction inglomerular filtration rate. The transtubular potassium gradient [2]is usually low in patients with this disorder ()the collecting tubule is not responding appropriately to theprevailing hyperkalernia. In such patients, a unique dysfunction ofpotassium and acid secretion by the collecting tubule is attributedto either hypoaldosteronism [42, 431 or to a decrease in effective-ness of aldosterone [44]. Thus, in patients presenting with thisconstellation of findings, an evaluation of renin-aldosterone elab-oration is indicated. A classification of the clinical disordersassociated with hyperkalemia and acidification defects is proposedin Table 3. 596NephrologyFomm: Hyperkalemic hyperchloremic metabolic acidosisTable 3. Pathophysiologic classification of disorders associated withhyperkalemic hyperchloremic metabolic acidosisMineralocorticoid deficiencyPrimaryGeneralized (Addison's disease)Isolated (selective aldosterone deficiency)SecondaryHyporeninemicPharmacologicMineralocorticoid resistancePseudohypoaldosteronism type 1Renal tubular dysfunctionVoltage defectsSecondary to drugs that interfere with Na channel functionPseudohypoaldosteronism type 2Secondary to tubulointerstitial disease (pseudohypoaldosteronismtype 3)Hydrogenion pump defectAldosterone deficiencyObstructive uropathyMineralocorticoid deficiency. Destruction of the adrenal cortexby hemorrhage, infection, invasion by tumors, or autoimmuneprocesses results in Addison's disease, that is, combined glucocor-ticoid and mineralocorticoid deficiency. Causes of Addison'sdisease include tuberculosis, autoimmune adrenal failure, 21-hydroxylase deficiency, fungal infections, adrenal hemorrhage,metastasis, lymphoma, AIDS, amyloidosis, and drug toxicity (ke-toconazole, fiuconazole, phenytoin, rifampin, and barbiturates)[44, 45]. Addison's disease is manifest by hypoglycemia, anorexia,weakness, hyperpigmentation, and a failure to respond to stress.These defects can occur in association with renal salt wasting andhyponatremia, low plasma aldosterone levels, high levels ofplasma renin activity (PRA), and hyperkalemia and metabolicacidosis [42—451. The metabolic acidosis of mineralocorticoiddeficiency results from a decrease in hydrogen ion secretion in thecollecting duct secondary to decreased H-ATPase number andfunction. The hyperkalemia of mineralocorticoid deficiency, inturn, decreases ammonium production and excretion.In contrast to the less common primary adrenal disorder,patients with hyporeninemic hypoaldosteronism exhibit lowplasma renin activity. Patients are usually older (mean age, 65years) and commonly exhibit mild to moderate renal insufficiency(70%) and acidosis (50%) in association with chronic hyperkale-mia (5.5—6.5 mEq/liter) [43, 46]. The most frequently associatedrenal diseases are diabetic nephropathy and tubulointerstitialdisease, systemic lupus erythematosus, and AIDS nephropathy[44]. For 80% to 85% of such patients, there is a reduction inplasma renin activity that cannot be stimulated by the usualphysiologic maneuvers. Aldosterone secretion, while low, can beincreased by administration of angiotensin II or ACTH. Sinceapproximately 30% of patients with hyporeninemic hypoaldoste-ronism are hypertensive, the finding of a low plasma renin in thesepatients suggests a volume-dependent form of hypertension withphysiologic suppression of renin production [461. In general,patients with more advanced renal insufficiency as a result ofglomerular disease rather than tubulointerstitial disease (forexample, diabetic nephropathy) are more commonly volumeexpanded [44, 471. Because either mild salt wasting or saltretention can occur in this disorder, the precise cause of thedecrease in plasma renin has not been established firmly. As Imentioned, impaired ammonium excretion is the combined resultof hyperkalemia, impaired ammoniagenesis, a reduction innephron mass, reduced proton secretion, and impaired transportof ammonium by nephron segments in the inner medulla [31, 41.Isolated hypoaldosteronism, which can occur in critically illpatients, particularly in the setting of severe sepsis or cardiogenicshock, is manifest by markedly elevated ACTH and cortisol levelsin association with a decrease in aldosterone elaboration inresponse to angiotensin II [48]. This hypoaldosteronism might besecondary to selective inhibition of aldosterone synthase as aresult of hypoxia, release of cytokines such as TNFs or IL-I, or,alternatively, high circulating levels of atrial natriuretic peptide(ANP) [49]. The features of hypoaldosteronism, including hy-perkalemia and metabolic acidosis, can be potentiated by theadministration of potassium-sparing diuretics, potassium loads inparenteral nutrition solutions, or heparin, which suppresses aldo-

sterone synthesis in the critically ill patient [50].Mineralocorticoid resistance. Pseudohypoaldosteronism (PHA)is characterized by renal resistance to the action of aldosterone.The clinical features include hyperkalemia (which can be attrib-uted to impaired potassium secretion), normal or elevated aldo-sterone levels, and hyperkalemic hyperchioremic metabolic aci-dosis. By definition, physiologic mineralocorticoid replacementtherapy does not correct the hyperkalemia. Mineralocorticoidresistance with hyperkalemia can be associated with salt retentionor with salt wasting. Pseudohypoaldosteronism type 1, first de-scribed in infants with severe salt wasting, hypotension, andhyperkalemia, is inherited as either an autosomal recessive ordominant disorder. The collecting tubule is unresponsive toaldosterone, and some patients demonstrate resistance in othertarget organs [51]. This disorder appears to be the result of aloss-of-function mutation in the c or 3 subunit of the epithelialsodium channel (ENaC) [521. Adult patients have been reportedwith hyperkalemia, hyperchloremic metabolic acidosis, hyperten-sion, normal renal function, undetectable plasma renin activity,and low aldosterone levels (pseudohypoaldosteronism type II)[44]. These patients generally have not exhibited glomerular ortubulointerstitial disease. The acidosis is mild and can be ac-counted for solely by the magnitude of hyperkalemia. Further-more, renal potassium secretion is resistant to mineralocorticoidadministration. Renin and aldosterone levels increase if volumeexpansion is corrected by diuretics or salt restriction. In carefulstudies, Schambelan et al demonstrated that potassium excretionresponds to sodium sulfate infusion but not to sodium chlorideinfusion [53]. They then suggested that this disorder results froman early distal tubule "chloride shunt." The "shunt" is viewed asthe result of increased chloride reabsorption in the early distaltubule, which would, by decreasing transepithelial voltage, reducethe driving force for potassium secretion. Thiazide diureticsconsistently correct the hyperkalemia and metabolic acidosis, aswell as the hypertension and plasma aldosterone and plasma reninlevels; thus it would be reasonable to assume that this disorderstems from increased activity of the thiazide-sensitive Na -Clcotransporter in the connecting tubule [44]. Nevertheless, geneticlinkage studies have failed to establish such a relationship [54].Pseudohypoaldosteronism can be distinguished from selectivehypoaldosteronism by the presence of hypertension and normalrenal function, the absence of diabetes mellitus and salt wasting,and the failure to exhibit a kaliuretic response to mineralocorti-coids. Some authors have described a third type of PHA, which Nephrology Forum: J-Iyperkalemic hyperchioremic metabolic acidosis597Table 4. Acquired renal tubular secretoly defects associated withhyperkalemic hyperchloremic metabolic acidosisaSickle-cell diseaseSLERenal transplant rejectionObstructive uropathyMedullary cystic diseaseDrug-induced interstitial nephritisAnalgesic abuse nephropathyHIV nephropathy1gM monoctonal gammopathy—a Usuallyassociated with decreased renin secretion and impaired re-sponse of collecting tubule to aldosterone.occurs in adults with overt salt wasting and tubulointerstitialdisease but without hypertension [55].Table 4 outlines a number of renal diseases that are commonlyassociated with deficient renin elaboration and an attenuatedresponsiveness to aldosterone, hyperkalemia, and on occasiontubulointerstitial involvement [55—57]. Although hyperkalemia ismore likely to be associated with metabolic acidosis and reducedammonium excretion when the GFR is below 20—50 mI/mm, inthese disorders significant hyperkalemia can occur with signifi-cantly less renal functional impairment. Hyperkalemia out ofproportion to the degree of renal insufficiency typically occurswith the nephropathies associated with sickle cell disease, sys-temic lupus erythematosus, and obstruction of the collectingsystem; with acute and chronic renal allograft rejection; in hypoal-dosteronism; and occasionally in patients with multiple myelomaand amyloidosis [53].An additional disorder that results in hyperkalemic hyperchlo-remic metabolic acidosis has been dubbed hyperkalemic distalRTA because of the coexistence of hyperkalemia with the pa-�tient's inability to acidify the urine (U 5.5) during spontane-ous acidosis or following an acid load [581.Thehyperkalemiaresults from impaired renal potassium secretion, and the TTKGor FEK is invariably lower than that expected for hyperkalemia[56]. Urinary ammonium excretion is reduced, but aldosteronelevels can be low, normal, or even increased. This variability inaldosterone levels assists in distinguishing this disorder fromselective hypoaldosteronism. Moreover, in selective hypoaldoste-ronism, the urinary pH is low, and the defect in urinary acidifica-tion is attributed to the decrease in ammonium excretion andhydrogen secretion. Hyperkalemic distal RTA, which is assumedto be the result of a "voltage defect" in the cortical collecting duct,appears to occur in a variety of renal diseases [56, 57]. Neverthe-less, in patients with hyperkalemic RTA and hypoaldosteronism,no evidence for such a defect could be demonstrated [58]. It wasconcluded that acidification was impaired as a result of abnormalfunction of the hydrogen ion secretory pump (H-ATPase).Drugs can impair renin or aldosterone elaboration or causemineralocorticoid resistance, which mimics the clinical manifes-tations of the acidification defect that occurs with the generalizedform of distal RTA with hyperkalemia (Table 5). Cyclo-oxygenaseinhibitors (nonsteroidal anti-inflammatory drugs) can generatehyperkalemia and metabolic acidosis as a result of inhibition ofrenin release [53]. Beta-adrenergic antagonists cause hyperkale-mia because they alter potassium distribution and interfere withthe renin-aldosterone system. Heparin impairs aldosterone syn-thesis as a result of direct toxicity to the zona glomerulosa, leadingTable 5.Drug-inducedhyperkalemiaImpaired renin-aldosterone elaborationCyclo-oxygenase inhibitors13-adrenergic antagonistsConverting-enzyme inhibitorsHeparinInhibitors of renal potassium secretionPotassium

-sparing diureticsTrimethoprimPentamidineCyclosporine ADigitalis overdoseLithiumAltered potassium distributionInsulin antagonists (somatostatin, diazoxide)/3-adrenergic antagonistsa-adrenergic agonistsHypertonic solutionsDigitalisSuccinylcholineArginine hydrochloride, lysine hydrochlorideto inhibition of aldosterone synthase. Angiotensin-converting-enzyme (ACE) inhibitors interrupt the renin-angiotensin systemand cause hypoaldosteronism with hyperkalemia and acidosis,particularly in patients with advanced renal insufficiency and thosewith a tendency to develop hyporeninemic hypoaldosteronism,such as those with diabetic nephropathy [56]. The combination ofpotassium-sparing diuretics and ACE inhibitors should beavoided assiduously.Spironolactone, a competitive inhibitor of aldosterone, fre-quently causes hyperkalemia and metabolic acidosis when admin-istered to patients with renal insufficiency [44, 561.Similarly,amiloride and triamterene can have the same effect [53, 59]. Bothpotassium-sparing diuretics block the apical Na-selective chan-nel in the collecting duct principal cell and thus alter the drivingforce for potassium secretion. Amiloride is the prototype for agrowing number of agents that cause hyperkalemia, includingtrimethoprim and pentamidine, particularly in patients withAIDS. Trimethoprim and pentamidine are related structurally toamiloride and triamterene; all these agents are heterocyclic, weakbases that exist primarily in the protonated forms in an acidicurine [60]. Kleyman and Ling demonstrated that the protonatedforms of both trimethoprim and pentamidine inhibit the epithelialsodium-selective channel in A6 distal nephron cells [60, 61]. Thiseffect in A6 cells has been verified in rat late distal tubulesperfused in vivo [62]. Hyperkalemia has been observed in 20% to53% of HIV-infected patients receiving high-dose trimethoprim(TMP)-sulfamethoxazole (SMX) or TMP-dapsone for the treat-ment of opportunistic infections, and in as many as 100% ofpatients with AIDS-associated infections (Pneumocystis carinii)receiving pentamidine for more than 6 days [62]. Because bothTMP and pentamidine decrease the electrochemical driving forcefor both potassium and hydrogen secretion in the CCT (Fig. 3),metabolic acidosis frequently accompanies the hyperkalemia,even in the absence of severe renal failure, adrenal insufficiency,severe tubulointerstitial disease, or hypoaldosteronism. While ithas been assumed widely that such a "voltage" defect couldexplain the decrease in hydrogen ion secretion, it is likely thathyperkalemia plays a significant role in the development of themetabolic acidosis through a decrease in ammonium production 598NephrologyForum: Hyperkalemic hyperchloremic metabolic acidosisFig. 3. Mechanisms of drug-induced voltage defects in the cortical collectingtubule that inhibit K secretion. Apical Na channel blocked by amiloride,triamterene, trimethoprim (TMP), and pentamidine, and basolateral Na,K-ATPase by cyclosporine A (CsA).and excretion. Ammonium excretion has not been systematicallyinvestigated in patients with hyperkalemia receiving either TMPor pentamidine thus far, however.Cyclosporine A (CsA) is associated with hyperkalemia in thetransplant recipient. This drug inhibits basolateral Na,K-ATPase, thereby decreasing intracellular [K] and the transepi-thelial potential, which together decrease the driving force forpotassium secretion [631.Sandset al have suggested that thespecific mechanism of CsA inhibition of the sodium pump isthrough inhibition by this agent of calcineurin activity [64].Cyclosporine also could decrease the filtered load of potassiumthrough hemodynamic mechanisms, such as vasoconstriction,which decrease GFR and alter the filtration fraction. Additionalmechanisms by which CsA might cause hyperkalemia and meta-bolic acidosis include mineralocorticoid resistance [65], a distalhydrogen pump defect [66], and impairment of the apical mem-brane potassium-selective channel.TreatmentDocumentation of the background upon which hyperkalemichyperchloremic metabolic acidosis is expressed provides the basisfor appropriate therapy. Of particular importance is a careful drugand dietary history. Contributing or precipitating factors include:low urine flow or decreased distal sodium delivery, a rapid declinein GFR (especially in acute superimposed on chronic renalfailure), hyperglycemia or hyperosmolality, and unsuspectedsources of exogenous potassium intake. The workup shouldinclude evaluation of the 3TFKG and its response to furosemideand fludrocortisone, an estimate of renal ammonium excretion(urine net charge, urine osmolar gap, and urine pH), and evalu-ation of PRA and aldosterone secretion. The last can be obtainedunder stimulated conditions utilizing dietary salt restriction andfurosemide-induced volume depletion.The decision to treat often is based on the severity of thehyperkalemia. A reduction in serum potassium often improves themetabolic acidosis by increasing ammonium excretion as potas-sium levels return to the normal range [44]. Patients with com-bined glucocorticoid and mineralocorticoid deficiency shouldreceive both adrenal steroids in replacement dosages. Patientswith hyporeninemic hypoaldosteronism usually respond to acation-exchange resin (sodium polystyrene sulfonate), alkali ther-apy, or treatment with a ioop diuretic. Volume depletion shouldbe avoided unless the patient is volume overexpanded or hyper-tensive. Supraphysiologic doses of mineralocorticoids are some-times necessary but should be administered cautiously in combi-nation with a loop diuretic to avoid volume overexpansion oraggravation of hypertension [56]. Infants with pseudohypoaldo-steronism type I should receive a salt supplement in amountssufficient to correct the syndrome and allow normal growth;patients with pseudohypoaldosteronism type IT should receivethiazide diuretics along with dietary salt restriction [441. Althoughit seems prudent to discontinue drugs identified as likely causes ofthe hyperkalemia, this is not always feasible in the patient with alife-threatening disorder, for example, during TMP-SMX or pen-tamidine therapy in the atovaquone (Mepron)-sensitive AIDSpatient with Pneumocystis carinii pneumonia. The mechanism bywhich TMP and pentamidine cause h

yperkalemia (voltage defect)might lead one to reason that the delivery to the cortical collectingduct of a poorly reabsorbed anion might improve the electro-chemical driving force, favoring potassium and hydrogen secre-tion. The combined use of acetazolamide with sufficient sodiumbicarbonate to deliver significant amounts of bicarbonate to thecortical collecting tubule, thereby increasing the negative trans-epithelial voltage, theoretically could increase potassium andhydrogen ion secretion [63]. Obviously, with such an approach,aggravation of metabolic acidosis by excessive acetazolamide orinsufficient sodium bicarbonate administration must be avoided.This approach, however, has not yet been corroborated by clinicaltrials.In summary, hyperkalemia plays a pivotal role in the kidney'sresponse to metabolic acidosis. Hyperkalemia impairs ammoniumproduction and transport in the proximal tubule, as well asammonium transport in the thick ascending limb of Henle's loopand the medullary collecting duct. The competitive inhibition byhyperkalemia of thick limb absorption in turn impairs the coun-tercurrent system in the renal medulla, which reduces ammoniumentry into the inner medullary collecting duct, thus decreasing netacid excretion. This cascade of events is most commonly seen inhyperkalemic hyperchloremic metabolic acidosis as a result ofimpaired renin release or angiotensin formation, aldosteronesecretion or action, or collecting tubule sodium transport. Thesedisorders, often associated with a generalized defect in potassiumand hydrogen ion secretion in the collecting duct, are sometimesdesignated collectively as type-TV distal renal tubular acidosis.Several drugs related structurally to amiloride, such as tn-methoprim and pentamidine, impair sodium absorption by inter-fering with the epithelial sodium channel in the apical membraneof the collecting duct. As a result of impaired sodium absorption,a "voltage" lesion occurs that compromises potassium secretionAmilorideTriamtereneTMPPentamidineNaKCCT ( principal cell) Nephrology Forum: Hyperkalemic hyperchioremic metabolic acidosis599secondarily. Clinical assessment of potassium secretion is pro-vided by calculation of the fractional excretion of potassium or thetranstubular potassium gradient. In addition, the ammoniumconcentration in the urine can be estimated in the acidotic patientby calculation of the urine net charge. Treatment should bedirected toward correction of the hyperkalemia, restoration ofeuvolemia, adequate alkali therapy, loop diuretics, and dietarypotassium restriction. In severe hypoaldosteronism, the effect ofloop diuretics can be augmented significantly by administration ofsmall doses of mineralocorticoid. Fludrocortisone should be usedcautiously, however, and avoided in the presence of hypertensionor congestive heart failure.Questions and answersDR. NIc0LA0s E. MADIAS (Chief Division of Nephrology, NewEngland Medical Center, Boston, Massachusetts): As you know,Ang II itself has now been regarded as an important modulator ofproximal acidification processes as well as of renal ammoniaproduction and transport [67, 68]. Could you please comment onthe potential direct effects of Ang II in the acidification defect ofhyporeninemic hypoaldosteronism?DR. DuBosu: Clearly, more investigation is needed to elucidatethe role of angiotensin II in the development of metabolic acidosisin association with hyporeninemia. As you mentioned, it is wellestablished that Ang II increases bicarbonate absorption in the Ssegment by inhibiting adenylyl cyclase. The principal pathway forselective addition of ammonium into the proximal tubule isthrough substitution of ammonium on the apical sodium/protonexchanger. Nagami has shown that luminal Ang II increasesselective ammonium addition in the proximal tubule in parallelwith stimulation of luminal acidification [68]. In hyporeninemichypoaldosteronism per se, one would anticipate a marked reduc-tion in ammonium secretion in the proximal tubule. In our modelof selective aldosterone deficiency, we did not specifically examineproximal ammonium secretion. Angiotensin II inhibitors may beassociated with hyperkalemia and metabolic acidosis, of course,but this complication appears to occur when there has beentubulointerstitial disease and/or a significant reduction in renalmass. Our models predict that in this setting, both potassium andhydrogen ion secretion are impaired in the collecting duct.Hyperkalemia further compromises net acid excretion by inhibit-ing both ammonium production and excretion. The effect of AngII on ammonium transport in the inner medulla has not beenexamined directly, however.DR. MADIAS: In reflecting on the acidification defect of hy-poreninemic hypoaldosteronism, one would expect features of avoltage-dependent acidification defect. Yet patients with hy-poreninemic hypoaldosteronism don't exhibit sodium wastingunless placed on severe sodium restriction, and they maintain theability to acidify the urine normally. Do you think that the basicacidification defects underlying hyporeninemic hypoaldosteron-ism and hyperkalemic distal RTA are essentially the same butquantitatively more expressed in the latter syndrome?DR. DUBOSE: I agree that these disorders appear to represent acontinuum of the same defect. In the model of selective aldoste-rone deficiency, I believe that the bulk of the evidence is morecompatible with a "pump" defect rather than with a "voltage"defect. Support for this point of view is provided by studies in boththe turtle bladder and the isolated perfused collecting tubule. It isclear that aldosterone directly and specifically impairs protonpump function in the outer medullary collecting duct, where"voltage" would not be a factor. Aldosterone deficiency thereforeappears to compromise the function or reduce the number ofproton pumps in the membrane.DR. JOHN T. HARRINGTON (Dean ad interim, Tufts UniversitySchool of Medicine, Boston, Massachusetts): I have two questions.First, in regard to the TTKG, could you tell us the validity,sensitivity, specificity, etc., supporting the concept that a ratio ofless than 8 signifies renal hyperkalemia, or is that figure simply anarbitrary definition? Second, why do only 50% or so of patientswith hyporeninemic hypoaldoste

ronism actually have metabolicacidosis? Shouldn't it be 100%?DR. DuBosE: The TTKG represents a simple computation thathelps place the pathophysiology of hyperkalemia into perspective.It should be viewed as nothing more than a diagnostic aid, not adefinitive test. The validity of the TT'KG has not been testedrigorously, but there are numerous assumptions and potentialpitfalls.To answer your second question, I don't know why 100% ofpatients with isolated hypoaldosteronism don't develop metabolicacidosis. There are a number of mechanisms by which ammoniumexcretion can be increased to compensate for simple hyperkale-mia. The simultaneous occurrence of metabolic acidosis andhyperkalemia seems to require either aldosterone deficiency orresistance, a decrease in functional renal mass, or impairedtubular transport.DR. MADIAs: Are there good observations on the level ofplasma aldosterone in patients with hyporeninemic hypoaldoste-ronism following correction of the hyperkalemia?DR. DUBOSE: A few studies have examined this questionindirectly. Szylman and colleagues first showed that correction ofhyperkalemia by Kayexalate was associated with repair of meta-bolic acidosis [69]. This repair of acidosis occurred in tandem withan increase in urinary ammonium excretion. These changes tookplace irrespective and independent of aldosterone; pharmacologicdoses of fludrocortisone did not correct the hyperkalemia or theacidification defect.DR. JAMES STROM (Division of Nephrology, St. Elizabeth's Hos-pital, Brighton, Massachusetts): What distinguishes pseudohypoal-dosteronism type II from hyporeninemic hypoaldosteronism? If Iunderstood your categorization correctly, the groups have similarhormone levels and may require similar supraphysiologic replace-ment of mineralocorticoid.DR. DUBOSE: Patients with hyporeninemic hypoaldosteronismoften have significant renal insufficiency. Pseudohypoaldosteron-ism type II can occur with a normal GFR. These latter patients areresistant to large doses of mineralocorticoid and are volume-expanded. The primary abnormality in PHA type II has beenproposed to be of tubular origin, the "chloride shunt," or in-creased CL transport, which causes the volume expansion, hyper-tension, and hyperkalemic metabolic acidosis.DR. AJAY SINGH (Division of Nephrology, New England MedicalCenter): Among patients with hyporeninemic hypoaldosteronism,there appear to be a significant number in whom aldosteronesecretion cannot be independently stimulated by ACTH or angio-tensin II [70, 71]. This would suggest that other factors play a rolein suppressing aldosterone. Indeed, one study suggests that ANPis this factor [72]. Can you comment on this phenomenon andupdate us on any recent studies that might have examined thisissue? 600NephrologyForum: Hyperkalemic hyperchloremic metabolic acidosisDR. DUBOSE: The consensus has been that when renin isdepressed as a result of volume expansion, aldosterone produc-tion will not increase in response to hyperkalemia. Insulin defi-ciency also might play a role. As you point out, a deficientresponse of ANP to volume expansion can accompany PHA typeII or Gordon's syndrome, but there is no direct information tosupport such an association.DR. GEETHA NARAYAN (Division of Nephrology, St. Elizabeth sHospital): It has been postulated that severe volume depletion canhelp perpetuate even simple metabolic acidosis, such as acidosisresulting from diarrhea. It also has been postulated that this mayresult from a reversible distal acidification defect, resulting indecreased ammonium excretion, Can you comment on this, and isthis defect caused by decreased sodium availability in the distalnephron, coupled with more efficient chloride absorption, produc-ing a voltage-type defect?DR. DUBOSE: Yes, in the scenario you describe, decreaseddelivery of sodium ultimately could lead to impaired potassiumand hydrogen ion secretion, which would reduce net acid excre-tion and cause metabolic acidosis. However, it is not necessary tosuperimpose a chloride shunt, because this lesion causes volumeexpansion. Extremely low distal sodium delivery or impairedsodium absorption should be sufficient.DR. NARix: Is the effect of hyperkalemia on ammoniumproduction still believed to be mediated, at least in part, throughtranscellular cation exchange and intracellular alkalosis?DR. DUBOSE: Hyperkalemia decreases the uptake of glutamineand decreases production of ammonium from glutamine precur-sors. Although the cellular mechanism has not been definedprecisely, it is known that uptake is enhanced by intracellularacidosis. Whether intracellular alkalosis develops in the proximaltubule cell during hyperkalemia, or whether it plays a role in thedecrease in ammonium excretion, is not known. In a model ofhyperkalemic metabolic acidosis, David Good and I attributed thedecrease in ammonium excretion to changes in transport beyondthe proximal tubule [26]. This defect was later localized to therenal medulla [27].Whileammonium production was reduced by50% and excretion by 40%, there was no change in net ammo-nium transport in the proximal tubule. Such a disassociation ofproximal production and transport conceivably could occur inresponse to intracellular alkalosis. Nevertheless, this effect couldbe modulated by the peritubular potassium concentration, whichclearly alters proximal ammonium secretion.DR. MAIMAS: As you know, some work suggests that ammoniacan exert injurious effects on the kidney by augmenting interstitialfibrosis via complement activation [73]. In this context, thedecreased ammonia production in tubulointerstitial diseasesmight be teleologically adaptive. On the other hand, metabolicacidosis has its own adversity [74]. Can you reflect a bit on thisissue?DR. DUBOSE: The interesting study you mentioned [73] sug-gested that ammonia in the inner medulla could activate thealternate complement pathway, which could lead to tissue injuryin the form of tubulointerstitial disease. One might speculate thatin tubulointerstitial disease a decrease in ammonium absorptionin the thick limb and a decrease in ammonia accumulation in theinner medulla might be beneficial in slowing perpetuation of theinjury, as you suggest. This model has not been developed further,however, and additional work is needed i

n this area.DR. HARRJNGTON: I have two questions regarding the conceptof direct ammonium secretion. First, how do you specificallymeasure direct ammonium secretion across the distal nephron?Second, of the total amount of ammonium excreted, how much isvia direct ammonium secretion versus ammonia combining withhydrogen to be excreted as ammonium?DR. DUBOSE: The precise delineation of which moiety issecreted requires consideration of the pK' of ammonium, the pHof tubule fluid, the total ammonia concentration, and transepithe-hal potential. To answer your question, each nephron segment'scontribution would have to be defined in each physiologic condi-tion. This type of study has not been performed. Since theammonium excreted is ultimately derived from the ammoniumsecreted by the proximal tubule, I would speculate that thecontribution from the "direct" route exceeds the trapping route byat least twofold.DR. HARRINGTON: Drawing from your own experience or theliterature, can you quantitate the common causes of hyperkalemiain hospital patients? That is, how much of the hyperkalemia is dueto disease, and how much is due to administered potassium?DR. DUBOSE: We have not quantitated such data from ourhospital. The sources of exogenous potassium are significant, butI think that the incidence of drug nephrotoxicity as a cause ofhyperkalemia is increasing. Patients often receive potassium loadsand drugs that are inappropriate for the degree of renal function.Examples include nonsteroidal anti-inflammatory drugs, ACEinhibitors, and potassium-sparing diuretics. At least three groupsof patients are at particular risk: (1) Patients with HIV or AIDS,with a life-threatening infection, who are receiving pentamidine ortrimethoprim. (2) Patients with diabetes mellitus; very often thesepatients receive a multitude of drugs including nonsteroidalagents and potassium-sparing diuretics, as well as ACE inhibitors,all of which, especially in combination, can cause severe hyperka-lemia and metabolic acidosis. (3) Hypoxic patients in the criticalcare setting who develop selective hypoaldosteronism with orwithout heparin administration.DR. SINUH: As you stated, one of the drugs implicated as a causeof drug-induced hypoaldosteronism and hyperkalemia is heparin,which appears to inhibit 18-hydroxylase in the adrenal gland,thereby impeding aldosterone biosynthesis [75, 76]. Given thatheparin is used so frequently, I'm surprised that we do not seemore hyperkalemia among our patients in the hospital. Is it adose-related effect? Are other factors important?DR. DUBOSE: Heparin-associated hyperkalemia is limited tocritically ill patients and is not dose-related. As I said, the mostcommon co-morbidity is hypoxemia.DR. MADIAs: My experience is that even low-dose heparin canproduce hyperkalemia but not in the presence of normal renalfunction. Almost all the patients I have seen have had underlyingrenal dysfunction or another independent reason for hyperkale-mia.DR. SINGH: Is the development of heparin-induced hyperkale-mia observed in patients treated with low-molecular-weight hep-arm?DR. DuBosc: I am not aware of any reports to this effect.DR. IOANNIs GIATRAS (Renal Fellow, New England MedicalCenter): Does heparin administered during hemodialysis decreasethe level of plasma aldosterone?DR. DuBosE: I am unaware of any measurements in theliterature.DR. MADIAS: Have you had the opportunity to study the effectsof lithium?DR. DuBosE: We have studied lithium nephrotoxicity in the rat, Nephrology Forum: Hyperkalemic hyperchioremic metabolic acidosis601but the lithium model was not associated with hyperkalemia.These animals developed metabolic acidosis as a result of a"pump" defect.DR. HARRINGTON: You discussed hypokalemia and the impacton ammonium secretion in experimental animals that also hadchloride-depletion metabolic alkalosis. What is the effect ofdiet-induced hypokalemia on ammonium secretion quantitativelyin the absence of chloride-depletion metabolic alkalosis?DR. DUBOSE: We have not studied pure hypokalemia, and I amnot aware of any study in humans or in animal models onammonium transport in hypokalemia without alkalosis.DR. ANDREW J. KING (Division of Nephrology, New EnglandMedical Center): In the treatment of type-TV renal tubular acido-sis, is there a role for poorly absorbed anions? If you were goingto design an approach that employed poorly absorbed anions,what would you use?DR. DUBOSE: What you've proposed seems logical. Sincesulfate is not tolerated well, bicarbonate is reasonable therapy fora "voltage" lesion. One might need to combine bicarbonateadministration with acetazolamide. Furosemide, rather thanacetazolamide, might work, as it has other beneficial effects,including an increase in potassium excretion.DR. MITCHELL S. JACOBSON (Renal Fellow, Division of Nephrol-ogy, New England Medical Center): Trimethoprim has been de-scribed as an important cause of hyperkalemia in patients withAIDS. The dose of trimethoprim in HIV-infected patients is oftenmuch greater than that given for conventional infections. How-ever, in rare cases hyperkalemia has been described in elderlypatients given conventional doses. In fact, I can recall during myfellowship a diabetic patient taking oral trimethoprim-sulfa-methoxazole who developed life-threatening hyperkalemia. Canyou comment on the occurrence of hyperkalemia with conven-tional doses of trimethoprim? What conditions would predisposea patient to hyperkalemia in this setting?DR. DuBosu: The same phenomenon has been reported inchildren taking higher doses of trimethoprim in the absence ofsignificant renal insufficiency. These agents work from the luminalside, so I assume that it is the drug concentration in the urine thatmatters. It's more likely that hyperkalemia would be observed inprerenal conditions.DR. HARRINGTON: Could you define disequilibrium pH for theacid-base neophyte?DR. DuBosE: Disequilibrium pH is the difference between pHmeasured in situ and pH attained when the tubule fluid isremoved and allowed to come to equilibrium. Obviously thistechnique applies only to the micropuncture or microperfusionsetting.Reprint requests to Dr. T D. DuBose, Jr., University of Texas MedicalSchool-Houston, 6431 Fannin, MSB 4.136, Houston, Texas 77030, USAReferences1. WRIG

HT FS, GIEBISCH G: Regulation of potassium excretion, in TheKidney, Physiology and Pathophysiology, edited by SELDIN DW,GIEBIscH G, New York, Raven Press, 1992, p 22092. WEST ML, MARSDEN PA, RICHARDSON MA, ZETFLE RM, HALPERINML: New clinical approach to evaluate disorders of potassium excre-tion. Miner Electrol Metab 12:234—238, 19863. WEST ML, BENDZ 0, CHEN CB, SINGER GG, RICHARDSON RMA,SONNENBERG H, HALPERIN ML: Development of a test to evaluate thetranstubular potassium concentration gradient in the cortical collect-ing duct in vivo. Miner Electrol Metab 12:226—233, 19864. KAMEL KS, ETHIER JH, QUAGGIN S, LEVIN A, ALBERT S, CARLISLEEJF, HALPERIN M: Studies to determine the basis for hyperkalemia inrecipients of a renal transplant who are treated with cyclosporine. JAm Soc Nephrol 2:1279—1284, 19915. SAJO TM, GOLDSTEIN MB, SONNENBERG H, SM0 TM, GOLDSTEIN MB,SONNENBERG H, STEINBAUGH BJ, WILsON DR, HALPERIN ML: Sites ofammonia production to tubular fluid in rats with chronic metabolicacidosis. Kidney mt 20:353—358, 19816. GooD DW, BURG MB: Ammonia production by individual segmentsof the rat nephron. J Clin Invest 63:602—610, 19847. GooD DW, DUBOSE TD JR: Ammonia transport by early and lateproximal convoluted tubule of the rat. J Clin Invest 79:689—69 1, 19878. HWANG J-J, CURTHOYS NP: Effect of acute alterations in acid-basebalance on rat renal glutaminase and phosphoenolpyruvate carboxyki-nase gene expression. J Biol Chem 266:9392—9396, 19919. KAISER 5, CURTHOYS NP: Effect of pH and bicarbonate on phos-phoenolpyruvate carboxykinase and glutaminase on mRNA levels incultured renal epithelial cells. J Biol Chem 266:9397—9407, 199110. KINSELLA IL, ARONSON PS: Interaction of NH4 and Li with therenal microvillus membrane Na-H exchanger. Am J Physiol 241:C220—C226, 198111. NAGAMI GT, SONU CM, KUROKAWA K: Ammonia production byisolated mouse proximal tubule perfused in vitro. Effect of metabolicacidosis. J Clin Invest 78:124—129, 198612. HAMMLL,TRIGG D, MARTIN D, GILLESPIE C, BUERKERT J: Transportof ammonia in the rabbit collecting tubule. J Gun Invest 75:478—485,198513. GARVIN JL, BURG MB, KNEPPER MA: NH3 and NH4 transport byrabbit renal proximal straight tubules. Am J Physiol 21:F232—F239,198714. KURTZ I, STAR R, BALABAN RS, GARVIN JL, KNEPPER MA: Sponta-neous luminal disequilibrium pH in S3 proximal tubules. Role inammonia and bicarbonate transport. J Clin Invest 78:989—996, 198615. DUBOSE TD JR, LUCCI MS, HOGG RJ, PUCACCO LR, KoKEo JP,CARTER NW: Comparison of acidification parameters in superficialand deep nephrons of the rat. Am J Physiol 244:F497—F503, 198316. GooD DW, KNEPPER MA, BURG MB: Ammonia and bicarbonatetransport by thick ascending limb of rat kidney. Am J Physiol 247:F35—F44, 198417. GooD DW: Ammonium transport by the thick ascending limb ofHenle's ioop. Ann Rev Physiol 5 6:623—647, 199418. WATTS BA, GooD DW: Effects of ammonium on intracellular pH inrat medullary thick ascending limb: Mechanisms of apical membraneNH4 transport. J Gen Physiol 103:917—936, 199419. BOIM MA, Ho K, SHUCK ME, BIENKOWSKI MJ, BLOCK JH, SLIGHTOMIL, YANG Y, BRENNER BM, HEBERT SC: ROMK inwardly rectifyingATP-sensitive K channel II. Cloning and distribution of alternativeforms. Am J Physiol 268 (Renal Fluid Electrol Physiol 37):F1132—F1140, 199520. DUBOSE TD JR, GooD DW: Role of the thick ascending limb andinner medullary collecting duct in the regulation of urinary acidifica-tion. Semin Nephrol 11:120—128, 199121. GooD DW, CAFLISCH DR, DUBOSE TD JR: Transepithelial ammoniaconcentration gradients in inner medulla of the rat. Am J Physiol252:F491—F500, 198722. WALL SM: NH4 augments net acid secretion by a ouabain-sensitivemechanism in isolated perfused inner medullary collecting ducts. Am JPhysiol 270 (Renal Fluid Electrol Physiol) 39:F432—F439, 199623. DUBOSE TD JR, GOOD DW, HAMM LL, WALL SM: Ammoniumtransport in the kidney: New physiologic concepts and their clinicalimplications. JAm Soc Nephrol 1:1193—1203, 199124. GOLDSTEIN MB, BEAR R, RICHARDSON RMA, MARSDEN PA, HAL-PERIN ML: The urine anion gap: A clinically useful index of ammo-nium excretion. Am J Med Sci 292:198—202, 198625. CARLISLE EJF, DONNELLY SM, HALPERIN ML: Renal tubular acidosis(RTA): Recognize the ammonium defect and pHorget the urine pH.Pediatr Nephrol 5:242—246, 199126. DUBOSE TD JR, GooD DW: Effects of chronic hyperkalemia on renalproduction and proximal tubule transport of ammonium in the rat.Am J Physiol 260:F680—F687, 199127. DUBOSE TD JR, Gooo DW: Chronic hyperkalemia impairs ammo-nium transport and accumulation in the inner medulla of the rat.J Clin Invest 90:1443—1449, 199228. DUBOSE TD JR, Goon DW: Effects of chronic chloride depletion 602NephrologyForum: Hyperknlemic hyperchloremic metabolic acidosismetabolic alkalosis on proximal tubule transport and renal productionof ammonium. Am J Physiol 269 (Renal Fluid Electrol Physiol38):F508—F514, 199529. GooD DW: Active absorption of NH4* by rat medullary thickascending limb. Am J Physiol 255 (Renal Fluid Electrol Physiol38):F78—F87, 198830. DUBOSE TD JR, CAFLISCH DR: Validation of the U-B PCO2 gradientas an index of distal nephron acidification in experimental models ofdistal renal tubular acidosis. J Gun Invest 75:1116—1123, 198531. DUBOSE TD JR, CAFLISCH CR: Effect of selective aldosteronc defi-ciency on acidification in nephron segments of the rat inner medulla.J Clin Invest 82:1624—1632, 198832. HUSTED RF, STEINMETZ PR: The effects of amiloride and ouabain onurinary acidification by turtle bladder. J Pharmacol Exp Ther 2 10:264—268, 197933. DuBose TD JR: Experimental models of distal renal tubular acidosis.Semin Nephrol 10:174—179, 199034. LASKI ME, KURTZMAN NA: Characterization of acidification in thecortical and medullary collecting tubule of the rabbit. J Clin Invest72:2050—2059, 198335. SABATINI S, KURTZMAN NA: Enzyme activity in obstructive uropathy:The biochemical basis for salt wastage and the acidification defect.Kidney mt 37:79—84, 199036. PURCELL H, BASTANI B, HARRIS KPG, HEMKEN P, KLAHR S, GLUCK S:Cellular distribution of H ATPase following acute unilateral ureteralobstruction in rats. Am J Physiol 261:F365—F376, 199137. HULTER HN, ILNtou LP, HARBOTrLE JA, SEBASTIANA:Impairedrenal H secretion and NH3 production in mineralocorticoid-deficientglucocorticoi

d-replete dogs. Am J Physiol 326:F136—F146, 197938. AL-AWQATI Q,NORBYLH, MUELLER A, STEINMETZ PR: Character-istics of stimulation of H transport by aldosterone in turtle urinarybladder. J Gun Invest 58:35 1—358, 197639. SEBASTIAN A, SurroN JM, HULTER HN, SCHAMBELAN M, POLER SM:Effect of mineralocorticoid replacement therapy on renal acid-basehomeostasis in adrenalectomized patients. Kidney mt 18:762—769,198040. SZYLMAN P, BEITER OS, CHAIMOWITZ C, ROSLER A: Role of hyper-kalemia in the metabolic acidosis of isolated hypoaldosteronism.N Engi J Med 294:361—365, 197541. SEBASTIAN A, SCHAMBELAN M, LINDENFELD 5, MORRIS RC JR:Amelioration of metabolic acidosis with fludrocortisone therapy inhyporeninemic hypoaldosteronism. N EngI J Med 297:576—582, 197742. SCI-IAMBELAN M, SEBASTIAN A, BIGLIERI EG: Prevalence, pathogene-sis, and functional significance of aldosterone deficiency in hyperkale-mic patients with chronic renal insufficiency. Kidney mt 17:89—94, 198043. DEFRONZO RA: Nephrology Forum: Hyperkalemia and hyporenine-mic hypoaldosteronism. Kidney mt 17:118—134, 198044. DUBOSE TD JR, COGAN MG, RECTOR FC JR: Acid-base disorders, inThe Kidney (5th ed), edited by BRENNER BM, RECTOR FC JR,Philadelphia, Saunders, 1995, p 92945. OETLIKER OH, ZURBRUGG RP: Renal tubular acidosis in salt-losingsyndrome of congenital adrenal hyperplasia (CAH). J Clin EndocrinolMetab 31:447—450, 197046. WEIDMAN P, REINHART R, MAXWELL MH, ROWE P, COBURN JW,MASSRY SG: Syndrome of hyporeninemic hypoaldosteronism andhyperkalemia in renal disease. J Clin Endocrinol Metab 36:965—977,197347. OH MS, CARROL HJ, CLEMMONS JE, VAGNUCCI AH, LEVISON SP,WIJANG ESM: A mechanism for hyporeninemic hypoaldosteronism inchronic renal disease. Metabolism 23:1157—1166, 197448. DAVENPORT MW, ZIPSER RD: Association of hypotension with by-perreninemie hypoaldosteronism in the critically ill patient. ArchIntern Med 143:735—737, 198349. Amoi'niuAi I, WANG Y, HORTON R: Tumor necrosis factor andinterleukin-1 may regulate renin secretion. Endocrinolo 126:273—278, 199050. O'KELLY R, MAGEE F, MCKENNA TF: Routine heparin therapyinhibits adrenal aldosterone production. J C/in Endocrinol Metab56:108—112, 198351. ARMANINI D, KUFIALE U, STRASSER T, DoRR H, BUTENANDT I,WEBER PC, STOCKIGT JR, PEARCH P, FUNDER JW: Aldosteronereceptor deficiency in pseudohypoaldosteronism. N Engi J Med 313:1178—1183, 198552. CHANG SS, GRUNDER 5, HANAKOGLU A, ROSLER A, MATHEW PM,HANUKOGLU I, SCHILD L, LU Y, SHIMKETS RA, NELSONWILLIAMS C,ROSSIER BC, LIFTON RP: Mutations in subunits of the epithelialsodium channel cause salt wasting with hyperkalemic acidosis,pseudohypoaldosteronism type 1. Nat Genet 12:248—253, 199653. SCHAMBELAN M, SEBASTIAN A, RECTOR FC JR: Mineralocorticoid-resistant renal hyperkalemia without salt wasting (type II pseudohy-poaldosteronism): Role of increased renal chloride reabsorption.Kidney mt 19:716—727, 198154. SIMoN DB, FARFEL Z, ELLISON D, BIA M, Tucct J, LwrON RP:Examination of the thiazide-sensitive Na-Cl cotransporter as a candi-date gene in Gordon's syndrome. JAm Soc Nephrol 6:632A, 199555. RODRIGUEZ-SORIANO J, VALLO A, QUINTELA MJ, OLIvEROs R, UBE-TAGOYENE M: Normokalemic pseudohypoaldosteronism is present inchildren with acute pyelonephritis. Acta Pediatr 81:402—406, 199256. DUBOSE TD JR, ALPERN R: Renal tubular acidosis, in The MetabolicBasis of Inherited Disease (7th ed), edited by SCRIVER CR, BEAUDETAL, SLY WS, VALLE D, New York, McGraw-Hill, 1995, p 365557. MEHTA BR, CAVALLO T, REMMERS AR JR, DUBOSE TD JR: Hy-poreninemic hypoaldosteronism in a patient with multiple myeloma.Am J Kidney Dis 4:175—178, 198458. BATLLE DC: Segmental characterization of defects in collecting tubuleacidification. Kidney Int 30:546—554, 198659. PONCE SP, JENNINGS AE, MADIAS NE, HARRINGTON JT: Drug-induced hyperkalemia. Medicine 64:357—370, 198560. KLEYMAN TR, ROBERTS C, LtNG BN: A mechanism for pentamidine-induced hyperkalemia: Inhibition of distal nephron sodium transport.Ann Intern Med 122:103—106, 199561. SCHLANGER LE, KLEYMAN TR, LING BN: K-sparing diuretic actionsof trimethoprim: Inhibition of Na channels in A6 distal nephroncells. Kidney Int 45:1070—1076, 199462. VALAZQUEZ H, PERAZELLA MN, WRIGHT FS, ELLISON DH: Renalmechanisms of trimethoprim-induced hyperkalemia. Ann Intern Med1193:296—301, 199363. KAMEL KS, ETHIER JH, QUAGGIN 5, LEVIN A, ALBERT 5, CARLISLEEJF, HALPERIN ML: Studies to determine the basis of hyperkalemia inrecipients of a renal transplant who are treated with cyclosporine. JAm Soc Nephrol 2:1279—1284, 199164. SANDS JM, MCMAHON SJ, TUMLIN JA: Evidence that the inhibition ofNa'7K-ATPase activity by FK506 involves calcineurin. Kidney mt46:647—652, 199465. BANTLE JP, NATH KA, SUTHERLAND ER, NAJARIAN JS, FERRIS TF:Effects of cyclosporine on the rcnin-angiotensin-aldosterone systemand potassium excretion in renal transplant recipients. Arch InternMed 145:505—508, 198566. AGUILERA 5, DERAY G, DESJOBERTT H, BENHMIDA M, HOANG PL,JACOBS C: Effects of cyclosporine on tubular acidification function inpatients with idiopathic uveitis. Am J Nephrol 12:425—430, 199267. EIAM-ONG 5, HILDEN SA, JOHNS CA, MADIAS NE: Stimulation ofbasolateral Na-HCO3-cotransporter by angiotensin II in rabbitrenal cortex. Am J Physiol 265:F195—F203, 199368. NAGAMI GT: Effect of luminal angiotensin II on ammonia productionand secretion by mouse proximal tubules. Am J Physiol 269:F86—F92,199569. SZYLMAN P, BETITER OS, CHAIMOVITZ C, ROSLER A: Role of hyper-kalemia in the metabolic acidosis of isolated hypoaldosteronism.N Engl J Med 294:361—365, 197670. SCHAMBELAN M, SEBASTIAN A: Hyporeninemic hypoaldosteronism.Adv Intern Med 24:385, 197871. LEVEL M, GROSE JH: Angiotensin II effect on plasma steroids inselective hypoaldosteronism. Horm Metab Res 14:432—436, 198272. WILLIAMS GH: Hyporeninemic hypoaldosteronism. N Engi J Med314:1041, 198673. NATH KA, HOSTETIER MK, HOSTETTER TH: Pathophysiology ofchronic tubulo-interstitial disease in rats. J C/in Invest 76:667—675,198574. ALPERN RJ: Nephrology Forum: Trade-offs in the adaptation toacidosis. Kidney mt 47:1205—1215, 199575. PHELPS KR, OH MS, CARROLl. Hi: Heparin-induced hyperkalemia.Report of a case. Nephron 25:254, 198076. SHERMAN RA, RUDDY MC: Suppression of aldosterone production bylow-dose heparin. Am J Nephrol 6:16