View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Kidney International,Vol.33(1988), pp. 900--91 / NEPHROLOGY FORUM

Aminoglycoside nephrotoxicity Principal discussant: H. DAVID HUMES

Veterans Administration Medical Center, Ann Arbor, Michigan

Urinalysis showed granular casts and a substantial number of renal epithelial cells. Urine sodium concentration was now 45 mEq/liter; Editors urine osmolality, 300 mOsm/kg H,O; and urine FEN.,1.8%.At that JORDANJ.COHEN time, peak and trough serum gentamicin concentrations were 15 sg/ml and 5 g/ml, respectively. The..gentamicin dose was reduced to 70 mg JOHN T. HARRINOTON every 24 hours, but serum creatinine continued to rise slowly over the JEROME P.KASSIRER next 4 days to 5.7 mg/dl despite appropriate dosing adjustments to NIc0LAOs E. MADIAS maintain peak and trough levels below toxic serum levels. After 10 days, gentamicin therapy was discontinued, but the patient's serum creatinine rose to a maximal value of 7.8 mg'dl on day 14. At no time did Managing Editor the patient develop hyperkalemia, fluid overload, or uremic symptoms. CHERYL J. ZUSMAN The serum creatinine persisted at 7.8 mg/dl until day 17. when it began to decline; ultimately it reached a level of 1.2 mg/dl on the 28th day after the patient was discharged from the hospital. Universityof Chicago Pritzker School of Medicine and Discussion Tufts University School of Medicine DR.H. DAVIDHUMES(Chief, Medical Service, Veterans AdministrationMedical Center; and Professor and Associate Chairman, Department of Internal Medicine, Unii'ersitv of Michigan Medical Center, Ann Arbor, Michigan): Aminoglyco- side antibiotics, despite their nephrotoxicity, continue to be a Case presentation mainstay in the clinical management of gram-negative infec- tions. Gram-negative organisms account for the majority of A 74-year-oldman with a historyof peptic ulcer disease arrived at the hospital-acquired infections [I], and the occurrence of amino- hospital complaining of persistent midepigastric pain that had lasted 3 days but which had worsened over the previous 24 hours. Because he glycoside-induced acute renal failure remains cdmmonplace. A was nauseated, he had taken little by mouth except ginger ale over the recent prospective study of hospital-acquired acute renal failure 2 days prior to admission. Physical examination revealed a temperature in 2116 patients revealed that 11% of the cases probably were of 99.rF, an orthostatic blood pressure drop from 120/82mm Hg lying related to aminoglycoside use [2]. in a variety of clinical to 90/58 mm Hg standing, and an orthostatic pulse rate elevation from studies, the use of gentamicin has been associated with acute 90/minto 120/mm. He had abdominal tenderness without rebound in the midepigastrium. Laboratory data disclosed BUN, 40 mg/dl; serumrenal failure in 8% to 26% of patients [3]. The understanding of creatinine, 1.8 mgldl; and amylase, 190 Somogyi units/dI (normal, aminoglycoside nephrotoxicity therefore has substantial clinical 80—200 units). Urinalysis was normal and examination of a random relevance. In this Forum, I will summarize the renal handling of urine collection revealed a sodium concentration of 4 mEq/liter, urine the aminoglycosides, the pathogenetic mechanisms of nephro- osmolality of 750 mOsm/kg HO, and FEN.,of0.4c. The patient was treated with intravenous fluids and nasogastrictoxicity, and the clinical aspects of aminoglycoside-induced suction. Six hours after admission, he developed severe midepigastric acute renal failure. I will place particular emphasis on recent pain with rebound tenderness and a temperature of l02.5F. The clinical data that have increased our understanding of the interaction of diagnosis of perforated peptic ulcer was confirmed at operation and an aminoglycosides with the renal tubular cell and the effects of exploratory taparotomy was done. Immediately postoperatively, the this interaction on cellular function and integrity. BUN was 28 mg/dI and serum creatinine 1.4 mg/dl. The patient was given clindamycin, 500 mg every 6 hours, and gentarnicin. 80 mg every 8 hours. Three days later, blood cultures grew Pseudomonas aerugi- Renal handling of aminoglycosides nosa sensitive to gentamicin and tobramycin. Both antibiotics were Most experimental evidence indicates that net reabsorption continued. The patient stabilized and improved, maintaining a urine of aminoglycosides occurs in the proximal tubule by means of a output of 1500 mI/day. Six days after operation. laboratory studies revealed a BUN of 30 mg/dl and a serum creatinine of 2.7 mg/dl; no high-capacity transport system [4. 5]. Little direct evidence BUN and creatinine determinations had been done in the interval.supports increases in luminal concentration of aminoglycosides along the proximal tubule. Net drug secretion might occur in the early proximal tubule at high doses, whereas net secretion in Presentation of this Forum is made possible by grants from Merckthe late proximal tubule ol' juxtamedullary nephrons might Sharp & Dohme; Pfizer, Incorporated; Sandoz. Incorporated; and E. R. occur at lower doses [5]. This nephronal heterogeneity in drug Squibb & Sons, Incorporated. handling explains most features observed in the experimental © 1988 by the International Society of Nephrology setting, and I will discuss these later.

900 Nephrology Forum: Aminoglycoside nephroioxicily 901

Roth luminalreabsorption andbasolateraluptake probably Table I. Risk factors in aminoglycoside nephrotoxicity accountfor the high drug levels achieved in renalcorticaltissue 1. Dose and duration of drug treatment in experimental animals and in humans; variability in these 2. Recent aminoglycoside therapy processes, in turn, accounts for the variations in cortical drug 3. Preexisting renal insufficiency content noted among different aminoglycosides [6]. Luminal 4. Preexisting hepatic disease 5. Elderly age drug uptake occurs after aminoglycoside binding to brush- 6. Concomitant nephrotoxic drug administration border membrane sites and is followed by absorptive pinocyto- 7. Volume depletion sis, the dominant pathway for intracellular drug accumulation 8. Potassium depletion [7]. Drug uptake is followed by lysosomal processing, although 9. Magnesium depletion temporal and quantitative evidence suggests important interac- tions with other intracellular organelles [8]. After being taken up by the renal tubular cell, aminoglycosides reside in a poorly exchangeable pool; they therefore have prolonged tissue half- Proximal tubular transport processes also deteriorate during lives [9]. aminoglycoside toxicity. This alteration results in glycosuria, Aminoglycosides produce tubular cell necrosis, which isaminoaciduria, tubular proteinuria, and transport defects con- largely confined to the proximal convoluted tubule and parssistent with a Fanconi-like syndrome [13]. Aminoglycosides recta [10, 11]. The earliest lesion, seen by electron microscopy, also can induce renal potassium and magnesium wasting, with is an increase in the number and size of secondary lysosomes resulting hypokalemia and hypomagnesemia [15]. The selective (cystosegrosomes, phagosomes) [10]. The development of ly- effects on the renal handling of these intracellular cations may sosomal alterations during aminoglycoside administration does reflect a specificity of action of the aminoglycosides. namely, not necessarily progress to renal cell necrosis and organ failure, alteration of plasma membrane transport of, or permeability to, however [12]. Additionally, these morphologic alterations are these particular ions. not specific for aminoglycoside toxicity and can occur in other Polyuria and nephrogenic diabetes insipidus develop early in tissues alter focal cell injury induced by a variety of toxins [13]. aminoglycoside nephrotoxicity [15]. Urinary concentrating de- Following lysosomal injury, injury to renal proximal tubulefects and resistance to antidiuretic hormone characterize most cells continues, as reflected by mitochondrial swelling and states of tubulointerstitial injury and are due, in large part, to an increasing loss of brush-border membranes. The degree of cellinability of the injured kidney to maintain the hypertonicity of injury in the proximal tubule correlates reasonably well with the the medullary interstitium. For the aminoglycosides, inhibition decline in renal excretory function. The pattern of substantial of antidiuretic hormone activity (via interference with the involvement of S1 and S2 proximal tubular segments in amino- action of adenylate cyclase) may be an additional pathogenetic glycoside injury contrasts with the markedly greater involve-cause for this functional defect [19]. ment of the S3 segments in heavy-metal nephrotoxicity and A variety of factors predispose to aminoglycoside nephrotox- ischemia. Perhaps this difference is due to a much higher toxin icity (Table I). The dose and duration of drug administration are load presented to the more proximal S1 and S2 cells along the the two most important. Higher antibiotic doses result in higher tubule. Distal tubules and glomeruli reveal little damage by light serum drug levels and more rapid aminoglycoside accumulation and transmission electron microscopy [10, 14]. in the kidney. Prolonged therapy increases the risk that toxic Regeneration of proximal tubule cells begins during the stage concentrations will persist within renal parenchyma. A relation- of patchy necrosis. The relatively undifferentiated, immature, ship between serum levels, renal parenchymal concentration, regenerating cells subsequently begin to regain normal height and nephrotoxicity for any given aminoglycoside has been and structure. Foci of interstitial inflammatory infiltrates appear suggested [20, 21]. The risk of aminoglycoside toxicity is throughout the cortex during this stage and can become more increased not only in patients given continuous prolonged prominent with time [15]. Eventually, most areas of the affected therapy, but also in those given the drug in repeated courses kidney regain normal architecture and function, but residualseparated by a few days or weeks. scars containing collapsed, atrophic tubules can develop in the The probability that a given patient will develop nephrotox- cortex [16]. 1 should note the interesting fact that regeneration icity when given an aminoglycoside is increased by several can continue within the renal cortex even in the presence offactors including (1) increasing age, (2) preexisting renal insuf- continued aminoglycoside administration and toxic drug levels ficiency [22], (3) concomitant administration of intrinsically in the renal cortex [17]. nephrotoxic agents (such as cyclosporine [23]), (4) preexisting Aminoglycoside nephrotoxicity is characterized by a variety liver disease, (5) severe hypertension [24], and (6) fluid and of renal functional alterations. Renal excretory failure with near electrolye abnormalities. Volume depletion potentiates nephro- cessation of effective glomerular filtration rate is but the finaltoxicity, presumably because of elevated rates of aminoglyco- manifestation of this clinical disorder. Prior to this evenL more side accumulation in the renal cortex [25]. In experimental subtle nephronal derangements, dominated by proximal tubular animals, potassium depletion condenses the usual time frame abnormalities, occur in humans and animals. for aminoglycoside-induced acute renal failure [26]. Finally, The initial renal manifestation of aminoglycoside toxicity isgentamicin nephrotoxicity is potentiated by magnesium deple- enzymuria. As early as 24 hours after a single therapeutic dose tion [27]. of aminoglycoside, the urinary excretion of a variety of brush- border membrane enzymes increases and progressively rises as Diagnosis'and treatment therapy continues [18]. Increases in the urinary excretion of Aminoglycosidenephrotoxicity typically is associated with various lysosomal enzymes also can be seen. nonoliguric acute renal failure, that is, azotemia in the presence 902 NepI.rology Forwn: A ininog!vcoside nephrowxicity

of urine output between 1 and 2 liters/day.Urineoutput is, tial. Once established, aminoglycoside-induced acute renal fail- therefore,an unreliable marker for the development of amino-ure must be handled like any other type of acute renal failure, glycoside nephrotoxicity. Declines in glomerular filtration ratewith careful fluid and electrolyte management and, if necessary, and elevations in serum creatinine usually are not apparent untildialysis. after 7 to 10 days of aminoglycoside treatment. However, the presence of any of the risk factors I mentioned can condense Co,nparafive clinical nephrotox:city this time frame. Before concluding that a patient has nephro- Thecomparative nephrotoxicity of the aminoglycosides has toxic acute renal failure, or aminoglycoside-induced acute renalbeen established for gentamicin, tobramycin, amikacin, and failure, in particular, reversible prerenal and postrenal causesnetilmicin [131. Inferences concerning clinical toxicity based on must be excluded. animal studies, however, must be made with caution, because When using these potentially life-saving antibiotics, we areboth the onset and severity of aminoglycoside nephrotoxicity faced with competing goals: on the one hand we want to achievedepend on the species and strain of experimental animal uti- high peak and mean plasma aminoglycoside levels duringlized, gender, age, and dosing schedule [13]. treatment to eliminate the invading bacteria [28] and, on the The many clinical studies of aminoglycoside toxicity have other hand, we want to avoid nephrotoxicity. Unfortunately,been the subject of a recent review that analyzed 144 published effective therapeutic doses for antibacterial activity are verytrials involving nearly 10,000 patients [30]. The findings under- near potentially toxic doses. Because many factors such asscore the difficulty inherent in comparing relative drug nephro- renal function, age, and sodium balance can affect serum levelstoxicities among studies. The incidence of nephrotoxicity in of the aminoglycosides, assays monitoring antibiotic levels areprospective comparative trials varies considerably, thus em- required to maintain effective but nontoxic antibacterial con-phasizing the need for rigorously defined nephrotoxicity crite- centrations. Avoidance of peak serum concentrations of genta-ria, patient randomization, concurrent controls, and double- micin or tobramycin greater than 10 g/ml and trough serumblind evaluations. Overall, gentamicin and tobramycin appear concentrations greater than 2 jig/mI have been associated withto be equally nephrotoxic and are slightly more nephrotoxic lower rates of nephrotoxicity [21]. Undue emphasis on thethan are amikacin and netilmicin; amikacin and netilmicin trough aminoglycoside level may be inappropriate, however,appear to be equally nephrotoxic. When prospective compara- because elevations of trough serum aminoglycoside levels fre-tive trials are considered separately, gentamicin appears to be quently are the result, not the cause, of renal insufficiency [29]. the most nephrotoxic drug, followed in order of decreasing Various formulas have been devised for calculating dose alter-nephrotoxicity by tobramycin, amikacin, and netilmicin. It is ations in the presence of renal insufficiency so that serum druginteresting yet unexplained why the incidence of gentamicin levels can be maintained below these toxic concentrations.nephrotoxicity is greater in trials comparing the drug with These formulas provide only an approximation of the appropri-netilmicin than in those comparing gentamicin with other ami- ate dose. Serum level monitoring must be employed to ensure anoglycosides. Further, in a comparative trial with tobramycin, correct dosing schedule given the multiple factors besides renalgentamicin appeared to be more nephrotoxic than in an earlier function that can influence antibiotic excretion, serum levels,comparative trial with amikacin, even though the studies were and renal accumulation. These final adjustments in dose, basedundertaken by the same group of investigators using similar on serum antibiotic concentrations, are especially necessary incriteria [29]. the elderly, in. whom serum creatinine is not as accurate a These studies have prompted several suggestions regarding reflection of renal function as in younger patients. aminoglycoside administration in older patients and in those Every effort should be made to avoid those factors known towith preexisting renal insufficiency or liver disease. Because provoke toxicity. The patient receiving an aminoglycosidethese patients are at greater risk of toxicity, tobramycin rather should be well hydrated to avoid volume depletion, and shouldthan gentamicin is a reasonable choice when an aminoglycoside be potassium and magnesium replete, especially in the clinicalis required. Amikacin and, especially, netilmicin appear to be settings requiring concurrent treatment with potent diureticless nephrotoxic therapeutic alternatives to gentamicin and agents. Finally, if therapy was begun empirically, continuationtobramycin, but these should be reserved for bacterial strains of aminoglycoside therapy should be based on the results ofresistent to gentamicin and tobramycin. In contrast to amika- bacterial cultures. If the serum creatinine rises and glomerularcm, netilmicin is associated with substantially less cochlear and filtration rate declines, one should consider discontinuation ofvestibular toxicity [30]. I should emphasize that the choice of an the aminoglycoside if a different, but equally efficacious, anti-aminoglycoside antibiotic depends on the patient, clinical set- biotic can be substituted. If the aminoglycoside must be con-ting, local bacterial resistance patterns, and importantly, on tinued in the presence of developing acute renal failure, the onlyculture and sensitivity results when these are available, because reliable guide to proper dosage is the serum drug level. Even ifall aminoglycosides have the potential for producing nephro- the antibiotic is discontinued at the time the serum creatininetoxic complications. begins to rise, renal injury can continue and can be followed by severe acute renal failure. This phenomenon also occurs regu- Patho genesis larly in the animal models of gentamicin nephrotoxicity and Rena!aspects. The final common pathogenetic pathway for likely reflects the continued effects of the accumulation of highthe development of either ischemic or nephrotoxic acute renal antibiotic levels in tissue. Because even a modest decline infailure is renal tubular cell injury [31]. Progressive renal tubular glomerular filtration rate can be a signal of progressive renalcell injury initiates alterations at the nephronal level that damage, there is a premium on recognizing incipient renalultimately result in renal insufficiency. These alterations in- damage. To do so, daily assessment of renal function is essen-clude: (I) intratubular obstruction, (2) backleak of glomerular Nephrology Foru,n: A,ninog!yoside nepliroioxuitv 903 p'z.05 p< .05 150

0 C -r 0 (.) 100 —

C p<.00I p<.OI 0 C C.) 50 E C a) (D n=9 n=5 n6 fl=jfl9' Control Trypsin Chymotrypsin N-Ethylmal Phospholipase Phospholipase -elmide A C Fig. I. Specific binding of 3H gentamicin to isolated renal brush-border membranes aftertreatment with protein-modifying agentstrypsin, chymotrypsin, and N-ethylmaleimide or after treatmentwith phospholipasesA and C. Theseresults suggestthat the gentamicin-bindingsite is primarily phospholipid in nature. Adapted from Ref.37. filtrate through damaged tubular epithelium, and (3)reduction a function of the permeability of the glomerular capillary wall to in glomerular filtration rate of solute and water. Of these threewater and of the surface area available for filtration. Substantial factors, intratubular obstruction appears to be a major mecha-evidence has demonstrated that the intrarenal generation of the nism for a decline in glomerular filtration rate (GFR) in bothvasoconstrictor hormone angiotensin II (All) dramatically low- ischemic and nephrotoxic acute renal failure. Recent studiesers both renal blood flow (from renal arteriolar contraction) and have shown that if the intraluminal blockage to urine flow isKf (from mesangial cell contraction); these decreases lead to a diffuse enough to involve most functioning nephrons, thisreduction in glomerular capillary surface area [331. In principle, process will lead to a major decline in renal excretory functionboth alterations could diminish GFR. Despite much study, [31. 32]. Free-flow micropuncture studies of experimental ami-however, a definitive role for alterations in renal blood flow in noglycoside nephrotoxicity show that all functioning nephronsthe initiation or maintenance of nephrotoxic acute renal failure had elevated intratubular pressures compared with controlshas not been established. Manipulations of the renin-angioten- [30]. Single-nephron GFR (SNGFR) was diminished in involvedsin system have not yielded evidence of a role for this hormone tubules compared with controls and rose as intratubular pres-system in acute tubular necrosis. In this regard, the increase in sure was decreased. Microinfusion of involved tubules led toAll generation in the kidney is thought to represent a secondary marked increases in intranephronal pressures. Elevations ofderangement of lesser importance arising from the primary intratubular pressures in controls to levels observed in drug-process of renal tubular cell injury. The improvement of this treated rats led to reductions in SNGFR to levels observed insecondary change in GFR does not ameliorate the underlying injured tubules. primary tubular injury, and therefore any All-mediated changes The necrotic cells shed into the tubular lumen not only reducein glomerular hèmodynamics early in the process of renal cell renal excretory function by obstructing urine flow but also byinjury would be expected to be far overshadowed by the GFR leaving gaps along the tubular epithelia through which glomer-decline prompted by intratubular obstruction and backleak. ular filtrate can re-enter the circulation. With this disruption, Micropuncture studies have delineated several alterations in the effective GFR is compromised further. Recent studies.usingthe determinants of glomerular ultrafiltration in gentamicin microinjection techniques have correlated the magnitude ofnephrotoxicity in particular. Doses of gentamicin that did not backleak with the extent of severe tubular injury present inproduce severe tubular cell injury led to modest declines in acute tubular necrosis (ATN) [32]. GFR in the rat. The primary cause for this reduction in GFR The GFR also might decline in acute tubular necrosis becausewas a significant decline in Kç, although a small decline in of primary disturbances in the glomerular filtration process. Theglomerular plasma flow was noted in the experimental group reduction could be due to a fall in glomerular capillary hydro-given high doses. This decline in plasma flow contributed to the static pressure or renal blood flow, or to a decrease in thedecreased SNGFR 1341. Other data show that the angiotensin- glomerular capillary ultrafiltration coefficient, K1 [33J. The K isconverting enzyme inhibitor captopril almost completely pre- 904 Nephrology Forum: Aminoglycoside nephrotoxicity

vents the untoward effects of gentarnicin on glomerular hemo- Table 2. Membrane alterations in aminoglycoside nephrotoxicity dynamics [35]. Although these results are compatible with thePlasma membrane alterations hypothesis that gentamicin-jnduced functional derangements in 1. Phospholipid alterations: aminoglycoside binding to membrane glomerular hemodynamics are secondary to increased intrare- acidic phospholipids; changes in content and metabolism of mem- nal angiotensin II generation, 3 days of captopril treatment brane phospholipids; displacement of Ca +boundto phospholip- ids preceded this study. The possibility exists, therefore, that other2. Transport defects: inhibition of Na-K-ATPase and adenylate cy- determinants of nephrotoxicity, such as tissue drug concentra- clase; alterations in PAH transport; renal wasting of K and Mg*+ tion, are at play. in fact, further studies using gentamicin doses that produced both functional and morphologic evidence ofMitochondrial membrane alterations severe renal failure demonstrated that captopril treatment or1. In-viva gentamicin: decreased State 3 and DNP-uncoupled respira- tion pretreatment did not ameliorate renal functional, morphologic,2. In-vitro gentamicin: increased State 4, decreased State 3. and or glomerular ultrastructural features of aminoglycoside neph- DNP-uncoupled respiration secondary to alteration of Mg-con- rotoxicity [36]. The results of this study argue against a domi- trolled, monovalent cation permeability pathway; inhibition of nant role for the renin-angiotensin system in severe renal failure Ca transport induced by aminoglycosides. Lysosomal membrane alterations Cellularaspects. Anycomprehensive view of the pathogen- I. Lysosomal instability, in vitro and in vivo esis of acute tubular necrosis requires an explanation of the2. Diminished phospholipase activity biochemical alterations responsible for the loss of renal tubule cell integrity. Without tubule cell injury, the various factors responsible at the nephronal level for renal failure would not be initiated. Growing evidence suggests that nephrotoxins andtive binding studies of brush-border membrane vesieles and ischemia exert their primary detrimental effects on cells byliposomes with various aminoglycosides and polyamines lend disturbing the membranes of the cell [30]. Both the plasmafurther support for a predominant role of anionic phospholipids membrane and the intracellular organellar membranes are po-in drug binding [37, 39]. Study of in-vitro drug binding interac- tential sites of toxicity. Because of the transport properties oftions, however, fails to completely predict whole-organ neph- the renal tubular cell, its membranes are especially susceptiblerotoxicity [31]. Stepwise computer conformational analyses of to toxic insult. The transport of a multiplicity of compoundsaminoglycoside-phospholipid bilayer systems have suggested through the intracellular matrix of the renal tubular cell exposesthat hydrophobic interactions and lipid bilayer penetration are the membranes of its subcellular organdies, such as the mito-potentially important modifiers of simple electrostatic binding chondnal inner membrane, to potentially toxic concentrations[40]. Such sophisticated physical approaches might provide of a variety of substances. In addition, because the renal tubulegreater insight into the biochemical mechanisms of aminoglyco- avidly reabsorbs salt and water, substances are concentratedside nephrotoxicity, the creation of less nephrotoxic drugs, and within the tubular lumen, thereby exposing the luminal mem-an explanation for the failure of simple electrostatic formula- branes of renal tubular cells to potentially toxic concentrations.tions to predict binding and toxicity. These membrane alterations in the renal tubular cell produce Immediate and strong aminoglycoside-phospholipid binding multiple biochemical derangements that ultimately lead to cellcharacterizes the interaction of these drugs with the kidney. death. We are only now beginning to clarify the manner ofThe resulting membrane alterations therefore might figure im- interaction and the rank order of importance of the multipleportantly in the pathogenesis of aminoglycoside-induced acute biochemical abnormalities that are initiated by either toxins orrenal failure. Specific alterations in cellular phospholipids occur ischemia. in whole-animal, human, and cell-culture models of amino- Most recent studies have focused on the manner in whichglycoside nephrotoxicity [31]. Available evidence suggests that these antibiotics interact with the plasma membrane of the renalthese lipid compositional alterations involve all cellular organ- tubular cell and the important subcellular organeliar mem-elles and probably result from drug inhibition of multiple branes, primarily those of the mitochondria and the lysosomes.phospholipases of varying substrate specificity and cellular Investigation of the physiochemical nature of aminoglycoside-location as well as from redirection of glycerolipid biosynthesis membrane interactions, both in vivo and in vitro, led us to[41, 42]. Although a precise causal relationship between cellular identify acidic phospholipids as the molecular sites for amino-phospholipidosis and injury has not yet been demonstrated, glycoside binding [371. Other investigations also have demon-evidence linking cellular phospholipidosis to organ injury has strated global alterations in phospholipid metabolism and asso-been adduced in several models of drug toxicity involving ciated alterations in the lipid content of plasma and organellarcationic compounds [43]. membranes, altered plasma-membrane-bound enzyme activi- Several tubular cell membrane functional changes occur with ties, and disturbances in organellarfunction[l3]. The scope andaminoglycoside exposure. The activity of sodium-potassium implications of these findings as summarized in Table 2 will beATPase, the major regulator of the cation gradients across briefly considered. plasma cell membranes, critically depends on its surrounding Plasma membrane. Binding between aminoglycosides andphospholipid environment [44]. Neomycin inhibits sodium- membrane phospholipids appears to involve the electrostaticpotassium ATPase activity in vitro, and this effect is specifically interaction of the cationic, polybasic antibiotic and the acidicantagonized by phosphatidylinositol [45]. Administration of phospholipids, particularly the phosphoinositides [37, 38]. Alarge doses of gentamicin results in significant and reversible role for less-well-defined acidic phospholipid moieties in mem-declines in sodium-potassium ATPase activity in vivo [46]. brane binding has not been excluded, however [38]. Competi-Persistent decreases occur as a late effect and might he related Nep/tro!ogv !oru,n: A minoglycoside nephrozox,cuy 905

mitochondrial membrane for oxidative phosphorylation and its associated cation and anion transport process is thus a complex integrated function, depending on both the integrity and limited cation permeability of that membrane and the stability of the intracellular environment. Given the transporting and concen- trating properties of the renal tubular epithelium, substantial opportunity exists for the membrane-active aminoglycosides to 0C 60 affect bioenergetics. These effects occur both by means of direct effects on the mitochondrial membrane as well as by C50 means of alterations in the intracellular environment because of effects on other cellular membranes. These alterations in renal 40 cell bioenergetics probably contribute importantly to the patho- 30 genesis of cell injury produced by the aminoglycoside antibiot- ics. 20 The aminoglycosides have specific and reversible effects at the inner mitochondrial membrane, where membrane-bound 10 magnesium plays a critical role in the regulation of membrane transport processes. Gentamicin competes with magnesium for membrane sites, where magnesium limits membrane permeabil- 0.05 0.1 0.5 1.0 ity to monovalent cations. Gentamicin itself, however, is not as Molar multipleofgentamicin effective as magnesium in limiting monovalent cation perme- ability. This effect results in mitochondrial swelling and alter- ations in mitochondrial respiratory function [50, 51]. Other Fig.2. In-vitro and in-vivo dose responses of polyamino acid inhibition aminoglycoside antibiotics have mitochondrial effects similar to of membranebindingand nephrotoxicity of gentamicin. A refers tothose produced by gentamicin with a relative potency that polyasparagine membrane-binding inhibition: refers to polyaspara- gine nephrotoxicity inhibition; •refersto polyaspartic acid membrane- correlates with their cationicity [52]. Magnesium also acts at the binding inhibition; and 0 refers to polyaspartic acid nephrotoxicityinner mitochondrial membrane to limit mitochondrial calcium inhibition. These data demonstrate that these polyamino acids bothuptake. In this case, the effects of gentamicin are additive to the inhibit gentamicin binding to isolated renal brush-border membranes effects of magnesium on calcium uptake [53]. These alterations and protect against gentamicin nephrotoxicity at similar molar ratios. Reprinted from Ref. 77. in mitochondrial membrane alterations appear to relate, at least in part, to aminoglycoside-induced stimulation of the generation of reactive oxygen metabolites by renal cortical mitochondria [54]. to generalized membrane deterioration. The activity of the Aminoglycosides also have well-defined effects on lysosomal membrane-bound enzyme adenylate cyclase also has beenmembranes in vitro. Aminoglycosides inhibit the activity of shown to depend on its phospholipid environment [47]. Amino-lysosomal phospholipases A and C from renal cortex [55. 561. glycoside exposure inhibits fluoride-stimulated adenylate cy-At low concentrations, aminoglycosides, as a function of their clase of rat proximal tubular basolateral membrane in vivo andcationicity, stabilize isolated lysosomes, whereas at high con- inhibits toad bladder adenylate cyclase in vitro [19, 48]. Ami-centrations they increase lysosomal lability [57]. noglycosides also have selective effects on the organic acid These effects of aminoglycosides on subcellular membranes secretory transport process. Gentamicin treatment increasesalso appear to occur in vivo [60]. Data are available on the both renal clearance of para-aminohippurate (PAH) and renalin-vivo effects of gentamicin in a number of subcellular sys- slice uptake of PAH: the latter effect was observed after a singletems. To gain insight into pathogenetically important processes, dose [49], Both PAH transport and organic base transportit is important that attention be given to phenomena that occur decline at later stages of nephrotoxicity; this fall probably is aprior to massive cell necrosis and disruption. Mitochondria nonspecific effect of progressive renal cell injury. isolated from rats treated with gentamicin are functionally Considerable evidence indicates that aminoglycosides areabnormal prior to severe tubular cell injury as compared with presented to the tubular cell in high concentrations, leading tocontrols [58]. However, the pattern of mitochondrial functional phospholipid binding and subsequent alterations in cell phos-effects seen after in-vivo gentamicin administration differs from pholipid composition. Additionally, these antibiotics provokethat seen with in-vitro exposure. Thus it is difficult to say with early and specific alterations in plasma membrane permeabilitycertainty whether the in-vivo effects are due to direct interac- and transport, many of which are mediated by membranetions of gentamicin with mitochondria or are secondary to phospholipid binding and alterations. While the resulting changesanother early intracellular process evoked by gentamicin. In in the intracellular environment can have indirect detrimentaleither case, function of this vital intracellular organelle is effects on intracellular processes and organellar function, addi-compromised early enough during the course of aminoglycoside tional data have demonstrated aminoglycosides' potential forexposure to potentially contribute to the pathogenesis of renal altering organellar membranes directly. cell injury. Although nephrotoxin-related alterations in renal Alterations in organellar structure and function. Intact mito—cell bioenergetics appear to be a potential, common biochemi- chondrial oxidative phosphorylation is essential to the mainte-cal pathway in the pathogenesis of nephrotoxic acute renal nance of normal renal tubular cell function. 'I'he capacity of thefailure [59]. current evidence does not assign mitochondrial 906 Nepl?rologyForum: A minoglycoside m'phrotoxicily dysfunction a unique role in the early pathogenesis of amino- Table 3. Modification of experimental aminoglycoside nephrotoxicity glycoside nephrotoxicity. Important questions remain. Potentiating factors Lysosomal alterations occur soon after exposure of renal I. Hypercalcemia tubule cells to aminoglycosides. Lysosomes within proximal 2. Potassium depletion tubule cells increase in size and number, and lysosomal enzym- 3. Magnesium depletion 4. Metabolic acidosis uria occurs [11], accompanied by increases in renal cortical 5. Endotoxemia content of at least one lysosomal enzyme, NAG [42]. Lysoso- 6. Pyelonephritis mal enzyme activity, including sphingomyelinase and phospho- lipase A from renal cortex, is inhibited early in the course ofProtective factors experimental aminoglycoside nephrotoxicity [60, 611. Both drug- 1. Uncontrolled streptozotocin-induced diabetes mellitus 2. Parathyroidectomy induced lysosomal alterations (with resulting impairment in 3. Thyroxine administration membrane degradation and recycling) and toxic lipid sequestra- 4. Phosphate depletion tion (with subsequent release of lysosomal enzyme contents) 5.Oralcalcium supplementation might contribute to the pathogenesis of renal cell injury. Al- 6. Polyamino acid administration though it currently is impossible to ascribe a predominant pathogenetic role to any one aminoglycoside-induced mem- brane or organellar impairment in the causation of renal failure,dude alterations in phosphorus and divalent cation intake or these differing adverse organellar effects probably are closelyrenal handling, potassium or magnesium deficiency, hypercal- interrelated in the pathogenesis of aminoglycoside nephrotox-cemia, metabolic acidosis, the presence of uncontrolled diabe- icity. tes mellitus, hypothyroidism, parathyroid hormone deficiency, Clinical determinants. Clearly, a structural component of thepyelonephritis, and endotoxemia. These observations regarding aminoglycosides imparts toxic potential to these compounds.modulation of nephrotoxicity are important, because they might Recent evidence supports the thesis that the cationicity derivedprovide insights into the pathogenesis of renal cell injury and from ionizable amino groups is a major molecular determinantcould have important implications for the prevention of clinical of aminoglycoside membrane action. The toxic potential of anaminoglycoside-induced acute renal failure. aminoglycoside antibiotic is attributable in large part to its Uncontrolled streptozotocin-induced diabetes mellitus in the ability to interact with membranes and to alter membranerat protects against gentamicin-induced acute renal failure [64]. structure and function. It is therefore not surprising that a roughThis effect can be observed as early as 5 days after streptozo- correlation between cationicity and whole-organ toxicity existstocin treatment and is uniformly associated with renal cortical [31]. A single group of aminoglycosides with identical cationicaminoglycoside levels lower than in controls. Although the changes, however, can have appreciable differences in toxicityprecise mechanism of protection remains obscure, elevated [62]. In fact, more refined analyses of aminoglycoside-mem-GFR and decreased renal cortical drug binding sites, with brane interactions have suggested significant effects on bindingresulting diminished cortical drug accumulation, seem to be by amino-group orientation and by hydrophobic interactionsimportant [65]. A partial role for solute diuresis and urine flow [40]. rate also seems likely. Nonetheless, there is little doubt that the cationic amino Treatment with thyroxine of normal and parathyroidectom- groups are the structural determinants of toxicity. Polydis-ized rats prevents aminoglycoside-induced alterations in renal persed neutral dextran consists of linearly branched chains offunction [66]. This protective effect of thyroxine could not be glucose residues connected by glycosidic linkages of varyingexplained by alterations in GFR or calcium excretion. The molecular size. This compound has no nephrotoxic activity oractivity of renal cortical sodium-potassium ATPase, however, deleterious effects on mitochondrial membrane function. Thecorrelated with total plasma thyroxine in both control and addition of cationic diethylaminoethyl (DEAE) substitutiongentamicin-treated animals. This protective effect of thyroxine groups to the glucose residues of neutral dextran yields alacks specificity and has been observed both in ischemic renal preparation of DEAE dextran remarkably reminiscent of thefailure and in other models of nephrotoxicity [67]. Parathyroid- structure of the aminoglycoside antibiotics. This compound hasectomy in Fisher 344 rats fed a standard diet conferred partial substantial nephrotoxicity, producing severe necrosis in S1 andprotection from aminoglycoside nephrotoxicity compared with S2 segments of the proximal tubule within 48 hours aftersham-operated controls [68]. Another study, however, did not administration [63]. In fact, a variety of low-molecular-weightreport a protective effect from parathyroidectomy on amino- organic compounds that possess significant cationic chargeglycoside nephrotoxicity in rats [66]. density derived from ionizable amino groups can induce irre- Electrolyte alterations also modulate aminoglycoside neph- versible cell injury [31]. rotoxicity. Aluminum hydroxide-induced phosphorus depletion ameliorates gentamicin nephrotoxicity in the rat [69], and acute Modification of experimental a,ninoglvcoside nephrotoxicitv phosphate infusion causes both functional and morphologic Numerous conditions modify both the time course and sever-worsening of ischemic acute renal failure [70]. Production of ity of experimental aminoglycoside nephrotoxicity. As I notedmild hypercalcemia with I ,25(OH)D1 markedly potentiates the earlier, these factors include the animal's age, gender, speciesseverity of aminoglycoside nephrotoxicity [71]. Experimental and strain, and the aminoglycoside dose and dosing intervalmagnesium deficiency promotes aminoglycoside nephrotoxicity [31]. Within a well-defined experimental setting, various sys-in Sprague-Dawley rats [27]. Studies in rodents and dogs clearly temic and renal alterations also have been demonstrated todemonstrate that hypokalemiaexacerbates aminoglycoside neph- modulate aminoglycoside nephrotoxicity (Table 3). These in-rotoxicity [261. Metabolic acidosis significantly worsens genta- Nepl,ro!ogv Forum:4minogivcosule nephrotoxicilv 907 micin-induced acute renal failure in the rat [72].Renalinfectionfirst noted to be elevated. Furthermore, BUN and serum or systemic infection complicated by endotoxemia potentiatescreatinine should have been monitored daily. aminoglycoside nephrotoxicity [73]. This potentiation is due, at The serum creatinine followed a typical triphasic pattern: a least in part, to more rapid accumulation of aminoglycosides inrapid increase, a gradual plateau, and a slow to moderate renal parenchyma. improvement to normal. As I noted before, a continued in- The mechanisms by which these renal or systemic factorscrease in serum creatinine after discontinuation of the antibiotic alter aminoglycoside nephrotoxicity are not well understood.is common, because renal cell injury continues when toxic Most are likely nonspecific and probably affect other forms ofconcentrations of the drug persist in the renal parenchyma. nephrotoxic injury as well. Those factors that specificallyFortunately, the nonoliguric state helped maintain normokale- influence aminoglycoside nephrotoxicity appear to act by in-mia and volume homeostasis without uremic symptoms, and creasing the rate of antibiotic accumulation within the renaldialysis was not required. cortex. Approaches that may ameliorate aminoglycoside nephrotox- Questions and answers icity are promising. Among these attempts, the most exciting consists of identifying compounds that are competitive inhibi- DR. JEROME P. KASSIRER (Associate Physician-in-Chief, De- tors of aminoglycoside membrane binding. Small organic poly-partinent of Medicine, Nest' England Medical Center, Boston, cations were initially tested, but these had appreciable nephro-Massachusetts): How important is the cationic nature of the toxicity of their own [74]. Because calcium is an effectiveaminoglycosides with respect to their effect on bacteria? If you competitive inhibitor of aminoglycoside binding to biologicinterfere with one effect, you might, in principle at least, membranes [75], oral calcium supplementation was used tointerfere with the other. increase delivery of the ion to the kidney. Oral calcium loading DR. HUMES: The cationicity of the aminoglycosides probably markedly protected against gentamicin nephrotoxicity as mea-dictates their membrane interactions with bacterial membranes sured by both renal functional and biochemical indices of renalas well as their transport into bacteria. Thus, the cationicity of cell injury [75, 76]. However, the protective effect of calciumthese antibiotics also is an important determinant for their was not accompanied by a change in either peak renal corticalantimicrobial activity. Of note, the report using large-molec- gentamicin levels or the time taken to reach them. Thus it isular-weight polyamino-substituted derivatives demonstrated a possible that the effects of calcium are mediated either by theprotective effect on aminoglycoside nephrotoxicity but did not inhibition of gentamicin's action within renal cells rather thanalter the antimicrobial activity of aminoglycosides in vitro [77]. on their surfaces, or another metabolic alteration might occurThis dissociation might relate to the large size of the polyamino- during oral calcium loading. Because of the large amounts ofsubstituted derivative. calcium required to achieve this protective effect, this maneu- DR. MICHAEL BARZA (Division of Geographic Medicine and ver probably is not clinically applicable. Infectious Diseases, New England Medical Center): Gram- This type of therapeutic approach has been extended to thenegative bacteria have an outer membrane containing pores use of various polyamino acids, including polylysine, polyaspa-through which most substances that are not highly lipid-soluble ragine, and polyaspartic acid [77]. These compounds inhibitmust pass in order to enter the cytoplasm. Large molecules aminoglycoside binding to renal brush-border membranes. Fur-such as the ones you are describing would be unlikely to pass thermore, these agents, when coadministered with aminoglyco-through the outer membrane. sides to rats, completely protect against nephrotoxicity but DR. HUMES:Yourexplanation appears very reasonable. have no effect on the antibacterial activity of these antibiotics in DR. KASSIRER: Is there any effect of these large cationic vitro [77]. These findings demonstrate that we can identifymolecules on other cells or organs? I ask this question because interventions for ameliorating nephrotoxicity as soon as weof the extent of anionic sites in various glycoproteins and other understand the cellular pathophysiology of the toxic process. structural proteins throughout the body. Is there evidence of Now let us return to the patient presented today. The initialtoxicity in these tissues? decrease in the patient's renal function on admission was the DR. HUMES: I am not aware of potential toxicity of these result of the volume depletion as evidenced by the normalagents on other tissues. urinalysis, low UNa, and high urine osmolality. The subsequent DR. BARZA: Aminoglycosides only appear to be concentrated decrease in renal function 6 days after operation was caused byin two sites inthe body: the proximal renal tubule and the inner acute tubular necrosis induced by gentamicin, as reflected byear. Those are the only two sites that seem to have these the abnormal urinalysis, high UNa and FENa, and low urinepolyphosphoinositol substances that act as receptors for ami- osmolality. Acute tubular necrosis occurred despite the use of anoglycosides. It may be that, like aminoglycosides, other poly- "normal" dose of gentamicin, probably because of the patient'scationic substances fail to be taken up by most tissues of the age and the coexisting volume depletion. These factors shouldbody. have prompted different therapeutic choices and clinical follow- Some years ago. we studied proximal tubular preparations up: euvolemia should have been aggressively maintained through-and showed that the uptake of radiolabeled gentamicin ap- out the patient's course. Because the patient was elderly,peared to he inhibited by 2,4-dinitrophenol and cold tempera- tobramycin rather than gentamicin should have been used.tures. This finding suggested an active transport system. Did Serum gentamicin levels should have been measured on the firstyou allude to that? and second days after therapy was begun so that dose adjust- DR. HUMES: No, I did not allude to that phenomenon, ments could have been made as appropriate rather than 6 daysAminoglycosides are taken up by proximal tubule cells via an into the patient's course, when the serum creatinine level wasendocytotic process, which is an energy-dependent process. 908 NephrologvForum: A,ninoglvcoside ,,ephrotoxkity

Thus this transport process is influenced by maneuvers that DR. STROM: And a clinical question: Wouldn't it be unwise to interrupt energy transduction processes. measure aminoglycoside levels on the first day, before steady- Da. RONALDD.PERRONE (Division of Nephrology, Newstate levels have been achieved? England Medical Center): Youindicated that pretreatment of DR. HUMES: It is best to measure aminoglycoside levels after rats with calcium ameliorated aminoglycoside-induced acutea steady state has been achieved. The point I would like to renal failure. Have you looked further at magnesium or othermake is that this measurement should be done after 24 or 48 physiologic cations in similar circumstances? hours following initiation of aminoglycoside therapy rather than DR. HUMES: Yes, we initially studied magnesium supplemen-after 5 days. tation. This maneuver, however, resulted in severe diarrhea in DR. VINCENT J. CANZANELLO (DivisionofNephrology, New the animals, as well as subsequent volume depletion andEngland Medical Center): Manyof our patients have preexist- electrolyte imbalances. These disturbances would alter theing renal disease. When we give aminoglycosides to these development of aminoglycoside-induced renal dysfunction andpatients, how should we administer the drug? obscure the final results. We therefore undertook the studies DR. HUMES: Preexisting renal insufficiency probably is a risk using calcium supplementation. factor for any nephrotoxin, because of an increase in delivery of DR. JOHN T. HARRINGTON: (Chief of Medicine, Newton-We!-the potential nephrotoxin per functioning nephron. For ami- lesley Hospital, Newton, Massachusetts):I would like to follownoglycosides, animal studies have demonstrated that for any up on Dr. Perrone's question. Does calcium protect againstgiven total dose, adjustments downward in dose are more gentamicin nephrotoxicity at a luminal site or at an intracellularnephrotoxic than are adjustments upward in dosing intervals site? [79]. Other studies, however, suggest that better antimicrobial DR. HUMES: The animals that were given oral calciumefficacy is accomplished with dose adjustments rather than with supplements developed hypercalciuria but not hypercalcemia. Idosing interval alterations. The final decision therefore will therefore would speculate that the protection develops from adepend on the individual clinical situation and the risk/benefit of luminal membrane event. Oral calcium supplementation alsoeither approach. produces other alterations such as hypophosphatemia, which DR. MADIAS: How good is the correlation between antibiotic might ameliorate aminoglycoside toxicity by other mechanismsblood levels and nephrotoxicity in animal and clinical studies? separate from competitive inhibition of membrane binding. DR. HUMES: There is only an approximate correlation, The DR. HARRINGTON:Giventhat calcium's protection againstproblem is that a variety of factors can influence the patient's aminoglycoside nephrotoxicity seems to be on the luminal side,risk of developing nephrotoxicity. If the patient is septic and what do you think mediates the effect of volume depletion involume depleted, aminoglycoside toxicity still can develop, no promoting gentamicin toxicity? Is it a sodium-dependent phe-matter how well the physician controls the serum levels. nomenon or might it be calcium dependent as well? Da. MADIAs: Is there any clinical utility for monitoring the DR. HUMES: The routine answer has always been that volumeenzymuria associated with the onset of aminoglycoside neph- depletion leads to increased concentration of aminoglycosidesrotoxicity? in the renal cortex, but aminoglycoside transport by renal DR. HUMES: No. Enzymuria is nonspecific and too sensitive epithelia is not a sodium-dependent process. Instead, I thinka marker to be diagnostically or prognostically beneficial. this potentiation might relate to the influence of volume deple- DR. KASSIRER: Some evidence was presented several years tion on renal blood flow. It's been shown that brief renalago that aminoglycosides produce a reduction in K1. Given the ischemia potentiates nephrotoxic injury [78]. huge amount of polyanions in the glomerular capillary wall and DR. NIcOLAOs E. MADIAS (Chief, Division of Nephrology,particularly in the basement membrane, does a reduction in Kf NewEngland Medical Center):In some studies, metaboliccontribute to the reduction in GFR in gentamicin nephrotoxic- acidosis has been shown to potentiate aminoglycoside nephro-ity? toxicity. I noted that in your calcium protective studies you DR. HUMES: As I mentioned, arninoglycosides produce a used calcium carbonate; is it possible that part of the protectionmodest decline in K1. This decline is not the predominant cause actually derived from the administered alkali load? of the decline in renal excretory function. The predominant DR. HUMES: These studies incorporated a group of animalspathophysiologic process is the toxic injury to the proximal that were supplemented with equivalent amounts of sodium tubule cell produced by the aminoglycoside. bicarbonate to calcium carbonate. Sodium bicarbonate supple- D. PERRONE: Although there is no accepted role for diuret- mentation was not protective. ics in the prevention of acute renal failure, can you envision any Da. JAMES STROM (Chief, Division of Nephrology, St. Eli- point in this process at which a diuretic might have an impact in zaheth'sHospital, Boston):What is the link between the aminoglycosides' polycationic membrane interaction and thea mechanistic way to prevent either uptake or binding of an displacement of magnesium? aminoglycoside? DR. HUMES: Magnesium and calcium are bound to all cellular D. HUMES: No; in fact, aggressive diuresis can aggravate membranes and are important in the regulation of variousthe potential nephrotoxicity because of the resulting volume transport processes. depletion and potassium depletion. DR. STROM: And is this caused specifically by the polycati- DR. KASSIRER: How often does aminoglycoside nephrotox- onic interaction? icity produce permanent renal failure? DR. HUMES: The displacement of these divalent cations by DR. HUMES: Like most toxic injury to the kidney, it is a the aminoglycoside antibiotics produces membrane dysfunc-reversible process because of the regenerative ability of the tion, at least in part, by this effect. renal epithelium to reline the tubule. Interstitial fibrosis can Nep/zrologv Forum: Aminoglycoside nephrotoxicitv 909 develop after nephrotoxicity, however, hut not sufficiently to Transport of gentamicin in rat proximal tubule. LabInvest 48:212, cause renal insufficiency. 1983 9. FABREJ,RUDHARDTM,BLANCHARDP.REGAMEYL:Persistence D, KASSIRER: Given the mechanisms you proposed for the ofsisomicin and gentamicin in renal cortex and medulla compared initiation of acute renal failure by aminoglycosides, I do not with other organs and serum of rats. Kidney mt 10:444-449. 1976 understand the mechanism of the recovery. If obstruction and 10. KOSEKJC,MAZZE RI, CousiNs MJ: Nephrotoxicity of gentamicin. backleak are the major factors, the repair of the tubular Labinvest 30:48—57,1974 II. WELLWOODJM,LOVELLD,THOMPSONAE,TIGHEJR:Renal epithelium should result in a reduced backleak and a greater damage caused by gentamicin: A study of the effects on renal quantity of fluid remaining in the tubule in the presence of morphology and urinary enzyme excretion. J Pat/iol 118:171—182, continued obstruction. If anything, the retrograde pressure on 1976 the glomerulus would increase and GFR would fall further. Do 12. GIULIANORA,PAULUSGJ,VERPOOTENGA,PATTYNUM, we know anything about how obstruction is relieved and about POLLETDE,NOUWENEJ,LAURENTG,CARLIERMB,MALDOGUE P,TULKENSPM,DEBROEME:Recovery of cortical phospholipi- the relation between relief of obstruction and reduction in dosis and necrosis alter acute gentamicin loading in rats. Kidney mt tubular backleak? 26:838, 1984 DR. HUMES: No one has carefully looked at the recovery 13. HUMESHD,WEINBERGJM,KNAUSSTC:Clinical and pathophys- phase of acute renal failure either from a micropuncture or a iologic aspects of aminoglycoside nephrotoxicity. Am J KidneyDis biochemical approach. This recovery phenomenon is an area 2:5—29,1982 14. HOUGHTONDC,HARTNETTM,CAMPBEL-BOSWELLM,PORTER0, waiting to be explored. BENNETT W:A light and electron microscopic analysis ofgenta- DR. MADIAS: I believe there are some good data suggesting micin nephrotoxicity in rats. Am J Pathol 82:589—612, 1976 that the recovery process of aminoglycoside nephrotoxicity can 15. CRONINRE,BULGERRE,SOUTHERNP,HENRICHWC:Natural express itself in the presence of continued exposure to the history of aminoglycoside nephrotoxicity in the dog. J Lab Clin antibiotic. Is there an understanding of this phenomenon? Med95:463.-474,1980 16. HOUGHTONDC,PLAMPCE,DEFEHRJM,BENNETTWM,PORTER DR. HUMES: Studies have demonstrated that renal epithelia G,GILBERTD:Gentamicin andtobramycinnephrotoxicity. AmJ during the recovery phase of nephrotoxic injury are resistant to Pathol 93:137—152, 1978 a subsequent toxic event [80]. Aminoglycosides are a clear17. GILBERTDN,HOUGHTONDC,BENNETTWM.PLAMPCE,ROGER example of this phenomenon. The explanation for this process K,PORTERG:Reversibility of gentamicin nephrotoxicity in rats: Recovery during continuous drug administration. ProcSoc Exp currently is not understood, but it might relate to the resistance Biol Med 160:99—103,1979 of immature, regenerating renal epithelial cells to toxic injury. 18. MONDORFAW,BRIERJ,HANDUSJ,SCHERBERICHJE. DR. HENRY YAGER (Associate Chief of Medicine, Newton- MACKENRODT0, SHAH P.STILLEW,SCHIEPPEW: Effect of Wellesley Hospital): In clinical practice, we often use aminogly- aminoglycosideson proximal tubular membranes of the human cosides in combination with other antibiotics. The literature has kidney. EuriC/in Pharmacol13:133—142,1978 19. HUMESHD,WEINBERGJM:The effect of gentamicin on antidiure- some conflicting data about the synergistic effect of cephalo- tic hormone-stimulated osmotic water flow in the toad urinary sporins on aminoglycoside toxicity. Could you comment on bladder. J Lab ClinMed 101:472—478.1983 that? 20. SCHENTAGii,Juso WJ: Renal clearance and tissue accumulation DR. HUMES: Two cephalosporins seem to potentiate amino- of gentamicin. Cliii PharmacolTher 22:364—370,1977 21. DAHLGRENJG,ANDERSONET,HEWLITTWL:Gentamicin blood glycoside nephrotoxicity: cephaloridine and cephalothin. The levels: A guide to nephrotoxicity. AntimicrobAgentsChemother newer cephalosporins fortunately do not appear to have this 8:58—62, 1975 ability. 22. LANEAZ,WRIGHTGE,BLAIRDC:Ototoxicity and nephrotoxicity of amikacin. AmJMed62:9l1—918,1977 23. WHITINGPH,SIMPsoNJO,DAVIDSONRJL,THOMSONAW:The Reprintrequests to Dr. H.D. Humes. Medical Service, Veterans toxic effects of combined administration of cyclosporine A and AdministrationMedical Center, 22/5 Fuller Road, Ann Arbor, Michi- gentamicin. BrJ Exp Pathol63:554-561, 1982 gan 48105, USA 24. MOORERD.SMITHCR,LIPSKYTJ,MELLITSED,LIETMANPS: Risk factors for nephrotoxicity in patients treated with aminogly- References cosides. AnninternMed 100:352,1984 25. BENNETTWM,HARTNETFMN,GILBERTD,HOUGHTONDC, 1. FINLANDM:Changing ecology of bacterial infection as related toPORTERGA:Effectof sodiumintake on gentamicin nephrotoxity in antibacterial therapy. J Infect Dis 122:419—431, 1970 therat.ProcSoc Exp Biol Med 151:736—738,1976 2. Hou SH, BUSHINSKYDA.WIsI-I JB. COHENJJ,HARRINGTONJT: 26. BRINKERKR,BULGERRF,DOBYANDC,STACEYTR,SOUTHERN Hospital-acquired renal insufficiency: A prospective study. Am J PM,HENRICHWL,CRONINRE:Effect ofpotassiumdepletion on Med74:243—248,1983 gentamicin nephrotoxicity. JLab C/in Med 98:292.1981 3. KAHLMETER0. DAIILAGER J:Aminoglycoside toxicity—A review 27. RANKINLi,KROus H, FRYERAW.WHANGR:Enhancement of of clinical studies published between 1975 and 1982. JAnti,nicrob gentamicin nephrotoxicity by magnesium depletion in the rat. Chemother 13:9A,1984 MinerElectrolyte Metab 10:199—203.1984 4, PASTORIZA-MUNOZE.BOWMANRL,KALOYANIDESGJ:Renal 28. MooRERD, SMITH CR,LIETMANPS:Association of aminoglyco- tubular transport of gentamicin in the rat. Kidneymt16:440—450. side plasma levels with therapeutic outcome in gram-negative 1979 pneumonia. AmJMed 77:657. 1984 5. SENEKJIANHO,KNIGHTTF,WEINMANEJ:Micropuncture study 29. LIETMANPS.SMITHCR:Aminoglycoside nephrotoxicityin hu- of the renal handling of gentamicin in the rat. Kidney Jut 19:416— mans. RevInfect Dis 5:S284.1983 423. 1981 30. NEUGARTENJ,AYNEDJIANHS, BANK N: Role of tubularobstruc- 6.PASTORIZA-MUNOZ F. TIMMERMAN D.KALOYANIDESGi:Renal tion in acute renal failure due to gentamicin. KidneyJut24:33t). transport of netilmicin in the rat, JPhurinacolEvp Ther 228:65, 1983 1984 31. HUMESlID.WEINBERGJM:Alterations in renal tubular cell 7. SILVERBLATTFJ,KUEHNC:Autoradiology of gentamicin uptake metabolismin acute renal failure.MinerElecirolve Metab 9:29(-. by the rat proximal tubule cell. Kid,,ev mt 15:335—345. 1979 305. 1983 8. WEDEENRP.BATUMANV.CHEFKSC, MARQUET E.SOIIELIL 32. DONOHOEiF. VENKATACHALAM MA.BERNARDDB. LEVINSEY 910 Nephrology Forum: Aminoglycosidenephrotoxici!y

NG: Tubular leakage and obstruction after renal ischeinia: Struc- pholipases AandC by aminoglycoside antibiotics: possible mech- tural-functional correlations. KidneyIn!13:208—222, 1978 anism of aminoglycoside toxicity. ProcNail Acad Sci USA 33. BRENNER BM, ISCHIKAWA I, DEEN WM: Glomerular filtration, in 79:1663,1982 The Kidney(vol 1,2nd ed), edited by Brenner BM, Rector FC Jr, 56.LAURENT G, CARLIER MB, ROLLMAN B, VAN HOFF F, TULKENS Philadephia, Saunders, 1981 P: Mechanism of aminoglycoside-induced lysosomal phospholipi- 34. BAYLIS C, RENNKE HR, BRENNER BM: Mechanisms of the defect dosis: in vitro and in vivo studies with gentamicin and amikacin. in glomerular ultrafiltration associated with gentamicin administra- Biochem Pharmacol 3 1:3861, 1982 tion. KidneyIn! 12:344—353,1977 57. POWELL JH,REIDENBERGMM:in vitroreponse of rat and human 35.SCHORN, ICHIKAWA I, RENNKE HG,TROYJ, BRENNER BM: kidney lysosomes to aminoglycosides. Biochem Pharmacol 31: Pathophysiology of altered glomerular function in aminoglycoside- 3447,1982 treated rats. Kidney mt 19:288—296, 198! 58. SIMMONSCF JR, BOGUSKY RT, HUMES HD: Inhibitory effects of 36.LUFTFC,ARONOFFGR,EVANAP,CONNORSBA,WEINBERGER gentamicin on renal mitochondrial oxidative phosphorylation. J MH, KLEIT SA: The renin-angiotensin system in aminoglycoside- Pharmacol Exp Ther 214:709, 1980 induced acute renal failure. J Pharmaco! Exp Ther 20:433, 1982 59.HUMESHD, WEINBERG JM: Cellular energetics in acute renal 37. SASTRASINH M, KNAUSS TC, WEINBERG JM, HUMES HD: Identi- failure,Acute inRenal Failure, edited byBRENNERBM, LAZARUS fication of the aminoglycoside binding site in rat renal brush border JM,Philadelphia,Saunders, 1983, p 47 membranes. J Pliarmaco! Exp Ther 222:350—359, 1982 60.AUBETT-TULKENSG,VAN HOOF F, TULKENS F: Gentamicin- 38. SCHACHT J: isolation of the aminoglycoside receptor from guinea induced lysosomal phospholipidosis in cultured rat fibroblasts: pig inner ear tissues and kidney. Arch Otorhinolaryngol 224:129— quantitative ultrastructural and biochemical study. Lab Invest 134, 1979 40:481, 1979 39. LODHI S, WEINER ND, SCHACHT J: Interactions of neomycin and 61.MORINJP,VIOTTE 0,VANDEWALLE A,VANHOOF F,TULKENSF, calcium in synaptosomal membranes and polyphosphoinositide mono- FILLASTRE JP:Gentamicin-inducednephrotoxicity: a cell biology layers. Biochim Biophys Ada 426:781—785, 1976 approach. Kidney mt 18:583, 1980 40. BRASSEUR R, LAURENT G, RAYSACHAERT JM, TULKENS P: Inter- 62. SOBERON L, BOWMAN RL, PASTORIZA-MUNOZ E, KALOYANIDES actions of aminoglycoside antibiotics and negatively charged lipid GJ: Comparative nephrotoxicities of gentamicin, netilmicin and layers: Biochemical and conformational studies. BiochemPharma- tobramycin in the rat. I Pharmacol Exp Ther 210:334-343, 1979 col 33:629, 1984 63. SIMMONS CF JR. RENNKE H, HUMES HD: Acute renal failure 4!. SCJ-IWERTZ DW, KREISBERG ii, VENKATACHALAM MA: Effects of indiced by diethylaminoethyl dextran: Importance of cationic charge. aminoglycosides on proximal tubule brush border membrane phos- Kidneymt19:424—430, 1981 phatidylinositol-specific phospholipase C. J Pharmacol Exp Ther 64. TEIXEIRA RB, KELLEY J, ALPERT H, PARDO V. VAAMONDE CA: 231:48, 1984 Complete protection from gentamicin-induced acute renal failure in 42. KNAUSS TC, WEINBERG JM, HUMES HD: Alterations in renal thediabetesmellitus rat. Kidney mt 21:600,1982 cortical phospholipid content induced by gentamicin: Time course, 65.VAAMONDE CA, BIER RT, GOUVEA W, ALPERT H, KELLEY J, specificity and subcellular localization. Am I Physiol 244:F535— PARDOV: Effect of duration of diabetes on the protection observed F546, 1983 in the diabetic rat against gentamicin-induced renal failure. Miner 43. KACEW S: Gentamicin or chlorphentermine induction of phospho- Electrolyte Metab 10:209,1984 lipidosis in the developing organism: Role of tissue and species in 66.CRONIN RE. NEWMAN JA: Protective effect of thyroxine but not manifestation of toxicity. J Pharmaco! Exp Ther 232:239, 1985 parathyroidectomyon gentamicin nephrotoxicity. AmI Ph.'siol 44. DAFIL JL, HOKIN LE: The sodium-potassium adenosine triphos- 248:F332.1985 phatase. Annu Rei' Biochem 44:327—356, 1974 67. GAUDIOKM,KASHGARIANM,SIEGELNJ:Thyroxin accelerates 45. LIPSKY JJ, LIETMAN PS: Neomycin inhibition of adenosine triphos- recovery from ischemic acute renal failure. GunRes 32:533A, 1984 phatase: Evidence for a neomycin-phospholipid interaction. Anti- 68.PULLIAMJP.HOUGHTONDC.BENNETTWM:Modification of microb Agents Chemother 18:532—535, 1980 gentamicinnephrotoxicity by selective parathyroidectomy. Am J 46. CR0NIN R,Nix K, FERGUSONE,SOUTHERNPM, HENRICH WL: Phvsiol 249:F832—F835,1985 Renalcortex ion composition and Na-K-ATPaseactivityin early 69. EKNOYANGL, GENTRY R, BULGER R, DOBYAN D: Attenuation of gentamicin nephrotoxicity. Am I Physiol 242:F477-F483, 1982 gentamicin-induced renal failure by phosphate depletion. Kidney 47. RUBALCAVA B, RODBELL M: The role of acidic phospholipids in mt 25:229. 1984 glucagon action on rat liver adenylate cyclase. J Biol C/zem 70. ZAGER RA,JOHANNESGA,SHARMAHM:Hyperphosphatemia:A 248:3831—3837, 1973 factorthat promotes severe experimental acute renal failure. I Lab 48. WILLIAMS PD,HOLOHAMPD,RossCR: Gentamicin nephrotox- Clin Med 100:230, 1982 icity. II. Plasma membrane changes. Toxicol App! Pharrnacol 71. COHEN RA, JOHNSON K,HUMESHD: Vitamin D induced hyper- 61:243—251, 1981 calcemia potentiates gentamicin nephrotoxicity. Miner Electrolyte 49. BENNETT WM, PLAMP CE, PARKER RA, GILBERT DN, HOUGHTON Metab, in press DC, PORTER GA: Renal transport of organic acids and bases in 72. Hsu CH, KURTZ TW, EASTERLING RE, WELLER JM: Potentiation aminoglycoside nephrotoxicity. Antimicrob Agents Chemother 16: of gentamicin nephrotoxicity by metabolic acidosis. Proc Soc Exp 231, 1979 Biol Med 146:894, 1974 50. WEINBERG JM,HUMES HD: Mechanismsof gentamicin-induced 73. ZAGER RA, PRIOR RB: Gentamicin and gram-negative bacteremia: dysfunction of renal cortical mitochondria. I.Effectsofmitochon- Asynergismfor thedevelopment of experimental nephrotoxic drial respiration. Arc/i Biochem Biophys 205:222—231, 1980 acute renal failure. J Clin Invest 78:196—204, 1986 51. WEINBERG JM, HARDING PG. HUMES HD: Mechanisms of genta- 74. JOSEPOVITZ JC, PASTORIZA-MUNOZ E. TIMMERMAN D, SCOTT M, micin-induced dysfunction of renal cortical mitochondria. 11. Ef- FELDMAN S, KALOYANIDES Gi: Inhibition of gentamicin uptake in fects on mitochondrial.monovalent cation transport. Arch Biochem rat renal cortex in vivo by aminoglycosides and organic polycat- Biop/,vs205:232—239,1980 ions. J Pharmacol Exp Ther 223:3 14. 1982 52.WEINBERG JM, SIMMONS CF JR. HUMES HD: Alterations of 75. HUMES HD. SASTRASINH M. WEINBERG JM: Calcium is a compet- mitochondrjal respiration induced by aminoglycoside antibiotics. itive inhibitor of gentamicin-renal membrane binding interactions Ri's Comm Chem Pathol Pharm 27:521—532, l980 and dietary calcium supplementation protects against gentamicin 53. SASTRASINH M, WEINBERG JM, HUMES HD: Effect of gentamicin nephrotoxicity. I Cli,, Invest 73:134, 1984 on calcium uptake by renal mitochondria. LifeSci30:2309—2315, 76. BENNETT WM,ELLIOTTCW, HOUGHTON DC, GILBERT DN. 1982 DEFEHR J.MCCARRONDA: Reduction of experimental gentamicin 54. WALKER PD,SHAHSV:Gentamicinenhanced production of hy- nephrotoxicity in rats by dietarycalcium loading. Antimicrob drogen peroxide by renal cortical mitochondria. Am I Phvsiol, in Agents Che,not/,er 22:508—512, 1982 press 77. WILLIAMS PD, HOTTENDORF GH, BENNETT DB: Inhibition of 55. HOSTETLER KY, HALL LB: Inhibition of kidney lysosomal phos- renal membrane binding and nephrotoxicity of aminoglycosides. J Nephrology Foru,n: Anzinoglvcoside iu'phrotovicitv 911

Pliarinacol Exp Ther 237:919—925, 1986 of peak serum levels from renal failure. J Infect Dis 140:576-581, 78. ZAGLR RA: Gentamicin increases renal susceptibility to an acute 1979 ischemic insult. J LabC/inMcd101:670—678,1987 80. OKEN DE, MENDE CW, TARABA I, FLAMENBAUM W: Resistance to 79. BENNETT WM. PLAMP CE, GILBERT DM: The influence of dosage acute renal failure afforded by prior renal failure: Examination of regimen on experimental gentamicin nephrotoxicity: Dissociation the role of renal renin content. Nep/,ron15:131-142, I975