BRIEF COMMUNICATIONS

Anesthetic Effects on the Glycerol Model of Rhabdomyolysis- Induced Acute Renal Failure in Rats

KAREN M. LOCHHEAD,* EVAN D. KHARASCH,t and RICHARD A. ZAGER Departments of *Medicine and Anesthesiology, University of Washington, and Fred Hutchinson Cancer Research Center, Seattle, Washington.

Abstract. Isoflurane, the most widely used inhalational anes- acute intrarenal heme loading (cast formation), and BP during thetic, releases inorganic fluoride during its metabolism by the the initiation phase of renal injury (0 to 4 h after glycerol cytochrome P450 system. Recent experimental data indicate injection) were also assessed. Glycerol induced severe ARF that when cultured proximal tubular cells are exposed to inor- under pentobarbital anesthesia (Cr, 2.8 ± 0.3 mg/dl; severe ganic fluoride, they become relatively resistant to myoglobin- tubular necrosis). Somewhat worse azotemia, but comparable and ATP depletion- mediated attack. The present study was tubular necrosis, resulted with desfiurane use. Conversely, undertaken to assess whether isoflurane anesthesia might con- glycerol plus isoflurane anesthesia induced only mild renal fer in vivo cytoprotection, possibly by causing renal tubular damage (Cr, 0.9 ± 0.1, minimal tubular necrosis; P < 0.01). inorganic fluoride exposure, thereby mitigating a combined This reduction apparently was not due to differences in degrees myoglobin/ATP depletion model of acute renal failure (glyc- of muscle necrosis, hemolysis, acute renal heme loading, or BP erol-induced ARF). Rats were injected with hypertonic glyc- during the initiation phase of ARF, suggesting that a direct erol (50%; 9 ml/kg, intramuscularly) while undergoing 4 h of renal mechanism was operative. These results: (1) underscore isoflurane anesthesia. Glycerol-injected rats anesthetized with that differing anesthetics can profoundly alter the expression of a virtually nondefluorinated inhalational anesthetic (desflu- experimental renal injury; (2) raise the intriguing possibility rane) or with a nonfluorinated anesthetic (pentobarbital) served that isoflurane could potentially protect surgical/trauma pa- as controls. The severity of ARF was assessed 24 h later (blood tients from rhabdomyolysis-induced ARF; and (3) further sup- urea nitrogen, plasma creatinine [Cr], and renal histology). port the concept that renal fluoride exposure may confer prox- Anesthetic effects on extrarenal injury (plasma creatine phos- imal tubular cytoprotective effects. (J Am Soc Nephrol 9: phokinase, lactate dehydrogenase, and hematocrit levels), 305-309, 1998)

The early postoperative period is one of the most common sia, far bess injury resulted in the former group (2). Although clinical settings for the onset of acute renal failure (ARF). the explanation for this observation has remained elusive, it Indeed, some degree of renal insufficiency, or severe ARF, has nevertheless graphically demonstrates that even structurally been reported in 0. 1 to 30% of patients after various forms of related anesthetics can have profoundly different effects on the surgery (1). It traditionally has been assumed that perioperative expression of superimposed acute renal damage. hypotension and/or nephrotoxin-induced acute tubular necrosis Volatile fluorinated anesthetics (e.g., isoflurane, desfiurane, are the dominant causes of postoperative ARF. However, the sevoflurane, and enflurane), rather than barbiturates, are the high degree of variability in the incidence of postoperative current mainstays of clinical anesthetic practice (3). Although renal failure suggests that multiple factors, in addition to hy- structurally similar (all fluorinated methylethyl ethers), they potension and nephrotoxin exposure, may be involved. have differing degrees and sites of metabolism (3-5). One of An issue of considerable clinical relevance is whether the the major differences is the extent to which they are defluori- type of anesthesia used during surgery can modulate the ex- nated via the cytochrome P450 system (4,5). Methoxyflurane, pression of superimposed acute renal tubular damage. Obser- a previously used fluorinated anesthetic, undergoes marked vations from Finn almost 20 years ago provide experimental hepatic and renal defluorination, resulting in the accumulation support for this possibility. They found that when rats were of purportedly nephrotoxic inorganic fluoride levels (>50 M) subjected to 60 mm of renal artery occlusion under mactin (a (6-8). Indeed, this has been widely implicated as the expla- thiobarbiturate) or pentobarbital (an oxybarbiturate) anesthe- nation for an outbreak of ARF cases induced by methoxyflu- rane during its widespread use in the 1960s (culminating in its elimination from clinical practice). Isoflurane, now the most Received June 23, 1997. Accepted August 21. 1997. commonly used inhalationab anesthetic, undergoes signifi- Correspondence to Dr. Richard A. Zager, Fred Hutchinson Cancer Research cantly less intrahepatic, and presumably intrarenal, defluorina- Center, Room M621, 1 124 Columbia Street. Seattle, WA 98104. tion than methoxyflurane (e.g., approximately 10 iM plasma 1046-6673/0902-0305$03.00/0 Journal of the American Society of Nephrology fluoride levels in rats) (9,10). This may explain why it has little Copyright 0 1998 by the American Society of Nephrology or no nephrotoxic effect. In contrast, desfiurane, a relatively 306 Journal of the American Society of Nephrology new fluorinated anesthetic, is virtually completely resistant to sion, and exsanguination was performed by aortic puncture followed cytochrome P450-mediated defluorination reactions (3,10). by kidney resection. Blood urea nitrogen and plasma creatinine con- A recent in vitro study from this laboratory indicates that centrations were determined by autoanalyzer technology. Kidneys although high doses of inorganic fluoride can induce proximal (from S to 8 rats per group, randomly selected) were bisected and fixed with I 0% buffered formalin. Four-micrometer paraffin-embed- tubular cell necrosis, in low concentrations it can exert a ded sections were then stained with hematoxylin and eosin for histo- profound cytoprotective effect (1 1). For example, when cub- logic assessments, as described below. tured human proximal tubular (HK-2) cells were exposed to subtoxic inorganic fluoride challenges, a marked increase in Assessment ofAnesthetic Effects on the Extent of resistance to concomitant nephrotoxic (myoglobin)- and ATP Extrarenal Tissue injury depletion- mediated tubular cell death resulted (an approximate The following experiment was undertaken to test for the possibility 50% decrease in cell death). This observation raises the in- that differing anesthetic effects on the severity of glycerol-induced triguing possibility that the use of fluorinated anesthetics that ARF were due to differences in the extent of glycerol-induced: (1) release subtoxic levels of inorganic fluoride might confer an in muscle necrosis; (2) hemolysis: (3) intrarenal heme protein loading; vivo cytoprotective effect. (4) hemoconcentration (due to fluid “third” spacing into injured The purpose of the present study was to gain experimental muscle); or (5) differences in BP during the induction phase of this support for this hypothesis. To this end, rats were subjected to ARF model. A total of 12 rats was injected with glycerol under intramuscular glycerol injection, which induces a combined isoflurane, desfiurane, or pentobarbital anesthesia, as described above nephrotoxic (myohemoglobinuric)/ATP depletion form of tu- (n 4 per group). The animals that received isoflurane or desfiurane bular cell death ( I 2). Because we previously found that inor- had their systolic BP monitored during the 4-h anesthesia period via tail BP cuff technology ( I 3). [Note: This method is not suitable for ganic fluoride exposure attenuates each of these forms of gauging systemic BP under pentobarbital anesthesia ( 14, 15). Hence. tubular cell injury in vitro (I 1 ), we postulated that isoflurane in pilot experiments, we demonstrated that BP, assessed by direct anesthesia, which produces subtoxic renal fluoride accumuba- intracarotid artery monitoring (16), is stable under pentobarbital an- tion (9,10), might limit the severity of glycerol-induced ARF. esthesia ± glycerol injection.] After completing the 4-h anesthesia As controls, rats were subjected to glycerol injection under protocols. the rats were killed by aortic puncture. and plasma was either desfiurane or pentobarbitab anesthesia. The rationale for obtained for quantification of: ( I) creatine phosphokinase (CPK: as an this was that: (I) desflurane would provide a virtual noninor- index of muscle necrosis; Sigma, 520C. St. Louis, MO); (2) lactate ganic fluoride-generating fluorinated anesthetic control (3,10); dehydrogenase (LDH: as a combined index of muscle necrosis/hemo- and (2) the use of pentobarbital anesthesia would allow for the lysis); and (3) hematocrit (as an index of hemoconcentration, recog- assessment of whether fluorinated anesthetics in general (i.e., nizing that a degree of hemolysis complicates interpretation of this independent of defluorination reactions) alter the expression of assessment). The kidneys were resected, and tissue sections were prepared as noted above and then used to microscopically assess the the glycerol model of ARF (by comparing the desflurane results extent of intraluminal heme pigment loading/cast formation during the with those obtained with pentobarbital, a nonfluorinated agent). induction phase of ARF.

Materials and Methods Statistical Anah’ses Glycerol Model of Mvohemoglobinuric ARF Group values are given as mean ± 1 SEM. Statistical comparisons were made by ANOVA with after testing by unpaired t test with Male Sprague Dawley rats ( I 75 to 200 g: B &K Universal, Kent, Bonferroni correction. The extent of morphologic injury at 24 h WA) maintained under standard laboratory conditions with free food postglycerol injection was gauged semiquantitatively and in a blinded and water access were used for all experiments. After being placed in manner (by Dr. Zager) by assessing the approximate percentage of an airtight 22-L chamber, they were anesthetized for 15 mm with cortical proximal tubular segments that manifested necrosis, as fol- either isoflurane (n = 15) or desfiurane (,z = 8). These agents were lows: 0, none; 1+, l0%; 2+, 10 to 25%; 3+, 25 to 50%; 4+, delivered in 100% oxygen at a flow rate of 2 L/min using agent- 50%. Histologic comparisons were made by Wilcoxon rank sum test specific vaporizers. The vaporizers were set to maintain chamber with Bonferroni correction. isoflurane and desfiurane concentrations (monitored by infrared anal- ysis) at I .4 and 5.7%, respectively (equivalent anesthetic dosages) (3). A comparable level of anesthesia (defined as a complete inhibition of Results and Discussion spontaneous, nonrespiratory body movements) was induced in addi- Rats subjected to glycerol injection under pentobarbital de- tional rats (ii = I 2) by intramuscular pentobarbital injection (75 mg/ veloped severe ARF, as denoted by marked blood urea nitro- kg). Fifteen minutes after inducing anesthesia, each rat was subjected gen and serum creatinine increments (Figure 1 ). This result to intramuscular glycerol injection (50% solution, 9 ml/kg), adminis- was not simply due to a uniquely adverse effect of pentobar- tered in equally divided doses into each upper hind limb. Anesthesia bital on the glycerol model, because even worse azotemia was then maintained for an additional 3 3/4 h (the pentobarbital- resulted when the rats were challenged with glycerol under treated rats requiring a maintenance dose of 50 mg/kg at the 2-h time desfiurane anesthesia (Figure 1 ). In contrast, rats subjected to point). Body temperature. assessed with a rectal probe. was main- glycerol injection under isoflurane anesthesia manifested a tamed at 36 to 37#{176}Cusing an adjustable heating surface. After completing anesthesia, the rats were placed into standard cages per- marked attenuation of glycerol-induced ARF (approximately mitting rapid recovery from anesthesia. Free food and water access two-thirds to three-fourths reduction in the degree of azotemia, were provided. Twenty-four h later, the rats were reanesthetized (75 compared with the other anesthetic groups). Renal histologic mg/kg pentobarbital), the abdomens were opened via a midline mci- assessments at 24 h indicated that isoflurane use was associated Anesthetic Effects on Acute Renal Failure 307

<0.0001 <0.0001 200 1- 4.5 <0.0001 <0.035 II p 4 I ‘175 <0.001 3.5 j)150 ‘- 3 - 125 . 2.5 z 100 2 jl.5

L)0.5 0 Iso Des Pent Iso Des Pent

Figure 1. Blood urea nitrogen (BUN; left panel) and plasma creatinine (right panel) values 24 h after anesthesia under isoflurane (Iso: n = 15), desfiurane (Des; ,, = 8), or pentobarbital (Pent; n = I 2) anesthesia. Whereas desfiurane and pentobarbital anesthesia were associated with severe acute renal failure (ARF), glycerol injection under isoflurane anesthesia induced only mild ARF. (For comparison. normal BUN and creatinine values for rats, as obtained previously in this laboratory, are 10 to 20 mg/dl and 0.3 to 0.5 mg/dl, respectively). The attenuation of glycerol-induced injury under isoflurane or pentobarbital anesthesia was not simply due to the use of an inhalational or a noninhalational agent, because desfiurane (an inhalational agent) was associated with slightly worse ARF than pentobarbital anesthesia (P < 0.05).

with morphologic, and not simply renal functional, protection. induction phase of injury, because CPK, LDH, and hematocrit Whereas glycerol injection induced severe and equivalent renal values at this time did not statistically differ among the three morphologic injury (12) under conditions of pentobarbital and experimental groups (Figure 4). Given the similar CPK and desfiurane anesthesia (heme cast formation, extensive proxi- LDH increments, it was not surprising that no discernible mal tubule necrosis; 24-h assessments), the isoflurane/glycerol differences in the degree of intraluminal heme accumulation group sustained only mild morphologic damage (Figure 2 and (cast formation) were apparent among the three experimental Figure 3, A through C). The isoflurane-associated decrease in groups at 4 h postglycerol injection (Figure 3D). Finally, renal injury could not simply be attributed to a reduction in systolic BP after glycerol injection were highly comparable for muscle necrosis, hemolysis, or hemoconcentration during the the desfiurane (95 ± 2 mmHg) and isoflurane (91 ± 4 mmHg; NS) groups, consistent with comparable systemic hemody- namic profiles. (Nevertheless, this does not rule out the possi- <0.01 bility that differences in intrarenal hemodynamics could have

4. U. U #{149}#{149} existed among the experimental groups. potentially affecting the induction of renal damage.) 3. Although it is virtually impossible to assign direct “cause- U. U . .. and-effect” relationships in whole animal experiments, the C present results are at beast consistent with the hypothesis that

(I) 2- U isoflurane anesthesia, via its release of inorganic fluoride, can elicit a direct proximal tubular cytoprotective effect. In this z regard, it is noteworthy that fluoride affects at least three 1- U.... pathways that we have previously documented to be important mediators of myoglobin cytotoxicity. First, it exerts a ouabain- 0- U. like effect on proximal tubule Na,K-ATPase activity ( 17), an -I action that can directly attenuate the cytotoxic effects of myo- globin (1 8). Second, it can alter mitochondrial electron trans- Iso Des Pent port (17). In this regard, it is noteworthy that both rotenone and antimycin A, mitochondrial respiratory chain inhibitors, can Figure 2. Acute tubular necrosis (ATN) scores 24 h after glycerol mitigate myoglobin cytotoxicity in a cell culture system. pre- injection. Whereas glycerol induced extensive ATN under pentobar- sumabby by decreasing mitochondrial-driven oxidant stress bital (Pent) or desfiurane (Des) anesthesia (3 to 4+ scores, or >25 to ( 1 8). Third, inorganic fluoride can acutely deplete proximal 50% necrotic tubule segments; see text), only modest ATN (0 to 2+ scores, or 0 to 25% necrotic tubules) resulted with isoflurane anes- tubule cytosolic phospholipase A2 activity (1 1), an action that thesia (P < 0.01). (Eight, five, and six randomly selected kidneys can induce or contribute to a cytoprotected state (I 1). Thus, from the isoflurane, desfiurane, and pentobarbital groups. respec- although it remains to be proven that the observed differences tively. were prepared for histologic examination and scoring.) in glycerol-induced renal injury were mediated by inorganic 308 Journal of the American Society of Nephrology

Figure 3. (A through C) Representative renal histology 24 h after glycerol injection. Minimal tubular necrosis is apparent in renal cortex from an isoflurane/glycerol-treated rat (A) compared with a pentobarbital (B)- or desfiurane (C)-treated rat. (D) Intraluminal heme accumulation 4 h after glycerol injection. Marked heme accumulation is apparent. the extent of which did not noticeably differ among the three experimental groups (isoflurane depicted).

NS NS NS 400 50 1500

1250 =‘ 40 E E300 .. ‘U 1000 30 = 200

,., 20 500 :

250 10

0 0 0 Iso Des Pent Iso Des Pent Iso Des Pent Figure 4. Plasma creatine phosphokinase (CPK), lactate dehydrogenase (LDH), and hematocrit levels 4 h after glycerol injection for the three experimental groups. No significant differences were observed. Dotted lines represent normal mean values (CPK, 70 ± 22: LDH, 21 ± 5: hematocrit, 40 ± 1%). Abbreviations as in Figure 1.

fluoride, the available data are at least consistent with this fluoride levels were needed to produce cytoresistance in our hypothesis. previous cell culture experiments (1 1), it might appear that the It is noteworthy that isoflurane anesthesia in rats generates inorganic fluoride levels generated from isoflurane would be approximately 5- to 10-pM plasma fluoride bevels (9, 10) (our far too low to induce an analogous cytoresistant state. How- unpublished observations). Because low millimolar inorganic ever, two points must be noted in this regard. (1) An approx- Anesthetic Effects on Acute Renal Failure 309 imate I 000: 1 extracelbular:intracellular inorganic fluoride gra- 4. Kharasch ED, Hankins DC. Thummel KE: Human kidney me- dient exists; thus, low millimolar inorganic fluoride additions thoxyflurane and sevoflurane metabolism: lntrarenal fluoride to cell culture likely produce only bow micromolar intracellular production as a possible mechanism of methoxyflurane nephro- fluoride concentrations (1 1,17). (2) Isoflurane is highly mem- toxicity. Anesthesiology 82: 689-699. 1995 brane-permeable. allowing for high micromolar/low millimolar 5. Kharasch ED. Thummel KE: Identification of cytochrome P450 intracellular concentrations to be achieved ( 17). This presum- 2El as the predominant enzyme catalyzing human liver micro- somal defluorination of sevoflurane, isofiurane, and methoxyflu- ably allows for micromolar (i.e. , cytoprotective) intracellular rane. Anesthesiology 79: 795- 807, 1993 fluoride levels to be achieved as a result of intracellular P450 6. Crandall WB, Pappas SG, McDonald A: Nephrotoxicity associ- metabolism. Given these considerations, it is apparent that ated with methoxyflurane anesthesia. Anesthesiology 27 : 591- intracellular fluoride generation, rather than tubular fluoride 607, 1966 uptake from plasma or urine, is presumably the critical deter- 7. Mazze RI, Trudell JR. Cousins MI: Methoxyflurane metabolism minant of the cytoprotective effects of fluoride. and renal dysfunction. Anesthesiology 35: 247-252. 1971 In conclusion, the results of the present studies indicate that 8. Hollenberg NK, McDonald FR, Cotran R, Galvanek EG, Wahol isoflurane, an agent that releases subtoxic doses of inorganic M, Vandam LD, Merrill JP: Irreversible acute oliguric renal fluoride, can markedly attenuate the expression of myohemo- failure: A complication of methoxyflurane anesthesia. N Engl globinuric ARF, compared with a minimally defluorinated J Med 286: 877-879. 1972 (desflurane) or a nonfluorinated (pentobarbital) anesthetic. 9. Rice SA, Fish KJ: Anesthetic metabolism and renal function in This finding has a number of potential implications. (1) It obese and non obese Fischer 344 rats following enfiurane or underscores that the type of anesthetic used in studies of isoflurane anesthesia. Anesthesiology 65: 26-34. 1986 experimental ARF can have a profound impact on the results 10. Koblin DD, Eger El II. Johnson BH. Konopka K. Waskebb L: obtained. This makes it mandatory to scrutinize anesthetic 1-653 resists degradation in rats. Anestli Analg 67: 534-538. regimens when interpreting and comparing results in the ex- I988 perimental literature. (2) It raises the intriguing possibility that II . Zager RA, Iwata M: Inorganic fluoride: Divergent effects on isoflurane could potentially decrease the risk of ARF in se- human proximal tubular cell viability. Ant J Pat/zol 150: 735- lected surgical patients (e.g. . those with rhabdomyolysis), at 745, 1997 least relative to other anesthetics. (3) It suggests that our I 2. Zager RA: Rhabdomyolysis and myohemoglobinuric acute renal previous in vitro observations of inorganic fluoride-induced failure [Editorial Review]. Kidney hit 49: 314-326, 1996 proximal tubule cytoprotection might, in fact, have a clinically 13. Pfeffer JM, Pfeffer MA, Frohlich ED: Validity of an indirect tail relevant in vivo correlate. We believe that each of these issues cuff method for determining systolic arterial pressure in normo- provides a compelling rationale for further exploration of flu- tensive and spontaneously hypertensive rats. J Lab Cliii Med 78: 957-962. 1971 oride effects on proximal tubular homeostasis. 14. Seyde WC, McGowan L. Lund N, Duling B. Longnecker DE: Acknowledgments Effects of anesthetics on regional hemodynamics in normov- olemic and hemorrhaged rats. Am J Phvsiol 249: H164-H173, The authors thank Ms. K. M. Burkhart for her expert technical I 985 assistance. This work was supported by grants from National Insti- 15. Johnson R, Fowler IF, Zanebli GD: Changes in mouse blood tutes of Health (DK 38432, DK 0772 1 . and GM 487 12) and the pressure. tumor blood flow, and core and tumor temperatures Northwest Kidney Foundation. following nembutal or urethane anesthesia. Radiology I I 8: 697- 703, 1976 References 16. Zager RA: Myogbobin depletes renal adenine nucleotide pools in 1. Novis BK, Roizen MF, Aronson 5, Thisted RA: Association of the presence and absence of shock. Kidney In! 39: 1 1 1- I I 9, 1991 preoperative risk factors with postoperative acute renal failure. 17. Lochhead KM. Kharasch ED. Zager RA: Spectrum and subcel- Anesth Analg 78: 143-149, 1992 2. Finn WF: Nephron heterogeneity in polyuric acute renal failure. lular determinants of fluorinated anesthetic mediated proximal J Lab Cliii Med 98: 21-29, 1981 tubular injury. Am J Pathol 150: 2209-2221, 1997 3. Marshall BE, Longnecker DE: General anesthetics. In: Goodman 18. Zager RA, Burkhart K: Myogbobin toxicity in proximal human and Gilman ‘s The Pharmacologic Basis of Therapeutics, 9th Ed.. kidney cells: Roles of Fe2, Ca2. H2O2, and terminal mito- New York, McGraw-Hill, 1996, pp 307-330 chondrial electron transport. Kidney mt 5 1: 728 -738. 1997